Dustin & Baldari, 2017 - The Immune Synapse PDF

Methods in
Molecular Biology 1584
Cosima T. Baldari
Michael L. Dustin Editors
The Immune
Synapse
Methods and Protocols
Methods in Molecular Biology
Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, AL10 9AB, UK
For further volumes:

http://www.springer.com/series/7651
The Immune Synapse
Methods and Protocols
Edited by
Cosima T. Baldari
Department of Life sciences, University of Siena, Siena, Siena, Italy
Michael L. Dustin
University of Oxford, Kennedy Institute of Rheumatology, Headington, Oxford, UK
Editors
Cosima T. Baldari Michael L. Dustin
Department of Life sciences University of Oxford, Kennedy Institute
University of Siena of Rheumatology
Siena, Siena, Italy Headington, Oxford, UK
ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-4939-6879-4 ISBN 978-1-4939-6881-7 (eBook)
DOI 10.1007/978-1-4939-6881-7
Library of Congress Control Number: 2017931687
© Springer Science+Business Media LLC 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been made.
The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Humana Press imprint is published by Springer Nature

The registered company is Springer Science+Business Media LLC
The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.
Preface
Initiation of the T cell-mediated adaptive immune response to pathogens is crucially depen-

dent on the assembly of a highly specialized signaling platform that forms at the interface
of a T cell and an antigen-presenting cell (APC) bearing specific peptide antigen associated
with major histocompatibility molecules, known as the immune synapse. From its initial
description as a membrane domain characterized by the segregation in concentric subdo-
mains of specific receptors that is accompanied by the polarization of the microtubule-
organizing center towards the APC contact area, our understanding of the structure,
dynamics, and function of the immune synapse has rapidly evolved. It is now clear that the
mature bull’s eye synapse marks the final phase of an extremely dynamic process where
microclusters of receptors and signaling mediators converge as they signal towards the cen-
ter of the IS, where they are either internalized to be targeted for degradation or released
as microvesicles to convey information and instructions to the APC. Vesicular traffic has
emerged as a central player in ensuring not only polarized delivery of cytokines and enzymes
to target cells by T cell effectors but also sustained signaling at the immune synapse and
modulation of the APC during naive T cell activation. Moreover the T cell immune synapse
has recently emerged as a paradigm for a variety of immune cell interactions that include
synapses formed by B cells, NK, and mast cells. The remarkable progress in this rapidly
moving area has required the development of powerful techniques and tools of analysis,
ranging from super-resolution microscopy and electron tomography, to the generation of
highly specific micropatterned surfaces for studying the dynamics of microclusters and sin-
gle molecules, to a variety of molecular probes to image signaling dynamics, to the imaging
of immune cell interactions in vivo, to robust computational methods to address the spa-
tiotemporal complexity of the immune synapse. This book has collected all the essential
protocols that are currently used to study the immune synapse, addressing (1) methods for
the study of the dynamics of immune synapse assembly; (2) methods for the study of
vesicular traffic at the immune synapse; (3) new high resolution imaging, biophysical, and
computational methods for the study of the immune synapse; (4) methods for the study of
effector immune synapses; (5) methods for the study of B cell, NK, and mast cell immune
synapses; and (6) methods for the study of immune interactions in vivo. This timely and
exhaustive collection of protocols is expected to be of interest to immunologists and, at a
more general level, to cell biologists, biophysicists, and computational biologists.
Siena, Italy Cosima T. Baldari

Headington, Oxford, UK Michael L. Dustin
v
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
1 The Immune Synapse: Past, Present, and Future . . . . . . . . . . . . . . . . . . . . . . . 1

Michael L. Dustin and Cosima T. Baldari
2 Analyzing Actin Dynamics at the Immunological Synapse . . . . . . . . . . . . . . . . . 7
Katarzyna I. Jankowska and Janis K. Burkhardt
3 Analysis of Microtubules and Microtubule-Organizing Center
at the Immune Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Noelia Blas-Rus, Eugenio Bustos-Morán, Francisco Sánchez-Madrid,
and Noa B. Martín-Cófreces
4 Analyzing the Dynamics of Signaling Microclusters . . . . . . . . . . . . . . . . . . . . . 51
Akiko Hashimoto-Tane, Tadashi Yokosuka, and Takashi Saito
5 Reconstitution of TCR Signaling Using Supported Lipid Bilayers . . . . . . . . . . . 65
Xiaolei Su, Jonathon A. Ditlev, Michael K. Rosen, and Ronald D. Vale
6 Plasma Membrane Sheets for Studies of B Cell Antigen Internalization
from Immune Synapses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Carla R. Nowosad and Pavel Tolar
7 Studying the Dynamics of TCR Internalization at the Immune Synapse . . . . . . 89
Enrique Calleja, Balbino Alarcón, and Clara L. Oeste
8 T Cell Receptor Activation of NF-κB in Effector T Cells: Visualizing
Signaling Events Within and Beyond the Cytoplasmic Domain
of the Immunological Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Maria K. Traver, Suman Paul, and Brian C. Schaefer
9 Imaging Vesicular Traffic at the Immune Synapse . . . . . . . . . . . . . . . . . . . . . . . 129
Jérôme Bouchet, Iratxe del Río-Iñiguez, and Andrés Alcover
10 Analysis of TCR/CD3 Recycling at the Immune Synapse . . . . . . . . . . . . . . . . . 143
Laura Patrussi and Cosima T. Baldari
11 Simultaneous Membrane Capacitance Measurements and TIRF
Microscopy to Study Granule Trafficking at Immune Synapses . . . . . . . . . . . . . 157
Marwa Sleiman, David R. Stevens, and Jens Rettig
12 Mathematical Modeling of Synaptic Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Anastasios Siokis, Philippe A. Robert, and Michael Meyer-Hermann
13 Super-resolution Analysis of TCR-Dependent Signaling: Single-Molecule
Localization Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Valarie A. Barr, Jason Yi, and Lawrence E. Samelson
14 Förster Resonance Energy Transfer to Study TCR-pMHC Interactions
in the Immunological Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Gerhard J. Schütz and Johannes B. Huppa
vii
viii Contents
15 Two-Dimensional Analysis of Cross-Junctional Molecular Interaction

by Force Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Lining Ju, Yunfeng Chen, Muaz Nik Rushdi, Wei Chen,
and Cheng Zhu
16 Studying Dynamic Plasma Membrane Binding of TCR-CD3 Chains
During Immunological Synapse Formation Using Donor-Quenching
FRET and FLIM-FRET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Etienne Gagnon, Audrey Connolly, Jessica Dobbins,
and Kai W. Wucherpfennig
17 Revealing the Role of Microscale Architecture in Immune Synapse
Function Through Surface Micropatterning . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Joung-Hyun Lee and Lance C. Kam
18 Spatial Control of Biological Ligands on Surfaces Applied
to T Cell Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Haogang Cai, David Depoil, James Muller, Michael P. Sheetz,
Michael L. Dustin, and Shalom J. Wind
19 Probing Synaptic Biomechanics Using Micropillar Arrays . . . . . . . . . . . . . . . . . 333
Weiyang Jin, Charles T. Black, Lance C. Kam, and Morgan Huse
20 Microchannels for the Study of T Cell Immunological Synapses and Kinapses . . . 347
Hélène D. Moreau, Philippe Bousso, and Ana-Maria Lennon-Duménil
21 Purification of LAT-Containing Membranes from Resting
and Activated T Lymphocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Claire Hivroz, Paola Larghi, Mabel Jouve, and Laurence Ardouin
22 Quantitative Phosphoproteomic Analysis of T-Cell Receptor Signaling . . . . . . . 369
Nagib Ahsan and Arthur R. Salomon
23 Imaging Asymmetric T Cell Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
Mirren Charnley and Sarah M. Russell
24 Ultrastructure of Immune Synapses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Jaime Llodrá
25 Systems Imaging of the Immune Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
Rachel Ambler, Xiangtao Ruan, Robert F. Murphy,
and Christoph Wülfing
26 Comprehensive Analysis of Immunological Synapse Phenotypes
Using Supported Lipid Bilayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Salvatore Valvo, Viveka Mayya, Elena Seraia, Jehan Afrose,
Hila Novak-Kotzer, Daniel Ebner, and Michael L. Dustin
27 Studying Immunoreceptor Signaling in Human T Cells
Using Electroporation of In Vitro Transcribed mRNA . . . . . . . . . . . . . . . . . . . 443
Omkar Kawalekar, Carl H. June, and Michael C. Milone
28 A Protein Expression Toolkit for Studying Signaling in T Cells . . . . . . . . . . . . . 451
Ana Mafalda Santos, Jiandong Huo, Deborah Hatherley, Mami Chirifu,
and Simon J. Davis
29 Imaging the Effector CD8 Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
Gordon L. Frazer, Yukako Asano, and Gillian M. Griffiths
Contents ix
30 The Mast Cell Antibody-Dependent Degranulatory Synapse . . . . . . . . . . . . . . 487

Salvatore Valitutti, Régis Joulia, and Eric Espinosa
31 Measurement of Lytic Granule Convergence After Formation
of an NK Cell Immunological Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
Hsiang-Ting Hsu, Alexandre F. Carisey, and Jordan S. Orange
32 Studying the T Cell-Astrocyte Immune Synapse . . . . . . . . . . . . . . . . . . . . . . . . 517
George P. Cribaro, Elena Saavedra-López, Paola V. Casanova,
Laura Rodríguez, and Carlos Barcia
33 Aberrant Immunological Synapses Driven by Leukemic
Antigen-Presenting Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
Fabienne McClanahan Lucas and John G. Gribben
34 Studying the Immune Synapse in HIV-1 Infection . . . . . . . . . . . . . . . . . . . . . . 545
Iratxe del Río-Iñiguez, Jérôme Bouchet, and Andrés Alcover
35 In Vivo Imaging of T Cell Immunological Synapses and Kinapses
in Lymph Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
Hélène D. Moreau and Philippe Bousso
36 Studying Dendritic Cell-T Cell Interactions Under In Vivo Conditions . . . . . . 569
Nicholas van Panhuys
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585
Contributors
Jehan Afrose • University of Oxford, Oxford, UK

Nagib Ahsan • Brown University and Rhode Island Hospital, Providence, USA
Balbino Alarcón • Universidad Autónoma de Madrid, Madrid, Spain
Andrés Alcover • Institut Pasteur, Paris, France
Rachel Ambler • University of Bristol, Bristol, UK
Laurence Ardouin • Institut Curie, Paris, France
Yukako Asano • Cambridge Institute for Medical Research, Cambridge, UK
Cosima T. Baldari • University of Siena, Siena, Italy
Carlos Barcia • Universidad Autónoma de Barcelona, Barcelona, Spain
Valerie A. Barr • National Cancer Institute, Bethesda, USA
Charles T. Black • Brookhaven National Laboratory, New York, USA
Noelia Blas-Rus • Universidad Autónoma de Madrid, Madrid, Spain
Jérôme Bouchet • Institut Pasteur, Paris, France
Philippe Bousso • Institut Pasteur, Paris, France
Janis K. Burkhardt • University of Pennsylvania, Philadelphia, USA
Eugenio Bustos-Morán • Centro Nacional Investigaciones Cardiovasculares (CNIC),
Madrid, Spain
Haogang Cai • Columbia University, New York, USA
Enrique Calleja • Universidad Autónoma de Madrid, Madrid, Spain
Alexandre F. Carisey • Texas Children’s Hospital and Baylor College of Medicine,
Houston, USA
Paola V. Casanova • Universidad Autónoma de Barcelona, Barcelona, Spain
Mirren Charnley • Swinburne University of Technology, Hawthorn, VIC, Australia;
Peter MacCallum Cancer Centre, East Melbourne, VIC, Australia
Yunfeng Chen • Institute of Technology, Atlanta, USA
Wei Chen • Zhejiang University, Hangzhou, Zhejiang, China
Mami Chirifu • University of Oxford, Oxford, UK
Audrey Connolly • University of Montreal, Montreal, Canada
George P. Cribaro • Universidad Autónoma de Barcelona, Barcelona, Spain
Simon J. Davis • University of Oxford, Oxford, UK
Iratxe del Río-Iñiguez • Institut Pasteur, Paris, France
David Depoil • University of Oxford, Oxford, UK
Jonathon A. Ditlev • Marine Biological Laboratory, Woods Hole, USA; University of Texas,
Texas, USA
Jessica Dobbins • Dana-Farber Cancer Institute and Harvard Medical School, Boston, USA
Michael L. Dustin • University of Oxford, Oxford, UK; New York University School
of Medicine, New York, USA
Daniel Ebner • University of Oxford, Oxford, UK
Eric Espinosa • University of Toulouse, Toulouse, France
Gordon L. Frazer • Cambridge Institute for Medical Research, Cambridge, UK
Etienne Gagnon • University of Montreal, Montreal, Canada
xi
xii Contributors
John G. Gribben • Queen Mary University of London, London, UK

Gillian M. Griffiths • Cambridge Institute for Medical Research, Cambridge, UK
Akiko Hashimoto-Tane • RIKEN Center for Integrative Medical Sciences, Yokohama, Japan
Deborah Hatherley • University of Oxford, Oxford, UK
Claire Hivroz • Institut Curie, Paris, France
Hsiang-Ting Hsu • Texas Children’s Hospital and Baylor College of Medicine, Houston, USA
Jiandong Huo • University of Oxford, Oxford, UK
Johannes B. Huppa • Technical University of Vienna, Vienna, Austria
Morgan Huse • Memorial Sloan-Kettering Cancer Center, New York, USA
Katarzyna I. Jankowska • University of Pennsylvania, Philadelphia, USA
Weiyang Jin • Columbia University, New York, USA
Régis Joulia • University of Toulouse, Toulouse, France
Mabel Jouve • Institut Curie, Paris, France
Lining Ju • University of Sydney, Sydney, Australia
Carl H. June • University of Pennsylvania, Philadelphia, PA, USA
Lance C. Kam • Columbia University, New York, USA
Omkar Kawalekar • University of Pennsylvania, Philadelphia, PA, USA
Paola Larghi • University of Milan, Milan, Italy; Istituto Nazionale Genetica Molecolare,
‘Romeo ed Enrica Invernizzi’, INGM, Milan, Italy
Joung-Hyun Lee • Columbia University, New York, USA
Ana-Maria Lennon-Duménil • Institut Curie, PSL Research University, Paris, France
Jaime Llodrá • University of Bern, Bern, Switzerland
Noa B. Martín-Cófreces • Universidad Autónoma de Madrid, Madrid, Spain
Viveka Mayya • University of Oxford, Oxford, UK
Fabienne McClanahan Lucas • Queen Mary University of London, London, UK; The
Ohio State University, Columbus, OH, USA
Michael Meyer-Hermann • Helmholtz Centre for Infection Research, Braunschweig,
Germany
Michael C. Milone • University of Pennsylvania, Philadelphia, PA, USA
Hélène D. Moreau • Institut Curie, PSL Research University, Paris, France; Institut
Pasteur, Paris, France
James Muller • New York University School of Medicine, New York, USA
Robert F. Murphy • Carnegie Mellon University, Pittsburgh, USA
Hila Novak-Kotzer • Kennedy Institute of Rheumatology, Nuffield Department of
Orthopedics, Rheumatology and Musculoskeletal Sciences, The University of Oxford,
Oxford, UK
Carla R. Nowosad • Francis Crick Institute, London, UK
Clara L. Oeste • Universidad Autónoma de Madrid, Madrid, Spain
Jordan S. Orange • Texas Children’s Hospital and Baylor College of Medicine, Houston, USA
Laura Patrussi • University of Siena, Siena, Italy
Suman Paul • Uniformed Services University, Bethesda, USA
Jens Rettig • Universität des Saarlandes, Homburg/Saar, Germany
Philippe A. Robert • Helmholtz Centre for Infection Research, Braunschweig, Germany;
Université Montpellier II, Montpellier, France
Laura Rodríguez • Universidad Autónoma de Barcelona, Barcelona, Spain
Michael K. Rosen • Marine Biological Laboratory, Woods Hole, USA; University of Texas,
Texas, USA
Xiangtao Ruan • Carnegie Mellon University, Pittsburgh, USA
Contributors xiii
Muaz Nik Rushdi • Institute of Technology, Atlanta, USA

Sarah M. Russell • Swinburne University of Technology, Hawthorn, VIC, Australia;
Peter MacCallum Cancer Centre, East Melbourne, VIC, Australia; University
of Melbourne, Parkville, VIC, Australia
Elena Saavedra-López • Universidad Autónoma de Barcelona, Barcelona, Spain
Takashi Saito • RIKEN Center for Integrative Medical Sciences, Yokohama, Japan
Arthur R. Salomon • Brown University and Rhode Island Hospital, Providence, USA
Lawrence E. Samelson • National Cancer Institute, Bethesda, USA
Francisco Sánchez-Madrid • Universidad Autónoma de Madrid, Madrid, Spain
Ana Mafalda Santos • University of Oxford, Oxford, UK
Brian C. Schaefer • Uniformed Services University, Bethesda, USA
Gerhard Schütz • TU Wien, Vienna, Austria
Elena Seraia • University of Oxford, Oxford, UK
Michael P. Sheetz • Columbia University, New York, USA; National University
of Singapore, Singapore, Singapore
Anastasios Siokis • Helmholtz Centre for Infection Research (HZI), Braunschweig, Germany
Marwa Sleiman • Universität des Saarlandes, Homburg/Saar, Germany
David R. Stevens • Universität des Saarlandes, Homburg/Saar, Germany
Xiaolei Su • Marine Biological Laboratory, Woods Hole, USA; University of California
San Francisco, San Francisco, USA
Pavel Tolar • Francis Crick Institute, London, UK; Imperial College London, London, UK
Maria K. Traver • Uniformed Services University and Henry M Jackson Foundation for
the Advancement of Military Medicine, Bethesda, USA
Ronald D. Vale • Marine Biological Laboratory, Woods Hole, USA; University of
California, San Francisco, USA
Salvatore Valitutti • University of Toulouse, Toulouse, France
Salvatore Valvo • University of Oxford, Oxford, UK
Nicholas van Panhuys • Sidra Medical and Research Center, Doha, Qatar
Shalom J. Wind • Columbia University, New York, USA
Kai W. Wucherpfennig • Dana-Farber Cancer Institute, Boston, USA; Harvard Medical
School, Boston, USA
Christoph Wülfing • University of Bristol, Bristol, UK
Jason Yi • Columbia University, New York, USA
Tadashi Yokosuka • RIKEN Center for Integrative Medical Sciences, Yokohama, Japan
Cheng Zhu • Georgia Institute of Technology, Atlanta, USA
Chapter 1
The Immune Synapse: Past, Present, and Future

Michael L. Dustin and Cosima T. Baldari
Abstract
Immunological synapses are specialized cell-cell junctions characterized by (1) close apposition of the
immune cell membrane with the membrane of another cell driven by adaptive or innate immune recogni-
tion, (2) adhesion, (3) stability, and (4) directed secretion. This phenomenon was first recognized in the
1970s and the early 1980s through electron microscopy of ex vivo functioning immune cells. Progressive
advances in fluorescence microscopy and molecular immunology in the past 20 years have led to rapid
progress on understanding the modes of cell-cell interaction and underlying molecular events. This vol-
ume contains a diverse range of protocols that can be applied to the study of the immunological synapses
and related immune cell junctions both in vitro and in vivo; and in disease settings in animal models and
humans. We have also included chapters on critical molecular tools such as protein expression and mRNA
electroporation that underpin or expand imaging approaches, although they are not specific to the study
of immune synapses. We hope that these chapters will be of use to people entering the field as well as sea-
soned practitioners looking to expand their repertoire of methods.
Key words Science history, Fluorescence, Affinity, Modeling, Microscopy
1 Introduction
Phagocytosis and antibodies were described within a decade of each

other at the turn of the nineteenth to twentieth century [1, 2], but
while the direct physical role of the phagocyte in phagocytosis was
immediately evident, it took another 60 years to recognize that
lymphocytes made antibodies [3]. The existence of two types of
lymphocytes and the need for their cooperation in antibody pro-
duction led to studies in the 1970s on the physical interaction of T
lymphocytes that accompanied T cell help and cytotoxicity [4]. The
role of macrophages and dendritic cells in the generation of T cell
help was discovered in this same period with the genetic evidence
for MHC restriction of T cell responses [5]. These studies initially
relied on electron microscopy to reveal the close membrane align-
ment between lymphocytes that was well described in solid organs
as representing adhesive junctions [6, 7]. The molecules that medi-
ated the adhesion were discovered through function blocking
Cosima T. Baldari and Michael L. Dustin (eds.), The Immune Synapse: Methods and Protocols, Methods in Molecular Biology,
vol. 1584, DOI 10.1007/978-1-4939-6881-7_1, © Springer Science+Business Media LLC 2017
1
2 Michael L. Dustin and Cosima T. Baldari
monoclonal antibodies [8]. The field’s recognition of directed

secretion as a critical process in T cell cytotoxicity and the role of
cytoplasmic Ca2+ evaluation in the signaling process was synthesized
by Norcross as suggesting a “synaptic basic for T cell activation”
[9]. Seder and Paul further elaborated on this, summarizing 10 years
of work on T cell help for B cells and the role of directed secretion
of cytokines with specific label of “immunological synapse” in a
commentary in Cell [10]. Within a year of this, Kupfer presented
his first images of the supramolecular activation clusters that were
revealed by the application of wide-field fluorescence microscopy
with deconvolution to conjugates of T cells with B cell tumors [11].
Parallel work to address the measurement of 2D affinity using sup-
ported planar bilayers (SLB) advanced to reconstitution of T cell
activation with convergence on the time-dependent evolution of
the same pattern, which was defined as a mature immunological
synapse [12]. A reviewer of a predeceding paper that sought to
define an immunological synapse raised the caveat that “immuno-
logical synapse” was too broad a term to apply to a structure formed
by T cells, as other immune cells might use similar strategies [13].
This was of course correct and we can now consider the mature
immunological synapse to be the result of a common strategy
applied by many immune cell types that use immunoreceptors
including mast cells, multiple types of T cells, NK cells, B cells, neu-
trophils, macrophages, and dendritic cells [14–19]. In the subse-
quent years, the field has continued to evolve with technology and
many features of immunological synapses have been discovered,
including imaging of interaction of T cells and antigen presenting
cells in situ and in vivo [20–22]. The immunological synapse pres-
ents an outstanding opportunity in basic cell biology as T cells can
be triggered by well-defined inputs to display multiple modes of
motility and polarization [23–25]. The immunological synapse is
disrupted in primary immunodeficiency diseases [26, 27] and auto-
reactive T cells form defective immunological synapses [28]. The
immunological synapse concept has guided studies leading to life-
saving therapies, particularly in cancer immunotherapy [29, 30].
There are still many questions remaining and this book is meant to
provide a current and forward-looking set of methods that will help
to address the next level of questions and allow further application
to improvement of human health.
2 Materials
This book is composed of 35 chapters (excluding this introductory

chapter) that present methods relevant to the characterization of
the immunological synapse. Some chapters present multiple proven
approaches to study a particular phenomenon within the immuno-
logical synapse or type of immunological synapse. Others present
The Immune Synapse: Past, Present, and Future 3
details of technical approaches that can be applied to the multiple

types of immunological synapses and other related biology. Yet
others present enabling technologies that are quite general in
applications in life sciences, such as methods for efficient expres-
sion of exogenous proteins in primary cells or recombinant pro-
teins expression and purification. While of broad utility, we invited
them here because they are key enabling technologies for future
studies on immune synapses.
As a matter of MIMB style, the authors have been discouraged
from providing detailed information on suppliers for common
materials due to concerns about regional differences in chemical
supply. However, we have broken with this style in some instances
to identify suppliers for what appear to be common items (for
example, microscope coverslips or glass bottom 96-well plates)
when the authors have taken great effort to screen many potential
suppliers of similar items and identified particular sources that out-
performed others in direct comparisons. These instances may be
further highlighted in the Notes section to describe the criteria for
selecting particular suppliers. This should be helpful in case any
reader has difficultly accessing particular suppliers in their regions.
The relevant screening criteria can be reapplied if necessary to find
a suitable alternative supplier. Furthermore, if individuals reading
these chapters run into problems with applying the protocols
included here, all of the authors are happy to be contacted by
email, included in the corresponding authors list, and will try their
best to provide additional guidance.
3 Methods
3.1 Elements Chapters 2–12 deal with methods to investigate particular subsys-
of Immunological tems that are likely to be applicable to any type of immunological
Synapses synapse. These include cytoskeleton, immunoreceptor microclus-
ters, receptor trafficking in vesicles, cytoplasmic signaling com-
plexes, and interfacial patterns. In some cases, the experimental
examples focus on Jurkat T cells, a common model system because
somatic variants lacking key signaling molecules are available and
they are readily transfectable to generate stable or transiently
expressing cell lines. But others provide examples with primary
cells. In one instance, the focus is on cell-free reconstitution of
signaling, which nicely complements in situ analysis of signaling
microclusters. This group also includes a chapter on mathematical
modeling of molecular patterns in the immunological synapse.
3.2 Technologies Chapters 13–27 focus on technologies that can be applied to the
study of any immunological synapse. These include single mole-
cule imaging and interaction measurements, fluorescence reso-
nance energy transfer (FRET), force measurement, micro and
4 Michael L. Dustin and Cosima T. Baldari
nano-fabrication methods, proteomics, asymmetric cell division,

electron tomography, and systematic imaging methods. In some
instances these modalities are combined as in single molecule
FRET and force measurements using micro- or nano-fabricated
surfaces. Different approaches to systematic analysis of immune
synapses are highlighted. A powerful method for gene expression
in primary cells based on mRNA electroporation is described.
Finally, a chapter is provided on methods for recombinant protein
expression in bacterial or mammalian cells that provide greatly
accelerated pathways to milligram amounts of proteins, which are
enabling for many of the reconstitution approaches.
3.3 Biological Chapters 28–36 focus on some compelling biological examples

Examples studied using a variety of cutting edge methods. These include
analysis of killer cells in action, neuro-immune synapses, and con-
sequences of pathological situations like cancer and infection for
immune synapses. The book ends with two protocols for in vivo
imaging of T cell-dendritic cell interactions in vivo, which is critical
for basic understanding and also to help guide the in vitro efforts
toward greater future relevance.
We are very excited to have these state-of-the-art methods,
most of which have already been featured in outstanding primary
publications, described in step by step detail in one volume. It is
our hope that this collection will accelerate the reproduction of key
results, prime new biological observations, and technical innova-
tions. Best wishes for success with your experimental and/or mod-
eling efforts.
Acknowledgments
We thank Éva Culleton-Oltay for assistance with editing the chap-

ters and all the authors for their hard work on this project.
References
1. Karnovsky ML (1981) Metchnikoff in Messina: components by sensitised T lymphocytes in
a century of studies on phagocytosis. N Engl lymphocytic choriomeningitis. Nature
J Med 304(19):1178–1180 251(5475):547–548
2. Llewelyn MB, Hawkins RE, Russell SJ (1992) 6. Lipsky PE, Rosenthal AS (1975) Macrophage-
Discovery of antibodies. BMJ 305(6864): lymphocyte interaction. II. Antigen-mediated
1269–1272 physical interactions between immune guinea
3. McGregor DD, Gowans JL (1963) The anti- pig lymph node lymphocytes and syngeneic
body response of rats depleted of lymphocytes macrophages. J Exp Med 141:138
by chronic drainage from the thoracic duct. 7. Geiger B, Rosen D, Berke G (1982) Spatial rela-
J Exp Med 117(2):303–320 tionships of microtubule-organizing centers and
4. Raff MC (1973) T and B lymphocytes and the contact area of cytotoxic T lymphocytes and
immune responses. Nature 242(5392):19–23 target cells. J Cell Biol 95(1):137–143
5. Zinkernagel RM, Doherty PC (1974) 8. Sanchez-Madrid F, Krensky AM, Ware CF,
Immunological surveillance against altered self Robbins E, Strominger JL, Burakoff SJ,
The Immune Synapse: Past, Present, and Future 5
Springer TA (1982) Three distinct antigens required for extensive, functional T cell con-
associated with human T-lymphocyte-mediated tacts. J Leukoc Biol. doi:10.1189/
cytolysis: LFA-1, LFA-2, and LFA-3. Proc Natl jlb.2A0215-050RR
Acad Sci U S A 79(23):7489–7493 20. Stoll S, Delon J, Brotz TM, Germain RN
9. Norcross MA (1984) A synaptic basis for (2002) Dynamic imaging of T cell-dendritic
T-lymphocyte activation. Ann Immunol cell interactions in lymph nodes. Science
135D(2):113–134 296(5574):1873–1876
10. Paul WE, Seder RA (1994) Lymphocyte 21. Bousso P, Robey E (2003) Dynamics of CD8+
responses and cytokines. Cell 76:241–251 T cell priming by dendritic cells in intact lymph
11. Monks CR, Freiberg BA, Kupfer H, Sciaky N, nodes. Nat Immunol 4(6):579–585
Kupfer A (1998) Three-dimensional segrega- 22. Azar GA, Lemaitre F, Robey EA, Bousso P
tion of supramolecular activation clusters in T (2010) Subcellular dynamics of T cell immu-
cells. Nature 395(6697):82–86 nological synapses and kinapses in lymph
12. Grakoui A, Bromley SK, Sumen C, Davis MM, nodes. Proc Natl Acad Sci U S A
Shaw AS, Allen PM, Dustin ML (1999) The 107(8):3675–3680
immunological synapse: a molecular machine 23. Jacobelli J, Bennett FC, Pandurangi P, Tooley
controlling T cell activation. Science AJ, Krummel MF (2009) Myosin-IIA and
285(5425):221–227 ICAM-1 regulate the interchange between two
13. Dustin ML, Olszowy MW, Holdorf AD, Li J, distinct modes of T cell migration. J Immunol
Bromley S, Desai N, Widder P, Rosenberger F, 182(4):2041–2050
van der Merwe PA, Allen PM, Shaw AS (1998) 24. Finetti F, Paccani SR, Riparbelli MG,
A novel adapter protein orchestrates receptor Giacomello E, Perinetti G, Pazour GJ,
patterning and cytoskeletal polarity in T cell Rosenbaum JL, Baldari CT (2009) Intraflagellar
contacts. Cell 94:667–677 transport is required for polarized recycling of
14. Davis DM, Chiu I, Fassett M, Cohen GB, the TCR/CD3 complex to the immune syn-
Mandelboim O, Strominger JL (1999) The apse. Nat Cell Biol 11(11):1332–1339
human natural killer cell immune synapse. Proc 25. Burkhardt JK, Carrizosa E, Shaffer MH (2008)
Natl Acad Sci U S A 96(26):15062–15067 The actin cytoskeleton in T cell activation.
15. Batista FD, Iber D, Neuberger MS (2001) B Annu Rev Immunol 26:233–259
cells acquire antigen from target cells after syn- 26. Clark R, Griffiths GM (2003) Lytic granules,
apse formation. Nature 411(6836):489–494 secretory lysosomes and disease. Curr Opin
16. Stinchcombe JC, Bossi G, Booth S, Griffiths Immunol 15(5):516–521
GM (2001) The immunological synapse of 27. Dupre L, Aiuti A, Trifari S, Martino S, Saracco
CTL contains a secretory domain and mem- P, Bordignon C, Roncarolo MG (2002)
brane bridges. Immunity 15(5):751–761 Wiskott-Aldrich syndrome protein regulates
17. Carroll-Portillo A, Spendier K, Pfeiffer J, lipid raft dynamics during immunological syn-
Griffiths G, Li H, Lidke KA, Oliver JM, Lidke apse formation. Immunity 17(2):157–166
DS, Thomas JL, Wilson BS, Timlin JA (2010) 28. Schubert DA, Gordo S, Sabatino JJ Jr,
Formation of a mast cell synapse: Fc epsilon RI Vardhana S, Gagnon E, Sethi DK, Seth NP,
membrane dynamics upon binding mobile or Choudhuri K, Reijonen H, Nepom GT,
immobilized ligands on surfaces. J Immunol Evavold BD, Dustin ML, Wucherpfennig KW
184(3):1328–1338 (2012) Self-reactive human CD4 T cell clones
18. Goodridge HS, Reyes CN, Becker CA, form unusual immunological synapses. J Exp
Katsumoto TR, Ma J, Wolf AJ, Bose N, Chan Med 209(2):335–352
AS, Magee AS, Danielson ME, Weiss A, 29. Egen JG, Allison JP (2002) Cytotoxic T lym-
Vasilakos JP, Underhill DM (2011) Activation phocyte antigen-4 accumulation in the
of the innate immune receptor Dectin-1 upon immunological synapse is regulated by TCR

formation of a 'phagocytic synapse'. Nature signal strength. Immunity 16(1):23–35
472(7344):471–475 30. Pentcheva-Hoang T, Chen L, Pardoll DM,
19. Malinova D, Fritzsche M, Nowosad CR, Armer Allison JP (2007) Programmed death-1 con-
H, Munro PM, Blundell MP, Charras G, Tolar centration at the immunological synapse is
P, Bouma G, Thrasher AJ (2015) WASp- determined by ligand affinity and availability.
dependent actin cytoskeleton stability at the Proc Natl Acad Sci U S A 104(45):
dendritic cell immunological synapse is 17765–17770
Chapter 2
Analyzing Actin Dynamics at the Immunological Synapse

Abstract
T cell signaling is inextricably linked to actin cytoskeletal dynamics at the immunological synapse (IS). This
process can be imaged in living T cells expressing GFP actin or fluorescent F-actin binding proteins.
Because of its planar nature, the IS provides a unique opportunity to image events as they happen, moni-
toring changes in actin retrograde flow in T cells interacting with different stimulatory surfaces or after
pharmacological treatments. Here, we described the imaging methods and analytical procedures used to
measure actin velocity across the IS in T cells spreading on planar stimulatory surfaces.
Key words Actin, Cytoskeleton, Kymograph, Immunological synapse, T-cells, Integrin, Planar lipid
bilayer, Mobile ligands, Spinning disk, Live cell imaging
1 Introduction
The formation of the immunological synapse (IS) between a T cell

and an antigen presenting cell (APC) depends on actin dynamics
downstream of T cell receptor (TCR) and integrin engagement
[1–4]. TCR signaling activates the Arp2/3 complex-dependent
polymerization of branched actin filaments at the edges of the cell-
cell contact site, driving initial spreading of the T cell on the APC
surface and subsequent centripetal flow of the acto-myosin network.
This process corresponds to the retrograde actin flow that occurs at
the leading edge of a migrating cell. Centripetal actin flow drives
the ongoing assembly and function of TCR signaling complexes,
and ultimately shuttles these complexes to the center of the IS,
where signal extinction takes place [5, 6]. Centripetal flow of the T
cell actin network also regulates integrin conformational change,
thereby promoting adhesion to ligands on the APC surface, as well
as outside-in signals that costimulate T cell activation [7, 8]. Actin
dynamics thus function as an essential part of a key feedback loop
that coordinates T cell signaling events at the IS. Thus, measuring
actin flow at the IS is valuable for u nderstanding the fundamental
mechanisms that drive and fine-tune T cell activation.
7
8 Katarzyna I. Jankowska and Janis K. Burkhardt
Much of what is known about actin dynamics at the IS comes

from studies of T cells responding to coverslips or planar lipid
bilayers coated with stimulatory antibodies or ligands [9]. While
these planar stimulatory surfaces do not recapitulate the complex
undulations that are formed at a T cell-APC interface [10, 11],
they represent a powerful tool because they allow investigators to
image the movements of fluorescently tagged cytoskeletal elements
and signaling molecules in the microscope’s X–Y plane. In con-
junction with these planar stimulatory surfaces, several labs have
taken advantage of super resolution approaches such as structured
illumination (SIM), stimulated emission depletion (STED), and
single molecule localization techniques (PALM/STORM) to
examine the molecular architecture at the IS [12–16]. Recently,
lattice light sheet technology has also proven to be valuable [17].
Nonetheless, simpler techniques for live cell imaging such as total
internal reflection (TIRF) and spinning disk confocal microscopy
continue to be the best ways to answer many biological questions.
TIRF optics are often used to image movement of signaling mol-
ecules at the IS. This modality is favored because it focuses analysis
on events occurring at or very near the plasma membrane (within
about 100 nm), and offers low background noise. However, the
actomyosin network extends more than 1 μm into the cell at early
times after TCR engagement [11]. Even at later times when T cells
are well spread, the actin-rich lamellipodia are nearly 0.5 μm thick
[18]. Thus, only a subset of actin filaments is captured within the
TIRF plane. Indeed, it is nearly impossible to gain an overall sense
of the acto-myosin network using TIRF optics. Because we are
interested in the actin cytoskeleton as a functional unit, we prefer
to use spinning disk confocal microscopy. As detailed below, we
typically capture three planes spanning a total distance of 0.5 μm.
This usually captures the entire thickness of the lamellipodium of a
spreading T cell. By generating a maximum intensity projection or
a 3-dimensional rendering of the three planes, we can analyze the
behavior of the lamellipodial actin network as a whole.
Armed with suitable video sequences, it is relatively straight-
forward to carry out quantitative analysis of actin flow rates by
tracking the movement of small structures within the actin net-
work. The most common technique to visualize motion from
sequential 2-D imaging is kymographic analysis [19, 20]. To gen-
erate a kymograph, one first selects a narrow region of interest and
extracts this region from each image in a time series. The selected
region is then laid side-by-side for all time points, generating a
picture (kymograph) that displays movement of objects within the
selected region over time, such that one axis represents space and
the other axis represents time. Movement within these space-time
plots is seen as diagonal lines of bright or dark features, and the
speed of movement can be determined based upon the slope of
these lines.
Analyzing Actin Dynamics at the Immunological Synapse 9
In studies using planar surfaces to analyze protein dynamics at

the IS, one important factor is the mobility of stimulatory ligands.
Neither glass surfaces bearing immobilized ligands nor planar lipid
bilayers bearing freely mobile ligands faithfully recapitulate the
biology of bona fide antigen presenting cells, where some ligands
exhibit constrained mobility and others are freely mobile [21].
However, these simplified systems provide a means of exploring
the ways in which ligand mobility affects actin flow and signaling
through actin-coupled receptors.
There are several good protocols in the literature for the prep-
aration of T cell stimulatory surfaces suitable for microscopy [22–
24]. Here, we describe our procedures for the preparation of both
immobile and mobile stimulatory surfaces, followed by our meth-
ods for imaging and measuring lamellipodial actin flow in T cells
interacting with these surfaces. We provide details for viewing actin
movements in a predetermined plane with optimal spatial and tem-
poral resolution, and testing the effects of ligating specific recep-
tors in the absence of other stimuli and altering actin dynamics
using pharmacological agents.
2 Materials
All solutions should be prepared using reverse osmosis (Milli-Q)

water and analytical grade reagents. Avoid repeated freeze-thawing
of protein solutions and inhibitors; make small aliquots before
freezing. Follow institutional safety guidelines in handling and dis-
posing of hazardous chemicals including Hellmanex III and
Piranha solution and biohazardous materials such as recombinant
viruses and human cells.
2.1 Generation 1. Coverslips for Sticky-Slides (Ibidi): # 1.5H (170 μm ± 5 μm)

of Stimulatory D 263 M Schott glass, 25 mm × 75 mm.
Surfaces 2. 2% Hellmanex III detergent (Hellma Analytics): dissolve 20 ml
2.1.1 Preparation of detergent in 980 ml of water. Other alkaline glassware deter-
of Glass Coverslips gents may substitute for Hellmanex III (e.g., Linbro 7×), but
optimization of concentration, time, and rinsing requirements
is needed.
3. Piranha solution: Working in a fume hood, mix 300 ml of sul-
furic acid with 100 ml of hydrogen peroxide in a glass beaker
[7] (see Note 1).
4. Bath sonicator.
5. Plasma cleaner.
2.1.2 Preparation 1. Bottomless Sticky-Slide VI 0.4 6-channel chambers (recom-

of Imaging Chambers mended supplier –Ibidi, Martinsried, Germany). Various home-
made flow chambers may be substituted. The desired channel
width and length is ~5 × 25 mm with a height of 250 μm.
2.1.3 Coating Surfaces 1. Mouse anti-human T cell receptor: OKT3 (recommended

with Immobilized Ligands supplier—BioXCell, Burlington, VT). Store at 4 °C.
2. Human VCAM-1: 1 mg/ml in PBS. Available as a soluble Fc
fusion from recommended suppliers Sino Biological or R&D
Systems. His-tagged protein also works for nonspecific adsorp-
tion. Store at −20 to −80 °C in small aliquots.
3. Human ICAM-1: 1 mg/ml in PBS. Available as a soluble Fc
fusion from recommended suppliers Sino Biological or R&D
Systems. His-tagged protein also works for nonspecific adsorp-
tion. Store at −20 to −80 °C in small aliquots.
4. L15 imaging medium: Supplement L15 with 2 mg/ml d-(+)
glucose by adding 1 g of d-glucose to 500 ml L15 medium
and filtering through a sterile filter unit (0.2 μm pore size).
Optionally, use phenol red-free RPMI-1640 enriched with 25
mM HEPES by adding 12.5 ml of sterile 1 M HEPES solution
to 500 ml RPMI medium.
5. Phosphate buffered saline (PBS)- standard formulation with or
without d-glucose.
6. Multichannel pipette.
7. Poly-l-Lysine hydrobromide (PLL) (mol. wt. 30,000–
70,000): Dissolve 5 mg in 10 ml of water. Add 40 μl of 5%
NaN3. Freeze in aliquots or store at 4 °C for 1–2 weeks.
2.1.4 Coating Surfaces 1. 50 ml glass round-bottom flask.

with Stimulatory Supported 2. Glass syringes (10, 50, and 500 μl, Hamilton).
Planar Lipid Bilayers
3. 2% Hellmanex III detergent: see Subheading 2.1.1, item 2.
4. Acetone (HPLC grade).
5. Chloroform (HPLC grade).
6. Lipids (recommended supplier—Avanti Polar Lipids, see Note 2):
DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine) 25 mg/ml
in chloroform, DGS-NTA(Ni) (1,2-dioleoyl-sn- glycero-3-
[(N-(5-amino-1-carboxypentyl)iminodiacetic acid)succinyl]
(nickel salt)) 10 mg/ml in chloroform, DSPE- PEG(2000)
Biotin (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-
[biotinyl(polyethylene glycol)-2000] (ammonium salt) ) 5
mg/ml in chloroform. Store at −20 °C.
7. Extruder set (recommended supplier—Avanti Polar Lipids):
excluder with two glass syringes and 50 nm pore membranes.
8. Bath sonicator.
9. Compressed air (house air is fine, but place a filter on the out-
put to remove contaminants).
10. Vacuum dessicator.
11. PBS.
12. Biotinylated OKT3 (0.5 mg/ml). Store at 4 °C.

13. Streptavidin, NeutrAvidin or NeutrAvidin –TexasRed conju-
gate: 1 mg/ml in PBS. Store at −20 °C.
14. Human VCAM-1-His tagged (recommended supplier—Sino
Biological Inc): 1 mg/ml in PBS. Store at −20 to −80 °C in
small aliquots.
15. Human ICAM-1-His tagged (recommended supplier—Sino
Biological): 1 mg/ml in PBS. Store at −20 to −80 °C in small
aliquots.
16. L15 imaging medium: L15 supplemented with 2 mg/ml d-(+)
glucose.
17. Multichannel pipette.
2.2 Imaging Actin 1. Jurkat T cell growth medium: RPMI 1640 enriched with 5%
Dynamics in Living fetal bovine serum and 5% newborn calf serum, 2 mM L-alanyl-
T Cells l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin
(see Note 3).
2.2.1 Preparation
of Cells Expressing 2. Jurkat T cells stably expressing fluorescent actin probes (fluo-
Fluorescent Actin Probes rescent protein-labeled actin, Lifeact, or F-tractin) (see Notes 3
for Live-Cell Imaging and 4 for details on the use of these probes).
Jurkat T-Cells
Primary Human Peripheral 1. Primary T cell growth media: RPMI 1640 supplemented with
Blood CD4+ T Cells 10% fetal bovine serum (Atlanta Biologicals), 2 mM l-alanyl-l-
glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and
50 U/ml of human rIL-2 (see Note 5).
2. Primary human CD4+ T cells expressing fluorescent actin
probes (see Notes 3 and 6).
2.2.2 Fluorescence Below is the description of the setup we use to acquire the images.
Microscopy Other vendors also provide similar systems and there are many
options for analysis software.
1. Inverted microscope (Zeiss Axiovert 200 with Piezo Z-focus).
2. Yokagawa spinning disk head (PerkinElmer Ultraview ERS6
with Photokinesis unit).
3. CCD camera: Orca ER camera, Hamamatsu (see Note 7).
4. Objective: 63× Plan Apo 1.4 NA, oil immersion.
5. Solent Scientific environmental chamber.
6. Multi laser module (laser lines 405, 440, 488, 514, 561, 640 nm).
7.
Emission filters (455/60, 485/60, 527/55, 587/125,
615/70, 705/90).
8. Vibration isolation table (Vibraplane kinetic systems).
9. Image acquisition software (Volocity v. 6.3, Perkin Elmer).
2.3 Image Analysis 1. Volocity v. 6.3 imaging software (Perkin Elmer).

2. Microsoft Excel.
3 Methods
3.1 Generation 1. Sonicate the glass coverslips in 2% Helmanex detergent for

of Stimulatory 30 min.
Surfaces 2. Wash vigorously with water to remove all detergent.
3.1.1 Preparation 3. Air dry.
of Glass Coverslips 4. Dip into fresh Piranha solution and incubate for 30 min (see
(See Note 8) Note 1).
5. Wash with water.
6. Air dry.
7. For studies involving lipid bilayers, we recommend plasma
cleaning the glass for 3 min just before the addition of the
small unilamellar liposome vesicle (SUV) solution. This step is
dispensable for studies where coverslips are coated with stimu-
latory proteins without lipid bilayers.
3.1.2 Preparation 1. Peel the protective paper off the Sticky-Slide 6-channel cham-
of Imaging Chambers bers, revealing the self-adhesive underside. Mount cleaned
(See Note 9) coverslip, aligning carefully. Take care to touch only the edges
of the coverslip.
2. Press down carefully, using a pen or the back of a pair of twee-
zers to secure the seal between each well. To prevent leakage,
make sure that the tape sticks well to the coverslip. If desired,
one can further protect the sides from leaking by applying nail
polish around all sides.
3.1.3 Coating Surfaces Two types of stimulatory surfaces can be prepared: ligand can be
with Immobile or Mobile immobilized by adsorption onto the glass coverslip (Subheading
Ligands “Coating with Immobile Ligands by Protein Adsorption to the
Coverslips”) or ligand can be attached to lipid bilayers where it will
have high lateral mobility (Subheading “Coating Surfaces with
Stimulatory Supported Planar Lipid Bilayers”).
Coating with Immobile We typically apply OKT3, a monoclonal antibody that reacts

Ligands by Protein with an epitope on the epsilon-subunit within the human CD3
Adsorption complex [25], in the presence or absence of the adhesion mole-
to the Coverslips cules VCAM-1 or ICAM-1, which bind to the integrins VLA-4
and LFA-1, respectively. Depending on the experiment, other
stimulatory ligands can be used, and concentrations can be varied
(see Note 10).
1. Dilute OKT3 in PBS to a concentration of 10 μg/ml.
2. Using the multichannel pipette, add 100 μl of 10 μg/ml OKT3

to each chamber well. To create flow in one direction, it is
important always to pipette into one side of the wells (intake
ports), and out from the other side (outtake ports). To facili-
tate this, mark the intake side of the chamber (the side to which
OKT3 was added) with an arrow.
3. Incubate 2 h at RT or overnight at 4 °C.
4. Wash each well three times with PBS. To do this, use a multi-
channel pipette to remove 90% of the solution in the wells
(withdraw from the outtake ports). Then pipette in 150 μl of
PBS to the marked intake side. Repeat two more times. Never
let the wells dry out.
5. At this point, another ligand can be added. We use 2 μg/ml of
VCAM-1 or ICAM-1 in PBS. Withdraw 100 μl of PBS from
the outtake wells and replace with 100 μl of 2 μg/ml VCAM-1
or ICAM-1. Incubate for another 2 h at 37 °C, and wash three
times with PBS, all as in steps 2–5 (see Note 11).
6. Exchange the solution to L-15 imaging medium by washing
three times with pre-warmed L-15 medium (37 °C).
7. Incubate the chamber on the microscope stage at 37 °C for
about 10 min before adding cells.
Coating Surfaces with In order to facilitate specific binding of ligands to the lipid bilayer,
Stimulatory Supported functionalized lipid must be incorporated into the lipid mixture
Planar Lipid Bilayers during vesicle preparation. Many functionalized lipids are com-
mercially available. We used biotinylated lipids and lipids with a
Ni-NTA group, allowing us to attach biotinylated OKT3 (via a
Streptavidin bridge) as well as His-tagged ICAM-1 or VCAM-1.
The following procedure has two stages: Steps 1–13 describe
preparation of lipid vesicles in chloroform in a desired mol% ratio.
We use 5 mM DOPC:DSPE-PEG(2000) biotin:DGS-NTA(Ni)
(nickel salt) in a 98:1:1 mol% ratio. Steps 14–23 describe the use
of these vesicles to generate planar bilayers.
1. Sonicate the 50 ml glass round-bottom flask and the extruder
set in 1% Hellmanex III solution for 10 min.
2. Thoroughly rinse the flask and extruder set with water to com-
pletely remove the residual detergent.
3. Air dry.
4. Rinse the flask in acetone and then chloroform, vortexing to
be sure to cover all surfaces. It’s fine to leave a little chloroform
in the flask.
5. Wash each glass syringe thoroughly by passing chloroform
through it five to ten times.
6. Use the glass syringes to add 200 μl chloroform to the flask.
Again using the glass syringes, add 30.6 μl DOPC, 6 μl
DSPE-PEG(2000), and 1 μl DGS-NTA into the flask, pipet-

ting each into the chloroform. Vortex to mix (see Note 12).
7. Gently dry the lipid solution with compressed air while rotat-
ing the round bottle to make a uniform lipid film.
8. Place the round bottle into a vacuum desiccator and dry for 2
h or overnight.
9. Rehydrate the lipid film with 200 μl of PBS (or the buffer of
your choice) to bring the final lipid concentration to 5 mM,
and then sonicate for 5 min to produce micelles.
10. Assemble the mini-extruder as per Avanti instructions, using
the 50 nm pore membrane.
11. Pass PBS through the extruder a few times to ensure that the
assembly does not leak. Monitor the volume that comes across
the extruder after five to ten passes. If volume is lost, reassem-
ble the set.
12. Extrude the lipid solution through the membrane at a con-
stant, steady rate 21 times, creating a lipid vesicle mixture (see
Note 13).
13. Transfer the lipid vesicle mixture to a 1.5 ml conical microcen-
trifuge tube. This mixture can be kept at 4 °C for 1 week.
14. For each chamber slide, mix 150 μl of the lipid vesicle mixture
with 600 μl PBS to get 1 mM liposome suspension. Vortex to
mix (see Note 14).
15. Using freshly prepared plasma cleaned Sticky-Slide chambers
(see Subheading 3.1.2), and a multichannel pipette, add about
100 μl of diluted lipid vesicle mixture to each chamber and
incubate for 30 min at RT.
16. Rinse the wells thoroughly with PBS to remove the excess ves-
icles. This should be done by sequential addition of PBS to the
intake well and removal of flow-through on the other side.
About 150 μl of PBS can be added at a time for a total of three
to five times. Never allow the wells to dry out or air to enter
the channel containing the bilayer.
17. Remove the final PBS wash and pipette in Streptavidin or fluo-
rescent Neutravidin (1 μg per 100 μl PBS for each well).
Incubate for 20 min at RT.
18. Wash thoroughly as before.
19. Incubate the chambers with biotinylated OKT3 and His-
tagged ICAM-1 or VCAM-1. This should be done sequen-
tially, incubating for 20–30 min and washing three times with
PBS after each addition (see Note 15).
20. Exchange the PBS with three washes of L-15 imaging medium.
21. Bilayer surfaces should be tested for ligand mobility (see Note 16)
and used the same day. We usually transfer them to the microscope
environmental chamber (see Subheading 3.2.2). Ligand mobility

and surface quality may decrease with longer storage.
22. Incubate the chamber on the microscope stage at 37 °C for
about 10 min before adding cells.
3.2 Imaging Actin 1. Culture T cells expressing fluorescent actin probes as detailed
Dynamics in Living in Notes 3–6. Ensure that the culture is growing well, and that
T Cells cells exhibit high viability at the time of analysis.
3.2.1 Preparation
of Cells Expressing
Fluorescent Actin Probes
for Live-Cell Imaging
3.2.2 Preparation 1. Set the environmental chamber on the microscope to 37 °C,

of the Microscope and allow it to equilibrate for at least 1 h prior to imaging.
2. Place all chambers, reagents, etc. into the environmental cham-
ber to allow equilibration.
3.2.3 Preparation 1. Pipette about 5 ml of cells into a 15 ml tissue culture grade
of Cells for Live-Cell conical tube.
Imaging 2. Centrifuge the cell suspension at 250 × g for 5 min at room
temperature.
3. Aseptically aspirate or decant the supernatant without disturb-
ing the cell pellet.
4. Resuspend the cell pellet in 5 ml of L-15 medium.
5. Determine the total number of cells using a hemocytometer.
6. Centrifuge the cells again to remove residual serum.
7. While the cells are in the centrifuge, calculate the volume of
L-15 imaging medium needed to resuspend the cells at the
desired density. We typically resuspend the cells at 1 × 106/ml.
8. Resuspend the washed cells in L-15 medium at the desired
final density. Maintain at 37 °C until imaging (see Note 17).
3.2.4 Imaging Cells 1. Open imaging software and set all basic parameters. Configure
by Spinning Disk Confocal the time-lapse settings; we usually collect a z-stack of three
Microscopy planes spaced 0.25 μm apart every 0.5–1 s, over a total time of
about 4 min.
2. Inject 50 μl of cell suspension into the intake well of a chamber
coated with stimulatory ligands (prepared as described in
Subheading 3.1).
3. Mix and distribute the cells by gently removing about 50 μl of
flow-through from the outtake port and adding back to the
intake port. Repeat this three to five times.
4. Place the chamber under the microscope and allow the cells to
interact with the stimulatory surface for about 5 min before
imaging. In some cases, it is desirable to image the early phases
of T cell contact with the coverslip. The easiest way to do this
is to add only a few cells initially, allow those cells to settle, and
adjusting settings as described in step 5. Then wait for addi-
tional cells to touch down, adding more if needed, and analyze
that population.
5. Meanwhile, adjust the focus and set imaging parameters (expo-
sure time and laser power) based on the brightness of cells as
they come into contact with the surface. Use the lowest pos-
sible intensity and time to minimize photobleaching.
6. Choose a field with individual cells that are not contacting
other cells. If the population of cells is heterogeneous in bright-
ness, take care not to select very bright cells that may overex-
press fluorescent actin probes, as overexpression may perturb
actin dynamics.
7. Focus on the bottom of a cell, just above the coverslip. This
will be the region where actin-rich lamellipodia form as the cell
spreads on the stimulatory surface.
8. Collect a time-lapse series.
9. Image as many fields as needed from one chamber for up to
20 min or when cells start to deform and detach from the sur-
face (see Note 18).
10. If desired, inhibitors can be added to test effects on actin
dynamics (see Note 19, which includes a table of commonly
used inhibitors). After imaging the untreated cells for 1–2 min,
add the desired inhibitor using a gel-loading tip. Mix gently by
pipetting in and out of the intake port, or by removing 50 μl of
medium from the outtake port and adding back to the intake
port. Take care not to disturb the cells or bump the stage.
Resume imaging as soon as possible after adding the inhibitor.
3.3 Image Analysis Actin flow rates are calculated based on kymographic analysis.
Here, we describe the procedure using Volocity v. 6.3. Other soft-
ware packages have similar capabilities (see Note 20).
1. Select a video sequence of a cell for analysis.
2. Draw a ray from the center of immunological synapse (IS) to
the periphery (see Fig. 1a, yellow line in top panel).
3. Generate kymograph. (Go to “tool” in Volocity and choose
“kymograph”).
4. In Volocity, you can set the time units so that one pixel equals
1 s. This way, the y axis of the kymograph (displayed in pixels)
is equal to time in seconds (see Fig. 1a, bottom panel).
5. Choose the line tool and draw lines along the diagonal intensity
maxima (see the white dashed lines in Fig. 1a, bottom panel)
(see Note 21).
6. Go to the “Measurements” section displayed above the kymo-
graph image. In the Measurements view you will find the
length of each of your drawn lines, the location of each line
(start position x, start position y, end position x, end position
y), and the line angles in degrees. The output you will get is
shown in Table 1.
7. Select all the measurements for lines and copy to the Microsoft
Excel folder. You will need angle values (displayed in degrees)
a b
100
OKT3
80 OKT3+VCAM
Flow rate [nm/s]
60
LP
40
20
0
0.0 0.2 0.4 0.6 0.8 1.0
Radius
β1
c
***
Time
Flow rate [nm/s]
100
β2
β3 50
β4
0 0
0 x4 x3 x2 x1 OKT3 OKT3+VCAM
Distance
Fig. 1 Characterization of F-actin dynamics in Jurkat T lymphoma cells. T cells were allowed to interact with
coverslips coated with anti-CD3 (OKT3) ± VCAM-1 and imaged for 4 min. (a) Single time point of a responding
cell stimulated on anti-CD3 (top) and the corresponding kymograph of F-actin dynamics generated along the
yellow line (bottom). The dotted lines trace the paths taken by distinct features along the distance xn with their
corresponding angles βn. (b) Kymographic analysis of F-actin flow in Jurkat T cells, showing the distribution
of F-actin velocity across the immunological synapse. The area marked by the dashed box displays the periph-
eral lamellipodial region (LP). (c) Actin flow rates within the LP region for Jurkat T cells responding to OKT3 in
the absence or presence of VCAM-1. Means ± SD are shown (n = 20–40 cells per condition), ***, P < 0.001.
Scale bar 10 μm
18
Table 1
Data from the measurements panel in Volocity for the white dashed lines in Fig. 1a
Total
line Start Start End End Angle
ID Item Name Name Population Type length position x position y position x position y (degrees)
1 Jurkat_Kymograph Lines 1 Lines Line 28.32 98 39 119 58 132.14

to calculate actin flow rates. If you are interested in calculating

actin flow rates as a function of position along the cell radius as
described in steps 12–19, you will also need end position x.
8. In Volocity, “angle” ranges from 0 to 180. 0 means pointing
“up the screen,” along the y axis. 180 means pointing “down”
the screen. Thus, to calculate the angle of the lines you have
drawn (see β1 − n in Fig. 1a, bottom panel), you need to sub-
tract the angle obtained from Volocity from 180 (β_degree =
180-Angle[degree]).
9. Convert each angle β_degree to radians (multiply by π/180°).
10. Calculate flow rate from the slope: V[μm/s]=tan(β_radian) ×
(tpixel [pixels]/ts[s]) × Cf[μm/pixel], where Cf is a conversion
factor that will convert distance in pixels to metric scale. Note
that if you set the time units so that one pixel equals 1 s, in step 4,
this conversion is unnecessary because tpixels/ts = 1.
11. Average the flow rates calculated from all kymographs and cal-
culate standard deviation (SD) and standard error (SE).
12. Since actin flow decelerates with centripetal movement, it is
valuable to calculate actin flow rates as a function of position
along the radius of the immunological synapse. To do this,
for each line drawn in step 5, record the position (x end
point), based upon the distance from IS center (see x1 − n in
Fig. 1a, bottom panel). For each cell, record the radius (r)
(see Note 22).
13. Create a table like that shown in Table 2. In addition to mea-
sured values from steps 8–10, your table should include nor-
malized distances (Dn = position xn/cell radius r).
14. Calculate actin flow rates as in step 10.
15. Create additional kymographs and repeat the analysis of
actin slopes and flow rate calculations for 10–30 cells per
condition.
Table 2
Calculation of actin velocity across the IS as described in Subheading 3.3
End Normalized β (radian) = V [μm/s] =

position distance Angle β = 180–Angle β (degree) × Slope = slope × Cf ×
ID x D = xn/r (degrees) (degrees) π/180° tan(βradian) tpixel/ts
1 119 0.915 132.14 47.86 0.83 1.10 0.116
2 93 0.715 138.37 41.63 0.73 0.89 0.093
3 76 0.585 153.43 26.57 0.46 0.50 0.053
4 58 0.446 171.87 8.13 0.14 0.14 0.015
n xn Dn angle β_degree β_radian slope Vn
16. Bin the values based on where each measurement was made
along the cell radius (after normalizing from 0 to 1). Units of
0.1 radius work well. Average flow rate values for each bin.
17. Create a graph like that shown in Fig. 1b, showing actin flow
rate vs normalized position (from 0 to 1). This graph will show
the distribution of F-actin velocity across the immunological
synapse, grouped into ten equally spaced bins.
18. As an alternative to analyzing actin flow across the entire
immunological synapse, it is sometimes sufficient to analyze
flow where it is fastest, i.e., within the outer lamellipodial
region. Based on morphology and localization of actin and
myosin, we define this region as the outer 20% of the radius
[18]. To obtain this value, bin measurements as described in
step 16 and average measurements for all data points in the
radius range 0.8–1 (see Fig. 1 c).
19. Calculate statistical significances using Student’s T test for
unpaired samples.
4 Notes
1. Piranha solutions are extremely energetic and may result in

explosion or skin burns if not handled with extreme caution.
Work in a fume hood and wear a lab coat, acid apron, safety
goggles, and heavy, nitrile gloves. When preparing Piranha
solution, always add the peroxide to the acid. Dispose of
Piranha solution in a sturdy container, sitting within a second-
ary container. Maintain a loose cap; the solution generates
gases for several days and could explode if not allowed to vent.
It is important to avoid adding any other waste into this con-
tainer. Mixing with organic compounds like toluene, chloro-
form, and phenol will result in a violent reaction. The Piranha
waste container must be well marked and lab members should
be instructed about the associated hazards. We post a promi-
nent sign whenever working with Piranha solution, to mini-
mize traffic and distractions in that area of the laboratory.
2. Chloroform evaporation can cause inaccuracy in measurement,
thus the volumes of lipid solutions should be noted and losses
in volume should be replenished before mixing different lipid
solutions.
3. Jurkat cells are grown in suspension, in a 37 °C degree incuba-
tor containing 5% CO2. The cells should not be allowed to
become too dense; ideally, the culture should be kept between
1x105/ml and 2x106/ml. Typically, these cells will need to be
split 1:10 every 2 days. Jurkat cells sometimes become unre-
sponsive when passaged multiple times, and we typically thaw
a fresh vial about once a month. Thus, it is important to main-

tain a good supply of frozen vials. Thaw an early freeze and
expand again when running low on backup vials. An important
caveat to the use of Jurkat T cells is that these cells have known
signaling abnormalities, particularly in inositol lipid regulatory
pathways [26, 27]. Nonetheless, these cells are a valuable tool
because of their large size and the ability to generate stable
transfectants. For our studies, we typically use a Jurkat T cell
line that was stably transfected with GFP-actin [28]. These
cells were screened to rule out changes in TCR signaling, and
to ensure that GFP-actin is not grossly overexpressed. Plasmids
and recombinant viral particles expressing fluorescent actin are
available from Clontech and can be expressed in Jurkat T cells
by transfection or lentiviral transduction. Fluorescently tagged-
actin can perturb actin dynamics, and can be excluded from
lamellar networks and filopodia [29–31]. Moreover, this strat-
egy labels both polymeric and monomeric pools. Thus, we
sometimes use cells expressing fluorescent-Lifeact, a 17 amino
acid peptide that binds selectively and reversibly to actin fila-
ments (F-actin), and has little or no effect on actin dynamics
[32]. Plasmids and recombinant viral particles expressing fluo-
rescent Lifeact are available from Ibidi. Alternatively, trans-
genic Lifeact mice are available [33] and can be used as a source
of T cells; such mice are particularly valuable for studies requir-
ing naive T cells, since activation is typically required for effec-
tive viral transduction. Other investigators use F-tractin, the
actin-binding domain from rat inositol triphosphate 3-kinase,
which facilitates visualization of actin arcs at the lamellar region
[34, 35]. Care should be taken in choosing an actin probe,
since each has the potential to bias the structures being imaged
[30]. Where possible, controls should be performed using
multiple approaches.
4. Plasmids expressing fluorescent F-tractin are available from
Addgene (Cambridge, MA). Regardless of the probe used,
cells can be stably transfected by electroporation or lentiviral
transduction, and selected with appropriate antibiotics. Stable
cell lines should either be screened periodically for expression,
or highly stable lines should be generated by two rounds of
single cell cloning. Care should always be taken that lines do
not change over time, and multiple vials of early passage freezes
should be maintained and thawed whenever there is evidence
of change.
5. We obtain primary human peripheral blood CD4+ T cells from
the University of Pennsylvania’s Human Immunology Core
under an Institutional Review Board approved protocol.
However, they can be readily prepared from human peripheral
blood using commercial kits based on depletion of other cell
types. Purified CD4+ T cells are activated with CD3/CD28

magnetic beads (Dynabeads, Life Technologies) and cultured
a humidified 37 °C incubator with 5% CO2. Beads are mag-
netically removed on day 7 after initial stimulation, and cells
are then cultured for an additional day in the presence of 10
U/ml of IL-2. Primary T cells should be used at days 8 and 9
after activation. During this window, the cells are optimally
“rested” after the initial activation, and are not yet showing
signs of activation-induced cell death.
6. To generate primary human CD4+ lymphocytes expressing
Lifeact-GFP, cells are cultured for 24 h with human T-Activator
CD3/CD28 magnetic beads, and then transduced with
GFP lentivirus. Lentivirus and 8 μg/ml polybrene
Lifeact-
(Sigma-Aldrich) is mixed with 5–10×106 T cells in the wells in
a 5 ml round-bottom polystyrene tube and centrifuged at
1200 × g for 2 h at 37 °C. Lentivirus-containing media is then
replaced with T cell culture media, and the cells are returned
to culture. Two days after transduction, the media is supple-
mented with 2 μg/ml puromycin, and cells are cultured for an
additional 4 days before magnetic removal of the activator
beads. Cells are cultured for an additional 1–2 days in media
with 2 μg/ml puromycin and 10 U/ml IL-2 before use (day
8–9 after activation). A similar approach can be used for expres-
sion of other actin probes listed in Note 3.
7. While many investigators choose a more sensitive EM CCD
camera, our system was configured with an ER CCD because
it has a smaller pixel size, permitting higher resolution imag-
ing. In addition, the ER camera allows a combination of fluo-
rescence and white light (DIC) imaging. Scientific CMOS
cameras also work well for the same reasons, and have a larger
field size and improved signal-to-noise over the older ER CCD
technology.
8. Proper cleaning is especially important for the generation of
uniform stimulatory surfaces. This is especially true for studies
using lipid bilayers, since any impurities may disturb the lipid
bilayer and its mobility. Thus, it is important to use a deter-
gent solution followed by an acid bath. We have used both
Piranha solution (detailed above), and Nochromix solution
successfully. As detailed in Note 1, Piranha solution is quite
hazardous. Nochromix, a commercial formulation that con-
tains ammonium persulfate, is a safer alternative. Nochromix
solution should be prepared in a fume hood according to
package directions. Usually, Nochromix powder is mixed with
water and added to sulfuric acid, forming a clear solution. In
this case incubate Hellmanex-cleaned glass slides in Nochromix
solution overnight. Following cleaning with either Piranha or
Nochromix, some investigators plasma clean the glass for
3 min just before the addition of SUV solution.
9. It is best to prepare chambers just before usage and to use

freshly cleaned coverslips. In place of the sticky slides, one can
use any high coverslip-bottomed imaging chamber (e.g., Mat-
Tek dishes). We prefer the Ibidi sticky slides because they pro-
vide a very small controlled volume and access for perfusion of
inhibitors. The multi-well format also allows imaging of mul-
tiple samples on a single glass surface, thus speeding imaging
and promoting reproducibility.
10. Some investigators recommend coating glass coverslips with
poly-l-lysine prior to addition of stimulatory ligands. We have
tried the procedure with and without poly-l-lysine, and have
not observed a difference in results. However, this may vary
depending on the ligand used. To use poly-l-lysine, add 120 μl
of poly-l-lysine solution to wells before step 1, incubate for
30 min at RT, and wash three times with water. Remove the
water using an aspirator and air dry. Proceed with step 1 as
before.
11. As an alternative to sequential coating with ligands, we have
coated the glass with a mixture of OKT3 and VCAM-1 or
ICAM-1 overnight at 4 °C. We have not seen a difference in
experimental results. Nonetheless, we prefer sequential coating
since we want to hold the levels of OKT3 constant, and mixing
could affect the amount of OKT3 adsorbed onto the surface.
12. If desired, the lipid mixture can be prepared in larger quantity,
and stored in single use aliquots at −20 °C. Note the volume,
and upon thawing, add chloroform as necessary to account for
evaporation.
13. Always do uneven number of passes so that your final micelles
are in the opposite syringe than the original, this ensures more
uniform size and quality.
14. Oversaturation of glass surface with lipid vesicles will result in
partial fusion of vesicles to glass and diminish the mobility of
the bilayers.
15. Biotinylated and His-tagged ligands should be added in excess
to cover the surface completely. We use 1 μg of OKT3-biotin
and 2 μg of ICAM-1–His per chamber. Optimal concentrations
of ligands should be determined empirically. Saturating all the
biotin binding sites of the bilayer bound Neutravidin will also
help prevent crosslinking by the multibiotinylated OKT3. If
lower densities of anti-CD3 are desired then using monovalent
and monobiotinylated UCHT1 Fab’ is an alternative [36].
16. High concentration of ligand or Streptavidin may cause ligand
aggregation and affect its mobility. Thus, it is important to
perform control experiments for bilayer quality and ligand
mobility. The simplest way to test bilayer or ligand mobility is
to perform FRAP experiments. For this, lipids or ligands have
to have fluorescent dye. For lipid bilayers we have used Texas

Red DHPE (0.3%, about 2 μl of 1 mg/ml stock) but any lipid
dye will work. To test ligand mobility we have used Neutravidin-
Texas Red, OKT3-FITC, and VCAM-Alexa Fluor 647 that
was labeled using a microscale protein labeling kit. When the
surface is ready, image the surface and photobleach a small area
until the area appears black and next record its recovery. If
recovery is observed (the black spot will get brighter over
time) then the surface is mobile, but if the bleached area
remains dark then that will indicate that the ligands are unable
to diffuse and are immobilized on the surface. In this case, the
surface has to be prepared again. We have observed ligand
immobility even in cases when lipid bilayer was mobile, due to
ligand aggregation. Thus, it is the best to test both bilayer and
ligand mobility at the same time. We have observed Neutravidin
aggregation when higher concentrations of DSPE-PEG(2000)
Biotin lipid are used (above 1%) but this may vary with specific
lipid composition. Thus, optimal concentrations of ligands
should be determined empirically.
17. Because L-15 medium is buffered to maintain pH in an air
environment; cells do not need to be kept in a CO2 incubator
during this time. Instead, it is often convenient to keep the
working tube of cells in the microscope-mounted environmen-
tal chamber. The L-15 imaging medium is serum-free to elimi-
nate molecules that might interfere with cell responses to
surface bound ligands. This means that the cells awaiting imag-
ing are undergoing serum deprivation. This can be useful, as it
tends to suppress basal levels of cell signaling. However, care
should be taken not to maintain cells in L-15 for more than
about 2 h. If longer times are needed, it is best to take another
batch of cells from tissue culture.
18. Monitor cells during imaging, making focus corrections as
needed. This is especially important if your system lacks an
autofocus correction mechanism. Even small changes in envi-
ronmental temperature may cause focus drift. To minimize
drift, we use an incubator chamber that encloses the entire stage
and objective and preincubate the chamber before imaging.
Systems employing a heated dish and an objective heater can be
used, but we find that they cause more problems with focus.
19. Optionally, a syringe pump may be used to introduce an
inhibitor into the chamber. Acute treatment with inhibitors
by addition during imaging only works for fast-acting inhibi-
tors; some drugs such as Y27632 and blebbistatin require lon-
ger to take effect, so cells must be pre-treated prior to adding
to the imaging chamber. Conversely, agents that depolymer-
ize actin filaments abolish T cell spreading, so can only be
used acutely, after the cells have interacted with stimulatory
surfaces. Table 3 lists commonly used inhibitors.
Table 3
Inhibitors used to perturb dynamics of the acto-myosin network
Inhibitor Dose Effect on T cell actin

name Mode of action (μM)a dynamics References
Cytochalasin Inhibits actin 10 Induced the accumulation of [21]
D polymerization by disordered F-actin–rich
binding to growing zones
ends of actin nuclei
and filaments (F-actin),
and preventing
addition of monomers
(G-actin) to these sites
Latrunculin Inhibits actin 1 Led to complete depletion of [21]
B (Note A) polymerization the F-actin network
through association
with actin monomers
Jasplakinolide Disrupts actin filaments 1 Arrested F-actin retrograde [8, 18, 34]
(Jas) and induces flow within 30 s. The
polymerization of network formed a tight
monomeric actin into band that constricted inward
amorphous masses. due to myosin II activity
Blebbistatin Inhibits myosin II 50 T cell spreading and [8, 18, 34,
(Bleb) ATPase activity and centripetal actin flow were 37]
(Note B) slows down phosphate unaffected by blebbistatin
release pretreatment. At late time
points, myosin-inhibited
cells failed to contract
normally and became
irregularly shaped.
Disrupted organization of
actin arcs
Y27632 Rho kinase inhibitor, 25 T cell spreading and [8, 18]
(Y27) blocks phosphorylation centripetal actin flow were
of myosin light chain unaffected. Effects at late
at S19 and inhibits time points were similar to
myosin II filament blebbistatin
assembly
Jas+Bleb / Inhibitor cocktail See Arrested F-actin flow and [8, 18]
Jas+Y27 above caused a slow, partial
collapse of the actin network
SMIFH2 Inhibits formin-mediated 2.5–10 Initially slowed centripetal [38, 39]
actin assembly flow of actin arcs. Eventually
arrested arc generation from
the cell edge
(continued)
Table 3
(continued)
Inhibitor Dose Effect on T cell actin

name Mode of action (μM)a dynamics References
CK666 Arp2/3 complex 100 Altered lamillipodial actin [38, 39,
inhibitor; stabilizes the architecture and slowed 40]
inactive state of centripetal flow. Drove a
Arp2/3 complex and lamellipodial-to-filopodial
prevents shape change. Blocked the
conformational formation of TCR-
changes required for associated actin foci
activation (invadopodia like
projections within the
central region of the
immunological synapse)
CK689 Inactive homologue of 100 Actin structure was [39, 40]
CK666 indistinguishable from
untreated cells
Note A: Latrunculin B is serum sensitive [41]. Thus, a higher dose may be required in the presence of serum
Note B: Blebbstatin is light sensitive. Extended exposure to blue light (450–490 nm) may cause degradation of bleb-
bistatin to an inactive product via cytotoxic intermediates. Thus, this drug should be handed in dim lighting and red
fluorophores should be used. A photo-stable analog of (−)-blebbistatin, (S)-nitro-blebbistatin, is stable to prolonged
irradiation at 450–490 nm. (S)-nitro -blebbstatin is commercially available from Cayman Chemicals and can be used in
the same dose as the photo-inactivated form [42].
a
Final concentration. Inhibitors are usually prepared as 200–500 × stocks, dissolved in DMSO

20.
Several software packages offer kymography modules.
MetaMorph has been used successfully for studies similar to
ours [19, 34]. ImageJ can also be used after installation of
plugins that are available online.
21. In cases when actin speckles are not readily visible, recorded
movies can be processed. We used the Smart Sharpen filter in
Adobe Photoshop, with 300% amplification of local maxima
within a 3 pixel radius. This facilitated identification of indi-
vidual GFP-actin speckles, which can be used as fiduciary marks
for analysis. In control studies, analysis of sharpened and
unprocessed movies yielded similar results. As an alternative
approach, a portion of the F-actin network can be photo-
bleached to induce a synchronous wave of bleached GFP-actin
propagating toward the center of the IS. In both cases, a ray
was struck from the center of the IS to the periphery, and verti-
cal kymographs were generated in Volocity analyzed as
described previously (see Subheading 3.3).
22. The depletion angles (slopes) are sufficient to calculate rates
of actin flow. However, actin flow decelerates as the network
moves toward the center of the IS. Thus, we typically calculate
actin flow rates as a function of the position within the synapse
where each measurement was made. Because the diameter of

spread T cells varies somewhat, it is convenient to set the
radius of each spread T cell to 1 and normalize all positions
accordingly.
References
1. Kumari S, Curado S, Mayya V, Dustin ML ognition is facilitated by invadosome-like pro-

(2014) T cell antigen receptor activation and trusions formed by memory/effector T cells.
actin cytoskeleton remodeling. Biochim J Immunol 188(8):3686–3699. doi:10.4049/
Biophys Acta 1838(2):546–556. doi:10.1016/ jimmunol.1102594
j.bbamem.2013.05.004 11. Roybal KT, Mace EM, Mantell JM, Verkade P,
2. Comrie WA, Burkhardt JK (2016) Action and Orange JS, Wulfing C (2015) Early signaling in
traction: cytoskeletal control of receptor trig- primary T cells activated by antigen presenting
gering at the immunological synapse. Front cells is associated with a deep and transient
Immunol 7:68. doi:10.3389/fimmu.2016. lamellal actin network. PLoS One 10(8):
00068 e0133299. doi:10.1371/journal.pone.0133299
3. Le Floc'h A, Huse M (2015) Molecular mech- 12. Brown AC, Dobbie IM, Alakoskela JM, Davis
anisms and functional implications of polarized I, Davis DM (2012) Super-resolution imaging
actin remodeling at the T cell immunological of remodeled synaptic actin reveals different
synapse. Cell Mol Life Sci 72(3):537–556. synergies between NK cell receptors and integ-
doi:10.1007/s00018-014-1760-7 rins. Blood 120(18):3729–3740. doi:10.1182/
4. Beemiller P, Krummel MF (2013) Regulation blood-2012-05-429977
of T-cell receptor signaling by the actin cyto- 13. Sherman E, Barr V, Samelson LE (2013)
skeleton and poroelastic cytoplasm. Immunol Super-resolution characterization of TCR-
Rev 256(1):148–159. doi:10.1111/imr.12120 dependent signaling clusters. Immunol Rev
5. Cemerski S, Das J, Giurisato E, Markiewicz 251(1):21–35. doi:10.1111/imr.12010
MA, Allen PM, Chakraborty AK, Shaw AS 14. Mace EM, Orange JS (2014) Visualization of
(2008) The balance between T cell receptor the immunological synapse by dual color time-
signaling and degradation at the center of the gated stimulated emission depletion (STED)
immunological synapse is determined by anti- nanoscopy. J Vis Exp 85. doi:10.3791/51100
gen quality. Immunity 29(3):414–422. 15. Ashdown G, Pandzic E, Cope A, Wiseman P,
doi:10.1016/j.immuni.2008.06.014 Owen D (2015) Cortical actin flow in T cells
6. Lee KH, Dinner AR, Tu C, Campi G, quantified by spatio-temporal image correla-
Raychaudhuri S, Varma R, Sims TN, Burack tion spectroscopy of structured illumination
WR, Wu H, Wang J, Kanagawa O, Markiewicz microscopy data. J Vis Exp 106:e53749.
M, Allen PM, Dustin ML, Chakraborty AK, doi:10.3791/53749
Shaw AS (2003) The immunological synapse 16. Lillemeier BF, Mortelmaier MA, Forstner MB,
balances T cell receptor signaling and degrada- Huppa JB, Groves JT, Davis MM (2010) TCR
tion. Science 302(5648):1218–1222 and Lat are expressed on separate protein
7. Springer TA, Dustin ML (2012) Integrin islands on T cell membranes and concatenate
inside-out signaling and the immunological during activation. Nat Immunol 11(1):90–96.
synapse. Curr Opin Cell Biol 24(1):107–115. doi:10.1038/ni.1832
doi:10.1016/j.ceb.2011.10.004 17. Ritter AT, Asano Y, Stinchcombe JC,
8. Comrie WA, Babich A, Burkhardt JK (2015) Dieckmann NM, Chen BC, Gawden-Bone C,
F-actin flow drives affinity maturation and spa- van Engelenburg S, Legant W, Gao L, Davidson
tial organization of LFA-1 at the immunologi- MW, Betzig E, Lippincott-Schwartz J, Griffiths
cal synapse. J Cell Biol 208(4):475–491. GM (2015) Actin depletion initiates events
doi:10.1083/jcb.201406121 leading to granule secretion at the immuno-
logical synapse. Immunity 42(5):864–876.
9. Balagopalan L, Sherman E, Barr VA, Samelson doi:10.1016/j.immuni.2015.04.013
LE (2011) Imaging techniques for assaying
18. Babich A, Li S, O'Connor RS, Milone MC,
lymphocyte activation in action. Nat Rev
Freedman BD, Burkhardt JK (2012) F-actin
Immunol 11(1):21–33. doi:10.1038/nri2903
polymerization and retrograde flow drive sus-
10. Sage PT, Varghese LM, Martinelli R, Sciuto tained PLCgamma1 signaling during T cell
TE, Kamei M, Dvorak AM, Springer TA, activation. J Cell Biol 197(6):775–787.
Sharpe AH, Carman CV (2012) Antigen rec- doi:10.1083/jcb.201201018
19. Doggett TM, Breslin JW (2011) Study of the of actin network architecture. Bioarchitecture
actin cytoskeleton in live endothelial cells 4(6):189–202. doi:10.1080/19490992.2014.1
expressing GFP-actin. J Vis Exp 57. 047714
doi:10.3791/3187 31. Doyle T, Botstein D (1996) Movement of
20. Chiba K, Shimada Y, Kinjo M, Suzuki T, yeast cortical actin cytoskeleton visualized
Uchida S (2014) Simple and direct assembly of in vivo. Proc Natl Acad Sci U S A 93(9):
kymographs from movies using 3886–3891
KYMOMAKER. Traffic 15(1):1–11. 32. Riedl J, Crevenna AH, Kessenbrock K, Yu JH,
doi:10.1111/tra.12127 Neukirchen D, Bista M, Bradke F, Jenne D,
21. Comrie WA, Li S, Boyle S, Burkhardt JK (2015) Holak TA, Werb Z, Sixt M, Wedlich-Soldner R
The dendritic cell cytoskeleton promotes T (2008) Lifeact: a versatile marker to visualize
cell adhesion and activation by constraining F-actin. Nat Methods 5(7):605–607. doi:
ICAM-1 mobility. J Cell Biol 208(4):457–473. 10.1038/nmeth.1220
doi:10.1083/jcb.201406120 33. Riedl J, Flynn KC, Raducanu A, Gartner F,
22. Bunnell SC, Barr VA, Fuller CL, Samelson LE Beck G, Bosl M, Bradke F, Massberg S, Aszodi
(2003) High-resolution multicolor imaging of A, Sixt M, Wedlich-Soldner R (2010) Lifeact
dynamic signaling complexes in T cells stimu- mice for studying F-actin dynamics. Nat
lated by planar substrates. Sci STKE 2003(177): Methods 7(3):168–169. doi:10.1038/
PL8 nmeth0310-168
23. Vardhana S, Dustin M (2008) Supported pla- 34. Yi J, Wu XS, Crites T, Hammer JA 3rd (2012)
nar bilayers for the formation of study of Actin retrograde flow and actomyosin II arc
immunological synapses and kinapse. J Vis Exp contraction drive receptor cluster dynamics at
19. doi:10.3791/947 the immunological synapse in Jurkat T cells.
24. Dustin ML, Starr T, Varma R, Thomas VK Mol Biol Cell 23(5):834–852. doi:10.1091/
(2007) Supported planar bilayers for study of mbc.E11-08-0731
the immunological synapse. Curr Protoc 35. Johnson HW, Schell MJ (2009) Neuronal IP3
Immunol Chapter 18:Unit 18 13. 3-kinase is an F-actin-bundling protein: role in
doi:10.1002/0471142735.im1813s76 dendritic targeting and regulation of spine
25. Nguyen K, Sylvain NR, Bunnell SC (2008) T morphology. Mol Biol Cell 20(24):5166–
cell costimulation via the integrin VLA-4 inhib- 5180. doi:10.1091/mbc.E09-01-0083 E09-
its the actin-dependent centralization of signal- 01-0083 [pii]
ing microclusters containing the adaptor 36. Schubert DA, Gordo S, Sabatino JJ Jr,
SLP-76. Immunity 28(6):810–821. doi: Vardhana S, Gagnon E, Sethi DK, Seth NP,
10.1016/j.immuni.2008.04.019 Choudhuri K, Reijonen H, Nepom GT,
26. Abraham RT, Weiss A (2004) Jurkat T cells and Evavold BD, Dustin ML, Wucherpfennig KW
development of the T-cell receptor signalling (2012) Self-reactive human CD4 T cell clones
paradigm. Nat Rev Immunol 4(4):301–308. form unusual immunological synapses. J Exp
doi:10.1038/nri1330 Med 209(2):335–352. doi:10.1084/
27. Bartelt RR, Cruz-Orcutt N, Collins M, jem.20111485
Houtman JC (2009) Comparison of T cell 37. Zhang X, Moore SW, Iskratsch T, Sheetz MP
receptor-induced proximal signaling and (2014) N-WASP-directed actin polymerization
downstream functions in immortalized and pri- activates Cas phosphorylation and lamellipo-
mary T cells. PLoS One 4(5):e5430. dium spreading. J Cell Sci 127(Pt 7):1394–
doi:10.1371/journal.pone.0005430 1405. doi:10.1242/jcs.134692
28. Gomez TS, McCarney SD, Carrizosa E, Labno 38. Henson JH, Yeterian M, Weeks RM, Medrano
CM, Comiskey EO, Nolz JC, Zhu P, Freedman AE, Brown BL, Geist HL, Pais MD,
BD, Clark MR, Rawlings DJ, Billadeau DD, Oldenbourg R, Shuster CB (2015) Arp2/3
Burkhardt JK (2006) HS1 functions as an complex inhibition radically alters lamellipodial
essential actin-regulatory adaptor protein at the actin architecture, suspended cell shape, and
immune synapse. Immunity 24(6):741–752. the cell spreading process. Mol Biol Cell
doi:10.1016/j.immuni.2006.03.022 26(5):887–900. doi:10.1091/mbc.E14-07-
29. Spracklen AJ, Fagan TN, Lovander KE, Tootle 1244
TL (2014) The pros and cons of common actin 39. Vitriol EA, McMillen LM, Kapustina M,
labeling tools for visualizing actin dynamics Gomez SM, Vavylonis D, Zheng JQ (2015)
during Drosophila oogenesis. Dev Biol 393(2): Two functionally distinct sources of actin
209–226. doi:10.1016/j.ydbio.2014.06.022 monomers supply the leading edge of lamelli-
30. Belin BJ, Goins LM, Mullins RD (2014) podia. Cell Rep 11(3):433–445. doi:10.1016/j.
Comparative analysis of tools for live cell imaging celrep.2015.03.033
40. Kumari S, Depoil D, Martinelli R, Judokusumo latrunculin B on mechanical properties of cells.

E, Carmona G, Gertler FB, Kam LC, Carman J Cell Sci 114(Pt 5):1025–1036
CV, Burkhardt JK, Irvine DJ, Dustin ML 42. Ponsaerts R, D'Hondt C, Bultynck G, Srinivas
(2015) Actin foci facilitate activation of the SP, Vereecke J, Himpens B (2008) The myosin
phospholipase C-gamma in primary T lympho- II ATPase inhibitor blebbistatin prevents
cytes via the WASP pathway. eLife 4. thrombin-induced inhibition of intercellular
doi:10.7554/eLife.04953 calcium wave propagation in corneal endothe-
41. Wakatsuki T, Schwab B, Thompson NC, Elson lial cells. Invest Ophthalmol Vis Sci 49(11):
EL (2001) Effects of cytochalasin D and 4816–4827. doi:10.1167/iovs.07-1533
Chapter 3
Analysis of Microtubules and Microtubule-Organizing

Center at the Immune Synapse
Noelia Blas-Rus, Eugenio Bustos-Morán, Francisco Sánchez-Madrid,
and Noa B. Martín-Cófreces
Abstract
The immune synapse (IS) is a specialized structure that enables cell-cell communication between immune
cells. As such, it involves direct cell-to-cell contact. It is sustained by cytoskeletal components that allow
the intracellular polarization of different organelles and the surface re-organization of signaling and adhe-
sion receptors. The tubulin-based cytoskeleton is a key player in IS formation and signaling. We describe
methods to analyze through Western blot and microscopy analysis the polarization to the IS of the centro-
some, also known as microtubule-organizing center (MTOC), the dynamics of microtubule growth and
polymerization from the MTOC to the IS and the activation of signaling molecules.
Key words Immune synapse, Cytoskeleton, Signaling, T Cell receptor, Mitochondria, Centrosome,
Microtubules
1 Introduction
The immune synapse (IS) is a cell-cell contact between a T cell and

an Antigen Presenting Cell (APC) that enables the activation of
the T cell receptor (TCR) and its downstream signaling pathways.
During the formation of the IS, the TCR and its associated mole-
cules segregate into a central area at the interface with the APC,
surrounded by adhesion molecules that help to close the extracel-
lular milieu into a synaptic cleft [1, 2]. During this process, the
tubulin cytoskeleton undergoes dramatic changes, promoting the
translocation of the microtubule-organizing center (MTOC) to
the IS. The translocation of the MTOC is crucial for proper T cell
activation as it orchestrates microtubule growth. The microtubular
network at the IS controls the organization of multiple organelles,
*
The chapter authors, Drs. Blas-Rus and Bustos-Moran have contributed equally as first authors, while the
last two authors, Drs. Sanchez-Madrid and Martin-Cofreces, have contributed equally as senior authors.
31
32 Noelia Blas-Rus et al.
such as the Golgi Apparatus (GA), the endolysosomal system,

including the multivesicular bodies (MVB), and the mitochondria.
In fact, the majority of vesicular traffic at the IS depends on the
microtubules beneath the plasma membrane [3, 4].
The analysis of the tubulin cytoskeleton during the IS has been
carried out using a varied array of approaches. Protocols including
cell lysis, subcellular fractionation, and/or specific immunoprecipi-
tation and subsequent Western blot analysis has allowed the detec-
tion of posttranslational modifications related to the microtubules
during T cell activation. For example, acetylation at Lys40 and
tyrosination/detyrosination (Glu) of the C-terminus of tubulin
monomers inform on their stability and dynamics [5, 6]. MTOC
translocation to the IS has been analyzed by microscopy (light,
confocal and total internal reflection fluorescence microscopy
(TIRFm)) in fixed T cell-APC conjugates or live T cells allowed to
spread over stimulation surfaces or by in vivo imaging [7, 8].
Indeed, the localization of microtubule- or centrosome-associated
proteins is also important during IS assembly, due to their role in
different processes, such as TCR activation or vesicular trafficking
[7]. Here, we describe different approaches to analyze the changes
to microtubule modification, localization, and dynamics in T cells
in response to specific peptide antigen, super-antigen, or poly-
clonal activation. We take advantage of different protein markers
and methodologies that allow the study of TCR activation, micro-
tubule dynamics, and IS formation.
2 Materials
2.1 Cells 1. Transfected cell lines Jurkat lymphoblastoid cells; E1–6 (Vαl.2
Vβ8+ TCR) or CH7C17 (HA1.7 Vβ3+ transgenic αβTCR,
specific for HA peptide) cells.
2. B cell lines as antigen presenting cells for conjugates: Raji B
(Burkitt lymphoma) and Hom2 (HLA-DRB1*0101 positive,
EBV-transformed) lymphoblastoid cells with MHC-II com-
patible for Jurkat E1–6 cells and CHC17 cells, respectively.
3. Primary T lymphocytes from human healthy donors (purified
CD4+ or SEE-specific blasts).
4. Primary CD4+ T lymphocytes from mouse lymph nodes and
spleen.
2.2 Reagents 1. Stimulatory antibodies and recombinant proteins: anti-CD3

and anti-CD28 antibodies with stimulatory activity. Human:
purified OKT3 (eBiosciences), T3b (produced at the labora-
tory), or HIT-3a (Biolegend) monoclonal antibodies for CD3ε
and CD28.2 for CD28 (BD Biosciences). Mouse: 2c11 clone
for CD3ε and clone 37.51 for CD28 (BD Biosciences).
Recombinant ICAM1-Fc of human or mouse origin (R&D).
Analysis of Microtubules and Microtubule-Organizing Center at the Immune Synapse 33
Recombinant SEE (Staphylococcus aureus enterotoxin E; Toxin

Technologies) and synthetic HA307–319 peptide (sequence:
PKYVKQNTLKLAT).
2. Cytokines: human recombinant IL-2 cytokine. Mouse recom-
binant IL-7.
3. Plasmids: EB3-GFP (a kind gift from Dr. Anna Akhmanova)
[8], tubulin-mcherry (a kind gift from Dr. Draber) [9], and
Ensconsin-GFP [10].
4. Antibodies and probes for immunofluorescence: anti-Tubulin
(alpha or gamma), anti-CD3, fluorochrome-conjugated
Phalloidin, fluorochrome-conjugated highly crossabsorbed
secondary antibodies, CMAC (7-Amino-4-Chloro-
methylcoumarin).
5. Poly-l-lysine hydrobromide 75,000 > Mw > 150,000,
γ-irradiated for cell culture.
6. Fibronectin from human plasma suitable for cell culture.
7. Coverslip-bottom dishes for imaging. The chambers may be
commercial (35 mm diameter Mat-Tek Corporation) or home-
made, but always use no. 1.5 for coverslip thickness to opti-
mize image quality (see Note 1).
8. Petri dishes for cell adhesion.
9. Flasks for cell culture.
2.3 Media 1. Complete medium: RPMI 1640 supplemented with Glutamine

(100 mM), nonessential aminoacids, Hepes (25 mM), FCS
(Fetal calf serum; 10%), β-mercaptoethanol (1 mM; only for
mouse cells).
2. Incomplete medium: RPMI 1640, l-Glutamine (100 mM),
nonessential aminoacids, HEPES (25 mM).
3. Wash solution: Hank’s Balanced Salt Medium (HBSS).
4. Isolation wash solution: HBSS, 1% FCS, 1 mM EDTA.
5. Saline solution: NaCl (154 mM).
6. Transfection medium: Optimem I (Gibco-Invitrogen).
7. Lymphocyte separation medium: any commercial media such
as Ficoll Histopaque.
8. Coating buffer: Bicarbonate-carbonate medium. NaHCO3
(0.1 M), Na2CO3 (0.032 M), pH: 9.6.
9. Imaging medium: HBSS, 25 mM Hepes (pH: 7.4), 1% FCS.
10. Lysis buffer: 50mM Tris–HCl (pH 7.4), 1% NP40, 0.2%
Triton X-100, 150 mM NaCl, 2 mM EDTA, 1.5 mM MgCl2
and phosphatase and protease inhibitors.
11. TBS (Tris-buffered saline): Tris-HCl 50 mM (pH: 7.4), NaCl
(154 mM).
12. PHEM (2×): 120 mM Pipes, 50 mM Hepes, 20 mM EGTA, 4

mM MgCl2; pH 6.9.
13. Fixation solution: PHEM (1×), 4% paraformaldehyde (PFA),
0.12 M sucrose.
14. Immunofluorescence (IF) blocking solution: PHEM (1×),
bovin serum albumine (BSA) 3%, human γ-globulin 100 μg/
ml, sodium azide 0.2% (see Note 2).
15. IF Blocking and permeabilizing solution: PHEM (1×), BSA
3%, human γ-globulin 100 μg per ml, sodium azide 0.2%, 0.2%
Triton Tx-100.
2.4 Equipment 1. Confocal imaging: TCS SP5 confocal laser scanning unit with
spectral detection and resonant scanner, attached to an inverted
epifluorescence microscope (DMI6000) fitted with an HCX
PL APO 63× 1.40NA −0.6 oil objective (see Note 3).
2. TIRFm imaging: Leica AM TIRF MC M system mounted on
a Leica DMI 6000B microscope coupled to an Andor-DU8285
VP-4094 camera and fitted with a HCX PL APO 100× 1.46
NA oil objective.
3. Microscopes are mounted into microscope environmental
chamber with heat (Temperature regulator TempControl-37-2
digital) and humidity and CO2 gas controllers (CTI-Controller
3700 digital); in particular, TIRFm from Leica microsystems is
coupled to a BLX incubator for warm air and CO2 control.
2.5 Software 1. LAS-AF 2.6.0. Build 7266 for image acquisition.

2. Imaris software 7.2.2 (Bitplane) for image analysis.
3 Methods
3.1 Purification 1. Isolate the PBMLs from Buffy coat preparations (450 ml of
of CD4+ T Cells peripheral blood from normal healthy human donors) or from
from Human PBLs complete blood (50–200 ml) through a Ficoll Histopaque
gradient.
2. Once the cells are recovered from the interphase with the
Ficoll, wash them with saline solution four to six times to drain
the platelets.
3. Purify CD3+CD4+ cells by magnetic beads-based, negative
selection and afterward transfect (see Subheading 3.3). A
cocktail of antibodies and Streptavidin-conjugated beads for
Automacs is recommended (Miltenyi Biotech).
3.2 Generation 1. Isolate the PBMLs from Buffy coat preparations (450 ml
of SEE-Specific peripheral blood from normal healthy human donor) or from
Lymphoblasts complete blood (50–200 ml) through a Ficoll Histopaque
from Human PBLs gradient.
2. Once the cells are recovered from the interphase with the
Ficoll, wash them with saline solution four to six times to drain
the platelets.
3. Deplete monocytes and granulocytes by plate adhesion in
complete medium (two rounds at least) (see Note 4).
4. Count cells and plate them at 2 × 106 per ml in complete
medium.
5. Add SEE (0.01 μg/ml) and incubate for 48–72 h (see Note 5).
6. Spin the cells (1000 × g) and grow them in complete medium
+ IL2 (20–50 Units/ml). Add IL2 every 2 days (approx).
7. A week later, SEE-treated cells can be restimulated with SEE
(0.1 μg/ml) and PHA (0.4 μg/ml) for 18–24 h. Then spin
them and grow in complete medium with IL2 as above. Wait
for 18–24 h before using them again. The percentage of Vβ8+
cells could then be approximately 40–60%, as measured by
flow cytometry (anti-Vβ8-FITC, BD Biosciences, see Note 6).
8. Restimulation with SEE and PHA may be performed every 15
days. SEE-lymphoblasts may be frozen (107/ml). After cryo-
genization, an effective transfection is more difficult, but they
can be restimulated before transfection (see Note 7).
3.3 Transfection 1. Pre-warm medium and cell culture plates.

of T Cells 2. Count the cells. Use 107 cells maximum per electroporation.
3.3.1 Amaxa 3. Spin the cells and discard supernatant.
Nucleofection (Human 4. Wash cells in cold HBSS or saline solution twice.
SEE-Specific T
5. Aspirate all the washing media and add the transfection solu-
Lymphoblasts)
tion (VPA-1002) with the DNA or RNA included (pre-warm
the mix) (see Note 8).
6. Put the cell mix into the pre-warmed cuvette to be nucleo-
fected immediately.
7. Use programme T23 (Nucleofector I-Amaxa, see Note 9).
8. Recover the cells at once and plate them in incomplete medium
(complete medium without FCS and antibiotics) for 4–6 h at
2 × 106 cells/ml. Then add FCS (10%) and IL2 (20 U/ml).
Do not leave the cells with the transfection medium for more
than 3–5 min.
3.3.2 Amaxa 1. Pre-warm medium and cell culture plates.

Nucleofection (Mouse 2. Count the cells. Use 5 × 106 cells maximum per nucleofection.
CD4+ T Cells)
3. Spin the cells and discard supernatant.
4. Wash cells in cold HBSS or saline solution twice.
5. Aspirate all the washing media and add 100 μl of the transfec-
tion solution (Optimem I) with the DNA or RNA included
(pre-warm the mix) (see Note 8).
6. Put the cell mix into the pre-warmed cuvette and nucleofect
immediately.
7. Use Programme X-01 (Nucleofector I-Amaxa, see Note 9).
Recover the cells at once and plate them in complete medium
at 4 × 106 cells/ml. Then, add IL7 (5 ng/ml).
3.3.3 Electroporation 1. Count the cells. Use 10–20 × 106 cells per sample.
(Cell Lines and Human 2. Spin the cells and discard supernatant.
SEE-Specific T
Lymphoblasts)
3. Wash cells twice with cold HBSS.
4. Wash cells once with cold Optimem I.
5. Resuspend in 400 μL of Optimem I with the DNA or RNA.
Cuvette of 0.4 ml. Do not use more than 10 μl of DNA or
RNA. Cells may be stored at 4 °C with Optimem I and the
DNA/RNA until transfection (not more than 30 min).
6. Electroporation: 240 V, 975 mΩ (usual time: 27.5–29 ms in a
GenePulser II (Bio-Rad)).
7. Recover the cells and plate them in incomplete medium for
4–6 h at 2 × 106 cell/ml.
8. Add FCS and IL2. FCS can be added at 5% for the first 18–24
h and then supplemented to 10%.
9. Plate them immediately upon electroporation. Up to ten sam-
ples may be electroporated at once.
3.4 Preparation of B 1. Count the cells. Use the appropriate number of cells for each
Lymphoblastoid condition. Use 0.4 × 106, 0.15 × 106, and 0.5 × 106 for each
as Antigen Presenting condition for cell lysis and immunoblotting (IB), immunofluo-
Cells rescence (IF), and confocal live cells imaging protocols,
respectively.
2. Spin the cell culture and discard supernatant.
3. Wash cells with HBSS.
4. Separate B cells in two pools; those without antigen will be
used as negative control for IS formation. For IB, the control
pool is only one sample per T cell condition (corresponds to
no stimulation or time zero). For IF and confocal live cell
imaging, the control and antigen-preloaded pools are similar.
5. For IB, preload the required number of Raji cells with SEE
(0.3 μg/ml) or Hom2 cells with SEB (5 μg/ml) or HA
peptide (200 μg/ml) in 400 μl of complete medium for
30 min at 37 °C. To preload Hom2 cells with HA peptide,
increase the time of incubation to 2 h at 37 °C. Do not use
a cell concentration larger than 10 7/ml (see Note 10).
Wash the cells twice with HBSS to exclude excess of anti-
gen and resuspend in complete medium (4 × 10 6/ml). Use
100 μl for each time condition.
6. For IF, preload B cells with a cell tracker such as CMAC (1

μM) in incomplete medium simultaneously with SEE or SEB/
HA peptide (Raji or Hom2, respectively); after 30 min of incu-
bation, add FCS (5%) to the samples (see Note 11). Wash the
cells twice with HBSS to exclude excess of antigen and resus-
pend in complete medium (3 × 106/ml). Use 50 μl for each
coverslip preparation.
7. For confocal live imaging and MTOC localization, preload B
cells with a cell tracker such as CMAC (1 μM) in incomplete
medium; after 30 min of incubation, add FCS (5%) to the sam-
ples (see Note 11). To avoid endocytosis of the stimulus, a first
aliquot of APCs can be preloaded simultaneously with SEE or
SEB/HA peptide (Raji or Hom2, respectively), and keep the
rest only with CMAC labeling. Wash the cells twice with HBSS
to exclude excess of antigen and resuspend in complete
medium (107/ml). Use 50 μl for each coverslip imaged.
3.5 Formation Conjugate Raji or Hom2 B cells with Jurkat or CH7C17 T cells,
of Conjugates respectively, at 1:5 ratio.
and Immunoblotting
3.5.1 Preparation 1. Count the cells. Use 2 × 106 cells for each sample.
of Jurkat, CH7C17 T Cells, 2. Spin cell culture; decant cell culture medium and wash cells
or Human SEE-Specific T twice with HBSS.
Lymphoblast Cells
3. Resuspend cells at 107/ml of complete medium; 200 μl will be
used for each sample.
3.5.2 Preload of Raji 1. Use 100 μl of a 4 × 106/ml cell suspension in complete medium
and Hom2 B Cells per condition.
with Antigen (See Step 5
of Subheading 3.4)
3.5.3 Formation 1. Mix T and B cells and centrifuge at low speed to facilitate the
of Conjugates formation of conjugates.
2. Incubate cells for the required time conditions at 37 °C.
3. Stop activation by incubation of cells at 4 °C; spin the cells for
5 min at 4 °C (1000 × g) to collect cells. Discard supernatants.
3.5.4 Lysis and IB 1. Gently resuspend cells in lysis buffer (50 μl/106 cells).
2. Incubate for 20 min at 4 °C.
3. Spin lysates at 21,000 × g for 10 min at 4 °C to remove debris
and nuclei.
4. Remove the supernatant and place it in a clean tube. Mix it
with Laemmli solution and β-mercaptoethanol (final concen-
tration 0.15 M).
5. Boil samples for 5 min at 95 °C.

6. Separate proteins by SDS–PAGE and perform wet electro-
transfer for IB with nitrocellulose membranes.
7. Block membranes with TBS containing 0.2% TWEEN and 5%
BSA.
8. Blot membranes with primary antibodies (o/n at 4 °C) and
peroxidase-conjugated corresponding secondary antibodies
(30 min). Wash with TBS containing 0.2% Tween at least
three to four times each antibody. Detection of chemilumines-
cence signal may be performed with different imaging systems
(see Note 12).
3.6 Coating of 1. Mix anti-CD3 and anti-CD28 (3:1 ratio) antibodies in coating
Coverslips and buffer. Add ICAM1-Fc (1 μg/ml).
Coverslip-Bottom 2. Pipette 50–100 μl of antibody mix per coverslip-bottom dish
Dishes and incubate o/n at 4 °C.
3.6.1 Preparation 3. Wash twice with washing buffer.
of Stimulating Surfaces 4. Pre-warm at 37 °C in imaging medium before use. Do not
allow to dry.
3.6.2 Preparation 1. Incubate Poly-l-Lys hydrobromide in sterile water (50 μg/ml).

of poly-l-Lysine-Coated 2. Pipette 50–100 μl of the mix per coverslip and incubate o/n at
Surfaces 4 °C.
3. Wash twice with sterile water.
4. Store in HBSS until used. They can be frozen at this step. Do
not allow to dry.
3.6.3 Preparation 1. Incubate fibronectin in incomplete medium (10 μg/ml).

of Fibronectin-Coated 2. Pipette 50–100 μl of the mix per 13 mm coverslip or coverslip-
Surfaces bottom dish or 300 μl for 35 mm coverslip and incubate o/n
at 4 °C.
3. Wash twice with HBSS or incomplete medium.
4. Store in HBSS until used. They can be frozen at this step. Do
not allow to dry.
3.7 Formation 1. Count the cells. Use 0.15 × 106 for each coverslip.
of Conjugates 2. Spin the cell culture, decant the cell medium.
and Immuno
3. Resuspend in complete medium (3 × 106/ml). Use 50 μl for
fluorescence
each sample.
3.7.1 Preparation
of T Cells
3.7.2 Preparation of B 1. Use 50 μl of a 3 × 106/ml cell suspension in complete medium

Cells (See Step 6 per condition.
of Subheading 3.4)
3.7.3 Formation 1. Wash poly-l-Lysine-coated coverslip with 150 μl of HBSS and

of Conjugates aspirate.
2. Immediately, conjugate Raji or Hom2 B cells (APCs) with
Jurkat or CH7C17 T cells in a ratio 1:1, respectively. In par-
ticular, mix 50 μl of the T cell solution with 50 μl of B cell
solution directly onto the coverslip, to favor the formation of
conjugates.
3. Incubate cell conjugate for 30 min at 37 °C.
4. Carefully, stop the activation by adding 100 μl of fixation buf-
fer over the 100 μl cell mix drop by drop onto the coverslip to
avoid the separation of the cells. Let the fixative act for 10 min
at room temperature (RT).
3.7.4 Immunostaining 1. Permeabilize and fix the cells for 5 min at RT with a mix of
of MTOC 50% fixation buffer + 50% IF blocking buffer and 0.2% Triton
X-100.
2. Block the cells with the blocking and permeabilizing buffer for
30 min at RT.
3. Add the primary antibodies diluted in the blocking and per-
meabilizing buffer o/n at 4 °C or 1 h at 37 °C depending on
the antibody affinity. For MTOC staining, add an anti-tubulin
antibody (e.g., DM1A clone from Sigma).
4. Wash the cells 5 × 3 min with TBS.
5. Add the secondary antibody for the corresponding species
diluted in the blocking and permeabilizing buffer for 30 min at
37 °C. Avoid species cross-reactivity and fluorescence dyes
overlapping.
6. Wash the cells 5 × 3 min with TBS. Proceed with a final wash
with distilled water.
7. Dry the water drop of the coverslip and mount the coverslip by
adding 8–10 μl of mounting medium (e.g., Mowiol or Prolong)
to the coverslip and putting the coverslip over the microscope
slide. Be careful to avoid air bubble formation.
3.8 Analysis This method allows the measurement of the distance from the
of MTOC Translocation MTOC to the contact area with the APC in a 3D system. This is an
with IMARIS Software objective manner to measure MTOC translocation, since it takes
(See Fig. 1) into account the 3D localization of the MTOC through the a nalysis
of a confocal XYZ-stack. By measuring the distance to the contact
area along the volume of the APC, small changes in MTOC trans-
location toward the IS can be detected.
3.8.1 Preparation 1. Image cells at confocal microscope after labeling of samples

of Images prepared as described in Subheading 3.7 with a marker for the
MTOC (e.g., Tubulin), a cell tracker for APC volume (CMAC)
Fig. 1 Imaris analysis of MTOC translocation. (a) Initial confocal image of T cells and APC conjugates. Z stack
with staining of α-tubulin (green), F-actin (red), and CMAC for the APC (cyan). (b) Generation of the APC mask
(Surface) using CMAC as reference (cyan). (c) Creation of the MTOC mask (Spots) manually selected (yellow
spheres). (d) Measurement of the distance between the MTOC masks and the APC mask in a 3D system. Scale
bar 5 μm
and a marker for proper immune synapse formation (e.g.,

F-actin or CD3 accumulation).
2. Take a thick Z stack of the cell by imaging some additional
slices over both ends of the APC.
3. Image the samples with maximal pixel resolution possible.
3.8.2 Generation 1. Choose the “Surface” tool to generate a volume. Select auto-
of an APC Mask matic creation and indicate the channel associated with the
with IMARIS Software APC volume.
(Bitplane) 2. Establish the values of “Smooth,” “Absolute Intensity,” and
“Background (Bg) subtraction” parameters to generate an
appropriate mask.
3. Go to next step. Once the histogram of masks is generated,
remove the surfaces that are too small to correspond to any
APC.
4. Individual surfaces can also be eliminated by selecting the
“Pencil tool” and pressing the chosen surface + Shift. APCs that
are not in contact with any T cell or those that are not generat-
ing a proper conjugate (by using the channel of the IS marker,
e.g., actin or CD3) can be removed.
5. Go to the last step and save results.
3.8.3 Creation 1. Select in tools “Distance transformation” and press OK.

of the Distance Channel 2. Choose “Outside surface object” and press OK.
3. A new channel should have been created named “Distance
channel.”
3.8.4 Generation 1. Select the “Spots” tool. Select manual creation and indicate the
of the MTOC Mask channel associated with the MTOC specific channel (tubulin).
2. In order to select the different MTOCs, shift between the
“Select” tool to select the MTOC of a cell and the “Navigate”
tool to move through the image. Use the scroll wheel to adjust
the volume of the mask to the size of all the MTOCs. The
mask should be the same size for the different MTOCs.
3. Select the “Center Point” tool to automatically set the center
of the MTOC mask in the point of maximal intensity in the
tubulin channel.
3.8.5 Generation 1. Select “Statistics” and then “Intensity Mean” of the channel
of the Distance Statistics corresponding to the “Distance channel.” This will measure
the distance of MTOC mask to the closer point of the APC
mask.
2. Export results as an Excel or Txt file to generate graphs.
3.9 Live Imaging 1. Transfect T cells with Tubulin-mCherry or Ensconsin-GFP

MTOC Translocation plasmids 24 h before imaging. Both proteins are localized at
Studies the centrosome and stain the MTOC very well, allowing a very
good imaging of its dynamics.
3.9.1 Preparation
of T Cells 2. Count the cells. Use 0.5 × 106 cells for each video.
3. Spin the cell culture and wash cells twice with HBSS.
4. Resuspend T cells in imaging media (107/ml).
3.9.2 Preloading of B 1. To perform different videos with fresh antigen-pulsed B cells,
Cells with Antigens (See load B cells with the corresponding stimulus at different times
Step 7 of Subheading 3.4) for imaging as described in step 7 of Subheading 3.4. Use 50
μl of a 107/ml cell suspension in complete medium per
condition.
3.9.3 Conjugate 1. Pre-warm the microscope chamber at 37 °C and 5% CO2.

Formation and Microscope 2. Wash the excess of fibronectin with imaging buffer and mount
Conditions the 35 mm coverslip onto the microscope ring adapter. Add
300 μl of imaging medium to the chamber formed.
3. Put a drop of immersion oil onto the 63× objective and place
the coverslip with the adapter over it.
4. Gently add the T cells as little drops all over the imaging sur-
face (50 μl per coverslip). Let them settle for 5 min.
5. Localize transfected cells and set the microscope conditions
before adding the APCs. Take the images at 512 × 512 pixel
resolution and 400 Hz or at 1024 × 1024 pixel resolution
and 1000 Hz of scanning speed with minimal zoom to image
several cells per field. Establish the Z stack range as narrow
as possible to image the MTOC and the cell some microns
over and below it. Capture the different channels at the same
time by using the “between lines” scaning mode instead of
the “between frames” mode. Try to reduce the acquisition
time below 30 s and set the total imaging time to 30–45 min.
Try to reduce the laser power as low as possible to avoid

photo-bleaching and cell damage.
6. Add the APCs (50 μl per coverslip) carefully, drop by drop. Try
to add them just above the objective position where the imag-
ing is going to be performed.
7. Quickly, but carefully, close the stage and start the imaging (see
Note 13).
8. During imaging, pulse a different aliquot of CMAC-preloaded
APC with the corresponding stimulus. For imaging the forma-
tion of other conjugates, mount a new fibronectin-coated cov-
erslip onto the ring adapter and add new T cells from the stock.
Use the freshly pulsed APC to form new synapses.
3.10 Imaging The growth of microtubules can be studied in live cells through
of +Tips in Live T Cells the observation of the incorporation of proteins involved in the
by TIRF Microscopy process to the end of microtubules. EB3-GFP incorporates and
decorates the end of microtubules, since it accumulates at this
position [3]. There, it helps the incorporation of heterodimers of
αβ-tubulin into the microtubule. Tracking of EB3-GFP-decorated
+end of microtubules (+tips) allows study of microtubule
dynamics.
3.10.1 Preparation 1. Collect transfected cells expressing EB3-GFP, spin (500 × g),
of Cells and resuspend in 2 ml of HBSS. Add 1 ml of Ficoll to the
bottom of the tube and spin cells for 5 min (1200 × g) with-
out brake at RT. Recover the live cells from the interface with
the Ficoll.
2. Wash cells twice with HBSS and resuspend in imaging medium
(106 cells/ml). Pre-warm at 37 °C and 5% CO2 until used (see
Note 14).
3. Pre-warm anti-CD3+ anti-CD28 antibody-coated dishes with
2 ml of imaging medium.
3.10.2 Image Acquisition 1. Pre-warm TIRFm stage at least 4 h before image acquisition at
37 °C. Adjust CO2 to 5% and humidity of the stage. Immersion
media must be also pre-warmed. Put a drop of it onto the
objective (100×; 1.46 NA) and introduce the coverslip-bot-
tom dish in the stage. Focus and align the laser beam with the
coverslip.
2. Add 20 μl of cells to the dish, localize the transfected ones with
the oculars, and place them at the center of the imaging area.
3. Set the laser power needed and the best angle for imaging a
homogeneous evanescent field.
4. Upon adhesion to the stimulating surface, MTOC transloca-
tion takes about 1–2 min. If MTOC is not correctly imaged, few
EB3-GFP decorated +tips will be imaged. Search for another

cell or start again (see Note 15).
5. Acquire an initial image including EB3-GFP fluorescence and
bright field for localization of the cell margins and MTOC.
The lamella indicating cell adhesion can be observed in the
bright field image.
6. Time-lapse for correct EB3-GFP tracking range about 300 ms
for human T cells and 200 ms for mouse T cells. Low laser
power is recommended. Recommended laser penetration is
150 nm to detect MTOC; 200 nm can also be used. Changes
in localization of MTOC in mouse T cells are often observed
during recording. Human T cells usually stabilize their MTOC
at shorter times.
7. Acquire time-lapse for 4–5 min.
8. Acquire a final image including EB3-GFP fluorescence and
bright field for localization of the cell margins and MTOC.
3.10.3 Image Analysis 1. Open the file from TIRFm with Imaris software. Fluorescence
(See Fig. 2) image as in Fig. 2a will be observed. Crop time and XY dimen-
sions to analyze only the desired time range and cell of interest.
Usually, 30 s to 2 min for tracking is sufficient to have repro-
ducible results. Create a new channel for EB3-GFP tracking
with the “Surface” tool. Select the manual adjust of parameters
and the “Track surfaces over time” option.
2. Measure the dimensions of the detected EB3-GFP-decorated
+tips. Select the “Smooth” option and use the larger diameter
to indicate the maximal size of the object that fits into the sur-
faces to be detected in the “Background subtraction” option.
The “Surface area detail” is usually one-tenth of the maximal
diameter.
3. Adjust the threshold for background subtraction to make the
surfaces detected from the fluorescence of your particles. The
number of voxels determines the minimal size of the particles
to be analyzed (usually 3–5).
4. Select the “connected components” algorithm to calculate the
tracks for detected surfaces. Indicate the minimal duration for
calculation. The resulting surface and track calculation should
be similar to that observed in Fig. 2a.
5. Select the MTOC and split it from the tracks. It connects sev-
eral tracks at a time.
6. At this point, the statistics calculated by the program contain
all the tracks detected, even if they are branched (Fig. 2b and
Table 1).
7. Some of the tracks will need manual separation of surfaces that
have been detected as a unique fluorescent object.
Fig. 2 Tracking of +tips from TIRFm imaging. (a) Fluorescence image from a time-lapse of an EB3-GFP
expressing T cell. (b–e) Set of images from different steps of the surface generation from fluorescence time-
lapse images and tracks. (b) Initial surfaces and tracks including MTOC. (c) Surfaces and track excluding
MTOC. (d) Processed surfaces and tracks excluding MTOC, peripheral tracks and tracks <2 μm. (e) Tracks
from (d). Bar, 2 μm. Analysis corresponds to 30 s from the acquired time-lapse. (f). Graphs showing the indi-
cated information from the analysis of tracks from the cell shown in (c) excluding the MTOC. (g). Graphs show-
ing the indicated information from the analysis of tracks from the cell shown in (e). Duration is expressed in s,
length in μm, and speed in μm.s−1
8. Select the tracks that arise from the MTOC and that have more
than 1 or 2 μm for mouse or human T cells, respectively, to
avoid +tips of microtubules that are out of focus. Tracks that
pertain to +tips that started or finished their movement before
or after the time range analyzed, respectively, can be ignored
Table 1
Parameters for +tip tracking from EB3-GFP expressing cells through TIRFm imaging. Image analysis
and parameters were generated with Imaris software and statistics analyzed with Excel
Displacement Length Displacement Speed

Sample N Value (μm) (μm) Duration (s) speed (μm.s−1) (μm.s−1)
Initial tracks 240 Mean 1.53 2.70 8.64 0.17 0.32
detected Median 0.86 1.83 6.00 0.17 0.30
SD 1.65 2.61 8.47 0.09 0.09
Tracks emerging 90 Mean 5.18 3.11 16.27 0.33 0.20
from the Median 4.83 2.90 14.38 0.33 0.20
MTOC (2 μm) SD 2.26 1.50 8.08 0.06 0.05
for calculation of track length, but they can serve to calculate

the number of microtubules growing from the centrosome per
time or the speed of growing. Indeed, tracks at the periphery
of the immune synapse-like can also be neglected. Statistics
calculated at this point are depicted in Fig. 2c and Table 1.
3.11 Imaging 1. Collect and prepare transfected cells as for TIRFm imaging.
of +Tips in Live T Cells 2. Pre-warm anti-CD3+ anti-CD28-coated dishes with 2 ml of
by Confocal imaging medium.
Microscopy
3.11.1 Preparation
of Cells
3.11.2 Image Acquisition 1. Pre-warm confocal stage for at least 1–2 h before image acqui-
sition at 37 °C. Adjust CO2 to 5% and humidity of the stage.
Immersion media must be also pre-warmed. Put a drop of it
onto the objective (63×; 1.4 NA; maximal zoom to be used is
3×) and introduce the coverslip-bottom dish in the stage. Add
20 μl of cells to the dish, localize the transfected ones with the
oculars, and place them at the center of the imaging area.
2. To allow high speed scanning, use a specific device such as the
resonant scanner (8000 Hz) and the bidirectional scanning
mode. Use 1024 × 512 px resolution to reduce acquisition
time. Use the “between lines” mode of scanning for the acqui-
sition of different fluorophores.
3. Acquire EB3-GFP fluorescence and bright field at same time.
Include the whole cell in the Z stack. Take images every 0.25–
0.5 μm. This will allow a frame-lapse of 1 s to 1.2 s.
4. Acquire time-lapse for 2–3 min.
3.11.3 Image Analysis 1. Image analysis is similar to that of the +tips from TIRFm
(See Fig. 3) imaging for tracking, but for confocal Z stack time-lapse a
background subtraction and a Gaussian filtering is needed to
Fig. 3 Analysis of +tips from confocal Z stacks imaging. (a) Fluorescence and (b)
bright field images (BF) from a Z-projection from a single time of an EB3-GFP
expressing T cell. (c–e) Set of images from different steps of the surface genera-
tion from fluorescence time-lapse images and tracks. (c) Surfaces and tracks
including MTOC. (d) Surfaces and tracks excluding MTOC. (e) Surfaces for EB3-
GFP fluorescence included in +tips. (f) Surface for EB3-GFP total fluorescence in
whole cell. Bar, 5 μm
clearly detect the objects. Resonant scanning does not allow

correctly calculating mean fluorescent intensities. Therefore,
ratios of fluorescence must be established with respect to the
total fluorescence for the whole cell. A ratiometric analysis of
the incorporated fluorescence in +tips vs the whole cell indi-
cates the amount of microtubule polymerization.
2. Open the file from the confocal microscope with Imaris soft-
ware. Fluorescence image as in Fig. 3a will be observed. Crop
time and XY dimensions to analyze only the desired time range
and cell of interest. Usually, 30 s to 2 min for tracking is enough
to have reproducible results. Create a new channel for EB3-
GFP tracking with the “Surface” tool as described above for
TIRFm imaging (Fig. 3).
3. For calculation of cell surface, select the Surface’ tool, but do
not track surfaces over time. At the “Smooth” option, the value
for the surface area detail is the same as the one used before for
+tips-generated surface, but the cell major diameter is to be
used as the value for the maximal size of the object that fits into
the surfaces to be detected in the “Background subtraction”
option (10–20 μm). The minimal number of voxels should be
10 or more to avoid little surfaces. Delete those surfaces that
do not represent the cell volume manually (Fig. 3).
4. For statistics, use the sum of intensities from +tips-generated
surfaces and from whole cell-generated surfaces to calculate
the ratio of incorporated fluorescence into +tips as a measure
for microtubule polymerization.
4 Notes
1. Use 10 or 13 mm diameter coverslips to reduce the amount of

antibodies or substrate used.
2. Human γ-globulins are used to block any unspecific cross-
reaction through lateral interactions of antibodies with other
molecules or with Fc receptors present in the cell sample for
immunofluorescence protocols.
3. Confocal imaging with a 63× objective with a high NA is rec-
ommended for better resolution. The use of a 100× objective
will improve the resolution of the image, but will absolutely
need a stronger signal from fluorochromes to be detected.
Therefore, using the optical zoom available from your system
can be a better option. A 100× objective with a high NA is a
required objective to acquire images by TIRFm with high
resolution.
4. Do not use antibody and magnetic beads isolation to purify
specific populations at this point. The small proportion of
APCs will be used to present the SEE to CD3+ positive cells

(SEE in solution favors apoptosis).
5. SEE binds to the extracellular β2 loop of the MHC class II and
specifically to the extracellular Vβ8 region on the TCR, acting
as a clamp between the APC and the T cell to activate the T
cell in a specific manner.
6. Cells will show their usual polarized shape on the 5th day from
the first stimulation. Extent of polarization depends mostly on
the donor (and of course on the treatment). The more polar-
ized cells you observe, the more Vβ8+ CD3+ cells there are in
the cell culture. The percentage may be about 20% of cells at
this stage.
7. Cells may be nucleofected every time with the Amaxa system
(human T cell kit, programme T23) or electroporated in
Optimem medium (240 V, 975 mΩ, GenePulser II, Bio-Rad)
after the first restimulation, with either DNA or RNA. Cells
can be purified by magnetic beads-based, negative selection
and afterward transfected to use only CD3+CD4+ (mostly of
them Vβ8+) (see Notes 5 and 6).
8. Avoid using more than 5 μl of DNA or RNA in 100 μl of mix.
RNA is usually used at 1–4 μM. Best results: 1.5–2.5 μl in 100
μl of Amaxa mix. Frequently, about 50% of cells die due to the
transfection procedure. The rest of them divide correctly
(unless your construct or knock-down affects the normal cell
cycle).
9. Nucleofect and plate a single cell mix at a time to improve cell
survival.
10. Raji B cells express the MHC class II specific for Jurkat E1–6
T cells, whereas Hom2 B cells express the MHC-II for
CH7C17 cells. The TCR expressed by E1–6 T cells presents a
Vβ8 region that binds to SEE. Indeed, CH7C17 cells are a
β-deficient, E1–6 cells-derived clone (13–31) reconstituted
with a transgenic αβ-TCR that specifically recognizes the HA
peptide. This TCR presents a Vβ3 region, specific for SEB.
11. Serum esterases from FCS will cut off the ester residues of
CMAC that allow its entry into cells.
12. ImageQuant LAS-4000 chemiluminescence and fluorescence
imaging system (Fujifilm).
13. Let it image for 30 min at least, because usually it takes some
time for the APC to contact the T cell. Once the conjugate
forms, it takes between 2 and 10 min for MTOC
translocation.
14. Cells can be sorted before imaging if the transfection efficiency
is very low, to improve timing and microscope use.
15. Detection of fluorescence through TIRFm imaging is only

possible very near the objective, where there is effective eva-
nescent field. The working distance range is between 70 and
300 nm. If MTOC is not close to the glass surface, its fluores-
cence and the corresponding fluorescence from microtubules
emerging from it will not be detected.
Acknowledgments
We thank Dr. Miguel Vicente Manzanares for critical reading of

the manuscript. Optical microscopy experimentation has been
conducted at the Microscopy & Dynamic Imaging Unit of the
CNIC (Centro Nacional de Investigaciones Cardiovasculares) and
at the Microscopy Facility of the IIS-IP (Instituto Investigación
Sanitaria-Instituto Princesa), Madrid, Spain. This study was sup-
ported by grants SAF2014-55579-R from the Spanish Ministry of
Economy and Competitiveness, INDISNET-S2011/BMD-2332
from the Comunidad de Madrid, ERC-2011-AdG 294340-
GENTRIS Red Cardiovascular and RD 12-0042-0056 from
Instituto Salud Carlos III (ISCIII). The Centro Nacional de
Investigaciones Cardiovasculares (CNIC, Spain) is supported by
the Spanish Ministry of Science and Innovation and the Pro-CNIC
Foundation.
References
1. Dustin ML, Olszowy MW, Holdorf AD, Li J, 6. Serrador JM, Cabrero JR, Sancho D,
Bromley S, Desai N, Widder P, Rosenberger F, Mittelbrunn M, Urzainqui A, Sanchez-Madrid
van der Merwe PA, Allen PM, Shaw AS (1998) F (2004) HDAC6 deacetylase activity links the
A novel adaptor protein orchestrates receptor tubulin cytoskeleton with immune synapse
patterning and cytoskeletal polarity in T-cell organization. Immunity 20:417–428
contacts. Cell 94:667–677 7. Blas-Rus N, Bustos-Morán E, Pérez de Castro
2. Monks CR, Freiberg BA, Kupfer H, Sciaky N, I, de Cárcer G, Borroto A, Camafeita E, Jorge I,
Kupfer A (1998) Three-dimensional segrega- Vázquez J, Alarcón A, Malumbres M, Martín-
tion of supramolecular activation clusters in T Cófreces NB, Sánchez-Madrid F (2016) Aurora
cells. Nature 395:82–86 A drives early signalling and vesicle dynamics
3. Martin-Cofreces NB, Baixauli F, Sanchez- during T-cell activation. Nat Commun 7:11389
Madrid F (2014) Immune synapse: conductor 8. Grigoriev I, Akhmanova A (2010) Microtubule
of orchestrated organelle movement. Trends dynamics at the cell cortex probed by TIRF
Cell Biol 24:61–72 microscopy. Methods Cell Biol 97:91–109
4. Finetti F, Patrussi L, Masi G, Onnis A, Galgano 9. Vinopal S, Cernohorska M, Sulimenko V,
D, Lucherini OM, Pazour GJ, Baldari CT Sulimenko T, Vosecka V, Flemr M, Draberova
(2014) Specific recycling receptors are targeted E, Draber P (2012) gamma-Tubulin 2 nucleates
to the immune synapse by the intraflagellar microtubules and is downregulated in mouse
transport system. J Cell Sci 127:1924–1937 early embryogenesis. PloS One 7(1):e29919
5. Andres-Delgado L, Anton OM, Bartolini F,
10. Faire K, Waterman-Storer CM, Gruber D,
Ruiz-Saenz A, Correas I, Gundersen GG, Masson D, Salmon ED, Bulinski JC (1999)
Alonso MA (2012) INF2 promotes the forma- E-MAP-115 (ensconsin) associates dynamically
tion of detyrosinated microtubules necessary with microtubules in vivo and is not a physio-
for centrosome reorientation in T cells. J Cell logical modulator of microtubule dynamics.
Biol 198:1025–1037 J Cell Sci 112:4243–4255
Chapter 4
Analyzing the Dynamics of Signaling Microclusters

Akiko Hashimoto-Tane, Tadashi Yokosuka, and Takashi Saito
Abstract
T cell antigen receptor (TCR) stimulation induces recruitment and accumulation of various types of signal-
ing molecules and forms signaling microclusters. The dynamics of the microclusters are important for
regulating the quality and quantity of T cell activation. We describe here our protocols for analysis of sig-
naling microclusters by using supported planar bilayers.
Key words TCR, Signal, Phosphorylation, Retroviral expression system, Supported planar bilayers,
Time-lapse imaging, TIRF microscopy, Confocal microscopy
1 Introduction
The in vitro analysis of T cell activation provides considerable

information about intracellular signaling events involved in this
process. Triggering of the T cell receptor (TCR) by peptide-MHC
induces formation of TCR microclusters and clustering of various
signaling molecules to mediate phosphorylation/activation. To
analyze the dynamics of such TCR and signaling microclusters in
living T cells, we have utilized the combination of TIRF fluores-
cence microscopy and a planar bilayer system for live cell imaging.
We choose retrovirus-mediated gene transfer to express fluores-
cently labeled signaling molecules in normal T cells. Because it is
quite difficult to transfect genes into primary T cells by lipofection
or electroporation, virus transduction is a powerful tool for prepar-
ing cells expressing the desired fluorescently labeled molecules.
For the activation of T cells for real-time imaging by micros-
copy, we use a supported planar bilayer system. The planar bilayer
mimics the membrane surface of antigen presenting cells (APC) by
the expression of freely mobile GPI-anchored MHC proteins, inte-
grins, and other co-stimulatory molecules. A supported planar
membrane containing only peptide-MHC and integrins is able to
induce T cell activation. Under these conditions, the i mmunological
synapse is generated on the surface of the T cells at the interface
51
52 Akiko Hashimoto-Tane et al.
with the flat surface of the planar bilayer and this structure can be
easily analyzed, making this the most suitable system for micro-
scopic analysis. To confirm the bilayer data under physiological
conditions with authentic APC, we also perform T cell stimulation
using activated B cells as APC. We present here our protocols for
the preparation of T cell transductants and for live cell imaging by
fluorescence microscopy.
2 Materials
2.1 T Cell Isolation 1. TCR transgenic mice for primary T cells: AND [1] or 5C. C7
[2], both of which are cytochrome c-specific, I-Ek restricted T
cells, DO11.10 [3] or OT-1 [4]. Mice for APC: C3H, B10BR,
BALB/c, or C57BL/6.
2. Medium: RPMI 1640 supplemented with 8.5 μM of
2-Mercaptoethanol (2-ME), 100 mg/L of streptomycin, 57.6
mg/L of penicillin and 10% (v/v) fetal bovine serum (FBS).
3. Glass homogenizer (16 × 160 mm, 5 mL, e.g., Abe Kagaku
Inc.).
4.
Ammonium-chloride-potassium (ACK) buffer: 150 mM
NH4Cl, 10 mM KHCO3, and 0.125 mM EDTA, adjusted to
pH 7.6 with HCl.
5. BioMag Goat Anti-Mouse IgG (Qiagen) or equivalent.
6. Biotinylated mAb cocktail: 2.5 μg/mL of biotin-labeled
CD11b, CD11c, CD45/B220, CD105, Gr.1, I-A/I-E (MHC
II), TER-119/Erythroid, TCR-γδ, ΝΚ1.1, CD19 and
CD24 in RPMI 1640 with 2% FCS.
7. Mojo buffer: 0.5% (w/v) bovine serum albumin (BSA) and 2
mM EDTA in phosphate buffered saline (PBS).
8. Mojo streptavidin beads (BioLegend) or equivalent.
9. Antigen peptide: MCC peptide (88–103) for AND and 5C. C7
tg mice, MCC peptide mutant (K99A) for AND tg mice, OVA
peptide (323–339) for DO11.10 cells, and OVA peptide
(257–264) for OT-I tg mice.
10. IL-2 containing medium: 1.5 or 3.1 ng/mL IL-2 (Peprotech)
in RPMI 1640 with 10% FCS.
11. B220 MACS beads (Miltenyi Biotec) or equivalent.
2.2 Retrovirus 1. Plasmids: we recommend pMX retroviral expression vector

Preparation series [5] (Cell Biolabs); pMX or pMX-IRES-human CD4/8
(with deleted cytoplasmic region).
2. Packaging cells: we recommend Phoenix-ECO (ATCC),
Platinum-E (Plat-E) [6] (Cell Biolabs).
Signaling Microclusters 53
3. Medium: Dulbecco’s modified Eagle’s medium (DMEM) sup-

plemented with 8.5 μM 2-ME, 100 mg/L streptomycin, 57.6
mg/L penicillin, and 10% (v/v) FBS.
4. Medium for lipofection: Opti-MEMs (ThermoFisher) supple-
mented with 50 μM 2-ME and 0.9 mM CaCl2 or equivalent.
5. Lipofection reagent: 2% (w/v) Polyethylenimine “Max” (MW
40,000) (PEImax) (Polyscience) in sterilized water, filtered
and stored at 4 °C or equivalent.
6. Transduction reagent: 10 μg/mL polybrene in sterilized ultra-
pure water.
2.3 Liposome 1. Plasmids: I-Adα-GPI, I-Adβ-GPI, H-2Kb-GPI, CD80-GPI,

Preparation CD86-GPI, or PD-L1-GPI was inserted in the pMX expres-
sion vector [7–9].
2. CHO or BHK cell lines.
3. Agarose beads: Cyanogen bromide activated agarose.
4. Antibodies for affinity purification of proteins: 14–4-4 for I-Ek,
YN1/1 for ICAM-1, MK-D6 for I-Ad, 20–8-4 for H-2Kb,
16-10A1 for CD80, PO3 for CD86, and MIH5 for PD-L1.
5. Coupling buffer: 0.1 M NaHCO3, 0.5 M NaCl pH 8.3–8.5.
6. Tris buffer: 25 mM Tris, 0.15 M NaCl adjusted to pH 8.0 with
HCl.
7. Lysis buffer: 1% (w/v) TitonX-100, 25 mM Tris, 0.15 M
NaCl, protease inhibitor cocktail [0.2% (v/v) aprotinin, 50
μg/mL antipain, 50 μg/mL chymostatin, 20 μg/mL leu-
peptin, 10 μg/mL pepstatin A] pH 8.0.
8. OG wash buffer: 1% (w/v) n-octyl-Beta-D-glucopyranoside
(OG), 25 mM Tris, 0.15 M NaCl, pH 8.0.
9. Elution buffer: 50 mM triethylamine, 0.15 M NaCl, 1% OG
(pH 11.5) or 50 mM Glycine, 0.15 M NaCl, 1% OG (pH 3.0).
10. Neutralizing solution: 1 M Tris pH 7.0, 1% OG.

11.
1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) (e.g.,
Avanti Polar Lipids).
12. Dialysis tubing (14000 MWCO).
13. Argon gas tank and regulator.
2.4 Lipid Bilayer 1. Chromic sulfuric acid mixture.

2. Cover glass, 40 × 50 mm, thickness 0.12–0.17 mm.
3. Micro cover glass, diameter 12 mm, thickness 0.12–0.17 mm.
4. Silicone sheet with a thickness 0.2 mm, stiffness of 20 was cut
into strips ~4 mm × 25 mm.
5. Silica beads 4.86 μm diameter (e.g., Bangs Laboratory).
6. Lab-marker felt tip pen.

7. Loading buffer: 1% (w/v) BSA, 50 μg/mL gentamycin in PBS
adjusted to pH 4.5 with 2.3 M citric acid.
8. Peptide solution: MCC peptide, 1% EDTA, 0.2% (v/v) apro-
tinin and 20 μg/mL leupeptin in loading buffer.
9. Acrylic tubing 35 mm outside and 25 mm inside diameters,
cut to 10 mm height.
10. High Vacuum grease (e.g., Dow Corning Toray).
11. HEPES buffered saline stock solution (10 × HBS): 200 mM
Hepes, 1.37 M NaCl, 50 mM KCl, 7 mM Na2HPO4, 60 mM
D-glucose, 10% (v/v) BSA, adjusted to pH 7.2 with hydro-
chloric acid (HCl).
12. 1× HBS: 10% (v/v) 10× HBS, 1 mM CaCl2, 1 mM MgCl2,
10% (v/v) FBS, prepared when needed.
13. 4% Paraformaldehyde (PFA) in PBS.

14.
Mounting medium: Permafluor, Prolong Gold
(ThermoScientific) or equivalent.
15. Slide glass 76 × 26 mm, thickness 1.2–1.5 mm.
2.5 Cell Staining 1. H57(Fab)-Dy549: Purified Fab fragment from anti-mouse

TCRβ chain mAb (H57) labeled with Dy549.
2. Halo tag TMR ligand (Promega).
3. Permeabilizing buffer for fixed cells: 0.05% Digitonin (stock 1
mg/mL in DMSO) or 0.1% TritonX-100 in PBS containing
2% BSA, 0.02% NaN3.
4. Abs for intracellular staining: Abs were labeled with Alexa
series or DyLight series and diluted in permeabilizing buffer.
2.6 Microscope The information below describes commercial microscope systems

and Software and software we have used successfully. Many other options are
available.
1. A TIRFM system with dual laser (488, 561 nm) (Olympus),
UAPON 100× OTIRF lens (NA1.49, Olympus) and CCD
camera (ORCA-Flash, Hamamatsu photonics), auto-focus
system ZDC (Olympus) were attached to an Olympus IX81
inverted microscope.
2. Confocal microscope: Leica TSC SP5 system with an oil
immersion objective (HCX PLAPO 100×, NA1.44, oil, Leica),
four lasers (405, 488, 561, 633 nm), and a HyD detector.
3. Software: Metamorph (Molecular Devices) for Olympus IX81
inverted microscope, Leica Application Suite (Leica) for Leica
TSC SP5 system, Fiji (ImageJ) (freeware), and Imaris (Bitplane)
for statistical analysis.
3 Methods
3.1 Cell Preparation 1. T cell preparation: Take spleen and lymph nodes from a TCR
transgenic mouse and homogenize in 5 mL RPMI 1640. Treat
the cell suspensions with ACK solution (2 mL/mouse) to lyse
red cells. From the cell suspension, deplete B cells with 2 mL/
mouse of BioMag Goat Anti-Mouse IgG and other non-T cell
populations by staining with a biotinylated-Ab cocktail and
Mojo streptavidin magnetic beads (5 μL/107 cells) and then
sort T cells on a magnetic stand. The purity of the isolated T
cells is between 85 and 95%.
2. APC preparation: Take spleen cells from a mouse for APC and
lyse red cells with ACK solution. The remaining cell suspen-
sion is then irradiated with 15 Gy of gamma radiation and used
for APC. Alternatively, incubate the cells with 10 μg/mL
Mitomycin C for 2 h and wash well.
3. Cell culture: Culture 3.0 × 106 purified T cells with 3.0 × 106
of the APC in 500 μL RPMI 1640 containing antigen peptide
and 1.5 ng/mL IL-2 per well in a 48-well culture plate. We
always prepare two to four wells of T cells for the analysis of
each signaling molecule.
4. B cell preparation: For preparation of B cells as APC, B220
positive cells sorted from spleen by using B220 MACS beads
were cultured at 3.0 × 106 in 600 μL RPMI 1640, adding 10
μg/mL of LPS to activate the cells and pulsing with 100 nM
to 10 μM MCC peptide for 12–24 h.
3.2 Virus 1. Prepare retrovirus vectors for expressing fluorescently labeled

Transduction signaling molecules. The position to insert the fluorescent
molecule should be optimized for each target molecule.
Practically, we always try both N- and C-termini with or with-
out some linker, and choose the best one in terms of function
and/or brightness. For example, some transmembrane mole-
cules like CD3ζ, CD28, and LAT were optimally labeled at the
C-terminus. Certain intracellular molecules such as Lck,
ZAP70, PKCθ, CTLA-4, dynein light chain, Pyk2, and paxillin
were labeled at the C-terminus, but some others such as SLP76
and α-tubulin were at the N-terminus. GFP (pEGFP-N1, e.g.,
Addgene) is our first choice for a fluorescent protein because it
provides sufficiently bright fluorescence and has less effect on
the function or localization of the target protein. Red fluores-
cent proteins are potential candidates as a second color; how-
ever, based on our experience with T cells, red fluorescent
molecules such as mRFP, mCherry, mStrawberry, and Kusabira
orange have a tendency to form artificial bright aggregates.
Although we know that these red proteins work well in other
cell types, they are sometimes problematic, even in the mono-

meric form, in T cells for unknown reasons. Therefore, we are
using a combination of Halo-tag (pHTC HaloTag, Promega)
and TMR ligand or CFP (pECFP-N1, Addgene) as the coun-
ter color for GFP [7–13].
2. Packaging cells: 1 day before lipofection, seed Phoenix cells at
1 × 106/well in a 6-well culture plate. If using Plat-E, plating
3 × 106/well 3 hours before lipofection is also acceptable.
3. Lipofection: Mix 3–10 μg of pMX plasmid with 250 μL of
Opti-MEM. In another polystyrene tube, mix 10 μL PEImax
and 250 μL of Opti-MEM, gently mix by taping, and then
leave for 5 min. Next, add the plasmid solution to the PEImax
solution and wait for 20–30 min. Remove culture supernatant
(sup) from the packaging cells until 500 μL/well remains and
then add the 500 μL plasmid/PEImax solution very gently
dropwise. Immediately, add 1.0 mL DMEM gently, incubate
for 2–4 h, and then add an additional 3.0 mL DMEM.
4. Harvesting Virus: On the next day (day 2), remove culture sup
and add fresh DMEM, 11 mL/well. Harvest the virus-
containing culture sup on day 3 and add fresh media to harvest
again on day 4. Combine the sups in a 50 mL tube and con-
centrate virus by overnight centrifugation, 8,000 × g at 4
°C. After discarding the supernatant by decantation, add 1.0
mL of fresh RPMI 1640 and use for transduction, or freeze at
−80 °C.
5. Viral infection: 24–36 h after initiation of T cell cultures, col-
lect the cells and replate in RPMI 1640 (50 to 100 μL/well)
in the same plate. Add 250–500 μL/well virus sup supple-
mented with 10 μg/mL polybrene and 3.2 ng/mL IL-2 for
each well and centrifuge in a pre-warmed centrifugation at
1,400 × g, 25–32 °C for 2 h. Then add 400 μL/well RPMI
1640 with 3.2 ng/mL IL-2 and incubate for another 2 days.
If you intend to infect two differently labeled target mol-
ecules into the same cells, prepare virus separately and mix
them at the transduction step.
6. Cell sorting: Harvest cells and isolate the fluorescent positive
infected cells with a cell sorter (see Note 1). If you use pMX-
IRES- hCD4/8, you can use magnetic beads to enrich the
infected cells. Incubate the sorted cells for 1 day in RPMI
1640 without IL-2.
3.3 Liposome 1. Prepare the expression construct in the pMX vector engineered
Preparation to express the GPI anchored form of the molecule of interest.
The extracellular domains of I-Adα, I-Adβ, CD80, CD86, or
PD-L1 were fused with a GPI-motif derived from human DAF
and the expression constructs were transfected into CHO or
BHK cells. CHO cells expressing I-Ek-GPI and BHK cells

expressing ICAM1-GPI were kindly provided by Dr. Michael
Dustin [14].
2. Cell culture: Culture the CHO or BHK cells expressing the
GPI-constructs, harvest, wash the cells with PBS, and keep fro-
zen at −80 °C. We usually repeat the cell harvest several times
until reaching a 3.0 g cell pellet to collect a large amount of the
molecules of interest (see Note 2).
3. Affinity purification: Extract GPI-anchored molecules from
the cell pellets by affinity purification. Prepare 1.0 mL agarose
beads coupled with 2.0 mg of mAb specific for the target pro-
tein. Lyse the cells by ultrasonic treatment in Lysis buffer.
Ultracentrifuge the cell lysates containing GPI-anchored mol-
ecules and then preclear using agarose beads coupled with
nonspecific IgG. Incubate the cell lysates overnight at 4 °C
with beads coupled with the specific mAb. After washing the
beads, the GPI-anchored molecules are eluted with elution
buffer, pH 11.5, or pH 3.0 (the optimal pH should be deter-
mined for each target protein). Collect eluates in
450-μL-fractions in tubes containing 50 μL neutralizing solu-
tion. Do not mix the fractions that have different concentra-
tions of eluted GPI-anchored proteins. We use the fractions
whose OD280 is more than 0.08. If the OD280 is less than 0.08,
concentrate using pre-washed centrifugal filter tubes.
4. DOPC preparation: Prepare a 4.0 mM DOPC micelle stock
solution by making a thin film of 6.28 mg DOPC on a glass
bottle by evaporating chloroform, then add 2.0 mL of Tris
buffer with 1% OG, sonicate and filter. Mix the micelle solu-
tion with nine volumes of each purified GPI-anchored protein
fraction, which results in 0.4 mM DOPC, and dialyze against
Tris buffer. We always prepare protein-free 0.4 mM DOPC
liposomes for diluting GPI-anchored proteins in the next step.
5. Quantitation on glass beads: Check the concentration of GPI-
anchored proteins by FACS analysis after staining of reconsti-
tuted glass beads. Put 1.5 μL of glass beads on the bottom of a
PCR tube and let it stand at an angle for 3 min to allow the
beads to sink to the bottom side (Fig. 1a). Remove the fluid
from the beads (Fig. 1b) (do not allow them to dry) and add
3.0 μL of dialyzed liposome solution. Mix and invert the tube,
keeping the fluid as a hanging drop for 30 min (Fig. 1c) with
gentle tapping every 10 min. Wash the liposome-coated beads
gently with HBS, then incubate with RPMI 1640–10% FCS for
blocking. Next, stain the beads with fluorescently labeled mAbs
specific for the GPI-anchored proteins (Fig. 1d). Based on the
fluorescent intensity and the fluorescent-protein (F/P) ratio of
the mAbs, evaluate the amount of GPI-anchored proteins in
the liposomes. Dilute each liposome to a predetermined ideal
Fig. 1 Preparation of lipid bilayers reconstituted on glass beads for quantification of GPI-anchored proteins and
T cell stimulation. (a) Put 1.5 μL of glass beads in the bottom of a PCR tube. (b) Lean the tube for 3 min to allow
the beads to sink to the bottom side and remove the fluid. (c) Add 3.0 μL of liposome solution, mix very gently,
and invert the tube. Incubate for 30 min at room temperature, tapping gently every 10 min. Wash the liposome-
coated beads gently with HBS. (d) Add RPMI 1640-FBS medium to the beads for blocking, and stain with fluo-
rescently labeled mAbs for FACS analysis. (e) Incubate with MCC antigen peptide in loading buffer for 12–24
h and use for T cell simulation
concentration (molecules/μm2) and store at 4 °C in screw-

capped tubes with air replaced by Argon gas. In the case of
MHC-GPI and ICAM-1-GPI, we prepare 800 molecules/μm2
as stock.
3.4 Setting 1. Wash glass: Place glass (40 × 50 mm) and cover glass (diameter
Up the Planer Bilayer 12 mm) in chromic acid for at least 1 day (we routinely store
these items in a chromic acid chamber), rinse with water, 95%
ethanol and ultra-pure water, then air dry.
2. Reconstitution of the lipid bilayer: Mix the MHC (e.g., I-Ek)
and integrin ligand (e.g., ICAM-1) liposomes to make a final
concentration of 200 molecules/μm2 each [14]. Add other co-
stimulatory molecules (e.g., CD80) as needed and incubate for
several minutes at room temperature. Set three silicone sheets
in parallel about 6 mm apart on the 40 × 50 mm glass and put
1.2–1.5 μL of mixed liposome solutions in between them
(Fig. 2a). Place a cover glass on each liposome and confirm
that the liposome drop touches both the slide glass and the
cover glass. We always make two to four spots per glass. Mark
the drop position with a felt tip lab marker and incubate at
room temperature for 30 min, taking special care to prevent
A cover glass B Lab marker C

liposomes loading buffer
silicon glass lipid bilayer

sheet
acrylic tubing
D E F
culture dish grease
paper cell acrylic tubing
napkin HBS
grease
wet filter paper objective
RPMI 1640
Fig. 2 Preparation of planer bilayer for live cell imaging. (a) Set three silicone sheets in parallel approximately
6 mm apart on the glass and put 1.2–1.5 μL of mixed liposomes in between the two sheets. (b) Place cover
glasses on each liposome, mark the position with a felt tip lab marker, and incubate for 30 min at room tem-
perature. (c) Wash each drop by flowing pre-warmed loading buffer between the glass and the cover glass,
and replace the peptide solution. (d) Incubate the glasses on wet filter paper in a culture dish at 37 °C for
24–48 h. (e) Wash the cover glasses with pre-warmed RPMI 1640 and remove silicone sheets from the cover
glasses. Wipe excess liquid with paper towels, apply grease to the acrylic tubing, and pour in pre-warmed HBS.
(f) Set the chamber on a microscope, put cells on the lipid bilayer area, and collect images at the interface
drying (Fig. 2b). Warm loading buffer and peptide solution to

avoid emerging bubbles. After prepareing the spots as recon-
stituted lipid bilayers, flow more than 3.0 mL/spot of loading
buffer between glass and cover glass, and then peptide solu-
tion (Fig. 2c). Take special care to prevent bubbles (see Note
3). Put the glass on wet filter paper in a culture dish and incu-
bate at 37 °C for 24–48 h (Fig. 2d).
3. Setting up chambers: Put grease uniformly on one side of the
cut acrylic tubing (Fig. 2e). Wash the position of the lipid
bilayers with pre-warmed RPMI1640 and remove the cover
glasses from the silicone sheets (see Note 4). Submerge the
cover glasses in pre-warmed HBS in a 12-well culture dish with
the lipid bilayer side up (Fig. 3a). Wipe the area surrounding
the spots with a paper napkin (e.g., Kimwipe) to approximate
the size of acrylic tubing (Fig. 2e). Push the greased side of the
acrylic tubing onto the wiped area of the glass to make an assay
chamber (Fig. 2e). Confirm the absence of air bubbles in the
grease from the bottom side and immediately add pre-warmed
HBS (Fig. 2f ). You can set up many chambers at once and
keep them at 37 °C. For fixed samples, keep the cover glasses
in a 12-well culture dish at 37 °C.
A cell B C
mounting
antibody solution medium
HBS
cover glass parafilm slide glass

with lipid bilayer
D
slide glass
nail polish
mounting medium
cells
objective cover glass
Fig. 3 Analysis of fixed cells on a planar bilayer. (a) Immerse the lipid bilayer-
reconstituted cover glasses in 400 μL of HBS in a 12-well culture dish with the
bilayer side up, then add the T cell suspension dropwise and incubate at 37
°C. Add 800 μL of a PFA solution to fix the cells and then replace with PBS. (b)
Place the cover glass with the cell side up on Parafilm, followed by treatment
with permeabilizing buffer. Stain the cells with 150 μL of the mixed Ab solution
and then wash. (c) Drop 10 μL of mounting medium on the slide glass and put
the cover glass on the medium with the cell side down. (d) Seal the edge with
transparent nail polish and examine on a microscope by setting the cover glass
side to the objective lens
3.5 Imaging of Live T 1. Cell preparation: Collect the sorted GFP- (or other fluorescent
Cells by Microscopy protein) positive T cells as described in step 6 of Subheading
3.2 and resuspend in 1.0–2.0 μL of 10 μg/mL H57
(Fab)-Dy549. When Halo tag is used for the red color, incu-
bate the cells with 1.0 μM TMR ligand for 15 min. Wash the
cells and incubate them in fresh medium for at least 30 min,
then collect and resuspend them in 1.0–2.0 μL HBS at a con-
centration between 2.5 × 104 and 2.5 × 105/μL.
2. Planar bilayer assay on the microscope: Warm the microscope
system. We use a custom-made acrylic box to warm the entire
microscope (see Note 5). Set the assay chamber on the micro-
scope, drop a very small volume of cell suspension on the pla-
nar bilayer area, and adjust focus and laser direction (Fig. 2f ).
The TIRF-M instrument from Olympus has independent laser
illumination paths for 488 and 561 nm. Therefore, the best
angle for each laser can be adjusted. Choose the illumination
conditions and resolution from the viewpoint of image clarity
and bleaching. Put 0.5 μL of the cells on the planar bilayer and
start collecting images. We take one image every 5 s for
10–15 min at a resolution of 99.6 nm/pixel and area of 1024
× 1024 pixels on the TIRF-M [11] (see Note 6).
3. APC stimulation on the microscope: For T cell stimulation by

B cells as APC, prepare a glass bottom dish precoated with 0.5
μg/mL B220 Ab. Incubate the LPS-activated and MCC-
loaded B cells on the glass for at least 20 min, change the
medium to HBS buffer, and move onto the microscope. After
confirming that the B cells are immobilized on the B220
Ab-coated glass, fluorescent-labeled T cells are added. T cells
scan the surface of the B cells and eventually stop, forming a
firm conjugate with the B cells and begin activation. Determine
a good time point and position to analyze molecular dynamics
at the interface [10, 12].
3.6 Imaging of Fixed 1. Planar bilayer: Change the buffer in the 12-well culture plate
T Cells described in step 3 of Subheading 3.4 to 400 μL of fresh pre-
on the Microscope warmed HBS. Drop 1.0–2.0 μL of T cell suspension (~106
cells/μL) on the planar bilayer area and incubate for 1.5 min for
the analysis of the initial activation phase or for 5 min for the
analysis of cSMAC formation (Fig. 3a). To fix the cells, quickly
add 800 μL of pre-warmed 4% PFA solution and keep at 37 °C
for 10–20 min. We prepare many samples at once on the same
culture plate by sequential handling. Remove the PFA solution
by decantation, replace with PBS, and keep at 4 °C.
2. Stimulation with APC: Coat cover glasses with 0.5 μg/mL
B220 Ab and then immerse them in RPMI1640 in a 12-well
culture plate with the Ab-coated side up. Add a drop 2.0 × 106
LPS-activated and MCC-loaded B cells and incubate for at
least 20 min. Change the medium to pre-warmed HBS, and
add 1.0–2.0 μL of T cell suspension (~106 cells/μL) dropwise
onto the immobilized B cell area. The following procedure is
the same as in step 1 of Subheading 3.6.
3. Ab staining: Place the cover glasses with the fixed cells side up
on Parafilm, followed immediately by a drop of the permeabi-
lization buffer and then incubation for 15 min (Fig. 3b). Stain
the cells by replacing the buffer with 150 μL of the mixed mAb
solution and incubate at room temperature for 20 min to sev-
eral hours, depending on the target molecule. Choose the
mAb color combinations according to the microscope settings.
Wash the cover glass with PBS.
4. Mounting on slide glass: Drop 10 μL of mounting medium
onto a slide glass and put a cover glass on the medium with the
cell side down (Fig. 3c). Wait for 30 min and seal the edge with
nail polish (Fig. 3d).
5. Confocal microscopy: Set the objective lens above the cover
glass and detect fluorescent molecules at the interface between
the cell and the cover glass (Fig. 3d). We first adjust the focus
onto the TCR microclusters and then adapt the position to
other signaling molecules.
3.7 Data Analysis Use Fiji (ImageJ) and Imaris for the analysis of fluorescent intensi-
ties, number of microclusters, cell adhesion area, and so on (see
Note 7).
3.8 Bioassay To measure cellular responses under the same stimulation condi-
tions by microscopy, we prepare glass beads as in step 4 of
Subheading 3.2 and incubate with antigen peptide in loading buf-
fer for 12–24 h (Fig. 1e). The beads can be used in the same man-
ner as APC for measurement of cytokine production or cell growth.
4 Notes
1. Infection efficiency is dependent on both virus titer and cell

viability. If the virus titer is low, several points should be
improved: reconstruct the plasmid, functionally select the
packaging cell lines by adding selection drug/antibiotics, or
simply increase virus titer per cell. If the viability of T cells
upon stimulation is low, check TCR expression and increase
the peptide concentration for stimulation. Sometimes, removal
of cell debris derived from irradiated spleen by density gradient
centrifugation increases the frequency of infection.
2. To collect 3.0 g of cell pellet of CHO or BHK cells, we repeat
cell harvesting two to three times from cultures in 30 of
150 mm diameter dishes. To increase the quality of purified
GPI m olecules, it is critical to avoid overgrowth, to wash
medium out before freezing and to tap well after thawing the
cells.
The quality of the planar bilayer is critical for the dynamics
of all stages of IS formation. If the planer bilayer is badly pre-
pared, the area of cell adhesion and movements of signaling
molecules are altered and one often fails to observe the typical
IS, particularly the cSMAC. Some reasons for failure are shorter
incubation time of liposome drop, stains on the surface of the
glass, bilayer broken by air bubbles, and oxidization of lipo-
somes, which prevents motility of the molecules. The motility
of molecules in liposomes can be monitored by fluorescence
recovery after photobleaching (FRAP). If the motility is not
sufficient, it is better to make a new preparation.
3. When setting up the chamber, pay caution to exposure to air.
If the planar bilayer is exposed to air for even a short period, it
will be broken. Although the acrylic tubing we use is nondis-
posable, the grease is ethanol-soluble and easily removed, thus
the acrylic tubing can be reused.
4. We use a TIRF microscope for live cell imaging and a confocal
microscope for fixed cell analysis. The reasons are as follows:
live cell imaging requires fast shutter speed, higher sensitivity,
and relatively less harmful conditions for the target cells.

Confocal microscopy has advantages of many color channels
and application to 3-D reconstitution analysis; however, it is
less sensitive and harmful by laser illumination than the TIRF
system.
5. Temperature strongly affects focus. The problem is serious,
especially for time-lapse imaging by TIRF microscopy. Because
this is due to thermal expansion of metals, we warm the whole
microscope system to keep the temperature constant. In addi-
tion, it is recommended to use an auto-focus system ZDC.
6. If T cells on the planer bilayer keep moving even after stimula-
tion with a higher concentration of antigen peptide, they may
still be in an activated state. We incubate T cells in media with-
out extra IL-2 for 1 day, sometimes longer, before assay to let
them return to the resting state. It is necessary to check the cell
status by cell size and the surface expression of activation
markers.
7. Recently, sophisticated statistical analysis of imaging studies is
often requested by reviewers for publication. In the case of the
analysis of TCR microclusters, the definition of a microcluster
is not fixed. We use criteria that a TCR microcluster is a dot,
which can be fitted by Gaussian fitting and is more than 10%
brighter than a nonclustered area. Although this is not a uni-
versal criterion, the resulting count of our analysis is similar to
other studies.
Acknowledgment
This work was supported by KAKENHI (Grant-in-Aid for Scientific

Research (S) 24229004 for T.S. and (C) 10415226 for A.H.-T.),
and partly by Takeda Science Foundation, Novartis Foundation
(Japan), and Hayashi Memorial Foundation for Female Natural
Scientists.
References
1. Kaye J, Hsu ML, Sauron ME, Jameson SC, 3. Murphy KM, Heimberger AB, Loh DY (1990)
Gascoigne NR, Hedrick SM (1989) Selective Induction by antigen of intrathymic apoptosis
development of CD4+ T cells in transgenic of CD4+CD8+TCRlo thymocytes in vivo.
mice expressing a class II MHC-restricted anti- Science 250:1720–1723
gen receptor. Nature 341:746–749 4. Hogquist KA, Jameson SC, Heath WR,
2. Seder RA, Paul WE, Davis MM, de St. Groth Howard JL, Bevan MJ, Carbone FR (1994) T
BF (1992) The presence of interleukin 4 dur- cell receptor antagonist peptides induce posi-
ing in vitro priming determines the lymphokine- tive selection. Cell 76:17–27
producing potential of CD4+ T cells from T 5. Kitamura T (1998) New experimental
cell receptor transgenic mice. J Exp Med approaches in retrovirus-mediated expression
176:1091–1098 screening. Int J Hematol 67:351–359
6. Morita S, Kojima T, Kitamura T (2000) Plat-E: microclusters are essential for T cell activation.
an efficient and stable system for transient pack- J Exp Med 213(8):1609–1625
aging of retroviruses. Gene Ther 11. Hashimoto-Tane A, Yokosuka T, Sakata-
7:1063–1066 Sogawa K, Sakuma M, Ishihara C, Tokunaga
7. Yokosuka T, Kobayashi W, Sakata-Sogawa K, M, Saito T (2011) Dynein-driven transport of
Takamatsu M, Hashimoto-Tane A, Dustin ML, T cell receptor microclusters regulates immune
Tokunaga M, Saito T (2008) Spatiotemporal synapse formation and T cell activation.
regulation of T cell costimulation by Immunity 34:919–931
TCR-CD28 microclusters and protein kinase C 12. Hashimoto-Tane A, Yokosuka T, Ishihara C,
theta translocation. Immunity 29:589–601 Sakuma M, Kobayashi W, Saito T (2010) T-cell
8. Yokosuka T, Kobayashi W, Takamatsu M, receptor microclusters critical for T-cell activa-
Sakata-Sogawa K, Zeng H, Hashimoto-Tane tion are formed independently of lipid raft clus-
A, Yagita H, Tokunaga M, Saito T (2010) tering. Mol Cell Biol 30:3421–3429
Spatiotemporal basis of CTLA-4 costimulatory 13. Yokosuka T, Sakata-Sogawa K, Kobayashi W,
molecule-mediated negative regulation of T Hiroshima M, Hashimoto-Tane A, Tokunaga
cell activation. Immunity 33:326–339 M, Dustin ML, Saito T (2005) Newly gener-
9. Yokosuka T, Takamatsu M, Kobayashi-Imanishi ated T cell receptor microclusters initiate and
W, Hashimoto-Tane A, Azuma M, Saito T sustain T cell activation by recruitment of
(2012) Programmed cell death 1 forms negative Zap70 and SLP-76. Nat Immunol
costimulatory microclusters that directly inhibit 6:1253–1262
T cell receptor signaling by recruiting phospha- 14. Grakoui A, Bromley SK, Sumen C, Davis MM,
tase SHP2. J Exp Med 209:1201–1217 Shaw AS, Allen PM, Dustin ML (1999) The

10. Hashimoto-Tane A, Sakuma M, Ike H, immunological synapse: a molecular machine
Yokosuka T, Kimura Y, Ohara O, Saito T controlling T cell activation. Science
(2016) Micro adhesion rings surrounding TCR 285:221–222
Chapter 5
Reconstitution of TCR Signaling Using Supported Lipid

Bilayers
Xiaolei Su, Jonathon A. Ditlev, Michael K. Rosen, and Ronald D. Vale
Abstract
Biochemical reconstitution has served as an important tool for understanding the mechanisms of many
cellular processes including DNA replication, transcription, translation, vesicle trafficking, and ubiquitin-
mediated proteolysis. Here, we demonstrate that biochemical reconstitution can be applied to studying a
complex signaling pathway involving as many as 12 proteins or protein complexes acting at the surface of
model membranes. We show that a temporal sequence of events in activated T cells beginning with phos-
phorylation of the T cell receptor and culminating in the activation of actin polymerization can be repli-
cated in vitro. Our reconstitution demonstrates the sufficiency of these proteins in producing many of the
complex behaviors observed during T cell activation. The ability to manipulate all of the components,
measure reaction rates, and observe molecular behaviors, including at single molecule resolution, has
enabled us to gain insight into some of the important biochemical features of this signaling pathway such
as microcluster formation. The same system could be adapted to study other membrane-proximal signaling
pathways, including growth factor receptors, death receptors, and Eph receptors.
Key words TCR, Microcluster, Reconstitution, Supported lipid bilayer, Multivalency, Phase sep-
aration, Actin, LAT
1 Introduction
T cell receptor (TCR) signaling mediates T cell activation during

adaptive immune responses. A prominent feature of this signaling
pathway is that multiple signaling components form submicron- to
micron-sized clusters on the plasma membrane, though the mech-
anisms and functional consequences of forming these clusters
remained poorly understood. It had been suspected that microclu-
sters are hot spots for signaling because they are enriched in phos-
phorylated proteins, though direct evidence had been lacking [1,
2]. To understand the mechanisms of T cell microcluster forma-
tion, we developed methods for reconstituting microclusters using
purified proteins assembled on supported lipid bilayers in vitro. We
initially focused on LAT (Linker for activation of T cells) [3], a
transmembrane adaptor protein that connects upstream TCR
65
66 Xiaolei Su et al.
activation to multiple downstream signaling pathways including

calcium mobilization, Ras activation, and actin polymerization
[4, 5]. We recombinantly expressed a C-terminal fragment of LAT
containing the four tyrosine residues that have been shown to be
sufficient for mediating TCR signaling [6–8], phosphorylated the
purified protein using the syk family kinase ZAP70, labeled it with
maleimide-conjugated Alexa488, and attached it to planar lipid
bilayers through the interaction of an N-terminal His8 tag on LAT
with Ni-NTA-functionalized lipids incorporated into the bilayer.
pLAT-Alexa488 alone was uniformly distributed on membranes.
However, when Grb2 and Sos1, two prominent, multivalent bind-
ing partners of LAT, were added, submicron-sized clusters formed
and gradually grew in size and intensity (Fig. 1). Physical analyses
Fig. 1 Reconstitution of LAT microclusters on supported lipid bilayers. Top: Schematic of the clustering assay.
Phosphorylated LAT (pLAT) was attached to Ni-NTA functionalized synthetic supported lipid bilayers through an
N-terminal His8 tag. The SH2 and SH3 domains of Grb2 bind the phospho-tyrosines in LAT and proline-rich
motifs (yellow pointy brackets) of Sos1, respectively, thereby creating a network of multivalent interactions.
Bottom: Total internal reflection fluorescence (TIRF) microscopy imaging of LAT clustering. Clusters were
formed after Grb2 (0.5 μM) and Sos1 (0.25 μM) were added to membrane-bound pLAT-Alexa488 (300 mole-
cules/μm2) at 0 min. Scale bar: 5 μm
Reconstitution of TCR Signaling Using Supported Lipid Bilayers 67
of these clusters indicated that they formed through a process of

multivalent assembly and phase separation [9, 10].
Next we extended this reconstitution to a multi-step TCR sig-
naling pathway. The first step in this pathway is phosphorylation of
the TCR (CD3ζ subunit) by the src family kinase, Lck.
Phosphorylated TCR recruits the kinase ZAP70 from solution and
activates it on the membrane. ZAP70, in turn, phosphorylates
LAT on the membrane [11, 12]. Phosphorylated LAT then recruits
Grb2, SOS and the additional multivalent adaptor proteins Gads
and SLP-76, which further recruit the actin regulators Nck,
N-WASp, and the Arp2/3 complex to promote actin polymeriza-
tion on the membrane [13, 14]. We fluorescently labeled ZAP70,
LAT, and actin to serve as reporters for TCR phosphorylation,
LAT clustering, and actin polymerization, respectively. We prein-
cubated all the components in the experimental system. After add-
ing ATP to the reaction to initiate CD3ζ phosphorylation by Lck,
we observed a rapid recruitment of ZAP70 to the membrane,
reflecting its binding to phosphorylated CD3ζ. This translocation
of ZAP70 was then followed by the formation of LAT clusters.
Actin then co-localized with the LAT clusters on the membrane
and eventually polymerized and formed bundles (Fig. 2).
These data demonstrate that a multi-step signaling pathway
can be reconstituted on supported lipid bilayers. The high spatial
and temporal resolution of the tracking of this reconstituted sys-
tem allows us to understand some critical aspects of the signal
transduction process, including microcluster formation, exclusion
of tyrosine phosphatases, and organization of actin regulators [15].
Moreover, the reconstitution itself demonstrates the sufficiency of
the components to produce many of the complex behaviors
observed in vivo during T cell activation.
2 Materials
1. Proteins: LAT, Grb2, Sos1, Gads, SLP-76, Nck, N-WASp,

Arp2/3 complex, and actin (see ref. 15 for detailed protocols
for producing these proteins).

2.
Lipid components: 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphocholine (POPC) (for example (e.g.), Avanti, 850457C),
1,2-dioleoyl-sn-glycero-3-[(N-(5-amino-1-carboxypentyl)imi-
nodiacetic acid)succinyl] (DOGS-NTA) (e.g., Avanti,
790404C), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-
N-[methoxy(polyethylene glycol)-5000] (PEG-5000 PE) (e.g.,
Avanti, 880230C) (see Note 1).
3. Selected 96-well glass bottom plate (flat glass, upper plastic
structure with low fluorescence background and resistant to
detergent, acid, and base treatment, e.g., Matriplate
MGB096-1-2-LG-L).
Fig. 2 Reconstituting a TCR-LAT-actin pathway. Top: Schematic of the components involved in a reaction
designed to reconstitute a signaling pathway from CD3ζ/TCR phosphorylation to actin polymerization. ZAP70-
505-Star, LAT-Alexa647, and actin-Rhodamine serve as reporters for TCR phosphorylation, LAT clustering, and
actin assembly, respectively. Lck, CD3ζ, and LAT were attached to the membrane and incubated with other
components in solution. Bottom: TIRF microscopy revealed ZAP70 membrane recruitment, LAT clustering, and
actin polymerization in the reconstituted assay after the addition of ATP at time 0. Input: 2.5 nM His-tagged
Lck, 5 nM His-tagged CD3ζ, 10 nM His-tagged pLAT-Alexa647, 10 nM ZAP70-505-Star, 250 nM Gads, 500 nM
SLP-76, 1000 nM Nck, 500 nM N-WASp, 5 nM Arp2/3 complex, 1000 nM actin (5% Rhodamine labeled), and
0.5 mM ATP-Mg. Scale bar: 5 μm
4. Adhesive PCR Sealing Foil Sheets (e.g., Thermo Fisher

AB-0626).
5. Glass vials (e.g., National Scientific B7800-2).
6. Hellmanex III (e.g., Sigma Z805939).
7. Gastight Syringes 25 μL (e.g., Hamilton 80,275), 250 μL
(e.g., Hamilton 81,175).
8. Argon.
9. Chloroform (e.g., Electron Microscopy Sciences 12,550).
10. PBS buffer: 155 mM NaCl, 3.0 mM Na2HPO4, 1.1 mM
KH2PO4, pH 7.4.
11. NaOH.
12. Basic buffer: 50 mM HEPES-Na+, 150 mM NaCl, and 1 mM
Tris(2-carboxyethyl)phosphine (TCEP), pH 7.4.
13. Clustering buffer: 50 mM HEPES-Na+, 150 mM NaCl, 1 mM
TCEP, and 1 mg/mL BSA (heat shock fraction, ≥98%), pH 7.4.
14. Signaling buffer: 35 mM HEPES, 60 mM NaCl, 30 mM KCl,
0.4 mM TCEP, 1 mM MgCl2, and 0.4 mg/mL BSA, pH 7.2.
15. High speed centrifuge and polycarbonate centrifuge tubes.
16. Heat Block.
17. Glucose.
18. β-mercaptoethanol.
19. Glucose oxidase.
20. Catalase.
3 Methods
3.1 Preparation 1. Clean the glass vial with 5% Hellmanex III, rinse thoroughly
of Small Unilamellar with ultrapure water (18.2 MΩ at 24 °C).
Vesicles (SUV) 2. Warm up individual lipid stocks to room temperature
(see Note 2).
3. Use glass syringes to prepare a DOGS-NTA lipid mix in the
glass vial as follows: rinse the vial with chloroform, add ~1 mL
of chloroform, and then individual lipids (see Note 3). The
total lipids add up to 4 μmol.
Lipid composition:
98% POPC
2% DOGS-NTA (see Note 4)
0.1% PEG-5000 PE
4. Dry the lipid mix with a stable flow of argon. Use an ~45 °C
water bath during drying (see Note 5). You will see multiple
white layers adhered to the wall of the vial after the lipids
are dried.
5. Continue to dry the lipids completely in a desiccator over 2 h.

6. Resuspend the dried lipids in 1.5 mL of PBS and vortex.
7. Transfer the resuspension into two 1.5 mL conical microcen-
trifuge tubes (750 μL each).
8. Freeze the resuspension in liquid nitrogen and thaw in a water
bath at room temperature. Repeat the freeze-thaw for 30
cycles. The cloudy solution will become clear over the freeze-
thaw cycles (see Note 6). The cleared resuspension can be
stored at −80 °C for future use.
9. Centrifuge the resuspension at 33,500 × g for 45 min at 4 °C.
10. Transfer the SUV-containing supernatant to a clean tube.
Avoid disturbing the white pellet at the bottom of the tube
during the transfer. In the new tube, cover the SUV solution
with argon atmosphere. Store it at 4 °C. Use the SUV within
2 weeks (see Note 7).
3.2 Preparation It is absolutely essential that clean glass is used when making sup-
of Supported Lipid ported lipid bilayers. Glass that is not extremely clean will cause
Bilayers defects in the bilayers and will impair lipid mobility and reproduc-
ibility of experiments. Here, we introduce the use of Hellmanex
III and NaOH to clean the glass.
1. Immerse a 96-well imaging plate into 1 L of 5% Hellmanex III
in a 1 L beaker. Microwave to 50 °C. Stir and incubate over-
night at room temperature. The next morning, rinse each well
with ultrapure water ten times. Blow-dry each well. Seal the
plate with adhesive foil.
2. Use a blade to open the wells that will be used. Add 250 μL of
freshly made 5 M NaOH to each well, incubate at 50 °C for 1
h on a heat block. Remove NaOH, repeat the cleaning twice
(see Note 8).
3. Rinse each well with 500 μL of ultrapure water twice, and then
500 μL of basic buffer twice.
4. Add 200 μL of basic buffer to each well, and then 5 μL of SUV
solution. Tap the tube containing the SUV solution gently to
mix before transferring to the wells. Incubate the plate at 37
°C for 1–1.5 h to allow the bilayer to form. From here on,
make sure the bilayers are always covered with buffer (a mini-
mum volume of ~50 μL in the well).
5. Wash each well with 500 μL of basic buffer three times.
6. Block the bilayers with clustering buffer and incubate for
20 min at 37 °C.
3.3 Imaging LAT 1. Dilute His8-tagged, pre-phosphorylated, and Alexa488 (or

Microcluster other dye)-labeled LAT in clustering buffer and add the mix to
Formation the well and incubate for 2 h at 30 °C (see Note 9).
2. Wash each well with 500 μL of clustering buffer three times.
3. Prepare an oxygen scavenger solution by mixing 0.2 mg/mL
glucose oxidase, 0.035 mg/mL catalase, 25 mM glucose, and
70 mM β-mercaptoethanol in clustering buffer. Add the solu-
tion to the well. The total solution volume in well is 90 μL.
4. Start a time-lapse movie using TIRF (Total Internal Reflection
Fluorescence) microscopy. At the beginning, LAT will appear
evenly distributed on the supported lipid bilayer (see Note 10).
5. Add 0.5 μM Grb2 and 0.25 μM Sos1 in 10 μL of clustering
buffer during the time-lapse acquisition. Pipette up and down
gently several times to mix well. Avoid touching the bilayer
with the pipette tips. Monitor the cluster formation process
(Fig. 1).
3.4 Imaging 1. Add His10-tagged Lck, CD3ζ, and fluorescently labeled His8-
a Signaling Pathway LAT in clustering buffer to the well and incubate for 3 h at 30
from TCR to Actin °C on a heat block or in an incubator.
2. Wash each well with 500 μL of clustering buffer two times,
followed by 500 μL of signaling buffer once.
3. Prepare an oxygen scavenger mix by including 0.2 mg/mL
glucose oxidase, 0.035 mg/mL catalase, 25 mM glucose, and
70 mM β-mercaptoethanol in signaling buffer. Add Alexa505-
ZAP70, Gads, SLP-76, Nck, N-WASp, Arp2/3 complex, and
Rhodamine-actin at desired concentrations to the scavenger
mix (see Fig. 2 legend for recommended concentrations). Then
add the entire mix to the well. The total solution volume in
well is 90 μL.
4. Start a time-lapse movie using TIRF microscopy.
5. Add 0.5 mM ATP-Mg in 10 μL of signaling buffer to the reac-
tion mix during the time-lapse acquisition. Pipette up and
down gently several times to mix well. Avoid touching the
bilayer with the pipette tips. Monitor the sequential signaling
steps from TCR activation, to LAT clustering, finally resulting
in actin polymerization (Fig. 2).
3.5 Analyzing 1. Prior to image analysis, capture five images of an empty well
Cluster Formation (see Note 11). These will be used to determine the noise of the
with FIJI (ImageJ) camera without fluorescence.
2. Also capture five images of an SLB coated with His8 pLAT-
Alexa488 (see Note 12) using microscope settings identical to
those used for time-lapse acquisition. These will be used to
create a flat-field image.
3. Prior to image analysis, set up the proper measurements in

FIJI. The three essential measurements are the integrated fluo-
rescent intensity of the image (the sum of all pixel intensities
within the opened image), maximum intensity, and the stan-
dard deviation of the pixel intensities. To set these go to the
Set Measurements menu (found in the Analysis menu) and
select “Standard deviation,” “Min & Max gray value,” and
“Integrated density.”
4. To determine the camera noise, open the images from step 1
using FIJI, create a stack using the “Images to Stacks” func-
tion (found in the Image menu under Stacks), Z-project the
stack (also found in the Image menu under Stacks) using the
“Average Intensity” projection type, and measure (Ctrl-M or
found in the Analyze menu) the averaged image to determine
the average pixel intensity. This value will be subtracted from
each image used for flat-field correction as well as each frame
of the time-lapse.
5. To flat-field correct the time-lapse, we first create a “ratio
image” that is used to flatten the intensity within each frame of
the time-lapse (see Note 13). Open the five images of His8
pLAT-Alexa488-coated SLBs from step 2 using FIJI, subtract
the noise value determined in step 3 from each image using
the Subtract command (found in the Process menu under
Math), create a stack using the “Images to Stacks” function
(found in the Image menu under Stacks), Z-project the stack
(also found in the Image menu under Stacks) using the
“Average Intensity” projection type, and determine the maxi-
mum pixel intensity in the averaged image using the measure
command (Ctrl-M or found in the Analyze menu). Divide the
averaged image by the maximum measured intensity (the
Divide command is found in the Process menu under Math) to
obtain the “ratio” image. Keep the “ratio” image open and
open the time-lapse that will be analyzed. Open the Calculator
Plus (found in the Process menu) and select the “Divide”
operation using the time-lapse as i1, the “ratio” image as i2,
leaving k1 as 1.0, and k2 as 0.0. Upon completion of this pro-
cess, the time-lapse will now be flat-field corrected.
6. We have used two approaches to analyze cluster formation in
our assays. In the first approach, we measure the variance of
clustering in the frame of interest from a time-lapse of cluster
formation (see Note 14). The variance of the image is simply
the standard deviation in pixel intensities from the image
squared (Variance = Standard Deviation2). To measure vari-
ance in FIJI, open the time-lapse to the desired frame, measure
the standard deviation (Ctrl-M or found in the Analysis menu
using the Measure command), and square the standard devia-
tion. We use variance measurements to determine whether
clusters have formed under different sets of experimental con-

ditions. If the variance of an image containing clustering agents
is significantly different than the variance of an image of a
bilayer coated with an unclustered His8 pLAT-Alexa488 (t-tests
or ANOVA can be used depending on the number of condi-
tions being compared), we consider the critical concentration
for phase separation has been reached. This is important for
determining whether the conditions being used result in clus-
ter formation that can lead to downstream signaling.
7. In the second approach, we measure the fractional intensity of
fluorescent molecules found in clusters (see Note 14). This
requires two independent measurements. First, measure the
total intensity of the image with the measure command (Ctrl-M
or found in the Analysis menu) and record the integrated den-
sity of the image. Next, use the triangle algorithm to set a
threshold for the image (Ctrl-shift-T or found in the Adjust
tab in the Image menu) (see Note 15) and measure the total
intensity of the thresholded region (to measure the integrated
density within only the thresholded region, go the to Set
Measurements menu in the Analysis menu and select “limit to
threshold”). Divide the thresholded integrated density by the
total integrated density to obtain the fractional intensity. We
use fractional intensity as a metric for how many molecules are
located within clusters versus outside clusters. This measure-
ment provides a metric for comparing the level of clustering
between different experimental conditions (see refs. 9, 10, 15)
as examples of how clustering can be compared between differ-
ent experiments).
4 Notes
1. New users of lipids should read FAQ and tips for storage and
handling of lipids from Avanti Polar Lipids, Inc.
http://avantilipids.com/tech-support/faqs/
http://avantilipids.com/tech-support/storage-handling-
of-lipids/
2. Once a lipid vial is opened, store the remaining unused lipids
in a clean capped glass vial. Fill the space above lipids with
argon to prevent potential oxidation of lipids. Store the vial at
−20 °C. Alternatively, for a modest price one can request bot-
tling service from Avanti Polar lipids, which provides small ali-
quots of lipids of a specific volume (chosen based on your
experimental needs).
3. When transferring small amount of lipids, rinse syringe with
the lipids first to avoid any dilution of lipids.
4. We have found that supported lipid bilayers containing 2%

DOGS-NTA lipids provide more stable and reproducible
attachment of His-tagged proteins than bilayers containing 1%
DOGS-NTA. We have also found that bilayers containing 5%
DOGS-NTA or more tend to become less fluid (longer recov-
ery time when tested by FRAP).
5. Adjust the argon flow to allow only small ripples during the
drying process. Usually, it takes ~5 min to dry 2 mL of lipids
in chloroform. It is highly recommended to immerse the lower
half of the vial in a warm water bath (~40 °C) because evapora-
tion of chloroform will take heat away and cool the bottle
down, which could potentially reduce the solubility of lipids or
cause phase separation.
6. Using bath sonication can greatly reduce the cycle numbers.
7. The duration that SUVs can be stored at 4 °C depends on the
stability of the specific lipids used in an experiment. It is worth-
while testing the quality of bilayers by FRAP (fluorescence
recovery after photobleaching) to determine the maximal stor-
age time for a specific lipid mixture.
8.
Keep the temperature of the heat block below 50
°C. Overheating will cause the glue of the plate to melt and
solutions will leak into neighboring wells.
9. In our experience, an input of 3 nM phosphorylated LAT with
an incubation of 2 h will result in a membrane density around
300 molecules/μm2. The final membrane density scales
roughly linearly with the input concentration in a range of
0.1–10 nM. For any new proteins, we recommend measuring
the membrane density [15] at the range needed for the assay.
10. When using a new membrane-bound protein, it will be neces-
sary to determine the mobility of that protein by FRAP to
ensure that the protein behaves well on the membrane.
Sometimes, black holes appear on the bilayer when imaging
fluorescently labeled protein, indicating membrane defects.
Membrane defects are usually caused by insufficient cleaning of
the glass.
11. It is important to know (and subtract) the baseline noise for
the camera used to capture time-lapse images of cluster forma-
tion when quantifying fluorescent intensities within the time-
lapse. The easiest way to determine the noise of the camera is
to capture images in an empty well.
12. There are multiple methods for creating a flat-field image.
Alternative methods can be found at http://imagej.net/
Image_Intensity_Processing.
13. TIRF microscopy often results in an uneven illumination field.
To accurately quantify clustering, we need to “flatten” the
uneven field so that the intensity of both clustered and non-

clustered regions of the image can be accurately measured.
14. We use both variance and fractional intensity to compare the
degree of clustering in any given set of experimental conditions
(see refs. 9, 10, 15 as examples of how clustering can be com-
pared between different experiments). Variance is a good mea-
sure of how clustering changes the distribution of molecules,
from a uniform density to small, intense regions upon the clus-
tering of molecules. Larger, more intense clusters within
images will result in an increased variance. Likewise, an
increased number of clusters within an image will also result in
an increased variance. Fractional intensity is a good measure of
the percentage of molecules that are clustered within any given
image. As the number of clusters within an image or the size
and intensity of clusters within an image increases, the frac-
tional intensity of clusters will also increase.
15. The triangle thresholding algorithm yielded similar results to
an iterative manual procedure that was used to identify pixels
with intensities greater than three standard deviations above
the mean of regions outside of clusters [10].
Acknowledgment
We thank Marcus Taylor and Enfu Hui for their assistance in devel-
oping the reconstitution assays. This work was supported by grants
from the HCIA program of HHMI, the NIH (R01-GM56322 to
M.K.R.), and Welch Foundation (I-1544 to M.K.R.). X.S. was
supported by CRI Irvington postdoctoral fellowship. J.A.D. was
supported by NRSA F32 award 5-F32-DK101188.
References
1. Dustin ML, Groves JT (2012) Receptor sig- linker for activation of T cells is recruited to
naling clusters in the immune synapse. Annu microclusters and is active in signaling.
Rev Biophys 41:543–556 J Immunol 190:3849–3853
2. Seminario MC, Bunnell SC (2008) Signal ini- 6. Zhang W, Trible RP, Zhu M, Liu SK, McGlade
tiation in T-cell receptor microclusters. CJ, Samelson LE (2000) Association of Grb2,
Immunol Rev 221:90–106 Gads, and phospholipase C-gamma 1 with
3. Zhang W, Sloan-Lancaster J, Kitchen J, Trible phosphorylated LAT tyrosine residues. Effect
RP, Samelson LE (1998) LAT: the ZAP-70 of LAT tyrosine mutations on T cell angigen
tyrosine kinase substrate that links T cell receptor-mediated signaling. J Biol Chem
receptor to cellular activation. Cell 92: 275:23355–23361
83–92 7. Lin J, Weiss A (2001) Identification of the
4. Abraham RT, Weiss A (2004) Jurkat T cells minimal tyrosine residues required for linker
and development of the T-cell receptor signal- for activation of T cell function. J Biol Chem
ling paradigm. Nat Rev Immunol 4:301–308 276:29588–29595
5. Balagopalan L, Barr VA, Kortum RL, Park AK, 8. Zhu M, Janssen E, Zhang W (2003) Minimal
Samelson LE (2013) Cutting edge: cell surface requirement of tyrosine residues of linker for
activation of T cells in TCR signaling and thy- kinase that associates with the TCR zeta chain.
mocyte development. J Immunol 170: Cell 71:649–662
325–333 13. Bunnell SC, Hong DI, Kardon JR, Yamazaki
9. Li P, Banjade S, Cheng HC, Kim S, Chen B, T, McGlade CJ, Barr VA, Samelson LE (2002)
Guo L, Llaguno M, Hollingsworth JV, King T cell receptor ligation induces the formation
DS, Banani SF, Russo PS, Jiang QX, Nixon of dynamically regulated signaling assemblies.
BT, Rosen MK (2012) Phase transitions in the J Cell Biol 158:1263–1275
assembly of multivalent signalling proteins. 14. Barda-Saad M, Braiman A, Titerence R,
Nature 483:336–340 Bunnell SC, Barr VA, Samelson LE (2005)
10. Banjade S, Rosen MK (2014) Phase transitions Dynamic molecular interactions linking the T
of multivalent proteins can promote clustering cell antigen receptor to the actin cytoskeleton.
of membrane receptors. Elife 3:e04123 Nat Immunol 6:80–89
11. Reth M (1989) Antigen receptor tail clue. 15. Su X, Ditlev JA, Hui E, Xing W, Banjade S,
Nature 338:383–384 Okrut J, King DS, Taunton J, Rosen MK, Vale
12. Chan AC, Iwashima M, Turck CW, Weiss A RD (2016) Phase separation of signaling mol-
(1992) ZAP-70: a 70 kd protein-tyrosine ecules promotes T cell receptor signal trans-
duction. Science 352:595–599
Chapter 6
Plasma Membrane Sheets for Studies of B Cell Antigen

Internalization from Immune Synapses
Carla R. Nowosad and Pavel Tolar
Abstract
Surrogate planar and membrane systems have been employed to study the architecture of immune syn-
apses; however, they often do not recapitulate trans-synaptic extraction and endocytosis of ligands by the
immune cells. Transendocytosis (or trogocytosis) of antigen from immune synapses is particularly critical
for antigen processing and presentation by B cells. Here we describe a protocol for preparation of plasma
membrane sheets (PMSs), which are flexible and fluid membrane substrates that support robust B cell
antigen extraction. We show how to attach B cell antigens to the PMSs and how to investigate antigen
extraction and endocytosis by fluorescent microscopy and computational image analysis. These techniques
should be broadly applicable to studies of transendocytosis in a variety of cellular systems.
Key words B cells, B cell receptor, Immune synapse, Signaling, Cytoskeleton, Antigen internaliza-
tion, Plasma membrane sheets, Transendocytosis, Trogocytosis
1 Introduction
B cell responses are initiated by binding of antigens to B cell recep-

tors (BCRs). This cognate interaction triggers not only B cell intra-
cellular signaling but also endocytosis of the antigen, and its
processing and presentation to helper T cells. Both BCR signaling
and T cell help are required for fully protective antibody responses.
While BCR signaling has been increasingly well characterized, anti-
gen endocytosis is still poorly understood. Importantly, because
most physiologically relevant antigens are recognized by B cells in
immune synapses with antigen presenting cells (APCs), to endocy-
tose the antigens, B cells need to physically extract them from the
APCs [1–3]. This transfer of membrane-attached material from
one cell to another through direct contact is also called transendo-
cytosis or trogocytosis and has been implicated in the function of a
variety of cell–cell contacts [4–9].
In vitro studies of B cell synapses have initially focused on
models where antigen is presented on planar substrates, such as
77
78 Carla R. Nowosad and Pavel Tolar
plastic, glass or glass-supported synthetic lipid bilayers [10, 11].

The advantage of such models is that they allow high-resolution
imaging of the organization of the B cell signaling and cytoskele-
ton proteins in the synapse. Although some of these models lead to
detectable antigen presentation by B cells expressing very high
affinity BCRs [10, 12], one major disadvantage is that these sub-
strates support internalization of only a small fraction of the anti-
gen accumulated in the synapse [13]. In this sense these models do
not recapitulate normal physiology—the prominent acquisition of
antigen that is seen in B cell contacts with live APCs [14]. Reasons
for the poor endocytosis are not completely understood, but it is
likely that the antigen extraction is limited by the flat, rigid nature
of these substrates, and by the experimental necessity to use strong
tethering of the antigen to the substrate in order to prevent spon-
taneous release.
Here we describe a protocol for preparation of membrane sub-
strates, called plasma membrane sheets (PMSs), which support
robust endocytosis of antigen from B cell synapses [13]. PMSs
were originally developed for studies of endocytosis using reconsti-
tuted cell-free assays [15, 16]. They are sheets of plasma mem-
brane left over after mechanical sheering of adherent cells. Similarly
to synthetic lipid bilayers, the PMSs contain a fluid lipid bilayer,
albeit with a reduced protein diffusion. Unlike glass-supported
synthetic bilayers, however, the lipid bilayers of the PMSs are
attached to the coverslip sparsely and indirectly via transmembrane
proteins and residuals of the extracellular matrix. The advantage is
that the physical properties of the PMSs resemble the viscoelastic
flexibility of the plasma membrane. This allows the interacting cells
to pull on the PMSs and pinch of vesicles for endocytosis as has
been observed with live APCs [13].
A notable caveat of using PMSs for mimicking cell–cell inter-
faces is that they lie on the substrate with their original cytoplasmic
side facing up and therefore extracellular ligands have to be
attached to them artificially. Care should be taken to exclude the
possibility that the cells of interest are recognizing a nonphysiolog-
ical ligand present on the cytosolic side of the PMSs. The PMSs are
also still a passive viscoelastic substrate and may not mimic APCs
that are themselves mechanically active and regulate their mem-
brane stiffness by the cortical cytoskeleton.
However, apart from these caveats, PMSs are easy to make
and can be readily functionalized with desired ligands. We
envisage that they can be used to study transendocytosis in a
wide range of experimental systems. Here we focus on a basic
protocol of preparing PMSs from mammalian cultured cells,
their functionalization using annexin V linkage and their use in
imaging and computational quantification of B cell antigen
endocytosis.
Plasma Membrane Sheets for Studies of Antigen Internalization 79
2 Materials
All buffers and media should be filtered through a 0.22 μm bottle-

top filter. This is not only to ensure sterility but also to remove
debris that can compromise imaging. Buffers and media may be
used for up to a month or longer if they have been stored appro-
priately, although they should be filtered before each use. All buf-
fers containing BSA should be stored at 4 °C. Stock solutions are
prepared using ultrapure water.
2.1 Buffers 1. Running buffer: Phosphate buffered saline solution (PBS) sup-
and Media plemented with 0.5% Bovine serum albumin (BSA) and 2 mM
ETDA. Stir with a magnetic stirrer until all BSA has
dissolved.
2. Hank’s Balanced Salt Solution (HBSS) with Calcium and
Magnesium and without phenol red supplemented with 0.1%
BSA (HBSS/BSA): Dissolve BSA as above.
3. PBS supplemented with 2% BSA: (PBS/BSA).
4. Full RPMI: Roswell Park Memorial Institute (RPMI) medium
1640 supplemented with 10% fetal bovine serum (FBS), 1%
MEM Non- Essential Amino Acids, 2 mM l-glutamine, 50 μM
2- mercaptoethanol, 100 U/ml Penicillin and 100 μg/ml
Streptomycin.
5. Full DMEM: Dulbecco’s Modified Eagles Medium (DMEM)
supplemented with 10% FBS, 1% MEM Non-Essential Amino
Acids, 2 mM l-glutamine, 100 U/ml Penicillin and 100 μg/
ml Streptomycin.
6. Sodium carbonate buffer, pH 9.3. Prepare by mixing 3 ml
0.1 M NaCO with 17 ml 0.1 M NaHCO.
2.2 PMS Generation 1. Coverslip-bottom 8-well imaging chambers (Nalgene Nunc

Lab-Tek catalog number 155411, Thermo Fisher Scientific,
Waltham, MO).
2. 0.01% Poly-l-Lysine solution in water. Keep solution sterile,
store at 4 °C.
3. HEK293 cells (see Note 1).
4. Sterile PBS.
5. KimWipes (Kimberley-Clark, Irving, TX).
2.3 Loading PMSs 1. Annexin-V-biotin (BioVision, Milpitas, CA).

with Fluorescent 2. Streptavidin (Jackson ImmunoResearch, Westgrove, PA).
Antigens
3. Goat anti-mouse Igκ F(ab’) (Southern Biotech, Birmingham,
AL, USA).
4. EZ-Link Sulfo-NHS-LC-LC-biotin (Thermo Fisher Scientific,

Waltham, MO).
5. Fluorescent Cy3 or Cy5 monovalent dye packs (GE Healthcare,
Little Chalfont, UK) or other lyophilized reactive fluorescent
dyes (e.g., Atto dyes or Alexa dyes). Where possible use a pre-
weighted dye pack (available for Cy5 or Cy3 dyes).
6. Anhydrous Dimethyl Formamide (DMF).
7. Zeba spin desalting columns, 7 K MKWCO, 2.5 ml (Thermo
Fisher Scientific, Waltham, MO).
2.4 Preparing 1. ACK red cell lysis buffer (Thermo Fisher Scientific, Waltham,
and Adding B Cells MO).
to PMSs 2. Anti-CD43 microbeads or anti-biotin microbeads (Miltenyi
Biotech, Gladbach, Germany).
3. Alfa Aesar Paraformaldehyde 16% w/v solution (Thermo
Fisher Scientific, Waltham, MO), aliquoted and stored at
−20 °C.
2.5 Equipment 1. Tissue culture hood.

2. Tissue culture incubator (37 °C, 5% CO2).
3. 37 °C waterbath.
4. Aspirator.
5. Probe sonicator: Jencons Scientific model GE 50 20 kHz
Ultrasonic Processor or similar, fitted with a solid 3.2 mm
probe (VWR International, Radnor, PA). Probe affixed to an
adjustable stand and housed in a soundproof cabinet.
6. Centrifuge (Eppendorf, Hamburg, Germany).
7. (Optional) NanoDrop (ThermoFisher Scientific, Waltham,
MO) or other spectrophotometer that can measure UV/Vis
absorbance.
8. Inverted microscope fitted with a 10× objective for cell
culture.
9. AutoMACS cell separator (Miltenyi Biotech, Gladbach,
Germany) (see Note 2).
10. Inverted epifluorescence or confocal microscope with a 100×
objective and motorized z movement (such as IX81, Olympus,
Tokyo, Japan).
11. Image acquisition software capable of z stack imaging, (such as
MetaMorph, Molecular Devices, Sunnyvale, CA).
12. (Optional) AnalyzeInternalizationGUI software tool (available
from http://www.crick.ac.uk/pavel-tolar)
3 Methods
3.1 Cell Seeding Prepare chambers the day before the intended imaging day
for PMSs Production (see Note 3).
1. In a tissue culture hood, remove an 8-well chamber from pack-
aging and lay onto a KimWipe (see Note 4). Pipette 250 μL of
Poly-l-Lysine solution into each well of the chamber, cover with
the chamber lid, and incubate for 1 h. at room temperature.
2. Count HEK 293 cells, wash and resuspend at 1 × 106 cells/ml
in full DMEM prewarmed to 37 °C.
3. Wash each well with five changes of PBS and completely
remove all PBS via aspiration.
4. Add 250 μL of the prepared HEK 293 cells to each well, this
equates to 2 × 105 cells per well.
5. Mark with a pen where the liquid level is, this will serve as a
useful guide for later. Cover chamber with lid.
6. Store the chamber in an incubator set to 37 °C, 5% CO2
overnight.
3.2 Making As a model antigen for polyclonal mouse B cell stimulation, we are
Fluorescently Labeled using F(ab’)2 anti-Igκ. The antigen will be fluorescently labeled
Antigens and biotinylated, which will allow attachment to the PMSs via
annexin V-biotin-streptavidin linkage. Both modifications can be
done easily using commercially available amine-reactive biotin and
fluorescence dyes.
1. Equilibrate the biotin vial to room temperature, then weight
out 2 mg into a tube. Dissolve the biotin in 300 μl of ultrapure
water. This is a 10 mM biotin solution.
2. Add 5 μl–500 μl of 0.5 mg/ml solution of F(ab’)2 anti-Igκ (see
Note 5). This is a ~20-fold molar excess of biotin (see Note 6).
3. Incubate the antigen and biotin together for 30 min at 25 °C.
4. During biotinylation, equilibrate a Zeba spin column with
sodium carbonate buffer, pH 9.3. Use manufacturer’s instruc-
tions for buffer volumes, speed, and duration of centrifugation.
5. Spin the antigen through the column to remove excess biotin
and simultaneously buffer exchange into the carbonate buffer
for dye labeling.
6. Dissolve dye in a small amount of DMF to make a 10 mM
solution.
7. Add 5 μl of the dye solution to 500 μl of the biotinylated anti-
gen, this results in a ~10-fold molar ratio of dye to antigen.
8. Incubate at room temperature in an aluminum foil-covered
tube for 30 min, vortexing every 10 min.
9. Equilibrate a Zeba spin column with PBS and spin labeled

antibody through the column to remove excess dye.
10. The prepared antigen should be visibly colored upon flow
through the column. Optionally, the number of dyes per anti-
gen molecule can be determined by measuring absorbance of
the dye.
3.3 Generating PMSs Once cells are seeded in the coverslip chambers, care must be taken
with pipetting buffers in and out of the wells. Liquids should be
gently applied to the center of the well with a submerged tip, with-
out touching the bottom of the well. The bottom of the wells
should not become exposed to air at any point.
1. Remove pre-seeded chamber from incubator and look at the
cells under an inverted microscope at 10× magnification. Cells
should be close to 100% confluent and well spread (see Note 7
for more information of cell seeding). When the chamber is
held to the light and viewed from underneath, the coverslip
should appear cloudy by the monolayer of cells.
2. Using continual washing (Fig. 1, also see Note 8), wash off the
media from the cells with 10 ml PBS. Fill wells completely to
the top with PBS after washing.
3. Clean the sonicator probe and place the chamber on a suitable
KimWipe-covered stand under the sonicator probe. Carefully
insert the probe into the chamber so that the probe hangs
5 mm above the glass of the chamber. Top up the well with
PBS if any liquid has spilled out.
4. Sonicate for 10 s in each corner of the well with the sonicator
at 20% power (see Note 9 for more sonication tips).
5. Remove the chamber and view the glass from underneath. The
cloudy cell monolayer should be mostly cleared out with small
patches of cells around the edges only. If any large patches of
cells remain, sonicate in the relevant area for 10 s again.
6. Repeat this process for the remaining wells of the chamber
using timings from the first well as your guide.
7. After sonication, wash wells with 10 ml PBS using continual
washing.
8. With a 200 μl pipette tip fitted to an aspirator, carefully remove
PBS until the liquid drops to the marked 250 μl guide level.
9. Block the wells by adding 250 μl of 2% PBS/BSA to each well.
Incubate for at least 20 min, preferably up to 1 h.
10. After blocking, wash wells as in steps 7 and 8.
11. Buffer exchange wells into HBSS/BSA by continually washing
with 5 ml per well. Leave 250 μL HBSS/BSA in wells.
10 ml pipette fitted with a disposable tip
8-well chamber on a KimWipe
Aspirator fitted with a disposable tip
-Opposite corners
-Tips touch liquid
-Tips do not touch glass
-All wells have liquid in
Fig. 1 Setup for continual washing. An 8-well chamber was prepared as detailed in
Subheading 3.1. Top panel shows a 10 ml pipette fitted with a disposable pipette
tip, an 8-well chamber resting on a KimWipe and an aspirator fitted with a dispos-
able tip. Lower panel shows the position of the pipette tips during washing
12. Add annexin-V-biotin at 24 nM (see Note 10). Prepare a 2×

stock by diluting 0.5 mg/ml annexin-V-biotin stock 1:250 in
HBSS/BSA. Add 250 μl annexin dilution to 250 μL HBSS/
BSA presently in the well, making the final dilution 1:500.
Incubate for 20 min at room temperature.
13. Wash wells with 10 ml HBSS/BSA using continual washing.
Leave 250 μl of buffer in the wells.
14. Prepare a 6× stock solution of streptavidin at 6 μg/ml in
HBSS/BSA; add 50 μL to each well, final concentration 1 μg/
ml. Incubate at room temperature for 20 min.
15. Wash wells as in step 13.
16. Prepare antigen in a similar way to streptavidin. Make a 6× stock
solution in HBSS/BSA. Add 50 μl of this per well (see Note 11).
Incubate antigen on sheets at room temperature for 20 min.
Fig. 2 An example image of PMSs loaded with anti-Igκ-Cy5 antigen. PMSs were pre-
pared as described in this protocol and imaged with 640 nm laser illumination and a
100× objective fitted to an Olympus IX81 fluorescent microscope. Scale bar 5 μm
From now on, keep the chamber covered in aluminum foil to

protect it from light and avoid photobleaching.
17. Wash wells as in step 13.
18. Check sheets under a fluorescent microscope before loading
with cells. See Fig. 2 for an example image of antigen-loaded
PMSs.
3.4 Measuring Primary B cells are isolated from spleens of mice. Generally, pri-
Internalization mary B cells should be prepared immediately before imaging, kept
in Primary B Cells cold at all times and warmed to 37 °C before imaging.
1. Mash a spleen through a 70 μm cell strainer using the plunger
from a 2 ml syringe.
2. Rinse cells through the strainer with 2 ml ACK buffer and
incubate at room temperature for 5 min to lyse red blood cells.
3. Collect cells in 10 ml cold, sterile running buffer and spin
down at 300 g for 10 min at 4 °C.
4. Resuspend cell pellet in 450 μL running buffer and add 50 μl
anti-CD43 microbeads (see Note 12).
5. Incubate at 4 °C on a rotator for 15 min.
6. Wash cells, resuspend in 1 ml running buffer and run through
an AutoMacs cell separator, collecting the “depleted” fraction,
which contains enriched B cells.
7. Count B cells (one spleen yields around 30 × 106 naive B cells).
8. Wash B cells in running buffer before resuspending them in

warm full RPMI at 5 × 106 cells/ml.
9. Rest cells in an incubator set to 37 °C for 10–15 min before
imaging.
10. Meanwhile, warm imaging chamber with PMSs at 37 °C in the
incubator for at least 10 min.
11. To image, remove 0.5 × 106–1 × 106 cells per imaging well and
spin.
12. Resuspend cells in 50 μl/well warm HBSS/BSA and warm up
briefly at 37 °C.
13. Pipette cells slowly into the wells and return the chamber to
the incubator and incubate at 37 °C for 20 min.
14. Fix cells by adding 50 μl 16% PFA directly to each well and
incubating at room temperature for 20 min.
15. Optional: Steps 15–17 are optional to stain the B cells for
surface markers. Wash PFA off cells with HBSS/BSA “by
hand”—filling wells and removing liquid carefully with a
pipette. Take care not to disturb the B cells adhered to the
PMSs. Repeat washing by hand at least five times. Add normal
mouse serum at 5% v/v and incubate at room temperature for
30 min to block the wells.
16. Stain B cells with fluorescently labeled primary antibodies
added directly into blocked wells at 1 μg/ml and incubated at
room temperature for 20 min.
17. Wash by hand and fix again as in step 14.
18. Transfer chamber to the stage of the microscope for image
acquisition. Using fluorescence of the antigen, find a field of
view with PMSs and look for attached B cells. See Fig. 3 for an
Antigen B Cell Merge
Top
View
Antigen B Cell Merge

Side
View
Fig. 3 An example of B cell internalization of antigen from PMSs. Red shows fluorescent antigen (anti-Igκ-Cy5), cyan
shows surface B cell staining (anti-B220-FITC). Top panels show a single z-plane with the PMS in focus. The B cell
has gathered antigen on the PMS, forming an immune synapse. Antigen can be seen inside the B cell in a side-view
maximum projection (lower panel), as generated by the AnalyzeInternalizationGUI software tool. Scale bar 5 μm
example image of B cell synapse formation and antigen

internalization. Acquire z stacks covering a volume starting
several planes below the PMSs and ending several planes above
the B cells with z-spacing of 200–500 nm.

19. Analyze antigen internalization using the AnalyzeInter
nalizationGUI tool. For details on usage, follow instructions
included with the software.
4 Notes
1. We use HEK293 cells that are adapted for growth in suspension.

These cells are maintained in shaker flasks using Pro293 media
(Lonza, Slough, UK) supplemented with 3% FBS. Adaptation
of adherent HEK293 cells to suspension culture can be done
using protocols from Lonza. If adherent HEK293 cells are used
for making PMSs, use slightly lower numbers of cells (1.5 × 105
per imaging well). In general, the quality of the PMSs depends
on the health of the cells. Keep cells at a low passage (below 35)
and monitor cell viability regularly.
2. An AutoMacs is not strictly required, manual separation of
cells using a magnet and columns also works, and may even be
preferred, as it is easier to maintain sterile conditions for subse-
quent cell culture.
3. Seed cells to make PMSs as late in the day as is practical.
4. Always lay the imaging chamber on a KimWipe. Laying the
imaging chamber directly on the bench or hood may dirty or
scratch the coverslip, compromising the quality of the
imaging.
5. If you are using a different antigen, make sure it is in a buffer
that is free of additives or protein carriers. Buffers with small
molecular weight additives can be exchanged for PBS using a
desalting column. Protein carriers, such as BSA have to be
removed using chromatography.
6. Calculations of molar ratios have been done using the molecu-
lar weight of the F(ab’)2 anti-Igκ fragment. If you are using
antigen of a different molecular weight, adjust the volume of
the biotin or dye solutions accordingly to maintain the indi-
cated molar ratio.
7. If the HEK293 cells are less than confluent, the quality of
PMSs will not be affected, although they will be sparser. If cells
are overgrown and less spread, they may peel away entirely
from the coverslip during sonication, leaving large areas of
glass without PMSs. The PMSs will also be smaller. In our
experience, most problems with PMSs’ preparation are because
of poor cell health or incorrect cell density before sonication.
8. Continual washing refers to exchanging buffers in the chamber

without emptying the wells. This involves slowly adding buffer
to one corner of the well using a 10 ml pipette fitted with a
disposable 1000 μl pipette tip, and simultaneously aspirating
buffer from an opposite corner of the well with an aspirator
also fitted with a disposable pipette tip (Fig. 1). Ideally, this
creates a laminar flow through the chamber, exchanging the
buffer. At no point during the washing should the buffer sur-
face drop to a level that would expose the PMSs to air. 10 ml
of continuous washing is usually sufficient to exchange buffers
in a well of this imaging chamber, 20 ml may be required if
more extensive washing is needed.
9. Power settings can vary between sonicator models and may
need to be adjusted empirically. During sonication the liquid in
the well should quiver but not froth. Small cell patches may be
visibly sonicated from the bottom of the well. If cells disappear
from the bottom very quickly, lower the power of the sonica-
tor. Refer to Note 7 for tips on cell health and density.
Oversonicating can lead to a loss of PMSs, undersonicating can
leave some intact HEK293 cells on the coverslip, which may
interfere with imaging.
10. Annexin V requires calcium for binding to phosphatidylserine
in the exposed leaflet of the PMSs. Buffer exchange into a cal-
cium-containing buffer like HBSS is important before the
addition of annexin-V-biotin to PMSs. The PMSs should be
kept in a calcium-containing buffer from now on.
11. Some optimization of antigen loading concentration may be
necessary. Loading of PMSs often requires higher antigen con-
centration than loading of planar lipid bilayers. Care should be
taken with wash steps at higher concentrations to remove all
soluble antigen.
12. Anti-CD43 microbeads are useful for rapid B cell isolation as
they offer a one-step isolation protocol. However, enrichment
of more specific B cell populations can be achieved using a
combination of biotinylated antibodies followed by anti-biotin
microbeads. It is always better to use negative selection and
leave cells of interest free of antibodies and microbeads before
imaging on PMSs.
Acknowledgements
This work was funded by the Francis Crick Institute, which receives
its core funding from Cancer Research UK, the UK Medical Research
Council and the Wellcome Trust. In addition, this work was sup-
ported by the European Research Council Consolidator Grant num-
ber 648228 and by the EMBO Young Investigator Programme.
References
1. Carrasco YR, Batista FD (2007) B cells acquire lytica contributes to cell killing and tissue inva-
particulate antigen in a macrophage-rich area at sion. Nature 508:526–530. doi:10.1038/
the boundary between the follicle and the subcap- nature13242
sular sinus of the lymph node. Immunity 27:160– 9. Steele S, Radlinski L, Taft-Benz S et al (2016)
171. doi:10.1016/j.immuni.2007.06.007 Trogocytosis-associated cell to cell spread of
2. Phan TG, Green JA, Gray EE et al (2009) intracellular bacterial pathogens. Elife
Immune complex relay by subcapsular sinus 5:e1000211. doi:10.7554/eLife.10625
macrophages and noncognate B cells drives 10. Fleire SJ, Goldman JP, Carrasco YR et al
antibody affinity maturation. Nat Immunol (2006) B cell ligand discrimination through a
10:786. doi:10.1038/ni.1745 spreading and contraction response. Science
3. Junt T, Moseman EA, Iannacone M et al 312:738–741. doi:10.1126/science.1123940
(2007) Subcapsular sinus macrophages in 11. Tolar P, Hanna J, Krueger PD, Pierce SK
lymph nodes clear lymph-borne viruses and (2009) The constant region of the membrane
present them to antiviral B cells. Nature immunoglobulin mediates B cell-receptor clus-
450:110–114. doi:10.1038/nature06287 tering and signaling in response to membrane
4. Marston DJ, Dickinson S, Nobes CD (2003) antigens. Immunity 30:44–55. doi:10.1016/j.
Rac-dependent trans-endocytosis of ephrinBs immuni.2008.11.007
regulates Eph-ephrin contact repulsion. Nat 12. Batista FD, Neuberger MS (2000) B cells
Cell Biol 5:879–888. doi:10.1038/ncb1044 extract and present immobilized antigen:
5. Qureshi OS, Zheng Y, Nakamura K et al implications for affinity discrimination. EMBO
(2011) Trans-endocytosis of CD80 and CD86: J 19:513–520
a molecular basis for the cell-extrinsic function 13. Natkanski E, Lee W-Y, Mistry B et al (2013) B
of CTLA-4. Science 332:600–603. cells use mechanical energy to discriminate
doi:10.1126/science.1202947 antigen affinities. Science 340:1587–1590.
6. Martínez-Martín N, Fernández-Arenas E, doi:10.1126/science.1237572
Cemerski S et al (2011) T cell receptor inter- 14. Batista FD, Iber D, Neuberger MS (2001) B
nalization from the immunological synapse is cells acquire antigen from target cells after syn-
mediated by TC21 and RhoG GTPase- apse formation. Nature 411:489–494
dependent phagocytosis. Immunity 35:208– 15. Moore MS, Mahaffey DT, Brodsky FM,
222. doi:10.1016/j.immuni.2011.06.003 Anderson RG (1987) Assembly of clathrin-
7. Meloty-Kapella L, Shergill B, Kuon J et al coated pits onto purified plasma membranes.
(2012) Notch ligand endocytosis generates Science 236:558–563
mechanical pulling force dependent on dyna- 16. Wu M, Huang B, Graham M et al (2010)
min, epsins, and actin. Dev Cell 22:1299– Coupling between clathrin-dependent
1312. doi:10.1016/j.devcel.2012.04.005 endocytic budding and F-BAR-dependent

8. Ralston KS, Solga MD, Mackey-Lawrence NM tubulation in a cell-free system. Nat Cell Biol
et al (2014) Trogocytosis by Entamoeba histo- 12:902–908. doi:10.1038/ncb2094
Chapter 7
Studying the Dynamics of TCR Internalization

at the Immune Synapse
Enrique Calleja, Balbino Alarcón, and Clara L. Oeste
Abstract
Establishing a stable interaction between a T cell and an antigen presenting cell (APC) involves the forma-
tion of an immune synapse (IS). It is through this structure that the T cell can integrate all the signals
provided by the APC. The IS also serves as a mechanism for TCR downregulation through internalization.
Here, we describe methods for visualizing MHC-engaged T cell receptor (TCR) internalization from the
IS in human cell lines and mouse primary T cells by confocal fluorescence microscopy techniques.
Key words T cell, TCR internalization, Immune synapse, Antigen presenting cell, MHC-peptide,
SMAC, Confocal fluorescence microscopy
1 Introduction
When a T cell contacts an APC, a tight junction is formed between

the two cells: the immune synapse. The IS serves different pur-
poses for the function of the T cell such as enhancing and sustain-
ing TCR signaling, as well as precluding T cell circulation by
upregulating the expression of adhesion molecules such as LFA-1
or ICAM-1. This subcellular structure is also responsible for the
secretion of lytic molecules and cytokines [1].
Within seconds of the T cell:APC interaction, the IS begins to
form via a structure known as the “bull’s-eye” [2]. Shortly there-
after, the contacted TCRs translocate to the center of the IS. Finally,
a pattern appears that divides the IS into three differentiated and
concentric rings named supramolecular activation complexes
(SMAC). The d-SMAC (distal) is the outermost rim and contains
CD43 and CD45. The p-SMAC (peripheral), or intermediate rim,
is rich in adhesion molecules, whereas the c-SMAC (central) is
where the contacted TCRs are located. This final configuration can
last several hours [3–6].
89
90 Enrique Calleja et al.
The TCR forms aggregates termed microclusters (MC) during

the very first seconds after T cell:APC contact. These MCs translo-
cate to the c-SMAC, where it seems that they are not phosphory-
lated, i.e., inactive. This observation gave rise to the view that the
IS can act as an “off-switch” to terminate TCR signaling.
Furthermore, this hypothesis is enhanced by the fact that the phos-
phatase CD45 is excluded from peripheral MCs [7, 8]. The forma-
tion of the c-SMAC therefore accounts for the downregulation of
the TCR, which terminates its signaling [7]. In resting T cells, the
TCR is continuously endocytosed and reexpressed on the plasma
membrane [9]. When the TCR is engaged by a peptide-coupled
MHC (MHCp), the TCR is downregulated, and driven toward its
degradation. It has been proposed that the IS boosts TCR trigger-
ing when the MHCp is present at low concentrations, whereas
high concentrations of MHCp lead to TCR downregulation [10].
During IS formation, two different types of TCR internaliza-
tion can be distinguished. When internalized, the TCR can follow
both a clathrin-dependent and a clathrin-independent pathway,
although the former can occur in both triggered and nontriggered
TCRs [11]. The first one allows recycling and relocation of the
TCR (clathrin-dependent), whereas the second one arises from the
c-SMAC and is associated with a degradative outcome, or down-
regulation (clathrin-independent).
In this Chapter, we set forth confocal microscopy approaches
for studying TCR internalization. We describe a method to track
the TCR and other proteins of interest in live cells at the IS formed
between human T and B cells (Jurkat and Raji cell lines, respec-
tively). A second protocol is presented to quantify TCR internal-
ization at the IS formed between primary T cells (isolated from
transgenic mice) and APCs by differential extra- and intracellular
TCR staining. Both methods have been previously employed to
elucidate the role of small GTPases in TCR internalization [12],
and will hopefully help to shed light on the involvement of other
proteins of interest in this crucial cellular event for immune
regulation.
2 Materials
2.1 Cell Culture Carry out all procedures in a laminar flow hood for an aseptic
environment.
1. Cell lines: Jurkat human T cell lymphoma clone E6-1 cells
(ATCC® TIB-152™), Raji human lymphoblastoid B cells
(ATCC® CCL-86™), DCEK mouse fibroblasts (see Note 1).
2. Mice for primary T cell isolation: C57BL/6 background TCR
AND transgenic mice (see Note 1).
TCR Internalization at the IS 91
3. Complete medium for Jurkat and Raji cell culture: Roswell

Park Memorial Institute medium (RPMI 1640) supplemented
with 5% Fetal Bovine Serum (FBS), 2 mM l-glutamine, 100
U/mL penicillin, and 100 U/mL streptomycin.
4. Complete medium for DCEK cell culture: Dulbecco’s
Modified Eagle’s Medium (DMEM) supplemented with 10%
FBS, 2 mM l-glutamine, 100 U/mL penicillin, and 100 U/
mL streptomycin.
5. Complete medium for primary T cell culture: RPMI supple-
mented with 10% FBS, 2 mM l-glutamine, 100 U/mL peni-
cillin, 100 U/mL streptomycin, 10 mM sodium pyruvate, 20
μM β-mercaptoethanol, and 5 pg/mL of interleukin 2 (IL-2,
Peprotech®) (see Note 2).
6. Cell culture dishes (100 mm, 6-well and 24-well plates) and
flasks (T-75), autoclaved pipette tips, sterile glass pipettes,
Neubauer cell counting chamber, and light microscope.
7. CellTracker™ Blue CMAC Dye (ThermoFisher Scientific) (see
Note 3).
8. Staphylococcus aureus enterotoxin E (SEE, Toxin Technologies,
Inc.) (see Note 4).
2.2 Jurkat Cell 1. Transfection medium: RPMI supplemented with 20% FBS and
Transfection 20 mM Hepes pH 7.4.
2. Cell culture plates for microscopy: MatTek 35 mm glass bot-
tom dishes (10 mm glass diameter).
3. Fibronectin at 20 μg/mL to cover glass bottom dishes (see
Note 5).
4. Bio-Rad Gene Pulser Electroporator.
5. 0.4 mm electroporation cuvettes (Bio-Rad).
6. Plasmids: pEYFP-N1-CD3ζ for TCR tracking and other plas-
mids of choice (see Note 6).
2.3 Murine Cell 1. CD4+ T Cell Isolation Kit (Miltenyi Biotec©) (see Note 7).
Immunofluorescence 2. Phosphate-Buffered Saline (PBS) to wash cells. Supplement
with 1% Bovine Serum Albumin (BSA) for antibody staining.
3. 2% Formaldehyde in PBS for cell fixation.
4. 0.1% Saponin in PBS for fixed cell permeabilization.
5. Hank’s Balanced Salt Solution (HBSS) to resuspend cells for
microscopy observation.
6. Cell culture plates for microscopy: MatTek 35 mm glass bot-
tom dishes (10 mm glass diameter).
7. Fibronectin at 20 μg/mL to cover glass bottom dishes (see
Note 5).
8. Primary antibodies: polyclonal rabbit anti-mouse CD3ζ anti-

body 448 [13] and biotinylated hamster anti-mouse βTCR
antibody (H57-597, Immunotools, Bioscience).
9. Secondary antibodies: anti-rabbit Alexa 488 and Streptavidin
Alexa 555 (Invitrogen).
10. Mowiol® 4-88 and Dabco® 33-LV (Sigma-Aldrich).
2.4 Confocal 1. Zeiss confocal Laser Scanning Microscope LSM710 coupled

Fluorescence to an AxioObserver inverted microscope with 63× and 100×
Microscopy PlanApochromat oil immersion DIC M27 objectives, 1.4
numerical aperture (see Note 8).
2. Thermostatic chamber for live cell observation at 37 °C where
specified.
3. ImageJ software [14].
4. Metamorph software 6.2r6 (Universal Imaging).
3 Methods
3.1 Tracking TCR The protocol below describes the steps needed to follow immuno-
Internalization logical synapse (IS) formation in Jurkat-Raji cell conjugates by
at the Immune time-lapse fluorescence confocal microscopy in live cells. Transient
Synapse in Jurkat-Raji transfection of Jurkat cells with pEYFP-N1-CD3ζ by electropora-
Conjugates tion allows tracking of the TCR upon contact with Raji cells that
have previously been loaded with superantigen (SEE). Jurkat cells
may be cotransfected with a protein of interest and its variants to
elucidate the degree of localization to the IS and/or its involve-
ment in TCR internalization.
1. Grow Jurkat and Raji cells in 100 mm cell culture dishes or
larger flasks in complete medium at 37 °C in an atmosphere
with 5% CO2 to a maximum concentration of 1 × 106 cells/
mL.
2. Transient transfection of Jurkat cells with pEYFP-N1-CD3ζ by
electroporation
(a) Count Jurkat cells on a Neubauer chamber and collect the
necessary volume to obtain 10 × 106 total cells per trans-
fection condition in a 15 mL polypropylene tube.
Centrifuge at 500 × g for 5 min.
(b)
Discard supernatant and resuspend the total 10 × 106
Jurkat cells in 500 μL of pre-warmed transfection medium.
(c) Place Jurkat cells in a 0.4 mm electroporation cuvette, add
40 μg of pEYFP-N1-CD3ζ (mix gently), and incubate for
10 min at room temperature (see Note 9).
(d) Give one single electroporation shock at 960 μF capaci-
tance and 260 mV. Immediately transfer Jurkat cells to a
100 mm culture dish with 10 mL of complete growth

medium (see Note 10).
(e) Maintain Jurkat cells in this medium for 16–24 h, after
which transfection efficiency should be checked on a flow
cytometer (see Note 11). If Jurkat cells are sufficiently
transfected, carry out experiment at this time.
3. Count Raji cells on a Neubauer chamber, centrifuge and resus-
pend at a final concentration of 3 × 106 cells/mL in RPMI
without serum.
4. Stain Raji with 10 μM CellTracker™ Blue CMAC Dye for
30 min at 37 °C (see Note 12).
5. Wash Raji cells twice with complete medium (see Note 13).
6. Load Raji cells with SEE at 1 μg/mL final concentration in
complete medium for 20 min at 37 °C.
7. Resuspend transfected Jurkat cells at 0.3 × 106 cells/mL in
HBSS supplemented with 2% FBS.
8. Seed 1 mL of Jurkat cells on a fibronectin-coated, glass bot-
tom 35 mm dish and allow cells to adhere for 15 min at 37 °C.
9. Place the dish on a 37 °C thermostatized chamber in the
LSM710 confocal microscope and focus in on Jurkat cells of
interest. Obtain sequential images of fluorescence and bright
fields every minute, establishing three to four planes and
acquiring z stacks in each fluorescent channel.
10. One minute after starting to record Jurkat cells, add 60 μL of
preloaded Raji cells near the area that has been focused on,
close to the objective, to increase chances of contact between
the chosen Jurkat and the incoming Raji.
11. Obtain sequential images of fluorescence and bright fields

every minute (see Fig. 1 for example images).
12. Process images for contrast, brightness, background subtrac-
tion, z stack merging, or single plane splitting, using Image J
software.
13. Analyze data to obtain relative values of internalization of the
TCR from the IS by scoring according to the presence of
CD3ζ in vesicles immediately below the IS using Metamorph
software (see Note 14).
3.2 Quantification of The following protocol is based on conjugates formed by naive T

TCR Internalization cells extracted from lymph nodes of transgenic TCR AND mice
from the Immuno- and DCEK cells loaded with their corresponding antigen (MCC)
logical Synapse as presenting cells (see Note 15). To track external TCR that is
in Primary Mouse amenable to internalization, T cells are stained with a biotinylated
CD4+ T Cells antibody tagging the β-TCR chain extracellularly prior to contact
with DCEK. After T cell-DCEK incubation during which immune
synapses are formed, cells are fixed, permeabilized and an antibody
Fig. 1 Time-lapse video microscopy of Jurkat-Raji conjugates. Jurkat cells transiently cotransfected with GFP-
RRas2 and CD3ζ-mCherry were presented with SEE-loaded Raji cells stably expressing CFP. Note the colocal-
ization between the TCR and the small GTPase at the internalizing sites from the IS. Single Z-sections are
shown for every time point for the three fluorescence channels and a brightfield image, as well as the merged
image of the fluorescent channels
recognizing internal segments of the TCR is used, as well as fluo-

rescent streptavidin. In this way, any TCR that has been internal-
ized will be tracked according to its correlation with internal TCR
chains. The genetic background of mice from which T cells are
obtained provides flexibility for studying the effect on TCR dynam-
ics of alterations such as knock-outs, knock-ins, or deletions of
genes of interest.
1. Seed DCEK cells (antigen presenting cells (APC)) on round
glass coverslips placed in wells of a 24-well plate at 5 × 104
cells/well.
2. Incubate overnight in DMEM with 10% FBS adding MCC
peptide at 10 μM (see Notes 1 and 17).
3. Obtain CD4+ T cells from lymph nodes of transgenic TCR
AND mice using the CD4+ T Cell Isolation Kit according to
the manufacturer’s instructions and resuspend at 2 × 106 cells/
mL (see Note 7).
4. Stain purified T cells with biotinylated anti-βTCR antibody
(H57-597) at 1:100 in PBS with 1% BSA for 5 min at 0 °C (see
Notes 16 and 17).
5. Meanwhile, wash DCEK with PBS to do away with excess
peptide.
6. Add 250 μL of the purified T cell mix to each well containing

DCEKs and incubate for 30 min.
7. Fix cells in 2% formaldehyde in PBS at room temperature (RT)
for 10–15 min.
8. Permeabilize cells with 0.1% saponin in PBS at RT for 20 min.
9. Stain cells with anti-CD3ζ 448 antibody at 1:3000 in PBS
with 1% BSA at RT for 30 min.
10. Wash with 1% BSA in PBS.
11. Incubate cells with streptavidin Alexa 555 at 1:400 and anti-
rabbit Alexa 488 at 1:500 at RT for 30 min (see Note 16).
12. Wash twice in PBS and once in dH2O.
13. Mount coverslips with cells onto glass slides with Mowiol and
the anti-fading agent Dabco, according to the manufacturer’s
instructions.
14. Allow enough time for drying (overnight) and keep cells at 4
°C until observed on the microscope.
15. Place the slides in the LSM710 confocal microscope at RT and
search for T cell:DCEK conjugates. Obtain sequential images
of fluorescence and bright fields, establishing several planes
and acquiring z stacks for all channels.
16. Process images for contrast, brightness, background subtrac-
tion, z stack merging, or single plane splitting using Image J
software.
17. Analyze images on Metamorph software (see Note 18):
(a)
Quantify IS formation according to the percentage of T
cell:APC conjugates with CD3ζ concentrated at the synapse.
(b) Quantify TCR internalization from the IS according to the
percentage of T cells with internal vesicles positive for both
CD3ζ and H57 antibodies.
4 Notes
1. DCEK is a mouse fibroblast cell line that is stably transfected

with I-Ek (a class II MHC) and with CD80. This allows them
to act as APCs of the specific peptide, MCC (moth cytochrome
C, ANERADLIAYLKQATK). This antigen is recognized by T
cells from mice with transgenic AND TCR, which express
TCR chains Vα11.1 and Vβ3 that are MCC specific and I-Ek
restricted [15].
2. Interleukin 2 (IL-2) is added to the complete growth medium
of isolated murine T cells due to its ability to sustain their
long-term proliferation.
3. 7-amino-4-chloromethylcoumarin (CMAC) has blue excita-
tion/emission spectra (353/466 nm maxima) that can be
detected with 405 nm diode lasers on the confocal microscope

to track cells.
4. Staphylococcal enterotoxin E (SEE) is loaded onto Raji cells,
which recognize this protein through their MHC-II [16].
SEE-loaded MHC-II molecules are in turn recognized by
Jurkat TCRs, eliciting immune synapse formation.
5. Place a 50 μL drop of fibronectin at 20 μg/mL on the glass
area of a MatTek dish and carefully introduce dish into a 37 °C
incubator for 1 h. Remove with an aspiration tool and seed
cells directly on dish.
6. Jurkat cells can be cotransfected with other plasmids to track
colocalization or cointernalization of proteins of interest with
the TCR. Such approaches were used to determine the role of
certain small GTPases (i.e., RRas2 and RhoG) in TCR inter-
nalization by trogocytic mechanisms [12].
7. The CD4+ T Cell Isolation Kit is used to isolate CD4+ T cells
by depletion of non-CD4+ T cells (negative selection). Lymph
node cell suspensions are incubated with a mix of biotin-
conjugated antibodies against CD8a, CD11b, CD11c, CD19,
CD45R (B220), CD49b (DX5), CD105, Anti-MHC-class II,
Ter-119, and TCRγ/δ. These primary antibodies are then
magnetically labeled with Anti-Biotin MicroBeads, such that
the magnetically labeled nontarget cells are depleted by retain-
ing them on a MACS® Column in the magnetic field of a
MACS Separator. The unlabeled CD4+ target cells pass
through the column and can be used for further experiments
directly. See ref. 17 for a detailed protocol.
8. Ensure the confocal microscope is equipped with the adequate
lasers and filters for optimal fluorescence signal detection. The
microscope used for the experiments based on this protocol
contains the following lasers: Diode 405 nm, Argon
458/488/514 nm, DPSS 561 nm, and HeNe 633 nm. The
filters are GFP (Ex 450-490 Em 500-550, 38HE Zeiss) and
CFP/DsRed (Ex 400-440/525-580 Em 460-510/590-710,
55,018 Chroma).
9. If cotransfecting with another plasmid, use 20 μg of each such
that maximum total DNA is 40 μg. Preferably use purified
DNA from Maxipreps for optimum transfection efficiency.
10. Viscous, foamy material may form inside the electroporation
cuvette due to DNA from ruptured cells. Take care not to
transfer this material to the 100 mm culture plate. Use 200 μL
pipette tips to work around these dense strings and collect
only the soluble cell suspension.
11. Use a cytometer that is equipped with the lasers and filters
needed to detect fluorescence of the fusion protein used (in
this case, a 488 nm laser and a 530/30 nm filter for eYFP).
Transfection efficiencies of approximately 50% or more are
considered adequate to proceed with the experiment.
12. CellTracker™ Blue CMAC Dye is a transient dye that can be
used for short tracking times with Raji cells. However, for lon-
ger incubations, Raji cells stably expressing a soluble, untagged
fluorescent protein can be used to track cell movement. If so,
take note to use a fluorescent protein that does not interfere
with the channels used to detect other fluorophores. For
example, use CFP, with an excitation wavelength of 405 nm,
if detecting other proteins with maximum excitation peaks at
488 nm, 561 nm or 633 nm.
13. Washing cells with medium containing serum helps to stop
CMAC from further entering the cells and is a milder medium.
14. To quantify TCR internalization, score 15–20 single T cell:B
cell conjugates according to the presence of CD3ζ in vesicles
immediately below the IS. The presence of intracellular vesi-
cles derived from the IS is evaluated according to their dis-
tance to the IS (<1.4 μm) and the existence of a gap between
the vesicle and the IS ≥300 nm, to obtain the number of
TCR-containing vesicles beneath the IS in different condi-
tions [12].
15. It is also possible to use different T cell:antigen presenting cell
pairs, i.e., cells obtained from mice with transgenic TCRs con-
taining different chains that specifically recognize other anti-
gens, presented by different types of MHCs. For example, T
cells from lymph nodes of OT-I TCR transgenic mice bear
TCRs that are specific for a class I Kb-restricted OVA peptide
(OVAp, SIINFEKL). B cells derived from spleens of these
same mice can be loaded with OVAp and recognized by the
OT-I TCR to form conjugates that can be tracked following
the same methods described in this Chapter.
16. This biotinylated anti-βTCR antibody (H57-597) recognizes
the extracellular region of the TCR and has a low stimulatory
profile. Therefore, it will bind to TCRs present on the exter-
nal part of the T cell membrane prior to stimulation with
DCEKs. The IS will be formed upon contact with the pre-
senting cell (DCEK) and TCRs will become internalized.
After conjugate formation, cell fixation and permeabilization,
anti-CD3ζ 448 is used to recognize intracellular regions of
TCR complexes. Streptavidin Alexa 555 (for H57) and anti-
rabbit Alexa 488 (for 448) secondary incubations will then
highlight extra- versus intracellular TCRs. Since cells were
not permeabilized at the moment of H57 addition, any
detectable intracellular staining with H57 will be due to inter-

nalized TCR from the membrane.
17. As a negative control, DCEKs that have not been preloaded
with MCC are incubated with CD4+ T cells. This ensures that
T cell stimulation is due to antigen presentation by DCEKs
and not direct stimulation by the H57 antibody.
18. This quantification consists of first detecting T cell:DCEK

contacts and considering whether they are effective IS, i.e.,
positive for CD3ζ accumulation at that area. If so, the next
step is to detect whether TCRs that were on the plasma mem-
brane prior to DCEK presentation have become internalized,
i.e., to quantify vesicles that are positive for both antibodies
used for intra- and extracellular TCR staining (anti-CD3ζ and
anti-βTCR, respectively).
References
1. Dustin ML (2002) The immunological syn- and terminated in the central supramolecular
apse. Arthritis Res 4(Suppl 3):S119–S125 activation cluster. Immunity 25(1):117–127.
2. Monks CR, Freiberg BA, Kupfer H, Sciaky N, doi:10.1016/j.immuni.2006.04.010
Kupfer A (1998) Three-dimensional segrega- 9. Alcover A, Alarcon B (2000) Internalization
tion of supramolecular activation clusters in T and intracellular fate of TCR-CD3 complexes.
cells. Nature 395(6697):82–86. Crit Rev Immunol 20(4):325–346
doi:10.1038/25764 10. Cemerski S, Das J, Giurisato E, Markiewicz
3. Irvine DJ, Purbhoo MA, Krogsgaard M, Davis MA, Allen PM, Chakraborty AK, Shaw AS
MM (2002) Direct observation of ligand rec- (2008) The balance between T cell receptor
ognition by T cells. Nature 419(6909):845– signaling and degradation at the center of the
849. doi:10.1038/nature01076 immunological synapse is determined by anti-
4. Krogsgaard M, Li QJ, Sumen C, Huppa JB, gen quality. Immunity 29(3):414–422.
Huse M, Davis MM (2005) Agonist/endoge- doi:10.1016/j.immuni.2008.06.014
nous peptide-MHC heterodimers drive T cell 11. Monjas A, Alcover A, Alarcon B (2004)
activation and sensitivity. Nature Engaged and bystander T cell receptors are
434(7030):238–243. doi:10.1038/ down-modulated by different endocytotic
nature03391 pathways. J Biol Chem 279(53):55376–55384.
5. Ehrlich LI, Ebert PJ, Krummel MF, Weiss A, doi:10.1074/jbc.M409342200
Davis MM (2002) Dynamics of p56lck translo- 12. Martinez-Martin N, Fernandez-Arenas E,
cation to the T cell immunological synapse fol- Cemerski S, Delgado P, Turner M, Heuser J,
lowing agonist and antagonist stimulation. Irvine DJ, Huang B, Bustelo XR, Shaw A,
Immunity 17(6):809–822 Alarcon B (2011) T cell receptor internaliza-
6. Costello PS, Gallagher M, Cantrell DA (2002) tion from the immunological synapse is medi-
Sustained and dynamic inositol lipid metabo- ated by TC21 and RhoG GTPase-dependent
lism inside and outside the immunological syn- phagocytosis. Immunity 35(2):208–222.
apse. Nat Immunol 3(11):1082–1089. doi:10.1016/j.immuni.2011.06.003
doi:10.1038/ni848 13. San Jose E, Sahuquillo AG, Bragado R, Alarcon
7. Vardhana S, Choudhuri K, Varma R, Dustin B (1998) Assembly of the TCR/CD3 com-
ML (2010) Essential role of ubiquitin and plex: CD3 epsilon/delta and CD3 epsilon/
TSG101 protein in formation and function of gamma dimers associate indistinctly with both
the central supramolecular activation cluster. TCR alpha and TCR beta chains. Evidence for
Immunity 32(4):531–540. doi:10.1016/j. a double TCR heterodimer model. Eur
immuni.2010.04.005 J Immunol 28(1):12–21
8. Varma R, Campi G, Yokosuka T, Saito T, 14. Schneider CA, Rasband WS, Eliceiri KW
Dustin ML (2006) T cell receptor-proximal (2012) NIH Image to ImageJ: 25 years of
signals are sustained in peripheral microclusters image analysis. Nat Methods 9(7):671–675
15. Kaye J, Hsu ML, Sauron ME, Jameson SC, bacterial and viral proteins that manipulate the
Gascoigne NR, Hedrick SM (1989) Selective immune system. Annu Rev Cell Biol 9:101–
development of CD4+ T cells in transgenic 128. doi:10.1146/annurev.
mice expressing a class II MHC-restricted anti- cb.09.110193.000533
gen receptor. Nature 341(6244):746–749. 17. Thornton AM (2003) Fractionation of T and B
doi:10.1038/341746a0 cells using magnetic beads. Curr Protoc
16. Scherer MT, Ignatowicz L, Winslow GM, Immunol Chapter 3:Unit 3 5A.
Kappler JW, Marrack P (1993) Superantigens: doi:10.1002/0471142735.im0305as57
Chapter 8
T Cell Receptor Activation of NF-κB in Effector T Cells:

Visualizing Signaling Events Within and Beyond
the Cytoplasmic Domain of the Immunological Synapse
Maria K. Traver, Suman Paul, and Brian C. Schaefer
Abstract
The T cell receptor (TCR) to NF-κB signaling pathway plays a critical role in regulation of proliferation
and effector T cell differentiation and function. In naïve T cells, data suggest that most or all key cytoplas-
mic NF-κB signaling occurs in a TCR-proximal manner at the immunological synapse (IS). However, the
subcellular organization of cytoplasmic NF-κB-activating complexes in effector T cells is more complex,
involving signaling molecules and regulatory mechanisms beyond those operative in naïve cells. Additionally,
in effector T cells, much signaling occurs at cytoplasmic locations distant from the IS. Visualization of
these cytoplasmic signaling complexes has provided key insights into the complex and dynamic regulation
of NF-κB signal transduction in effector T cells. In this chapter, we provide in-depth protocols for activat-
ing and preparing effector T cells for fluorescence imaging, as well as a discussion of the effective applica-
tion of distinct imaging methodologies, including confocal and super-resolution microscopy and imaging
flow cytometry.
Key words Signal transduction, T lymphocyte, T cell receptor, POLKADOTS, Bcl10, NF-κB,
Confocal microscopy, Super-resolution microscopy, Imaging flow cytometry, Retroviruses
1 Introduction
In the natural world, multicellular organisms must survive constant

challenge from invading pathogenic species to maintain normal
functioning. Mechanisms utilized in pathogen defense include
both recognition of generic pathogenic signals, such as lipopoly-
saccharide found in bacterial membranes, and recognition of anti-
gens which are specific to a particular species. This antigen-specific
immunity, facilitated by the adaptive immune system, is mediated
in all vertebrates by a class of white blood cells called lymphocytes
[1, 2]. Among this class of immune cells, T lymphocytes mediate
an antigen-specific cellular immune response. Each individual T
lymphocyte expresses a remarkable cell surface receptor, the T cell
receptor (TCR), which is produced by an orchestrated series of
101
102 Maria K. Traver et al.
gene rearrangements and mutational processes during T cell

ontogeny [3]. As a result of this developmental process, each TCR
in the naïve repertoire has a unique sequence and antigen-binding
specificity [3]. When presented with a cognate peptide antigen in
the context of the major histocompatibility complex (MHC) found
on antigen-presenting cells, T lymphocytes initiate a complex sig-
nal transduction pathway, leading to activation of a number of key
transcription factors [4–6]. A cytoplasmic signaling pathway of
particular importance for T cell activation is the cascade which cul-
minates in the translocation of the transcription factor NF-κB to
the nucleus [7]. There, NF-κB initiates a gene expression program
that plays an essential role in expansion and differentiation of the
antigen-specific T cell into clonal effector cells, which direct the
cellular immune response to the specific threat [8].
Given its importance to the functioning of the adaptive immune
system, the TCR-to-NF-κB signaling pathway has been extensively
scrutinized, and much has been discovered about the mediators and
mechanisms of action of intermediate signaling steps (Fig. 1).
Activation of the TCR leads to phosphorylation and activation of
the θ isoform of protein kinase C (PKCθ), which phosphorylates
and activates a multi-domain scaffold protein called CARMA1 (also
known as CARD11) [7]. Upon activation, CARMA1 forms a
Antigen
Receptor
PKCθ/β CARMA1
BCL10 CBM
Forming MALT1 Complex
Auto-
phagosome
p62
1 MALT1 TR
U U MALT AF6 POLKADOTS
t
Filamen
BCL10 IKK Signalosome
UU MALT1MA
LT1 P
FIGURE LEGEND
activation signal
Gene P activatory
transcription phosphorylation
U K63-poly-
ubiquitination
Fig. 1 The T cell receptor-to-NF-κB signaling pathway. Adapted from [7]

T Cell Receptor Activation of NF-κB 103
unique aggregate structure which nucleates the formation of

filaments composed of monomers of the adapter protein Bcl10 and
its associated signaling partner, Malt [9]. These filaments form the
core of a signaling structure called the POLKADOTS (punctate
and oligomeric killing or activating domains transducing signals)
signalosome, which recruits signaling partners including the IκB
kinase (IKK), a complex of proteins containing two kinases, IKKα
and IKKβ, and a dimer of the non-catalytic subunit, IKKγ [10–12].
IKK is activated by phosphorylation of IKKβ and ubiquitination of
IKKγ [6, 13]. The activated IKK complex then phosphorylates
IκBα, the binding partner of NF-κB that sequesters this complex in
the cytoplasm, preventing its function as a transcription factor [8].
IKK-mediated phosphorylation triggers ubiquitination and degra-
dation of IκBα by the proteasome, which frees NF-κB to translocate
to the nucleus, initiating transcription of target genes [8].
Much of the above description of the TCR-to-NF-κB signal
transduction pathway is based on biochemical analyses. While valu-
able, a biochemistry-only approach to signaling yields only an
understanding of events averaged across millions of cells. A more
sophisticated and complete understanding of this complex signal-
ing pathway can be obtained through fixed and live cell micro-
scopic visualization of signaling events. Fluorescence microscopy
analysis of activated T cell lines and primary effector T cells has
demonstrated that TCR activation of NF-κB initiates a series of
protein redistribution events both proximal to and distal from the
immunological synapse (IS), the junction between the T cell and
the stimulatory antigen-presenting cell (APC) [14, 15]. Within
seconds following engagement of the TCR, PKCθ becomes
enriched at the cytoplasmic domain of the IS [14]. Within minutes
of TCR activation, CARMA1 and Bcl10 also become enriched at
the IS, colocalizing with PKCθ [16]. However, the IS enrichment
of Bcl10 is transient. During the time frame of approximately
10–60 min post-TCR activation, Bcl10 is primarily found enriched
in the POLKADOTS signalosome [10]. Assembly of the
POLKADOTS signalosome requires TCR-stimulated K63-
polyubiquitination of Bcl10, stimulating recruitment of the Bcl10-
Malt1 complex to preexisting cytoplasmic aggregates of the adapter
protein, p62, which are present throughout the cytoplasm [17].
Over time, the POLKADOTS signalosomes become aggregated
beneath the immunological synapse, during which time Bcl10 is
degraded by selective autophagy [17]. However, Malt1 persists in
the POLKADOTS signalosome for 3–4 hr post-TCR stimulation
as a single aggregate in the cytoplasmic domain beneath the IS
(unpublished data). Data suggest that activation of IKK and phos-
phorylation of IκBα occur exclusively at the POLKADOTS sig-
nalosome [12]. Our most recent imaging flow cytometry studies
demonstrate that continued NF-κB nuclear translocation can be
observed until the last remnants of POLKADOTS structures have
faded, indicating continuous activation of IKK, even as

POLKADOTS mature and change (unpublished data).
In this chapter, we will describe in detail methodologies useful
in the visualization of the TCR-to-NF-κB signal transduction
pathway in a murine T-lymphocyte-derived cell line. We will out-
line the steps necessary to derive clones which stably express fluo-
rescently tagged signaling proteins, various methods of activation,
antibody staining protocols, and preparations for imaging. We will
conclude with a discussion of the pros and cons of various imaging
methodologies, including the resolving power of confocal and
super-resolution microscopy and the statistical power of imaging
flow cytometry.
2 Materials
2.1 Generation 1. Retroviral vector with IRES-driven selectable marker(s),

of D10 T Cell Clones diluted in sterile, endotoxin-free water (see Note 1).
Expressing Tagged 2. HEK293T cell line.
POLKADOTS
3. HEK293T culture media: Dulbecco’s Modified Eagle Media
Components (DMEM), high glucose, sodium pyruvate, no l-glutamine,
2.1.1 Retrovirus supplemented with 10% (v/v) FBS, and antibiotics (penicillin,
Production streptomycin, and gentamycin).
4. Tissue-culture-treated flasks, sizes T25–T225, with vented caps.
5. Tissue culture incubator, 37 °C/5% CO2.
6. Hemocytometer or other cell-counting method.
7. 6-well tissue-culture-treated plates.
8. Coating solution: 100 μg/mL high molecular weight (MW
70,000–150,000) poly-d-lysine hydrobromide in distilled
water, sterile filtered.
9. Phosphate-buffered saline (PBS).
10. Trypsin solution: 0.05% trypsin with 0.2 g/L EDTA.
11. Transfection media: Iscove’s Modified Dulbecco’s Media

(IMDM), supplemented with 10% (v/v) FBS and antibiotics
(penicillin, streptomycin, and gentamycin).
12. Sterile, endotoxin-free water.
13. pCL-Eco helper plasmid, diluted in sterile, endotoxin-free

water. Available from Addgene.
14. 2.5 M calcium chloride in endotoxin-free water, sterilized via
0.2 μm filtration.
15. 2× HEPES solution: Create a 194 mM Na2HPO4 and 105
mM NaH2PO4 solution in endotoxin-free water. Dilute 100
μL of this solution to 20 mL in endotoxin-free water (final
concentration 1.5 mM sodium phosphate). Add 140 mM
NaCl and 50 mM HEPES (free acid). Adjust final pH of solu-

tion to exactly 7.05 and sterilize via 0.2 μm filtration.
2.1.2 Infection 1. D10 murine T cell line (see Note 2).

and Selection 2. D10 culture media: Eagle Hank’s Amino Acids (EHAA) cell
culture media [also called Click’s media], supplemented with
10% (v/v) FBS, 6% (v/v) supplement solution
(see Subheading 2.1.2, item 3), and 0.5% (v/v) IL-2 solution
(see Subheading 2.1.2, item 4).
3. Supplement solution: Mix 660 mL unsupplemented EHAA,
600 mg penicillin G, 1000 mg streptomycin sulfate, 500 mg
gentamycin sulfate, 5.84 g l-glutamine, 12 g sodium HEPES,
15.2 g HEPES (free acid), and 38 μL β-mercaptoethanol in a
1000 mL beaker in a chemical fume hood. Adjust pH to 7.4.
Filter solution through a 0.22 μm filter. Store aliquots at −20
°C and thaw at 37 °C.
4. IL-2 solution: 200 ng/mL murine IL-2 in EHAA with 10%
(v/v) FBS. Store aliquots long term at −20 °C or for up to 4
months at 2–8 °C.
8. Temperature-controlled centrifuge with appropriate adaptors
for conical tubes and 24-well plates.
9. 10 mg/mL polybrene in sterile distilled water (1000×).
10. 24-well tissue-culture-treated plates.
11. Zipper seal bags.
12. Selection antibiotics (see Note 3).
13. Freezing media: Heat-inactivated FBS with 10% (v/v) tissue-
culture-certified DMSO.
14. 2 mL cryotubes.
2.1.3 Subcloning 1. D10 murine T cell line (see Note 2).

2. D10 culture media: see Subheading 2.1.2, item 2.
6. 6-well and 96-well tissue-culture-treated plates.
7. Selection antibiotics (see Note 3).
8. Sterile pipet basin.
9. 8/12 channel pipettor.
2.2 Preparation 1. Petri dishes, 60 mm.

of Activated D10 T 2. Square glass coverslips, 18 × 18 mm, #1.5 thickness.
Cells for Imaging
3. Etching Solution: 1 M HCl in 70% ethanol.
2.2.1 Stimulation 4. Coating Solution: 100 μg/mL high molecular weight (MW
with Anti-CD3 70,000–150,000) poly-d-lysine hydrobromide in distilled water.
5. Tissue-culture-treated dishes, 35 mm.
6. Antibody solution: 100 μg/mL anti-CD3 antibody (clone
145-2C11) in sterile PBS.
7. D10 murine T cell line (see Note 2).
12. Freezer stock, poly-d-lysine-coated coverslips: Create coating
solution of 3 mg/mL high molecular weight (MW 70,000–
150,000) poly-d-lysine hydrobromide in distilled water. Fill a
100 mm petri dish with 10 mL of this coating solution. Place
approximately 30 coverslips (see Subheading 2.2.1, item 2) in
the dish; make sure each coverslip is fully submerged. Incubate
for 6–8 h at ambient temperature, then wrap dish in parafilm,
and store flat at −20 °C. Thaw at ≤37 °C. When opening,
ensure minimal water loss from evaporation and condensation
to prevent changes in poly-d-lysine concentration.
2.2.2 Stimulation 1. D10 murine T cell line (see Note 2).

with Antigen-Presenting 2. CH12 murine B cell line (see Note 4).
Cells
4. CH12 culture media: same as D10 culture media but without
IL-2 solution.
5. 10 mg/mL conalbumin in Hank’s balanced salt solution
(HBSS).
9. Freezer stock, poly-d-lysine-coated coverslips: see Subheading
2.2.1, item 12.
10. Petri dishes, 60 mm.
11. Tissue-culture-treated dishes, 35 mm.
2.2.3 Fixation, Staining, 1. Fix solution: 3% paraformaldehyde and 3% sucrose in PBS. To

and Slide Preparation make, heat 600–800 mL distilled water to 56 °C. Transfer to
chemical fume hood; while maintaining temperature and stir-
ring, add 30 g paraformaldehyde powder. Allow to mix thor-
oughly, then add a few drops of 10 N NaOH (just enough for
the majority of the paraformaldehyde to become soluble).
Remove from heat. Add 30 g sucrose and 100 mL of 10× PBS
and mix thoroughly. Adjust pH to about 7.5 using pH paper.
Add distilled water to a final volume of 1 L and filter through
a ≤0.45 μm filter. Store aliquots at −20 °C, and thaw at 37 °C
with careful mixing. Aliquots can survive multiple freeze-thaw
cycles if properly thawed.
2. Permeabilization solution: 0.2% Triton-X100 detergent in
1× PBS.
3. Blocking solution: HEK293T culture media without antibiotics
(see Subheading 2.1.1, item 3).
4. Refrigerated centrifuge with appropriate adaptors for 1.5 mL
microcentrifuge tubes, capable of spinning at 10,000 × g.
5. Glass microscope slides, 1 mm thick, 3 in × 1 in.
6. Mounting media: p-phenylenediamine (PPD) in 90% glycerol
(see Note 5). To make, freeze 1 mL aliquots of 20 mg/mL
PPD in 10× PBS at −80 °C, protected from light, for long-
term storage (note that PPD is toxic and light sensitive and
solubilizes slowly in PBS). Thaw one aliquot carefully, allow-
ing time for complete solubilization; add to 18 mL high qual-
ity, ≥99% pure glycerol in a conical tube and mix thoroughly
using a rotator, protected from light, for at least 10 min (add-
ing a stir bar to the tube enhances mixing, even in the absence
of magnetic stirring). Add 1 mL carbonate-bicarbonate buffer
(see Subheading 2.2.3, item 9) and mix thoroughly using a
rotator, protected from light. Test pH of medium using pH
paper (dilute medium in distilled water to ensure proper
absorption by the paper); adjust pH to approximately 8.5
using carbonate buffer (see Subheading 2.2.3, item 7) (note
that PPD at neutral pH is not an effective anti-fade reagent, so
pH adjustment is vital). Solution should be pale amber in
color. Store aliquots at −20 °C, protected from light, and
thaw at ambient temperature. Note that PPD degrades if kept
for more than a few hours at room temperature; if aliquots
become dark orange in color, discard.
7. Carbonate buffer: Add 2.12 g anhydrous sodium carbonate to
100 mL of distilled water for a 0.2 M solution.
8. Bicarbonate buffer: Add 1.68 g sodium bicarbonate to 100
mL of distilled water for a 0.2 M solution.
9. Carbonate-bicarbonate buffer: Mix 5.4 mL carbonate buffer
(see Subheading 2.2.3, item 7) and 4.6 mL bicarbonate
buffer (see Subheading 2.2.3, item 8).
10. Fine-tipped forceps and/or a sterile needle, as needed.
11. Blotting paper, cut into thin, ~20 mm strips.
12. Clear nail polish.
2.3 Preparation 1. Antibody solution: 100 μg/mL anti-CD3 antibody (clone

of Activated D10 Cells 145-2C11) in sterile 1× PBS.
for Imaging Flow 2. 6-well plates, tissue culture treated.
Cytometry
3. Zipper seal bags.
2.3.1 Stimulation 4. D10 murine T cell line (see Note 2).
with Anti-CD3
2.3.2 Sample 1. Hank’s balanced salt solution (HBSS).

Preparation 2. Refrigerated centrifuge with appropriate adaptors for 1.5 mL
microcentrifuge tubes.
3. Wash buffer: 2% (v/v) FBS in PBS.
4. 0.6 mL siliconized microcentrifuge tubes.
5. Fixation buffer: 4% paraformaldehyde in 1× PBS.
6. Permeabilization buffer: 2% (v/v) FBS and 0.1% (v/v)
Triton-X100 in 1× PBS.
3 Methods
3.1 Generation When visualizing the TCR-to-NF-κB pathway, it is often helpful to

of D10 T Cell Clones fluorescently tag proteins of interest. This is especially true when
Stably Expressing attempting super-resolution microscopy, as achieving sufficient
Fluorescently Tagged labeling density when using antibody staining can be more chal-
Proteins lenging than in conventional confocal microscopy. We strongly
suggest the use of cell lines which stably express the fluorescent
protein fusion, as opposed to repeated transient transfections, for
reasons of repeatability of experimental results. Once cell lines
which stably express fluorescently tagged protein(s) have been
derived, it is important to ensure that the fusion protein behaves
similarly to endogenous protein before proceeding. For example,
N-terminal fusions of fluorescent markers to Bcl10, which form
the filamentous core of Bcl10 filaments, are inefficiently recruited
to POLKADOTS and should not be attempted (unpublished
data). C-terminal fusions, however, behave indistinguishably from
endogenous Bcl10. In addition, when expressing fluorescently
tagged constructs in a cell, it is important to consider the effects of
single protein overexpression on the overall signaling pathway. For
example, overexpression of fluorescent Malt1, a component of the
POLKADOTS signalosome, without an equivalent overexpression
of Bcl10 leads to cells which inefficiently form POLKADOTS and
do not activate in high percentages. Expression of both Malt1 and
Bcl10 in cells leads to efficient activation, although POLKADOTS
filaments appear longer than those found in parental cell lines
(unpublished data). It is thus important to compare results

obtained by overexpression of fluorescently tagged protein with
results using antibody staining of endogenous protein to ensure
that no artifacts have been introduced to the system.
In this section, we outline the steps for using retroviral trans-
duction to derive D10 T cell lines which stably express fluorescent
proteins. We begin with the production of the recombinant retro-
virus and then describe the transduction and selection steps. Finally,
we discuss the generation of stable, monoclonal cell lines with
desired characteristics for imaging.
3.1.1 Retrovirus 1. Utilizing standard molecular biology techniques for DNA

Production manipulation, fuse your choice of fluorescent protein directly
to your desired POLKADOTS signalosome cDNA sequence.
Inclusion of a spacer of several amino acids is recommended.
2. Clone the fused sequence into a retroviral vector. Ensure that
vector design includes a drug-selectable marker cloned
downstream of the cDNA-fluorescent protein fusion via an
IRES cassette (see Note 1).
3. Begin a culture of HEK293T cells in HEK293T culture media
(see Subheading 2). Ensure cells are growing at a healthy and
stable rate before continuing with virus production.
4. Add 1.5 mL coating solution (see Subheading 2) to the desired
number of wells in a 6-well plate. Incubate for 10 min at ambi-
ent temperature.
5. Aspirate coating solution and rinse each well 5 times with
1× PBS.
6. Aspirate media from HEK293T cell culture and cover with
trypsin solution (see Subheading 2). Incubate at 37 °C/5%
CO2 for 5 min. Add an equal amount of media to inactivate
trypsin and pipet repeatedly to ensure complete cell detach-
ment. Transfer cells to a conical tube and determine cell
number.
7. Pellet cells at 500 × g for 5 min, and then resuspend in fresh
media at a density of 6 × 105 cells/mL. Plate 1 mL of cells in
each pre-coated well of the 6-well plate. Incubate overnight at
37 °C/5% CO2.
8. Prewarm transfection media to 37 °C (see Subheading 2).
Aspirate media from HEK293T cells and add 1.5 mL transfec-
tion media per well. Incubate at 37 °C/5% CO2 until step 11.
9. Assemble a master mix of transfection material in a 1.5 mL
microcentrifuge tube as follows and in the indicated order:
(a) 2.4 μg retroviral construct from step 2 per well.
(b) 0.6 μg pCL-Eco helper plasmid (see Subheading 2) per well.
(c) 7.5 μL 2.5 M calcium chloride solution per well.
(d) Endotoxin-free water (to a total volume of 75 μL per well).
10. Add 75 μL per well 2× HEPES solution (see Subheading 2) to

the master mix, blend thoroughly by pipetting, and incubate
for exactly 1 min at ambient temperature.
11. Add 150 μL master mix to each well of HEK293T cells from
step 8, dropwise, ensuring complete coverage of well surface
area. Incubate at 37 °C/5% CO2 for 24 h.
12. Prewarm DMEM media. Aspirate media from cells and replace
with 2 mL DMEM media. Incubate at 37 °C/5% CO2 for 24 h.
13. Collect media from cells and transfer to a conical tube(s).
These retrovirus-containing media can be stored at 4 °C for
up to 2 weeks or at −80 °C indefinitely.
14. If desired, a further 2 mL of prewarmed DMEM can be incu-
bated on cells for a further 24 h at 37 °C/5% CO2. This media
will contain approximately half the viral titer of that collected
on the first day.
3.1.2 Infection 1. Begin a culture of D10 T cells in D10 culture media (see
and Selection Subheading 2). Ensure cells are growing at a healthy and stable
rate, doubling once approximately every 24 h, before continu-
ing with transduction (see Note 2).
2. Determine density of D10 cell culture. Transfer 1 × 106 total
cells to a conical tube(s). Pellet cells at 500 × g for 5 min and
aspirate media.
3. Thaw 2 mL of retroviral supernatant generated in Subheading
3.1.1. Add 20 μL IL-2 solution (see Subheading 2) for a final
concentration of 2 ng/mL IL-2, and add 2 μL of 1000× poly-
brene solution (see Subheading 2).
4. Resuspend D10 cells in retroviral supernatant solution pre-
pared in step 3. Transfer to one central well of a 24-well plate
and enclose in a zipper seal bag, taking care to remove excess
air. Centrifuge plate, taking care not to allow any loose bag
edges to create friction or drag, at 1200 × g for 2 h at ≥ ambi-
ent temperature (32 °C is ideal; a temperature-controlled cen-
trifuge will reach 32 °C within a few minutes of running, due
to heat produced by the motor and friction) to promote viral
fusion. Centrifugation at lower temperatures inhibits viral
fusion and should be avoided.
5. Transfer D10 cells to a 1.5 mL microcentrifuge tube or 15 mL
conical tube. Pellet cells at 500 × g for 5 min and aspirate
media. Resuspend cells in fresh D10 culture media at a density
of approximately 1 × 105/mL and transfer to a tissue culture
flask. Incubate at 37 °C/5% CO2 for 24 h.
6. After 24 h, add selection antibiotics as appropriate (see Note
3). Incubate at 37 °C/5% CO2 for 3–7 days as appropriate to
select for uninfected cells. A parallel culture of uninfected cells
can be used to determine when uninfected cells have been

killed. Ensure culture does not exceed a density of 5 × 105/
mL; dilute with antibiotic-containing media to a density of 1 ×
105/mL if necessary.
7. Following selection, grow remaining D10 cell culture in selec-
tion antibiotics to a size suitable for passaging and/or freezing.
Depending on transduction efficiency (which can be low in
D10 cells) and possible growth-inhibiting effects of the retro-
virally encoded transgene, this may take several weeks.
8. D10 cells may be frozen by resuspending in freezing media (see
Subheading 2), transferring to appropriate cryotubes and incu-
bating at −80 °C for ≥24 h. Following initial freezing, cells
should be stored long term in liquid nitrogen. To thaw, rapidly
warm cryotube to 37 °C, and then transfer contents to a 15
mL conical tube. Add 10 mL D10 culture media (see
Subheading 2). Pellet cells by centrifuging at 500 × g for 5 min
and aspirate media. Gently resuspend cells in 5 mL D10 cell
culture media, pellet cells by centrifuging at 500 × g for 5 min,
and aspirate media. Resuspend in 10 mL D10 cell culture
media and transfer to a T25 tissue-culture-treated flask.
Incubate at 37 °C/5% CO2 for 24 h, and then count cells. Split
if necessary and add selection antibiotics as appropriate.
Maintain culture as normal.
3.1.3 Subcloning 1. Ensure the D10 polyclonal cell line created in Subheading
3.1.2 is growing at a healthy and stable rate, doubling once
approximately every 24 h, before continuing (see Note 2).
2. Determine density of D10 cell culture. Transfer 1 × 104 total
cells to 1.5 mL microcentrifuge tube. Adjust volume of media
to 100 μL; if too much media, concentrate by centrifuging at
500 × g for 5 min, followed by aspiration of the supernatant
media down to 100 μL; if too little media, dilute by adding
additional D10 cell culture media. Resuspend cells thoroughly.
3. Add 900 μL D10 cell culture media, for a total volume of 1.0
mL of cells at a density of 1 × 104 cells/mL. Resuspend cells
thoroughly.
4. Transfer 100 μL of cell suspension from step 3 to a new 1.5
mL microcentrifuge tube. Add 900 μL D10 culture media, for
a total volume of 1.0 mL of cells at a density of 1 × 103 cells/
mL. Resuspend cells thoroughly.
5. Transfer 100 μL of cell suspension from step 4 to a new 1.5
mL microcentrifuge tube. Add 900 μL D10 culture media, for
a total volume of 1.0 mL of cells at a density of 1 × 102 cells/
mL. Resuspend cells thoroughly.
6. Transfer 500 μL of cell suspension from step 5 to a 50 mL

conical tube. Add 49.5 mL D10 culture media, for a total
volume of 50 mL of cells at a density of 1 cell/mL or 0.1
cell/100 μL. Resuspend cells thoroughly and transfer to a
sterile pipet basin.
7. Using a multichannel pipettor, plate 100 μL (0.1 cell) per well
of cell suspension from step 6 into each well of five 96-well
plates. If desired, set up one or two additional plates at a den-
sity of 1 cell/well, in case an insufficient number of clones are
recovered at 0.1 cells/well; to do this, add 100 μL of cell sus-
pension from step 4 into 9.9 mL D10 culture media, transfer
to a sterile pipet basin, and plate 100 μL (1 cell) per well in a
96-well plate. Incubate all plates at 37 °C/5% CO2 for 1 week
to allow cells time to divide and grow.
8. Check plates for wells with growing cell clones by examining
the underside of each plate. Wells with growing clones will
appear cloudy/opaque, with cells often concentrated along
one side of the well; media color change is also a good indica-
tion of growth. Make note of these wells and examine under a
microscope to confirm cell growth. If no growth is visible,
culture plates at 37 °C/5% CO2 for a further 3 days and reex-
amine. The number of wells per plate containing growth varies
significantly with the ability of D10 T cells to tolerate the ret-
rovirally encoded transgene. Also note that growth rates
between individual wells can vary significantly. Continue cul-
turing/examining plates until 3–10 clones per plate (ideally)
demonstrate healthy growth. In general, it is best not to pick
clones from plates in which more than ~20 wells show growth;
this will ensure that the great majority of clones originated
from a single founder cell.
9. Choose 15–20 wells with growing cells. Detach cells from well
bottom by gentle pipetting and transfer media and cells to
individual wells in 12-well tissue-culture-treated plates. Add
1–2 mL of D10 culture media containing the appropriate
selection of antibiotics to each well. Culture plates at 37
°C/5% CO2 for 2 days.
10. Examine cultures under a microscope to determine cell

clones with continued expansion. Transfer expanding clones
to tissue culture flasks and continue culturing at 37 °C/5%
CO2, maintaining cultures at a density between 1 × 105 and
5 × 105 cells/mL.
11. When cultures are growing at a healthy rate, examine via fluo-
rescence microscopy or flow cytometry to determine which
clones have the best activation rates and/or fluorescent con-
struct expression (see Subheading 3.2 for appropriate method-
ologies). Discard clones which do not express desired
fluorescent constructs and/or do not respond to stimulation
with anti-CD3.
3.2 Preparation Initiation of the TCR-to-NF-κB signaling pathway is achieved by

of Activated D10 T stimulating the TCR. The TCR found on the D10 cell line can be
Cells for Imaging activated by direct stimulation using antibodies against either the
TCR itself or its co-receptor CD3. We find no phenotypic difference
in cells stimulated with either of these antibodies; for the sake of
simplicity, we assume in the following sections that an anti-CD3
antibody is used. In addition, D10 T cells can be activated in a more
physiologically relevant manner by pairing them with antigen-
presenting cells (APCs) that have processed the antigen conalbumin,
which is recognized by the D10 T cell receptor. We use the murine
B cell line, CH12, as APCs for D10 cells, but other I-Ak-expressing
APCs can be substituted. For the translocation events discussed
below, we have so far observed no phenotypic difference in cells
activated by antibody versus antigen, and the choice of activation
method is largely based on experimental considerations. The earliest
steps of activation (from 5 to 20 min post-activation) are more easily
visualized via antigen stimulation, as D10-APC coupled cells adhere
more quickly to treated coverslips. For visualizing later activation
steps, including both POLKADOTS structures (visible from approx-
imately 20 to 60 min post-activation) and aggregated POLKADOTS
remnants (visible from approximately 60–180 min post-activation),
antibody stimulation is often easier and more convenient. Also, for
antibody staining experiments, expression of target antigens in APCs
can interfere with clean discrimination of signaling complexes in T
cells, particularly for complexes proximal to the IS.
In this section, we describe the methodology for both anti-
body and APC-mediated stimulation. We then describe methods
for antibody staining of cell components and preparation of sam-
ples for imaging via optical microscopy. Finally, we include a short
discussion of confocal vs. super-resolution microscopy techniques
and when each is appropriate for visualizing the TCR-to-NF-κB
signaling pathway.
3.2.1 Stimulation Day 0

with Anti-CD3 Antibody 1. Place the appropriate number of coverslips into a petri dish,
one coverslip per experimental sample (60 mm works well for
5–20 coverslips). Add 2–6 mL of etching solution (see
Subheading 2), enough to cover the coverslips. Agitate gently
to ensure complete coverage. Incubate at ambient temperature
for 15 min.
2. Rinse coverslips thoroughly (5×) with distilled water.
3. Add 2–6 mL of coating solution (see Subheading 2) to the dish,
enough to cover the coverslips. Agitate gently to ensure com-
plete coverage. Incubate at ambient temperature for 15 min.
4. Individually aspirate coverslips to dryness and place in 35 mm
dishes, one coverslip per dish (see Note 6). Ensure coverslips
are not touching the side of the dish, then carefully pipet 150
μL of antibody solution (see Subheading 2) on top of each
coverslip and cover plates. Plates can be stacked in pipet tip

box lids for stability. Insert plates with coverslips and antibody
solution inside a zipper seal bag (taking care not to displace
solution). Incubate at 4 °C overnight (16–24 h).
5. In addition to above steps, set up a flask with the proper vol-
ume of the D10 cell clone you plan to image at a density of 1
× 105/mL in D10 culture media (see Subheading 2). D10 cells
will double in density approximately every 24 h; ensure you
will have enough cells to plate 1.5 × 105 cells per coverslip with
enough left over to maintain your culture. Incubate at 37
°C/5% CO2 overnight.
Day 1
6. Rinse coated coverslips 6× with PBS. Leave final PBS wash on
coverslips to prevent drying of coverslip surface (which can
denature the antibody); aspirate this final wash immediately
before adding D10 T cells. Store plates at ambient tempera-
ture until step 9.
7. Thaw freezer stock of poly-d-lysine-coated coverslips (see
Subheading 2). Transfer the appropriate number of coverslips
for your inactivated controls to a petri dish, and rinse 5× with
distilled water (see Note 7). Individually aspirate coverslips to
dryness and place in 35 mm dishes, one coverslip per dish.
Store plates at ambient temperature until step 9.
8. Concentrate your D10 cell culture to a density of 1 × 107 cells/
mL and a volume of [150 μL × (# of coverslips) + 200 μL]. To
do this, first determine the density of the overnight D10 cell
culture. Transfer the appropriate volume of culture to a 15 mL
conical tube and pellet cells by centrifuging at 500 × g for
5 min. Aspirate supernatant media down to the appropriate
volume and resuspend cells.
9. Aspirate final wash from antibody-coated coverslips, ensuring a
dry zone exists around the entire perimeter of the coverslip (see
Note 8). Plate 150 μL of concentrated cell culture onto each
coverslip and cover plates. Plates can be carefully stacked in
pipet tip box lids for stability. Transfer all plates to a 37 °C/5%
CO2 cell culture incubator, being careful not to displace cells
from coverslips.
10. Incubate plates for the appropriate amount of time to observe
the desired signaling pathway step(s). Ensure an incubation
time of at least 20 min to allow for sufficient cell adherence.
11. Following incubation, remove plates and proceed with fixation
and processing steps found in Subheading 3.2.3.

with Antigen-Presenting 1. Set up appropriate overnight cultures, including:
Cells
(a) Flask with CH12 antigen-presenting cells at a density of 1

× 105/mL in proper media (see Subheading 2).
(b) Flask with CH12 antigen-presenting cells at a density of 1
× 105/mL in proper media plus 250 μg/mL conalbumin
(1:40 dilution stock, see Subheading 2).
(c) Flask with D10 T cells at a density of 1 × 105 cells/mL in
proper media (see Subheading 2).
Ensure you will have enough cells to plate 1.5 × 105 D10 cells
+ 1.5 × 105 CH12 cells per coverslip with enough left over to
maintain your cultures. Incubate all flasks overnight (12–24 h)
at 37 °C/5% CO2.
Day 1
2. Thaw freezer stock of poly-d-lysine-coated coverslips (see
Subheading 2). Transfer the appropriate number of coverslips
(one per experimental sample) to a petri dish, and rinse 5×
with distilled water. Individually aspirate coverslips to dryness
and place in 35 mm dishes, one coverslip per dish. Store plates
at ambient temperature until step 5.
3. Determine densities of overnight cultures. Add an appropriate
number of D10 cells (1.5 × 105 cells per coverslip) to each of
two 1.5 mL microcentrifuge tubes; add an equivalent number
of CH12 cells with or without conalbumin to each tube.
Ensure that no tube receives more than 1.5 mL total volume;
concentrate cells if necessary by centrifuging at 500 × g for 5
min, aspirating media down to an appropriate volume, and
resuspending cells.
4. Spin CH12 and D10 mixtures at ≥500 × g for 30 s to initiate
conjugate formation. A mini-microcentrifuge which acceler-
ates and decelerates very quickly is ideal, as this limits time
added to this centrifugation step. Incubate for the appropriate
amount of time to observe the desired signaling pathway
step(s) in a 37 °C water bath. If examining the earliest time
points (5–10 min post-activation), skip this incubation.
5. Ten minutes prior to the end of the incubation, spin both
CH12 and D10 mixtures at 500 × g for 1 min. Aspirate media
of both samples down to 150 μL per desired coverslip and
resuspend cells thoroughly to dissociate large cell clumps.
Pipet 150 μL of cells onto coverslips from step 2. Cover plates,
and incubate at 37 °C/5% CO2 for 10 min (or 5 min if analyz-
ing the 5 min time point).
6. Following incubation, remove plates and proceed with fixation
and processing steps found in Subheading 3.2.3.
3.2.3 Fixation, Staining, 1. Following activation as outlined in Subheadings 3.2.1 and

and Slide Preparation 3.2.2, add 1 mL of fix solution (see Subheading 2) to the edge
of each plate and swirl gently to cover the coverslip surface.

Incubate at ambient temperature for 10 min. Aspirate fix
solution and wash 3× with 1× PBS (be very careful not to add
fix solution or washes directly on top of coverslips as this will
dislodge most cells from the coverslip).
2. If no antibody staining is required, proceed to step 8. Otherwise,
add 1 mL permeabilization solution (see Subheading 2) to all
plates and incubate for 5 min at ambient temperature.
3. Aspirate permeabilization solution and wash 2× with 1×
PBS. Aspirate excess PBS and ensure a dry zone exists around
the entire perimeter of the coverslip (but it is crucial not to
allow the coverslip itself to dry; it is therefore safest to perform
this step only 2 coverslips at a time) (see Note 8). Carefully
add 150 μL of blocking solution (see Subheading 2) on top of
each coverslip, and incubate for ≥30 min at ambient
temperature.
4. Dilute primary antibody(s) to the proper concentration
(empirically determined) in ice-cold blocking solution.
Centrifuge at 10,000 × g for 10 min to precipitate any particu-
late matter. Transfer solution to a fresh 1.5 mL microcentri-
fuge tube, taking care to avoid disturbing the particulate pellet
(containing antibody aggregates) at the bottom of the tube.
5. Gently tilt plates and carefully aspirate blocking solution
directly from coverslips, taking care not to disturb cells. Add
150 μL antibody solution prepared in step 4 to top of cover-
slips. Incubate at ambient temperature, protected from light,
for ≥1 h; note that many antibodies work best if incubated
overnight (12–16 h), and initial testing of antibodies with
overnight incubation is therefore recommended.
6. Wash coverslips with 1× PBS, taking care not to dislodge cells
from coverslip, and leave final rinse on cells. If secondary
antibody(s) is not necessary, skip to step 8; otherwise, dilute
secondary antibody(s) to the proper concentration in ice-cold
blocking solution. Centrifuge at 10,000 × g for 10 min to pre-
cipitate any particulate matter. Transfer solution to a fresh 1.5
mL microcentrifuge tube, taking care to avoid disturbing the
particulate pellet at the bottom of the tube.
7. Aspirate excess PBS and ensure a dry zone exists around the
entire perimeter of each coverslip. Add 150 μL antibody solu-
tion prepared in step 6 to top of coverslips. Incubate at ambi-
ent temperature, protected from light, for ≥1 h. Wash coverslips
3× with 1× PBS and leave final wash on cells.
8. If necessary, prewash slides with distilled water and dry care-
fully with a Kimwipe. Pipet 15 μL of mounting media
(see Subheading 2) onto slide surface (see Note 5). Rinse one
coverslip with distilled water; do not aspirate rinse. Using
fine-tipped forceps and/or a sterile needle, carefully lift cov-

erslip from plate (keep surface with cells facing up). Wipe the
bottom (non-cell side) with a Kimwipe or equivalent lab-
grade absorbent tissue to remove excess water. Tilt coverslip
at a 45° angle, with cell side facing down, and touch to
mounting media droplet on slide, allowing media to spread
along the edge of the coverslip. Slowly lower coverslip to the
slide surface, allowing media to spread along the entire cov-
erslip area. If necessary, gently push on top of coverslip with
forceps to eliminate bubbles that are present in the mount-
ing media between the slide and coverslip. Using blotting
paper strips (see Subheading 2), absorb excess mounting media
and residual water from the sides of coverslip.
9. Repeat step 8 for all coverslips; note that up to two coverslips
can be mounted on one slide. Once all coverslips are mounted
and excess water is removed, paint edges of coverslips with
clear nail polish. Allow all slides to dry at ambient temperature
for 10–15 min (a chemical fume hood works well to speed
drying). Store slides at −20 °C until ready for imaging.
3.2.4 Imaging The most appropriate optical microscope technology for imaging
Methodologies D10 cells depends on the level of resolution a particular experi-
ment requires. We do not, in general, recommend wide-field fluo-
rescent imaging of POLKADOTS structures, as the level of
resolution provided is suboptimal, given the smaller cell volume
and larger cell nucleus found in most T cell lines. For visualizing
aggregated POLKADOTS remnants found at later time points
post-activation, confocal microscopy is an appropriate tool and
provides sufficient resolution for colocalization studies. However,
while confocal microscopy is able to resolve some POLKADOTS
filaments, delineation of filament subregions requires utilizing
super-resolution microscopy techniques. We recommend struc-
tured illumination microscopy (SIM), as sample preparation is
identical to wide-field and confocal sample preparation as outlined
in the above sections, yet the resolution provided is approximately
double the maximum possible resolution of conventional confo-
cals. Commercial SIM-capable microscopes are available from sev-
eral major manufacturers. Figure 2 demonstrates the differences in
confocal and SIM images of POLKADOTS filaments, illustrating
that the higher resolution obtained using SIM microscopy pro-
vides better visualization of the nanoscale organization of the
POLKADOTS structure. It should be noted that SIM microscopy
uses more powerful lasers and an increased number of scans in
comparison to standard confocal microscopes, and thus great care
should be taken in choosing strong anti-fade reagents to protect
the sample.
3.3 Preparation Confocal and/or super-resolution microscopy provides the oppor-

of Activated D10 Cells tunity to visualize the TCR-to-NF-κB signaling pathway in high
for Imaging Flow resolution, allowing for detailed analysis of signaling structures and
Cytometry their subdomains. However, obtaining statistically significant
quantitative data from microscopy, such as colocalization analysis
or percentage of NF-κB nuclear translocation, can be a long and
tedious endeavor involving capturing and separately analyzing
hundreds of images. Technological advancements in the last decade
have led to a new imaging technology, imaging flow cytometry,
which allows the capture of hundreds of images per minute, along
with analysis software packages which produce quantitative mea-
surements that can be applied to the entire datasets in a matter of
minutes. However, due to the speed of imaging, a limitation is that
only wide-field images can be obtained, so measurements requir-
ing very fine detail are not possible. We find, however, that analysis
of measurements such as NF-κB nuclear translocation in response
to various treatments is greatly facilitated by the use of imaging
flow cytometry.
In this section, we outline the steps for activation of D10 T
cells and preparation for imaging using an Amnis ImageStream®X
imaging flow cytometer, one of two currently commercially avail-
able imaging flow cytometers. We then discuss in more detail how
to analyze NF-κB nuclear translocation using the Amnis IDEAS®
software package.
Brightfield Bcl10 Malt1 Merged
Zeiss 710
Confocal
Zeiss ELYRA
SIM
Fig. 2 Comparison of confocal and structured illumination microscopy (SIM) methodologies for visualizing
POLKADOTS. D10 murine T cells expressing Bcl10-CFP and Malt1-YFP were activated for 20 min and then fixed
and processed for imaging according to the procedures outlined in this chapter. Cells were imaged on a Zeiss 710
confocal microscope or a Zeiss ELYRA PS.1 microscope in SIM mode, as indicated. Scale bars = 5 μm

with Anti-CD3 Antibody 1. Add 1.5 mL of antibody solution (see Subheading 2) to the
appropriate number of wells in a tissue-culture-treated 6-well
plate(s). Insert plate(s) inside a zipper seal bag. Incubate at 4
°C overnight (16–24 h).
2. Select a D10 cell clone for analysis, and seed at a density of 1 ×
105/mL in the proper media (see Subheading 2). D10 cells will
double in density approximately every 24 h; ensure you will
have enough cells to have 3 × 106 cells per sample with enough
left over to maintain your culture. Incubate at 37 °C/5% CO2
overnight.
Day 1
3. Wash anti-CD3-coated wells 6× with 1× PBS. Leave final PBS
wash on wells to prevent drying of well surface (which can
denature the antibody). Aspirate the last wash immediately
before adding D10 T cells. Ensure you have enough uncoated
wells for any inactivated control samples. Store all plates at
ambient temperature until step 6.
4. Concentrate the D10 cell culture to a density of 1.5 × 106
cells/mL and a volume of (2 mL × (# of wells) + 1 mL). To do
this, first determine the density of the overnight D10 cell cul-
ture. Transfer the appropriate volume of culture to a conical
tube(s) and pellet cells by centrifuging at 500 × g for 5 min.
Aspirate supernatant media down to the appropriate volume
and resuspend cells. If the purpose is to examine NF-κB trans-
location to the nucleus, it is helpful to aspirate all supernatant
media and resuspend cells in media without IL-2 at this stage
to ensure a more physiological basal NF-κB translocation rate.
5. Plate 2 mL of cells in each desired well; ensure you have
enough wells for your experimental samples and for single-
stained and unstained controls. Transfer all plates to a 37
°C/5% CO2 cell culture incubator. Incubate plates for the
appropriate amount of time called for by your experiment (the
useful range for such experiments is generally 5 min–4 h).
Following incubation, proceed with sample preparation steps
outlined in Subheading 3.3.2.
3.3.2 Sample 1. Following activation, incubate plates on ice for ≥10 min.

Preparation 2. Pipet cells from the bottom of the dish into the media, taking
particular care for cells activated for 0–60 min as these cells are
particularly adherent. Transfer media and cells to 1.5 mL micro-
centrifuge tubes. Rinse wells carefully with HBSS to ensure
complete recovery and transfer rinse to microcentrifuge tubes.
3. Centrifuge microcentrifuge tubes at 500 × g for 5 min at 4 °C
to pellet cells. Aspirate supernatants, leaving approximately 15
μL to avoid discarding any cells. Resuspend in ice-cold wash
buffer (see Subheading 2) and consolidate like samples into 0.6

mL siliconized microcentrifuge tubes. Siliconized microcen-
trifuge tubes are necessary to ensure minimal loss of cell pellets
during wash steps. 0.6 mL microcentrifuge tubes are the pre-
ferred size as they minimize the amount of supernatant vol-
ume required, which minimizes cell loss. Keep tubes on ice as
much as possible throughout the remainder of this protocol,
except where otherwise noted (see Note 9).
4. Centrifuge all tubes at 500 × g for 5 min at 4 °C to pellet cells.
Aspirate supernatant to a total volume of 50 μL. If no cell
surface markers are to be stained, skip to step 5; otherwise,
dilute desired cell surface antibody(s) to twice the proper con-
centration (empirically determined) in ice-cold wash buffer
(see Note 10). Add 50 μL of either antibody solution or wash
buffer to tubes and carefully resuspend cells. Incubate at ambi-
ent temperature, protected from light, for 20 min.
5. Centrifuge all tubes at 500 × g for 5 min at 4 °C to pellet cells.
Aspirate supernatants, leaving approximately 15 μL. Resuspend
cells in 100 μL fixation buffer (see Subheading 2) and incubate
at ambient temperature for 10 min.
6. Add 500 μL wash buffer to each tube and centrifuge at 500 ×
g for 5 min at 4 °C to pellet cells. Aspirate supernatants.
Repeat. If necessary, at this step, tubes can be filled with wash
buffer and stored at 4 °C overnight.
7. Dilute desired primary antibody(s) for intracellular antigens
(e.g., NF-κB subunits) to twice the desired concentration in
ice-cold permeabilization buffer (see Subheading 2). Centrifuge
tubes at 500 × g for 5 min at 4 °C to pellet cells. Aspirate
supernatants down to a total volume of 50 μL and add 50 μL
of either antibody dilution or wash buffer. Carefully resuspend
cells and incubate at ambient temperature, protected from
light, for 20 min.
8. Add 500 μL permeabilization buffer to each tube and centri-
fuge at 500 × g for 10 min at 4 °C to pellet cells. The longer
spin is necessary to ensure minimal cell loss following
permeabilization.
9. If no secondary antibody is needed, skip to step 10. Otherwise,
dilute desired secondary antibody(s) to twice the desired con-
centration in ice-cold permeabilization buffer. Aspirate super-
natants down to a total volume of 50 μL, and add 50 μL of
either antibody dilution or wash buffer. Carefully resuspend
cells and incubate at ambient temperature, protected from
light, for 20 min.
10. Add 500 μL permeabilization buffer to each tube and centri-
fuge at 500 × g for 10 min at 4 °C to pellet cells. The longer
spin is necessary to ensure minimal cell loss following
permeabilization.
11. Aspirate supernatants. Add 500 μL wash buffer to each tube

and centrifuge at 500 × g for 5 min at 4 °C to pellet cells.
Repeat.
12. Aspirate supernatants. Add 500 μL wash buffer to each tube
and store at 4 °C until ready to image.
13. Just prior to imaging, centrifuge all tubes at 500 × g for 5 min
at 4 °C to pellet cells. Aspirate supernatants down to a total
volume of 25 μL. If DNA/chromatin staining is necessary,
dilute nucleic acid dye to twice the desired concentration in 1×
PBS. Add 25 μL of either 1× PBS or diluted nucleic acid dye to
each tube and carefully resuspend cells. Samples are now ready
for imaging (see Note 11).
3.3.3 Imaging There are currently two commercially available imaging flow
and Analysis Methods cytometers, the FlowSight® and the ImageStream®X, both from the
Amnis Corporation. We recommend using the ImageStreamX, as it
has the capability of providing higher magnification images, and
thus somewhat higher resolution, which is important when imag-
ing smaller cells such as T cells. The IDEAS® software that accom-
panies both instruments is excellent for image analysis and for basic
or common analyses is extremely intuitive and user-friendly. The
software comes with multiple built-in analysis “wizards” which
guide the user through commonly used analyses such as colocaliza-
tion, spot counting (useful for specific vesicle counts), and nuclear
translocation of a transcription factor (such as NF-κB). In addition,
the ability to make complex or custom analyses using a wide variety
of tools is available, although these analyses require a more
advanced understanding of the software.
We currently utilize the ImageStreamX mainly to examine
NF-κB nuclear translocation in response to various treatments.
The accompanying IDEAS® software has a built-in “nuclear local-
ization” analysis wizard, which is a useful tool for quickly deter-
mining qualitative differences in NF-κB translocation between two
populations. The nuclear localization wizard relies on two concur-
rent stains in the sample, one a nuclear dye of choice and the other
a stained component of the transcription factor of interest. We uti-
lize antibody staining of p65/RelA, a component of the canonical
NF-κB heterodimer, [8] and the nuclear dye DRAQ5. The analysis
wizard then determines the log-transformed Pearson’s correlation
coefficient comparing the staining intensity of NF-κB and the
nuclear dye over the area of the cell. Thus each cell is assigned a
number between approximately −2.5 and 4.0, with lower numbers
indicating greater dissimilarity (and therefore less nuclear NF-κB).
While this analysis is easy to perform for beginner users of the soft-
ware and provides a semiquantitative comparison of nuclear trans-
location, we prefer to use a custom analysis to measure the average
ratio of nuclear/cytoplasmic NF-κB for reporting in the
literature.
To perform this more advanced analysis, first use the built-in

“building block” analyses to separate out only images of cells which
are in focus, single, and stained for both NF-κB and the nucleus.
Next, use the mask manager to create separate nuclear and cyto-
plasmic masks. We have empirically found that the results of the
analysis are more accurate when masks are created with enough
spatial separation to prevent overlapping signals caused by the lim-
ited resolving power of wide-field microscopy from distorting the
analysis. To achieve this, create the following mask formulae:
Nuclear Mask Threshold 75: Threshold(M05, Ch05 DRAQ5, 75)
Nuclear Mask Threshold 30: Threshold(M05, Ch05 DRAQ5, 30)
Cell Area Threshold: Threshold(M02, Ch02 RelA, 70)
Cytoplasm Threshold: Cell Area Threshold And Not Nuclear Mask
Threshold 75
Note that these formulae assume that the NF-κB staining
intensity was measured in channel 2, the nuclear dye was measured
in channel 5, and the channels have been named using the “Image
Gallery Properties” menu; adjust the formulae as necessary to
reflect the channel names and numbers in your experiment.
Once these masks have been created, open the Feature Manager
menu and create three new features (the term the IDEAS software
uses for mathematical calculations), as follows:
Average Intensity Cytoplasm: Type, Single/Mean Pixel; Mask,
Cytoplasm Threshold; Image, Ch02 RelA
Average Intensity Nucleus: Type, Single/Mean Pixel; Mask,
Nuclear Mask Threshold 30; Image, Ch02 RelA
Nuclear to Cytoplasmic Ratio: Type, Combined; Description,
Average Intensity Nucleus/Average Intensity Cytoplasm
(Once again, remember to change the channel names/num-
bers to suit your experimental parameters.)
Together, these masks and features calculate the average pixel
intensity of NF-κB staining in two areas: first, over the most intense
30% of nuclear dye-stained pixels, which correspond to NF-κB in
the nucleus, and, second, over the most intense 70% of NF-κB-
stained pixels which do not overlap with the most intense 75% of
nuclear dye-stained pixels (i.e., NF-κB staining outside the
nucleus). These numbers were empirically determined to provide
the best separation between nuclear and cytoplasmic NF-κB stain-
ing, which enables the collection of data which are accurate and
repeatable.
Once the proper masks and features have been created and
calculated, the nuclear-to-cytoplasmic ratio will be calculated for
every image in the dataset. Data can be presented as histograms or
an average ratio of nuclear/cytoplasmic NF-κB. For an example of
the completed analysis, as performed on D10 cells, see Fig. 3.
Fig. 3 Analysis of nuclear translocation of NF-κB using the ImageStreamX. (a) Sample images of two cells. D10
murine T cells were activated for 90 min and then fixed and stained with anti-RelA antibody and the DRAQ5
nucleic acid dye according to the procedures outlined in this chapter. Cells were imaged using the ImageStreamX.
The first two columns display the compensated images obtained by the instrument. The third column merges
the RelA and nucleic acid images. The fourth column overlays the nuclear and cytoplasmic masks calculated
according to the procedures outlined in this chapter over the RelA staining image; the calculated ratio of
nuclear to cytoplasmic RelA staining intensity is listed in yellow. (b) Histograms of nuclear-to-cytoplasmic RelA
ratios over time. D10 murine T cells were activated for the indicated times, then fixed and stained with anti-
RelA antibody and the DRAQ5 nucleic acid dye according to the procedures outlined in this chapter, and
imaged using the ImageStreamX. The ratio of nuclear to cytoplasmic RelA staining intensity was calculated for
each cell according to the procedures outlined in this chapter. Histograms of the frequency of each ratio over
the population of cells at a given time point are displayed, along with tables presenting the number of cells
imaged per time point and the percentage of cells with a ratio ≥1.00, indicating a prevalence of nuclear RelA
4 Notes
1. We have had success using the previously described PE vector

[18]. Various commercially available vectors would also be
suitable.
2. We use a rapidly growing subclone of the TH2 clone D10.
G4.1 [19], which we refer to simply as D10 in this protocol.
We have also described protocols for performing such experi-
ments using primary murine effector T cells [20]. In principle,
any source of effector T cells can be used for these experi-
ments. Note that we strongly advise against using Jurkat T

cells, as the origin of this tumor cell line is uncertain, and the
subcellular organization of Jurkat cells is abnormal, as is the
case with many tumor lines. D10 cells should be cultured in
tissue-culture-treated flasks with vented caps at 37 °C/5%
CO2. Density of cells should be maintained between 1 × 105/
mL and 5 × 105/mL for optimal growth. Density should not
be allowed to exceed 9 × 105/mL at any time, as activation
phenotypes become permanently altered. Cells should not be
cultured longer than 6 weeks due to increased phenotypic
variation with longer culture times.
3. The empirically determined optimal concentration of antibiot-
ics for use with common selectable markers for D10 cells is as
follows: G418, 1 mg/mL; hygromycin, 0.3 mg/mL; and zeo-
cin, 25 μg/mL.
4. CH12 cells should be cultured in tissue-culture-treated flasks
with vented caps at 37 °C/5% CO2. Density of cells should be
maintained between 1 × 105/mL and 5 × 105/mL for optimal
growth. Cells should not be supplemented with IL-2 as this
causes CH12 cell death.
5. We have used multiple types of mounting media for visualiz-
ing the TCR-to-NF-κB signaling pathway, including several
commercially available media which are popular for fluores-
cence microscopy. We find that most media which contain
anti-fade reagents work well for confocal microscopy. However,
super-resolution microscopy requires the use of more power-
ful lasers than those utilized for confocals, and we find that
many commercially available anti-fade reagents are less able to
protect samples under these conditions than our own PPD-
containing mounting medium. In addition, we are wary of the
artifacts introduced into samples by hardening media, espe-
cially when examining thin, fragile structures (such as cyto-
skeletal filaments) using super-resolution microscopy. Ensure
that the mounting media you choose are appropriate for the
experimental goals you hope to achieve and be wary of intro-
duced artifacts.
6. Ensure total dryness of coverslips before transferring to plates;
any moisture will lead to coverslips becoming strongly adhered
to the bottom of the plate, making eventual removal for
mounting difficult.
7. Poly-d-lysine-coated coverslips are very sticky; never transfer
to completely dry plates, or they will become adhered to the
bottom and never be removable. Always coat plates with a
layer of distilled water prior to coverslip transfer, and rinse
excess poly-d-lysine with multiple washes of distilled water to
prevent complete adherence.
8. This dry zone ensures that no solution runs off the coverslip,
allowing the total volume to be in contact with the coverslip
surface. Solution not on the coverslip, i.e., solution that has
run off and is on the plate surface, is wasted. If solution does
run off the coverslip at any point during the incubation, espe-
cially cell solutions in the process of activation or expensive
antibody dilutions, the solution can be transferred to a micro-
centrifuge tube while the plate bottom is aspirated fully and
the dry zone reestablished. The solution can then be re-added
to the coverslip, where it should stay put.
9. The most difficult part of this entire procedure is minimizing cell
loss. Great care should be taken during all wash steps to avoid
aspiration of any part of the pellet. Even small losses add up over
the course of many washes. Some loss is inevitable, but the goal
is to have at least 1 × 106 cells per sample for imaging. Data
acquisition on the ImageStreamX is fastest and easiest when cells
are extremely concentrated; the lower the concentration, the
longer data acquisition takes, and thus significant cell loss can
extend data acquisition time for large numbers of samples by
several hours. It should be noted that cell loss appears greatest in
cells at early activation time points (between 20 and 90 min post-
activation), especially following the permeabilization step, so the
greatest care should be taken with these samples.
10. We have empirically noted that the proper antibody dilution
for imaging flow cytometry often differs significantly from
that used for fluorescence microscopy. Due to the significant
number of wash steps, cell surface markers often require a
higher concentration of antibody than usual to obtain suffi-
cient staining. However, staining of internal proteins often
requires significantly less antibody for sufficient labeling inten-
sity; this is especially true when several fluorescent channels
are utilized in an experiment, as smaller antibody concentra-
tions are less likely to cause fluorescence bleeding into other
channels and thus require less compensation post-processing.
11. Occasionally, certain samples (particularly inactivated samples)
can be far more densely concentrated than others, leading to
significant numbers of cell aggregates being imaged by the
ImageStreamX instead of individual cells. Samples which tend
toward dense concentration and aggregation, such as
inactivated samples, can be diluted further with PBS (100–
300 μL total volume instead of 50 μL often works well). Total
sample volume will not affect data collected by the
ImageStreamX. In addition, ensure that cells are thoroughly
resuspended to minimize cell aggregates before data collec-
tion begins. Briefly vortexing each sample just prior to loading
onto the ImageStreamX will help ensure uniform distribution
of cells in the sample and increase usable data. We have not
found that filtering cells prior to data collection, as is usual

practice in many types of flow cytometry, has any effect on
data collection quality.
Acknowledgments
Supported by a grant from the US National Institutes of Health

(GM109887-01 to BCS), an Irvington postdoctoral fellowship (to
MT) from the Cancer Research Institute, and predoctoral fellow-
ships (to SP) from the American Heart Association (10PRE3150039)
and the Henry M. Jackson Foundation for the Advancement of
Military Medicine. The opinions or assertions contained herein are
the private ones of the authors and are not to be construed as offi-
cial or reflecting the views of the Department of Defense, the
Uniformed Services University of the Health Sciences, or any other
agency of the US government.
References
1. Anderson MK, Pant R, Miracle AL, Sun X, Egelman EH, Wu H (2013) Structural archi-
Luer CA, Walsh CJ, Telfer JC, Litman GW, tecture of the CARMA1/Bcl10/MALT1 sig-
Rothenberg EV (2004) Evolutionary origins nalosome: nucleation-induced filamentous
of lymphocytes: ensembles of T cell and B cell assembly. Mol Cell 51(6):766–779
transcriptional regulators in a cartilaginous 10. Rossman JS, Stoicheva NG, Langel FD,
fish. J Immunol 172(10):5851–5860 Patterson GH, Lippincott-Schwartz J, Schaefer
2. Hirano M, Guo P, McCurley N, Schorpp M, BC (2006) POLKADOTS are foci of func-
Das S, Boehm T, Cooper MD (2013) tional interactions in T-cell receptor-mediated
Evolutionary implications of a third lympho- signaling to NF-kappaB. Mol Biol Cell
cyte lineage in lampreys. Nature 17(5):2166–2176
501(7467):435–438 11. Kingeter LM, Schaefer BC (2008) Loss of pro-
3. Nikolich-Zugich J, Slifka MK, Messaoudi I tein kinase C theta, Bcl10, or Malt1 selectively
(2004) The many important facets of T-cell impairs proliferation and NF-kappa B activa-
repertoire diversity. Nat Rev Immunol tion in the CD4+ T cell subset. J Immunol
4(2):123–132 181(9):6244–6254
4. Rudolph MG, Wilson IA (2002) The specific- 12. Paul S, Traver MK, Kashyap AK, Washington
ity of TCR/pMHC interaction. Curr Opin MA, Latoche JR, Schaefer BC (2014) T cell
Immunol 14(1):52–65 receptor signals to NF-kappaB are transmitted
5. Smith-Garvin JE, Koretzky GA, Jordan MS by a cytosolic p62-Bcl10-Malt1-IKK signalo-
(2009) T cell activation. Annu Rev Immunol some. Sci Signal 7(325):ra45
27:591–619 13. Wang C, Deng L, Hong M, Akkaraju GR,
6. Thome M, Charton JE, Pelzer C, Hailfinger S Inoue J, Chen ZJ (2001) TAK1 is a ubiquitin-
(2010) Antigen receptor signaling to dependent kinase of MKK and IKK. Nature
NF-kappaB via CARMA1, BCL10, and 412(6844):346–351
MALT1. Cold Spring Harb Perspect Biol 14. Monks CR, Freiberg BA, Kupfer H, Sciaky N,
2(9):a003004 Kupfer A (1998) Three-dimensional segrega-
7. Paul S, Schaefer BC (2013) A new look at T tion of supramolecular activation clusters in T
cell receptor signaling to nuclear factor- cells. Nature 395(6697):82–86
kappaB. Trends Immunol 34(6):269–281 15. Dustin ML, Chakraborty AK, Shaw AS (2010)
8. Oh H, Ghosh S (2013) NF-kappaB: roles and Understanding the structure and function of
regulation in different CD4(+) T-cell subsets. the immunological synapse. Cold Spring Harb
Immunol Rev 252(1):41–51 Perspect Biol 2(10):a002311
9. Qiao Q, Yang C, Zheng C, Fontan L, David L, 16. Schaefer BC, Kappler JW, Kupfer A, Marrack P
Yu X, Bracken C, Rosen M, Melnick A, (2004) Complex and dynamic redistribution of
NF-kappaB signaling intermediates in response 19. Kaye J, Porcelli S, Tite J, Jones B, Janeway CA Jr
to T cell receptor stimulation. Proc Natl Acad (1983) Both a monoclonal antibody and antisera
Sci U S A 101(4):1004–1009 specific for determinants unique to individual
17. Paul S, Kashyap AK, Jia W, He YW, Schaefer cloned helper T cell lines can substitute for anti-
BC (2012) Selective autophagy of the adaptor gen and antigen-presenting cells in the activation
protein Bcl10 modulates T cell receptor activa- of T cells. J Exp Med 158(3):836–856
tion of NF-kappaB. Immunity 36(6):947–958 20. Paul S, Schaefer BC (2015) Visualizing TCR-
18. Schaefer BC, Mitchell TC, Kappler JW, induced POLKADOTS formation and
Marrack P (2001) A novel family of retroviral NF-kappaB activation in the D10 T-cell clone
vectors for the rapid production of complex and mouse primary effector T cells. Methods
stable cell lines. Anal Biochem 297(1):86–93 Mol Biol 1280:219–238
Chapter 9
Imaging Vesicular Traffic at the Immune Synapse

Jérôme Bouchet, Iratxe del Río-Iñiguez, and Andrés Alcover
Abstract
Immunological synapse formation is the result of a profound T cell polarization process that involves
the coordinated action of the actin and microtubule cytoskeleton, as well as intracellular vesicle traffic.
Endosomal vesicle traffic ensures the targeting of the T cell receptor (TCR) and various signaling
molecules to the synapse, being necessary for the generation of signaling complexes downstream of
the TCR. Here we describe the microscopy imaging methods that we currently use to unveil how
TCR and signaling molecules are associated with endosomal compartments and deliver their cargo to
the immunological synapse.
Key words Immunological synapse, Vesicle traffic, Endosomes, Cytoskeleton, T cell activation, TCR
signaling
1 Introduction
Immunological synapses are characterized by the accumulation

and clustering of TCRs, co-signaling receptors like CD4 and
CD28, inhibitory receptors like CTLA4 and PD1, adhesion mol-
ecules like LFA-1, and a number of signaling molecules of the
TCR signaling cascade, including the tyrosine kinases Lck and
ZAP70, the adapter molecules LAT and SLP76, and the serine-
threonine kinases PKCθ and Erk1/2. Dynamic clustering of these
various molecules ensures TCR signal regulation, long-term T cell
activation, and effector functions, like cytokine production or the
expression of surface ligands by T helper cells or the release of
cytotoxic granule content by cytotoxic T cells [1, 2]. In order to
accumulate and cluster at the immunological synapse, those mol-
ecules need to be transported to and/or retained at the T cell-
antigen-presenting cell contact site. Two main mechanisms have
*
These authors contributed equally to this work.
129
130 Jérôme Bouchet et al.
been proposed to transport proteins to the synapse: cytoskeleton-

based movement at the plasma membrane [3] and transport via
intracellular vesicles [4]. While the former mechanism was shown
to transport the LFA-1 integrin, the latter was shown to transport
the TCR-CD3 complex subunits [5–7], the tyrosine kinase Lck [8,
9], the adapter LAT [10, 11], and the inhibitory receptor CTLA4
[12]. Different endosomal compartments and regulatory transport
proteins are involved. These different endosomal compartments
appear intermingled in the T cell cytoplasm, but they are distinct
with regard to endosomal markers and are differentially regulated
when targeting TCR and signaling molecules to the synapse [4,
13]. Thus, the TCR-CD3 complex was shown to be transported
via transferrin receptor-positive recycling endosomes [5], while the
TCRζ subunit is associated with Rab8 and intraflagellar transport
(IFT) protein-regulated compartments [7, 13, 14]. Moreover, Lck
traffic is regulated by MAL [15, 16], Unc119 [17], and Rab11 and
its effector FIP3 [13] (Bouchet et al., submitted). Finally, LAT was
found associated with Rab7, Rab27, and Rab37 endosomes and its
targeting to the synapse regulated by the SNARE VAMP7, the
calcium sensor synaptotagmin-7, and the IFT20 [11, 13, 18, 19].
Intracellular vesicle traffic and actin and microtubule cytoskeleton
rearrangements are tightly interconnected processes that ensure
immunological synapse architecture, molecular clustering, signal-
ing complex dynamics, and ultimately the regulation of T cell acti-
vation [4, 20, 21]. Interestingly, the subcellular localization and
activity of the GTPase Rac1, a master regulator of the actin cyto-
skeleton crucial for immunological synapse formation and func-
tion, are regulated by Rab11 and its effector FIP3 [21].
Here we describe the methods we have been developing to
localize TCR and signaling molecules together with endosomal
regulatory proteins at the immunological synapse and to decipher
the functional importance of this endosomal transport.
2 Materials
2.1 Cells 1. Jurkat T cell leukemia cells, J77 clone 20 cells (J77cl20), and
Raji B cell lymphoma cells have been previously described [5,
9, 22]. Cells are cultured in RPMI 1640 + GlutaMAX™ + phe-
nol red medium (Gibco®, Life Technologies™, No. 61870-
010) supplemented with 10% fetal calf serum and 10 mM
HEPES. Jurkat and Raji cells are cultured at a density average
of 0.5–1 × 106 cells/mL, splitting the cultures every 2–3 days.
2. Peripheral blood T cells from healthy donors are isolated by
centrifugation through Ficoll-Hypaque using Unisep Maxi
tubes (Eurobio, No. U-10) (see Note 1). CD4+ T cells are
further purified using the CD4+ T cell isolation kit (Miltenyi
Biotec, 130-096-533) (see Note 2). After isolation they are
Vesicle Traffic at the Immune Synapse 131
cultured at 2 × 106 cells/mL in RPMI 1614 medium supple-

mented with 10% FCS, 1 mM sodium pyruvate, and 1% MEM
nonessential amino acids.
2.2 Microscopy 1. Confocal microscope: LSM 700 confocal microscope (Zeiss)

Materials equipped with a Plan-Apochromat 63× objective and ZEN
software (Zeiss).
2. Square glass coverslips 20 × 20 mm. Coverslips are coated with
500 μL poly-l-lysine (0.002% w/v in water) during 20 min at
room temperature, washed with water, and air-dried before use
(see Note 3).
3. Round glass coverslips 12 mm diameter. Coating as in 2, using
150 μL per round coverslip.
4. Glass slides 76 × 26 mm.
5. ProLong® Gold Antifade mounting medium with DAPI
(Molecular Probes®, Life Technologies™, No. P36935).
2.3 Chemicals 1. Poly-l-lysine MW: 150–300 kDa, 0.1% (w/v) (Sigma-Aldrich®,

and Biological No. P8920). Coating solution 0.002% in water.
Products 2. Paraformaldehyde: (Electron Microscopy Sciences, No 15714,
aqueous stock solution 32%). Paraformaldehyde solution 8%
(w/v) in water (see Note 4).
3. Triton X-100 0.1% (v/v) in phosphate buffer pH 7.5, 150
mM NaCl (PBS).
4. Methanol.
5. Bovine serum albumin, 1% (w/v) in PBS (PBS-BSA). Store
solution at 4 °C.
6. Staphylococcus enterotoxin E superantigen (SEE, Toxin
Technology Inc), 10 μg/mL in PBS. Biohazard (see Note 5).
7. FITC-coupled phalloidin probe (Molecular Probes Inc) is used
at1/100 dilution.
8. Primary antibodies for immunofluorescence: mouse monoclo-
nal IgG2a anti-Rab11, clone 47, and IgG2b anti-Rac1, clone
102 (Becton Dickinson), are used at 25 μg/mL and 1 μg/mL,
respectively. Mouse monoclonal IgG2b anti-Lck, clone 3A5
(Santa Cruz Biotechnology), is used at 2 μg/mL. Mouse
monoclonal IgG1 anti-CD3ε, clone UCHT1 (BioLegend
Inc), is used at 10 μg/mL. Rabbit anti centrin-3, gift of
M. Bornens (Institut Curie, France) is used at 1/400 dilution.
Mouse IgG2b anti-β-tubulin, clone KMX1 (Millipore) is used
at 10 μg/mL. Rabbit anti-phospho-ZAP70 (Y319) (cell sig-
naling) is used at 1/100 dilution, and mouse monoclonal
IgG2a anti-phospho-TCRζ (Y142), clone K25-407.69
(Becton Dickinson), is used at 5 μg/mL.
9. Secondary antibodies for immunofluorescence: highly cross-

adsorbed Cy3-coupled goat anti-mouse IgG2a, anti-mouse
IgG2b, and anti-rabbit IgG (Jackson Immuno-Research
Laboratories) are used at 1/100 dilution. FITC-coupled goat
anti-mouse IgG1 (Southern Biotech) is used at 0.7 μg/
mL. Alexa Fluor 488-coupled goat anti-fluorescein (Molecular
Probes Inc) is used at 1 μg/mL.
10. Antibody used for pseudo-synapses: mouse IgG1 anti-CD3ε,
clone UCHT1 (Bio Legend) is used at 500 ng/mL.
11. Expression vectors: pEGFP-C1/FIP3 and FIP3-I738E mutant
expression vectors [23]. pEGFPN1/Rab11 WT, Rab11-Q70L,
and Rab11-S25N mutants provided by Dr. A. Echard (Institut
Pasteur, Paris). pEF-BOS/Rac1-cMyc WT, Rac1-G12V, and
Rac1-T17N mutants [24].
12. siRNA oligonucleotides: We provide here an example of our
recent work [21]. siRNA duplexes based on the human
RAB11-FIP3 sequence had been described previously by
others: (5′-AAGGGATCACAGCCATCAGAA-3′) [25] and
(5′-AAGGCAGTGAGGCGGAGCTGTT-3′) [26]. Stock
solutions are stored in aliquots at −20 °C and used 1 nmol per
transfection [21] (see Note 6).
2.4 Transfection 1. Neon™ Transfection System and Neon™ Transfection 100 μL

Systems Kit, containing electrolytic buffer E2, resuspension buffer R,
100 μL Neon™ tips, and Neon™ electroporation tubes
(Invitrogen™, Life Technologies™).
2. Nucleofector™ 2b Device and Primary Cell Nucleofector™
Kit, containing Nucleofector™ Solution, single-use pipette,
and Amaxa™ 100 μL aluminum electrode cuvettes (Lonza).
3 Methods
3.1 T Cell For expression vector’s transfection, cell synapse assays are done 24
Transfection h after T cell transfection. For siRNA treatment of T cells, two
by Electroporation sequential transfections at 24 h interval are performed, and cell
synapses are performed 72 h after the first siRNA transfection. In
case of DNA plasmid transfection on siRNA-transfected cells, two
sequential transfections at 24 h interval are performed for the
siRNA, followed by DNA plasmid transfection 48 h after the first
siRNA transfection. Synapses in this case are performed 24 h after
transfection of expression vectors (see Note 7).
3.1.1 Transfection of T Transfection is done with 5–10 × 106 cells. Use 10 μg of DNA
Cell Lines Using plasmid or 1 nM of siRNA for 10 × 106 cells. Complete with resus-
the Neon™ pension buffer R to equilibrate different volumes for the same
Transfection System amount of DNA or siRNA when preparing Eppendorf tubes with
plasmid DNA or siRNA:
1. Preincubate at 37 °C one flask with 10 mL T cell culture

medium (RPMI 1640 supplemented with 10% FCS and 10
mM HEPES) for each transfection. Prepare Eppendorf tubes
with DNA plasmid or siRNA.
2. Harvest the amount of cells required, centrifuge at 290 × g, at
20 °C for 4 min. Wash twice in PBS and aspirate supernatant.
3. Insert a new Neon™ tube into the Neon™ Pipette Station and
fill it with 3 mL electrolytic buffer E2. Turn on Neon™ device,
and select electroporation protocol (voltage, 1400 V; width,
10 ms; pulses, 3).
4. Resuspend the cell pellet in 100 μL resuspension buffer R per
required transfection.
5. Add 100 μL resuspended cells in each Eppendorf tube with
DNA plasmid or siRNA, and mix gently.
6. Take 100 μL of the cells-plasmid mix using the Neon™ pipette
and a 100 μL tip, avoiding air bubbles. Change Neon™ tip for
every transfection with different DNA or siRNA. Insert the
Neon™ pipette with the sample vertically into the Neon™ tube
and press start.
7. Once electroporation is complete, remove Neon™ pipette and
transfer the sample into the pre-warmed culture flask and incu-
bate at 37 °C, 5% CO2 for 24–48 h (see Note 8).
3.1.2 Transfection Transfection is done with 5–10 × 106 cells. Use 5 μg of DNA plas-
of Purified Primary CD4+ T mid or 30–300 nM of siRNA for 10 × 106 cells. Complete with
Cells Using Nucleofector™ resuspension buffer R to equilibrate different volumes for the same
2b Device amount of DNA or siRNA when preparing Eppendorf tubes with
plasmid DNA or siRNA:
1. Preincubate at 37 °C one flask with 1 mL primary CD4+ T cell
culture medium (RPMI 1614 medium supplemented with
10% FCS, 1 mM sodium pyruvate, and 1% MEM nonessential
amino acids) per 10 × 106 cells transfected. Prepare 1.5 mL
Eppendorf tubes with DNA plasmid or siRNA.
2. Harvest the amount of primary CD4+ T cells required, and
centrifuge at 453 × g at 20 °C for 10 min. Wash twice in PBS.
3. Turn on Nucleofector™ 2b Device and select electroporation
protocol (U014).
4. Add corresponding volume of DNA plasmid or siRNA to the
Amaxa cuvette per transfection required.
5. Resuspend the cell pellet in 100 μL Amaxa buffer per 10 × 106
CD4+ cells.
6. Take 100 μL of resuspended cells and add them to the cuvette
with DNA plasmid or siRNA. Mix gently, avoiding air
bubbles.
7. Insert cuvette in the Nucleofector™ 2b Device and electropor-

ate the cells using protocol U014.
8. Use single-use pipettes to recover cells and transfer to the
pre-warmed flask with medium. Incubate 10 min at 37 °C.
9. Count cells and resuspend to a final concentration of 2 × 106
cells/mL in RPMI 1614 medium supplemented with 10%
FCS, 1 mM sodium pyruvate, and 1% MEM nonessential
amino acids at 37 °C, 5% CO2 for 24–48 h (see Note 8).
3.2 Immunological 1. Pulse antigen-presenting cells with superantigen: Harvest Raji

Synapse Formation cells (5 × 106 cells/mL) by centrifuging at 290 × g at 20 °C for
Between T Cells 4 min. Resuspend in RPMI 1640 (without serum) supple-
and Antigen- mented with 10 μg/mL Staphylococcus enterotoxin E (SEE)
Presenting Cells superantigen and incubate for 30 min at 37 °C.
2. Immunological synapse formation: Harvest Jurkat J77 clone
20 cells by centrifuging at 290 × g at 20 °C for 4 min and
resuspend at 5 × 106 cells/mL RPMI 1640 (without serum).
Jurkat J77 clone 20 cells are incubated with pulsed Raji cells at
1:1 ratio, at 37 °C during 5, 15, or 30 min in RPMI 1640
medium (without serum) in 1.5 mL Eppendorf tubes at 37 °C
in a water bath.
3. Plating cells on coverslips: 160 μL of conjugated cells are
plated onto poly-l-lysine-coated square coverslips and kept for
4. Fixation: 160 μL of 8% paraformaldehyde are mildly dropped
onto the coverslips containing cell suspensions (final concen-
tration of paraformaldehyde 4%) and incubated for 20 min at
room temperature. After incubation, paraformaldehyde is
removed and coverslips washed with PBS, 1% (w/v) bovine
serum albumin (PBS-BSA) (see Notes 4 and 10).
5. Saturation: Nonspecific binding is prevented by 15 min incu-
bation in PBS-BSA.
6. First antibody preparation. Primary antibody (or mix of primary
antibodies in case of multiple staining), at the recommended
dilution, is suspended in PBS-BSA, 0.1% Triton X-100.
7. Cell immunostaining: 80 μL drops of primary antibody solu-
tion are deposited on Parafilm®, and coverslips with fixed cells
are turned over on drops (see Note 11). Incubate for 1 h at
room temperature, protected from light.
8. Washes: After primary antibody incubation, coverslips are
washed twice in PBS-BSA by submerging them several times
in a beaker using forceps to handle the coverslips.
9. Second antibody preparation: Secondary antibody (or mix of
secondary antibodies in case of multiple staining) is resus-
pended in PBS-BSA.
10. Staining of coverslips: 80 μL drops of secondary antibody

solution are deposited on Parafilm® and coverslips are turned
over on drops, as in step 7. Incubate for 45 min–1 h at room
temperature, protected from light.
11. Washes: After secondary antibody incubation, coverslips are
washed twice in PBS-BSA by submerging in a beaker using
forceps to handle the coverslips. Drain the washing solution
by setting the edge of the coverslip on absorbent paper, in
order to remove excess PBS-BSA before mounting.
12. Mounting: Drops of 20 μL of ProLong Gold Antifade mount-
ing medium with DAPI are deposited on 76 × 26 mm slides
and coverslips are turned over on the drops. Let dry mounted
slides overnight at room temperature, protected from light.
Once dried, slides can be stored at 4 °C and used during 2–4
weeks (see Note 12).
13. Microscopy analysis: Cells are observed under a LSM 700 con-
focal microscope (Zeiss) equipped with an oil-immersion Plan-
Apochromat 63× objective. Z-stack optical sections are
acquired at different increments depending on image analysis
performed afterward: 0.2 μm depth increments for deconvolu-
tion and colocalization analysis and 1 μm depth increments for
fluorescence intensity analysis. Green and red laser excitations
are intercalated to minimize fluorescence spill over different
channels. Image acquisition is done with ZEN software (Zeiss).
3.3 Pseudo-synapse 1. Coating coverslips with anti-CD3 antibodies. Prepare 10 μg/

Formation on Anti- mL anti-CD3 (UCHT-1) in PBS. Round poly-l-lysine-coated
CD3-Coated Coverslips coverslips are turned over 50 μL drops of anti-CD3 solution
and incubated 2 h at 37 °C in a humidified chamber or over-
night at 4 °C (see Note 13). Coverslips are washed once in PBS
and saturated with 200 μL RPMI 1640 with serum to prevent
nonspecific binding to poly-l-lysine.
2. Cell resuspension. Jurkat J77 clone 20 cells or CD4+ primary
T cells are washed twice by centrifugation at 290 × g at 20 °C
for 4 min and resuspended at 2 × 106 cells/mL in RPMI 1640
medium without serum.
3. Plating of cells on coverslips. 100 μL of cells are plated onto
anti-CD3-coated coverslips during 3, 5, or 15 min 37 °C in a
humidified chamber kept at 37 °C during the assay.
4. Fixation. 100 μL of 8% paraformaldehyde are mildly dropped
room temperature. After incubation PFA is removed and cov-
erslips washed with PBS-BSA (see Notes 4 and 10).
5. Saturation. Nonspecific binding is prevented by 15 min incu-
bation in PBS-BSA.
6. First antibody preparation. Primary antibody (or mix of primary

antibodies in case of multiple staining) at the recommended
dilution is suspended in PBS-BSA, 0.1% Triton X-100.
7. Staining of coverslips. 30 μL drops of primary antibody solu-
tion are deposited on Parafilm® and coverslips with fixed cells
8. Washes. After primary antibody incubation, coverslips are
washed twice in PBS-BSA by submerging several times in a
beaker using forceps to handle the coverslips.
9. Second antibody preparation. Secondary antibody (or mix of
pended in PBS-BSA.
10. Staining of coverslips. 30 μL drops of secondary antibody
solution are deposited on Parafilm® and coverslips are turned
over on drops, as done in step 7. Incubate for 45 min–1 h, at
11. Washes. After secondary antibody incubation, coverslips are
forceps to handle the coverslips. Drain the washing solution
by setting the edge of the coverslip on absorbent paper, in
order to remove excess PBS-BSA before mounting.
12. Mounting. Drops of 20 μL of ProLong Gold Antifade mount-
ing medium with DAPI are deposited on 76 × 26 mm slides
and coverslips are turned on the drops. Let mounted cover-
slips harden overnight at room temperature, protected from
light. Slides are then stored at 4 °C and may be used during
2–4 weeks (see Note 12).
13. Microscopy analysis. Cells are observed under a LSM 700 con-
acquired at different increments depending on desired image
analysis: 0.2 μm depth increments for deconvolution and colo-
calization analysis and 1 μm depth increments for fluorescence
intensity analysis. Green and red laser excitation are interca-
lated to minimize fluorescence spill over different channels.
Images acquisition is done with ZEN software (Zeiss).
3.4 Staining 1. Immunological synapses or pseudo-synapses are performed as

of Microtubules in Subheadings 3.2 and 3.3 until the fixation step.
2. For microtubule staining, 8% paraformaldehyde is mildly
dropped onto the coverslips with cells and incubated for only
10 min at room temperature. Directly after incubation, cover-
slips are placed in 6- or 24-well plates with ice-cold methanol
and stored at −20 °C for at least 1 h for further fixation and

permeabilization. Before staining coverslips are washed once
with PBS-BSA.
3. The remaining steps are performed as in Subheading 3.2 for
immunological synapses with antigen-presenting cells or
Subheading 3.3 for pseudo-synapses (Fig. 1).
3.5 Quantitative 1. Deconvolution of confocal images is used to improve image

Image Analysis rendering, especially for 3D reconstruction and prior to colo-
calization analysis. Deconvolution is performed on Z-stacks of
confocal optical sections obtained at 0.2 μm depth increments,
using Huygens Professional software (Scientific Volume
Imaging).
2. Colocalization analysis is performed with Fiji software (open
platform for scientific image analysis). JaCoP plugin and colo-
calization threshold option of Fiji are used.
Pearson’s correlation coefficient corresponds to the linear
relationship between intensities of pixels in the two analyzed
channels. Mander’s coefficient is defined as the ratio of the
summed intensities of pixels from one image for which the
intensity in the second channel is above the threshold [27].
Costes Automatic Threshold allows us to consider only the
pixels in each channel showing statistical correlation, as
explained in Costes et al. [28] (Fig. 2).
3. Fluorescence intensity analysis. Using Fiji, in a chosen channel,
create a Z-projection of the image (whole cell or a number of
optical sections corresponding to an intracellular compartment)
and select a region of interest (i.e., immunological synapse or
an intracellular compartment), depending on experimental
requirements. Then, in the other channel, measure fluorescence
intensity of the selected area of interest (Fig. 3).
Fig. 1 Microtubule detection at T cell “pseudo-synapses.” (a) Schematic representation of T cells spreading on
anti-CD3-coated coverslips. (b) Jurkat cells spread 3 min on an anti-CD3-coated coverslip. β-tubulin staining.
Confocal image treated by deconvolution. A 4 μm Z-projection close to the coverslip is shown. Scale bar = 5 μm
Fig. 2 Colocalization analyses of Rac1, Rab11, and Rab11-FIP3. (a, b) Immunostaining of Jurkat T cells for
endogenous Rab11 and Rac1 GTPases. A Z-stack of confocal images was treated by deconvolution. A 4 μm
Z-projection of the area where the intracellular pericentrosomal compartment is located is shown. Pearson’s
correlation coefficient (R) in the region framed was measured and the graph is shown on the right panel.
Examples of different experimental conditions. (a) Control cells. Colocalization between Rab11 and Rac1. (b)
Cells silenced for the expression of the Rab11 effector FIP3. Colocalization between Rab11 and Rac1. (c) Effect
of GFP-FIP3 (green) overexpression in endogenous Rac1 localization (red). Colocalization between FIP3 and
Rac1. Scale bar = 5 μm
Fig. 3 Quantification of fluorescence accumulation at the immunological synapse. Immunological synapse of a

Jurkat (J77cl20) lymphocyte with a Raji B lymphocyte pulsed with SEE superantigen (antigen-presenting cell,
APC). Confocal image was treated by deconvolution. A 4 μm Z-projection is shown. From left to right: Lck
(green), CD3 surface staining (red), merge of fluorescence images, and phase contrast image. Rectangles show
the region of interest in which quantification of the intensity of fluorescence is performed. It can be then referred
to the total fluorescence of the cell or to a region at the opposite side of the synapse. Scale bar = 5 μm
4 Notes
1. Density-gradient centrifugation to isolate peripheral blood

mononuclear cells (PBMCs). Ficoll separation: spin Ficoll
tubes 5 min at 453 × g at room temperature. Lay 28 mL of
blood per tube of 15 mL Ficoll-Hypaque-Lymphoprep.
Centrifuge 30 min at 805 × g at room temperature with no
brake. Aspirate plasma layer containing PBMCs and add it to
new empty falcon tubes (15 mL recovered PBMCs per 50 mL
falcon tube). Add 35 mL RPMI no FCS per 15 mL of recov-
ered PBMCs per tube. Spin 10 min at 453 × g at room tem-
perature. Discard supernatants and pool pellets together to
one tube. Wash twice with 50 mL RPMI no FCS (10 min at
453 × g). Resuspend pellet in 50 mL of RPMI no FCS and
count cells using cell counter or trypan blue vital staining and
Malassez chambers (Fig. 4).
2. Purification of CD4+ T cells using MACS CD4 cell isolation
kit II: follow manufacturer’s instructions.
3. Coverslip coating with poly-l-lysine: prepare a 1:50 dilution
of 0.01% poly-l-lysine solution in water. Cover a flat surface
with Parafilm and place round or square coverslips over it. Add
500 μL for square coverslips or 150 μL for round coverslips of
the 1:50 poly-l-lysine dilution over the coverslips, covering
the surface completely. Incubate 20 min at room temperature
and remove excess. Wash once with water and dry the cover-
slips before use.
4. Paraformaldehyde. We recommend purchasing a stock
commercial solution, avoiding manipulation of paraformal-
dehyde powder, in order to prevent toxicity.
Paraformaldehyde solution manipulation should be carried
out under a chemical hood.
Top layer: plasma
PBMCs
Ficoll layer: clear liquid
Red blood cells, PMNs
Fig. 4 Scheme of a Ficoll-gradient centrifugation tube after centrifugation

Day 1 Day 2 Day 3 Day 4
1st siRNA 2nd siRNA DNA plasmid Synapses or pseudo-

transfection transfection transfection synapses assay
Fig. 5 Schematic representation of the transfection time line
5. Staphylococcus enterotoxins are a biohazard and should be

manipulated as such. Inactivation of all solutions and materials
in contact with the toxin should be performed at the end of
experiments using an excess of 0.5% (w/v) sodium hypochlo-
rite for at least 1 h. Dispose as biohazard.
6. siRNA efficacy to inhibit protein expression depends on the
messenger RNA stability and the turnover of each protein. A
dose-response assay is useful to set the appropriate experimen-
tal conditions for different proteins.
7. Schematic representation of the transfection time line (Fig. 5).
8. Different plasmids might need different expression times in
order to reach enough protein levels. Expression may be dif-
ferent in Jurkat and in primary T cells.
9. Cover with Parafilm a flat surface and place the poly-l-lysine-
coated coverslips over it (poly-l-lysine side of the coverslips
facing up). Add cell suspension over the coverslip, trying to
distribute it evenly over the surface.
10. Once fixed and washed, round or square coverslips containing
cells may be stored in 6- or 24-well plates, respectively, in an
excess of PBS-BSA, for 1–2 weeks at 4 °C. Staining does not
require to be performed immediately after fixation, except for
protein phosphorylation staining, which requires freshly pre-
pared cells for optimum results.
11. Cover with Parafilm a flat surface and label areas for different
staining solutions planned. Deposit 80 μL drops of primary
antibody mix on the Parafilm. Take the coverslips with fixed
cells and drain the excess of PBS-BSA by setting the edge of
the coverslip on absorbent paper. This prevents dilution of
antibody staining solutions. Place coverslips over primary anti-
body drops on Parafilm, cells facing the drops. Create a
humidified chamber for the incubation by placing wet pieces
of paper around the Parafilm, and cover everything with a lid
or opaque surface, in order to protect from light.
12. Once slides have been analyzed, it is recommended to store

them at −20 °C, to prevent loss of fluorescence. While flouro-
chrome fluorescence tends to be stable, GFP fluorescence is
lost much faster.
13. Cover with Parafilm a flat surface and deposit 50 μL drops of
anti-CD3 antibody mix on the Parafilm. Create a humidified
chamber for the incubation by placing wet pieces of paper
around the Parafilm, and cover everything with a lid.
Acknowledgments
This work was supported by grants from the Agence Nationale de

Recherche (ANR, No 11 BSV3 025 01), the Agence National de
Recherche sur le SIDA et les Hepatitis Virales (ANRS), Institut
Pasteur, CNRS, and INSERM. JB has been supported by ANRS and
Roux-Institut Pasteur postdoctoral fellowships. IdRI is a scholar in
the Pasteur-Paris University (PPU) International PhD program and
is funded by the People Programme (Marie Curie Actions) of the
European Union’s Seventh Framework Programme FP7/2007–
2013/under the REA grant agreement n° 317057 HOMIN.
References
1. Fooksman DR, Vardhana S, Vasiliver-Shamis cell polarization and recruitment of signaling

G, Liese J, Blair DA, Waite J, Sacristan C, proteins but not formation of the synaptic
Victora GD, Zanin-Zhorov A, Dustin ML pattern. Immunity 17:389–399
(2010) Functional anatomy of T cell activation 7. Finetti F, Paccani SR, Riparbelli MG,
and synapse formation. Annu Rev Immunol Giacomello E, Perinetti G, Pazour GJ,
28:79–105 Rosenbaum JL, Baldari CT (2009) Intraflagellar
2. Agüera-Gonzalez S, Bouchet J, Alcover A transport is required for polarized recycling of
(2015) Immunological synapse. eLS. Wiley, the TCR/CD3 complex to the immune syn-
Chichester. doi: 10.1002/9780470015902. apse. Nat Cell Biol 11:1332–1339
a0004027.pub2 8. Ehrlich LIR, Ebert PJR, Krummel MF, Weiss
3. Wülfing C, Davis MM (1998) A receptor/ A, Davis MM (2002) Dynamics pf p56lck
cytoskeletal movement triggered by costimula- translocation to the T cell immunological syn-
tion during T cell activation. Science apse following agonist and antagonist stimula-
282:2266–2269 tion. Immunity 17:809–822
4. Soares H, Lasserre R, Alcover A (2013) 9. Thoulouze MI, Sol-Foulon N, Blanchet F,
Orchestrating cytoskeleton and intracellular Dautry-Varsat A, Schwartz O, Alcover A
vesicle traffic to build functional immunologi- (2006) Human immunodeficiency virus type-1
cal synapses. Immunol Rev 256:118–132 infection impairs the formation of the immu-
5. Das V, Nal B, Dujeancourt A, Thoulouze MI, nological synapse. Immunity 24:547–561
Galli T, Roux P, Dautry-Varsat A, Alcover A
10. Bonello G, Blanchard N, Montoya MC,
(2004) Activation-induced polarized recycling Aguado E, Langlet C, He HT, Nunez-Cruz S,
targets T cell antigen receptors to the immuno- Malissen M, Sanchez-Madrid F, Olive D,
logical synapse; involvement of SNARE com- Hivroz C, Collette Y (2004) Dynamic recruit-
plexes. Immunity 20:577–588 ment of the adaptor protein LAT: LAT exists in
6. Blanchard N, Di Bartolo V, Hivroz C (2002) two distinct intracellular pools and controls its
In the immune synapse, ZAP-70 controls T own recruitment. J Cell Sci 117:1009–1016
11. Larghi P, Williamson DJ, Carpier JM, 20. Lasserre R, Charrin S, Cuche C, Danckaert A,
Dogniaux S, Chemin K, Bohineust A, Danglot Thoulouze MI, de Chaumont F, Duong T,
L, Gaus K, Galli T, Hivroz C (2013) VAMP7 Perrault N, Varin-Blank N, Olivo-Marin JC,
controls T cell activation by regulating the Etienne-Manneville S, Arpin M, Di Bartolo V,
recruitment and phosphorylation of vesicular Alcover A (2010) Ezrin tunes T-cell activation
Lat at TCR-activation sites. Nat Immunol by controlling Dlg1 and microtubule position-
14:723–731 ing at the immunological synapse. EMBO
12. Linsley PS, Bradshaw J, Greene J, Peach R, J 29:2301–2314
Bennett KL, Mittler RS (1996) Intracellular 21. Bouchet J, Del Rio-Iñiguez I, Lasserre R,
trafficking of CTLA-4 and focal localization Agüera-Gonzalez S, Cuche C, Danckaert A,
towards sites of TCR engagement. Immunity McCaffrey M, Di Bartolo V, Alcover A (2016)
4:535–543 Rac1-Rab11-FIP3 regulatory hub coordinates
13. Soares H, Henriques R, Sachse M, Ventimiglia vesicle traffic with actin remodeling and T-cell
L, Alonso MA, Zimmer C, Thoulouze MI, activation. EMBO J 35(11):1160–1174
Alcover A (2013) Regulated vesicle fusion gen- 22. Roumier A, Olivo-Marin JC, Arpin M, Michel
erates signaling nanoterritories that control T F, Martin M, Mangeat P, Acuto O, Dautry-
cell activation at the immunological synapse. Varsat A, Alcover A (2001) The membrane-
J Exp Med 210:2415–2433 microfilament linker ezrin is involved in the
14. Finetti F, Patrussi L, Galgano D, Cassioli C, formation of the immunological synapse and in
Perinetti G, Pazour GJ, Baldari CT (2015) The T cell activation. Immunity 15:715–728
small GTPase Rab8 interacts with VAMP-3 to 23. Horgan CP, Hanscom SR, Kelly EE, McCaffrey
regulate the delivery of recycling T-cell recep- MW (2012) Tumor susceptibility gene 101
tors to the immune synapse. J Cell Sci (TSG101) is a novel binding-partner for the
128:2541–2552 class II Rab11-FIPs. PLoS One 7:e32030
15. Anton O, Batista A, Millan J, Andres-Delgado 24. Komuro, R, Sasaki T, Takaishi K, Orita S, Takai
L, Puertollano R, Correas I, Alonso MA Y (1996) Involvement of Rho and Rac small G
(2008) An essential role for the MAL protein proteins and Rho GDI in Ca2+-dependent
in targeting Lck to the plasma membrane of exocytosis from PC12 cells. Genes Cells 1:
human T lymphocytes. J Exp Med 943–951.
205:3201–3213 25. Jing J, Junutula JR, Wu C, Burden J, Matern H,
16. Anton OM, Andres-Delgado L, Reglero-Real Peden AA, Prekeris R (2010) FIP1/RCP bind-
N, Batista A, Alonso MA (2011) MAL protein ing to Golgin-97 regulates retrograde transport
controls protein sorting at the supramolecular from recycling endosomes to the trans-Golgi
activation cluster of human T lymphocytes. network. Mol Biol Cell 21:3041–3053
J Immunol 186:6345–6356 26. Wilson GM, Fielding AB, Simon GC, Yu X,
17. Gorska MM, Liang Q, Karim Z, Alam R Andrews PD, Hames RS, Frey AM, Peden AA,
(2009) Uncoordinated 119 protein controls Gould GW, Prekeris R (2005) The FIP3-Rab11
trafficking of Lck via the Rab11 endosome and protein complex regulates recycling endosome
is critical for immunological synapse formation. targeting to the cleavage furrow during late
J Immunol 183:1675–1684 cytokinesis. Mol Biol Cell 16:849–860
18. Purbhoo MA (2013) The function of sub- 27. Bolte S, Cordelieres FP (2006) A guided tour
synaptic vesicles during T-cell activation. into subcellular colocalization analysis in light
Immunol Rev 251:36–48 microscopy. J Microsc 224:213–232
19. Vivar OI, Masi G, Carpier JM, Magalhaes JG, 28. Costes SV, Daelemans D, Cho EH, Dobbin Z,
Galgano D, Pazour GJ, Amigorena S, Hivroz Pavlakis G, Lockett S (2004) Automatic and
C, Baldari CT (2016) IFT20 controls LAT quantitative measurement of protein-protein
recruitment to the immune synapse and T-cell colocalization in live cells. Biophys J 86:
activation in vivo. Proc Natl Acad Sci U S A 3993–4003
113:386–391
Chapter 10
Analysis of TCR/CD3 Recycling at the Immune Synapse

Laura Patrussi and Cosima T. Baldari
Abstract
Engagement of the T cell antigen receptor (TCR) by specific ligand bound to the major histocompatibility
complex is the primary event that leads to the assembly of the immune synapse (IS). Central to this process
is TCR clustering at the T cell-APC contact, which is achieved with the contribution of an endosomal pool
that is delivered to the IS by polarized recycling. As the TCR recycling process has not been fully eluci-
dated, we developed methods suitable to quantitate recycling to the plasma membrane of TCR/CD3
complexes that have been engaged at the cell surface and track their traffic through the intracellular vesicu-
lar compartments toward the IS.
Key words TCR, Recycling, Internalization, Immune synapse, Polarization, Acid stripping, Flow
cytometer, Confocal microscope
1 Introduction
Receptor recycling is the mechanism that ensures the internaliza-

tion of ligand-bound receptors and their return to the plasma
membrane without the need of new synthesis [1]. This mechanism
is particularly relevant in the context of the IS, where engaged
TCR/CD3 complexes concentrate and toward which an endo-
somal CD3+ compartment polarizes [2–5].
An ordered sequence of membrane traffic events, as yet not
fully understood, regulates TCR/CD3 recycling to the T cell
surface. In particular, a question that remains to be elucidated is
how, following their internalization, TCR molecules are routed
to recycling endosomes or are instead targeted for degradation
[1, 6, 7]. To address this issue, we developed new protocols that
allow us (i) to quantify the proportion of engaged TCR/CD3
that, following internalization, recycle back to the plasma mem-
brane and (ii) to visualize engaged TCR/CD3 both intracellularly
and at the cell surface. These procedures rely on one of the most
commonly used anti-CD3 antibodies, OKT3, which potently
143
144 Laura Patrussi and Cosima T. Baldari
stimulates TCR/CD3-dependent signaling pathways and mediates

its internalization and recycling [8].
The flow cytometry-based quantification of CD3 recycling is a
modification of the method developed by Margadant and col-
leagues [9]. The initial step of CD3 surface staining/stimulation is
followed by the internalization of a portion, which is often limited,
of the engaged surface CD3/antibody (Ab) complexes. The key
point of this protocol is the “acid stripping” step that, by removing
all the CD3/Ab complexes still on the surface, excludes from the
analysis the TCR/CD3 complexes that have not undergone inter-
nalization, making it possible to specifically track the engaged/
internalized complexes [10, 11] and follow their alternative rout-
ing either to the recycling endosomes to be returned to the plasma
membrane or to the lysosomes for degradation. We optimized the
protocol to specifically measure TCR/CD3 complexes that follow
the recycling route. After staining of the primary anti-CD3 Ab
with secondary fluorescently labelled antibodies, intact cells can be
easily analyzed by flow cytometry to detect solely the CD3/Ab
complexes that have returned to the cell surface (see Subheading
3.1). Alternatively, the inclusion of a permeabilization step permits
the visualization of the internalized CD3/Ab complexes during
their transit through the recycling compartment by confocal
microscopy, as initially described by our group in a recent report
[8]. After a first staining step with the anti-CD3 Ab, the cells are
allowed to internalize CD3/Ab complexes. However, as opposed
to the previous flow cytometry-based method, here we can visual-
ize the endosomes containing internalized CD3/Ab complexes by
confocal microscopy following staining with secondary fluores-
cently labelled antibodies (see Subheading 3.2).
The protocols developed for the analysis of TCR/CD3 recy-
cling have turned out to be valuable tools to study the process of
polarized recycling to the IS as they allow for the discrimination
between the surface TCR/CD3 complexes that cluster to the IS
and the ones that recycle to the IS from the intracellular pool in
conjugates formed between T cells and APCs loaded with staphy-
lococcal superantigens. These modified protocols also capitalize on
the “acid stripping” step prior to conjugate formation to clean the
T cell surface from CD3/Ab complexes that were not internalized.
Staining of non-permeabilized antigen-specific conjugates with
secondary fluorescently labelled antibodies allows to selectively
track the fate of engaged surface CD3 molecules which, after inter-
nalization, recycle to the contact area between T and B cells (see
Subheading 3.3). A similar analysis in permeabilized conjugates,
combined with co-staining with antibodies against markers of spe-
cific endosomal subpopulations, can provide information as to the
intracellular localization of the TCR/CD3 complexes undergoing
recycling, regarding both their polarization toward the centrosome
that localizes beneath the IS and the identity of the endosome
Analyzing TCR/CD3 Recycling 145
where they become trapped when the expression of specific traffic

regulators is disrupted [8, 12].
CD3 recycling is a highly complex mechanism which is not yet
fully understood. The development of these methods, combined
with RNA interference or gene targeting technologies, can be
helpful to dissect the molecular composition and spatiotemporal
regulation of this pathway and can be extended to the growing
variety of receptors that exploit the recycling pathways to localize
to the IS.
2 Materials
2.1 Cell Purification 1. Jurkat T cell line (acute T cell leukemia, clone E6-1, ATCC®
and Maintenance TIB-152™), Raji B cell line (Burkitt’s lymphoma, ATCC®
CCL-86™), and primary T cells from healthy donors.
2. Medium for cell maintenance: RPMI 1640 with 2 mM l-
glutamine added with penicillin (2 mg/mL) and 7.5% bovine
calf serum (BCS, HyClone).
3. Purification of primary T cells from peripheral blood of healthy
donors: incubation with T lymphocyte purification kit
(StemCell) and subsequent gradient centrifugation on
Lympholyte (Cedarlane).
2.2 Flow Cytometry- 1. Anti-CD3 OKT3 mAb purified from the supernatant of the
Based Quantification OKT3 mouse hybridoma (ATCC® CRL-8001™) (see Note 1)
of Membrane Receptor using the MabTrap purification kit (GE Healthcare).
Recycling 2. Non-sterile 96-well plates with round bottom (Sarstedt) (see
Note 2).
3. RPMI 1% BSA: add 5 g of non-sterile BSA suitable for cell
culture to 500 mL of RPMI, and filter using a 0.2-μm vacuum
filter.
4. Stripping solution: 100 mM glycine, 100 mM NaCl, pH 2.5
(see Note 3).
5. Secondary fluorescently labelled Ab: Alexa Fluor goat anti-
mouse 488 (Invitrogen) diluted 1:500 in phosphate-buffered
saline (PBS) (see Note 4).
6. Flow cytometer.
2.3 Evaluation 1. Non-sterile round bottom 96-well plate (Sarstedt).

of TCR/CD3 Recycling 2. Preparation and washing of 10-well diagnostic microscope
by Confocal slides (Thermo Scientific): absolute ethanol, poly-l-lysine
Microscopy diluted in 10 mM Tris–HCl pH 8.0, 3MM paper (Whatman),
and a Hellendahl-type dish containing PBS.
3. Fixation solution: 4% paraformaldehyde in PBS.
4. Permeabilization solution: PBS 0.1% BSA plus 0.01% Triton

X-100.
5. Secondary fluorochrome-conjugated Alexa Fluor goat anti-
mouse antibodies, Hanks’ balanced salts.
6. Slide mounting medium: PBS 90% glycerol. Coverslips 24 ×
60 mm (VWR) and conventional nail polish.
7. Staphylococcal enterotoxin E (SEE), staphylococcal entero-
toxin A (SEA), and staphylococcal enterotoxin B (SEB) (Toxin
Technology).
8. Stripping solution: 100 mM glycine, 100 mM NaCl, pH 2.5
(see Note 3).
3 Methods
3.1 Flow Cytometry- Use triplicate samples. This protocol can be applied to both Jurkat
Based Quantification T lymphocytes and primary T lymphocytes from healthy donors.
of Membrane Receptor
Recycling
3.1.1 Setup 1. Count 1 × 106 cells per sample, resuspend in 100 μL RPMI
of the Internalization Time 1640 supplemented with 1% BSA and transfer to 96-well
plates. Incubate for 30 min at 37 °C in a cell culture incubator
with 5% CO2.
2. Centrifuge the plate at 600 × g, discard the supernatant and
briefly vortex, transfer to ice and resuspend each sample in 20
μL RPMI 1% BSA containing saturating concentrations of
CD3-specific OKT3 antibodies (see Note 5). Incubate for
30 min on ice to allow binding.
3. Wash cells twice with ice-cold PBS, centrifuge the plate for 30
s at 600 × g and discard the supernatant by quickly inverting
the plate to remove any residual unbound Ab. Resuspend
samples in 200 μL cold RPMI 1% BSA, carefully pipette to
mix the solution and transfer 30 μL (sample 1) in a new
96-well plate which will be maintained on ice until the end of
the experiment.
4. Shift samples to 37 °C in the cell culture incubator (see Note 6).
Transfer 30 μL of solution in the cold plate every 15 min.
5. After transfer of the last sample wash ice-stored plate twice
with 200 μL of PBS and centrifuge the plate for 30 s at 600 ×
g to remove all the supernatant. Vortex briefly.
6. Resuspend samples by carefully pipetting 10 μL of PBS con-
taining Alexa Fluor goat anti-mouse 488 diluted 1:500 (see
Note 7). Incubate for 30 min on ice in the dark (see Note 8).
7. Wash samples twice with cold PBS and centrifuge the plate for
30 s at 600 × g to remove all the supernatant. Vortex briefly.
Resuspend each sample in 200 μL PBS and analyze by flow

cytometry.
8. Quantify the percentage of internalization by analyzing the
decrease of CD3 surface staining with respect to the staining of
sample at time 0, set as 100%.
3.1.2 Measuring 1. Count 1 × 106 cells per sample (see Notes 9 and 10), resus-
Recycling of Engaged pend in 100 μL of RPMI 1% BSA and transfer to 96-well
Surface TCR/CD3 plates. Incubate for 30 min at 37 °C in cell culture incubator.
Complexes 2. Centrifuge the plate for 30 s at 600 × g, discard the superna-
tant and briefly vortex, transfer on ice and resuspend each sam-
ple in 20 μL of RPMI 1% BSA containing saturating
concentrations of OKT3 antibodies (see Note 5). Incubate for
30 min on ice.
3. Wash cells twice with cold PBS, centrifuge the plate for 30 s at
600 × g and discard the supernatant by quickly inverting the
plate. Resuspend samples in 200 μL cold RPMI 1% BSA, care-
fully pipette to mix the solution and transfer 30 μL (sample 1)
to a new 96-well plate which will be maintained on ice until
the end of the experiment.
4. Shift the plate to 37 °C in the cell culture incubator for 40 min
(see Note 11), then mix well, and transfer 30 μL of the solu-
tion (sample 2) to the ice-cold plate. Centrifuge remaining
cells for 30 s at 600 × g, remove all the supernatant and vortex
briefly.
5. Drop 50 μL of stripping solution onto the samples and incu-
bate 30 s at room temperature (RT) (see Note 12), then
quickly add 100 μL RPMI 1% BSA and centrifuge the plate for
30 s at 600 × g, remove all the supernatant and perform a sec-
ond wash as above. Briefly vortex and resuspend in 200 μL of
RPMI 1% BSA. Mix well and transfer 60 μL of the solution
(sample 3) to the ice-cold plate.
6. Incubate samples at 37 °C to allow recycling of receptor-Ab
complexes to the plasma membrane. After 20 and 40 min (see
Note 13), mix well and transfer 60 μL of the solution (samples
4 and 5) to the ice-cold plate.
7. Wash ice-stored plate with 200 μL PBS, then centrifuge for 30
s at 600 × g, and remove the supernatant. Place the plate on ice
and resuspend each sample by pipetting in 10 μL PBS contain-
ing Alexa Fluor goat anti-mouse 488 Ab diluted 1:500 (see
Note 14). Incubate on ice in the dark for 30 min.
8. Wash samples twice with cold PBS and centrifuge the plate for
30 s at 600 × g, vortex briefly and resuspend each sample in
200 μL of PBS. Analyze samples by flow cytometry.
9. The data are presented as the percentage of the internalized recep-

tors that have recycled to the cell surface as described [9], calculated
(MFI t - MFI s )
using the formula x t = ´ 100
(MFI max - MFI s ) - (MFI n - MFI s )
where MFIt is the mean fluorescence intensity at time “t,”
MFIs is the MFI after acid stripping of surface-bound Ab,
MFImax is the MFI after incubation on ice with receptor-spe-
cific Ab, and MFIn is the MFI after receptor-Ab complexes are
internalized [8].
3.2 Visualization Set up duplicate samples:

of Receptors
1. Count 1.2 × 105 of T lymphocytes (see Note 10) per sample
Undergoing Recycling
and place in a 96-well plate. Centrifuge the plate for 30 s at
by Confocal 600 × g to remove all the cell culture medium, briefly vortex
Microscopy and equilibrate samples for 30 min at 37 °C in 100 μL RPMI
1% BSA, and then centrifuge again to discard the supernatant.
2. Resuspend samples in at least 50 μL (see Note 15) of RPMI 1%
BSA containing saturating concentrations of CD3-specific
antibodies and place the plate at 37 °C for 2 h (see Note 16).
3. Wash 10-well diagnostic microscope slides with absolute ethanol
and allow them to dry. Coat each required well with 30 μL of
50 μg/mL poly-l-lysine (see Note 17) and maintain at RT for
20 min in the dark, and then discard the solution and carefully
wash with Milli-Q. Air-dry the slides and maintain in the dark.
4. At the end of the 2-h incubation, immediately transfer all sam-
ples to the poly-l-lysine-coated slides, one sample/well. Keep
the slide safe from accidental hits. Allow samples to adhere for
15 min at RT.
5. Check if cells adhered to the slide by observing them at the
microscope.
6. Remove the supernatant using the pipette. Do not completely
dry the wells to avoid cell disruption. Carefully add fixation
buffer (see Subheading 2) and maintain for 20 min at RT in the
dark.
7. Wash slides by immersion into a Hellendahl-type dish contain-
ing PBS. Dry the spaces between wells using 3MM paper,
being careful to not completely remove PBS from the wells to
avoid cell disruption.
8. Permeabilize cells by dropping permeabilization solution (see
Subheading 2) onto the wells and incubate for 20 min at RT.
9. Wash slides (as described in step 7).
10. Add 20 μL/well of fluorochrome-conjugated Alexa Fluor goat
anti-mouse Ab diluted 1:400 in Hanks’ salts and keep the slide
at RT in the dark for 1 h.
11. Wash slides (as described in step 7).
12. Add slide mounting solution (see Subheading 2), one drop/well
(see Note 18), and then add the coverslip, remove the excess of
slide mounting solution by very carefully pushing onto the cov-
erslip, and then fix it using conventional nail polish.
13. Store the slides at 4 °C in the dark.
Internalized CD3-Ab complexes are visualized by confocal
microscopy (see Note 19). The number of vesicles that are positive
for each receptor can be determined on individual medial confocal
sections using ImageJ (the “analyze particles” function) to identify
and count objects, setting 0.005 mm2 as the lowest limit and
excluding the compact pericentrosomal compartment where
objects cannot be discriminated [8].
3.3 Analyzing TCR/ Set up all samples in duplicate:

CD3 Recycling
1. Count 0.8 × 105 T lymphocytes (see Note 20) per sample and
to the IS in T-B Cell
put in 1.5-mL Eppendorf tubes (see Note 21). Centrifuge at
Conjugates 16,000 × g for 30 s, remove all culture medium, briefly vortex
and incubate samples in 50 μL RPMI 1% BSA for 30 min at 37
°C, and then centrifuge again at 16,000 × g for 30 s and dis-
card the supernatant.
2. Resuspend samples in 50 μL (see Note 15) RPMI 1% BSA
containing saturating concentrations of OKT3 antibodies and
place the plate at 37 °C for 2 h.
3. In the meantime, count 0.8 × 105 Raji B lymphocytes per sam-
ple, centrifuge at 16,000 × g for 30 s, and resuspend in RPMI
without serum at the final concentration of 1 × 106/200 μL in
a 2-mL Eppendorf tube. Add staphylococcal enterotoxin E
(SEE) at the final concentration of 1 μg/mL (see Notes 22 and
23) and place the tube in the cell incubator for 30 min (see
Note 24). Carefully mix the solution every 10 min to avoid
sinking of the cells.
4. Wash T lymphocytes with 200 μL of PBS to remove excess of
primary Ab, centrifuge at 16,000 × g for 30 s, and briefly vortex.
5. Drop 50 μL of stripping solution (see Note 25) onto the sam-
ples and incubate for 30 s at RT, then wash by quickly adding
200 μL RPMI 1% BSA, and centrifuge samples at 16,000 × g
for 30 s. Remove all the supernatant and wash again as above.
Resuspend each sample in 15 μL of RPMI 1% BSA.
6. Centrifuge B lymphocytes at maximum speed for a few sec-
onds, remove all the supernatant, and carefully resuspend the
pellet by gently tapping the bottom of the tube. Add a total
volume of RPMI 1% BSA calculated multiplying 15 μL by the
number of samples.
7. Transfer 15 μL of the B cell solution into each 1.5-mL tube
containing resuspended T lymphocytes (as described in step 6)
and mix well by pipetting. Immediately transfer all samples to

37 °C and incubate for 15 min. Avoid shaking (see Note 26).
8. Immediately transfer all samples (a total volume of 30 μL) by
gently pipetting and slowly dropping them on the poly-l-
lysine-coated diagnostic microscope slides, one sample/well
(see Note 27). Keep the slide safe from accidental hits. Allow
samples to adhere for 15 min.
9. Check cell adhesion by observing the slide at the microscope.
10. Remove the supernatant using a pipette (see Note 28). Do not
completely dry the wells to avoid cell disruption. Carefully add
fixation buffer (see Subheading 2) and incubate for 20 min at
RT in the dark.
11. Wash wells by immersing the slides into a Hellendahl-type dish
containing PBS for 2 min, and carefully dry the space between
wells using 3MM paper, being careful to not completely dry
the wells.
12. Drop 30 μL permeabilization solution (see Subheading 2) onto
the wells and incubate for 20 min at RT (see Note 29). Wash
wells as described in step 11 (see Note 30).
13. Add 20 μL/well fluorochrome-conjugated Alexa Fluor goat
anti-mouse Ab diluted 1:400 in Hanks’ salts and maintain the
slide at RT in the dark.
14. Wash wells as described in step 11.
15. Add slide mounting solution (see Subheading 2), one drop/
well, and then add the coverslip, remove all the excess mount-
ing solution by very carefully pushing on the coverslip, and
then fix it using conventional nail polish.
16. Store the slides at 4 °C in the dark.
17. Analyze the slides by confocal microscopy using a 63× or
higher magnification objective (see example in Fig. 1).
Fig. 1 (continued) plasma membrane. (c) Cells are stripped to completely remove
all primary antibodies bound to the TCR/CD3 complexes that have remained at
the plasma membrane. Only internalized complexes are still bound to the pri-
mary antibodies. (d) T cells are incubated with antigen-loaded APCs for 15 min
at 37 °C to allow immune synapse formation. During this step, intracellular Fig. 1
(continued) vesicles containing TCR/CD3-Ab complexes polarize beneath the
synaptic membrane, and up to 50% of these are recycled to the plasma mem-
brane as assessed by flow cytometry (the pH of recycling endosomes is not
sufficiently low to allow for the dissociation of the antibodies from the TCR/CD3
complexes). (e) Conjugates are transferred to diagnostic microscope slides,
fixed, and then permeabilized. Fluorochrome-conjugated secondary antibodies
added to the samples stain TCR/CD3-Ab complexes both in the intracellular com-
partment and on the cell surface. (f) Conjugates treated as above were fixed and
stained with fluorochrome-conjugated secondary antibodies. Only TCR/CD3/
antibody complexes on the cell surface are stained
Fig. 1 Schematic representation of a method to specifically track TCR/CD3 com-

plexes that have recycled to the plasma membrane. (a) T lymphocytes are incu-
bated with saturating concentrations of anti-CD3 (OKT3) primary antibodies at
37 °C for 2 h. (b) During the incubation, some TCR/CD3-Ab complexes are inter-
nalized in an intracellular vesicular compartment, while the others remain at the
4 Notes
1. Store OKT3 hybridoma cells in liquid nitrogen. When

required, thaw and grow them in RPMI 7.5% BCS till they
reach the concentration of 1 × 106. Purify and titrate OKT3
using the MabTrap purification kit according to the manufac-
turer’s instructions.
2. 96-well plates with round bottom allow all staining and wash-
ing procedures to be performed using a multichannel pipette
and the centrifugation steps to be performed using a plate
centrifuge.
3. The stripping solution can be stored at 4 °C for several months
without losing its activity. Check pH once in a while.
4. The fluorescently labelled secondary Ab must be kept safe
from the light. For this reason, prepare dilutions and perform
incubations in the dark.
5. Determine the saturating concentration of OKT3 for every
new preparation. Stain T cells with serial dilutions of the Ab for
30 min on ice in 1.5-mL Eppendorf tubes. Then wash twice
with cold PBS, centrifuge for 30 s at 16,000 × g, and incubate
with 10 μL of secondary fluorochrome-labelled antibodies
diluted in PBS for 30 min on ice. Wash twice with cold PBS,
centrifuge for 30 s at 16,000 × g, and resuspend in 200 μL of
cold PBS. Analyze samples by flow cytometry to find the con-
centration of primary Ab that gives the plateau of surface bind-
ing. Due to the fact that the surface levels of TCR/CD3 are
different between cell lines and primary T cells, we usually test
new batch of OKT3 on each cell type.
6. This step allows internalization of receptor-Ab complexes.
7. The incubation with secondary fluorescently labelled Ab spe-
cifically detects receptor-Ab complexes which remained on the
cell surface.
8. Alexa Fluor-labelled secondary antibodies are diluted 1:500 in
Hanks’ salts and immediately used. Dilutions maintained for
prolonged periods at 4 °C lose their activity.
9. To determine the final number of cells per sample, take into
account the total number of samples for each experiment. A
standard time course of recycling requires at least five samples
of 1 × 105 cells each.
10. This experiment has been set using Jurkat T lymphocytes and
then adapted to purified primary T lymphocytes from healthy
donors. By simply adjusting internalization and recycling time
courses, it can be applied to almost all surface receptors [4, 8,
13, 14] and virtually all cell types.
11. We identified the optimal internalization time for TCR/CD3

at 40 min for both Jurkat and primary T lymphocytes, follow-
ing the protocol described in Subheading 3.1.1.
12. The low pH of the stripping solution removes residual surface-
bound antibodies.
13. After 20 and 40 min of incubation at 37 °C, the percentage of
TCR/CD3 recycling is approximately 40 and 50, respectively.
14. This step specifically detects receptor-Ab complexes that, after
acid stripping, recycled to the cell surface.
15. Use not less than 50 μL of solution to resuspend samples to
avoid dehydration due to prolonged incubation at 37 °C.
16. Put the plate in the cell incubator to maintain the cells in a
homeostatic environment.
17. A 50 mg/mL poly-l-lysine stock solution in 10 mM Tris–HCl
pH 8.0 can be maintained at −20 °C. When required, dilute it
in 10 mM Tris–HCl pH 8.0 to the final concentration and use
it fresh.
18. Dispense slide mounting medium dropwise onto the slide wells
using a Pasteur pipette.
19. The number of intracellular CD3+ vesicles is inversely propor-
tional to the extent of receptor recycling. This method can also
be used to identify the vesicular compartment where recycling
TCR/CD3 complexes accumulate in cells knocked down for
expression of potential regulators of recycling by RNA inter-
ference or CRISPR-mediated technology by staining various
members of the Rab family of small GTPases, specific markers
of different endosomal populations [8, 12, 15].
20. The experiment can be performed using both T cell lines and
primary T cells.
21. Each T cell sample must be prepared in a 1.5-mL Eppendorf
tube in order to be processed individually.
22. Staphylococcal enterotoxin E (SEE) used at the final concen-
tration of 1 μg/mL allows the maximal percentage of T-B cell
conjugate formation [5].
23. This protocol has been optimized for the analysis of conjugates
between Raji B cells and Jurkat T cells. In case of use of pri-
mary T cells, load B cells with a mixture of staphylococcal
enterotoxin A (SEA), staphylococcal enterotoxin B (SEB), and
SEE at the final concentration of 1 μL/mL each, to ensure the
maximum coverage of TCR Vβ specificities [5].
24. B cell loading with staphylococcal enterotoxins ranges from a
minimum of 30 min to a maximum of 3 h, with unchanged
efficiency of conjugate formation.
25. We introduced this step in the protocol in order to visualize

only TCR/CD3 complexes that are internalized from the sur-
face to the intracellular compartment and that polarize and/or
recycle to the IS during conjugate formation.
26. Cell resuspension during this step could prevent proper conju-
gate formation.
27. Sample transfer from 1.5-mL tube to the slide represents one
of the major critical points of this protocol. To avoid disrup-
tion of conjugates during their transfer, cut approximately
0.5 cm from the 200-μL tips, one per sample.
28. Wells can also be washed by immersing the slide into a
Hellendahl-type dish containing PBS. Then the space between
wells has to be carefully dried using 3MM paper.
29. Permeabilization of T-B cell conjugates allows the secondary
Ab to stain both CD3/Ab complexes which recycle to the
plasma membrane of the IS and CD3/Ab complexes which do
not reach the plasma membrane and remain entrapped in an
endosomal compartment possibly polarized beneath the IS. To
specifically track CD3/Ab complexes that have recycled at the
plasma membrane, this permeabilization step has therefore to
be skipped (Fig. 1).
30. One of the main hallmarks of IS formation is the polarization
of microtubule-organizing center (MTOC) beneath the IS
[16]. To distinguish between intrinsic trafficking defects in the
recycling pathway and defects caused by the failure of the
MTOC to polarize toward the IS, it is useful to introduce an
optional staining step with an anti-γ tubulin Ab, which specifi-
cally detects the MTOC, allowing therefore to check MTOC
polarization in every T-B cell conjugate analyzed.
Acknowledgments
This work was supported by AIRC, ITT-Regione Toscana, and

Telethon (Grant GGP11021).
References
1. Maxfield FR, McGraw TE (2004) Endocytic immune synapse: lessons from the primary cil-
recycling. Nat Rev Mol Cell Biol 5:121–132 ium. Traffic 16:241–249
2. Das V, Nal B, Dujeancourt A, Thoulouze MI, 4. Onnis A, Finetti F, Baldari CT (2016) Vesicular
Galli T, Roux P et al (2004) Activation-induced trafficking to the immune synapse: how to
polarized recycling targets T cell antigen recep- assemble receptor-tailored pathways from a
tors to the immunological synapse; involvement basic building set. Front Immunol 7:50
of SNARE complexes. Immunity 20:577–588 5. Finetti F, Paccani SR, Riparbelli MG,
3. Finetti F, Onnis A, Baldari CT (2015) Giacomello E, Perinetti G, Pazour GJ et al
Regulation of vesicular traffic at the T cell (2009) Intraflagellar transport is required for
polarized recycling of the TCR/CD3 complex 12. Finetti F, Patrussi L, Galgano D, Cassioli C,
to the immune synapse. Nat Cell Biol 11: Perinetti G, Pazour GJ et al (2015) The small
1332–1339 GTPase Rab8 interacts with VAMP-3 to regu-
6. Alcover A, Alarcón B (2000) Internalization late the delivery of recycling T-cell receptors to
and intracellular fate of TCR-CD3 complexes. the immune synapse. J Cell Sci 128:
Crit Rev Immunol 20:325–346 2541–2552
7. Geisler C (2004) TCR trafficking in resting and 13. Patrussi L, Capitani N, Martini V, Pizzi M,
stimulated T cells. Crit Rev Immunol 24:67–86 Trimarco V, Frezzato F et al (2015) Enhanced
8. Finetti F, Patrussi L, Masi G, Onnis A, Galgano chemokine receptor recycling and impaired
D, Lucherini OM et al (2014) Specific recy- S1P1 expression promote leukemic cell infiltra-
cling receptors are targeted to the immune syn- tion of lymph nodes in chronic lymphocytic
apse by the intraflagellar transport system. leukemia. Cancer Res 75:4153–4163
J Cell Sci 127:1924–1937 14. Onnis A, Finetti F, Patrussi L, Gottardo M,
9. Margadant C, Kreft M, de Groot DJ, Norman Cassioli C, Spanò S et al (2015) The small
JC, Sonnenberg A (2012) Distinct roles of talin GTPase Rab29 is a common regulator of
and kindlin in regulating integrin α5β1 func- immune synapse assembly and ciliogenesis. Cell
tion and trafficking. Curr Biol 22:1554–1563 Death Differ 22:1687–1699
10. Liu GY, Kulasingam V, Alexander RT, Touret 15. Zerial M, McBride H (2001) Rab proteins as
N, Fong AM, Patel DD et al (2005) Recycling membrane organizers. Nat Rev Mol Cell Biol
of the membrane-anchored chemokine, 2:107–117
CX3CL1. J Biol Chem 280:19858–19866 16. Fooksman DR, Vardhana S, Vasiliver-Shamis
11. Patel S, Mukovozov I, Robinson LA (2011) G, Liese J, Blair DA, Waite J et al (2010)
Assessment of the recycling of the membrane- Functional anatomy of T cell activation and
bound chemokine, CX3 CL1. Methods Mol synapse formation. Ann Rev Immunol 28:
Biol 748:143–153 79–105
Chapter 11
Simultaneous Membrane Capacitance Measurements

and TIRF Microscopy to Study Granule Trafficking
at Immune Synapses
Marwa Sleiman, David R. Stevens, and Jens Rettig
Abstract
Whole-cell capacitance measurements allow the direct measurement of exocytosis with high temporal reso-
lution. An added benefit of the whole-cell configuration is the possibility to control the cytosolic free
calcium concentration allowing examination of the role of intracellular calcium in a variety of processes.
We have coupled this method with imaging of cytotoxic granule release using total internal reflection fluo-
rescence microscopy (TIRFM) to identify the capacitance steps associated with cytotoxic granule release
identified by TIRFM. This requires the use of fluorescent granule markers to identify cytotoxic granules
and allows characterization of cytotoxic granule fusion and of the behavior of cytotoxic granules at the
immune synapse prior to fusion. Combination of these methods enables the study of a number of processes
relevant to the function of the immune synapse.
Key words Capacitance measurements, TIRF microscopy, Cytotoxic T lymphocytes, Cytotoxic

granules, Exocytosis
1 Introduction
The killing of virally infected cells or tumor cells by cytotoxic T

lymphocytes (CTLs) requires delivery of cytotoxic granules (CG)
to the immune synapse (IS) and subsequent release of their cyto-
toxic substances [1]. The cytotoxic contents of CG are released by
exocytosis and then act on the target cell, resulting in target cell
death. The biogenesis, trafficking, and fusion of CG with the
plasma membrane are fascinating processes that are still being
sorted out [2, 3]. It is a rapid and highly regulated process requir-
ing methods with high both spatial and temporal resolution.
The application of total internal reflection fluorescence micros-
copy (TIRFM) to the study of secretory granules has resulted in
important insights [4, 5]. TIRFM allows high-resolution imaging
in the X and Y planes with a small range of visibility in the Z axis
157
158 Marwa Sleiman et al.
with very good signal-to-noise characteristics. Since the evanescent

wave extends only about 200 nm from the coverslip into the cell,
the technique allows observation of granules at the plasma mem-
brane in the area where a cell adheres to the coverslip [6, 7].
Application of anti-CD3 onto the coverslips induces CTLs to react
to the coverslip as they would to a target cell, orienting the
cytotoxin-releasing immune synapse to the coverslip [8, 9]. Thus,
CG move to the coverslip, entering the TIRFM field where they
can be observed. This allows observation of prefusion activity and
of CG fusion.
The application of whole-cell capacitance measurements to the
recording of exocytotic events requires voltage clamping but deliv-
ers an even higher time resolution [10]. Though this method has
been used for many studies of neuronal and neuroendocrine cells
and was applied to mast cells [11] and eosinophils [12] at its incep-
tion, it has rarely been applied to T lymphocytes.
Combination of these two methods allows independent mea-
sures of exocytosis [13, 14]. The use of patch clamp allows simul-
taneous control of the intracellular milieu and allows application of
substances to the cytoplasm but gives no spatial information in the
whole-cell configuration.
Here we present the protocols used to prepare and transfect
human primary CTLs and the constructs we use to introduce fluo-
rescent probes for identification and observation of CG in
TIRFM. We describe the solutions used and possible pitfalls inher-
ent in patch clamping and in the application of whole-cell capaci-
tance measurements in CTLs.
2 Materials
2.1 Isolation, 1. Dynabeads® Untouched™ Human CD8 T cells kit (Thermo

Activation, and Fisher Scientific, Catalog no. 11348D) and Magnet (Invitrogen.
Transfection of Primary, com/magnet selection) or equivalent negative selection-based
Human CTLs purification method.
2. Isolation buffer: Ca2+- and Mg2+-free phosphate-buffered
saline (PBS) (e.g., Gibco) supplemented with 0.1% BSA and 2
mM EDTA.
3. Serum-free medium; e.g., AIM V medium (Gibco Life
Technologies, Catalog no. 12055-091).
4. Heat-inactivated fetal bovine serum (FBS)/fetal calf serum
(FCS).
5. Mixer for tilting and rotation; e.g., Reax 2 (Heidolph).
6. Dynabeads Human T-Activator CD3/CD28 kit (Life
Technologies): The kit components and buffers need to be
kept cold.
Simultaneous Membrane Capacitance Measurements and TIRF Microscopy to Study… 159
7. Human CD8+ T lymphocytes.

8. Electroporator; e.g., Amaxa Nucleofector™ 2b Device (Lonza).
9. Electroporation cuvettes and reagents; e.g., Human T cell
Nucleofector kit (Lonza).
10. Human granzyme B-mCherry cDNA construct (see Note 1).
2.2 Total Internal 1. Inverted epifluorescence equipped microscope; e.g., Axiovert

Reflection 200 microscope (Zeiss, Göttingen, Germany).
Fluorescence 2.
TIRF-optimized objective and TIRF illuminator; e.g.,
Microscopy (TIRFM) 100×/1.45 NA Fluar objective (Zeiss, Göttingen, Germany)
and Zeiss TIRF-slider for Axiovert 200.
3. Solid-state laser emitting at 561 nm; e.g., 85YCA010 (Melles
Griot, Carlsbad, CA).
4. Filter set containing a UV-reflecting dual-band dichroic mirror
and an emitter to observe mCherry fluorescence (excitation
maximum ~590 nm; [15]); e.g., supplied by AHF
Analysentechnik (Tübingen, Germany).
5. EMCCD camera; e.g., Andor iXonEM (Belfast, Ireland; see
Note 2).
2.3 Patch-Clamp 1. Extracellular solution (0 mM [Ca2+] Ringer’s solution; see

Membrane Note 3): 155 mM NaCl, 4.5 mM KCl, 5 mM HEPES, 3 mM
Capacitance MgCl2×6H2O, pH = 7.4. Store at 4 °C (see Note 4).
Measurements 2. Intracellular solution ([Ca2+]i ~ 2 μM): 5.9 mM EGTA, 5.4
mM CaCl2, 40 mM HEPES, 2 mM MgATP, 0.3 mM Na2GTP,
120 mM Cs+ Glutamate, pH 7.3 (see Notes 5–10).
3. 25 mm glass coverslips.
4. 0.1 mg/mL poly-l-ornithine (Sigma) (see Note 11).
5. Anti-CD3 antibody (mouse anti-human CD3, Cell Sciences
clone B-B11).
6. Phosphate-buffered saline (PBS)—155 mM NaCl, 2.7 mM
Na2HPO4, 1.54 mM KH2PO4, pH 7.2.
7. Patch-clamp setup: Patch-clamp amplifier (we use the Heka
line of amplifiers, either EPC9 or EPC10, with Pulse software),
microscope (see Subheading 2.2), micromanipulators for the
patch pipette placement (we use simple combined mechani-
cal/piezo manipulators in the X, Y, and Z axis) (Physik
Instrumente, Germany), a vibration isolation table (Newport
Instruments, USA), Faraday cage (homemade), fluorescence
light source (monochromator or appropriate lasers, we use the
polychrome IV or V monochromator built by Till Photonics)
(the company has since changed hands, now FEI), and Sutter
P-97 pipette puller (Sutter Instrument, USA). There are sev-
eral vertical pipette pullers that are adequate for patch-clamp
pipettes that are considerably less expensive.
3 Methods
3.1 Isolation This protocol is based on the negative isolation of untouched

and Activation CD8+ T cells from peripheral blood mononuclear cells (PBMC) by
of Primary, Human depleting B cells, NK cells, CD4+ T cells, monocytes, dendritic
CTLs cells, and granulocytes with a mixture of biotinylated monoclonal
antibodies against non-CD8+ T cells. This method is scaled for
~100 × 106 peripheral blood mononuclear cells.
1. Centrifuge at 235 × g for 8 min at room temperature (RT).
2. Resuspend pellet in 1 mL of cold Isolation Buffer with 200 μL
of FCS and 200 μL of the antibody mixture.
3. Mix well and rotate (e.g., Reax 2 Heidolph) at 20 rpm for
20 min at 2–8 °C.
4. During this time, do the washing step of Depletion MyOne
SA Dynabeads.
5. Resuspend beads in vial by vortexing for >30 s (or tilt and
rotate for 5 min).
6. Transfer 1 mL of beads to a tube.
7. Add an equal volume of Isolation Buffer and mix by pipetting
(without introducing bubbles!).
8. Place the tube in magnet for a few minutes and discard
supernatant.
9. Remove tube from magnet and resuspend beads with 1 mL of
Isolation Buffer and repeat step 8.
10. Keep in 1 mL of Isolation Buffer.
11. Wash the cells by filling the tube with Isolation Buffer. Mix
well by tilting the tube several times and centrifuge at 380 × g
for 8 min at 2–8 °C.
12. Discard supernatant of washed cells and resuspend in 1 mL of
Isolation Buffer with 1 mL of washed Dynabeads.
13. Incubate for 15 min at 18–25 °C with gentle tilting and

rotation.
14. Resuspend bead-bound cells by vigorously pipetting >10 times
using a pipette with narrow tip opening (e.g., 1000 μL pipette
tip).
15. Fill the tube with Isolation Buffer and place in magnet for
6 min (for optimal collection of non-CD8+ T cells).
16. Transfer supernatant containing the untouched CD8+ T cells
to a new tube.
17. Resuspend bead-bound cells with 5 mL of Isolation Buffer
and let stand in magnet for 6 min.
18. Combine the two supernatants.
19. Count the cells and then centrifuge at 235 × g for 8 min at RT.
20. Resuspend the cells in AIMV medium supplemented with 10%
FCS (1.5 × 106 cells/mL in 6-well plate).
21. Activate the cells with Dynabeads Human T-Activator CD3/
CD28 at a 1:0.7 CTLs to beads ratio.

22. Incubate at 37 °C, 5% CO2 for 48 h before doing
transfection.
3.2 Transfection 1. For transfection prepare 2 mL (for each transfection) of FCS-

of Human CTLs supplemented AIMV in 24-well plate in humidified 37 °C, 5%
CO2 incubator to equilibrate. Put Human T cell Nucleofector
solution at RT (when opened for the first time, supplied sup-
plement is added to Nucleofector solution).
2. Collect 5 × 106 CTLs in a tube.
3. Place tube in magnet for few minutes to detach beads.
4. Collect supernatant in new tube and centrifuge the cells at 157
× g for 8 min at RT.
5. Resuspend the pellet with 10 mL of pre-warmed Isolation
Buffer and centrifuge again (same as in step 4).
6. During this time, prepare supplied cuvettes and pipettes. Turn
on Amaxa Nucleofector™ and choose program T-023.
7. Remove supernatant without leaving any traces and resuspend
pellet in 100 μL of Nucleofector solution per transfection then
add 1 μg of plasmid DNA (granzyme B-mCherry; see Note 1).
8. Transfer total volume to cuvette without introducing any bub-
bles and seal with the cap.
9. Take out the 24-well plate out of incubator and collect some
of the pre-equilibrated media with the pipette.
10. Place cuvette in the Nucleofector™ Device and carry out the
electroporation.
11. After the end of electroporation (takes few seconds), take out
the cuvette and add the media from the pipette.
12. Aspirate the total volume of cells with media without bubbles
and add it gently to the initial volume of media in 24-well
plate.
13. Let CTLs rest in incubator for 6 h.
14. Centrifuge the 2 mL of cells at 157 × g for 8 min at RT.
15. Resuspend cells in 2 mL of pre-equilibrated AIMV medium.
16. Add 2 μL of pre-prepared aliquot of human IL-2 (6 μL of
1000 U/μL IL-2 + 54 μL of AIMV with 10% FCS).
17. Experiments can be started after 12–16 h of incubation in
humidified 37 °C, 5% CO2 incubator.
3.3 Preparation 1. Dilute poly-l-ornithine 1:10 in PBS.

of Anti-CD3-Coated 2. Add coverslips in 6-well plate.
Coverslips
3. Add 40 μL of diluted poly-l-ornithine to the center of each
coverslip.
4. Keep at RT for 20–30 min.
5. Remove residual fluid.
6. Dilute anti-CD3 antibody to 30 μg/μL with PBS (volume of
diluted antibody/coverslip = 40 μL; see Note 3).
7. Add 40 μL of antibody to the center of each coverslip.
8. Keep in incubator at 37 °C for 2 h.
9. Remove residual fluid and store at 4 °C.
3.4 Combined 1. Add 50 μL of transfected cell suspension to the center of

Capacitance coverslip. Allow the cells to attach for 5 min.
and TIRFM 2. Add 0.1 g glucose to 50 mL of extracellular solution (10 mM
Measurements glucose).
3. Carefully add 1 mL of extracellular solution.
4. Transfer chamber to microscope for TIRF and patch clamping.
5. Image acquisition: Images are acquired using software written
in house in the LabVIEW environment (National Instruments,
München, Germany; see Note 2). Pixels were 160 × 160 nm.
6. Images are acquired at 10 Hz with an exposure time of 75 ms.
7. Place coverslip in a recording chamber on the stage of the
microscope.
8. Choose a field of view with a high density of transfected cells.
9. Establish TIRF illumination.
10. A CTL is located and centered in the TIRF field and the foot-
print is brought into focus (see Note 12).
11. The illumination angle is adjusted to achieve TIRF. If

mCherry-labeled CG are present at the CTL-coverslip inter-
face in TIRFM, the cell is patch clamped (see Note 13).
12. For patch clamping human CTLs, we use pipettes with a resis-
tance (when filled with intracellular solution) of about 3–4
MΩ (see Note 14).
13. The pipette lumen is placed under positive pressure (14–18 cm
water) prior to immersion in the bath fluid, and this pressure
is maintained until seal formation.
14. Zero the patch-clamp amplifier (run the Setup macro in

PULSE (EPC9 or EPC10), Heka, Germany).
15. When the patch pipette reaches the target cell membrane,
release of pressure initiates seal formation, which can be
improved by light suction and/or a positive pipette potential
(see Note 15).
16. Once a gigaseal is reached, activate the ON cell macro to reset

the amplifier to zero and cancel the fast capacitance transient.
17. Start the video recording in TIRF mode and check the TIRF
angle.
18. Rupture the membrane under the patch pipette by abrupt
application of negative pressure (e.g., suction) for a short
period (see Notes 16–19). When whole-cell configuration is
established, compensate the cell capacitance (see Note 20).
19. Begin capacitance measurement (start Sine + DC protocol; see
Notes 21–28).
20. Adjust TIRF illumination if necessary.
21. Restart video acquisition.
4 Notes
1. Granzyme B was amplified from human cDNA with primers

5′-TAT ACT CGA GCC ACC ATG CAA CCA ATC CTG
CTT CTG-3′ and 5′-ATA TAT CCG CGG GTA GCG TTT
CAT GGT TTT CTT T-3′ which adds XhoI and SacII restric-
tion sites at the ends. The mCherry construct was a gift from
Prof. Roger Tsien. After digestion with XhoI and SacII, gran-
zyme B was ligated to vectors containing mCherry-N1 at the
C-terminal end of granzyme B and then purified.
2. Cells were recorded at RT. Tracking of LGs is done using a
centroid algorithm with software written in house [7]. Frame
to frame movement is calculated as the square root of the sum
of the squares of the X and Y displacements. This software was
also used to calculate caging diameters of tracked granules.
3. The production of solutions containing calcium concentra-
tions similar to that found in the extracellular solution is rela-
tively straightforward. However, when solutions approximating
the intracellular milieu are prepared, it is not so simple [16].
When weighing calcium salts for solutions with calcium con-
centrations in the 100–1000 nM range, the contamination
from calcium in the water used, in the glassware used, or in the
other components of the solution may be higher than the
added calcium. This issue is addressed by using high-quality
distilled and deionized water; plastic ware can be used instead
of glass and combinations of calcium chelators such as EGTA
(Kd ~ 171 nM) or BAPTA (Kd ~ 224 nM) and calcium salts,
such that the added Ca2+ is much greater than the unknown
calcium contribution, making the solutions less sensitive to
inevitable contamination [16, 17]. The choice of chelator
used depends on the targeted free [Ca2+]. A rule of thumb is
that the buffering range of the chelator should be within an
order of magnitude of the dissociation constant for the metal

of interest, but this already pushes the boundary for adequate
chelation.
4. Solutions are buffered with HEPES buffer. This is particularly
important when using EGTA as a chelator since its chelating
function is pH dependent (see below).
5. To prepare 10 mL of this solution: Weigh 0.176 g glutamic
acid (Sigma G8415–1006) and 0.095 g HEPES and add 6–7
mL of HPLC water. Stir for 1–2 h at low-medium speed with-
out heating. Glutamic acid dissolves slowly. To help dissolve,
add 200–300 μL of 50% CsOH. pH will increase to 4.5. Let
pH equilibrate until it remains stable (this could take hours).
Add slowly 10% CsOH with Pasteur pipette until pH reaches
7.2. Add EGTA and stir well and then add CaCl2. Then finally
add MgATP and Na2GTP. Adjust pH to 7.3 using dilute
CsOH. Let dissolve for few more minutes then transfer beaker
onto ice. Pour the total volume into a 10 mL volumetric flask.
Fill to 10 mL line with HPLC water. Filter with a 0.2 μm filter.
Measure osmolarity. It should have an osmolarity of 300–305
mOsm. Store in 200 μL aliquots at −20 °C.
6. Most calcium chelators bind magnesium almost as well as cal-
cium so even near the Kd the calculated free calcium must take
into account the amount of Mg2+ added to the solution as
well. EGTA is an exception which is one reason for its popu-
larity as a chelator for intracellular solutions.
7. Since the concentration of unbound magnesium is signifi-
cantly higher than that of calcium in the cytoplasm, this is a
very real problem which is typically overcome with the use of
EGTA, which has a good selectivity for calcium over
magnesium. This makes it a good choice in the low to mid-
nM range. A drawback of the use of EGTA is that its chelator
function is very strongly dependent on the pH. For instance,
if a mixture of EGTA and CaCl2 has a free [Ca2+] of 60 nM at
pH 7.0, its free [Ca2+] at pH 7.4 will be near 400 nM. When
using EGTA as a chelator, setting the pH accurately is
extremely important.
8. To improve the accuracy of calcium solutions containing
EGTA, it is advisable to determine its purity. This can be done
by titration of a nominal EGTA solution with a calcium solu-
tion produced by dilution of a known calcium stock solution,
while measuring the pH [18]. The EGTA solution must be
well buffered so that it remains at an alkaline pH (>8.2). Under
these conditions, binding of Ca2+ to the EGTA releases an
equal amount of H+, changing the pH until the EGTA is satu-
rated with Ca2+. Once the EGTA is saturated, the pH remains
constant producing a line with slope 0. Fitting a line to the pH
vs added-calcium data points should produce a line which

intersects the 0 slope component at the point at which EGTA
becomes saturated. The added calcium required to reach satu-
ration reflects the actual amount of EGTA present. The ratio
actual: nominal EGTA gives the purity of the EGTA used.
9. BAPTA is an alternative to EGTA. It is more expensive than
EGTA but is not as pH dependent. It is not as selective for Ca2+
as EGTA is and binds Mg2+ with a similar affinity as it does Ca2+.
The binding constants are experimentally established so the
amount of BAPTA needed for a given free [Ca2+] can be calcu-
lated if the Mg2+ present is considered. If the free [Ca2+] should
be >~2 μM, a different chelator should be chosen, though the
higher the targeted calcium concentration goes, the less critical
the use of a chelator becomes. A number of candidate calcium
chelators are available and their properties have been discussed
[17]. For intracellular patch solutions with free calcium in the
low to mid-nM range, they are required.
10. The use of calcium-sensitive electrodes to test solutions, when
available, is an alternative, but their accuracy is also dependent
on the accuracy of electrode calibration solutions. The lower
the calcium concentration, the less accurate. It must be appre-
ciated that at very low nominal concentrations, the free cal-
cium concentration is an approximate value.
11. We use polyornithine plus anti-CD3, which induces strong
attachment and immune synapse formation at the CTL-
coverslip interface allowing total internal reflection imaging of
granule exocytosis during patch-clamp experiments. Since
CTLs are very active and tend to wander around in the cham-
ber, the use of a proper substrate can be quite helpful. For
TIRFM study of immune synapses, the CTL is enticed to
adhere to the coverslip by adding anti-CD3 (or some other
appropriate substrate) to the coverslip, which leads to strong
adhesion, arresting the T cells and activating them. The use of
target cells is not possible since the immune synapse will not
be visible in the TIRFM plane.
12. Cells that are in contact with the anti-CD3 antibodies adhere
and change shape, going from flat, highly mobile cells to
rounded, adherent cells during the activation process. These
cells tend to flatten out over time, increasing the difficulty of
patch clamping them. T cell receptor (TCR) binding to the
anti-CD3 antibody leads to its clustering which is required for
activation [19].
13. Use of TIRFM to identify CG fusion events in CTLs induced
to form an IS at a treated coverslip is well established [6, 20,
21]. Granzyme B was chosen as the CG marker because it is
present in human CTLs and its expression increases upon acti-
vation [22]. It was coupled to mCherry [15]. Seeding of CTLs

on anti-CD3/CD28-coated coverslips results in microtubule
reorientation; accumulation of actin, CD3, and cytotoxic
granules at the CTL-coverslip interface; and CG release
consistent with IS formation [9].
14. When filling patch pipettes, air bubbles can be a nuisance,
depending on tip shape: We use glass with a glass fiber in the
lumen which facilitates filling. Filling solutions should not be
injected into the tip of the pipette which increases the likeli-
hood of bubbles in the tip, nor should they be injected at the
open end of the pipette; rather, they should be loaded around
the middle and allowed to move to the tip by capillary action.
Once the tip has been filled, the remaining fluid can be encour-
aged to move to the tip by rapidly swinging the pipette while
holding the open end. Air bubbles can be removed after filling
by tapping with the fingernail or by pulling the pipette through
the jaws of a serrated forceps which sets up vibrations which
loosen the bubbles so they can float upward away from the tip.
We do both.
15. Patch clamping isolated human or mouse CTLs is not trivial.
For experimenters with patch-clamp experience and a well
thought-out setup, it should not be too daunting.
16. Patch-clamp recordings were carried out in the whole-cell

configuration [23]. Patch pipettes were pulled from thick-
walled borosilicate glass (GB150F-8P, Science Products,
Germany). After a GΩ seal forms, the focus is adjusted and
TIRF-video acquisition is begun. In addition to voltage clamp,
the whole-cell configuration allows control of the intracellular
fluids, including the free Ca2+ concentration.
17. Since a stable increase in intracellular [Ca2+] is required for
effector function in CTLs [24, 25], we supply ~2 μM free
intracellular Ca2+ via the patch pipette. The experiments are
carried out in a nominally calcium-free extracellular solution
to slow CTL activation and to prevent the release of granules
prior to recording. To examine the calcium requirement for
exocytosis, the intracellular free calcium can be adjusted.
18. Calcium-containing intracellular solutions: Changes in free

intracellular calcium were carried out by changing the amounts
of CaCl2 and EGTA. Calculation of free intracellular calcium
concentration in intracellular recording solutions for patch
clamp was carried out using the Maxchelator programs http://
web.stanford.edu/~cpatton/webmaxcS.htm.
19. When CTLs are recorded in the whole-cell configuration with
an adequate intracellular free calcium concentration and are
activated by contact to anti-CD3/CD28, they will release
cytotoxic granule content which can be followed in TIRFM.
20. The capacitance was compensated and the capacitance record

acquisition started. Cells were clamped at a holding potential
of −70 mV using an EPC-9 patch-clamp amplifier controlled
with the “lock-in” extension of PULSE software (Heka,
Germany). For whole-cell capacitance measurements which
we have used to measure the capacitance steps associated with
exocytosis in human CTLs, we use the sine + DC method with
a sine wave frequency of 1000 Hz and amplitude of 50 mV
RMS. The voltage-clamp data are acquired as a single continu-
ous record. The camera trigger signal is recorded to allow cor-
relation of video with capacitance data.
21. The capacitance is proportional (constant 1 μF/cm2) to the
surface area. Exocytosis results in addition of the granule mem-
brane to the plasma membrane resulting in an increase in the
membrane capacitance, so total release can be measured. If the
signal-to-noise ratio is low enough, single granule fusion events
can be recorded allowing granule size to be determined.
22. Capacitance recordings are carried out using the software

Lock-in module of the Pulse suite of programs (Heka,
Germany) controlling an EPC9 or EPC10 patch-clamp ampli-
fier. The cell capacitance can be determined since the cell
membrane has a capacitive component which shows up in
voltage-clamp records as an additional current component as
the capacitance is charged. The capacitive current flows when-
ever the membrane potential changes, for instance, at the
beginning and end of a step depolarization in voltage clamp.
This transient can be used to estimate cell capacitance.
23. There are alternative methods available for capacitance mea-
surements (for a description, see [26]). We use the Sine + DC
method which is well suited for whole-cell capacitance mea-
surements. For high time resolution capacitance measure-
ments, the voltage stimulus used is a sine wave, and changes in
capacitance produce a phase shift in the resulting current trace.
The Lock-in software reads and tracks the amplifier variables
required for the calculation and calculates the membrane
capacitance (a description of the method and its use can be
found at http://www.heka.com/support/tutorials/tutori-
als_down/pm_tutorial_lockin.pdf). Since a value is calculated
for each cycle, the time resolution depends on the frequency
of the sine wave.
24. Depending on the conditions, the cells being recorded and
the series resistance between the pipette and the cell, frequen-
cies in the thousands of Hz range can be applied which allows
very high temporal resolution. This is one reason sine wave-
based methods have become very popular. The sine + DC
method is quite robust and has the advantage that it can record
absolute changes in capacitance rather than relative differences
reported by the alternative (piecewise-linear) method. This is

possible due to the use of the added DC holding potential.
Technically for the method to function, the reversal potential
of the active membrane currents should be known, but under
typical whole-cell conditions in which the membrane resis-
tance is greater than the series resistance, the errors in the esti-
mate of reversal potential for the membrane currents are very
small and can be ignored [27], and zero mV is usually assumed
as the reversal potential.
25. If the series resistance is high, it can lead to decreased perfor-
mance of the voltage clamp and to errors in calculating the
capacitance. We have tested a range of sine wave frequencies in
test cells to confirm that the different sine wave frequencies
generate consistent Cm values and found that as long as series
resistance stayed low (<10–15 MΩ), the results were consis-
tent with frequencies up to about 5 kHz. Since the series resis-
tance can change during a recording and millisecond resolution
is more than sufficient for cytotoxic granule exocytosis, we use
1000 Hz, which leaves adequate room for error.
Counterintuitively, increasing the amplitude of the sine wave
voltage stimulus often improves the capacitance noise.
26. Since CTLs do not exhibit strong voltage-dependent mem-
brane conductance, there is no reason not to use relatively high
sine wave amplitudes (e.g., 50 mV RMS, coupled with a nega-
tive holding potential of −50 to −70 mV) for the stimulus.
27. Sampling frequencies should be chosen which allow a mini-
mum of nine sampled points per sine wave cycle or accuracy
will be sacrificed.
28. Fusion events produce discrete jumps in membrane capacitance
[28]. In order to improve accuracy of measurement of the
amplitude of capacitance steps, the capacitance traces are fil-
tered using a low-pass filter implemented in Igor Pro 6
(WaveMetrics, USA). This allows us to distinguish steps down
to ~1 fF. The accompanying membrane conductance and access
conductance measurements are also recorded. Stepwise
increases in capacitance are discarded if they are accompanied
by abrupt changes in either the access or membrane conduc-
tance, since such events, in particular those in the access con-
ductance trace, can cause a change in the capacitance trace [27].
Acknowledgment
Work in our lab was supported by grants from the Deutsche

Forschungsgemeinschaft (SFB 894 to JR).
References
1. de Saint Basile G, Menasche G, Fischer A cytotoxic T lymphocytes. PLoS One

(2010) Molecular mechanisms of biogenesis 10:e0135994
and exocytosis of cytotoxic granules. Nat Rev 15. Shaner NC, Steinbach PA, Tsien RY (2005) A
Immunol 10:568–579 guide to choosing fluorescent proteins. Nat
2. Page LJ, Darmon AJ, Uellner R et al (1998) L Methods 2:905–909
is for lytic granules: lysosomes that kill. Biochim 16. Bers DM, Patton CW, Nuccitelli R (1994) A
Biophys Acta 1401:146–156 practical guide to the preparation of Ca2+ buf-
3. Stinchcombe JC, Page LJ, Griffiths GM (2000) fers. Methods Cell Biol 40:3–29
Secretory lysosome biogenesis in cytotoxic T 17. Patton C, Thompson S, Epel D (2004) Some
lymphocytes from normal and Chediak Higashi precautions in using chelators to buffer metals in
syndrome patients. Traffic 1:435–444 biological solutions. Cell Calcium 35:427–431
4. Oheim M, Loerke D, Stühmer W et al (1998) 18. McGuigan JAS, Stumpff F (2013) Calculated
The last few milliseconds in the life of a secre- and measured [Ca(2+)] in buffers used to cali-
tory granule. Docking, dynamics and fusion brate Ca(2+) macroelectrodes. Anal Biochem
visualized by total internal reflection fluores- 436:29–35
cence microscopy (TIRFM). Eur Biophys 19. Choudhuri K, Dustin ML (2010) Signaling
J 27:83–98 microdomains in T cells. FEBS Lett
5. Oheim M, Stühmer W (2000) Interaction of 584:4823–4831
secretory organelles with the membrane. 20. Halimani M, Pattu V, Marshall MR et al (2014)
J Membr Biol 178:163–173 Syntaxin11 serves as a t-SNARE for the fusion
6. Martina JA, Wu XS, Catalfamo M et al (2011) of lytic granules in human cytotoxic T lympho-
Imaging of lytic granule exocytosis in CD8+ cytes. Eur J Immunol 44:573–584
cytotoxic T lymphocytes reveals a modified 21. Bertrand F, Muller S, Roh K-H et al (2013) An
form of full fusion. Cell Immunol initial and rapid step of lytic granule secretion
271:267–279 precedes microtubule organizing center polar-
7. Pasche M, Matti U, Hof D et al (2012) ization at the cytotoxic T lymphocyte/target cell
Docking of LDCVs is modulated by lower synapse. Proc Natl Acad Sci 110:6073–6078
intracellular [Ca2+] than priming. PLoS One 22. Grossman WJ, Verbsky JW, Tollefsen BL et al
7:e36416 (2004) Differential expression of granzymes
8. Schwarz EC, Kummerow C, Wenning AS et al A and B in human cytotoxic lymphocyte sub-
(2007) Calcium dependence of T cell prolifera- sets and T regulatory cells. Blood
tion following focal stimulation. Eur J Immunol 104:2840–2848
37:2723–2733 23. Hamill OP, Marty A, Neher E et al (1981)
9. Pattu V, Qu B, Marshall M et al (2011) Improved patch-clamp techniques for high-
Syntaxin7 is required for lytic granule release resolution current recording from cells and
from cytotoxic T lymphocytes. Traffic cell-free membrane patches. Pflügers Arch
12:890–901 391:85–100
10. Lindau M (2012) High resolution electrophys- 24. Hogan PG, Lewis RS, Rao A (2010) Molecular
iological techniques for the study of calcium- basis of calcium signaling in lymphocytes: STIM
activated exocytosis. Biochim Biophys Acta and ORAI. Annu Rev Immunol 28:491–533
1820:1234–1242 25. Pores-Fernando AT, Zweifach A (2009)
11. Fernandez JM, Neher E, Gomperts BD (1984) Calcium influx and signaling in cytotoxic
Capacitance measurements reveal stepwise T-lymphocyte lytic granule exocytosis.
fusion events in degranulating mast cells. Immunol Rev 231:160–173
Nature 312:453–455 26. Gillis KD (2000) Admittance-based measure-
12. Lindau M, Nüsse O, Bennett J et al (1993) ment of membrane capacitance using the
The membrane fusion events in degranulating EPC-9 patch-clamp amplifier. Pflügers Arch
guinea pig eosinophils. J Cell Sci 104(Pt 439:655–664
1):203–210 27. Gillis KD (1995) Techniques for membrane
13. Nofal S, Becherer U, Hof D et al (2007) capacitance measurements. In: Single-channel
Primed vesicles can be distinguished from recording. Plenum, pp 155–198
docked vesicles by analyzing their mobility. 28. Neher E, Marty A (1982) Discrete changes of
J Neurosci 27:1386–1395 cell membrane capacitance observed under
14. Ming M, Schirra C, Becherer U et al (2015) conditions of enhanced secretion in bovine
Behavior and properties of mature lytic gran- adrenal chromaffin cells. Proc Natl Acad Sci U
ules at the immunological synapse of human S A 79:6712–6716
Chapter 12
Mathematical Modeling of Synaptic Patterns

Anastasios Siokis, Philippe A. Robert, and Michael Meyer-Hermann
Abstract
During antigen recognition by T cells, a specific spatial structure is formed at the contact face to an
antigen-presenting cell (APC), called an immunological synapse (IS). The IS supports bidirectional signal-
ing and release of effector molecules and is widely studied both biologically and numerically, in order to
understand the process of T cell activation and signaling. This specialized structure harbors a central area
(central supramolecular activation cluster, cSMAC) populated by T cell receptor-peptide-major histocom-
patibility complex (TCR-pMHC) interactions, hedged by a peripheral ring (peripheral supramolecular
activation cluster, pSMAC) of integrin lymphocyte function associated-1 interactions with its immuno-
globulin superfamily ligand intercellular adhesion molecule-1 (LFA-1-ICAM-1). These two regions form
the “bull’s eye” pattern characteristic of the mature IS.
In theoretical studies, different modeling architectures, including partial differential equations (PDE)
and agent-based models, have been developed with the purpose to answer mechanistic questions about the
IS dynamics. In this chapter, we explain possible physiological mechanisms that lead to the formation of
ISs and technical issues that may occur in the course of development of agent-based models.
Key words Immunological synapse, Patterns, Agent-based modeling, Partial differential equations
(PDEs), Mechanics, Computational biology
1 Introduction
The study of the IS initially began at the end of the 1990s and begin-
ning of the 2000s [1], and the description of this phenomenon has
undergone continuous improvement with advances in microscopy
and molecular methods [2–4]. While biologists focused on details of
cytoskeletal and vesicular transport as candidate mechanisms, physi-
cists envisioned important contributions of membrane bending, and
mathematicians took unbiased approaches to the discovery of mini-
mal forces acting between molecules that could lead to these patterns.
In this chapter, we will focus on the mathematical approaches to build
minimal systems, explaining in depth the agent-based approach.
Inquiries regarding this work may be addressed to Anastasios Siokis <Anastasios.Siokis@helmholtz-hzi.de> or

Michael Meyer-Hermann <mmh@theoretical-biology.de>.
171
172 Anastasios Siokis et al.
In the face of biologists’ arguments that actomyosin-based con-

traction and vesicular trafficking could generate immunological
synapse patterns, Qi et al. established a partial differential equation
(PDE) model in order to examine whether or not the formation of
the synapse could be driven by self-assembly involving extracellular
receptor interactions and membrane bending [5, 6]. The model
consists of a set of reaction diffusion equations coupled to one that
describes potential motion toward a free energy minimum for the
membrane shape changes [6]. In this work, they conclude, with
assistance from physicochemical analysis of relevant processes, that
the formation of the IS may simply emerge from self-assembly pro-
cesses. Also, this model could be extended to test different systems,
like many different peptides on the APC surface or natural killer
(NK) cells, instead of T lymphocytes, with the limitation that recep-
tors and ligands are embedded in opposed deformable membranes
whose properties could be derived from the cytoskeleton [7].
Burroughs and Wülfing created another PDE model, trying to
prove that the mechanism of the IS formation is supported by the
length difference in the different kinds of complexes (TCR-pMHC
and LFA-1-ICAM-1), by incorporating reaction kinetics, thermo-
dynamics, and elasticity of the complexes and the membranes [8].
They concluded that the different bond lengths induce a signifi-
cant contribution to the free energy of the interaction, which
means that the membrane deformations are important for the
dynamics of the IS formation.
On the other hand, Weikl et al. introduced a statistical mechanics
model [9, 10], which includes thermal fluctuations. This model
incorporates different forces, including membrane deformations, in
an energy dependent manner, which describes the global configura-
tion of densities at each site of the lattice. Densities are locally updated
by virtue of a Monte Carlo simulation and depend on the variation
in the free energy. An extension of this model would be to explicitly
distinguish agents, which would not change the outcome of the sim-
ulations. The authors observe that the “bull’s eye” pattern is initially
reversed, with TCR-pMHC complexes forming at the periphery of
the contact region, while LFA-1-ICAM-1 complexes are present in
the cSMAC. They confirm the findings of [5, 6], but in contrast to
other theoretical studies, this model needs the cytoskeletal transport
process, in order to obtain the final pattern of the IS.
More recently, Carlson and Mahadevan introduced an updated
PDE model [11]. This model takes into account the membrane
mechanics, the protein binding kinetics and motion, as well as the
fluid flow in the synaptic cleft between the two cells. In comparison
with experimental data, the authors were able to show that passive
elastohydrodynamics and protein binding kinetics are able to
describe the formation and organization of protein clusters.
Interestingly this model does not invoke active cytoskeletal trans-
port processes.
Mathematical Modeling of Synaptic Patterns 173
Here, we will focus on how to create an agent-based model in

a way to describe physiological mechanisms that lead to the IS
structure. The basis of an agent-based model is that individual
interacting agents represent molecules that form patterns in the IS
and follow specific rules. The following description of agent-based
techniques for IS modeling is based on the work by Figge and
Meyer-Hermann [12, 13].
This methodology was further used by [14, 15] for modeling
a B cell-APC interaction and investigating whether the transport of
molecules by the cytoskeleton toward the center is a potential
mechanism for IS formation. The working hypothesis of this model
is that the diffusion of the molecules is biased toward the center of
the contact region, which reproduces the experimentally observed
IS for realistic time and affinity levels.
In the following, we present a step-by-step method to repre-
sent agents and to apply agent-based modeling to the IS at a level
to be easily reproduced by the community. We highlight specific
challenges of this approach and propose new features to avoid
common biases.
2 Materials
The following model is presented in an algorithmic spirit, which

enables the reader to formulate it in any programming language.
The results presented here were programmed using object-oriented
programming in C++ language. The code is separated in two
classes: one representing the agents and, the other, the lattice
including the rules applied to these agents. The completion time of
a simulation takes about 6 h of CPU time (using a 3.47 GHz CPU)
for 30 min of IS formation. The random generator used is
mersenne twister engine (mt19937), and the plots were generated
using the Qt framework.
3 Procedure
3.1 Dynamical ISs are complex cell-cell junctions that involve many interactions,
Properties but we will build the agent-based model around a simplified exper-
of the Immunological imental setup that recapitulates IS formation from the first contacts
Synapse to the formation of the “bull’s eye” pattern. In this experiment,
the APC is replaced by a supported lipid bilayer (SLB) coated with
laterally mobile and fluorescently tagged pMHC and ICAM-1
molecules. T lymphocytes interact with the SLB surface at 37
°C. An advantage of the SLB system is that the number of mole-
cules in the bilayer and the number of interactions with receptors
can be quantified. If TCRs in nascent contacts recognize pMHC,
this triggers activation of LFA-1 and leads to rapid spreading and

then formation of an IS [4].
Emerging dynamical properties of the IS can be observed in
the T cell-SLB system. The TCR-pMHC and LFA-1-ICAM-1
interactions are initiated by forming microclusters in the periphery
of the IS that merge into bigger clusters and finally travel toward
the center of the IS. Throughout this process, the T cell forms
protrusions, in order to anchor on the SLB surface, in which new
TCR-pMHC microclusters are formed [4, 16]. The LFA-1-
ICAM-1 microclusters undergo similar centrally directed move-
ment but generally disengage inside of the ring, and the components
are further recycled to the periphery to maintain the adhesion ring
[3]. Around 10 min after the initial contact, the IS “bull’s eye”
reaches a dynamic steady state. This pattern can remain stable for
minutes to hours [1, 2].
The two main unresolved questions around this pattern are,
firstly, the local forces acting on the molecular constituents that
lead to this specialized structure and, secondly, how signaling is
initiated and sustained. The central TCR accumulation has been
associated with signal enhancement and signal termination [2].
The main aim, for now, is to discuss possible mechanisms that lead
to the IS structure and how to model and simulate the IS utilizing
an agent-based approach. Special attention has to be taken on how
to implement the rules and hypotheses for agent movement and
interactions, in order to avoid common mistakes.
3.2 Lattice Firstly, a representation of the T cell surface and the SLB is needed.
Implementation In order to achieve that, a lattice is created for both of them. For the
sake of simplicity, the whole process will be modeled in two dimen-
sions (2D); thus, the lattice will be a square mesh, as depicted in Fig.
1a (see Notes 1 and 2). The nodes are called gridpoints, which are
initially empty. Each gridpoint on the lattice has eight neighbors,
four on the vertical and horizontal direction, called the “nearest
neighbors,” and the other four, the diagonal neighbors, which are
the “next-nearest neighbors.” This is called a Moore neighborhood,
and the movement of the agents is confined to these gridpoints.
Other possible approaches could represent the cell surface, for
example, a lattice with movement in only the four nearest neigh-
bors (von Neumann neighborhood) or even a square plane where
the agents can move in any random direction. But in comparison
with the Moore neighborhood, the von Neumann neighborhood
would require two moves in order for an agent to move diagonally,
which would make the simulation slower. Comparing with the sec-
ond case, utilizing a Moore neighborhood circumvents the use of
a collision detection algorithm. The usage of square lattices has the
advantage of being computationally fast but requires some mea-
sures that repair the artificial discretization of space which bears the
risk of generating artifacts.
a b c
Fig. 1 Lattice implementation. (a) Initialization of the lattice. A square mesh is created, which consists of grid-
points (nodes). (b) On the lattice, agents are created at random positions (x, y) and associated with the grid-
points. (c) The agents/molecules can randomly diffuse or move by interactions on the gridpoints, if and only if
the decided gridpoint position is free
3.3 Creation Now that the working space is ready, it has to be filled with agents.
of the Agents These agents will represent the different kinds of molecules that
are present on the surface of each cell. In order to have a minimal
model, we will restrict the kinds of molecules included in the model
to the most important ones. Note that as for a real cell surface, the
number of different types of molecules can be much larger. The
case examined here is restricted to two types of molecules on each
surface, TCR and LFA-1 on the T cell, and pMHC and ICAM-1
on the SLB surface. The agents are part of the lattices by filling the
gridpoints (see Notes 3 and 4) and will be randomly distributed at
the beginning (Fig. 1b). In the following sections, the agents will
acquire appropriate rules, in order to endorse them with physical
properties.
3.4 Diffusion Molecules diffuse on the surface of cells due to thermally induced
of Agents stochastic motion, which means that it is a random process and can
happen toward any direction. However, because of the Moore
neighborhood representation, diffusion can only happen toward
one of the eight neighbors. Therefore, a random direction genera-
tor will choose one of these neighbors in a probabilistic manner,
provided the selected position is empty, such that the movement
can be accomplished.
Diffusion is represented as a random walk in the lattice, where
the probability to move is defined as the ratio of the simulation
time step to the time an agent (molecule or complex) needs to dif-
fuse on a neighboring gridpoint on the 2D lattice with a type-
specific diffusion constant DX. Here X denotes one of the
molecules (TCR, LFA-1, pMHC, or ICAM-1) or complexes
(TCR-pMHC or LFA-1-ICAM-1) under consideration (see Note 5).
The diffusion of complexes, generated through binding of two
opposite molecules (see Subheading 3.5), is performed by moving

both molecules in the same direction, which happens less often,
justifying the use of a reduced diffusion rate for complexes.
The Moore neighborhood, though, leads to a quantitative
problem. If an agent keeps on moving diagonally, it will cover a
bigger distance than an agent that moves either vertically or hori-
zontally in a specific time interval. This means that the speed of the
first agent will be bigger than that of the second one. In order to
avoid that, the probabilities of moving either a free molecule or
complex to one of the next-nearest (diagonal) neighbors are
reduced by the geometrical factor 1 / 2 . In this way, the differ-
ence in the distance between different gridpoints on the lattice is
taken into account.
In order to account for engaged TCR cross-linking [17], an
adhesive force is implemented between neighboring TCR-pMHC
complexes, by further reducing the diffusion probability. Thus, a
phenomenological adhesive factor f(Nnn) is introduced to the
probability of TCR-pMHC diffusion:
1 / (1 + N nn ), N nn < 4
f (N nn ) = { ,
0, N nn ≥ 4
where Nnn is the actual number of the nearest neighbors to the
site under investigation (see Note 7).
3.5 Molecule Binding While molecules diffuse on the lattices, they are allowed to check
Kinetics the agent at the exact same position on the opposite lattice. If the
agent is appropriate, a rule of complex formation needs to be
defined. This is formulated using a receptor-ligand binding proba-
bility, based on the on-rates for the specific molecules under consid-
eration, Avogadro’s number, and the volume of the complex about
to form. On the other hand, the probability of unbinding a com-
plex into free molecules is based on the off-rates of each individual
type of complex, TCR-pMHC or LFA-1-ICAM-1 (see Note 6).
3.6 Complex By utilizing the model described so far, i.e., including diffusion
Interactions and binding kinetics, the simulations would simply exhibit a ran-
dom movement of molecules and complexes, without any particu-
lar patterning (Fig. 2a). Thus, there is something missing in the
model. It is known that TCR-pMHC complexes have a length of
LTM = 15 nm, while LFA-1-ICAM-1 complexes are three times
longer, with LLI = 45 nm. Due to this length difference, the larger
complexes, although they have some elasticity, cannot fit in the
same region as the smaller ones. Thus, membrane bending occurs
due to surface tension, forcing the longer complexes to segregate
from the shorter ones.
This rule must be applied to the agents, but because the whole
model is in 2D, membrane bending cannot be included. As a
Fig. 2 Step-by-step explanation of the implemented interactions in the model. (a) All molecules and complexes
are governed only by diffusion. There are no patterns emerging, and everything is well intermixed. (b) A repul-
sive interaction between LFA-1-ICAM-1 and TCR-pMHC complexes together with diffusion is implemented in
the model. The two types of complexes segregate, but the emerging multifocal pattern is missing the key
element in order to form a “bull’s eye” pattern. (c) Instead of the repulsive interaction, the model takes into
account an attractive interaction between TCR-pMHC complexes. The resulting pattern is a “dirty bull’s eye”
pattern, where LFA-1-ICAM-1 complexes are trapped in the cSMAC, but now the two types of molecules are
not clearly segregated. (d) When the model takes into account both the repulsive and attractive interactions
together with the diffusion, the “bull’s eye” pattern is generated with a clear segregation of the central and
peripheral SMAC. The formation steps follow that of the experimental observations. The final IS pattern is
formed within 10 min of contact and can stay stable for hours. TCR-pMHC (green) and LFA-1-ICAM-1 (red)
phenomenological approach, another way to impose membrane

bending is to model a repulsive interaction between the different
kinds of complexes. This interaction will be characterized by a
weight, wrep, and a typical length, Lrep. The following protocol is
then used: If the randomly selected agent is a complex, either
TCR-pMHC or LFA-1-ICAM-1, a force is calculated. For each
complex of the other kind within a distance Lrep, its contribution to
the force is computed as the unit vector in its direction and is mul-
tiplied by the weight wrep < 0, which results in a repulsive effect.
The forces are summed and the latter vector will point to the direc-
tion the force would push the agent. This force is converted into a
movement of the agent on the lattice into the direction of one of
the eight neighboring gridpoints, including the probabilistic
weighting of the non-nearest neighbors as discussed in Subheading
3.4. The conversion is explained in details in Note 7 .
This simplified approach is sufficient for the emergence of mul-
tifocal patterns on the contact surface of the two cells, character-
ized by TCR-pMHC microclusters, encircled by LFA-1-ICAM-1
complexes, with the two types of complexes being clearly
segregated [8]. However, the microclusters fail to merge into a
cSMAC, and the desired “bull’s eye” pattern cannot be repro-
duced (Fig. 2b), meaning that there is still a force missing.
The formation of actin foci at the site of TCR-pMHC micro-
clusters may have a significant effect in T cell signaling and activa-
tion [18]. A combination of the cortical actin contraction and
membrane deformations may also play an important role in the

“bull’s eye” pattern formation [19]. Recently, it has been observed
that cortical actin forms concentric arcs in the T cell during the
contact with an APC, which move inward, toward the center of the
contact region. By contraction, the cortical actin can pull the intra-
cellular domain of TCRs with it [19], which may be a possible
mechanism for TCR-pMHC complexes to move to the center of
the contact region. Again, though, to keep the model as simple as
possible, the agents need to be endorsed with a rule to achieve the
movement toward the cSMAC, without introducing new agents
such as the cortical actin.
This new rule is added as an attractive force between TCR-
pMHC complexes with a characteristic weight, watt, and a typical
length, Latt. The protocol of the simulation is similar to the previ-
ous case. If a TCR-pMHC complex is selected, then the weighted
vectors toward all TCR-pMHC complexes within the characteris-
tic distance Latt are calculated, weighted watt > 0, and summed up.
This direction will be converted into one of the eight neighbor-
ing gridpoints, as described in Note 7. By considering only this
attractive force together with the diffusion of agents, TCR-
pMHC accumulation at the center successfully occurs, but the
cSMAC region is occupied by both TCR-pMHC and LFA-1-
ICAM-1 complexes. Furthermore, there is no clear separation
gap between the central and peripheral SMAC as is experimen-
tally observed (Fig. 2c).
The use of this agent-based model allows us to characterize the
contribution of forces to pattern formation separately, and we then
asked if a combination of all the abovementioned forces is neces-
sary to recapitulate the major properties of the IS. When all mecha-
nisms are incorporated and the forces for attraction and repulsion
are summed before choosing a direction of movement (see Note
7), the model is able to fully reproduce the experimentally observed
“bull’s eye” pattern (Fig. 2d). It further recapitulates all the
dynamical steps of the immunological synapse formation as
observed by fluorescence microscopy: Initially, microclusters form
in the contact region, travel throughout the contact surface, and
merge into bigger clusters. At around 10 min of contact (see Note
6), these clusters merge and form the cSMAC, while integrins are
forming an adhesion ring or pSMAC.
The presented agent-based formalism has the particular advan-
tage of consisting of simple physical rules, which can be easily con-
trolled, while being able to reproduce complex emerging properties
which are physiologically relevant, meaning that simple forces are
enough to explain IS formation.
Importantly, the presented methodology only includes local
rules, in contrast to [9, 10, 12, 13], and shows how to avoid pos-
sible biases regarding movement of the molecules and complexes
on lattices.
4 Notes
In the following section, we provide additional details that are

important for the construction of the agent-based simulations used
to generate the presented results. Some of the most common pit-
falls that may appear during program development are discussed,
together with possible tricks to solve them.
1. Synapse Boundaries: The lattice constant is α = 0.07 μm, which
is the closest distance between two neighboring gridpoints
(nearest neighbors). Starting from a square lattice, the grid-
points that are outside a radius R > 4.9 μm from the central
point are blocked as inaccessible by agents. In this way, a circu-
lar surface is created, representing an established contact sur-
face between the two cells.
2. Barriers: It is possible to add barriers as a new kind of gridpoint
in the structure of the lattices if needed, to represent geometrical
repatterning experiments with chromium barriers to block dif-
fusion in SLB [20]. Similar barriers may exist on APC surfaces.
Molecules are confined within the barriers and complexes on the
same lattice cannot pass over them, whereas un-ligated particles
on the opposing cell can freely pass over them.
3. Density of agents: Each gridpoint is occupied by only one
agent. The number of created agents for the simulations has to
be kept realistic. For the results shown, 100 molecules/μm2
for TCR and pMHC and 400 molecules/μm2 for LFA-1 and
ICAM-1 are used.
4. Extension: This model can be updated in order to include
more kinds of molecules at either cell, but the rules applied
there must be extended to them and selected with attention.
5. Speed optimization: To optimize the speed, it is possible to
store the agents separately in a list, which is shuffled, and at
each time step, the agents are picked in a random order and
updated. Each of the steps described, diffusion, binding kinet-
ics, and interactions between complexes, is tested once per
time step per agent.
6. Parameter values: All the parameters used in the simulations are
taken from [15], in accordance with experimental values. Based
on these parameter values, the model achieves a proper IS for-
mation with the right timing as published in the literature. The
estimated time step for the simulations presented here is given
by the ratio of the lattice constant squared, a2, to the diffusion
coefficient of a free molecule, Dm, which is t=0.012 s.
7. Movement discretization: There are several options to convert
the unit vector direction of a force into the movement to a
gridpoint, which bears the risk of introducing unphysiological
biases into the simulation:
Fig. 3 Different approaches to define the movement of complexes based on their

interactions. (a) Utilizing the rounding of the final vector. The effect of the square
lattice is visible. (b) Case where the chance of going diagonally is reduced, utiliz-
ing the same factor as done for the probability of diagonal move. There is no
improvement in comparison to (a)
(a) As a first strategy, a rounding scheme can be used: If the

final vector points from (x, y) to a position (x + δ, y + ε),
the two coordinates are rounded to the closest integer
leading to δ , ε = ± 1 or δ , ε = 0. The resulting synapse
formation is affected by having a square lattice, leading to
a square “bull’s eye” pattern (Fig. 3a).
(b) Another technique is to use the factor 1 / 2 , as was done
for the probabilities to move diagonally by diffusion, but
now for the angle θ of the force. This technique allows
choosing a position with a higher chance of moving verti-
cally or horizontally compared to diagonally. The resulting
synapse is again affected by the square lattice as in the pre-
vious case (Fig. 3b).
(c) These two methods are deterministic, meaning that if an
agent followed the same force for several time steps, it
would move to the same of the eight possible neighbors
each time, which would make it follow a straight move-
ment into a direction imposed by the lattice rather than by
the direction of the force vector. In order to avoid this
problem, a new function is introduced, which decides
where an agent will move in a probabilistic manner. This
function gives the probability to move from (x, y) to
potential new positions (x + δ, y + ε), where δ = − 1 , 0 ,
1 and ε = − 1 , 0 , 1. For example, if the forces that act on
a complex in position (x, y) point to a position between 0
and 45 degrees, there are two options for the agent to
move. Either it will move to position (x + 1, y) or to posi-
tion (x + 1, y + 1). Thus, the probability of δ = 1 is P(δ =
1) = 1. On the other hand, ε can either be ε = 0 or ε = 1.
The way to decide between the two options is to project
the force vector to the base vectors formed by the two

possible movement vectors. It gives the ratio on how
many times they should be used in order to follow the
direction of the force, in average. Here, it leads to
sin(q )
P (e = 1) = = tan(q ) . In accordance, if, for example,
cos(q )
the vector of the forces points into the interval 45 ° ≤ θ ≤
cos(q ) 1
90°, then P(ε = 1) = 1 and P (d = 1) = - =- .
sin(q ) tan(q )
The formula giving the probabilities for δ and ε is depicted
in Fig. 4a, b and is derived from the tangent and the cotan-
gent of the angle θ, which is the angle of the force vector
with respect to the x-axis. Finally, as the diagonal move-
ments cover more distance, and because the ratio between
diagonal and straight moves has to be kept, the probability
to move by forces is reduced from 1 to 1/L where L is the
length of the “average movement vector” generated by
this method. It is following the unit vector force but its
length depends on the angle, Fig. 4c.
a Probability for x variable b Probability for y variable

1 1
P(δ=-1) P(δ=1) P(δ=-1) P(ε=-1) P(ε=1)
-Π - 3Π -Π -Π 0 Π Π 3Π Π -Π - 3Π -Π -Π 0 Π Π 3Π Π
4 2 4 4 2 4 4 2 4 4 2 4
c Length of average movement
2
-Π - 3Π -Π -Π 3Π
0 Π Π Π
4 2 4 4 2 4
Fig. 4 The probability functions for δ and ε variables. The decision between picking a value for δ and ε to be
either zero or one (a) depends on the functions ±1/tan(θ) for the δ variable or (b) on the functions ± tan(θ) for
the ε variable. (c) The length of the average movement vector, L. This algorithm was used to generate the
results in Fig. 2
Acknowledgments
We thank M. Dustin for editing the manuscript. This work was sup-
ported by the Human Frontier Science Program (RGP0033/2015).
References
1. Grakoui A, Bromley SK, Sumen C, Davis 12. Figge MT, Meyer-Hermann M (2006)
MM, Andrey SS, Allen PM, Dustin ML Geometrically repatterned immunological syn-
(1999) The immunological synapse: a molec- apses uncover formation mechanisms. PLoS
ular machine controlling T cell activation. Comput Biol 2(11):e171
Science 285(5425):221–227 13. Figge MT, Meyer-Hermann M (2009)
2. Alarcón B, Mestre D, Martínez-Martín N Modeling receptor-ligand binding kinetics in
(2011) The immunological synapse: a cause or immunological synapse formation. Eur Phys
consequence of T-cell receptor triggering? J D 51(1):153–160
Immunology 133(4):420–425 14. Tsourkas PK, Baumgarth N, Simon SI,
3. Tabdanov E, Gondarenko S, Kumari S, Liapis Raychaudhuri S (2007) Mechanisms of B-cell
A, Dustin ML, Sheetz MP, Kam LC, Iskratsch synapse formation predicted by Monte Carlo
T (2015) Micropatterning of TCR and LFA-1 simulation. Biophys J 92(12):4196–4208
ligands reveals complementary effects on cyto-
15. Tsourkas PK, Raychaudhuri S (2010)
skeleton mechanics in T cells. Integr Biol Modeling of B cell synapse formation by
7(10):1272–1284 Monte Carlo simulation shows that directed
4. Alvaro Ortega-Carrion A, Vicente-Manzanares transport of receptor molecules is a potential
M (2016) Concerning immune synapses: a formation mechanism. Cell Mol Biol
spatiotemporal timeline. F1000Res 310(5751): 3(3):256–268
1191–1193 16. Sage PT, Varghese LM, Martinelli R, Sciuto
5. Qi SY, Groves JT, Chakraborty AK (2001) TE, Kamei M, Dvorak AM, Springer TA,
Synaptic pattern formation during cellular recog- Sharpe AH, Carman CV (2012) Antigen rec-
nition. Proc Natl Acad Sci 98(12):6548–6553 ognition is facilitated by invadosome-like pro-
6. Lee S-JE, Hori Y, Groves JT, Dustin ML, trusions formed by memory/effector T cells.
Chakraborty AK (2002) The synapse assembly J Immunol 188(8):3686–3699
model. Trends Immunol 23(10):500–502 17. Bromley SK, Burack WR, Johnson KG,
7. Hori Y, Raychaudhuri S, Chakraborty AK Somersalo K, Sims TN, Sumen C, Davis MM,
(2002) Analysis of pattern formation and phase Shaw AS, Allen PM, Dustin ML (2001) The
separation in the immunological synapse. immunological synapse. Annu Rev Immunol
J Chem Phys 117(20):9491–9501 19(1):375–396
8. Burroughs NJ, Wülffng C (2002) Differential 18. Kumari S, Depoil D, Martinelli R, Judokusumo
segregation in a cell-cell contact interface: the E, Carmona G, Gertler FB, Kam LC, Carman
dynamics of the immunological synapse. CV, Burkhardt JK, Irvine DJ, Dustin ML
Biophys J 83(4):1784–1796 (2015) Actin foci facilitate activation of the
9. Weikl TR, Groves JT, Lipowsky R (2002) phospholipase C-γ in primary T lymphocytes
Pattern formation during adhesion of multi- via the wasp pathway. Elife 4:e04953
component membranes. Europhys Lett 59(6): 19. Yi J, Wu XS, Crites T, Hammer JA (2012)
916–922 Actin retrograde flow and actomyosin II arc
10. Weikl TR, Lipowsky R (2004) Pattern forma- contraction drive receptor cluster dynamics at
tion during T-cell adhesion. Biophys the immunological synapse is Jurkat T cells.
J 87(6):3665–3678 Mol Biol Cell 23(5):834–852
11. Carlson A, Mahadevan L (2015) 20. Mossman KD, Campi D, Groves JT, Dustin
Elastohydrodynamics and kinetics of protein ML (2005) Altered TCR signaling from geo-
patterning in the immunological synapse. PLoS metrically repatterned immunological synapses.
Comput Biol 11(12):e1004481 Science 310(5751):1191–1193
Chapter 13
Super-resolution Analysis of TCR-Dependent Signaling:

Single-Molecule Localization Microscopy
Valarie A. Barr, Jason Yi, and Lawrence E. Samelson
Abstract
Single-molecule localization microscopy (SMLM) comprises methods that produce super-resolution
images from molecular locations of single molecules. These techniques mathematically determine the cen-
ter of a diffraction-limited spot produced by a fluorescent molecule, which represents the most likely loca-
tion of the molecule. Only a small cohort of well-separated molecules is visualized in a single image, and
then many images are obtained from a single sample. The localizations from all the images are combined
to produce a super-resolution picture of the sample. Here we describe the application of two methods,
photoactivation localization microscopy (PALM) and direct stochastic optical reconstruction microscopy
(dSTORM), to the study of signaling microclusters in T cells.
Key words Single-molecule localization microscopy, Super-resolution microscopy, Photoactivation

localization microscopy, Direct stochastic optical resolution microscopy, T cell, Microclusters
1 Introduction
In many biological systems, ligand binding to specific cell surface

receptors initiates the complex process of signal transduction.
Often signal propagation within cells is characterized by the forma-
tion of complexes composed of many interacting proteins that pro-
mote and regulate the distal signaling events that then lead to
specific cellular responses and functional outcomes. In T cells,
engagement of the T cell receptor (TCR) by a cognate antigen
displayed by an antigen-presenting cell (APC) causes the rapid for-
mation of cell surface microclusters, which contain the macromo-
lecular signaling complexes consisting of enzymes and adapter
molecules required to initiate T cell activation [1–4]. Once activa-
tion begins, the T cell undergoes a number of dramatic changes.
Actin polymerization leads to large-scale morphological changes as
the T cell spreads and forms contacts with the APC [5]. The micro-
clusters themselves move and rearrange, eventually leading in some
cases to the formation of a larger complex structure, the
183
184 Valarie A. Barr et al.
immunological synapse (IS), at the contact surface between the T cell and
APC [6]. The IS contains a central region containing concentrated
TCR and other signaling proteins surrounded by an integrin-rich
ring, which in turn can be surrounded by very large glycoproteins.
The exact function of the IS is complex and includes both enhance-
ment and downregulation of signaling [7].
The microclusters involved in T cell activation have been
examined using light microscopy [8, 9]. However, the resolution
of conventional optical microscopes is limited by diffraction of the
light coming through the lens that recombines to form a magnified
image. Light from a point source appears as a large blurred spot in
an image so it is impossible to see molecular details using standard
optical instruments. However, in the last few decades, super-
resolution techniques that allow visualization of much smaller
detail have been developed, many of which are now available in
commercial systems [10–13]. A number of methods, collectively
referred to as single-molecule localization microscopy (SMLM),
allow visualization of single molecules and hold the promise of
defining the structure and heterogeneity of signaling complexes
[14, 15]. These techniques share a common strategy of imaging a
limited number of isolated fluorescent molecules and then using
mathematical techniques to determine the location of the fluoro-
phore (Fig. 1). The key is to visualize a small number of
Acquisition Photobleaching and

decoding
Diffraction limited Image Calculated localizations

Combine frames
X
X
X
X Rendered image
X
First round
X
X
X
X
X
Second round
Repeat many times
Fig. 1 Principle of single-molecule localization microscopy. A small number of fluorescent molecules are
imaged in each frame. They must be well dispersed so that the diffraction-limited spots do not overlap. Then,
the molecules are photobleached or moved to a permanent dark state. The center of each spot is calculated.
This process is repeated thousands of times to build up an image containing the localization of thousands of
molecules. Finally, the localizations are combined and displayed in a single super-resolution image
Single Molecule Localization Microscopy 185
ell-separated fluorescent molecules so that only one molecule is

w
present in a diffraction-limited spot. An image of these well-
separated spots is captured, and then the imaged molecules are
photoswitched or photobleached. The process can be repeated
many times to visualize thousands of single molecules in a sample.
Mathematical methods are then used to determine the center of
each diffraction-limited spot corresponding to the likely position
of a single molecule, thus building up an image composed of
molecular peaks or localizations. The first of these techniques,
named photoactivation localization microscopy (PALM) [16],
uses illumination with activating light to induce a photoactivatable
protein to become fluorescent or to change the emission proper-
ties of a photoswitchable protein. Modulating the strength of the
activating beam controls the number of protein molecules that are
capable of fluorescing. Another method, stochastic optical recon-
struction microscopy (STORM) [17], takes advantage of energy
transfer between cyanine dye molecules to produce an activated
state. A variation of this technique, direct stochastic optical resolu-
tion microscopy (dSTORM) [18–20], uses a single dye molecule
that can cycle between a dark and fluorescent state. There are now
many methods that use different strategies to produce a limited
number of fluorophores [21], including ground-state depletion
with single-molecule return (GSDIM) [22], PALM with indepen-
dently running acquisition (PALMIRA) [23], and point accumula-
tion imaging in nanoscale topography (PAINT) [24, 25], but the
underlying principle is the same. The final image contains a num-
ber of points representing the calculated position of each resolved
molecule, generally presented in a manner in which the size of a
point in the rendered image contains information on the precision
of each localization. The localization precision then limits the reso-
lution of the image.
SMLM has been applied to a large number of biological sam-
ples, including cells involved in immune responses [26]. IgE-FcεRI
complexes have been examined in mast cells [27] as well as recep-
tors in B cells [28], but most immune cell studies have focused on
the organization of signaling molecules and complexes in T cells
[29–33]. Because of topological constraints and the difficulty of
imaging the surface between two cells, SMLM imaging has been
performed using various model systems in which the activating
surface can be visualized in a single plane. These investigations
have revealed nanoscale organization of the TCR and other impor-
tant components of the signal transduction pathway. This chapter
will focus mainly on using PALM and dSTORM with Jurkat T
cells, a cell line derived from human T lymphocytes. We will also
touch briefly on applying PALM to human peripheral blood lym-
phocytes. In these experiments, the T cells were activated by con-
tact with an antibody-coated cover glass, thus producing signaling
complexes in a single plane near the coverslip. Our studies with
these techniques have focused on investigating complexes contain-

ing two essential adaptor proteins, linker for activation of T cells
(LAT) and the SH2 domain-containing protein of 76 kDa (SLP-
76), which will be presented as examples in this chapter. A number
of factors must be considered before embarking on any SMLM
experiment. In addition to describing our experimental methods,
we have tried to incorporate information on the decisions that
must be made when applying SMLM to a biological problem.
2 Materials
We assume that readers are familiar with general cell culture tech-
niques and provide only a very brief description of the reagents
used to culture Jurkat T cells.
2.1 Reagents 1. Culturing of Jurkat T cells and imaging buffer: Cells are grown
for Culturing, Plating, in culture flasks in RPMI medium 1640 with 10% fetal calf
and Fixing Cells serum and penicillin/streptomycin. Stably transfected cell lines
are maintained in the same medium with the addition of 1.3
mg/ml geneticin sulfate G418 (KSE Scientific). Jurkat T cells
are available through ATCC.
2. Plating cells for microscopy: Lab-Tek II Chambered Cover
Glasses, with four or eight wells (Nalge Nunc International),
are cleaned with acidic ethanol made by adding 5 ml of 10 M
HCl to 45 ml 70% ethanol, made from 100% ethanol. Cleaned
chambers are coated with 0.01% polylysine made by diluting
0.1% poly-l-lysine with purified distilled water. The chambers
are then coated with either murine IgG1 antibody against
CD3ε, clone UCHT1, the equivalent IgG2a clone Hit3a, or, a
non-activating control reagent, murine IgG1 antibody against
CD45 (clone HI30) in PBS (pH 7.4). Human PBLs are acti-
vated with a combination of anti-CD3ε and murine IgG1 anti-
CD28 antibodies (clone CD28.2). PBS is used to rinse the
chambers. Cells are resuspended in RPMI without phenol red
containing 10% FCS and 20 mM HEPES pH 7.0 before
plating.
3. Fixation of samples: 4% paraformaldehyde is made by dissolv-
ing paraformaldehyde powder in warm (70 °C) PBS (pH 7.4).
2.2 Equipment 1. SMLM can be performed with commercial imaging systems

and Reagents Needed that are equipped to detect single molecules (see Note 1 and
for All SMLM Fig. 2). Most systems use an electron-multiplying charged-
Techniques coupled device (EMCCD) camera with high-speed image cap-
ture for detection. The system should be equipped with high
numerical aperture (NA) objectives and high-quality filters and
dichroic mirrors. High-power lasers are needed, particularly
a) Light path for PALM Imaging
TIRF Objective
Dichroic Mirror
Dichroic Mirror Excitation laser
Microscope Activation laser

Body Mirror
Emission filter
Detector
b) Light path for TIRF Objective

Sample Evanescent
Field
Emitted
Light
Excitation
Beam
Fig. 2 Light paths used in SMLM. (a) For PALM or STORM imaging, the microscope must have two lasers, one
for activation and one for imaging. For dSTORM imaging, only a single high-power laser is needed. The detec-
tor must be capable of detecting single molecules; usually an EMCCD camera is used. Most SMLM is per-
formed with TIRF excitation to limit excitation to a single z section near the coverslip. (b) Most TIRF systems
use an objective that brings the excitation light through the objective so that it reaches the sample at the criti-
cal angle needed for TIRF
for dSTORM imaging. If possible, 70–125 mW are recom-

mended for laser lines to be used in dSTORM imaging. The
computer must be capable of acquiring and storing a large
number of frames. Most SMLM is performed with a total
internal reflection (TIRF) microscope to improve z resolution
and to reduce out-of-focus light. Additional hardware can be
added to gain resolution in three dimensions; most commonly
an extra lens is added to deform the point spread function in z,
thus gaining information on z positioning. Interferometer-
based systems can also be used to determine z position [34];
however, this feature is not yet commercially available. Of par-
ticular note, Zeiss has licensed the use of PALM, while Nikon
holds the patents for STORM imaging. We use a Nikon Eclipse
Ti inverted m icroscope, AOTF modulated LUNB solid-state
lasers (70 mW at 488 nm, 70 mW at 561 nm), and a 60× SR

Apochromat TIRF lens with an Andor iXon DU888 EMCCD
camera (1024 × 1024 pixels, 8 μm pixel) for PALM imaging.
For dSTORM imaging, we use the 647 nm line from the same
system (125 mW at 647 nm), a 100× SR Apochromat TIRF
objective lens (1.49 NA) with an Andor iXon Ultra 897
EMCCD camera (512 × 512 pixels, 16 μm pixel), and the 1.5
magnifier lens in place to increase the pixel resolution which
leads to better fitting of the localization peaks.
2. Several kinds of fiducial markers can be used in SMLM.
50–100 nm gold beads (microspheres-nanospheres) are popular;
however, some researchers use small multiwavelength fluorescent
beads such as 100 nm TetraSpeck beads (Invitrogen). We are
exploring the use of 100 nm negatively charged, nitrogen-
vacancy-center nanodiamonds (Adamas Nanotechnologies) in
SMLM.
3. Software is needed to calculate the super-resolved molecular
positions from the raw diffraction-limited images. Both Zeiss
and Nikon provide complete software packages as part of their
specialized SMLM systems. We use PeakSelector version 5 to
identify individual molecules in PALM images and the Fiji-/
ImageJ-based program ThunderSTORM to identify individual
molecules in dSTORM images. However, there are a large
number of commercial and free programs available. A Google
search for SMLM software yielded over two million entries
(also see a list of programs at http://bigwww.epfl.ch/smlm/
software/). A comparison of these methods is beyond the
scope of this chapter, but there are many publications explain-
ing the various algorithms [35–37], and a recent study com-
pared over 30 packages using metrics such as detection rate,
accuracy, and resolution to help users choose which is the best
for their purposes [38].
4. Statistical techniques can be used to further study the distribu-
tion of the localized molecules. Often, the images are analyzed
using spatial statistical tools such as second-order statistics,
point pattern sorting, or nearest neighbor clustering algo-
rithms. We use a published algorithm by Wiegand and Moloney
for second-order statistics [39] and customized MATLAB code
for the nearest neighbor analysis of clustering [30]. Minkowski
functionals can also be used to sort the point patterns generated
by LAT PALM imaging [40]. However, there are many other
approaches to these problems with a wide range of analysis
options including open-source code, stand-alone freeware, and
many commercial variations [26] (see Note 2).
2.3 Reagents 1. PALM imaging requires the use of proteins that either change
for PALM Imaging from dark to fluorescent state (photoactivatable) or change
emission (photoswitchable) in response to illumination by acti-
vating light. These genetically encoded tags must then be con-

jugated to a protein of interest, and their corresponding
cDNAs are usually expressed by a strong, constitutive pro-
moter. We use three photoactivatable proteins in our PALM
studies: Dronpa (ex 488 nm/em 505 nm) (MBL International
Corporation), PA-mCherry (ex 561 nm/em 565 nm) [41],
and PA-GFP (ex 488 nm/em 505 nm) [42]. The constructs
used in published studies [30] were generated in EGFP-N1 or
EGFP-C1 vectors (Clonetech) that contain a CMV promoter.
2. Transfection and expression of proteins. Any standard transfec-
tion system can be used that allows reasonable expression in
the cell line being studied. We generally transfect E6.1 Jurkat
T cells using a Nucleofector shuttle system and the Amaxa T
kit (Lonza). The same system is used to transfect human PBLs.
3. Production of stable cell lines. We make stable cell lines express-
ing photoactivatable constructs by single-cell cloning with
antibiotic selection. The EGFP-N1 and C1 vectors we use
contain a neomycin-resistance cassette allowing the use of
geneticin (G418) as the selection agent.
4. A system capable of sterile cell sorting is usually needed to
enrich samples for PALM imaging of transiently transfected
cells and to sort cells for two-color imaging. We use a MoFlo
Astrios EQ (Beckman Coulter Life Sciences). It is necessary to
activate PA-mCherry and PA-GFP prior to sorting. We use
400 nm illumination light, in our case using a light-emitting
diode (CoolLED) for photoactivation. We find it useful to
have stable cell lines available as standards that can be used to
set the fluorescence levels in the gates used to collect samples.
2.4 Reagents Reagent grade chemicals may be obtained from Sigma-Aldrich

for dSTORM Imaging unless otherwise noted.
1. Labeling of primary antibodies. Alexa647 dye (or other dyes
like Cy5 may be used with individually optimized conditions)
is conjugated to antibodies through lysine residues using suc-
cinimidyl esters of the dyes.
2. Staining of dSTORM samples. Samples are permeabilized with
0.1% Triton-X diluted in water. 1% fish-scale gelatin in PBS is
used as a blocking buffer. Samples are rinsed and stored in
PBS.
3. dSTORM buffers. The dSTORM imaging buffer requires a
GLOX solution stock of 14 mg glucose oxide and 50 μl 17
mg/ml catalase in 200 μl 10 mM Tris–HCl (pH 8.0)/50 mM
NaCl. 10 μl of GLOX solution is added to 960 μl 50 mM Tris–
HCl (pH 8.0)/10 mM NaCl/10% glucose (method B from
the Nikon N-STORM protocol) with the addition of 100 mM
2-mercaptoethanol, 20 mM cysteamine, and 2 mM cycloocta-
tetraene (see Note 3).
3 Methods
3.1 Cleaning 1. Chambers are incubated with 0.5 ml (4-well chambers) or

and Coating Chambers 0.25 ml (8-well chambers) acidic ethanol for 15 min followed
by the removal of the solution by aspiration and drying at
45 °C for at least 30 min.
2. The chambers are then incubated with the same volume of
0.01% polylysine for 15 min followed by the removal of the
solution by aspiration and drying at 45 °C for at least 30 min.
Gold beads are added to cleaned chambers before polylysine
(see Note 4). 100 nm gold beads are sonicated and diluted in
methanol and then plated into chambers. The chambers are
dried and then coated with polylysine as above. Nanodiamond
fiducial markers are added to the clean chambers after polyly-
sine coating, but before antibody coating. Fiducials should be
added at a high enough concentration to insure that at least
five fiducial markers are present in each image.
3. The polylysine-coated chambers are then coated with either
stimulatory anti-CD3ε antibodies or non-stimulatory anti-
CD45 antibodies at 10 μg/ml using 0.4 ml for 4-well cham-
bers and 0.2 ml for 8-well chamber with an overnight
incubation at 10 °C. Chambers for activating human PBLs are
coated with a combination of 10 μg/ml anti-CD3ε antibodies
and 10 μg/ml anti-CD28 antibodies. After the antibody solu-
tion is removed, the chambers are rinsed three times in PBS
and stored in PBS until used (see Note 5).
3.2 Plating 1. Stable Jurkat lines expressing photoactivatable proteins for

and Fixing Cells PALM or untransfected cells for dSTORM are typically pas-
saged the day before use so that the cells are in mid-log phase
on the day of the experiment.
2. Sorted cells transiently transfected with photoactivatable pro-
teins are usually used 24 h after sorting. Imaging buffer is
equilibrated at 37 °C in a humidified 5% CO2 tissue culture
incubator, and the antibody-coated chambers are also warmed.
3. The chambers are rinsed once with imaging buffer and 0.3 ml
(4-well chambers) or 0.15 ml (8-well chambers) of imaging
buffer is placed in each well. The cells are resuspended in
imaging buffer at these concentrations: 2.5 × 106 cells/ml for
cells to be fixed after 2.5 min of incubation or 4 × 106 cells/ml
for cells to be fixed within 2–2.5 min (see Note 6). If possible,
no fewer than 1 × 105 cells should be plated in each chamber.
100 μl (4-well chambers) of the resuspended cells is plated by
placing the pipette tip on the bottom of the chamber and gen-
tly pipetting onto the bottom of the chamber (Fig. 3). For
8-well chambers, 50 μl of resuspended cells is plated into each
chamber.
a) Plating Jurkat T cells in Chambered coverglass
b) Jurkat cells spreading on activating coverslip
Fig. 3 Preparing activated Jurkat T cells for SLML imaging. (a) Cells are plated
into a prepared chamber by careful pipetting. (b) Jurkat T cells contacting and
spreading on an antibody-coated coverslip. Fiducial markers are shown in red
under the antibody coating
4. The chamber is placed in a 37 °C humidified 5% CO2 tissue

culture incubator for the desired time (see Note 7). At the
appropriate time, cells are fixed by gently adding 600 μl (4-well
chambers) or 300 μl (8-well chambers) of 4% paraformalde-
hyde to the chamber.
5. The cells should be incubated in fixative for 30 min at 37 °C.
The fixative is removed and the chamber is rinsed three times
with PBS. The fixed cells should be imaged as soon as possible
(see Note 8).
6. If multiwavelength fluorescent beads are used as fiducial markers,
they should be added just before imaging, allowing sufficient
time for the beads to settle in the chamber before imaging.
3.3 PALM Imaging 1. Because we already have plasmids containing many of the sig-
naling proteins found in microclusters conjugated to fluorescent
proteins, we usually produce PALM reagents by replacing the
fluorescent protein tag with a photoactivatable version using
standard methods. Generally, the Age1 and BsrG1 sites are
used to exchange the fluorescent proteins [30]. This strategy
was used to produce LAT-Dronpa and SLP-76-PA-mCherry
from versions expressing yellow fluorescent protein. We have
always used probes conjugated to full-length proteins, but
shortened versions containing the domains of interest could
also be used (see Note 9).
2. Cells can be transfected with any method that will allow imag-
ing of the cells. For Jurkat T cells and human PBLs, we trans-
fect cells with the LONZA Nucleofector shuttle system,
program H-10, and the Amaxa T kit. Either transiently trans-
fected cells or stable cell lines can be used for imaging (see
Note 10). It is important to have a very high percentage of
transfected cells in the sample.
3. Generally, the transfection efficiency in lymphocytes is low, so
we routinely sort transiently transfected cells to generate sam-
ples containing mainly transfected cells. Dronpa has basal fluo-
rescence so it can be sorted using conditions for GFP
fluorescence (488 nm excitation, 500–520 emission). PA-
mCherry and PA-GFP must be activated before sorting. We
illuminate the cells with a 400 nm light-emitting diode (LED)
source (CoolLED) for 10 min before sorting. Windows can be
set using Dronpa or GFP stable cell lines for the green window
and PA-mCherry or Cherry for the red window to select an
appropriate expression level. Transiently transfected cells are
sorted at 24 h, allowed to recover overnight and then plated
and fixed 48 h after transfection.
4. Alternatively, single-cell cloning can be used to produce stable
cell lines that do not require sorting. We usually produce cell
lines by single-cell cloning in the presence of 1.3 mg/ml G418.
This method was used to produce a stable cell line expressing
LAT-Dronpa as well as a TCR-PA-mCherry cell line that was
used to set the sorting gates for all other cells expressing
PA-mCherry.
5. The expression level of the protein of interest should be deter-
mined in sorted cells or cell lines using Western blotting or a
comparable method. We generally image cells expressing about
two times endogenous levels of the protein of interest.
6. Multiplexed PALM imaging can also be performed in cells
expressing two photoactivatable proteins either by transfecting
cells with two plasmids or super-transfecting a stable cell line
with a second protein. In both cases, it is usually necessary to
sort the cells to obtain a sufficient number of cells expressing
both photoactivatable proteins (see Note 11).
7. The activation protocol depends on the photoactivatable pro-
tein. Dronpa can be easily activated by low-power 405 nm
light or 360 nm light. We use simultaneous illumination with
a DAPI cube (excitation 340–380 nm) and arc lamp for
Dronpa. PA-mCherry and PA-GFP require stronger illumina-
tion to photoactivate. We use 10–20 s of 405 light or maximal
intensity of the arc lamp with a CFP cube (excitation 426–446
nm) for PA-mCherry and 60 s to activate PA-GFP.
8. All imaging is performed in TIRF mode. Since Dronpa can be
activated and imaged at the same time, a large number of
frames can be collected in each sequence. We generally collect

2500 frames/sequence and at least two sequences per sample
using continuous activation illumination and simultaneous
imaging with an optical configuration suitable for visualizing
GFP (excitation 488 nm, emission 505–525 nm) (Fig. 4a).
PA-mCherry must be activated and then imaged separately
with an optical configuration suitable for mCherry visualiza-
tion (excitation 561 nm, emission 580–630 nm). Often, no
fluorescence will be visible after 500–700 frames. We typically
activate and image 500 frames/sequence and collect 5
sequences/sample although often the pool of excitable mole-
cules will be exhausted before all the sequences are captured.
PA-GFP is imaged after activation using the same optical con-
figuration as Dronpa. In our hands, PA-GFP fluoresces for
1000–1500 frames after activation. We generally collect 1250
frames/sequence and 4 sequences/sample (see Note 12).
9. For two-color PALM imaging, we use a Dronpa-conjugated
molecule such as LAT-Dronpa paired with a PA-mCherry-
conjugated molecule such as SLP-76-PA-mCherry (Fig. 4b).
The Dronpa construct is always imaged first as described
above. Once all the Dronpa images are collected, the sample is
illuminated with a stronger activating light, and then the
PA-mCherry images are captured. The photoactivation and
imaging of PA-mCherry can be repeated several times.
3.4 dSTORM Imaging 1. Primary antibodies should be directly conjugated to the dye of
interest (see Note 13). We conjugate antibodies with Alexa647
with a standard kit using the manufacturer’s recommended
protocol (Molecular Probes/Invitrogen).
2. Cell staining. Fixed cells are permeabilized for 5 min with
0.1% Triton-X, incubated in blocking buffer for 30 min and
then with the Alexa647-conjugated primary antibody (50
ng/ml) at room temperature for 1 h (see Note 14). Stained
samples are washed five times with PBS and stored in PBS at
4 °C in the dark. The dilution of the primary antibody
depends on the avidity of the particular antibody; we usually
use 50 ng/ml. Induced blinking of Alexa647 requires a reac-
tion with a thiol-containing compound so after five washes
with PBS, dSTORM imaging buffer is added in excess to the
samples, that is, at 2 ml/4-well chamber and 1 ml/8-well
chamber. Samples are sealed with a glass coverslip to protect
them from air.
3. Imaging sequence. The sample is illuminated with a high-
power laser to induce dye blinking. The cell sample is imaged
for 20,000–30,000 frames in dSTORM imaging buffer
(Fig. 4c).
3.5 Analysis 1. Determining the molecular localizations from the raw

of SMLM Data diffraction-limited images is a complicated process. Most
SMLM software packages process the raw images, correct for
stage drift, identify intensity peaks representing candidate mol-
ecules, determine the position of the molecules producing the
peaks, and finally create a rendered image of the localizations.
Each package has its own algorithms and required inputs for
performing these steps, so rather than explain how to use a
particular one, we will highlight some of the considerations for
these steps. However, all choices may not be available for every
step in every package.
2. Preprocessing steps. Background subtraction is strongly
advised, followed by noise reduction to enhance the detection
of local maxima corresponding to emissions from candidate
molecules while reducing false positives. SLML software pack-
ages generally include a number of standard techniques, such
as average, median, erosion, and Gaussian filter masks, which
can be used to reduce the noise in the image (see Note 15).
SMLM imaging sequences generally last for many minutes and
stage drift will affect the images. Drift correction can be per-
formed either by tracking fiducial markers present on each
frame of the imaging sequence [16, 17] or with cross-correla-
tion algorithms that use the sample structure to correct for
movement during the image sequence [43] (see Note 16). We
prefer fiducial markers for correcting drift. If more than one
color is used in SMLM, chromatic aberration will be a problem
[44]. In this case, fiducial markers can be used to align chan-
nels and to monitor the accuracy of the correction.
3. The user has to set a number of parameters, including thresh-
olds for peaks (local or global). The user may also need to set
the criteria for rejecting peaks using measures such as the
amplitude of the fitted function, symmetry, or skewness. Some
thought should be given to these parameters as they will
directly influence the quality of the SMLM data.
4. Most programs report an estimated localization precision or σ
that shows the error in the calculated position of each mole-
cule. This information can be used to judge the quality of both
the original images and the SMLM localizations. The most
important factors are the number of photons and noise (see
Note 17).
5. In both PALM and dSTORM, some fluorophores move to a
long-lived dark state instead of photobleaching. These mole-
cules can then blink or return to fluorescent state. Thus, a sin-
gle molecule could be counted multiple times in the final
SMLM image [45, 46]. This is easier to correct in PALM
imaging as the amount of blinking is lower. We collect localiza-
tions with a spatial radius 50% larger than the sigma if they
occur within 15 frames and group them into one localization.
The number of localizations found as the time gap (number of

frames) increases fits a negative exponential function; thus, for
larger time intervals, the improvement becomes asymptotically
lower. Our grouping conditions lower the number of detected
molecules to 115% of the asymptotic plateau. Other methods
such as scanning for temporal clusters can also be used to
account for blinking [45]; however, grouping errors are com-
mon in SMLM and bedevil all of the analyses (see Note 18).
6. Most SMLM software also contains algorithms to translate the
list of positions of localized molecules into an image. The final
image is a point pattern formed by placing filled spots at the
location of each molecule. Most simply, each localization can
be drawn as a single dot of a set size, but generally the spot
size, intensity, or color is used to convey information on error
in the calculated location of that particular molecule. In the
most common representation, each spot in the final image is
formed by Gaussian functions centered at the molecular posi-
tion with an intensity that represents the number of photons
and a width that depends on the localization precision. Smaller,
brighter spots represent better localizations. In this representa-
tion, the edges of molecules can overlap. Alternatively, images
can be made where the intensity corresponds to the probability
density of the fitted Gaussian distribution. Finally, spots can be
produced by binning localizations into an appropriate 2D grid.
Binned representations can be improved by using jittered 2D
histograms where the jitter is proportional to the localization
accuracy [47]. The density of the label also influences the accu-
racy of the image (see Note 19). In our PALM images that are
produced by the PeakSelector software, the spot intensity
reflects the probability density (Fig. 4a, b). In our dSTORM
images that are produced by ThunderSTORM, the spots are
produced by Gaussian functions (Fig. 4c).
7. The distribution of points in SMLM images can be analyzed to
determine if there is a pattern. Because these images are not
pixel-based intensity maps, traditional image processing algo-
rithms are not appropriate in most cases. Many SMLM studies
have used second-order statistics that compare the points in an
image to a null or random model to determine nonrandom
distributions [29–31]. Many of these methods have been bor-
rowed from other fields that have developed methods to exam-
ine distribution of objects. Care is needed to pick the correct
method and null model [39]. We use a heterogeneous null
model to account for variations in the membrane contact
surface in our analyses of single-color PALM images (Fig. 5).
The details of this analysis have been published [30]. Very dif-
ferent results can be obtained from the same data depending
on how the analysis is performed [48]. Second-order statistics
provide an aggregate answer to whether there is clustering and
a) Single color PALM imaging:LAT-Dronpa

Raw, diffraction limited Rendered
100 nm Molecule prob/nm2

0.00025 0.00322
2.0 µm 2.0 µm Molecule prob/nm2
0.0 0.00189
b) Two color PALM imaging:LAT-Dronpa and SLP-76-PA-mCherry

Raw, diffraction limited LAT-Dronpa
Rendered:LAT-Dronpa green, SLP-PA-mCherry red
2.0 µm
Raw, diffraction limited SLP-76-PA-mCherry
Molecule prob/nm2
0.00011 0.00170
200 nm
0.00020 0.00132
Molecule prob/nm2
0.00006 0.00118
2.0 µm
0.00012 0.00189
2.0 µm
c) Single color dSTORM imaging:

TIRF Image Rendered Composite
Fig. 4 Jurkat T cells imaged by PALM and dSTORM. (a) PALM imaging of LAT-Dronpa. The left panel shows the
sum of all the diffraction-limited images (5000 frames total). The center panel is a rendering of the localiza-
tions found in the 5000 frame series. The color scale shows the probability density of the localizations; a larger
probability density means there is a greater chance of finding a localization in a given volume. Areas with a
high localization density are small and white. As the probability density decreases, the rendered spots become
larger and darker red. The white circle indicates a representative fiducial marker. In this image the fiducials
a) Rendering of ROI b) Pixel map of ROI c) Density map of ROI

20 nm density
hard shell
Molecule prob/nm2 1.0 µm

0.000 0.00146
d) Univariate pair correlation function

2.0
1.8
1.6
g(r)
1.4
1.2
1
100 200 300 400 500 600 700 800 900 1000
r [nm]
Fig. 5 Analysis of a one color PALM image. (a) Rendering of a central ROI chosen from the cell pictured in
Fig. 4a. (b) Pixelation of the localizations in the ROI using a 20 nm hard-shell model. (c) Density map of ROI
showing heterogeneity of the density of molecules in the ROI. (d) Univariate pair correlation function g (r). The
blue line shows the sample PCF while the black dotted lines show the highest and lowest PCFs from 19 Monte
Carlo simulations of a random distribution using a heterogeneous Poisson process based on the density map
shown in (c). A sample PFC lying between these lines would be considered a random distribution with no
clustering. The sample PCF lies above the highest PCF from the random simulations showing that LAT-Dronpa
is more clustered than a random distribution
Fig. 4 (continued) were 100 nm TetraSpeck beads. The right panel shows a magnification of the boxed area
from the middle panel. (b) PALM imaging of LAT-Dronpa and SLP-76-PA-mCherry. 5000 frames were obtained
of LAT-Dronpa followed by imaging 2500 frames of SLP-76-PA-mCherry. The left panels show the sum of all
the diffraction-limited images of LAT-Dronpa (top) and SLP-76-PA-mCherry (bottom). The middle panel shows
the localizations from all of these images combined into a single two-color rendering. The white circle indi-
cates a representative fiducial marker. In this image the fiducials were 100 nm TetraSpeck beads. The right
panel shows a magnification of the boxed area from the middle panel. (c) dSTORM imaging of phosphorylated
SLP-76 using antibodies to anti-phosphorylated SLP-76 directly conjugated to Alexa647. The left panel shows
a diffraction-limited TIRF image of the stained sample. The middle panel shows the localizations from a 20,000
image series. This rendering is produced using Gaussian profiles centered at the molecular positions. The
white circle indicates a representative fiducial marker. In this image the fiducials were 100 nm negatively
charged, nitrogen-vacancy-center nanodiamonds. The right panel shows the TIRF image superimposed on the
dSTORM rendering
what is the length scale of the clustering. These statistics are

not meant to determine cluster size. An alternative approach
uses algorithms that examine the proximity of any given mol-
ecule to its nearest neighbors to study cluster distribution and
size [26]. There are many variations of this approach, and the
choices of thresholds and other variables can influence the out-
come. Obviously, accurate cluster analysis also requires an
accurate correction for multiple localizations from a single
molecule. Finally, SLML techniques, as currently implemented,
rarely give accurate counts of the number of molecules,
although methods are being developed to improve counting
[49, 50]. Both overcounting and undercounting errors are
common [48].
8. Analysis of SMLM images of two or more components is even
more complicated. As mentioned earlier, if different optical
configurations are used for the different species, chromatic
aberration can introduce additional errors in the accuracy in
assigning each position. For two-component images, second-
order statistics such as the bivariate pair correlation function
(PCF) or Ripley’s K-functions can be used to compare the dis-
tributions in an image to random patterns or random mixing.
We use the bivariate pair correlation with a random sampling
null model to analyze two-color images as described earlier
(Fig. 6) [30]. Colocalization measures for pixel-based images
including Pearson’s or Mander’s are generally not suited for
point patterns. Furthermore, these techniques can only ana-
lyze pairs of molecules, so additional methods are needed for
analysis of patterns and relationships between three or more
components.
4 Notes
1. While a SMLM can be performed on commercial microscopes,

the system must be sensitive, contain high-quality optics,
employ lasers of sufficient power, and be protected from ther-
mal instability and vibration to reduce drift. References are
available that describe how to assemble a SMLM system [18,
51, 52]. If TIRF illumination is used, nonuniform illumina-
tion may cause systematic variation in the localization accuracy
as areas with fainter illumination may produce fewer photons
per molecule. In extreme cases, molecules in some parts of the
sample may not be detected. We use only a small area in the
center of the field where the illumination is most uniform for
imaging and collects no more than three T cells per field for
PALM images and only one T cell per field for dSTORM
imaging.
a) Rendering of ROI b) Pixel map of ROI c) 1 case of random mixing

LAT-Dronpa green, SLP-PA-mCherry red 20 nm
hard shell
Molecule prob/nm2
0.00144 1.0 µm
0.00147
d) Bivariate Pair Correlation

3.5
2.5
g12(r)
1.5
0.5
100 200 300 400 500 600 700 800 900 1000
r [nm]
Fig. 6 Analysis of a two-color PALM image. (a) Rendering of a central ROI chosen from the cell pictured in
Fig. 4b. (B) Pixelation of the localizations in the ROI using a 20 nm hard-shell model. LAT-Dronpa is shown in
green and SLP-76-PA-mCherry is shown in red. (c) An example of a random mixing simulation. Red and green
spots were placed randomly into any position that was marked with a spot in (b). The spot locations are the
same as in the sample, and the number of green and red spots corresponds to the numbers found in (b). (d)
Bivariate pair correlation function g12(r). The blue line shows the sample PCF, while the black dotted lines show
the highest and lowest PCFs from 19 Monte Carlo simulations of the random mixing model. A sample PFC lying
between these lines would indicate random mixing of the two molecules in the sample. The sample PCF lies
below the lowest PCF from the random simulations showing that LAT-Dronpa and SLP-76 are more segre-
gated than would be expected in a random distribution
2. These statistical methods are applicable to one- and two-color
point patterns but would require sequential pair-wise analysis
of SMLM images containing more than two components.
New methods are needed to examine the relationships between
three or more proteins.
3. There are a wide variety of dSTORM buffers in use ranging
from simple buffers and commercial mounting [53] to mix-
tures optimized for maximum photon count [54]. It is worth
performing some preliminary experiments to determine the
best buffer for a particular application before beginning a
dSTORM project.
4. The fiducial marker should be chosen before preparing the

chambers, as different fiducials are added at different times in
the protocol. Many researchers have used gold beads as stable
fiducials. For optimal adherence, they should be added to the
chamber before the polylysine coating, although gold rods
will adhere reasonably well to an antibody-coated coverslip.
However, we have found that the signal diminishes at shorter
wavelengths, so gold fiducials sometimes fail as green or mul-
ticolor markers. Fluorescent beads make good fiducials, and
because they can be added after the sample is in place, they are
less likely to overlap with localizations from the sample.
However, they are not strongly bound to the surface and can
be slightly mobile in the chamber. Beads that fluoresce at mul-
tiple wavelengths are readily available. Unfortunately, they can
be bleached when performing STORM imaging. We have
found that negatively charged nitrogen-vacancy-center nano-
diamonds are very good fiducial markers. They produce stable
fluorescence in multiple channels, and because the fluores-
cence is an integral part of the crystal structure, they do not
photobleach. Nanodiamonds must be added before the anti-
body coating as they will not adhere to the coated surface.
5. Dry polylysine-coated chambers may be stored at room tem-
perature for many weeks. Antibody-coated chambers should
be coated shortly before use although the plated antibodies
are generally active for at least 1 week.
6. The number of cells securely attached to the coverslip is lower
at early timepoints, so more of the plated cells are removed in
the washes. Increasing the number of cells helps compensate
for this effect.
7. Once the cells have been plated, the chambers should be
moved carefully to avoid dislodging the activated cells.
8. PALM samples are light sensitive. Once the cells are plated
into the chamber, the sample should be protected from light.
Storing the samples below 10 °C diminishes the activation of
photoactivatable proteins. PALM samples should be stored at
room temperature.
9. There are many choices for photoactivatable and photoswitch-
able proteins [55, 56]. For single-molecule imaging, photon
yield is one of the most significant considerations. In addition,
photostability and the spontaneous photoactivation rate
should also be considered [10]. It is important to choose fluo-
rescent proteins that do not form aggregates either by them-
selves or as conjugated molecules [57]. The photoswitchable
protein tdEos is often used because it is very bright, but
because it is a tandem dimer, care should be taken to insure
that it does not affect the localization and function of the pro-
tein being studied. For multicolor PALM, it is important to

choose suitable pairs [48].
10. The relative merits of transient transfections and stable cell
lines are generally the same for PALM imaging as other
experiments. Stable cell lines offer a more uniform expres-
sion but require a substantial time investment. In several
cases, we found additional mutations in Jurkat cell lines
selected for s table expression of signaling proteins, so all cell
lines should be carefully evaluated to be sure they behave
normally. If a single construct will be used for many PALM
experiments, the advantages of a stable cell line may out-
weigh the disadvantages.
11. When a stable cell line is super-transfected, the levels of expres-
sion of the two proteins tend to be inversely related, so it can
be difficult to find cells with acceptable levels of both photoac-
tivatable proteins. Cells that are dually transfected with both
plasmids tend to give proportional labeling of both photoacti-
vatable proteins, making it easier to locate cells that can be
imaged.
12. The exact conditions needed for imaging will depend on

expression level as well as the photoactivatable protein that is
being used. As mentioned in Note 9, there is a wide range in
the number of photons produced by different photoactivat-
able proteins. Also, as the expression level increases, the
amount of activating light may need to be decreased to avoid
activating too many molecules. Generally, the strategy in
PALM imaging is to vary the activation energy to control the
number of activated molecules so that well-spaced single mol-
ecules are imaged. To a large extent, the optimal length of the
collection series following activation is controlled by the char-
acteristic production of photons by the fluorescent protein.
Increasing the excitation laser power may speed up the imag-
ing and bleaching of the activated molecules; however, as the
excitation power increases, there is increased risk of bleaching
without capturing fluorescence. When we image PA-mCherry
or PA-GFP, by the end of our imaging series, we have visual-
ized almost all of the molecules that we can activate in that
particular sample, so increasing the number of cycles would
not produce many more localizations.
13. Using indirect immunostaining puts two immunoglobulin

molecules between the protein being studied and the fluoro-
phore being imaged by dSTORM. This increases the distance
between the detected molecule and the actual target, signifi-
cantly increasing the uncertainty in the actual location of the
protein being studied. Therefore, all dSTORM probes should
be directly conjugated to as small an entity as possible. If Fab
fragments or single-domain antibodies can be found, these are

the best choices for primary antibodies for use in dSTORM.
14. Many fluorophores can be used for dSTORM imaging.

Alexa647 and Cy5 have the optimal combination of photon
count and duty cycles [56]. Other dyes are usable, but with
decreased performance [58].
15. Noise reduction filters have been evaluated for use in

SMLM. One report preferred median smoothing [59], while
a second study recommended convolution with a Gaussian
kernel [47].
16. Drift correction is implemented differently in different soft-
ware packages. We prefer methods based on fiducial markers.
Fiducial markers are required for channel correction when
performing multicolor SMLM.
17. There are several different measures that are important in

assessing the quality of SMLM data [36, 60]. The most com-
monly reported measure is σ or localization precision which
gives the error in the calculated position. Most studies use a
standard formula to estimate the error that takes into account
the number of photons, the pixel size, and the noise in the
image [61]. Increasing the number of photons/localization
greatly increases the localization precision. dSTORM has a
great advantage in photon production as the best dyes,
Alexa647 and Cy5, produce 6000 photons/burst, while pho-
toactivatable proteins are generally in the range of 200–600
photons/burst [15, 16, 62]. The image resolution or the abil-
ity to distinguish two different points in the image can be
affected by experimental errors so a given image may not actu-
ally achieve a resolution equal to σ. Many papers on SMLM
techniques show images to demonstrate that they can resolve
a known structure such as microtubules [22, 58, 63]. In addi-
tion to these intrinsic standards, DNA origami has been used
to produce molecular rulers that could be used to verify reso-
lution [64]. A technique used in electron microscopy, Fourier
ring correlation, has been applied to SMLM to determine the
true resolution [65, 66]. However, to analyze microclusters,
most researchers are interested in accurately determining the
position of each protein in a multi-protein complex.
Unfortunately, the localization error is not an adequate mea-
sure of the error in the actual position of the molecule. If the
error is distributed normally, the probability of finding a single
molecule in a circle with a radius of σ is only 33%. That is, if
the σ is 20 nm, 66% of the molecules will be outside a circular
area with a diameter of 40 nm. A circle of probable location
with a radius of 3σ or 60 nm will contain the molecule of inter-
est 99% of the time. However, a probable location somewhere
within an area of 11,000 nm2 is not sufficient to define the
molecular structure of a signaling complex. Moreover, when a

single molecule blinks, the calculated positions of the different
blinks may not overlap [46]. The true location of the molecule
could be anywhere within the territory defined by the set of
blinks.
18. The difficulty in properly assigning multiple localizations to
the correct molecule or the grouping of localizations remains
one of the most stubborn problems in SMLM [67]. Without
this crucial correction, it is impossible to perform a detailed
molecular analysis of microclusters and the immune synapse.
19. As the spatial resolution improves, it becomes harder to pro-
duce a labeling density that meets the Nyquist density require-
ment of two measurements per resolution unit. If too many
molecules are missing because of low labeling efficiency or low
label detection, the image will be incomplete and inaccurate
[68]. Labeling density in dSTORM is limited by antibody
binding affinities and steric hindrance of antibodies when the
epitopes are close together. In the crowded milieu of a signal-
ing complex, it may be quite difficult to label every protein of
interest. PALM offers the theoretical possibility of labeling
every protein in the absence of unlabeled endogenous pro-
teins. However, because PALM reagents produce fewer pho-
tons, it may be difficult to visualize every photoactivated
protein, which will reduce the density of detected molecules
and limit the accuracy of the image.
Acknowledgment
We thank Eilon Sherman for generating the algorithms used for

our PALM analysis and continued advice on imaging methods.
This research was supported by the Intramural Research Program
of the NIH, National Cancer Institute (NCI), and Center for
Cancer Research.
References
1. Bunnell SC, Hong DI, Kardon JR, Yamazaki microclusters initiate and sustain T cell activa-
T, McGlade CJ, Barr VA et al (2002) T cell tion by recruitment of Zap70 and SLP-76. Nat
receptor ligation induces the formation of Immunol 6:1253–1262
dynamically regulated signaling assemblies. 4. Balagopalan L, Coussens NP, Sherman E,
J Cell Biol 158:1263–1275 Samelson LE, Sommers CL (2011) The LAT
2. Campi G, Varma R, Dustin ML (2005) Actin story: a tale of cooperativity, coordination, and
and agonist MHC-peptide complex-depen- choreography. Cold Spring Harb Perspect Biol
dent T cell receptor microclusters as scaf- 3:89–109
folds for signaling. J Exp Med 202: 5. Bunnell SC, Kapoor V, Trible RP, Zhang WG,
1031–1036 Samelson LE (2001) Dynamic actin polymer-
3. Yokosuka T, Sakata-Sogawa K, Kobayashi W, ization drives T cell receptor-induced spread-
Hiroshima M, Hashimoto-Tane A, Tokunaga ing: a role for the signal transduction adaptor
M et al (2005) Newly generated T cell receptor LAT. Immunity 14:315–329
6. Monks CR, Freiberg BA, Kupfer H, Sciaky N, (2011) Direct stochastic optical reconstruction
Kupfer A (1998) Three-dimensional segrega- microscopy with standard fluorescent probes.
tion of supramolecular activation clusters in T Nat Protoc 6:991–1009
cells. Nature 395:82–86 19. Heilemann M, van de Linde S, Schuttpelz M,
7. Dustin ML, Chakraborty AK, Shaw AS (2010) Kasper R, Seefeldt B, Mukherjee A et al (2008)
Understanding the structure and function of Subdiffraction-resolution fluorescence imaging
the immunological synapse. Cold Spring Harb with conventional fluorescent probes. Angew
Perspect Biol 2:a002311 Chem Int Ed Engl 47:6172–6176
8. Balagopalan L, Sherman E, Barr VA, Samelson 20. Endesfelder U, Heilemann M (2015) Direct
LE (2011) Imaging techniques for assaying stochastic optical reconstruction microscopy
lymphocyte activation in action. Nat Rev (dSTORM). In: Verveer JP (ed) Advanced flu-
Immunol 11:21–33 orescence microscopy: methods and protocols.
9. Yokosuka T, Saito T (2010) The immunologi- Springer, New York, pp 263–276
cal synapse, TCR microclusters, and T cell acti- 21. Walter NG, Huang CY, Manzo AJ, Sobhy MA
vation. In: Saito T, Batista DF (eds) (2008) Do-it-yourself guide: how to use the
Immunological synapse. Springer, Heidelberg, modern single-molecule toolkit. Nat Methods
Berlin, pp 81–107 5:475–489
10. Sengupta P, van Engelenburg SB, Lippincott- 22. Folling J, Bossi M, Bock H, Medda R, Wurm
Schwartz J (2014) Superresolution imaging of CA, Hein B et al (2008) Fluorescence nanos-
biological systems using photoactivated copy by ground-state depletion and single-
localization
microscopy. Chem Rev molecule return. Nat Methods 5:943–945
114:3189–3202 23. Egner A, Geisler C, von Middendorff C, Bock
1 1. Thorley JA, Pike J, Rappoport JZ (2014) H, Wenzel D, Medda R et al (2007)
Super-resolution microscopy: a comparison Fluorescence nanoscopy in whole cells by asyn-
of commercially available options. In: chronous localization of photoswitching emit-
Cornea A, Conn PM (eds) Fluorescence ters. Biophys J 93:3285–3290
microscopy: super-resolution and other 24. Sharonov A, Hochstrasser RM (2006) Wide-
novel techniques. Academic, New York, NY, field subdiffraction imaging by accumulated
pp 199–212 binding of diffusing probes. Proc Natl Acad Sci
12. Nienhaus K, Nienhaus GU (2016) Where do U S A 103:18911–18916
we stand with super-resolution optical micros- 25. Giannone G, Hosy E, Levet F, Constals A,
copy? J Mol Biol 428:308–322 Schulze K, Sobolevsky AI, Rosconi MP,
13. Sydor AM, Czymmek KJ, Puchner EM, Gouaux E, Tampe R, Choquet D, Cognet L
Mennella V (2015) Super-resolution micros- (2010) Dynamic superresolution imaging of
copy: from single molecules to supramolecular endogenous proteins on living cells at ultra-
assemblies. Trends Cell Biol 25:730–748 high density. Biophys J 99:1303–1310
14. Knight AE (2017) Super-resolution fluores- 26. Curthoys NM, Parent M, Mlodzianoski M,
cence microscopy, localization microscopy. Nelson AJ, Lilieholm J, Butler MB et al (2015)
In: Lindon J, Tranter GE, Koppenaal D (eds) Chapter three—dances with membranes:
Encyclopedia of spectroscopy and spectrom- breakthroughs from super-resolution imaging.
etry, 3rd edn. Reference module in chemis- In: Anne KK (ed) Current topics in mem-
try, molecular sciences and chemical branes. Academic Press, New York, pp 59–123
engineering. Academic, New York, NY, pp 27. Shelby SA, Holowka D, Baird B, Veatch SL
325–330 (2013) Distinct stages of stimulated
15. Allen JR, Ross ST, Davidson MW (2013) FcepsilonRI receptor clustering and immobili-
Single molecule localization microscopy for zation are identified through superresolution
superresolution. J Optics 15:094001 imaging. Biophys J 105:2343–2354
16. Betzig E, Patterson GH, Sougrat R, Lindwasser 28. Maity PC, Blount A, Jumaa H, Ronneberger
OW, Olenych S, Bonifacino JS et al (2006) O, Lillemeier BF, Reth M (2015) B cell anti-
Imaging intracellular fluorescent proteins at gen receptors of the IgM and IgD classes are
nanometer resolution. Science clustered in different protein islands that are
313:1642–1645 altered during B cell activation. Sci Signal
17. Rust MJ, Bates M, Zhuang X (2006) Stochastic 8:ra93
optical reconstruction microscopy (STORM) 29. Lillemeier BF, Mortelmaier MA, Forstner MB,
provides sub-diffraction-limit image resolu- Huppa JB, Groves JT, Davis MM (2010) TCR
tion. Nat Methods 3:793–795 and Lat are expressed on separate protein
18. van de Linde S, Loschberger A, Klein T, islands on T cell membranes and concatenate
Heidbreder M, Wolter S, Heilemann M et al during activation. Nat Immunol 11:90–96
30. Sherman E, Barr V, Manley S, Patterson G, beling of proteins and cells. Science
Balagopalan L, Akpan I et al (2011) Functional 297:1873–1877
nanoscale organization of signaling molecules 43. Huang B, Wang W, Bates M, Zhuang X (2008)
downstream of the T cell antigen receptor. Three-dimensional super-resolution imaging
Immunity 35:705–720 by stochastic optical reconstruction micros-
31. Purbhoo MA, Liu H, Oddos S, Owen DM, copy. Science 319:810–813
Neil MAA, Pageon SV et al (2010) Dynamics 44. Erdelyi M, Rees E, Metcalf D, Schierle GS,
of subsynaptic vesicles and surface microclus- Dudas L, Sinko J et al (2013) Correcting chro-
ters at the immunological synapse. Sci Signal matic offset in multicolor super-resolution
3:ra36 localization microscopy. Opt Express
32. Hsu C-J, Baumgart T (2011) Spatial associa- 21:10978–10988
tion of signaling proteins and F-actin effects on 45. Annibale P, Vanni S, Scarselli M, Rothlisberger
cluster assembly analyzed via photoactivation U, Radenovic A (2011) Identification of clus-
localization microscopy in T cells. PLoS One tering artifacts in photoactivated localization
6:e23586 microscopy. Nat Methods 8:527–528
33. Rossy J, Pageon SV, Davis DM, Gaus K (2013) 46. Sengupta P, Jovanovic-Talisman T, Skoko D,
Super-resolution microscopy of the immuno- Renz M, Veatch SL, Lippincott-Schwartz
logical synapse. Curr Opin Immunol J (2011) Probing protein heterogeneity in
25:307–312 the plasma membrane using PALM and pair
34. Shtengel G, Galbraith JA, Galbraith CG, correlation analysis. Nat Methods 8:
Lippincott-Schwartz J, Gillette JM, Manley S 969–975
et al (2009) Interferometric fluorescent 47. Krizek P, Raska I, Hagen GM (2011)
super-resolution microscopy resolves 3D cellu- Minimizing detection errors in single molecule
lar ultrastructure. Proc Natl Acad Sci U S A localization microscopy. Opt Express
106:3125–3130 19:3226–3235
35. Ober RJ, Ram S, Ward ES (2004) Localization 48. Sherman E, Barr VA, Samelson LE (2013)
accuracy in single-molecule microscopy. Resolving multi-molecular protein interactions
Biophys J 86:1185–1200 by photoactivated localization microscopy.
36. Abraham AV, Ram S, Chao J, Ward ES, Ober Methods 59:261–269
RJ (2009) Quantitative study of single mole- 49. Fricke F, Beaudouin J, Eils R, Heilemann M
cule location estimation techniques. Opt (2015) One, two or three? Probing the stoichi-
Express 17:23352–23373 ometry of membrane proteins by single-
37. Small A Stahlheber S (2014) Chapter 16—the molecule localization microscopy. Sci Rep
role of image analysis algorithms in super- 5:14072
resolution localization microscopy. In: 50. Lee S-H, Shin JY, Lee A, Bustamante C (2012)
Fluorescence microscopy, Academic Press, Counting single photoactivatable fluorescent
New York, pp 227–242 molecules by photoactivated localization
38. Sage D, Kirshner H, Pengo T, Stuurman N, microscopy (PALM). Proc Natl Acad Sci U S A
Min J, Manley S et al (2015) Quantitative eval- 109:17436–17441
uation of software packages for single-molecule 51. Almada P, Culley S, Henriques R (2015)
localization microscopy. Nat Methods PALM and STORM: Into large fields and
12:717–724 high-throughput microscopy with sCMOS
39. Wiegand T, Moloney KA (2004) Rings, circles, detectors. Methods 88:109–121
and null-models for point pattern analysis in 52. Holm T, Klein T, Loschberger A, Klamp T,
ecology. Oikos 104:209–229 Wiebusch G, van de Linde S et al (2014) A
40. Parker J, Sherman E, van de Raa M, van der blueprint for cost-efficient localization micros-
Meer D, Samelson LE, Losert W (2013) copy. Chemphyschem 15:651–654
Automatic sorting of point pattern sets using 53. Olivier N, Keller D, Rajan VS, Gönczy P,
Minkowski functionals. Phys Rev E Stat Nonlin Manley S (2013) Simple buffers for 3D STORM
Soft Matter Phys 88:022720 microscopy. Biomed Opt Express 4:885–899
41. Subach FV, Patterson GH, Manley S, Gillette 54. Olivier N, Keller D, Gönczy P, Manley S
JM, Lippincott-Schwartz J, Verkhusha VV (2013) Resolution doubling in 3D-STORM
(2009) Photoactivatable mCherry for high- imaging through improved buffers. PLoS One
resolution two-color fluorescence microscopy. 8:e69004
Nat Methods 6:153–159 55. Shcherbakova DM, Sengupta P, Lippincott-
42. Patterson GH, Lippincott-Schwartz J (2002) Schwartz J, Verkhusha VV (2014)
A photoactivatable GFP for selective photola- Photocontrollable fluorescent proteins for
superresolution imaging. Annu Rev Biophys 62. Ha T, Tinnefeld P (2012) Photophysics of flu-
43:303–329 orescent probes for single-molecule biophysics
56. Allen JR, Ross ST, Davidson MW (2013) and super-resolution imaging. Annu Rev Phys
Sample preparation for single molecule local- Chem 63:595–617
ization microscopy. Phys Chem Chem Phys 63. Vaughan JC, Jia S, Zhuang X (2012)
15:18771–18783 Ultrabright photoactivatable fluorophores cre-
57. Wang S, Moffitt JR, Dempsey GT, Xie XS, ated by reductive caging. Nat Methods
Zhuang X (2014) Characterization and develop- 9:1181–1184
ment of photoactivatable fluorescent proteins for 64. Schmied JJ, Gietl A, Holzmeister P, Forthmann
single-molecule–based superresolution imaging. C, Steinhauer C, Dammeyer T et al (2012)
Proc Natl Acad Sci U S A 111:8452–8457 Fluorescence and super-resolution standards
58. Dempsey GT, Vaughan JC, Chen KH, Bates based on DNA origami. Nat Methods
M, Zhuang X (2011) Evaluation of fluoro- 9:1133–1134
phores for optimal performance in localization- 65. Banterle N, Bui KH, Lemke EA, Beck M
based super-resolution imaging. Nat Methods (2013) Fourier ring correlation as a resolution
8:1027–1036 criterion for super-resolution microscopy.
59. Wolter S, Schuttpelz M, Tscherepanow M, Van J Struct Biol 183:363–367
De Linde S, Heilemann M, Sauer M (2010) 66. Nieuwenhuizen RPJ, Lidke KA, Bates M, Puig
Real-time computation of subdiffraction- DL, Grunwald D, Stallinga S et al (2013)
resolution fluorescence images. J Microsc Measuring image resolution in optical nanos-
237:12–22 copy. Nat Methods 10:557–562
60. Deschout H, Zanacchi FC, Mlodzianoski M, 67. Erdélyi M, Sinkó J, Kákonyi R, Kelemen A,
Diaspro A, Bewersdorf J, Hess ST et al (2014) Rees E, Varga D et al (2015) Origin and com-
Precisely and accurately localizing single pensation of imaging artefacts in localization-
emitters in fluorescence microscopy. Nat
based super-resolution microscopy. Methods
Methods 11:253–266 88:122–132
61. Thompson RE, Larson DR, Webb WW (2002) 68. Patterson G, Davidson M, Manley S,
Precise nanometer localization analysis for Lippincott-Schwartz J (2010) Superresolution
individual fluorescent probes. Biophys imaging using single-molecule localization.
J 82:2775–2783 Annu Rev Phys Chem 61:345–367
Chapter 14
Förster Resonance Energy Transfer to Study TCR-pMHC

Interactions in the Immunological Synapse
Gerhard J. Schütz and Johannes B. Huppa
Abstract
T-cell antigen recognition is remarkably efficient: when scanning the surface of antigen-presenting cells
(APCs), T-cells can detect the presence of just a few single antigenic peptide/MHCs (pMHCs), which are
often vastly outnumbered by structurally similar non-stimulatory endogenous pMHCs (Irvine et al.,
Nature 419(6909):845–849, 2002; Purbhoo et al., Nat Immunol 5(5):524–530, 2004; Huang et al.,
Immunity 39(5):846–857, 2013). How T-cells achieve this is still enigmatic, in particular in view of the
rather moderate affinity that TCRs typically exert for antigenic pMHCs, at least when measured in vitro
(Davis et al., Ann Rev Immunol 16:523–544, 1998). To shed light on this in a comprehensive manner, we
have developed a microscopy-based assay, which allows us to quantitate TCR-pMHC interactions in situ,
i.e., within the special confines of the nascent immunological synapse of a T-cell contacting a planar-
supported lipid bilayer functionalized with the costimulatory molecule B7-1, the adhesion molecule
ICAM-1, and pMHCs (Huppa et al., Nature 463(7283):963–967, 2010) (Fig. 1). Binding measurements
are based on Förster resonance energy transfer (FRET) between site-specifically labeled pMHCs and
TCRs, which are decorated with recombinant site-specifically labeled single-chain antibody fragments
(scFV) derived from the TCRβ-reactive H57-597 antibody (Huppa et al., Nature 463(7283):963–967,
2010). FRET, a quantum-mechanical phenomenon, involves the non-radiative coupling of dipole moments
of two adjacent fluorophores, a donor molecule and an acceptor molecule. FRET efficiency is inversely
proportional to the sixth power of the inter-dye distance. Hence, it can be employed as a molecular ruler
(Stryer and Haugland, Proc Natl Acad Sci, USA 58(2):719–726, 1967) or, as is the case here, to score for
interactions of appropriately labeled molecules. To facilitate both quantitative and single-molecule read-
out, it is important to conjugate donor and acceptor dyes in a site-specific manner.
While SLBs mimic some but certainly not all properties of a plasma membrane of a living cell, their
use features a number of operational advantages: SLBs can be prepared in a fluid state, thereby facilitat-
ing the spatial rearrangements that accompany the formation of an immunological synapse (Grakoui
et al., Science 285(5425):221–227, 1999). The imaging of a three-dimensional binding process is
reduced to two dimensions, which saves time and fluorophore-emitted photons and allows for fast mea-
surements. Furthermore, images can be acquired in noise-attenuated total internal reflection (TIR)
mode, so far a necessity for single-molecule detection within the immunological synapse. Importantly,
the stimulatory potency of pMHCs is very well preserved compared to cell surface-embedded pMHCs.
Hence, while in principle artificial, SLBs are still a good approximation of the physiologic scenario a
T-cell encounters when approaching an APC. Vice versa, the reconstitutive approach offers unique
opportunities to interrogate the influence of accessory molecules on T-cell antigen recognition in a
highly quantitative manner.
In this chapter we will provide recommendations for the production of proteins used for SLB decora-
tion as well as hands-on protocols for the production of SLBs. We will describe in detail how to perform
207
208 Gerhard J. Schütz and Johannes B. Huppa
and analyze FRET-based experiments to determine synaptic binding constants. In the “Notes” section, we
will provide some information regarding the microscope setup as well as the mathematical and biophysical
foundation underlying data analysis.
Key words Affinity, Cytoskeleton, Fluorescence resonance energy transfer, Force, Single molecule
1 Materials
1.1 Proteins For stable SLB attachment, proteins need to be equipped with a
Employed to Decorate C-terminal (for type I membrane proteins) or N-terminal (for type
the SLB II membrane proteins) polyhistidine tag comprised of 10–12 histi-
dines. For proteins consisting of two membrane-anchored subunits
such as MHC class II molecules, each subunit can be extended
with six histidine residues. Due to space limitations, we will only
provide general recommendations as well as references.
1.1.1 B7-1-12H, Polyhistidine-tagged extracellular domains of accessory molecules

ICAM-1-12H like B7-1 or ICAM-1 can be easily expressed using a baculovirus-
based insect cell expression system and purified from culture super-
natant via affinity chromatography. Additional purification steps
(anion exchange chromatography, gel filtration) are required as
outlined [1]. If desired purified proteins can be labeled for later
visualization and density measurements via lysine residues using
NHS-modified fluorochromes.
1.1.2 pMHC For quantitative ensemble FRET measurements and the calcula-
tion of synaptic KDs, it is critical to employ site-specifically and
quantitatively labeled pMHCs. We have best results with pMHCs
refolded from E. coli inclusion bodies in vitro in the presence of a
space-holder peptide, which can later be exchanged for fluorescence-
labeled peptides [2, 3]. Please keep in mind that this strategy works
only for MHC class II molecules (and not necessarily for MHC
class I molecules) as their peptide-binding cleft is open at both
ends. For more detailed information, please refer to Axmann et al.
[1]. For site-specific labeling of MHC class I molecules, it is pos-
sible to introduce a cysteine residue in place of a serine residue in
close proximity to the peptide C-terminus for coupling dyes via
maleimide (unpublished observation). However it should be veri-
fied, e.g., by surface plasmon resonance measurement, that TCR
binding is unaffected by this modification.
1.2 Site-Specifically To obtain site-specifically as well as stoichiometrically labeled H57-

Fluorescence-Labeled 597 scFVs, you may want to introduce an unpaired cysteine residue
H57-597 Single-Chain for dye coupling via maleimide as published [4]. scFV-cysteine
Antibody Fragments mutants can be readily refolded from E. coli inclusion bodies [5]
(scFVs) and purified and modified as outlined [1]. Finally, the desired pro-
tein to dye ratio of 1 should be verified via spectrophotometry.
FRET-based Assay to Study Synaptic TCR-pMHC Binding 209
1.3 Microscope Custom-built systems are probably the most versatile and cost-
Setup efficient choice for conducting the experiments described below.
We would like to emphasize that any biologist with a high school
education in beam optics should be capable to set up such a sys-
tem. Critical elements are an inverted microscope equipped with a
TIR objective (numerical aperture ≥1.45); a fast EM-CCD camera
with single-molecule detection capabilities, tunable diode, or solid-
state lasers featuring 514 nm or 532 and 640 nm wavelengths and
allowing for shutting within the sub-millisecond to millisecond
range; and an emission beam splitter for simultaneous acquisition
of both the FRET donor and FRET (acceptor) channel. We have
recently published a more detailed description of a useful micro-
scope design [6].
1.4 Other Please refer to Table 1 for a list of other components required.
Components
2 Protocol
2.1 Lipid Bilayer Note: We have successfully worked with SLBs containing predomi-
Reconstitution nantly 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC),
which features one saturated fatty acid and one mono-unsaturated
fatty acid. For protein attachment via polyhistidine tails, POPC is
mixed with the synthetic lipid 1,2-dioleoyl-sn-glycero-3-{[N(5-amino-
1-carboxypentyl)iminodiacetic acid] succinyl}nickel salt (Ni-DOGS
NTA) to 1–10% (Fig. 1).
SLBs form spontaneously when unilamellar vesicles encounter
clean glass surfaces. Small unilamellar vesicles (SUVs) (between 20
and 100 nm in size) can be generated from dried lipid through
bath sonication. Lipid extrusion gives rise to large unilamellar ves-
icles (LUVs) (between 100 and 1000 nm in diameter). For protec-
tion from atmospheric oxygen, dried lipids are resuspended in
degassed PBS under inert gas until a whitish lipid suspension has
formed.
2.1.1 Mixing Lipids 1. Mix 45 mg of POPC and 6.9 mg Ni-DOGS NTA (Avanti
Polar Lipids) dissolved in chloroform in a 250 ml round bot-
tom flask.
2. Evaporate the chloroform using a Büchi rotary evaporator
(Büchi Labortechnik, Switzerland) or by carefully blowing it
off with inert gas such as nitrogen or argon (to be performed
inside a chemical hood). Do this while constantly turning the
flask, which leads to equal deposition of the lipid on the lower
half of the round bottom flask.
3. Attach the flask to vacuum overnight to get rid of the chloro-
form quantitatively.
Table 1
List of required components
Name of material Company Catalog number Comments

#1.5 glass slides VWR 631-0853
250 ml round bottom flask VWR 201-1357
Alexa Fluor®555 C2 maleimide Thermo Fisher A-20346 Site-specific protein
labeling via mutant
unpaired cysteines
Alexa Fluor®647 C2 maleimide Thermo Fisher A-20347 Site-specific protein
labeling via mutant
unpaired cysteines
Alexa Fluor®488-NHS Thermo Scientific A-20000 Protein labeling via lysine
residues
residues
residues
Autoclave tape VWR 489-1312 Or any other heat-stable
sticky tape
Avanti Mini-Extruder Avanti Polar Lipids 610000 Alternative to the bath
sonicator
Bath sonicator QSONICA Q700 QSONICA Q700
Büchi rotary evaporator VWR 531-0837
Conc. hydrogen peroxide Roth 9683.1
Conc. sulfuric acid Roth X944.2
Cy3 maleimide GE Healthcare PA23031 Site-specific protein
labeling via mutant
unpaired cysteines
Cy5 maleimide GE Healthcare PA25031 Site-specific protein
labeling via mutant
unpaired cysteines
Epoxy glue Uhu 45705 Uhu plus sofortfest 2 min
Fetal bovine serum (research HyClone SV30160.03 T-cell media supplement
grade )
Hank’s Balanced Salt Solution Thermo Fisher 10225362 Imaging buffer
plus calcium/magnesium
Lab-Tek chambers VWR 734-2062
Lipid DGS Ni-NTA Avanti Polar Lipids 790404C
(continued)
Table 1
(continued)
Name of material Company Catalog number Comments

Lipid POPC Avanti Polar Lipids 850457C Already in chloroform-
solved delivered
version which is
recommended
2-Mercaptoethanol Sigma Aldrich M6250 T-cell media supplement
Mouse interleukin-2 recombinant BPS Bioscience 90185-B T-cell media supplement
protein
Nonessential amino acid 100X HyClone SH30238.01 T-cell media supplement
Origin (analysis program) OriginLab http://www. Nonlinear fitting of two
originlab.com parameters (tauoff,
<ntlag>)
PBS Thermo Fisher 14190-136 Degas before usage
Penicillin/ Thermo Fisher 12000226 T-cell media supplement
streptomycin/L-glutamine
100x
Picodent twinsil 22 Picodent 13001000 Alternative to the epoxy
glue
Power meter LaserMate-Q Coherent Any power meter
sensitive in the used
spectral range is
sufficient
RPMI 1640, with l -glutamine Thermo Fisher 11554416 T-cell media
Syringe filter 0.2 μm Millipore GVWP04700
TetraSpeck™ microspheres, 0.1 Thermo Fisher T-7279
μm, fluorescent blue/green/
orange/dark red
Trichloromethane/chloroform Roth 3313.1
Tubes for ultracentrifugation Thermo Fisher 45237
polycarbonate 1.00 mm,
11 mm × 32 mm
Ultracentrifuge Thermo Fisher RC M150GX
Ultracentrifuge rotor Thermo Fisher S120-AT2
UV spectrophotometer Thermo Fisher
NanoDrop 2000c
Vacuum pump VWR 181-0248
Fig. 1 Principles underlying the FRET-based assay to measure TCR-pMHC binding in situ. (a) Schematic outline
of the planar glass-supported lipid bilayer (SLB) system. SLBs are composed of POPC (90–99%) and the syn-
thetic lipid DGS Ni-NTA (1–10%) and form spontaneously when clean glass surfaces are charged with small
unilamellar vesicles (SUVs) consisting of the corresponding lipids. Once formed, such SLBs can be functional-
ized with soluble polyhistidine-tagged extracellular portions derived from pMHCs, costimulatory B7-1 proteins, and
ICAM-1 adhesion proteins for T-cell stimulation. (b) Composite structure of a TCR in complex with an H57-597
2.1.2 Generation 1. Add 10 ml of degassed PBS to the dried lipid mixture within
of Unilamellar Vesicles the 250 ml round bottom flask.
from Dried Lipid 2. Fill the flask with nitrogen or argon and close it with a stopper.
by Sonication Seal the flask with autoclave tape to ensure that the stopper
remains tightly attached to the flask during sonication.
3. Bath-sonicate the lipid suspension in the flask at 120–170 W
for up to 60 min. It should have turned clear by the time you
finish.
4. You may want to monitor the progress of vesicle formation
spectrophotometrically. The absorption of the lipid emulsion
(use PBS as blank control) should remain constant at 234 nm
(as an approximate indicator for the amount of lipid present).
It should drop significantly at 550 nm due to reduced particle-
mediated light scattering.
5. Remove heavy non-unilamellar vesicles, which interfere with
the formation of a contiguous SLB, through centrifugation at
150,000 × g for 1 h at 25 °C followed by a second centrifuga-
tion of the supernatant at 288,000 × g (8 h, 4 °C).
2.1.3 SLB Formation 1. Immerse 24 × 50 mm #1 glass slides (German borosilicate,

and Functionalization e.g., Menzel-Gläser or Thermo Fisher) in a 1:1 mixture of
with Protein concentrated sulfuric acid and 30% hydrogen peroxide for
30 min.
2. Rinse glass slides under a stream of ddH2O out of a squirt
bottle, and place them on a Teflon stand (or on a similar device)
to let them dry for 10–30 min.
3. Take off the bottom glass slide of 8-well Lab-Tek chambers
(Nalge Nunc International), fill the bottom with epoxy glue
hardening within 5 min (Devon, Uhu) or dental glue
(Picodent), and carefully place one of the cleaned and air-dried
glass slides on the glue-covered chamber bottom. Wait for
10 min and remove excess glue from the bottom with a razor
blade or a scalpel.
4. Add 100 μl of a tenfold diluted SUV suspension into each well
and wait for 5 min to let a contiguous SLB form.
5. Rinse the SLB carefully at least twice with 15 ml PBS. Never
expose the formed SLB to air.
Fig. 1 (continued) single-chain fragment (scFV) and engaging a site-specifically labeled pMHC illustrates the
FRET-based approach described herein. Note the short distance of less than 45 Å separating the FRET dyes.
(c) FRET occurs when TCR and pMHC bind in situ, i.e., when receptor and ligand-associated FRET dyes are in
close enough proximity. The use of TIRF illumination reduces background substantially, which allows detection
of single-molecule fluorescence and FRET signals
6. For reproducibility of protein densities on the SLB, fill each

well all the way up with PBS. Take off 330 μl to leave 350 μl
PBS left in the well. Pipette 50 μl of a cocktail containing His-
tagged proteins to each well. Mix well and incubate at room
temperature for 60 min in the dark. Keep the incubation time
the same for all future experiments to ensure comparable pro-
tein densities on the SLB.
7. Rinse each well twice with 15 ml PBS.
Note: SLBs should be used for imaging no longer than 6 h after
the protein has been added. During this period, no loss in SLB-
associated protein can be detected and fluorophore recovery after pho-
tobleaching remains up to 95% (as long as SLBs are not in contact
with imaging buffer containing FBS).
2.2 Decorating 1. Pellet 106 T-cells taken from tissue culture for 2 min in a 5 ml
Murine T-Cells polypropylene round bottom tube (e.g., FALCON 352063),
with the H57-597 scFV e.g., employing a Clay Adams Sero-Fuge (Becton Dickinson)
or a similar centrifuge.
2. Take off the media, resuspend the cell pellet by it flicking gen-
tly, and add 0.5 μl of the H57-597 scFV (1 mg/ml) to the
suspended cell pellet. Employ Alexa Fluor 555- or Cy3-labeled
scFV for ensemble FRET measurements. When conducting
single-molecule FRET measurements, use a mixture of Alexa
Fluor 555-/Cy3-labeled scFVs (1 part) and unlabeled scFv
(5–9 parts).
3. Stain the cells on ice for 10–20 min to allow for quantitative
cell labeling.
4. Wash the cells twice using ice-cold imaging buffer (HBSS plus
1 mM CaCl2, 1 mM MgCl2, 0.5% ovalbumin, or 1% FBS).
Note: Primary T-cells can be stored on ice for up to 1 h without
significant loss of bound H57-597 scFV.
2.3 Measuring Note: You may refer to Fig. 2 for further guidance.
Protein Densities
1. To determine the average signal of individual fluorophores,
visualize them on an SLB. For this employ SLB-attached MHC
molecules loaded with stoichiometrically and site-specifically
Fig. 2 (continued) inspected for fluorescence intensity. To this end create a region
of interest (ROI) as shown in (c), which limits variability in intensity counts due to
inhomogeneous illumination resulting from the Gaussian laser intensity profile.
As shown in (d), determine the integrated intensity of 7 × 7 pixel-sized ROIs
placed around single-molecule signals (here, sm1, sm2, sm3, sm4) and corre-
sponding background (here, bg1, bg2, bg3, bg4). (e) Quantified average pixel
intensities are listed. To determine the single-molecule signal, integrated back-
ground intensities are subtracted from integrated signal intensities
A 55967
5µm
2615
B 3674
966
C 3674
966
D sm1 bg1 3674
sm2 sm3 bg4 sm4

bg2 bg3
bg4 sm4
966
E integrated integrated background

intensity intensity -corrected
sm signal background signal
1 79734 59884 19850
2 76014 61270 14744
3 77752 60828 16924
4 81855 64988 16867
Fig. 2 Quantitation of single-molecule signals. (a) A fluorescent SLB is imaged with the use of a slit aperture
placed within the excitation beam path of the microscope (for more detailed information on the custom-
assembled microscopy system used here, please refer to [6]), which allows defined fluorophore ablation with
the unmasked field. (b) Single fluorophores move into the previously bleached unmasked area and can be
fluorescence-labeled peptides as these pMHCs feature a

protein-to-dye ratio of 1. Bleach all labels on the SLB in the
field of view and allow the fluorescence recover partially to
visualize single fluorophores. Record many images in rapid
succession, for example, through application of a streaming
acquisition protocol.
2. Focus your analysis on the central illumination spot within the
bilayer, i.e., measure single-molecule and bulk fluorescence
only within this region of interest (ROI). For single-molecule
measurements, sum up the signal of 7 by 7 pixels surrounding
the single-molecule intensity peak located in the center.
Measure the background signal by integrating the intensity
counts of neighboring 7 × 7 pixels, which are free of any signal.
Subtract the background from the single-molecule signal to
arrive at the background-corrected single-molecule fluores-
cence value.
3. Repeat step 2 50 to ~1000 times to average the corrected sig-
nal. For this purpose you may also take advantage of open-
source software such as ThunderSTORM.
4. To be able to take non-saturated and unbleached images of the
bulk fluorescence of the SLB, place a neutral density filter of
two or higher into the excitation path. Determine the average
pixel intensity within the same ROI used for single-molecule
intensity measurements. Document the exposure time of the
bulk fluorescence measurements. For background subtraction
measure the signal within the same ROI imaged without
illumination.
5. To calculate the protein density within the SLB, multiply the
background-corrected bulk fluorescence average pixel inten-
sity determined in step 4 by the number of pixels per square
micron (e.g., 41.5), by 100 (if a neutral density filter of 2 was
employed), and by the exposure time used in step 2 and divide
it by the average single-molecule signal determined in step 3,
the exposure time used in step 4 (bulk measurement), and the
number of fluorophores per protein (e.g., 1).
2.4 Quantifying FRET Note: Since measurements rely only on the change in FRET donor
via FRET Donor channel intensity after FRET acceptor bleaching, no correction fac-
Recovery After FRET tors have to be employed, which renders this approach simple, robust,
Acceptor Bleaching and reliable. However, changes in FRET cannot be monitored over
time because FRET acceptor bleaching precludes repetitive measure-
ments of the same object. Fast acceptor bleaching is critical for keeping
measurement-associated noise low. Please refer to Fig. 3 for further
guidance.
1. Set up TIRF illumination and focus on the SLB.
2. Exchange PBS for imaging buffer (HBSS plus MgCl2/CaCl2
containing either 1% FBS or 0.5% ovalbumin).
A FRET donor channel FRET acceptor channel
H57-597 scFV - cy3 MHC-cy5
emission: 555-595 nm emission: 655-705 nm
10614 6790
1. excitation:
647 mn
1055 1823
10614 2104
2. excitation:
514mn
1055 1045
10614 65193
3. FRET acceptor
ablation
1055 22708
10614 2104
4. excitation:
514 mn
1055 1045
6. excitation: 10614 6790

647 mn
1055 5µm 1823
B
before after FRET acceptor ablation
10614
5µm 1187
average pixel intensities:

background 1187
synapse 3201 3518
TCR-cluster 6864 8721
FRET-yield:
synapse = (3518 - 3201) / (3201 - 1187) x 100% = 15.7%
TCR-cluster = (8721 - 6864) / (6864 - 1187) x 100% = 32.7%
Fig. 3 Bulk FRET yields as measured through FRET donor recovery after FRET acceptor bleaching. (a) Shown
is a representative measurement of synaptic FRET. As indicated a series of images was taken with the use of
an emission beam splitter giving rise to a FRET donor and a FRET acceptor channel (for more detailed informa-
tion on beam splitters, refer to Axmann et al. [6]). The line shown in the FRET acceptor and FRET donor chan-
nels indicates the boundary of the T-cell synapse. Note the increase in intensity in the FRET donor channel
after FRET acceptor bleaching (step 3). (b) FRET efficiencies can be determined for entire synapses and indi-
vidual synaptic regions. Images before and after FRET acceptor bleaching are shown with the use of two
lookup tables (green and physics). The boundary of the region used for background determination is marked
in blue, and the boundaries of the synapse and one representative TCR microcluster are colored in green (in
false color images only) and red (in both green and false color images), respectively
3. Transfer scFV-labeled T-cells to the well and let them settle

onto the SLB.
4. Fine-tune the excitation laser adjustment for best TIRF illumi-
nation. For this, adjust the focus slightly above the SLB plane.
When conditions for TIRF illumination are met, the illumi-
nated area turns blurry and no other parts of the cell’s H57-
597 scFV-decorated plasma membrane come into view.
5. Ensure superimposition of both laser profiles in the field of
view. If necessary, realign the green (514 nm or 532 nm) or
red (647 nm) laser.
6. To record FRET, take the following five images in rapid
succession:
• (1) 647 nm excitation (low power) to acquire an image of
the FRET acceptor (pMHC) prior to the bleach pulse
• (2) 514 nm/532 nm excitation (low power) to acquire an
image of the FRET donor (TCR) prior to the bleach pulse
• (3) 647 nm excitation (high power) to ablate the FRET
acceptor
• (4) 514 nm/532 nm excitation (low power) to acquire an
image of the FRET donor (TCR) following FRET acceptor
photobleaching
• (5) 647 nm excitation (low power) to confirm FRET accep-
tor ablation
• As an option you may acquire a DIC image before or after
this image sequence.
• For best results keep the time passed between the (2) and
(4) as short as possible. Also, minimalize FRET donor
bleaching by employing the excitation at the lowest possible
power level sufficient for FRET donor acquisition with
acceptable signal-to-noise ratio (≥10).
7. Focus on an ROI, e.g., a synapse or an individual TCR micro-

cluster; measure its intensity in (2) (=I(2)) and in (4) (=I(4)). For
background subtraction, choose an ROI of the same dimen-
sion outside the illumination spot in (2) or (4) and measure its
intensity (I(background)). To calculate the FRET yield, perform the
following operation:
FRET yield = {I (4) − I (2)} / {I (4) − I (background)}

2.5 Quantifying FRET Note: The emission of the FRET acceptor, as it occurs through FRET
via Sensitized donor excitation, is determined in the FRET channel, which makes it
Emission possible to follow changes in FRET over time. However, FRET donor
bleedthrough into the red-shifted acceptor channel and FRET accep-
tor cross excitation via direct donor excitation have to be carefully
determined for precise determination of FRET values. Corresponding

FRET donor and FRET acceptor channels have to be properly aligned.
Also, for best corrections for the FRET acceptor cross excitation, the
green and red laser excitation profiles have to be very well aligned.
1. Image alignment is very well achieved with the use of
TetraSpeck™ beads (0.1 μm in diameter, Thermo Fisher),
which emit in both the FRET donor and acceptor channels. In
TIRF mode only beads in contact with the glass surface pro-
duce a signal, and the spatial shift in register can be determined
by positioning individual beads within both fluorescence
channels.
2. Donor bleedthrough can be easily determined using an SLB
harboring the donor fluorophore alone (e.g., ICAM-1-Alexa
Fluor 555). First, determine the background outside the illu-
minated field of view and then subtract it from both channels.
Then determine the average background-corrected intensi-
ties of two corresponding ROIs (Idonor channel and Iacceptor channel).
Calculate FRET donor bleedthrough as follows:
Bleedthrough coefficient (BTC) = Iacceptor channel/Idonor channel
3. Acceptor cross excitation can be determined by exciting a

bilayer containing the acceptor fluorophore alone (e.g., MHC-
Alexa Fluor 647) first with donor excitation light (e.g., 514 nm
or 532 nm) and then with acceptor excitation light (e.g., 647
nm). Background-subtracted images within the FRET accep-
tor channel are then used to determine the cross excitation
correction matrix as follows:
[Cross excitation correction matrix] = [image, 647 nm excita-
tion, background-subtracted]/[image, 514/532 nm exci-
tation, background-subtracted].
To reduce statistical noise, this operation should be performed
for many image pairs. Use the median of the resulting
cross excitation correction matrix to plot the image gener-
ated solely through cross excitation.
[Image cross excitation] = [image, 647 nm excitation,
background-subtracted]/[median of cross excitation cor-
rection matrices]
4. The corrected FRET image can finally be calculated as
follows:
[Corrected FRET image] = [uncorrected FRET image,
background- subtracted] – [uncorrected FRET donor
image, background-subtracted] × BTC − [uncorrected
FRET image, background-subtracted]/[median of cross
excitation correction matrices].
time 0 ms 100 ms 200 ms 300 ms

300
Cy3
5 µm 0
75
Cy5
30
FRET
Cy5
FRET
Fig. 4 Single-molecule (sm) FRET events appear and disappear in single steps and co-localize with a single
FRET acceptor fluorophore. The trajectory of the smFRET event is shown. smFRET (annotated with a green
dashed circle) is visible in the first two frames and disappears in the third frame. Yellow arrows indicate
smFRET events matching in position with single FRET acceptor fluorophores. Image acquisition was performed
with a back-illuminated slow scan CCD camera in fast kinetics mode (for more information please refer to [6])
2.6 Single-Molecule Note: Please refer to the Note 1 in order to familiarize yourself with
FRET Measurements both the physics and mathematics behind the approach applied here.
For a guiding example, you may refer to Fig. 4.
1. Tune the power of both lasers to give rise to a power density of
1–5 kW/μm2 at the specimen. For more detailed information
on how to do this, refer to [6].
2. Decorate T-cells with the FRET probe as outlined above, i.e.,
with a mixture of unlabeled scFV (5–9 parts) and Alexa Fluor
555 7 Cy3-labeled scFV (1 part). In this fashion only a fraction
of TCRs is labeled, which certainly reduces the number of vis-
ible interactions. However, noise generated from donor
bleedthrough is also reduced substantially, which is required to
resolve individual single-molecule FRET events.
3. We advise to insert a slit aperture into the excitation pathway as

a means to mask the fluorescence-labeled SLB outside the syn-
apse. This way, even after repetitive FRET-based monitoring of
TCR-pMHC interactions within the same synapse, unbleached
FRET acceptor pMHC molecules can constantly recover into
the area of illumination to give rise to new TCR-pMHC inter-
actions, which can be tracked via single molecule FRET.
4. Add T-cells to the bilayer (immersed in imaging buffer) fea-
turing between 30 and 150 MHC-Alexa Fluor 647/-Cy5
molecules per μm2. Wait for synapses to appear in the field of
TIRF view.
5. Bleach the MHC molecules (=FRET acceptor) in view. Start
acquiring images when the MHC-related fluorescence on the
SLB is in the process of recovering.
6. Acquire rapidly a sequence of 10–20 image sets:
• Excitation 514 nm/532 nm
• Excitation 647 nm
• FRET donor and FRET acceptor fluorophores should be
exposed for 1–5 ms each and imaged with minimal delay
(e.g., 1 ms) separating the two acquisitions.
7. Single-molecule FRET events (Fig. 4) become visible in the
FRET channel without any need for donor bleedthrough and
acceptor cross excitation corrections. These can of course still
be performed, but should not be necessary for unambiguous
results.
8. To assess the identity of single-molecule FRET events, single-
molecule FRET events have to align with single acceptor mol-
ecules and should appear and disappear in one step.
2.7 TCR-pMHC Note: Prior to applying the following protocol, you may want to famil-
Off-Rate iarize yourself with the rationale underlying the approach, which is
Determination explained in detail in Note 1 . For further guidance please also refer
to Fig. 5 and Tables 2–4.
1. Record the trace lengths of single-molecule FRET events for at
least three acquisition time frames (tlags). In the example shown
in Fig. 5, we have measured the synaptic off-rate between the
5c.c7 TCR and IEk/K3 using four different delay times (42,
490, 1007, and 1989 ms).
2. Order FRET traces according to their trace lengths as done in
Table 2.
3. Transform Table 2 into an inverse cumulative decay function as
shown in Table 3 (colored numbers are taken from Table 2).
To normalize the function, divide the number of traces of
Table 3 by the sum of all traces in that particular group (as
done in Table 4).
Fig. 5 Extracting τoff = 1/koff through recording smFRET trajectories. (a) The fraction of detectable FRET events
(originating from H57-597 scFV-AF555-decorated TCR transgenic T-cell blasts recognizing AF647-labeled
weak agonist pMHC at 24 °C) is plotted for four different time lags (42, 490, 1007, and 1989 ms) against the
number of time frames after first smFRET detection (for additional guidance please refer to Tables 2–4). Mono-
exponential fit functions yield the corresponding negative inverse of the expectation values <n(tlag)>. (b)
Expectation values are plotted against delays tlag and fitted using equation <n(tlag)> = τoff/{(τoff/<nbleach>) + tlag}
to give rise to τoff and <nbleach>
Table 2
Frequency of trace lengths recorded with indicated delays
Number of Number of Number of

Trace length in traces at a traces at a traces at a delay Number of traces at
number of steps delay of 42 ms delay of 490 ms of 1007 ms a delay of 1989 ms
1 65 98 111 102
2 32 32 31 21
3 13 13 8 5
4 9 5 3 1
5 5 2 2 0
6 2 2 0 0
7 1 0 0 0
8 0 0 0 0
9 1 0 0 0
10 0 0 0 0
Table 3
Table 2 converted into an inverse cumulative decay function
Number of time Number of Number of Number of

frames survived Number of traces traces at a traces at a delay traces at a delay
after first detection at a delay of 42 ms delay of 490 ms of 1007 ms of 1989 ms
0 128 (=sum of 152 (=sum of 155 (=sum of 129 (=sum of
traces) traces) traces) traces)
1 63 (=128–65) 54 (=152–98) 44 (=155–111) 27 (=129–102)
2 31 (=63–32) 22 (=54–32) 13 (=44–31) 6 (=27–21)
3 18 (=31–13) 9 (=22–13) 5 (=13–8) 1 (=6–1)
4 9 (=18–9) 4 (=9–5) 2 (=5–3) 0 (=1–1)
5 4 (=9–5) 2 (=4–2) 0 (=2–2) 0 (=0–0)
6 2 (=4–2) 0 (=2–2) 0 (=0–0) 0 (=0–0)
7 1 (=2–1) 0 (=0–0) 0 (=0–0) 0 (=0–0)
8 1 (=1–0) 0 (=0–0) 0 (=0–0) 0 (=0–0)
9 0 (=1–1) 0 (=0–0) 0 (=0–0) 0 (=0–0)
4. Plot the normalized values against the number of time frames.

Omit the value of the last time frame (containing a zero) and
fit the observed decays with single exponents as shown in
Fig. 5a.
Table 4
Table 3 normalized by the sum of traces of the corresponding data group
Number of time
frames survived Fraction of
after first Fraction of traces traces at a delay Fraction of traces at Fraction of traces at
detection at a delay of 42 ms of 490 ms a delay of 1007 ms a delay of 1989 ms
0 1 (=128/128) 1 (=152/152) 1 (=155/155) 1 (=129/129)
1 0.492 (=63/128) 0.355 (=54/152) 0.283 (=44/155) 0.209 (=27/129)
2 0.242 (=31/128) 0.144 (=22/152) 0.0838 (=13/155) 0.0465 (=6/129)
3 0.14 (=18/128) 0.0592 (=9/152) 0.0322 (=5/155) 0.00775 (=1/129)
4 0.0703 (=9/128) 0.0263 (=4/152) 0.0129 (=2/155)
5 0.0312 (=4/128) 0.0131 (=2/152)
6 0.0156 (=2/128)
7 0.00781 (=1/128)
8 0.00781 (=1/128)
5. Plot the expectation value <n(tlag)>, i.e., the negative inverse

of the exponent of the decay function determined above,
against the delay time tlag employed (as shown in Fig. 5b, e, g,
x = tlag = 0.042 s and y = <n(tlag)> = 1/0.662 = 1.51057, x =
0.49 s and y = 1/0.902 = 1.10086, x = 1.007 s and y = 1/1.131
= 0.88417, x = 1.989 s and y = 1/1.591 = 0.62853). Now, fit
τoff and <nbleach> based on their relationship shown in Eq. 4 (see
Note 1). To this end employ a nonlinear fitting function of a
scientific data analysis program such as Origin (OriginLab).
Our example yields a τoff of 2.12 +/− 0.23 s and a <nbleach> of
1.53 +/− 0.06 s. The half-life of the interaction t1/2 off is
derived as follows:
t 1/ 2 off = t off ´ ln(2).

2.8 Determination Note: An example is shown in Fig. 6 for further guidance.

of Synaptic KDs
1. Measure FRET yields for individual TCR microclusters as
shown in Subheading 2.4 (Fig. 6a).
2. With the use of Eq. 8 (see Note 2), transform all individual
FRET yields into TCR occupancies (Fig. 6b).
Fig. 6 (continued) Note that C is a constant specific for the FRET system and the FRET fluorophores applied. In
this example it amounts to 1.995. Data were originally published in Huppa et al. [4] (H57-597 scFV-Cy3-deco-
rated TCRs of 5c.c.7 αβ TCR transgenic T-cells interacting with I−Ek/MCC-Cy5 at a density of 150 molecules
μm2 at 37 °C) and are displayed here in a new format
A B
N = 138 median = 0.518
0.4 0.2
fraction of clusters
median = 0.261 average = 0.532
0.3 average = 0.268 0.15
0.2 0.1
0.1 0.05
0 0
0.8
0.9
1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0
0.7
0.8
0.9
1
0.1
0.2
0.3
0.4
0.5
0.6
0
FRET yield / % TCR occupancy / %
C median = 139 molecules µ m-2

D median = 7.17 E-03 µ m2 molecules-1
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
E-01
E+00
E+00
E+01
E+01
E+02
E+03
E+03
E+04
E+04
E+05
E-05
E-05
E-04
E-04
E-03
E-02
E-02
E-01
E+00
E+00
E+01
2.50
1.53
6.10
2.44
9.77
3.91
1.56
6.25
2.50
1.00
4.00
1.60
6.40
2.56
1.02
4.10
1.64
6.55
2.62
1.00
4.00
1.60
effective synaptic KD / molecules µm-2 effective synaptic KA / µm2 molecules-1
E F 0.8
median = 4.48 E-02 µm2 molecules-1 s-1 C= 1.995
TCR occupancy
0.6
0.4
0.4
0.3
0.2 0.2
0.1
0
0 0 0.1 0.2 0.3 0.4
E+01
E-04
E-04
E-03
E-02
E-02
E-01
E+00
E+00
E+01
FRET yield
2.44
9.77
3.91
1.56
6.25
2.50
1.00
4.00
1.60
6.40
effective synaptic kon / µm2 molecules-1 s-1
Fig. 6 Determining effective synaptic KDs, KAs, and kons. (a) FRET yield data were determined for individual TCR
microclusters (N = 138) through FRET donor recovery after FRET acceptor bleaching. Numbers below histo-
gram bars indicate the upper limit within the interval. (b) Conversion of the FRET yields shown in (a) into TCR
occupancies a by multiplying measured FRET yields with the constant C determined in (f). Numbers below bars
indicate the upper limit within the interval. (c) Histogram (semilogarithmic, base = 4) showing the distribution
of synaptic KDs measured for individual TCR microclusters. Numbers below bars indicate the upper limit within
the interval. (d, e) The data shown in (a–c) were transformed into histograms displaying effective synaptic KAs
(d) and kons (e) (semilogarithmic, base = 4) employing the corresponding synaptic koff (6.25 s−1). (f) The cor-
relation between FRET yield values as determined by donor recovery after acceptor bleaching and TCR occu-
pancy was measured experimentally. TCR occupancy can be determined for individual TCR microclusters as
explained in Note 2. The slope of the linear fit is equal to the ratio C between a (TCR occupancy) and the FRET yield.
3. Employ Eq. 6 to transform all TCR occupancies into synaptic

KDs. As is indicated below, synaptic binding is inhomoge-
neous. A meaningful measure for synaptic KDs is the median
of all measured microclusters (Fig. 6c). As an adequate mea-
sure for the synaptic KA(= 1/KD, Fig. 6d), the average can be
used as well.
2.9 Calculation Synaptic kon values can now be calculated according to the law of
of Synaptic kons mass action with kon = koff/KD. The synaptic koff for the experiment
shown in the example (i.e., IEk/MCC interacting with the 5c.c7
TCR at 37 °C at an SLB density of 150 pMHC/μm2) is 6.25 s−1.
Hence, the synaptic KA plot (Fig. 6d) can be converted into a syn-
aptic kon plot (Fig. 6e).
3 Notes
1. Rationale behind off-rate determination.

Like most other fluorescence-based approaches, FRET-
based measurements are also susceptible to photobleaching.
This must be accounted for when extracting the half-lives of
synaptic TCR-pMHC binding events from single-molecule
FRET traces. The initial number of observable FRET signals
N(0) of a single donor-acceptor pair as a function of time t is
diminished by photobleaching of either the FRET donor or
the FRET acceptor and the off-rate of the interaction:
N (t ) = N (0) ´ exp(-t / t bleach ) ´ exp(-t / t off ) (1)

In the photobleaching term exp (−t/τbleach), we can replace
the time t by the product of the number of observations n and
the illumination time till because of the stroboscopic observa-
tion mode. Within the kinetic term exp −(t/τoff), we can
describe the time t by the product of the number of observa-
tions n and the time tlag for a single FRET observation.
N (t ) = N (0) ´ exp{-(n ´ t ill ) / t bleach } ´ exp(-n ´ t lag / t off ) (2)

The expectation value, <nbleach>, for the number of obser-
vations until bleaching occurs, is given by τbleach/till.
N (t ) = N (0) ´ exp(-n / ánbleach ñ) ´ exp(-n ´ t lag / t off )
(3)
= N (0) ´ exp{-n / [t off / (t off / ánbleach ñ + t lag )]}

The expectation value <n(tlag)> of the number of frames
with observable FRET signals N is therefore given by:
án(t lag )ñ = t off / {(t off / ánbleach ñ) + t lag } (4)

We control in the experiment tlag and we read out <n(tlag)>.

Two variables need to be extracted: the inverse of the koff =
1/koff = τoff, which describes the off-rate of the interaction, and
the bleach coefficient <nbleach>, which depends on the fluoro-
phores, excitation power, and illumination time employed.
With the relationship shown in Eq. 4, we can derive the two
missing constants τoff and <nbleach> by applying in the imaging
experiment at least three different tlags.
2. Rationale behind determining synaptic KDs and kons.
FRET yields are directly proportional to TCR occupancy
a, i.e., the ratio between bound TCRs and total TCRs. Of
note, this parameter can be transformed into a synaptic KD
according to Eq. 6 (see below) when the initial density of TCR
ligands (i.e., pMHCs) prior to the addition of T-cells to the
SLB is known. This is because SLBs provide an almost inex-
haustible reservoir of ligands and because of the high lateral
mobility of SLB-attached proteins.
a = FRET yield ´ C
(5)
with a = TCR occupancy and C = conversion factor
Synaptic KD = (1 / a - 1) ´ [pMHC initial ]
(6)
with [pMHC initial ] = initial density of pMHC prior to the addition of T-cells
C will need to be extracted from imaging experiments

since it varies with the FRET system and fluorophores used. In
the example summarized in Fig. 6, we determine it for the
FRET system consisting of H57-597 scFV-Cy3 (bound to the
5c.c7 TCR) and IEk/peptide(C)-Cy5.
At first we need to establish the relationship between the
corrected FRET intensity IFRET and a. For this we derive the
ratio R between the average fluorescence intensity of single
FRET donor fluorophores sm IFRETdonor and the average inten-
sity of a single-molecule FRET event sm IFRET. We can then
directly determine a according to Eq. 7.
a = [TCR ´ pMHC] / [TCRtotal ]
= bulk I FRET / sm I FRET ´ sm I FRETdonor / bulk I donor
= bulk I FRET / bulk I FRETdonor ´ sm I FRETdonor / sm I FRET (7)
= bulk I FRET / bulk I FRETdonor ´ R
with R = sm I FRETdonor / sm I FRET

R was determined as 1.45 for the H57-597 scFV- Cy3-
IEk/peptide(C)-Cy5 FRET system and leads to a = bulk
IFRET/bulk ITCR-cy3× 1.45.
Next we establish the relationship between TCR occu-
pancy a and the FRET yield as determined by FRET donor
recovery after acceptor bleaching. For this we plot both param-
eters against one another for a number of TCR microclusters

as shown in Fig. 6f.
The slope of the linear fit indicates the conversion factor C
(from Eq. 5). For the H57-597 scFV-Cy3-IEk/ peptide(C)-
Cy5 FRET system, C amounts to 1.995.
TCR occupancy a = FRET yield ´ 1.995 (8)
Applying Eqs. 6 and 8, we can now easily calculate the

synaptic KD between TCR and pMHC (see Subheading 2.8).
3. Choice of coverslips for single-molecule FRET measurements.
To be able to conduct single-molecule FRET measure-
ments, it is of great importance to use coverslips which do not
give rise to high background signals in the FRET channel
when exciting with donor excitation light. We noticed that the
use of coverslips manufactured/distributed by Menzel-Gläser/
Thermo Fisher does not produce spurious signals, while the
use of those of other manufacturers does. If you cannot get
access to Menzel-Gläser coverslips, you may want to verify the
quality of other coverslips or resort to quartz glass.
4. Verifying the functional integrity of the SLB.
Lateral mobility of all proteins attached to the lipid bilayer
is crucial for many of the assays shown. For example, the ratio-
nale underlying the determination of synaptic KDs rests on the
assumption that SLBs are fluid entities. Furthermore TCR-
pMHC binding is substantially reduced, as T-cells contacting
immobile SLBs require considerably higher antigen densities
for stimulation and TCR-pMHC binding (as quantified by
FRET). Hence, both SLB fluidity and T-cell antigen sensitivity
should be measured in initial experiments. The degree of
immobility is easily determined using a classical fluorescence
recovery after photobleaching (FRAP) experiment. For mea-
suring diffusion constants FRAP can be applied, yet single dye
tracing, once established in the lab, is the method of choice as
it also indicates the degree to which diffusion behavior is
homogeneous. Given the absence of any detergent during
bilayer formation and protein production, the SLB system
described herein excels in mobility and should give rise to an
immobile fraction lower than 5%. In case higher values are
measured, we recommend verifying the purity and the mono-
meric (i.e., non-aggregated) state of the protein preparations.
While proteins refolded from E.coli inclusion bodies are typi-
cally of high purity and right size after gel filtration (exception:
weakly expressed proteins accumulating in highly contami-
nated inclusion bodies), proteins expressed in insect or dro-
sophila cells require in addition ion exchange chromatography
to remove lipid contaminations, as these may considerably
reduce lateral mobility within SLBs.
The sensitivity of T-cells toward SLB-presented antigens

can be easily verified through measuring via Fura-2 the calcium
response of T-cells making SLB contact. T-cells should respond
to SLBs harboring five or fewer antigens per μm2. If this is not
the case, you may want to verify the presence and integrity of
accessory proteins such as ICAM-1 and, when working with
naïve or CD4+ helper T-cells, also B7-1.
Acknowledgments
This work has been supported by research grants of the Vienna

Science and Technology Fund (WWTF) LS13-030 (GS and JH)
and LS14-031 (JH).
References
1. Axmann M, Schutz GJ, Huppa JB (2015) 4. Huppa JB, Axmann M, Mortelmaier MA,
Measuring TCR-pMHC binding in situ using a Lillemeier BF, Newell EW, Brameshuber M,
FRET-based microscopy assay. J Vis Exp Klein LO, Schutz GJ, Davis MM (2010) TCR-
105:e53157. doi:10.3791/53157 peptide-MHC interactions in situ show acceler-
2. Toebes M, Coccoris M, Bins A, Rodenko B, ated kinetics and increased affinity. Nature 463
Gomez R, Nieuwkoop NJ, van de Kasteele W, (7283):963–967. doi:nature08746 [pii]
Rimmelzwaan GF, Haanen JB, Ovaa H, 5. Tsumoto K, Shinoki K, Kondo H, Uchikawa
Schumacher TN (2006) Design and use of M, Juji T, Kumagai I (1998) Highly efficient
conditional MHC class I ligands. Nat Med recovery of functional single-chain Fv frag-
12(2):246–251 ments from inclusion bodies overexpressed in
3. Xie J, Huppa JB, Newell EW, Huang J, Ebert Escherichia coli by controlled introduction of
PJ, Li QJ, Davis MM (2012) oxidizing reagent--application to a human
Photocrosslinkable pMHC monomers stain single-chain Fv fragment. J Immunol Methods
T cells specifically and cause ligand-bound 219(1–2):119–129
TCRs to be “preferentially” transported to 6. Axmann M, Schutz GJ, Huppa JB (2015)
the cSMAC. Nat Immunol 13(7):674–680. Single molecule fluorescence microscopy on
doi:10.1038/ni.2344 planar supported bilayers. J Vis Exp 105:
e53158. doi:10.3791/53158
Chapter 15
Two-Dimensional Analysis of Cross-Junctional Molecular

Interaction by Force Probes
Lining Ju, Yunfeng Chen, Muaz Nik Rushdi, Wei Chen, and Cheng Zhu
Abstract
Upon engagement with a specific ligand, a cell surface receptor transduces intracellular signals to activate
various cellular functions. This chapter describes a set of biomechanical methods for analyzing the charac-
teristics of cross-junctional receptor–ligand interactions at the surface of living cells. These methods com-
bine the characterization of kinetics of receptor–ligand binding with real-time imaging of intracellular
calcium fluxes, which allow researchers to assess how the signal initiated from single receptor–ligand
engagement is transduced across the cell membrane. A major application of these methods is the analysis
of antigen recognition by triggering of the T cell receptor (TCR). Three related methods are described in
this chapter: (1) the micropipette adhesion assay, (2) the biomembrane force probe (BFP) assay, and
(3) combining BFP with fluorescence microscopy (fBFP). In all cases, an ultrasoft human red blood cell
(RBC) is used as an ultrasensitive mechanical force probe. The micropipette assay detects binding events
visually. The BFP uses a high-speed camera and real-time image tracking techniques to measure mechanical
variables on a single molecular bond with up to ~1 pN (10−12 Newton), ~3 nm (10−9 m), and ~0.5 ms
(10−3 s) in force, spatial, and temporal resolution, respectively. As an upgrade to the BFP, the fBFP simul-
taneously images binding-triggered intracellular calcium signals on a single live cell. These technologies
can be widely used to study other membrane receptor–ligand interactions and signaling under mechanical
regulation.
Key words Dynamic force spectroscopy, Single molecule, T cell receptor, Micropipette,
Mechanotransduction, Biomembrane force probe, Red blood cell, GPIb, Integrin
1 Introduction
Cell-to-cell and cell-to-extracellular matrix (ECM) adhesion is

mediated by binding between cell surface receptors, ECM pro-
teins, and/or lipids. These physical interactions allow cells to form
functional structures, as well as to recognize, communicate, and
react to their microenvironment. Unlike soluble molecules (e.g.,
cytokines and growth factors) which bind to cell surface receptors
from a three-dimensional (3D) fluid phase, cell adhesion receptors
form bonds with their ligands across a narrow junctional gap to
bridge two opposing surfaces that constrain molecular diffusion to
231
232 Lining Ju et al.
a two-dimensional (2D) interface [1–4]. In the immune system,

T cells are the body’s sentinels, patrolling the body in search of
foreign threats such as infection by bacteria and viruses. TCR iden-
tify invaders by recognizing their specific antigens, i.e., peptides
bound to major histocompatibility complex (pMHC) molecules,
allowing the T cells to discriminate attackers from the body’s own
cells. When they recognize a threat, T cells can directly destroy it
or signal other elements of the immune system to confront the
invader.
There are many receptors on the surface of the T cell that can
initiate, propagate, and regulate cell activation. These receptors are
responsible for appropriate maturation of thymocytes, naïve T cell
activation, and initiation of effector immune functions. It is well
documented that once surface receptors engage their ligands on
the antigen-presenting cell, these molecules organize into a spa-
tially well-defined cluster and form a tight 2D junction termed the
immunological synapse (IS) [5]. This synapse ensures sustained
TCR engagement and downstream signaling. The 2D adhesion
techniques, developed by our group, allow the study of interaction
between TCR—of which there are as many as a million—and
pMHC protein molecules [3, 4]. In contrast to earlier 3D tech-
niques, which isolate the receptor molecules for study in a solution
environment, our 2D techniques provide a more realistic represen-
tation of their activity as they allow the study of receptors in their
native membrane environment (Fig. 1).
Our first 2D adhesion technique—micropipette adhesion
assay—was developed in 1998 to measure 2D receptor–ligand
binding kinetics [6]. The assay uses a human red blood cell (RBC)
as the adhesion sensor as well as to present ligands. It employs
micromanipulation to bring the RBC into contact with another
cell that expresses the receptors of interest with precisely controlled
contact area and time to enable bond formation. The adhesion
event is detected as an RBC elongation upon pulling the two cells
apart (cf. Fig. 2c). The adhesion probability is estimated from the
frequency of adhesion events in a sequence of repeated contact
cycles between the two cells for a given contact time. Varying the
contact time generates a binding curve. By adjusting the density of
ligands immobilized on the RBC surface, the steady-state proba-
bility of adhesion (where the binding curve saturates at sufficiently
long contact time) is kept in midrange between 20 and 80%. Fitting
a probabilistic model for receptor–ligand reaction kinetics to the
binding curve returns the 2D affinity and off-rate [6–8]. Since
then, the assay has been validated using interactions of Fcγ recep-
tors with IgG Fc [6, 9–12], selectins with glycoconjugate ligands
[13–15], integrins with ligands [16–18], platelet adhesion recep-
tor GPIb [19, 20], homotypical cadherin binding [21, 22], and,
most importantly, pMHC with CD8 [23, 24], pre-TCR [25],
TCR [26–30], or both TCR and CD8 [24, 31, 32].
Two-Dimensional Analysis of Molecular Interaction 233
Fig. 1 Instrumentation. Schematic drawing (a) and actual pictures (b–f) of the micropipette 2D analysis hard-
ware system. (a) The micropipette 2D analysis technique consists of a hardware system (optical, mechanical,
and electrical components) and a software system developed with LabVIEW. For concurrent fluorescence imag-
ing, a dual-cam system “DC2” is adapted. (b) The system overview. (c) Water pressure tower (manometer) for
adjustment of micropipette aspiration pressure. (d) Experiment chamber, adaptor, and two micropipette assem-
blies for the 2D adhesion assay. (e) The dual-cam system “DC2” (orange) onto which the high-speed camera
(blue) and a fluorescence camera (white) were mounted. (f) Three micropipette assembly for the BFP assay
Beyond merely providing physical linkage for cellular c ohesion,

adhesion receptors are a major component of the signaling machin-
ery for the cell to communicate with its surroundings. There has
been increasing interest in understanding how ligand engagement
of adhesion molecules initiates intracellular signaling and how the
initial signal is transduced inside the cell. Intuitively, properties of
receptor–ligand binding can impact the signals it induces. However,
it is difficult to dissect mechanistic relationships between the extra-
cellular interaction and intracellular signaling events using tradi-
tional ensemble of biochemical assays because of their many
limitations, e.g., a poor temporal resolution and the complete lack
of spatial resolution.
TCR’s antigen recognition process is highly dependent on
biomechanical cues that are generated from dynamic cytoskeleton
rearrangement, membrane bending, etc. Thus, to understand the
molecular mechanism of TCR triggering and antigen recognition,
we upgraded the 2D micropipette technique into a biomembrane
force probe (BFP)—an ultrasensitive force spectroscopic technique
Fig. 2 Micropipette adhesion assay setup. (a, b) Micrographs of the two micropipette setup in an experimental
chamber. (a) Micropipette assembly showing the probe micropipette (left) and target micropipette (right).
(b) Cell placement. A pMHC-coated RBC and a T cell are aspired on the probe and target pipettes, respectively.
(c) A micrograph depicting an adhesion event. A pMHC-coated RBC that previously contacted a T cell is being
retracted and elongated by the adhesion force, allowing binding to be unambiguously detected. (d) The molec-
ular binding between the ligand (pMHC on RBC) and receptor (TCR on T cell) pair that mediates the cross-
junctional adhesion event in the highlighted area in (c). (e) The deflection of RBC and the position of the target
(T cell) in a test cycle of micropipette 2D adhesion assay. A T cell (green) is brought into contact with a pMHC-
coated RBC (red). Upon T cell retraction, whether adhesion is present or not is detected by RBC elongation as
shown in (c). Adhesion frequency can be estimated from repeated contacts at a given contact duration
with high spatiotemporal resolution [7, 33]. The BFP enables

measurement of single-molecule 2D kinetics [7, 20, 26, 34],
mechanical properties [17, 35], conformational changes [36–38],
and regulation of 2D kinetics by biophysical factors, such as
the Brownian motion, separation distance, and diffusivity [34]. In
addition, we have further upgraded our BFP to allow the concur-
rent measurement of these characteristics with downstream signal-
ing (e.g., calcium flux) triggered by these single-bond events via
fluorescence imaging [39]. With this setup, in situ cell signaling
activities in the context of surface mechanical stimulation were
observed in T cells [27] and platelets [38]. The fBFP is versatile
and can be used for studies of cell adhesion and signaling mediated
by other molecules in other cells.
In summary, this chapter describes three 2D biomechanical

assays that serve different purposes: (1) the micropipette adhesion
assay to measure 2D binding kinetics under no pulling force;
(2) the biomembrane force probe (BFP) assay to measure force-
dependent off-rates of receptor–ligand dissociation, molecular
elasticity, and protein conformational changes; and (3) the fBFP
assay that integrates single-molecule mechanical characterization
with fluorescence microscopy to directly correlate receptor–ligand
kinetics and mechanics to the downstream signals so induced.
Unlike 3D bulk measurement technique (e.g., surface plasmon
resonance) that immobilizes receptors or ligands on a sensor chip,
these 2D assays allow movement and organization of T cell surface
receptors within their native membrane contexts, and they bring
surface molecules into contact with one another within a 2D cel-
lular junction with ligands immobilized on the surface of a RBC or
a glass bead. Moreover, the RBC probe or attached bead can be
coated with more than one ligand, allowing one to examine coop-
erativity between receptors on the cell surface [9]. The down-
stream effects of TCR triggering can be studied when a technique
is combined with fluorescence imaging [27, 38, 40]. Overall, the
2D biomechanical assays allow one to address the central questions
for T cell immunity and central tolerance—How do T cells recog-
nize subtle differences in TCR ligands, initiate distinct intracellular
signals, and induce the appropriate T cell function?
2 Materials
2.1 T Cells 1. Naïve CD8+ T cells were isolated from spleens of OT1
transgenic mice according to an Emory University IACUC-
approved protocol and placed into 3–5 ml HBSS 1× with Ca2+
and Mg2+.
2. For cell isolation: HBSS 1× with Ca, Mg.
3. Sterile 100 μm nylon cell strainer.
4. 3 ml syringe.
5. 6 cm Petri dish.
6. Mouse erythrocyte lysing kit/buffer.
7. Purified water.
8. Mouse CD8+ T cell enrichment kit (Stemcell Technologies).
9. EasySep magnet (Stemcell Technologies).
10. EasySep buffer: 1× PBS without Ca2+ and Mg2+, 2% FBS.
11. Cell culture media: RPMI + 1% FBS, l-glutamate, βME,
gentamicin.
2.2 Red Blood Cells 1. 5 ml blood from healthy donor according to a protocol
for Micropipette approved by the Institutional Review Board at the Georgia
Institute of Technology.
2. Experimental additive solution 45 (EAS45): 0.27 g/l adenine,
19.82 g/l d-glucose (dextrose), 10.02 g/l D-mannitol, 2.92 g/l
sodium chloride (NaCl), 84 g/l sodium phosphate dibasic
(Na2HPO4), 1.46 g/l l-glutamine.
3. Histopaque-1077 (Sigma-Aldrich).
4. 10 ml blood collection tubes with EDTA-based anticoagulant.
5. PBS buffer (pH 7.4, without Ca2+ and Mg2+).
2.3 Reagents 1. pMHC monomers. For RBC ligand coating, we have obtained
for Micropipette 2D synthesized recombinant pMHC extracellular domain mono-
Adhesion Assay mers from the National Institute of Health Tetramer Core
Facility at Emory University. In addition to chicken ovalbumin
(OVA) peptide for the OT-I system, an OVA-derived altered
peptide ligand (APL) sequence is tethered to a mutant MHC
haplotype of H-2Kb (replacing the α3 domain in mouse H-2Kb
with the α3 domain of human HLA-A2). It generates a p eptide
bound to an MHC molecule after the plasmids are expressed
in the clones. Here, we have the following sequences of
peptides: (a) chicken OVA-derived peptides OVA257-264

(SIINFEKL, agonist, and negative selecting ligand) and (b)
G4 (SIIGFEKL, weak agonist/antagonist, and positive select-
ing ligand [41]). OVA and G4 are recognized by OT1 TCR. All
of the pMHC monomers are engineered to have a biotin tag
on the C-terminus of the α chain [42].
2. L15 chamber media: 14.5 ml of L15, 0.5 ml of 30% BSA, and
75 μl of 1 M HEPES.
3. PBS buffer, pH 7.4, without Ca2+ and Mg2+.
4. FACS buffer, 1× PBS without Ca2+ and Mg2+, 5 mM EDTA,
1% BSA, 25 mM HEPES, and 0.02% sodium azide.
5. Experimental additive solution 45 (EAS45): 0.27 g/l adenine,
19.82 g/l d-glucose (dextrose), 10.02 g/l D-mannitol, 2.92 g/l
sodium chloride (NaCl), 84 g/l sodium phosphate dibasic
(Na2HPO4), 1.46 g/l l-glutamine.
6. Biotin-X-NHS.
7. Streptavidin powder.
8. Bovine serum albumin.
9. PE-conjugated anti-mouse TCR Vα2 monoclonal antibody
(mAb) (clone B20.1).
10. PE-conjugated anti-mouse H-2Kb (clone 25-D1.16).
11. PE-conjugated mouse IgG2a.
12. QuantiBRITE PE standard beads (340495, BD Biosciences)
or similar standard.
2.4 Reagents 1. Blood finger prick lancet device.

for Biomembrane 2. Carbonate/bicarbonate buffer (pH 8.5–9): 8.4 g/l sodium
Force Probe Assay carbonate (Na2CO3), 10.6 g/l sodium bicarbonate (NaHCO3).
3. Phosphate buffer (pH 6.5–6.8): 27.6 g/l sodium phosphate
monobasic (NaH2PO4⋅H2O), 28.4 g/l anhy. sodium phos-
phate dibasic (Na2HPO4).
4. N2–5% buffer (pH 7.2–7.4): 20.77 g/l potassium chloride
(KCl), 2.38 g/l sodium chloride (NaCl), 0.13 g/l potassium
phosphate monobasic (KH2PO4), 0.71 g/l anhy. sodium phos-
phate dibasic (Na2HPO4), 9.70 g/l sucrose.
5. Maleimide-PEG3500-N-hydroxysuccinimide.
6. Biotin-PEG3500-N-hydroxysuccinimide.
7. (3-Mercaptopropyl) trimethoxysilane (MPTMS).
8. Borosilicate glass beads (2 μm diameter in average).
9. Streptavidin−maleimide.
10. QuantiBRITE PE beads (340495, BD Biosciences) or similar
standard.
2.5 Micropipette 1. Capillary tube 0.7–1.0 mm × 30 in.

Fabrication 2. Flaming/Brown micropipette puller (P-97, Sutter Instrument)
or equivalent.
3. Pipette microforge (MF-900, Narishige) or equivalent.
2.6 Chamber 1. Mineral oil (Fisher Scientific).

Assembly 2. Microscope coverslip (40 mm × 22 mm).
3. Micro-injector (MF34G-5, World Precision Instruments) or
equivalent.
4. 1 ml tuberculin syringe.
5. Micropipette holder (HI-7, Narishige) or equivalent.
2.7 Fluorescence 1. Fura2-acetoxymethyl ester.

BFP Upgrade Materials 2. Dimethyl sulfoxide (DMSO).
2.8 Micropipette 2D Details of our instrument are provided below, but other hardware
Adhesion Assay can be adapted to this purpose.
Instrumentation
1. Acquire an inverted microscope (TiE, Nikon) with a numerical
aperture (NA) 0.85 condenser with a top lens and a CFI Plan
Fluor 40× objective (NA 0.75 WD 0.72 mm; Nikon) (Fig. 1b).
2. Implement the microscope onto an air anti-vibration table
(5′ × 3′; 77049089, TMC) isolating the mechanical vibrations
from the environment (Fig. 1b).
3. Customize the microscope (TiE, Nikon) by mounting (a) a set

of 3D mechanical manipulators (462-XYZ-M, Newport) on
the left side of the microscope stage to hold the probe micro-
pipette and (b) a set of 3D piezoelectric translators (M-105.3P,
Physik Instrumente) on the right side of the microscope stage
to hold the target micropipette (Fig. 1b) (see Note 1).
4. Mount a 1D piezo linear actuator (P-753.1CD, Physik Instru
mente) with capacitive feedback control and sub-nanometer
precision to the 3D translator.
5. Connect the target micropipette holder to the 1D linear actua-
tor, the driver of which is controlled by LabVIEW code that
allows one to move the target pipette along the axial direction
precisely and repeatedly in an adhesion test cycle.
6. Mount a normal speed CCD camera (30 fps, 1280 × 960,
Mono., CCD, 12 Bit ADC; GC1290, Prosilica) onto a camera
side port of the microscope by a video tube.
7. Build an in-house water pressure regulation system (manome-
ter, Fig. 1c) for aspiration control, which will hold the cells and
control the contact of a pMHC presenting an RBC onto a
T cell, using a fine mechanical positioner to precisely manipu-
late the height of the reservoir.
8. Connect a hydraulic tubing line between each micropipette
holder to its corresponding manometer reservoir.
9. Tune a customized LabVIEW program to control the 1D piezo
actuator and drive the axial movement of the target micro
pipette (see Note 2).
2.9 Biomembrane 1. Implement a mercury lamp with a focus tunable mercury lam-
Force Probe Upgrade phouse (HMX-4; Nikon) as the bright-field light source, which
can show a clear diffraction pattern of a glass bead and pro-
vides strong light and for high-speed camera grabbing in force
probe.
2. Add optical filters on the light path from the lamp to the cam-
era: (a) a neutral density filter (45 mm ND 8 or 16 A; Nikon)
to reduce the brightness for the protection of human eyes and
camera CCD; (b) a diffuser to keep the glass bead edge sharp
(45 mm; Nikon); and (c) a green light filter (45 mm
560 nm ± 20 nm, Chroma) to reduce chromatic aberrations
from the RBC (Fig. 1d).
3. Add a high-speed CCD camera (3000 fps, 640 × 480, GigE,
1/3″ CCD, mono GE680; Prosilica) onto another camera side
port of the microscope by a video tube.
4. Mount a hydraulic micromanipulator onto the microscope
stage with a remote fine control to position the probe bead
onto the apex of the red cell (PH400, Karl Suss).
5. Connect a hydraulic tubing line between helper micropipette

holder to its corresponding manometer reservoir.
6. Add a pressure sensor to the manometer connecting to the
probe pipette, which can quantitatively measure and indicate
aspiration pressure to determine the RBC’s spring constant.
7. Tune a customized LabVIEW program to control image acqui-
sition by the high-speed camera.
2.10 Concurrent 1. Set up a fluorescence light source (Lambda XL, Sutter

Fluorescence Imaging Instrument) that holds two excitation filters (340 nm ± 10 nm
Upgrade and 380 nm ± 10 nm; Chroma Technology) in its excitation
filter wheel.
2. Replace the green light filter (see effect in Fig. 1d) with a red
light filter (high pass 605 nm; Chroma Technology) (see effect
in Fig. 1f) in front of the bright-field light source (mercury
lamp; Nikon).
3. Implement a dual-cam system “DC2” that splits the light into
two and transmits them to the high-speed camera (Fig. 1e,
blue) and a fluorescence camera (ORCA-R2; Hamamatsu)
(Fig. 1e, white). The former will collect the probe image for
position determination, and the latter will collect fluorescence
images emitted from the target cell.
4. Implement a dichroic mirror (DC 565 LP, Chroma Technology)
and an emission filter set (T565LPXR, Photometrics) inside
the DC2.
5. Put a dichroic mirror (DC R488 nm, Chroma Technology)
into a filter wheel of the microscope to reflect the excitation
light with wavelength shorter than 488 nm to the sample and
to pass through all lights with wavelength longer than 488 nm
from the sample to the dichroic mirror inside of the DC2.
6. Tune the Micro-Manager software (ver. 1.4) to control fluores-
cence light source shutters and deliver alternating excitation
lights (340 and 380 nm) for Fura2 ratiometric imaging (Fig. 1a).
7. Check the optical settings to ensure (a) transmission red light
whose wavelength is longer than 605 nm is passed to the sam-
ple (Fig. 1f) and (b) emitted lights whose wavelength is longer
than 605 nm is guided to the high-speed camera and that
shorter than 605 nm to the fluorescence camera (Fig. 1a).
3 Methods
3.1 T Cell 1. Naïve CD8+ T cells were isolated from a mouse spleen accord-
Preparation ing to an Emory University IACUC-approved protocol.
2. Briefly, a midline incision is made on a sacrificed OT1 trans-
genic mouse. Skin is retracted and the spleen, located above
the peritoneum, is removed.
3. The spleen is placed in a 15 ml conical tube immersed with

3–5 ml 1× HBSS, depending on the size of the spleen.
4. Place a cell strainer in a 6 cm Petri dish, and pour the spleen
with HBSS into the strainer.
5. Remove the cylinder within the syringe and use the soft rubber
head to grind the spleen in the cell strainer.
6. Mix the grinded mixture with 1× HBSS to a final volume of
10 ml. Centrifuge at 500 × g for 5 min at room temperature
(RT). Discard the supernatant.
7. Follow the protocol of the erythrocyte lysing kit to remove all
red blood cells.
8. Follow protocol of mouse CD8+ T cell enrichment kit to
selectively isolate CD8+ T cells. Use 5 ml tube that fits the
EasySep magnet (Stemcell Technologies).
9. Take 10 μl of purified cell solution out to determine cell
concentration.
10. Centrifuge cells at 500 × g for 5 min at RT. Resuspend cells in
cell culture media to a concentration of 1 million cells/ml.
3.2 RBC Preparation 1. Recruit a healthy donor according to a protocol approved

from Whole Blood by the Institutional Review Board at the Georgia Institute of
Technology.
2. Draw 5 ml of blood from the median cubital vein into a 10 ml
tube containing EDTA and gently mix the blood with EDTA
immediately and thoroughly to avoid clotting.
3. Transfer the blood to a 50 ml centrifuge tube, add 10 ml of
cold, and sterile Histopaque-1077 to the bottom of the tube
and centrifuge for 5 min at 300 × g, 4 °C. Discard supernatant
with sterile plastic transfer pipette (see Note 3).
4. Add 10 ml cold and sterile PBS, centrifuge for 5 min at 300 × g,
4 °C, and then remove the supernatant. Repeat twice.
5. Wash with cold and sterile EAS45 (5 min at 300 × g, 4 °C)
twice. During the last washing, move RBCs into a new sterile
15 ml tube with 10 ml EAS45. Store at 4 °C for up to 2 h until
ready for biotinylation.
3.3 Functionalize 1. Remove all supernatant by centrifuging at 300 × g for 5 min at

RBC with Biotin 4 °C twice.
and pMHC 2. Use 10 μl of solid RBC pellet for each biotin concentration in
for Micropipette one vial. Add 90 μl 1× PBS for 10% hematocrit.
Adhesion Assay 3. Spin down RBC at 300 × g for 5 min at 4 °C.
4. Make fresh biotin-X-NHS dilutions according to the manufac-
turer’s instructions.
5. Mix isolated RBCs in titrated biotin-X-NHS solutions and

vortex immediately. Incubate at pH 7.2 for 30 min at RT. For
the specific details on mixture ratios, please refer to [8].
6. Wash each vial three times to remove the biotin-X-NHS with
800 μl of EAS45 for 2 min at 300 × g.
7. Resuspend the RBCs in EAS45 (final volume is 100 μl) for
storage (see Note 4).
8. Mix equal amounts (10 μl) of RBCs with streptavidin of the
saturating concentration (SA, 2 mg/ml), vortex immediately,
and incubate on rotator for 30 min at 4 °C.
9. Wash three times with 200 μl of EAS45 for 2 min at 300 × g.
10. Mix equal amounts (10 μl) of SA-coated RBCs with pMHC
solution (20 μg/ml), vortex immediately, and incubate on
rotator for 30 min at RT.
11. Wash two times with 200 μl of EAS45 (+1% BSA) for 2 min at
300 × g.
12. Resuspend in 10 μl of EAS45 (+ 1% BSA) and store at 4 °C or
on fresh ice until ready for use in micropipette.
3.4 Site Density 1. Incubate protein-bearing cells (i.e., T cells and pMHC-coated
Measurement RBCs) with saturating concentrations of primary mAbs (i.e.,
by Fluorescence- anti-Vα TCR clone β20.1 and anti-H-2Kb mAbs) at 10 μg/ml
Activated Cell Sorting concentration in 100 μl of FACS buffer at 4 °C for 30 min
(FACS) (see Note 5).
2. In separate vials, incubate cells with irrelevant isotype-matched
antibodies for control.
3. Analyze the fluorescent intensities of the prepared samples as
well as QuantiBRITE PE standard beads by the BD LSR II
flow cytometer.
4. Plot the fluorescence histograms of calibration beads (Fig. 3a,
pink) together with those of T cells (Fig. 3a, blue) and pMHC-
coated RBCs (Fig. 3a, green).
5. Specific anti-Vα TCR (blue) and anti-H-2Kb (green) mAb stain-
ings are shown in solid curves, and irrelevant isotype-matched
control antibody staining is shown in dotted curves in Fig. 3a.
6. Calculate Log10 for the geometric mean fluorescent intensity
(FI) of each peak value of four calibration bead histograms
from Fig. 3a (pink circles) and for the lot-specific PE molecules
per bead (from the manufacturer).
7. Build a linear regression model of Log10 PE molecules per
bead against Log10 fluorescence plotted (Fig. 3b). For T cells
the Log10 FI (y) values equal 4.22 (Fig. 3b, blue solid circle)
and 2.23 (Fig. 3b, blue open circle) for specific mAb and con-
trol antibody, respectively.
a b
400
5
300
4
Count
Log FI
200
3
100
0 2
102 2 3 4 5
0 103 104 105
Log PE/Bead
PE-A
Fig. 3 Determination of protein site density on cells/beads. (a) Fluorescence histograms of calibration beads
(pink) together with cells of interest (T cell in blue and pMHC-coated RBC in green). Specific primary mAb
staining is shown in solid curves and irrelevant isotype-matched control Ab staining is shown in dotted curves.
(b) Process of density quantification. Log10 was calculated for the geometric mean fluorescent intensity (FI) of
each peak value of four calibration bead histograms from panel (a) (pink circles) and for the lot-specific PE
molecules per bead (from the manufacturer). A linear regression of Log10 PE molecules per bead against Log10
fluorescence is plotted. By substituting Log10 FI of T cell populations into the regression model, the site densi-
ties for measured proteins will be determined. For T cells in this case, the Log10 FI are 4.22 (blue solid circle)
and 2.23 (blue open circle) for specific mAb and control Ab, respectively. The total number of TCR on T cells
was calculated as 16,400. Surface density was calculated to be 145 molecules/μm2, using 6 μm as the T cell
diameter. Similarly, the site density of pMHC on RBCs (green) was calculated to be 157 molecules/μm2
8. Solve the linear equation for x (values are plotted as green and
blue circles in Fig. 3b). x = Log10 PE/cell and, as PE:mAb ratio
was 1:1, the estimated total number of TCR on a T cell is cal-
culated as 16,400. Surface site density is calculated to be 145
molecules/μm2, using 6 μm as the T cell diameter.
9. Measure the density of pMHC on RBCs similarly using anti-
H-2Kb mAb, which equals 157 molecules/μm2.
3.5 Preparation 1. Cut long capillary glass tubes with a glass cutter into short
for Micropipette pieces of around 3 in. in length. Mount one piece onto the
and Cell Chamber Flaming/Brown pipette puller, click the “Pull” button so that
the middle of the capillary will be heated by the machine and
the capillary will be pulled on the two ends to make two capil-
laries with sharp tips (raw pipettes) (see Note 6).
2. Mount a raw pipette onto the MF-900 microforge and make a
micropipette by repeatedly melting and pulling off the very
top part to obtain the desired tip orifice. The examples micro-
pipette orifice sizes are 1.0–2.0 μm for a RBC in the micropi-
pette adhesion setup, 2.0–2.4 μm for a RBC in the BFP setup,
~1.5 μm for a bead, and ~2–4 μm for a T cell (see Note 7).
3. Build a new cell chamber for each experiment on a homemade

chamber holder, with two pieces of copper/aluminum squares
and a handle that links them (Fig. 1d).
4. Cut a 40 mm × 22 mm × 0.2 mm coverslip by a glass cutter to
make two 40 mm × 11 mm × 0.2 mm pieces (upper and lower
coverslips).
5. Glue two ends of the upper coverslip to the top of the chamber
holder with grease, and then similarly glue lower coverslip to
the bottom of the chamber holder (Fig. 1d), like making a
sandwich.
6. Inject ~200 μl of chamber media (e.g., L15) in between the
two coverslips. Make sure the buffer attaches to the upper cov-
erslip. Gently rotate and shake the chamber to let the buffer
connect the chamber-adaptor interfaces between both ends
along the longitudinal axis of the chamber (Fig. 1f). The buffer
zone will look like a dumbbell shape.
7. Carefully inject mineral oil to surround the buffer area and seal
the buffer from the atmosphere.
8. Inject concentrated two cell species (pMHC-coated RBCs and
T cells) at different locations within the buffer area.
3.6 Micropipette 1. Turn on the microscope in bright-field and place the chamber
Assembly onto the microscope platform.
Before an Experiment 2. Use a micro-injector to fill three micropipettes with experi-
mental chamber media.
3. Assemble two micropipettes (probe and target) to their respec-
tive micropipette holders, and then mount to respective hold-
ing stages (Fig. 1c, left, probe, to grab a RBC; right, target, to
grab a T cell) (see Note 8).
4. Push the micropipettes toward the chamber so that their tips
enter the chamber buffer area.
5. Adjust the positions of the micropipettes and find them under
the microscope field of view (Fig. 2a).
6. Move around the chamber holder stage to find the rough loca-
tions of injected two cell species one by one.
7. Adjust the position of the probe and target micropipettes by
twisting the knobs of the corresponding holding stages to let
the tips of the micropipettes approach their respective cell spe-
cies (RBCs and T cells).
8. Adjust the aspiration pressure inside the micropipettes to
aspire a RBC and a T cell for the two micropipettes, respectively,
and then put them into the same field of view (Fig. 2b)
(see Note 9).
Fig. 4 Adhesion frequency curves of the OT1 TCR on naïve OT1 T cells interacting
with OVA and G4 presented by H2-Kbα3A2 on RBCs measured by our micropi-
pette adhesion assay (reproduction of Fig. 2c from Huang et al., Nature, 2010).
Each cell pair was tested repeatedly at given contact duration tc to estimate an
adhesion frequency Pa, and 3–5 cell pairs were tested for each tc to calculate a
mean Pa ± s.e.m. The data (points) were fitted (color-matched solid curves) by the
model Pa = 1 − exp {mrmlAcKa[1 − exp (−kofftc)]},where mr and ml are the respec-
tive surface densities of TCR on OT1 T cells and pMHC on RBCs. The goodness-
of-fit was indicated by the R2 values. Color-matched dotted curves represent 95%
confidence intervals of the best-fit curves obtained by bootstrapping
3.7 Adhesion 1. Use computer-programmed piezoelectric linear actuator to drive

Frequency Assay a pMHC-coated RBC in and out of a normal-vector contact with
a T cell under a controlled contact area and time (Fig. 2c–e).
2. Ensure that the piezoelectric actuator travels at least 6 μm to
contact cells, in order to clearly observe RBC elongation.
3. Detect adhesion events by observing RBC deflection upon cell
separation (Fig. 2c).
4. Repeat the contact–retraction cycle 50–100 times for a given
contact time.
5. Record the observed adhesion events by adding “1” for adhe-
sion or “0” for no adhesion in a column of Excel spreadsheet
(see Note 10).
6. Record the adhesion frequency versus contact time curve using
at least three different cell pairs for each contact time to obtain
a mean and s.e.m. (Fig. 4).
7. Record the nonspecific binding curve for control by using RBCs
coated with irrelevant ligands (e.g., BSA) and/or blocking the
ligands or receptors using their specific functional blockade
mAbs. Specific adhesion frequency at each contact time point can
be calculated by removal of the nonspecific adhesion frequency.
3.8 2D Kinetics 1. Fit the specific adhesion frequency Pa versus contact time tc
Analysis data (Fig. 4) by a probabilistic model [6] that describes a
second-order forward and first-order reverse, single-step inter-
action between a single species of receptors and a single species
of ligands

{ ( )}
Pa = 1 - exp mrml Ac K a éë1 - exp -koff t c ùû

where Ka is the 2D effective binding affinity, koff is the off-rate,
mr and ml are the respective receptor and ligand densities mea-
sured in Subheading 3.4, and Ac is the contact area. The curve
fit has two parameters, AcKa and koff, as Ac and Ka are lumped
together and called collectively as effective 2D affinity. Its
product with the off- rate is the effective 2D on-rate:
Ackon = AcKa × koff.
2. The specific adhesion frequency Pa is calculated by subtraction
of the nonspecific adhesion fraction (Pn) from the total mea-
sured adhesion (Pt):
Pt - Pn
Pa =
1 - Pn
3.9 Bioengineer 1. Obtain 8–10 μl (one drop) of blood by finger prick lancet
the RBC device and add to 1 ml of the carbonate/bicarbonate buffer
into the Biomembrane (pH 8.5–9). Gently vortex or pipette the mixture and centri-
Force Probe (BFP) fuge for 1 min at 900 × g. Discard supernatant and wash once
more.
2. In a small beaker, weigh 3.5–4 mg of biotin-PEG3500-NHS
linker. Dissolve it in the carbonate/bicarbonate buffer to make
the final concentration 6 mg/ml (see Note 11).
3. Mix 171 μl of carbonate/bicarbonate buffer, 10 μl of RBC
pack, and 1049 μl of biotin-PEG3500-NHS linker solution
and incubate at RT for 30 min.
4. Wash the RBC with carbonate/bicarbonate buffer once and
then with N2–5% buffer (pH 7.2–7.4) twice.
5. Dilute nystatin into N2–5% buffer to make a final concentra-
tion of 40 μg/ml.
6. Mix 5 μl of biotinylated RBC with 71.4 μl of nystatin solution
and incubate for 1 h at 0 °C on ice (see Note 12).
7. Wash twice with N2–5% buffer and store with N2–5%
buffer + 0.5% BSA in the refrigerator (4 °C).
3.10 Glass Bead 1. Weigh out 50 mg of glass bead powder and resuspend them in
Silanization 500 μl of DI water.
and Thiolation 2. Mix 0.5 ml of 30% H2O2 with 9.5 ml of DI water in a 50 ml
beaker, then add 1 ml of concentrated NH4OH, and bring this
solution to a boiler on a hot plate.
3. Add the glass beads to the boiling solution and continue to

boil for another 5 min. Gently swirl the solution every minute.
After boiling, transfer 5 ml of this hot bead suspension into a
micro-centrifuge tube and top up with RT DI water.
4. Centrifuge at 3500 × g for 5 min, remove, and discard the
supernatant.
5. Transfer another 5 ml of hot bead suspension and add to the
washed beads, top up with more DI water, mix well, and cen-
trifuge again.
6. Repeat this procedure until about 50 ml of DI water is used.
7. Transfer the bead suspension into a 1 ml vial. Repeat washing
the beads with methanol by centrifugation at 17,000 × g for
5 min, three times, and finally resuspend the beads in 1 ml of
100% methanol.
8. In a 50 ml centrifuge tube, add the 1 ml bead suspension,
45.6 ml of methanol, 0.4 ml of acetic acid, 1.85 ml of DI
water, and 1.15 ml of 3-mercaptopropyltrimethoxysilane
(MPTMS), and then incubate at RT for 3 h.
9. After the reaction, remove all supernatants by washing once
with fresh methanol, and resuspend the beads in 500 μl of
methanol. Evenly divide this concentrated glass bead suspen-
sion into a set of 20 dry and clean glass vials with screw caps.
10. Evaporate off the methanol by using a jet of dry argon and
slowly rotate the vials so as to make a thin layer of dry beads on
the sides of each vial.
11. Place the vials of beads (referred to as “MPTMS beads” below)
into a preheated drying oven at 120 °C for 5 min and then take
out and quickly place the cap(s) loosely on. Place the vials in a
glass vacuum desiccator filled with drying agents on the bot-
tom and keep the vials desiccated until cooled.
12. Purge the vacuumed desiccator with dry argon to bring the des-
iccator to normal atmospheric pressure. Remove the desiccator
lid and quickly retighten the cap(s) on the vials. Seal the vials
with Parafilm and store them at RT in a dry dark storage box.
3.11 Bead 1. Take one vial of dry MPTMS beads and wash once with phos-
Functionalization phate buffer (pH 6.5–6.8). Resuspend into 50 μl of phosphate
buffer and store at 4 °C.
2. Take a certain volume (e.g., 2.5 μg) of the protein (e.g.,
pMHC) stock and mix with equal volume of carbonate/
bicarbonate buffer to make solution 1. The volume depends
on the desired final protein density on the beads’ surface.
3. In a small beaker, weigh 2–3 mg of MAL-PEG3500-NHS
linker and dissolve it with carbonate/bicarbonate buffer to
reach a final concentration of 0.231 mg/ml.
4. Mix solution 1 with an equal volume of this linker solution.

Incubate the mixture at RT for 30 min to make solution 2
(see Note 11).
5. (a) To functionalize beads with the protein of interest alone
(i.e., pMHC), mix 5 μl of MPTMS beads with solution 2 and
then add phosphate buffer to make a final volume of 250 μl.
(b) To functionalize beads with the protein of interest along
with streptavidin (SA) for RBC attachment (i.e., pMHC + SA),
mix 5 μl of MPTMS beads with solution 2 and 5 μl of 4 mg/
ml streptavidin−maleimide (SA-MAL) solution, and then add
phosphate buffer to make a final volume of 250 μl (Fig. 5c). (c)
To functionalize beads without any protein but SA alone, mix
5 μl of MPTMS beads with 5 μl of 4 mg/ml SA-MAL solution
and then add 140 μl phosphate buffer.
6. Incubate the beads overnight at RT and wash three times with
phosphate buffer. Finally, resuspend into 100 μl of phosphate
buffer and store at 4 °C for following experimental use.
7. If the protein of interest has been biotinylated, further mix 5 μl
of SA-coated beads (see Subheading 3.11, step 5c) with the pro-
tein (volume depending on the desired coating density) and then
add phosphate buffer to make the final volume to be 100 μl.
Then incubate the beads for 3 h at RT and wash three times with
phosphate buffer. Finally, resuspend into 100 μl of phosphate
buffer and store at 4 °C for following experimental use.
3.12 BFP Assembly 1. Inject the three concentrated cell/bead suspensions (biotinyl-
Before an Experiment ated RBCs, T cells, and beads coated with pMHC + SA) at
different locations within the buffer area.
2. Follow a similar procedure to Subheading 3.5 to assemble
the probe, target, and helper micropipettes to their respective
holding stages (Fig. 1f).
3. Adjust the positions of all three micropipettes and find them
under the microscope field of view (Fig. 5a).
4. Move around the chamber holder stage to find the rough loca-
tions of injected three bead/cell species one by one.
5. Adjust the position of the micropipettes (probe, target, and
helper) by twisting the knobs of the corresponding holding
manipulators to let the tips of the micropipettes approach their
respective cell/bead species (RBCs, T cells, and beads coated
with pMHC + SA).
6. Adjust the aspiration pressure inside the micropipettes to aspire
a bead or a cell for all three micropipettes, and then put them
into the same field of view (see Note 13).
7. Align the probe bead and RBC and then carefully approach the
probe bead to the apex of the RBC. Contact solidly and then
Fig. 5 Biomembrane force probe setup (reproduction of Fig. 1, S1 from Liu et al., Cell, 2014). (a, b) Micrographs
of BFP setting in an experimental chamber. (a) Micropipette assembly showing the probe pipette (left), target
pipette (upper right), and helper pipette (lower right). (b) Probe bead placement. A probe bead was manipulated
by a helper pipette and attached to a RBC apex to form a force probe. (c) Video micrograph depicting a force
probe (left) and a target T cell (right) aspirated by their respective micropipettes. The stationary force probe
consists of a swollen RBC and an attached ligand-bearing bead. The receptor-bearing T cell (target) is mounted
to a piezo actuator aligned opposite the probe. The region of interest (ROI) for tracking the bead edge is high-
lighted in green. The edge tracker is indicated in a blue line. The insert depicts the ligand (pMHC, bead side)
and receptor (TCR, T cell side) pair on the two opposing surfaces in the area marked in orange. (d) The intensity
profile of the bead edge in (c). The ROI in the x-direction is plotted as x-axis (in pixel number) versus the light
intensity (in gray scale value) averaged by binning 30 pixels along the y-direction. (e) The deflection of the RBC
and the position of the bead and the target (T cell) in a test cycle of force-clamp assay. The vertical and hori-
zontal dashed lines indicate the zero-force position of the RBC apex and the time course, respectively. The line
edge tracker of the RBC deformation is shown in blue in each panel. The same, yet fewer, steps are adopted
in thermal fluctuation assay (which lacks the step of “dissociate”)
retract slightly. Adjust the pressure of the helper micropipette

to gently blow the bead away, so that it will be left glued onto
the RBC apex (Fig. 5b). Move the helper micropipette away
and align the target and probe bead (Fig. 5c).
8. On the BFP control program (LabVIEW), in the vision field
window, use the tools in the program to measure the respective
radii of the probe micropipette (Rp), the RBC (R0), and the
circular contact area between the RBC and probe bead (Rc).
9. Enter the desired RBC spring constant into the program
(see Note 14), which will return a required aspiration pressure
in units of centimeter of water.
10. Adjust the height of the water manometer (Fig. 1c) that con-
nects probe micropipette holder until the required aspiration
pressure is reached.
11. Draw a horizontal line across the RBC apex (Fig. 5c), which
will yield a curve in the adjacent window indicating the bright-
ness (gray scale value) of each pixel along this line. Drag the
threshold line to be at around half the depth of the curve
(Fig. 5d) (see Note 15).
12. Select the desired experiment mode: adhesion frequency assay,
force-clamp assay, and thermal fluctuation assay; set the para
meters as desired (e.g., impingement force = 20 pN, loading
rate = 1000 pN/s, contact time = 1 s, clamping force = 20 pN
(for force-clamp assay) or 0 pN (for thermal fluctuation assay),
spring constant κ = 0.25–0.3 pN/nm (force-clamp assay [19,
20, 38]) or 0.15 pN/nm (thermal fluctuation assay [34])).
13. Start the program to begin data collection in repeated BFP
cycles.
3.13 BFP Test Cycles 1. A BFP experiment is composed of repeated test cycles that are
and Experimental performed sequentially. The fast-speed camera continuously
Modes monitors RBC deformations by tracking the probe bead edge,
deriving each cycle’s “force vs. time” signal.
2. At the beginning of a BFP cycle, the target T cell is driven to
approach, impinge, and contact with the pMHC-coated probe
bead by the program-controlled piezo actuator. The contact is
signified by the RBC indentation in the monitoring program
(Fig. 6a–c).
3. At the end of the contact duration, the piezo actuator retracts
the target T cell away from the probe to a preset position of
desired force or separation distance.
4. (a) In the case of no adhesion, no tensile force is generated by
the target retraction (Fig. 6a). The target will return to the
original position and begin the next test cycle. (b) In the pres-
ence of an adhesion, which is signified by an axial deflection of
the RBC toward the target, the retracting target will pull on
the probe until rupture (Fig. 6b) or until the desired force/
distance is reached (Fig. 6c), after which the force will be
clamped and exerted on the molecular bond.
5. This approach–impinge–contact–retract–clamp–dissociate test
cycle (Fig. 5e) will be repeated many times to acquire an
ensemble of data for statistical analysis [34, 38, 39].
6. For adhesion frequency assay (Fig. 4), record which cycles
contain an adhesion event (Fig. 6b, c) and which do not
(Fig. 6a), and summarize to yield an average adhesion fre-
quency. Also, the rupture force of each adhesion event, which
is the peak value of the linearly ramped force before bond rup-
ture, is collected.
7. For force-clamp assay (Fig. 6c), properties of each lifetime event
including the average force and lifetime elapse will be recorded
with the sequence number as well as the starting time and the
ending time of the lifetime event, which will allow one to draw
a cumulative lifetime curve. After a sufficient amount of lifetime
events have been collected under a range of forces, they can be
put together and grouped into different force bins, which will
produce an average lifetime in each force bin, and altogether
yield an “average lifetime vs. force” curve (Fig. 6d).
8. For thermal fluctuation assay, instead of retracting the cell to
generate a tensile force as in the force-clamp assay, the retr
action stops when the impingement force just vanishes and
receptor–ligand pairs are allowed to interact via BFP thermal
fluctuation. By analyzing the displacement and the standard
deviation of the bead movement (Fig. 6e, f), bond association
and dissociation events are identified from reduction and
resumption, respectively, of these fluctuations (see Note 16).
Thermal fluctuation assay measures bond lifetimes at zero
force, which are measured as the duration from fluctuation
reduction to resumption (Fig. 6f) (see Note 17). The interval
between two bond events reflects the reciprocal on-rate, while
the duration of the association events reflects the reciprocal
off-rate under zero force [34, 43, 44].
3.14 Concurrent BFP 1. Dissolve Ca2+ indicator Fura2-AM in DMSO at a stock con-
and Calcium Imaging centration of 10 mM.
2. Pre-load the T cells with Fura2 at a final concentration of
2 μM, and then incubate for 30 min at RT.
3. Wash once and then keep this fluorescently loaded cell suspen-
sion in the dark until use.
4. For concurrent fluorescence imaging, turn on the excitation
light source and the fluorescence camera, which are controlled
by a separate program (Micro-Manager, ver. 1.4).
Fig. 6 Biomembrane force probe data analysis (reproduction of Figs. 1, 2 from Chen et al., JoVE, 2015). (a–c)
Representative raw data (force-time traces) of a no-adhesion event (a), an adhesion-rupture event (b), and an
adhesion lifetime (c). Various phases of the cycle and the corresponding phases, respectively, are marked
in each panel. The force (y-axis) is derived from tracking the position change of the probe bead, as shown in
Fig. 5c. (a) No adhesion: the compressive (negative) force in the contact phase returns to zero upon retraction.
(b) Adhesion-rupture: a tensile (positive) force pulls via the receptor–ligand bond to elongate the RBC, which
ruptures (the instant is marked by *) during the retraction phase. (c) Adhesion lifetime: the bond persists until
the clamping force is reached (the instant is marked by *) and dissociates thereafter. (d) “Average lifetime vs.
force” curve of OT1 T cell interacting with its agonist OVA (green) and antagonist G4 (blue). The pooled data
are grouped into different force bins, and the mean ± s.e.m. of bond lifetimes is plotted versus force. (e, f)
Representative data of displacement vs. time (e) and 100-point sliding standard deviation of displacement vs.
time (f) curves of the thermal fluctuation assay. This assay differs from the force-clamp assay for bond lifetime
measurement in that the T cell retraction stopped at the zero-force position (e), so bond formation is not
detected by a tensile force (c). Instead, it is detected by the reduction of the sliding standard deviation below
a threshold (f). Further, bond dissociation is not detected by a sudden drop of the tensile force to zero (c), but
by the resumption of the sliding standard deviation above the threshold (f)
5. On the program, select the parameters for the fluorescence

imaging, including gain, exposure, excitation channels (in this
case, 340 and 380 nm light), etc.
6. Follow all preparations in the BFP experiment protocol,
including aligning the probe and the target, which will allow
for visualization of the target cell live fluorescent image excited
by 340 nm or 380 nm wavelength excitation light.
7. Use the sectioning tool to roughly section the area within
which the cell will stay during the entire recording period
(see Note 18).
8. Click on “Record” to allow the 340 and 380 nm light to alter-
nately excite the intracellular fluorescence dye (Fura2), and a
pair of corresponding fluorescence images will be alternately
recorded about once every second.
9. Simultaneously click on “Start” in the LabVIEW program to
begin the BFP experiment for analysis of molecular interaction
and the fluorescence imaging experiment to monitor intracel-
lular calcium signaling.
10. The system will produce a raw data file for the receptor–ligand
binding (Fig. 8a) and a series of fluorescent images in .tiff
format for the calcium signals (Fig. 8b).
3.15 Post- 1. Adjust the intensity threshold until the fluorescence images
Experiment Calcium show a clear contour of the cell in both 340 and 380 nm chan-
Imaging Analysis nels without background noise (Fig. 7a, b) (see Note 19).
2. Review the intracellular Ca2+ signal frame by frame with a
pseudo-color indicating the intensity level (Fig. 8b), which is
derived based on the intensity ratio of 340 nm/380 nm,
to generate the “normalized Ca2+ intensity vs. time” curve
(Fig. 8c).
3. Produce a movie that displays the fluorescence level second by
second from the pseudo-color fluorescence images.
4. Take the peak value of Ca2+ flux as the signaling readout to
seek its best predictor among various kinetic parameters,
including the number of adhesions, the force amplitude of
the binding, the average lifetime, the longest lifetime and the
cumulative lifetime of the bindings, etc.
5. Shown in Fig. 8 is an example of simultaneously recorded
individual bond lifetimes (where force was applied) and their
accumulation, together with the corresponding Ca2+ signal
curve. A systematic mathematical analysis of such data col-
lected from many individual cells revealed that the best correla-
tion of Ca2+ signaling intensity is lifetimes accumulated in the
first minutes of repeated TCR-pMHC interactions (refer to
ref. [27] for scientific details).
Fig. 7 Representative Ca2+ images excited at two wavelengths (reproduction of Fig. 5 from Chen et al., JoVE,
2015). (a, b) Correct image recognition of a T cell (indicated in red) in 340 nm (a) and 380 nm (b) channels
based on point-to-point screening using a properly assigned intensity threshold. (c) Inability to recognize the
fluorescence image of a T cell (indicated in red) in the 340 nm excitation channel, due to poor Fura2 loading
Fig. 8 Intracellular Ca2+ level (relative Fura2 ratio) and lifetimes of an OT1 T cell, during repeated touching with
an OVA-coated bead over 300 s (reproduction of Fig. 6 from Chen et al., JoVE, 2015). (a) A force curve showing
a sequence of non-adhesion, rupture force, and lifetime events generated by repeated contacts over time.
(b) Epi-fluorescence pseudo-color images of intracellular Ca2+ signals in the T cell at different time points. The
normalized Ca2+ level is indicated by the pseudo-color scale on the right. (c) Superimposition of the Ca2+ signal
curve (red) and the cumulative lifetime curve (yellow) on the same time course. The Ca2+ curve was plotted
based on the Ca2+ imaging. A Ca2+ flux is signified by a sharp elevation in the normalized Fura2 ratio. The time
when Ca2+ reaches the peak is indicated by a dashed line. The onset time of each lifetime event is marked on
the cumulative lifetime curve (solid triangle)
4 Notes
1. For mounting of micropipette holding sets, home-designed

mechanical parts and adaptor fabrications are built using CNC
machine tool in the machine shop of Georgia Tech’s Woodruff
School of Mechanical Engineering. All parts are customized
according to the computer-aided design (CAD) using 3D
CAD software (SolidWorks, Version 2012 SP5).
2. All control programs were developed using LabVIEW software
(National Instruments Version 2009) by Dr. Lining Ju. The
LabVIEW driver for the 1D piezo actuator was provided by
the manufacturer (Physik Instrumente).
3. The Histopaque-1077 sequesters all white blood cells in the
plasma/Histopaque-1077 interface. Process blood sample as
fast as possible so that the white blood cells do not pellet and
the repeated washing removes platelets. All the following steps
except centrifugation should be done under the hood to keep
the preparation sterile.
4. The biotinylated RBC aliquots can be store at 4 °C for further
use of up to 2 months.
5. If the primary antibodies are not fluorescently labeled, the
sample needs to be further incubated with fluorescently con
jugated secondary antibodies according to manufacturer’s
instructions.
6. According to the manual from the Flaming/Brown micropi-
pette puller, the desired morphology of the raw pipette has
6–8 mm taper, 0.1–0.5 μm tip, and 60–150 MΩ resistance.
7. The forge has a glass sphere that melts when heated. The heat-
ing and cooling of the glass sphere are controlled by the user
stepping on and releasing a foot switch pedal. When the glass
sphere melts, the tip of the raw pipette can be inserted inside.
Let the sphere cool down and become solid while the inserted
part stays inside. When pulled, the pipette will break from out-
side and leave its tip inside the sphere.
8. Take off the micropipette holder and put it at a lower position
to allow water to drip from the tip. Quickly insert the micropi-
pette into the holder tip and make sure no air bubbles get into
the micropipette during the connection, and then tighten the
holder screw.
9. Move around the chamber holder stage to find an open space
away from the colonies of injected cell/bead species where the
experiment will be performed. Turn on the software that con-
trols the normal speed camera to have a microscopic view on a
computer screen.
10. You may use a recording device, e.g., digital media or v ideotape,
to record microscopic images and facilitate the adhesion events
counting.

11.
MAL-PEG3500-NHS and biotin-PEG3500-NHS linkers
should be stored dry at −20 °C. For experimental preparation,
take it out from the freezer 30 min before reaction and leave at
room temperature to warm up before opening. Spooning out
these linker powder needs to be accomplished as fast as possi-
ble so that the remanent reagents inside the bottle will have
minimal exposure to the open air. After spooning out a portion
of the powder, place the bottle with loosened cap in a glass
vacuum desiccator filled with the drying desiccants on the bot-
tom and vacuum for 5 min, and then fill the desiccator with
argon. Tighten the cap and take the bottle out. Seal the bottle
with Parafilm (PM996, Bemis). Then place it into a container
filled with desiccant on the bottom and store in −20 °C.
12. The optimal incubation time depends on the RBC quality of
the donor. If the donor’s RBCs are easy to lyse, reduce the
incubation time and vice versa.
13. Move around the chamber holder stage to find an open space
away from the colonies of injected bead/cell species where the
experiment will be performed. Switch the microscope observa-
tion method to the “camera” mode and visualize the BFP
micrograph in the LabVIEW program on the computer screen
(Fig. 2a–c).
14. The spring constant of the BFP (κ) is determined by Evans’
p Rp Dp
model [7, 33]: k =
( ) (
1 - Rp / R0 ln éë 4R02 / Rp Rc ùû)
where Δp is the pressure difference aspired at probe pipette
tip. It follows from Hooke’s law that the binding force, F, can
be quantified by the product of spring constant and displace-
ment of the probe bead (Δd), i.e., F = κ × Δd (Fig. 5e, step 5).
Since we can adjust κ from 0.1 to 1 pN/nm, the BFP can
apply a very wide range of force from 1 to 1000 pN with a
very wide range of force loading rates from 10 to 104 pN/s.
15. The minimum point on the brightness curve below the thresh-
old line indicates the position of the bead edge; thus, only one
local minimum is allowed (Fig. 5d). If two or more local min-
ima are present, it indicates the image is not optimal (likely due
to the image being out of focus, or an underperformed align-
ment between the probe bead and the RBC).
16. Because of BFP’s high resolution and soft spring constant,
thermal fluctuation assay gives better measurement of koff than
adhesion frequency assay. To ensure this advantage, the BFP
spring constant is set to κ = 0.15 pN/nm [34].
17. One can use the thermal fluctuation level of the clamping
phase in the “force vs. time” signal to help distinguish associa-
tion and dissociation of a bond, since bond association leads to
a decrease in the thermal fluctuation amplitude (Fig. 6f).
18. Due to the use of the approach–contact–retraction cycle, the
cell will be moving forward and backward repetitively; thus,
the sectioned area should be much larger than the cell itself.
19. Take the ratio of the background subtracted fluorescence
images and calculate the relative intracellular Ca2+ level. For best
results, the specific fluorescence signals for 340 and 380 nm
excitation should be easily distinguished from the media back-
ground or background of cells not labeled with Fura2 AM
(Fig. 7a, b). If it is difficult to detect the cells in either channel
with non-negligible background noise, the cell labeling likely
needs to be improved (Fig. 7c).
References
1. Huang J, Meyer C, Zhu C (2012) T cell anti- 10. Williams TE, Nagarajan S, Selvaraj P, Zhu C
gen recognition at the cell membrane. Mol (2000) Concurrent and independent binding
Immunol 52:155–164 of Fcγ receptors IIa and IIIb to surface-bound
2. Edwards LJ, Zarnitsyna VI, Hood JD, Evavold IgG. Biophys J 79:1867–1875
BD, Zhu C (2012) Insights into T cell recogni- 11. Chesla SE (2000) The membrane anchor
tion of antigen: significance of two-dimensional influences ligand binding two-dimensional

kinetic parameters. Front Immunol 3:86 kinetic rates and three-dimensional affinity of
3. Zarnitsyna V, Zhu C (2012) T cell triggering: Fcgamma RIII (CD16). J Biol Chem 275:
insights from 2D kinetics analysis of molecular 10235–10246
interactions. Phys Biol 9:045005 12. Williams TE (2001) Quantifying the impact of
4. Zhu C, Jiang N, Huang J, Zarnitsyna VI, membrane microtopology on effective two-
Evavold BD (2012) Insights from in situ analy- dimensional affinity. J Biol Chem 276:
sis of TCR-pMHC recognition: response of an 13283–13288
interaction network. Immunol Rev 251:49–64 13. Long M, Zhao H, Huang KS, Zhu C
5. Grakoui A, Bromley SK, Sumen C, Davis MM, (2001) Kinetic measurements of cell surface
Shaw AS, Allen PM, Dustin ML (1999) The E-selectin/carbohydrate ligand interactions.
immunological synapse: a molecular machine Ann Biomed Eng 29:935–946
controlling T cell activation. Science 285: 14. Huang J, Chen J, Chesla SE, Yago T, Mehta P,
221–227 McEver RP, Zhu C, Long M (2004)
6. Chesla S, Selvaraj P, Zhu C (1998) Measuring Quantifying the effects of molecular orienta-
two-dimensional receptor-ligand binding kinet- tion and length on two-dimensional receptor-
ics by micropipette. Biophys J 75:1553–1572 ligand binding kinetics. J Biol Chem 279:
7. Chen W, Zarnitsyna VI, Sarangapani KK, 44915–44923
Huang J, Zhu C (2008) Measuring receptor– 15. Wu L, Xiao B, Jia X, Zhang Y, Lu S, Chen J,
ligand binding kinetics on cell surfaces: from Long M (2007) Impact of carrier stiffness and
adhesion frequency to thermal fluctuation microtopology on two-dimensional kinetics of
methods. Cell Mol Bioeng 1:276–288 P-selectin and P-selectin glycoprotein ligand-1
8. Zarnitsyna VI, Zhu C (2011) Adhesion fre- (PSGL-1) interactions. J Biol Chem 282:
quency assay for in situ kinetics analysis of 9846–9854
cross-junctional molecular interactions at the 16. Zhang F, Marcus WD, Goyal NH, Selvaraj P,
cell-cell interface. J Vis Exp 2:e3519 Springer TA, Zhu C (2005) Two-dimensional
9. Williams TE, Selvaraj P, Zhu C (2000) kinetics regulation of alphaLbeta2-ICAM-1
Concurrent binding to multiple ligands: kinetic interaction by conformational changes of the
rates of CD16b for membrane-bound IgG1 alphaL-inserted domain. J Biol Chem 280:
and IgG2. Biophys J 79:1858–1866 42207–42218
17. Fiore VF, Ju L, Chen Y, Zhu C, Barker situ TCR-peptide-bound MHC class II kinetics
TH (2014) Dynamic catch of a Thy-1– determine functions of CD4+ T cells.
α5β1+syndecan-4 trimolecular complex. Nat J Immunol 195:3557–3564
Commun 5:4886 29. Rosenthal KM, Edwards LJ, Sabatino JJ Jr,
18. Rosetti F, Chen Y, Sen M, Thayer E, Azcutia V, Hood JD, Wasserman HA, Zhu C, Evavold
Herter JM, Luscinskas FW, Cullere X, Zhu C, BD (2012) Low 2-dimensional CD4 T cell
Mayadas TN (2015) A lupus-associated Mac-1 receptor affinity for myelin sets in motion
variant has defects in integrin allostery and delayed response kinetics. PLoS One 7:e32562
interaction with ligands under force. Cell Rep 30. Sabatino JJ, Huang J, Zhu C, Evavold BD
10:1655-1664 (2011) High prevalence of low affinity peptide-
19. Ju L, Chen Y, Zhou F, Lu H, Cruz MA, Zhu C MHC II tetramer-negative effectors during
(2015) Von Willebrand factor-A1 domain polyclonal CD4+ T cell responses. J Exp Med
binds platelet glycoprotein Ibalpha in multiple 208:81–90
states with distinctive force-dependent dissoci- 31. Jiang N, Huang J, Edwards LJ, Liu B, Zhang
ation kinetics. Thromb Res 136:606–612 Y, Beal CD, Evavold BD, Zhu C (2011) Two-
20. Ju L, Dong J-F, Cruz MA, Zhu C (2013) The stage cooperative T cell receptor-peptide major
N-terminal flanking region of the A1 domain histocompatibility complex-CD8 trimolecular
regulates the force-dependent binding of von interactions amplify antigen discrimination.
Willebrand factor to platelet glycoprotein Ibα. Immunity 34:13–23
J Biol Chem 288:32289–32301 32. Casas J, Brzostek J, Zarnitsyna VI, Hong JS,
21. Chien YH, Jiang N, Li F, Zhang F, Zhu C, Wei Q, Hoerter JA, Fu G, Ampudia J,
Leckband D (2008) Two stage cadherin kinet- Zamoyska R, Zhu C, Gascoigne NR (2014)
ics require multiple extracellular domains but Ligand-engaged TCR is triggered by Lck not
not the cytoplasmic region. J Biol Chem associated with CD8 coreceptor. Nat Commun
283:1848–1856 5:5624
22. Tabdili H, Langer M, Shi Q, Poh YC, Wang N, 33. Evans E, Ritchie K, Merkel R (1995) Sensitive
Leckband D (2012) Cadherin-dependent force technique to probe molecular adhesion
mechanotransduction depends on ligand iden- and structural linkages at biological interfaces.
tity but not affinity. J Cell Sci 125:4362–4371 Biophys J 68:2580–2587
23. Huang J, Edwards LJ, Evavold BD, Zhu C 34. Ju L, Qian J, Zhu C (2015) Transport regula-
(2007) Kinetics of MHC-CD8 interaction tion of two-dimensional receptor-ligand asso-
at the T cell membrane. J Immunol 179: ciation. Biophys J 108:1773–1784
7653–7662 35. Choi YI, Duke-Cohan JS, Chen W, Liu B,
24. Liu B, Zhong S, Malecek K, Johnson LA, Rossy J, Tabarin T, Ju L, Gui J, Gaus K,
Rosenberg SA, Zhu C, Krogsgaard M (2014) Zhu C, Reinherz EL (2014) Dynamic control
2D TCR-pMHC-CD8 kinetics determines of β1 integrin adhesion by the plexinD1-
T-cell responses in a self-antigen-specific TCR sema3E axis. Proc Natl Acad Sci USA 111:
system. Eur J Immunol 44:239–250 379–384
25. Mallis RJ, Bai K, Arthanari H, Hussey RE, 36. Chen W, Lou J, Evans EA, Zhu C (2012)
Handley M, Li Z, Chingozha L, Duke-Cohan Observing force-regulated conformational
JS, Lu H, Wang J-H, Zhu C, Wagner G, changes and ligand dissociation from a single
Reinherz EL (2015) Pre-TCR ligand binding integrin on cells. J Cell Biol 199:497–512
impacts thymocyte development before αβTCR 37. Ju L, Lou J, Chen Y, Li Z, Zhu C (2015)
expression. Proc Natl Acad Sci U S A 112: Force-induced unfolding of leucine-rich
8373–8378 repeats of glycoprotein Ibα strengthens ligand
26. Huang J, Zarnitsyna VI, Liu B, Edwards LJ, interaction. Biophys J 109:1781–1784
Jiang N, Evavold BD, Zhu C (2010) The 38. Ju L, Chen Y, Xue L, Du X, Zhu C (2016)
kinetics of two-dimensional TCR and pMHC Cooperative unfolding of distinctive mechano-
interactions determine T-cell responsiveness. receptor domains transduces force into signals.
Nature 464:932–936 eLife 5:e15447
27. Liu B, Chen W, Evavold BD, Zhu C (2014) 39. Chen Y, Liu B, Ju L, Hong J, Ji Q, Chen W,
Accumulation of dynamic catch bonds between Zhu C (2015) Fluorescence biomembrane
TCR and agonist peptide-MHC triggers T cell force probe: concurrent quantitation of recep-
signaling. Cell 157:357–368 tor-ligand kinetics and binding-induced intra-
28. Hong J, Persaud SP, Horvath S, Allen PM, cellular signaling on a single cell. J Vis Exp
Evavold BD, Zhu C (2015) Force-regulated in 4:e52975
40. Pryshchep S, Zarnitsyna VI, Hong J, Evavold

42. Zarnitsyna VI, Huang J, Zhang F, Chien
BD, Zhu C (2014) Accumulation of serial Y-H, Leckband D, Zhu C (2007) Memory
forces on TCR and CD8 frequently applied in receptor- ligand-mediated cell adhesion.
by agonist antigenic peptides embedded in Proc Natl Acad Sci U S A 104:18037–18042
MHC molecules triggers calcium in T cells.
43. Chen W, Evans EA, McEver RP, Zhu C
J Immunol 193:68–76 (2008) Monitoring receptor-ligand interac-
41. Rosette C, Werlen G, Daniels MA, Holman tions between surfaces by thermal fluctuations.
PO, Alam SM, Travers PJ, Gascoigne NR, Biophys J 94:694–701
Palmer E, Jameson SC (2001) The impact of 44. Ju L, Wang YD, Hung Y, Wu C-FJ, Zhu C (2013)
duration versus extent of TCR occupancy on T An HMM-based algorithm for evaluating rates of
cell activation: a revision of the kinetic proof- receptor-ligand binding kinetics from thermal
reading model. Immunity 15:59–70 fluctuation data. Bioinformatics 29:1511–1518
Chapter 16
Studying Dynamic Plasma Membrane Binding of TCR-CD3

Chains During Immunological Synapse Formation Using
Donor-Quenching FRET and FLIM-FRET
Etienne Gagnon, Audrey Connolly, Jessica Dobbins,
and Kai W. Wucherpfennig
Abstract
Over the last decade, advancements in the time and space resolution of microscopy technologies have
enabled dissection of the molecular events involved in T cell Immunological Synapse (IS) formation.
Using a combination of Förster Resonance Energy Transfer (FRET) and Fluorescence Lifetime Imagining
Microscopy (FLIM), we have demonstrated dynamic plasma membrane binding by cytoplasmic domains
of T cell receptor (TCR)-associated CD3 chains and other T cell transmembrane receptors. We have devel-
oped methods for imaging such membrane binding both at steady state and during receptor triggering at
the IS. Plasma membrane binding by cytoplasmic domains may represent a novel mechanism for regulating
the signaling function of important receptors in the immune system.
Key words FLIM, FRET, Immunological synapse, TCR-CD3 complex, Membrane binding, Lipid
bilayers
1 Introduction
The mechanisms governing regulation and initiation T cell recep-

tor (TCR) triggering have been the subject of intense study for
many years. One proposed regulatory mechanism is based on the
discovery that the cytoplasmic domains of the TCR-associated
CD3ε and CD3ζ signaling subunits can interact with anionic lipid
head groups, which are enriched at the inner leaflet of the plasma
membrane [1–5]. An NMR structure showed that the critical
tyrosine residues of the CD3ε ITAM can dip into the hydro-
phobic core of the lipid bilayer, rendering them inaccessible to
Lck kinase activity [1]. Improved microscopy techniques allowed
us to directly examine cytoplasmic domain membrane binding
in a native lipid membrane environment in live cells using the
259
260 Etienne Gagnon et al.
steady-state q
uenching and dequenching FRET method described
here in Subheading 3.1.
The next essential question to address was how receptor trig-
gering altered membrane binding. TCR triggering results in the
formation of an immunological synapse (IS), which involves a
rapid and coordinated redistribution of surface receptors and
intracellular vesicles toward the contact site with an antigen pre-
senting cell (APC) or target cell [6–8]. Membrane binding by
cytoplasmic domains in the IS is not amenable to study by steady-
state quenching and dequenching FRET methods because these
rely on changes in local donor fluorescence intensity by the FRET
acceptor, and are therefore confounded by local changes in fluo-
rescence intensity caused by receptor clustering at the IS. Recent
advances in time- correlated single photon counting (TCSPC)
techniques have now made it possible to perform FRET measure-
ments using fluorescence lifetime imaging microscopy (FLIM),
which is unaffected by changes in fluorescence intensity [9–11].
Previously, the lengthy acquisition times required to generate reli-
able data using FLIM precluded application of this technique to
studies of the IS due to the rapid and dynamic receptor redistribu-
tion that takes place at early stages of IS formation [8]. This issue
has now been overcome with the development of novel detectors
combining GaASP and Photo-Multiplier Tube technology to cre-
ate hybrid detectors [12]. These detectors enable FLIM data
acquisition with increased photosensitivity and little to no after
pulsing, allowing signal acquisition for short periods of time and
minimizing noise accumulation [12].
We applied a FLIM-FRET technique to study changes in
membrane binding by the cytoplasmic domain of CD3ε during
TCR triggering induced at the IS by artificial antigen presenting
lipid beads (APLBs), as described in Subheading 3.2 [13]. These
APLBs provide a three-dimensional interaction surface that pres-
ents the key ligands required for initiation of T cell activation,
peptide-MHC complexes, and ICAM-1. These molecules are
bound to functionalized lipids on the beads, thus providing the
lateral mobility required for the formation of the typical IS [8, 14].
Using this system, we observed a reduction in membrane binding
by the CD3ε cytoplasmic domain selectively at the IS [13].
Membrane dissociation under receptor triggering conditions pro-
vides a mechanism for signal initiation.
Membrane binding and dissociation may represent a general
mechanism for regulation of signal initiation by other immune
receptors beyond the TCR. The two-part protocol described in
this chapter can be used to first assess membrane binding in live
cells using a basic steady-state quenching and dequenching
FRET system, and then to examine changes in membrane bind-
ing under receptor triggering conditions using FLIM-FRET
and APLBs.
The Study of Dynamic Protein-Plasma Membrane Binding in live Cells… 261
2 Materials
2.1 Jurkat Cell 1. pHAGE mammalian expression vector (Harvard Gene Therapy
Electroporation Initiative).
and Sorting 2. Teal fluorescent protein (mTFP1) cDNA (available from Allele
Biotechnology or Addgene).
3. Solution of 3 M sodium acetate, pH 5.2.
4. Ethanol, molecular-biology grade.
5. Nuclease-free, molecular-biology grade water.
6. Serum-free RPMI.
7. Jurkat cell complete medium: RPMI containing 10% fetal
bovine serum, 10 mM HEPES pH 7.4, and 2 mM
l-alanyl-l-glutamine.
8. Jurkat cell line, clone E6–1 (available from ATCC).

9. Electroporation cuvette, 0.2 cm gap (for example (e.g.),
BioRad, cat. No. 1652082).
10. Bio-Rad Gene Pulser II electroporator, or equivalent.
11. Cell sorter, such as BD FACSAria II, with plate adapter for
sorting, or equivalent.
12. Phosphate buffered saline (PBS).
13. FACS buffer: PBS containing 1% fetal bovine serum.
2.2 Imaging 1. Fluidics imaging chamber and accessories (diagramed in Fig.

Apparatus, Fluidics 1): RC-20H imaging chamber, PH-5 imaging platform, SA-
Components, 20PLIXR-AL stage adaptor, and CS-15R15 coverslips (Warner
and Buffers Instruments).
2. Peristaltic Pump: Econo-Pump Model EP-1 (BioRad), or
equivalent.
3. Syringe Injection Pump: Model SP1000 (Next Advance), or
equivalent.
4. Syringes: 5 and 1 mL syringes.
5. Tubings and connectors: 1.6 mm FPLC-grade tubing,
Kimble™ Kontes™ FlexColumn™ three-way Luer Stopcock,
tubing connectors, syringe adaptors (Fisher Scientific), or
equivalent.
6. Octadecyl Rhodamine B Chloride (R18) (e.g., cat. No. O246,
ThermoFisher) reconstituted in 100% ethanol as a 10 mg/mL
stock solution. Store at −20 °C with desiccant.
7. Wash buffer: PBS at 4 °C.
8. Waste collection system: vacuum trap flask connected to an in-
house vacuum line or another peristaltic pump with outlet tub-
ing into a bottle for biological waste containment.
Fig. 1 Fluidics chamber components. The fluidics chamber components required for these experiments are the
(a) RC-20H imaging chamber (RC-20 chamber shown here; **: inlet port; *: outlet port; arrow points to Teflon
gasket used to secure top coverslip), (b) PH-5 imaging platform (P-5 platform shown here; *: flaps used to
secure the imaging chamber, and (c) SA-20P LIXR-AL stage adaptor. Images taken from the Warner Instruments
website
9. Ice bucket with a mixture of ice and water (see Note 1).
10. 30 mL syringe with blunt-cut p200 pipette tip containing
Vaseline (see Note 2).
2.3 Microscopy 1. Microscope: Zeiss LSM880-Airyscan, Axio-Observer Z1 RP

Components Core 2012, 63× Objective PlanAchromat (1.4 N.A.), Argon
laser (458 nm, 477 nm, 488 nm, 514 nm), diode pumped
solid-state laser (561 nm), HeNe laser (633 nm), or
equivalent.
2. Basic image analysis software (e.g., Image J).
3. Multi-Photon Laser: Coherent Chameleon Ultra II (3.5 W;
690–1080 nm), or equivalent.
4. Time-correlated single photon counting (TCSPC) detector
and software suite: SPC-150 imaging module, HPM-100-40
detectors, SPC-Image software (Becker&Hickl), custom band-
pass filter cube for TFP detection (490/20) (Chroma), or
equivalent.
5. Black/Opaque microscope insulating box: Incubator XL S1
Multi dark (Zeiss, or equivalent) (see Note 3).
2.4 Human T cell See Schubert and Gordo et al. (2012) [15] for a detailed protocol
Clone Culture for culturing a human T cell clone.
and Transduction
1. Human T cell media: RPMI containing 10% fetal bovine
serum, 1% human serum (e.g., Valley Biochemical), 10 mM
HEPES pH 7.4, 2 mM l-alanyl-l-glutamine, and 5 U/mL
human rIL-2 (e.g., Roche).
2. PHA-L (e.g., Roche).
3. Feeder Cells: Peripheral blood mononuclear cells (PBMCs)
isolated from a fresh apheresis process that can typically be
obtained from blood banks with appropriate regulatory
approvals.
4. Ficoll-Paque Plus, density = 1.077 g/mL (e.g., GE Healthcare).
5. Human T cell clone (the HA-D7 clone used in our studies is
specific to the HA306–318 epitope presented on the HLA-DR0401
molecule).
6. Retronectin (human recombinant fibronectin, suggested sup-
plier Takara).
7. Solution of 3% BSA in PBS.
8. Polybrene (hexadimethrine bromide).
9. Titered lentivirus for delivery of the TFP-tagged protein of
interest (see Note 4).
2.5 Hydrophilized 1. Piranha solution: 70 mL concentrated H2SO4, 30 mL H2O2

Coverslips and APLBs (30% solution stock). IMPORTANT, HAZARDOUS! (Handle
with care, see Note 5).
2. Scissors-style forceps (e.g., Nalge Nunc cat No. 6320-0010).
3. Ultrapure water (e.g., from MilliQ system).
4. 2 × 150 mL beakers.
5. Peg-style test tube rack.
6. Silicate beads: acid washed non-functionalized silicate beads
4.5 μm diameter (e.g., Bangs Laboratory).
7.
Lipids: 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC)
(84.5%), 1,2-dioleoyl-sn-glycero-3-[(N-(5-amino-1-
carboxypentyl)iminodiacetic acid)succinyl] (nickel salt) (Ni-
NTA-DOGS) (15%),
1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(cap bio-
tinyl) (sodium salt) (cap-biotin-PE) (0.05%) lipids (e.g., Avanti
Polar Lipids) prepared in deoxygenated Tris-saline as described
in Dustin et al. (2007) [14].
8.
Purified proteins: Alexa647-labeled Streptavidin (e.g.,
ThermoFisher cat. No. S35357), mono-biotinylated
HLA-DR0401 loaded with HA306–318, and non-labeled mouse
ICAM-1-His12. Purified DR molecules are biotinylated using a
1:20 molar ratio of recombinant BirA enzyme to DR in a buf-

fer containing 100 μM biotin, 10 mM ATP, 10 mM magne-
sium acetate, 50 mM Bicine and protease inhibitors at pH 8.0.
The final HLA-DR concentration is adjusted to 2.5 mg/mL
with 10 mM Tris, pH 8.0, and the reaction is incubated for 2
h at 30 °C. Free biotin is removed by gel filtration chromatog-
raphy or extensive dialysis.
9. Lipid loading buffer: HBS (20 mM HEPES, 137 mM NaCl, 5
mM KCl, 0.7 mM Na2PO4, 6 mM d-glucose, 2 mM MgCl2, 1
mM CaCl2, pH 7.4).
10. Protein loading and lipid blocking buffer: HBS-HSA (HBS
containing 1% human serum albumin).
11. APLB wash buffer: HBS-HSA.
3 Methods
3.1 Steady-State This method can be used to assess steady-state plasma membrane
Quenching interaction of a cytoplasmic domain in the absence of receptor trig-
and Dequenching gering. We will use this example to explain the basic principles of
FRET FRET analysis in live cells before moving on to a more complex
technique that incorporates FLIM.
3.1.1 Construct Design 1. Using standard molecular cloning techniques, prepare an

expression construct (see Note 6) with the FRET donor
(mTFP1, teal fluorescent protein) fused to the C-terminus of
the protein of interest, which should be capable of surface
expression at a high level (see Note 7).
2. In the same expression vector, fuse the extracellular and trans-
membrane domains of KIR2DL3 (or another monomeric sur-
face protein) to TFP, separated by a 3 amino acid flexible linker
(sequence Gly-Gly-Ser) to be used as a positive control, or by
a flexible linker that is approximately the same length as the
cytoplasmic domain of the protein of interest to be used as a
negative control (see Note 8).
3.1.2 Electroporation 1. For each construct, transfer 100 μg of plasmid DNA to a

of Jurkat Cells (See Notes microcentrifuge tube.
9 and 10) 2. Add 1/10 volume of 3 M sodium acetate, pH 5.2.
3. Add 2.5 volumes of 100% ethanol. Invert the tube several
times to mix.
4. Incubate at −20 °C for 30 min.
5. Pellet DNA at maximum speed in a table-top microcentrifuge.
6. Discard supernatant and wash the pellet once with cold 70%
ethanol.
7. Air-dry the pellet completely.
8. Resuspend DNA in 10 μL ultrapure water pre-warmed to 55 °C.

9. For each construct, transfer 1 × 107 Jurkat cells to a centrifuge
tube.
10. Wash cells once with serum-free RPMI.
11. Resuspend the cells in 500 μL RPMI and transfer to an electro-
poration cuvette.
12. Add concentrated plasmid DNA to cells and incubate on ice
for 30 min.
13. Electrophorate at 250 V and 975 μF.
14. Incubate on ice for 5 min.
15. Push aside the froth of dead cells at the top of the liquid with
a pipette tip and recover as much liquid as possible from the
cuvette while avoiding the dead cells. Transfer cells to a T-25
flask with 7 mL pre-warmed Jurkat cell complete medium and
return to a 37 °C incubator.
3.1.3 Preparation 1. 48 h post-electroporation, collect the cells in a centrifuge tube.

of Jurkat Single Cell 2. Wash twice with FACS buffer and resuspend at 1 × 107 cells/
Clones mL for sorting.
3. Gate live cells based on FSC and SSC, and then gate on all
TFP+ cells (usually the transfection rate is less than 5% at this
stage). For this purpose, TFP can be measured using the same
settings as for GFP.
4. Sort the gated population of cells to high purity under aseptic
conditions (see Note 11).
5. Pellet sorted cells and resuspend at a density of 1 × 105 cells/
mL in pre-warmed Jurkat cell complete medium. Dispense the
cell suspension into a 24-well plate with 1 mL/well and return
to a 37 °C incubator.
6. Let the sorted cells expand for 7 days, and then repeat steps
1–5. Most cells lose transient TFP expression by this time, so
there should only be about 0.5% TFP+ cells.
7. Let the second round of sorted cells expand for 7 days, and
then repeat steps 1–2. At this point, a distinct population of
20–50% TFP+ cells should emerge that are stable transfectants.
Draw a sort gate covering approximately one log of the TFP
fluorescence intensity centered on this distinct population.
8. Perform single cell sorting into two 96-well U-bottom plates
containing Jurkat cell complete medium (200 μL/well). Also
perform a bulk-sort of stable transfectants to freeze as a back-
up and to use for preliminary FRET analysis, if desired.
9. Monitor the plates for outgrowth of clonal populations.
Expand to a 24-well plate when the cells begin to exhaust the
culture medium (i.e., when the well begins to turn yellow).
10. Continue expanding and maintaining cultures until several

clones grow out for each TFP fusion construct.
11. Collect a sample of 2 × 105 cells for each single cell clone for
flow cytometry analysis.
12. Wash the cells twice with FACS buffer and resuspend in 300
μL FACS buffer.
13. Measure the TFP fluorescence intensity of each clone by flow
cytometry. Ideally, the clones will have TFP expression approx-
imately 2–3 logs above background (i.e., the parental cell line).
Discard any clones that do not have a single, uniform TFP-
expressing population.
14. Select one to three clones for each construct with similar TFP
expression (see Notes 12 and 13) for FRET analysis.
15. The day before a planned imaging experiment, dilute a conflu-
ent culture 1:1 with fresh Jurkat cell complete medium to
ensure a high frequency of healthy cells with optimal morphol-
ogy for imaging.
3.1.4 Preparing The layout for the fluidics system used for the FLIM-FRET method
the Imaging Experimental is shown in Fig. 2. Some modifications can be made to this system
Setup to simplify the setup for steady-state quenching and dequenching
FRET with Jurkat cells. For clarity, we recommend reading about
the full setup in Subheading 3.2.4 before proceeding with the sim-
plified setup described here. As another alteration, we also recom-
mend performing this method at 37 °C to more closely approximate
physiological conditions, while the FLIM-FRET method is per-
formed at 4 °C for reasons described in Subheading 3.2.7.
1. Place the 1 L PBS bottle into a 37 °C water bath as a supply of
warm fluid for washing the cells in the flow chamber.
2. Submerge one end of a 12-in. section of tubing into the PBS
bottle and connect the other end to the inlet port of the peri-
staltic pump.
3. Connect a new section of tubing to the outlet port of the peri-
staltic pump. Connect the other end to injection port C1 on T
connector C.
4. Place the syringe pump beside T connector C so that the 5 mL
syringe can be directly connected to injection port C2. Set the
syringe pump to inject a volume of 800 μL at a rate of 0.5 mL/
min.
5. Connect a section of tubing to the remaining end of
T-connector C that is long enough to reach the fluidics cham-
ber once it is installed on the microscope.
6. Prime all of the tubing lines with PBS using the peristaltic
pump (Fig. 2).
Fig. 2 Fluidics system layout: Schematic representation of the components and the setup needed to perform
the fluidics injection experiments presented in Subheading 3.6
3.1.5 Preparation 1. Collect Jurkat cells from culture. Wash twice with PBS and
of Jurkat Cells and R18 resuspend in PBS at 2 × 106 cells/mL.
Labeling Buffer 2. Incubate on ice for 20 min (see Note 14).
3. During the incubation, dilute the R18 stock to 2 μg/mL in
PBS.
4. Collect 5 mL of the cell labeling buffer in a 5 mL syringe
through a 16-gauge needle, being careful to avoid introducing
foam or bubbles into the syringe.
5. Mount the R18 labeling syringe onto the syringe pump. Then,
gently and carefully connect the R18 labeling syringe to
T-connector C without introducing any air or generating any
backpressure (see Note 15).
6. Return to the Jurkat cell suspension on ice and transfer 50 μL
as a mounded drop onto a coverslip positioned in the depres-
sion at the center of the PH-5 platform. Incubate for 5 min at
room temperature to allow the cells to settle onto the
coverslip.
7. During the incubation, assemble the rest of the RC-20H
chamber. Extrude a thin line from the Vaseline syringe onto
the top side of the the flow cell, at the inside edge of the
opening onto the top coverslip (as shown in Fig. 3). Use a bent
p200 pipette tip to spread the Vaseline evenly (see Note 16).
8. Using fine dissection forceps, place the top coverslip over the
Vaseline coating.
9. Secure the top coverslip with the Teflon gasket provided by the
manufacturer, making sure that the Vaseline does not spread
toward the inside of the chamber (see Note 17) and that it
forms a complete seal.
Fig. 3 Fluidics imaging chamber setup: Schematic representation of the assem-

bled RC-20H imaging chamber onto the PH-5 platform and stage adaptor. Flow
direction is indicated by arrows, and placement of the extruded vaseline by
asterisk
10. Turn the RC-20H chamber upside down and use the Vaseline
syringe to extrude another thin line on the bottom side of the
the flow cell, at the inside edge of the opening onto the bot-
tom coverslip. Spread the Vaseline evenly as before.
11. Carefully affix the RC-20H chamber to the PH-5 platform,
matching the edges of the bottom coverslip (containing the
mounded drop of Jurkat cells) to the Vaseline coating on the
bottom of the chamber.
12. Secure the RC-20H chamber using the flaps on the PH-5 plat-
form and gently tighten the screws (see Note 18).
13. Install the PH-5 platform on the microscope stage. Insert the
inlet and waste tubings to the chamber as shown in Figs. 1
and 3.
14. Ensure that T connector C is in the correct position (PBS to
flow chamber) and then start the peristaltic pump at 0.3 mL/
min to begin washing the cells. This will remove any excess
cells that did not settle onto the coverslip and warm the cells in
preparation for R18 labeling (Fig. 3).
3.1.6 Data Acquisition FRET between the TFP donor at the C-terminus of the transmem-
for Steady-State brane protein of interest and R18 intercalated into the plasma
Quenching FRET membrane results in quenching of the TFP donor only when the
During Real-time R18 TFP is in close proximity to the inner leaflet of the plasma mem-
Labeling brane, either due to the short cytoplasmic linker (positive control)
or due to membrane interaction by the cytoplasmic domain of the
protein of interest. In a successful experiment, it should be possible
to observe a selective reduction in TFP fluorescence intensity at
the plasma membrane during live data acquisition while R18 labels
the plasma membrane. We recommend starting analysis with the
positive control to ensure that the system is working properly.
1. Prepare for data acquisition using the following general set-

tings in the microscope control software: 1x numerical zoom,
1024 × 1024 resolution, 600 Hz scan speed, 2 line average,
16-bit depth, 1 airy unit pinhole (see Note 19). Set up a
sequential scan protocol for measuring first TFP (excitation
458 nm, detection 466–526 nm) and then R18 (excitation
560 nm, detection 568–628 nm). Laser power and detector
gain will need to be determined empirically for each individual
microscope setup. Switch channels between frames and cap-
ture 14 image sets at 20 s intervals.
2. Use live scanning in the TFP channel to select a field of view in
which the cells are not overlapping or touching. Try to mini-
mize the scanning time to avoid causing any TFP photobleach-
ing due to the repeated excitation. Stop live scanning once the
field of view has been selected.
3. Stop the peristaltic pump and switch T connector C to send
fluid from the syringe pump to the flow chamber.
4. Start the sequential scan program in the microscope control soft-
ware. Start the syringe pump after the 2nd image set has been
acquired. Once the injection is complete, switch T connector C
back to the original position (peristaltic pump to flow chamber)
and start the peristaltic pump again. The cells will continue to
accumulate R18 at the plasma membrane during this wash step.
5. If desired, perform dequenching FRET analysis at this step (see
Subheading 3.1.7).
6. Otherwise, perform three R18 labeling replicates for each con-
struct to be tested. Also, collect a time-lapse series for TFP-
expressing cells with no R18 labeling to control for TFP
photobleaching.
3.1.7 Data Acquisition 1. Following the acquisition of quenching FRET data, change
for Steady-State settings in the microscope control software to 512x512 resolu-
Dequenching FRET Using tion and 3× numerical zoom.
R18 Photo-Bleaching 2. Select a field of view in which the cells are not overlapping or
touching, and have uniform R18 labeling compared to the rest
of the flow cell.
3. Use sequential scan to acquire an image set in both the TFP
and R18 channels (quenched donor images).
4. Perform R18 photobleaching using repeated R18 excitation
with high laser power and slow scan speed. As a starting point,
we used 100% power to the 561 nm laser with a scan speed of
100 Hz and performed live scanning for approximately 20 s to
achieve at least a 95% reduction in R18 signal intensity.
5. Revert all settings to those used in step 3 and acquire another
image set in both the TFP and R18 channels (dequenched
donor images).
6. Repeat steps 2–5 for other fields of view until the R18 dye
begins to label other membranes in the cell. Typically, five to
eight fields of view can be imaged during the photobleaching
step for each labeling experiment.
3.1.8 Data Analysis 1. Import images from a single time-lapse as a series into ImageJ
for Steady-State or other image analysis software.
Quenching FRET 2. Use the polygon tool to draw a region of interest (ROI)
encompassing a portion of the plasma membrane, as shown in
Fig. 4a (see Note 20).
3. Measure the TFP mean fluorescence intensity in the ROI for
each image in the series. Before measuring each image, adjust
the ROI as needed to accommodate for slight movements of
the cell or changes in plasma membrane morphology.
4. Calculate quenching FRET efficiency according to the formula
EFRET (%) = (TFP0 – TFPx)/TFP0 × 100 where TFP0 is TFP
fluorescence at t=0 s, and TFPx is at t = x s. Calculate average
EFRET for at least 20 cells per condition.
5. The positive control (3 amino acid cytoplasmic linker) should
yield high mean EFRET values of approximately 50–60% at late
time points (after maximal R18 labeling), as shown in Fig. 4.
When interpreting the results for the protein of interest, high
mean EFRET similar to the positive control construct (3 amino
acid cytoplasmic linker) indicates maximal membrane binding,
while low mean EFRET similar to the negative control construct
(flexible linker similar in length to the protein-of-interest) indi-
cates no specific membrane interaction. Also, apply the quench-
ing FRET efficiency calculation to the image set with no R18
labeling to control for false FRET caused by TFP photobleach-
ing (this should be less than 5%) (see Note 21) (Fig. 4).
3.1.9 Data Analysis 1. Import quenched and dequenched images as a series into
for Steady-State ImageJ or other image analysis software.
Dequenching FRET 2. Use the polygon tool to draw a ROI encompassing a portion
of the plasma membrane.
3. Measure the TFP mean fluorescence intensity in the ROI in
the quenched and dequenched images.
4. Calculate dequenching FRET efficiency according to the for-
mula EFRET (%) = (TFPDQ – TFPQ)/TFPDQ × 100 where TFPDQ
and TFPQ are equal to TFP mean fluorescence in the
dequenched and quenched image sets, respectively.
5. The mean EFRET calculated by the quenching and dequenching
methods should be internally consistent (i.e., they should be the
same for a given R18 labeling experiment), and should be simi-
lar across different R18 labeling replicates for a given c onstruct
provided that the R18 labeling efficiency is similar each time.
Fig. 4 Donor quenching FRET analysis methodology: HA-KIR-3TFP expressing T cells were labeled with R18
according to the protocol presented in Subheadings 3.9 and 3.10. Cells were analyzed according to the proto-
col presented in Subheading 3.11. (a) TFP and R18 fluorescence was captured at 20 s intervals during cell
labeling. To calculate Donor Quenching FRET efficiency, two sections of the plasma membrane were selected
(yellow) in both channels and signals were subtracted from background (purple). (b) Effects of R18 labeling on
TFP fluorescence were determined using the data obtained in (a). (c) FRET efficiency between TFP and R18
was calculated from the data obtained in A using the equation presented in Subheading 3.11
3.2 Measuring Once the membrane binding properties of the protein of interest
Dynamic Plasma have been established, similar principles can be applied to examine
Membrane Interaction alterations in membrane binding during receptor triggering at the
at the Immunological IS. It is not possible to use the basic quenching and dequenching
Synapse Using FRET methods to address this question because it relies on changes
FLIM-FRET in local TFP fluorescence intensity to assess FRET efficiency.
Dramatic receptor redistribution to the IS results in increased TFP
density and mean fluorescence intensity which therefore confounds
quenching and dequenching FRET calculations. Fluorescence life-
time is independent of fluorophore density, but is reduced in a
quantitative way for the donor fluorophore of a FRET pair. This
method will describe an adaptation that uses FLIM to measure
FRET at the IS of a human T cell clone upon stimulation with an
APLB as an artificial antigen presenting cell.
3.2.1 Transduction 1. Restimulate the human T cell clone according to the protocol
of Human T Cell Clone described in Schubert and Gordo et al. 2012 [15].
2. One day before transduction (on day 6 post restimulation),
add 250 μL/well of 20 μg/mL Retronectin in PBS to a 24-well
non-tissue-culture-treated plate. Incubate at 4 °C overnight.
3. The next day, remove the Retronectin solution from the wells
and replace with a solution of 3% BSA in PBS. Do not allow
the Retronectin-coated wells to dry out at any time. Incubate
the BSA-PBS solution in the wells for blocking while preparing
the T cells.
4. On day 7 post-restimulation, collect the T cells from two
96-well plates. Pellet the cells and resuspend in human T cell
media at 2 × 106 cells/mL. Remove the BSA-PBS solution
from the Retronectin-coated wells and quickly replace with
500 μL/well of the T cell suspension.
5. Add 500 μL of human T cell media containing titered lentivi-
rus (see Note 10) for a final MOI of 10 and polybrene for a
final concentration of 4 μg/mL.
6. Spin the cells for 60 min at 2000 (751 × g) rpm at 32 °C.
7. Return the cells to a 37 °C incubator for 1 h.
8. Add 1 mL of human T cell media per well. Return to the 37
°C incubator overnight.
9. Collect the transduced T cells from the 24-well plate. Pellet
the cells and resuspend in human T cell media at 2 × 106 cells/
mL. Transfer the cell suspension to a T-25 flask and return to
the 37 °C incubator.
10. At 14 day post-restimulation, collect the cells and prepare for
sorting as described in Subheading 3.1.3. Gate live cells based
on FSC and SSC (see Note 22). Draw a sort gate for TFP+ cells
covering approximately one log of TFP fluorescence intensity.
11. Restimulate sorted cells as described for general human T cell
clone culture [15].
12. T cells can be used for imaging on days 10–14 post-
restimulation. Cells can be kept in culture with restimulation
every 14 days for up to three restimulation cycles, after which
a new aliquot of frozen cells should be thawed (see Note 23).
13. Prior to imaging, recover cells from the plate and resuspend in
human T cell culture media without rIL-2 at a density of 1 ×
106 cells/mL. Incubate cells in IL-2 starvation conditions for
a minimum of 2 h prior to imaging (see Note 24).
3.2.2 Preparation Treatment of the glass surface with Piranha solution enables adhe-
of Hydrophilized Coverslips sion of the APLBs such that they are held in place during subse-
quent injections and during synapse formation. This facilitates
FLIM data acquisition and confocal image capture of the synapse

by minimizing movement of cell-bead conjugates.
1. Secure four to six coverslips within the grips of a scissor-type
forceps, with one coverslip per forceps.
2. Clean the coverslip with a Kimwipe to remove any dust
particulates.
3. Submerge the coverslip in a 150 mL beaker containing freshly
prepared Piranha solution, leaving the forceps handle accessi-
ble outside the beaker. Incubate for 15–20 min at room tem-
perature. EXTREME CAUTION: Full lab coat and eye
protection are essential at all times when manipulating Piranha
solution (see Note 5).
4. Remove the forceps-held coverslip and submerge in ulltrapure
water in a clean 150 mL beaker.
5. Gently agitate the beaker, taking care not to break the
coverslips.
6. Empty the beaker and add fresh ultrapure water.
7. Repeat steps 5 and 6 a minimum of three times to thoroughly
rinse the coverslips, and then remove them from the beaker.
8. Hold the coverslip steady with the forceps while removing any
clinging water droplets with a vacuum aspirator.
9. Insert the handle of the forceps into a peg-style test tube rack
in a chemical hood, leaving the coverslips exposed to air for
drying for at least 20 min.
3.2.3 Preparation Antigen presenting lipid beads approximate a three-dimensional

of Antigen Presenting Lipid artificial antigen presenting cell. The lipid environment on the
Beads (APLBs) coated bead surface provides lateral mobility for the loaded pro-
teins, which is important for the formation of the IS structure on
the T cell. Here, we describe the preparation of the most minimal-
istic APLBs for interaction with T cells, which present cognate
peptide-loaded MHC molecules and ICAM-1 as an adhesion mol-
ecule (Fig. 5).
1. Thoroughly resuspend silicate beads and then transfer 20 μL of
bead suspension into 500 μL HBS in an Eppendorf tube.
2. Invert the tube several times and then pellet the beads using a
pulse spin in a table-top centrifuge.
3. Carefully remove 500 μL of supernatant using a pipette, and
repeat the wash step with another 500 μL of HBS.
4. Add 40 μl of the 0.4 mM DOPC liposome preparation con-
taining Ni-NTA-DOGS (15%) and cap-biotin-PE (0.5%) (see
Note 25).
5. Incubate at room temperature on a tube rotator for 2–5 min.
Fig. 5 Antigen Presenting Lipid Beads (APLB): Schematic representation of the

surface of the APLB to mimic antigen presenting cells. Depicted here is the most
minimalistic surface capable of inducing immunological synapse formation in
primary T cells
6. Add 460 μL of HBS and wash the beads twice as in step 2.

Carefully remove 500 μL of supernatant after the last wash
step.
7. Add 500 μl HBS-HSA to the bead pellet from the last wash
step and incubate for 10 min on a tube rotator at room tem-
perature to block nonspecific protein binding sites.
8. Wash the beads with HBS-HSA as in step 2.
9. Resuspend the beads in 500 μL HBS-HSA containing 5 μg/
mL Alexa568 Streptavidin and incubate for 10 min on a tube
rotator at room temperature.
10. Wash the beads with HBS-HSA as in step 2.
11. Resuspend the beads in 500 μL HBS-HSA containing 1 μg/mL
mono-biotinylated pMHC (HA306–318-HLA-DR0401 for the
HA:D7 clone used in our studies) and 3 μg/mL ICAM-1-His12.
Incubate for 10 min on a tube rotator at room temperature.
12. Wash the beads with HBS-HSA as in step 2 and keep the pellet
in a minimal volume (approximately 20–40 μL) on ice for
same-day use.
3.2.4 Setup See Fig. 2 for a guide to the fluidics system layout.
of the Fluidics System
1. Place the 1 L PBS bottle into an ice bucket containing a mix-
for FLIM-FRET Microscopy
ture of ice and water to keep the solution cold.
Imaging
2. Connect T-connector A to a section of tubing using appropri-

ate tubing connectors. Typically, 2 in. of tubing is enough for
this section (see Note 26).
3. Connect T-connector B to the other end of the tubing from
the previous step using appropriate tubing connectors.
4. Connect the remaining end of T-connector B to a 12-in. sec-
tion of tubing using appropriate tubing connectors.
5. Connect T-connector C to the remaining end of the tubing
from the previous step using appropriate tubing connectors.
6. Submerge one end of a 12-in. section of tubing into the PBS
bottle and connect the other end to the inlet port of the peri-
staltic pump.
7. Connect a new section of tubing to the outlet port of the peri-
staltic pump.
8. Connect the other end of the tubing from the previous step to
T-connector C, which is already connected to the other
components.
9. Connect a section of tubing to the remaining end of
T-connector C that is long enough to reach the fluidics cham-
ber once it is installed on the microscope.
10. Once all the connections are made, the system is ready to be
primed with the various buffers (see Note 27). Switch
T-connector C valve to prime the PBS line toward the fluidics
chamber. The chamber should not be installed at this stage, so
simply collect the priming fluid in a waste beaker.
11. Turn on the peristaltic pump and prime the PBS lines at maxi-
mum speed. Once this line is filled, turn off the pump and
switch the T-connector C valve to close off this line.
12. Switch the T-connector A and B valves to prime the remaining
lines with HBS-HSA using injection port A2 and a 10 mL
syringe (see Note 28).
13. Once the lines are filled and no air bubbles are visible, the sys-
tem is ready.
3.2.5 Assembly 1. Assemble the RH-20C flow chamber and affix it to the PH-5
of the RH-20C Flow platform as described in Subheading 3.1.5 steps 7–12, except
Chamber and Injection be sure to use a hydrophilized coverslip for the bottom cover-
of the APLBs slip when performing this method with APLBs. In addition,
the T cells should not be pre-incubated on the bottom cover-
slip before assembling the flow chamber in this application.
2. Dilute 5 μL of the APLB suspension from Subheading 3.2.3
step 12 into 300 μL PBS.
3. Using a fine pipet tip, inject the APLBs into the assembled
fluidics chamber and let stand for 5 min. Typically, this yields a
medium density of beads on the coverslip at an ideal density

for imaging (see Note 29).
4. Flush the chamber slowly with 500 μl of HBS-HSA. The beads
should remain attached to the coverslip. The fluidics chamber
is now ready for imaging.
5. Secure the PH-5 platform onto the stage adapter and load
onto the microscope inside the dark incubator box.
6. Pass the tubings through the appropriate openings in the incu-
bator box and connect the inlet and outlet lines to the cham-
ber, as shown in Figs. 1 and 3.
3.2.6 Preparation 1. At this stage, the human T cell clones should have been in IL-2
of the T Cells and R18 starvation conditions for at least 2 h (Subheading 3.2.1, step 13).
Labeling Solution 2. Recover 1.5 mL of T cells from IL-2 starvation conditions into
a microcentrifuge tube. Pellet the cells and resuspend in 1.5
mL pre-warmed HBS-HSA.
3. Take up the T cells in a 1 mL syringe, being careful to avoid
introducing foam or bubbles into the syringe.
4. Gently and carefully connect the T cell syringe to T-connector
A (injection port A1) without introducing any air or generat-
ing any backpressure.
5. Dilute the R18 stock to 2 μg/mL in 10 mL cold PBS (from
the ice bucket).
6. Collect 5 mL of the cell labeling buffer in a 5 mL syringe
through a 16-gauge needle, being careful to avoid introducing
foam or bubbles into the syringe.
7. Mount the R18 labeling syringe onto the syringe pump. Then,
gently and carefully connect the R18 labeling syringe to
T-connector B without introducing any air or generating any
backpressure.
3.2.7 Injection of T Cells 1. Engage live wide field settings on the microscope to monitor
and Conjugate Formation the injection of T cells and conjugate formation with APLBs.
with APLBs 2. Ensure that all T-connectors are in the correct position for
fluid flow from injection port A1 (T cell syringe) to the flow
chamber.
3. Slowly and carefully inject 1 mL of T cells into the flow
chamber.
4. Monitor the injected T cells as they begin to land on the cov-
erslip and interact with the APLBs, which should occur approx-
imately 1 min. Following T cell injection. When sufficient
conjugate formation is observed, switch the valve on
T-connector A to inject from injection port A2. Slowly inject 1
mL of HBS-HSA into the fluidics chamber.
5. Immediately after, switch the valve on T-connector B to inject

from injection port B1. Activate the syringe pump to inject
800 μL of cell labeling buffer at a rate of 0.5 mL/min.
6. Once the injection is complete, switch the valve on T-connector
C to inject from injection port C1 and activate the peristaltic
pump to inject ice cold PBS at a flow rate of 0.5 mL/min.
Leave the peristaltic pump on while imaging as this will slow
down the movement of the cells, facilitating data acquisition
for FLIM analysis (see Note 30).
3.2.8 Image Acquisition 1. Engage both the Argon (458 nm) and diode pumped (561
for FLIM-FRET nm) lasers to visualize TFP (detection 466–526 nm) and R18
(detection 570–630 nm). Use a low laser power and a high
scanning speed to minimize photobleaching of the fluoro-
phores during cell selection. To compensate for low laser
intensity, use a high detector gain. This would normally com-
promise image quality, but that is not important at this stage
because the purpose of this initial scanning is only to select
good candidates for FLIM data acquisition and not to capture
quantitative images. Use a zoom of 1 and a resolution of 512
× 512.
2. Select cell-APLB conjugates in which the R18 labeling of the
cell is similar to the rest of the flow cell and TFP has accumu-
lated at the interface.
3. Use the crop/zoom tool to zoom to 8× centered on the cell of
interest. Adjust the focus if necessary and then stop image
acquisition.
4. Switch the deflection plate to send emitted photons to the
FLIM detector and then close all openings of the incubation
box to create a completely dark environment around the flow
chamber.
5. Activate the multi-photon laser and set at 820 nm with 6%
laser power.
6. Activate the HPM FLIM detector using the SPC-imaging
module and allow the detector to settle below 1000 photons/s
before engaging data acquisition (30 s is usually sufficient for
our instrumentation) (see Note 31).
7. Start recording on the FLIM detector and then rapidly engage
the imaging protocol within the microscope control software.
The FLIM detector and the excitation laser are controlled
by different software programs, and it is therefore essential
for proper data acquisition that certain key parameters be
matched between the two protocols (specifically, the image
resolution and the acquisition time) (see Note 32). The actual
start of data acquisition from the HPM-FLIM detector only
begins once photons reach the detector. The scanning speed
should be set at maximum to minimize pixel dwell time of the

laser; this not only reduces the risk of photobleaching of the
fluorophore, but also prevents a “photon avalanche” at the
detector, a condition in which multiple photons are captured
simultaneously in a given pixel and which confounds FLIM
calculations.
8. Once FLIM data have been acquired and scanning has stopped,
switch the deflection plate to have photons sent toward the
scan head.
9. Capture an image of the cell of interest in the TFP and R18
detection channels with optimized confocal microscopy set-
tings as in Subheading 3.1.6, step 1.
3.2.9 Data Analysis Here, we provide a detailed, step-by-step guide for FLIM data
for FLIM-FRET analysis using SPCimage software (v5.4, Becker & Hickl), which
accompanies the SPC-150 imaging module that we used for our
data acquisition. For assistance or a tutorial for data analysis using
other software, we recommend consultation with the manufac-
turer (Fig. 6).
1. Begin analysis with non-R18 labeled cells to establish the base-
line fluorescence lifetime of the donor fluorophore (i.e., TFP)
in the absence of FRET conditions. Import a FLIM image into
SPCimage software. The program will display two versions of
the image with raw data on the left and a color-coded image on
the right rendering the mean fluorescence lifetime for each
pixel in the image (Fig. 6b).
2. Select a pixel of interest within the plasma membrane using the
blue crosshair tool. The program will calculate a decay curve
for that pixel, which will be displayed as an overlay with the
raw photon count data as a function of time. The χ2 value
describes the goodness-of-fit of the calculated decay curve to
the raw data, with a value of 1.00 being a perfect fit (Fig. 6a).
3. In order to improve the quality of the curve fit, several param-
eters may need to be optimized (see Note 33). Begin by
increasing the number of photons considered in the decay
curve calculations by increasing the number of bins to have at
least 5000 photons in the trace (as shown in the lower right
corner of the trace window). Use a binning value below 4
when analyzing the plasma membrane to avoid including pixel
information from cytoplasmic vesicles adjacent to the mem-
brane. For the non-R18-labeled sample, only use a single
exponential decay function.
4. Perform a decay matrix calculation using the function in the
Calculate tab of the software. The program will return the
color-coded image described in step 1 (Fig. 6b), and a color-
coded histogram depicting the fluorescence lifetime for each
Fig. 6 SPCimage user interface and results examples: Basal FLIM images on HA-CD3ε-TFP expressing T cells
were acquired as presented in Subheading 3.14 and then analyzed by using SPCimage. (a) Fluorescence
decay trace (blue dots), fitted curve (red), and instrument response factor (obtained) after positioning blue
crosshair onto the fluorescent signal at the PM and engaging a 2× bin. Efficient curve fitting is observed with
a χ2 value optimal value of 1.00. (b) Images obtained after calculating decay matrix using the settings pre-
sented in a. The color-coded image on the right reflects the wavelengths depicted in c. (c) Relative distribution
of photons with indicated fluorescence half-life observed in figure presented in B
pixel in the field of view (Fig. 6c). Note that the fluorescence
lifetime for the donor fluorophore may differ across various
cellular compartments (e.g., TFP lifetime is lower in intracel-
lular vesicles) because local environmental conditions such as
pH and salt concentration can affect fluorescence lifetime. For
the purposes of this analysis, we are only concerned with fluo-
rescence lifetime at the plasma membrane.
5. Use the Define Mask tool to draw a ROI around a section of

plasma membrane and record the median fluorescence half-life
for pixels within the ROI (shown as a cross in Fig. 6c).
6. Repeat steps 1–5 to analyze a minimum of 20 cells to deter-
mine the average fluorescence half-life of the donor fluoro-
phore at the plasma membrane in the absence of quenching
acceptor.
7. Next, extend these analyses to a quenched donor sample by
repeating steps 1–5 with the following modifications to incor-
porate FRET conditions (see Note 34). Increase the number
of photons considered in the decay curve calculations by
increasing the number of bins to have at least 10,000 photons
in the trace (as shown in the lower right corner of the trace
window). A greater input of data is required for accurate fitting
to the more complex multiexponential decay function. As in
step 3, keep the binning value below 4. Increase the multiex-
ponential decay components to 2 and fix the t2 value to the
mean value previously determined in step 6 for the unquenched
donor at the plasma membrane.
8. Perform a decay matrix calculation using the Calculate tab in
the software. The program will use the best fit to a biexponen-
tial decay function to de-convolute the relative abundance of
the short-lived (FRET) and long-lived (non-FRET)
components within each pixel, which will be displayed as a1
and a2, respectively (see Note 35).
9. Use the Define Mask tool to draw a ROI around a section of
plasma membrane. See Fig. 7a, b for a side-by-side comparison
of the fluorescence half-life of the same donor fluorophore in
the presence or absence of FRET acceptor at the plasma
membrane.
10. Export the raw data for parameters of interest in text format
for further analysis using Excel or a similar program and to
generate appropriate figures for visual comparison. As an
example, we have depicted an overlay of the raw photon counts
and fitted fluorescence decay curves (Fig. 7c). The shift from a
single to a dual exponential decay function is readily apparent
from the curved nature of the red trace compared to the linear
blue trace. Additionally, we have analyzed differences in mean
fluorescence half-life for the same FRET donor fluorophore at
the plasma membrane in the presence or absence of FRET
acceptor (Fig. 7d).
11. In the absence of APLB interaction, this FLIM-FRET analysis
should confirm analysis of membrane binding by steady-state
quenching and dequenching FRET as described in Subheading
3.1. With the incorporation of APLBs, perform the same anal-
ysis described in steps 7–11 to assess changes in FLIM-FRET

(and hence changes in membrane binding of the protein of
interest) at the IS, where receptor triggering occurs, and at a
noninteracting region of the plasma membrane of the same
cell. Using this method, we have observed a reduction in mem-
brane binding by CD3ε cytoplasmic domain during receptor
triggering at the IS (Fig. 7) [13].
Fig. 7 Extending your FLIM data analysis for publication: Basal FLIM and FRET-FLIM images of HA-CD3ε-TFP
expressing T cells that were left untreated (Basal) or labeled with R18 (Quenched) were acquired as presented
in Subheading 3.14 and then analyzed using SPCimage. (a, b) Color-coded FLIM image of Basal-FLIM (a) and
Quenched FRET-FLIM (b) cells where a portion the PM was selected with the Define Mask tool. (c) Comparison
of the fluorescence decay traces and fitted curves of the Basal and Quenched cell PM presented in a and b.
The dual exponential nature of the Quenched fluorescence decay is noted by a nonlinear fitting onto an expo-
nential Y axis as compared to Basal fluorescence decay. (d) Relative distribution of photons found within the
masked regions of the PM presented in (a) according to their fluorescence lifetime. Decrease in fluorescence
lifetime and presence of multiple fluorescence components are indicative of active FRET
4 Notes
1. An ice bucket or expanded polystyrene box should be filled to

one-third of capacity with ice and contain enough water to cre-
ate a slurry to ensure proper cooling of the 1 L PBS bottle.
2. First check that large end of the p200 pipette tip fits securely
within the threads of the 30 mL syringe tip. Cut the tip to
increase the bore size (to about 2 mm) to extrude the petro-
leum jelly more easily. However, the opening should not be
made so large that too much petroleum jelly is extruded dur-
ing application, which will result in clogging of the fluidics
chamber inlets.
3. The Zeiss Incubator XL-S1 Multi Dark is a rather expensive
addition (around $12,000). The standard Incubator XL from
Zeiss can be purchased as an alternative for around $6000 and
modified with either black matte paint (the incubator should
be fully disassembled prior to painting), or opaque decals (full
disassembly also required). In both cases, ensure that all move-
able parts of the incubator retain full mobility once modifica-
tions have been made. Alternatively, an opaque box can be
custom-made from foam cardboard to fit tightly around the
microscope and tailored for the particular microscope setup.
4. Prepare titered lentivirus encoding the TFP fusion construct
for transduction of the human T cell clone according to a
protocol optimized for the vector of choice. In the case of the
pHAGE vector used for our studies, we packaged lentivirus in
HEK293T cells by co-transfection of 3 μg of pHAGE plasmid
DNA with 2.7 μg of pDR8.91 packaging plasmid and 0.2 μg
of VSVG envelope plasmid per 10 cm plate. Viral supernatants
were harvested at 48 h and 72 h post-transfection and concen-
trated by ultra-centrifugation. The concentrated lentivirus
stock was then titered by serial dilution on HEK293T cells.
5. Prepare Piranha solution in a clean Pyrex beaker (150 mL)
placed inside a Pyrex dish in a chemical hood. Slowly add 30
mL of H2O2 (30% solution) to 70 mL of concentrated H2SO4
while stirring gently with a clean glass rod. CAUTION: Use
extreme care when working with this reagent. Full lab coat
and eye protection must be worn at all times. Piranha solu-
tion waste continues to produce gases for a period of days.
A well-marked waste container (Pyrex glass bottle) should
be left with the cap loose (to avoid explosion) in a chemical
resistant tray (e.g., Pyrex). Introduction of any other type
of waste into this container, particularly organic compounds,
will result in a violent reaction and possible explosion. It is
important that the piranha solution waste container be well
labeled and that visible warnings be present. Contact the
institutional Environmental Safety office to determine the

appropriate method for waste disposal in accordance with
local laws and regulations.
6. We have used the pHAGE vector, which contains the EF1α
promoter, to express TFP fusion proteins for our studies.
Other expression vectors that have been validated for high
expression in the chosen cell type would also be suitable.
7. If the protein of interest is not normally well expressed at the
cell surface, membrane binding can still be assessed by fusing
the cytoplasmic domain of the protein of interest to the extra-
cellular and transmembrane domains of another surface-
expressed protein (e.g., the NK cell receptor KIR2DL3),
followed by the C-terminal TFP fusion as in the typical design.
8. The flexible linker allows the TFP donor to be positioned close
to or far from the plasma membrane depending on the confor-
mation of the linker, and the entire population of TFP mole-
cules measured experimentally should represent a random
distribution between these states. This results in random prox-
imity FRET, the intensity of which depends on the length of
the flexible linker. A shorter linker has a lower maximal dis-
tance from the membrane, shifting the population toward
greater membrane proximity. Therefore, the appropriate nega-
tive control must include a flexible linker that is the same
length as the cytoplasmic domain of the experimental protein.
The flexible linker can be composed of Gly-Gly-Ser repeated
units.
9. We have used Jurkat cells extensively for steady-state quench-
ing and dequenching FRET experiments because the bio-
chemical properties of the plasma membrane are likely to be
similar to that of primary T cells. Jurkat cells and other cell
lines provide an advantage in terms of ease-of-use because they
can support higher protein expression, which facilitates quan-
titative imaging experiments, and they are easier to maintain in
culture for imaging at any time.
10. Instead of electroporation, the TFP fusion constructs can also
be introduced to the cells by lentiviral transduction. For Jurkat
cells and other cell lines that are susceptible to transduction,
MOI of approximately 0.1 (resulting in approximately 10%
transduction efficiency) should be used to avoid introducing
multiple copies per cell. This results in fairly uniform expres-
sion of the transduced construct, which facilitates selection of
single-cell clones with matched expression for imaging. In
addition, when using the transduction protocol it is unneces-
sary to perform multiple rounds of sorting. Instead, skip
directly to the final (single-cell) sort described in steps 7–8 of
Subheading 3.1.3.
11. Contact a flow cytometry core facility for assistance if you are
unfamiliar with techniques for aseptic cell sorting.
12. Quenching and dequenching FRET measurements as described
here rely on measuring changes in the donor (TFP) fluores-
cence intensity. In order to draw comparisons between cells
expressing different TFP fusion constructs, it is essential that
the TFP expression level be as close to identical as possible
across all of the cell populations to be analyzed.
13. We perform most of our FRET measurements with one clone
for each TFP fusion construct to be analyzed, but reserve
another one or two clones to independently confirm the
results.
14. Only use cells that have been resuspended in PBS for less than
1 h. After that time, cells become unhealthy and are not suit-
able for imaging. Typically, a single batch of resuspended cells
can be used for about 3 R18 labeling trials before having to
prepare fresh cells.
15. Be very careful when connecting the syringe containing the
diluted R18 solution to the 3-way stopcock. Do not introduce
a bubble. Also, take care when tightening the syringe not to
introduce too much pressure: the bolus of pressure released
when opening the stopcock will wash the cells off the coverslip
in the imaging chamber.
16. Be careful to not spread extruded Vaseline into the inlet or
outlet openings in the fluidics chamber, which would cause
clogging and prevent fluid flow through the chamber. To
relieve slight clogging of the openings, insert and withdraw a
fine pipette tip (e.g., a 2–10 μL tip) or a 27-gauge needle bent
at a 90° angle at approximately 0.5 cm from the tip. It is rec-
ommended to perform this cleaning procedure as a precaution
against clogging each time a new coat of Vaseline is applied.
17. Here again, some spreading of the Vaseline is inevitable. Follow
the procedure explained in Note 16 to relieve clogging. Also,
remove any excess Vaseline from inside chamber by wiping
with a Kimwipe rolled onto fine dissecting forceps. Otherwise,
excess Vaseline may become dislodged during fluidics injec-
tions, interfering with R18 labeling and cell retention on the
coverslip, and possibly clogging the system.
18. Make sure that the assembled RC-20H chamber forms a tight
seal with the PH-5 platform, but do not tighten the screws on
the flaps too much as the chamber could split or break.
19. Image resolution and other general settings should be opti-
mized on the particular microscope system for the samples to
be analyzed. However, resolution must be high enough to dis-
tinguish the plasma membrane from other intracellular organ-
elles, as this is a key consideration for this FRET application.
20. Select cells for analysis that have a well-defined plasma mem-
brane in the focal plane. Avoid any cells that move substantially
or change morphology during the time-lapse series. Avoid any
membrane regions with adjacent or fusing vesicles. Draw one
to three regions-of-interest (ROI) around plasma membrane
regions spanning in total at least 33% of the cell circumference.
21. When comparing FRET efficiency across different fusion con-
structs using the same donor-acceptor pair, compensation at
the calculation step for donor fluorophore photobleaching
that occurred during image acquisition is not necessary, pro-
vided that the donor expression level is nearly identical across
all constructs. This is because the donor fluorophore should
photobleach to the same extent across the different fusion con-
structs if the same imaging settings are used for each. This may
result in a slight over-estimation of FRET efficiency (although
for our system the “false FRET” caused by TFP photo-bleach-
ing was less than 1%), but this factor will be the same across all
constructs. It is not recommended to attempt to make com-
parisons between constructs with markedly different donor
expression, but if this is unavoidable, and the donor imaging
protocol must be altered to compensate for it, then a compen-
sation factor for donor photo-bleaching can be incorporated
into the calculations. Perform the image acquisition protocol
in the absence of R18 cell labeling and measure the average
drop in fluorescence per frame. Calculate the “false FRET”
value using the same equation presented in Subheading 3.1.8
as an average of 20 cells for each construct. Subtract this value
from the calculated quenching FRET efficiency for the appro-
priate construct. Similar considerations apply to “false FRET”
during dequenching FRET calculations, except that with only
two images acquired for this protocol, donor photo-bleaching
should be minimal if proper microscope settings are used.
22. The human T cell clones are restimulated on irradiated feeder
cells, which should die within 24 h after plating. So although
the only live cells in the culture are the human T cells, the pres-
ence of the dead feeder cells can complicate gating solely based
on FSC and SSC. A live-dead discrimination stain (e.g.,
Biolegend Zombie Dye or 7-AAD) may be used to improve
live cell identification.
23. We recommend freezing human T cell clones either in 90%
fetal bovine serum with 10% DMSO or in Bambanker cell
freezing solution (Wako Chemicals) with 5–10 × 106 cells per
aliquot in 1 mL.
24. Resuspending the cells at 1 × 106 cells/mL is ideal for imaging
and will facilitate cell preparation just prior to injection into
the fluidics chamber. IL-2 starvation enhances robust immu-
nological synapse formation with APLBs. Additionally, other

cell treatments (e.g., with the Lck inhibitor PP2) can be per-
formed during the IL-2 starvation step [16].
25. Lipid stocks solutions have a limited stability and should be
handled with care (maximum storage for 6 months) [14].
Stock solutions containing chloroform, as well as any rinses
containing chloroform, must be disposed of in accordance
with institutional guidelines. This organic waste can be mixed
with other phenol/chloroform waste from other applications
(e.g., molecular biology and protein biochemistry).
26. It is important to use the minimal sufficient length of tubing
for all connections in the setup to reduce the dead volume.
Start with longer pieces to test the layout, and then cut these
down where possible to minimize total tubing length. The
final layout may differ from what is shown in Fig. 2 depending
on the configuration of the microscope and available lab bench
space.
27. The sequence of line priming is crucial because it ensures that
the appropriate buffer will be present in the lines during the T
cell injection step.
28. Make sure to prime injection ports A1 and B1 by toggling the
valves on T-connectors A and B while injecting HBS-HSA
from injection port A2 to form a liquid droplet at the opening
of the injection ports. These will later be connected to the T
cell injection syringe (port A1) and the R18 labeling buffer
(port B1), and having appropriate buffer already present in the
injection port helps to ensure that no air bubbles will be intro-
duced to the fluidics system during later connection of the
syringes.
29. The desired concentration of beads on the coverslip is around
1 bead per 50 μm radius to ensure that most injected T cells
will interact with a single cell. APLB dilution and injection
volumes may need to be optimized to adjust for variations in
the amount of bead loss during lipid and protein coupling.
30. Cooling the flow chamber to 4 °C helps to reduce the move-
ment of proteins, cells, and conjugates. Maintaining a constant
flow of cold PBS through the chamber ensures that the system
remains at 4 °C. However, some peristaltic pumps may not be
compatible with FLIM applications. The key parameter in this
case is liquid pulsing, which is the change in hydraulic pressure
induced during the positive displacement provided by the roll-
ers in the pump head. This must be kept at a minimum during
imaging because it will change the distance from the cell to the
objective due to slight deformation of the glass caused by the
periodic change in pressure. Due to the long acquisition times,
FLIM is especially sensitive to this issue, which will cause evi-
dent blurring in the acquired data and could lead to artifacts in

data analysis. The BioRad Econo pump used in this protocol
had little liquid pulsing when pumping cold PBS at a flow rate
of 0.5 mL/min. If a suitable pump is unavailable, the pump
can be stopped briefly for FLIM data acquisition as long as the
cells have been properly cooled prior.
31. Data acquisition using the SPC-imaging module will need to
be optimized for the particular microscope setup and experi-
mental conditions. The appropriate length of data acquisition
is entirely dependent on photon concentration per pixel. This
is fundamentally determined by the fluorescent protein quan-
tum yield as well as the local concentration. Pilot studies should
be performed on resting T cells prior to advancing to experi-
ments on synapse formation and membrane binding dynamics
to determine the best settings. Keep in mind that to produce
high-quality data for FLIM-FRET analysis, between 5000 and
10,000 photons are needed per pixel (or binned pixels) to reli-
ably calculate dual (or greater) exponential decay curves and
fluorescent lifetimes of the short- and long-lived components.
These values are important because they enable the user to
determine not only the actual true FRET efficiency, but also
the fraction of fluorophores in FRET state. These values may
ultimately provide additional insight into the studied process.
32. Although image capture protocols should optimized for each
experimental application, we will describe the settings we
developed for this application of FLIM-FRET to T cell-APLB
conjugates [13]. These settings may be a good starting point
when acquiring FLIM data for cells expressing mTFP1 as
donor fluorophore expressed at medium levels. The Coherent
Ti sapphire laser was set at 6% and at 820 nm wavelength. The
scan speed was set at 12 (maximum speed) for 45 s with con-
tinuous scanning, a numerical zoom factor of 8 and 512 × 512
pixel resolution. This resulted in a pixel dwell time of 0.79 ms
and a 487 ms frame rate. These settings typically yielded
between 3 × 104 and 4 × 105 photons/s/frame. HPM-FLIM
detection was set at a 512 × 512 resolution and photons were
captured in a time-correlated fashion for 40 s. This approach
typically yielded 5–10 × 103 photons per (2–4×) binned pixels.
Following FLIM acquisition, the deflection plate was removed
and an image of the FLIM-acquired cell-APLB conjugate was
captured using the Argon (458 nm) laser set at 8% to excite
mTFP1 and fluorescence was captured using spectral detection
PMT (466–526 nm) with a 710 voltage gain at 1 Airy unit.
R18 image was captured using the diode pumped solid-state
laser (561 nm) set at 3% and spectral detection PMT (570–630
nm) with a 710 voltage gain at 1 Airy unit.
33. There are many other parameters that can be modified to

improve curve fitting. However, further modifications should
only be attempted with a clear understanding of what these
parameters control and how they impact FLIM calculations.
Further information can be found in the BH TCSPC hand-
book (http://www.becker-hickl.com/literature.htm#handb).
34. If there is a contribution from an FRET component in the
sample, then the curve fit as assessed by χ2 value should be
substantially worse compared to data collected for the non-
R18-labeled sample. In order to improve curve fitting, several
parameters may need to be optimized.
35. The choice of donor fluorophore is very important when set-
ting up an FLIM-FRET assay. To simplify downstream analy-
sis, it is preferable to select a donor fluorophore that emits
photons in a single exponential decay manner. Proteins that
emit photons in a dual exponential decay manner, either due to
homo-FRET or intrinsic fluorescence properties, will require
an even more complex decay function (e.g., three or four com-
ponents) to accurately model behavior under FRET condi-
tions. The added complexity may preclude reliable analysis
without deep knowledge of additional parameters for
optimization.
Acknowledgments
E.G. is supported by CIHR grant MOP- 133726 and NSERC

grant RGPIN-436183 and receives salary support from FRQS
grant F27316.
References
1. Xu C, Gagnon E, Call ME, Schnell JR, multilayered control of T cell receptor phos-
Schwieters CD, Carman CV, Chou JJ, phorylation. Cell 142(5):669–671
Wucherpfennig KW (2008) Regulation of T 5. Wucherpfennig KW, Gagnon E, Call MJ,
cell receptor activation by dynamic membrane Huseby ES, Call ME (2010) Structural biology
binding of the CD3epsilon cytoplasmic of the T-cell receptor: insights into receptor
tyrosine-based motif. Cell 135(4):702–713 assembly, ligand recognition, and initiation of
2. Aivazian D, Stern LJ (2000) Phosphorylation signaling. Cold Spring Harb Perspect Biol
of T cell receptor zeta is regulated by a lipid 2(4):a005140
dependent folding transition. Nat Struct Biol 6. Varma R, Campi G, Yokosuka T, Saito T,
7(11):1023–1026 Dustin ML (2006) T cell receptor-proximal
3. Duchardt E, Sigalov AB, Aivazian D, Stern LJ, signals are sustained in peripheral microclusters
Schwalbe H (2007) Structure induction of the and terminated in the central supramolecular
T-cell receptor zeta-chain upon lipid binding activation cluster. Immunity 25(1):117–127
investigated by NMR spectroscopy. 7. Saito T, Yokosuka T, Hashimoto-Tane A
Chembiochem 8(7):820–827 (2010) Dynamic regulation of T cell activation
4. Gagnon E, Xu C, Yang W, Chu HH, Call ME, and co-stimulation through TCR-
Chou JJ, Wucherpfennig KW (2010) Response microclusters. FEBS Lett 584(24):4865–4871
8. Dustin ML (2010) Insights into function of 13. Gagnon E, Schubert DA, Gordo S, Chu HH,
the immunological synapse from studies with Wucherpfennig KW (2012) Local changes in
supported planar bilayers. Curr Top Microbiol lipid environment of TCR microclusters regu-
Immunol 340:1–24 late membrane binding by the CD3epsilon
9. van Munster EB, Gadella TW (2005) Fluorescence cytoplasmic domain. J Exp Med
lifetime imaging microscopy (FLIM). Adv 209(13):2423–2439
Biochem Eng Biotechnol 95:143–175 14. Dustin ML, Starr T, Varma R, Thomas VK
10. Shikawa-Ankerhold HC, Ankerhold R, (2007) Supported planar bilayers for study of
Drummen GP (2012) Advanced fluorescence the immunological synapse. Curr Protoc
microscopy techniques--FRAP, FLIP, FLAP, Immunol Chapter 18:Unit 18 13
FRET and FLIM. Molecules 17(4):4047–4132 15. Schubert DA, Gordo S, Sabatino JJ Jr,
11. De Los SC, Chang CW, Mycek MA, Cardullo Vardhana S, Gagnon E, Sethi DK, Seth NP,
RA (2015) FRAP, FLIM, and FRET: detection Choudhuri K, Reijonen H, Nepom GT,
and analysis of cellular dynamics on a molecular Evavold BD, Dustin ML, Wucherpfennig KW
scale using fluorescence microscopy. Mol (2012) Self-reactive human CD4 T cell clones
Reprod Dev 82(7–8):587–604 form unusual immunological synapses. J Exp
12. Becker W, Su B, Holub O, Weisshart K (2011) Med 209(2):335–352
FLIM and FCS detection in laser-scanning 16. Nyakeriga AM, Garg H, Joshi A (2012) TCR-
microscopes: increased efficiency by GaAsP induced T cell activation leads to simultaneous
hybrid detectors. Microsc Res Tech 74(9): phosphorylation at Y505 and Y394 of p56(lck)
804–811 residues. Cytometry A 81(9):797–805
Chapter 17
Revealing the Role of Microscale Architecture in Immune

Synapse Function Through Surface Micropatterning
Joung-Hyun Lee and Lance C. Kam
Abstract
The immune synapse has emerged as a compelling example of structural complexity within cell-cell inter-
faces. This chapter focuses on the use of microcontact printing to isolate and investigate how spatial orga-
nization of signaling molecules drives the function of immune cells. In the process detailed here, multiple
rounds of microcontact printing are combined to create patterned surfaces that control the relative spatial
localization of CD3 and CD28 signaling in T cells, effectively replacing an antigen presenting cell with an
engineered surface. A set of approaches used to address key issues of T cell activation are described and
discussed.
Key words Microcontact printing, Soft lithography, Antibodies, Immunological synapse, Cell
activation
1 Introduction
A striking aspect of the immune synapse (IS) is that even within

the limited extent of this cell-cell interface, membrane receptor
proteins and cellular structures organize into patterns covering a
range of spatial scales. Initial studies identified a self-organizing
“bullseye” motif that became associated with functional activation,
consisting of a peripheral supramolecular activation cluster
(pSMAC) rich in LFA-1 surrounding a central supramolecular
activation cluster (cSMAC) containing TCR [1, 2]. Subsequent
studies identified variations on this pattern in interfaces formed
between different cell types as well as additional micro- and nano-
scale structures that govern IS function [3–11]. Understanding
the molecular mechanisms that drive these structures is essential
for fully describing and ultimately directing cell-cell communica-
tion through the IS.
In this direction, we describe here a method for controlling the
microscale organization of signaling complexes within artificial IS
analogues formed by T cells, replacing the antigen presenting cell
291
292 Joung-Hyun Lee and Lance C. Kam
Fig. 1 (a) Patterning multiple ligands to cell surface receptors on a single surface provides a robust approach
to understanding how microscale architecture of the immune synapse (IS) drives cell function. In this example,
independent patterns of anti-CD3 (epsilon subunit, targeting TCR signaling), anti-CD28, and ICAM-1 are com-
bined and presented to T cells, forming artificial IS structures. (b) Implementation of the multicomponent,
micropatterned surface illustrated in panel A. This surface presents arrays of segregated (SEG) costimulation
sites, consisting of a central feature with anti-CD3 (red) surrounded by a constellation of smaller anti-CD28
(green) features. These patterns were created using two microcontact printing steps. The surface was sub-
sequently back-filled with ICAM-1. (c) Mouse T cells (mixed CD4+ and CD8+) adhere to and stop migration
upon features of anti-CD3 (pattern of larger dots, red). Microscopy-based observation of cells using probes
normally used in flow cytometry allows cell-by-cell analysis of cell identity and function (in this case, secretion
of the cytokines IL-2 and IFN-γ)
(APC) with a patterned surface (Fig. 1a). This approach is based

on microcontact printing, a soft-lithography method in which an
elastomer stamp is used to pattern a desired molecule onto a work-
ing surface. Since its introduction by the Whitesides group [12–
15], this method has expanded beyond the initial patterning of
alkanethiols to include proteins and other biomolecules onto a
wide variety of surfaces [16–21]. Applied to ligands recognized by
a target cell, this approach allows control over the spatial organiza-
tion of receptor activation to an extent that is difficult to obtain in
models based on cell-cell interactions.
The use of micropatterned surfaces to study the immune syn-
apse poses several challenges that are not fully developed in other
systems. Most prominently, microscale organization of the IS is
often described with regards to engagement of multiple biomo-
lecular ligands. The additive nature of microcontact printing is well
Micropatterning Artificial Immune Synapses 293
suited for creating multicomponent surfaces containing indepen-

dent patterns of biomolecules, each applied in a separate process-
ing step. However, this printing process involves drying of an
“inking” solution onto the stamp, a step that is detrimental to
many biomolecules [22, 23]. To address this issue, the method
described here incorporates an affinity capture approach. Several
proteins, including antibodies, Protein A, and streptavidin, retain
significant binding activity after patterning by microcontact print-
ing and also provide the ability to recognize an epitope engineered
into a bioactive ligand [24, 25]. By modifying ligands with an
affinity epitope tag, Fc region, or biotin, it is possible to pattern
these molecules by first microcontact printing one or more affinity
proteins onto a surface. The modified ligands are then patterned
onto the surface in a bath-application step. In addition, some
receptors of interest to the IS and cell activation (including CD3
and CD28) can be activated by crosslinking with antibodies; pat-
terning antibodies to these targets provides a direct way of control-
ling the localization of receptor activation. An additional
consideration of particular importance to the IS is that cell response
is often sensitive to the concentration of ligands, and frequently
biphasic, making control over the local surface density of each pro-
tein important. In principle, this can be controlled by changing the
concentration of protein in the inking solution. However, varia-
tions elastomer curing, humidity, stock solution concentration,
and other processing conditions make this an unreliable approach.
The protocol described here combines two experimental approaches
to circumvent these issues. First, the inking solution is prepared at
a concentration of protein that saturates the stamp surface and
subsequently the concentration of antibody transferred to the
working substrate; by remaining in a saturating regime, variations
in protein concentration, volume, or other factors are minimally
reflected in the amount of protein transferred onto the substrate.
Second, the surface concentration of an active antibody (such as
anti-CD3) is controlled by mixing with an inert antibody in the
inking solution. The ratio of the two variants in the stamped anti-
body reflects the relative composition in the stamping solution;
while exceptions to linear mixing have been reported [26] and the
final concentration must be verified, this provides a more repro-
ducible approach than stamping only the target antibody and vary-
ing inking concentration at sub-saturating levels.
We combined these approaches in a series of studies examining
how the microscale organization of CD3 and CD28 activation
influences T cell function. These studies used glass coverslips pat-
terned with arrays of costimulation sites, focusing on two specific
layouts of the sites. The first (Fig. 1a, b) consisted of a single, 2-μm
diameter feature presenting an activating antibody to CD3. This
was surrounded by a set of four features, each measuring 1 μm in
diameter, which presented an activating antibody to CD28. The
spaces between these features were backfilled with ICAM-1. The

microscale separation provided by these surfaces was intended to
capture the segregation of signaling complexes observed between
T cells and artificial APCs, and is denoted as a SEG pattern in this
chapter. In the second layout, anti-CD3 and anti-CD28 were pre-
sented together in a single, central feature; this colocalized (COL)
motif omitted the array of smaller features that made up the SEG
pattern. With these surfaces, we revealed that microscale separation
of CD3 and CD28 signaling abrogated functional activation of
conventional human CD4+ T cells, which were able to respond to
the COL pattern [27], as measured by cytokine secretion.
Surprisingly, cells isolated from mice were comparatively insensitive
to the relative position of CD3 and CD28 engagement [11].
Finally, regulatory T cells from mice were able to respond to segre-
gation/colocalization of these signals, in contrast to conventional
cells [28]. Notably, the use of microscopy-compatible coverslips
facilitated correlation of cytokine secretion with a specific cell and
pattern, an approach that can be expanded to include multiple bio-
markers (Fig. 1c). The protocol presented here describes with
practical detail how issues like activating antibody concentration
and the ability to work with small culture volumes can be used in
working with micropatterned surfaces. This example focuses on
proteins, concentrations, and procedures used for studies with
mouse cells, but can be readily adapted for other systems.
2 Materials
2.1 Stamp Masters The elastomer stamps used for patterning proteins are typically cast
from topological masters. For our studies, which involved rela-
tively small features of micrometer-scale dimension, these masters
were fabricated on silicon wafers by e-beam lithography with 1 μm
thick PMMA. One set of masters contained arrays of dots measur-
ing 2 μm in diameter while a second contained clusters of 1 μm
dots. Production of SEG surfaces combined stamps cast from both
masters, while the COL pattern used only the larger, 2 μm diame-
ter dots. Additional patterns, which served as control conditions,
used variations of these stamps and protein coating.
We do not describe here the procedures for master fabrication,
as they will vary extensively based on microfabrication equipment
available to the researcher. In addition, different procedures for
making these topological masters have been described previously
[15, 29–31]; the focus of this chapter is on adapting these methods
for the issues relevant to capturing the complexity of the immune
synapse.
2.2 Silanization 1. Trichloro(1H,1H,2H,2H,-perfluorooctyl)silane (e.g.,

Materials Sigma-Aldrich).
2.3 PDMS Stamps h-PDMS (Hard PDMS) component.

1. 7–8% vinylmethylsiloxane-dimethyl-siloxane copolymer (e.g.,
Gelest).
2. platinum-2,4,6,8-tetramethyl-2,4,6,8-tetravinylcyclotetrasi-
loxane complex solution (e.g., Sigma-Aldrich).
3. 1,3,5,7-tetravinyl-1,3,5,7-tetramethylcyclotetrasiloxane (e.g.,
Gelest).
4.
25–30% methylhydrosiloxane-dimethylsiloxane copolymer
(e.g., Gelest).
s-PDMS (Soft PDMS) component.
5. Sylgard 184 (Dow Corning) or equivalent polydimethylsilox-
ane formulation.
6. Biopsy punch, 7 mm and 12 mm (e.g., Acuderm).
2.4 Antibodies, 1. Anti CD3, (specific to mouse CD3) clone 145-2C11 (e.g.,
Proteins eBioscience).
2.4.1 Antibodies 2. Anti-CD28 (specific to mouse CD28), clone 37.51 (e.g.,
eBioscience).
3. Inert antibody, e.g., Goat anti-Chicken IgY (e.g., Novex).
4. Anti-phospho-Lck (pLck Tyr394, Rabbit polyclonal IgG, e.g.,
Santa Cruz Biotechnology, SC-101728).
5. Anti-PKC-θ (Rabbit polyclonal IgG, e.g., Santa Cruz
Biotechnology, SC212).
6. Donkey anti-Rabbit IgG Secondary antibody Alexa Fluor 488
conjugate (e.g., Molecular Probes).
2.4.2 ICAM-1 1. Recombinant ICAM-1/CD54- Fc Chimera (e.g., R&D Systems).
2.4.3 Fluorophors 1. Alexa Fluor 647 Carboxylic acid succinimidyl ester (e.g.,
for Labeling Antibodies Molecular Probes).
2. Alexa Fluor 488 Carboxylic acid succinimidyl ester (e.g.,
Molecular Probes).
3. Alexa Fluor 568 Carboxylic acid succinimidyl ester (e.g.,
Molecular Probes).
2.4.4 Dialysis 1. Slide-A-Lyzer Mini Dialysis Device, 10 k MWCO (e.g.,

ThermoFisher) or equivalent.
2. Dialysis Device Float Buoys (e.g., Thermofisher).
2.5 Glass Coverslips 1. 22 × 22 mm thickness no. 2 glass coverslip (e.g., Electron
Microscopy Sciences).
2. 7× cleaning solution (MP Biomedicals), or equivalent.
3. Ceramic rack (e.g., Thomas Scientific).
2.6 General 1. Small Benchtop Furnace (450 °C for >6 h), Barnstead/

Materials Thermolyne Type 1300 (VWR) or equivalent.
and Equipment 2. Centrifuge.
3. Microscope.
4. Rotary vaccum pump.
5. Round vaccum desiccator (Volume ~6 L).
6. Hot Plate/Stirrer.
7. 150 cm petri-dish.
8. Magnetic stir-bars.
9. Beakers (1 L, 150 mL).
10. pairs of Tweezers.
11. N2 gas to blow.
12. Weights for stamping (Steel Hex Nuts of about 5 g would
work well).
13. Biopsy Punches, 7 mm and 12 mm diameter (Acuderm Inc.)
or equivalent.
14. Falcon tubes, 15 and 50 mL.
15. Eppendorf tubes, 1.5 mL.
16. Phosphate buffered saline (PBS).
17. Deionized water (DIW).
18. Isopropyl alcohol (IPA).
19. Bovine Serum Albumin (BSA).
20. Parafilm.
3 Methods
This section first described the general steps needed to carry out
microcontact printing of antibodies. Subheading 3.8 describes
how the general steps are combined to prepare the SEG layout
(Fig. 1a, b). The last subsections describe modifications of this
process which are needed to create the simpler COL pattern, along
with other controls.
3.1 Silanization This step helps to release for cured PDMS from the master.
of Masters
1. Place a silicon master in a vacuum desiccator along with
0.5 mL of Trichloro(1H,1H,2H, 2H,-perfluorooctyl)silane in
a 25 mm petri-dish.
2. Close the lid of the desiccator and operate the rotary pump until
vacuum reaches about 25 in. Hg which takes about 20 min.
3. Leave it as is for an hour with valves closed and pump off.
3.2 Preparation Fabrication of a two-layer stamp (Fig. 2a) is recommended for

of PDMS Stamps molding PDMS stamps from masters that have the aspect ratio of
patterns higher than 0.5 or relatively large spacing between fea-
tures due to the deformation [32–34]. The first layer is composed
of thin hard PDMS and the second layer is soft PDMS
component.
1. Clean a master wafer by rinsing with isopropyl alcohol, blow
dry with N2 gas, and then leave in an oven at 60 °C for about
5 min or until completely dry.
2. Place the masters in an aluminum foil-lined 150 mm petri-dish
and set aside.
3. In a 50 mL falcon tube, add the h-PDMS prepolymer with
catalyst and modulator.
Fig. 2 (a) Process flow for the preparation of the SEG pattern by microcontact
printing. (b–e) Bench-scale examples of key microcontact printing steps, illus-
trating bench-level implementation. Panel B illustrates a chamber used to ink
stamps with antibodies, including a wet towel to reduce evaporation during the
coating process. Panel C illustrates the inking process, focusing on two of the
drops shown in Panel B. Panel D illustrates the use of a hex nut to apply force
to an inked stamp, printing antibodies onto a working surface. Once all micro-
contact printing processes are completed, a PDMS ring is applied onto the pat-
terned surfaces, allowing final protein modification steps (such as backfill or
affinity capture) or cell culture in a small, controlled volume
a. 3.5 mL of 7–8% vinylmethylsiloxane-dimethyl-siloxane

copolymer.
b. 18 μl of platinum-2,4,6,8-tetramethyl-2,4,6,8-tetravinylcy-
clotetrasiloxane complex solution.
c. 10 μl of 1,3,5,7-tetravinyl-1,3,5,7-tetramethylcyclotetrasilo-
xane.
7. Mix well by vortexing.
8. Add the second copolymer. 1 mL of 25–30% methylhydrosi-
loxane-dimethylsiloxane copolymer.
10. Mix well by vortexing.
11. Centrifuge at 1000 rcf for 3 min to degas.
12. Immediately form a thin layer (30–40 μm) of h-PDMS onto
the master cleaned above (step 2 in this section) by spin-
coating (1000 rpm, 40 s) (see Note 1).
13. Bake at 60 °C for 30 min.
14. Prepare 55 g of Sylgard 184 in a disposable cup (~ 300 mL) by
hand-mixing the base polymer and curing reagent in a 10:1
ratio by weight with a long rod and then degassing by centrifu-
gation at 1000 rcf for 3 min.
15. Pour the degassed s-PDMS prepolymer to about 4–5 mm
thickness (~ 50 g) on the h-PDMS-coated master in the
petri-dish.
16. Place the master in the petri-dish in a vacuum desiccator.
17. Maintain vacuum for about 30 min until no more bubbles are
generated from the s-PDMS.
18. Bake at 60 °C oven for about 3 h.
19. Pick up the aluminum foil attached PDMS and master from
the petri-dish and remove the aluminum foil carefully from the
master and PDMS.
20. Peel the completed 2-layered PDMS off the master very gently
and place the completed molded PDMS replica pattern-side up
in a 150 mm petri-dish (see Note 2).
3.3 Preparation 1. Place coverslips (22 × 22 mm, No. 2 thickness) on a ceramic

of Coverslips rack.
2. Make 1× detergent by diluting 7× detergent with DIW. Make
enough detergent to submerge coverslips in a 500 mL beaker.
(The detergent will be cloudy).
3. Carefully place a magnetic stir-bar and the rack in the beaker
with 1× detergent.
4. Place the beaker on a hot plate/stirrer and boil while stirring
at a speed of about 250 rpm.
5. Keep the solution boiling until the solution becomes transpar-

ent, which takes about 30–40 min. Maintain conditions for
another 30 min.
6. Wash coverslips in the rack in DIW by repetitive vigorous dip-
ping. Repeat while replacing DIW three times.
7. Gently blow N2 gas over the coverslips to remove water and
put the rack in a small beaker that will fit in the furnace.
8. Cover the beaker top with aluminum foil and place the beaker
in the furnace.
9. Heat the furnace to 450 °C and maintain this temperature for
at least 6 h.
10. Let the coverslips cool down to room temperature before use.
11. Store coverslips in a clean and dry environment. The coverslips
should be used preferentially within 2–3 weeks.
3.4 Preparation 1. Prepare 35 g of Sylgard 184 by mixing the base polymer and
of PDMS Rings curing reagent with the 10:1 ratio with a rod and then degas-
to Form Small Wells sing by centrifugation at 1000 rcf for 3 min.
2. Pour degassed PDMS prepolymer to about 2 mm thickness (~
30 g) in an empty 150 mm petri-dish.
3. Place the petri-dish in a vacuum desiccator.
4. Maintain vacuum for about 30 min until no more bubbles are
generated from the s-PDMS.
5. Bake at 60 °C oven for about 3 h.
6. Use 7 and 12 mm biopsy punches to make PDMS rings (7 mm
inner diameter and 12 mm outer diameter) out of the flat
PDMS in the dish.
7. Wash rings with IPA and blow dry with N2 gas.
8. These rings will be placed on antibody-stamped glass cover-
slips that will be described later in this paper, forming small-
volume wells (~80 μL) as shown in Fig. 2e.
3.5 Labeling Anti-CD3 (150 kDa, 1 mg/mL) and anti-CD28 (150 kDa, 1 mg/

of Antibodies mL) will be labeled with Alexa fluor 568 (MW 791.8, 10 mg/mL)
and Alexa fluor 647 (MW 1250, 10 mg/mL) respectively with
eight folds of molar ratio of fluorophore per antibody.
1. Mix 100 μL of anti-CD3 stock solution with 0.42 μL of Alexa
Fluor 568 in a 1.5 mL Eppendorf tube. Low vortexing follow-
ing vigorous pipetting up and down is recommended to mix
them thoroughly. Keep the mixture in room temperature for
1 h while protected from ambient light. In a separate tube, mix
100 μL of anti-CD28 stock solution with 0.60 μL of Alexa
Fluor 647 in 1.5 mL Eppendorf tube and keep in room tem-
perature for 1 h while protected from ambient light.
2. Place the antibodies in mini dialysis tubes on a float buoy.

3. Perform dialysis against 1 L of 1× PBS while stirring with a
magnetic stir bar at 4 °C for at least 6 h. Keep the beaker pro-
tected from light.
4. Collect the antibodies, aliquot them, and store them in
−20 °C. Avoid repeated freeze and thaw cycles.
Inert antibody (Goat anti-Chicken IgY, 2 mg/mL,
180 kDa) will be labeled and dialyzed the same way as above
with both fluorophores separately.
3.6 Preparation 1. COL ink (200 μL): mixture of 5 μg/mL anti-CD3 and 15 μg/
of Antibody mL anti-CD28 in PBS (total 20 μm/mL). To achieve this, add
and Protein Solutions the following in a 1.5 mL tube and mix well:
(a) 196 μL of PBS.
(b) 0.5 μL of Alexa-568 labeled anti-CD3.
(c) 0.5 μL of non-labeled anti-CD3.
(d) 1.5 μL of Alexa-647 labeled anti-CD28.
(e) 1.5 μL of non-labeled anti-CD28.
2. Anti-CD3 ink (400 μL): mixture of 5 μg/mL anti-CD3 and
15 μg/mL of inert antibody (anti-Chicken IgY) in PBS. To
achieve this, add the following in a 1.5 mL tube and mix well:
(a) 395 μL of PBS.
(b) 1 μL of Alexa-568 labeled anti-CD3.
(c) 1 μL of non-labeled anti-CD3.
(d) 1.5 μL of Alexa-568 labeled Goat anti-Chicken IgY.
(e) 1.5 μL of non-labeled Goat anti-Chicken IgY.
3. Anti-CD28 ink (400 μL): mixture of 15 μg/mL anti-CD28
and 5 μg/mL of inert antibody in PBS. To achieve this, add
the following in a 1.5 mL tube and mix well:
(a) 393 μL of PBS.
(b) 0.5 μL of Alexa-647 labeled Goat anti-Chicken IgY.
(c) 0.5 μL of non-labeled Goat anti-Chicken IgY.
(d) 3 μL of Alexa-647 labeled anti-CD28.
(e) 3 μL of non-labeled anti-CD28.
4. 1 mL of 2 μg/mL ICAM-1 in PBS. Prepare stock solution of
400 μg/mL ICAM-1 in PBS. Dilute 200:1 ratio in PBS before
use.
5. 4% BSA solution in PBS.
3.7 Stamping— 1. Line a 150 mm petri-dish with a 10 × 10 cm2 piece of parafilm.

Inking the Stamps 2. Place wet Kimwipes next to the walls of the petri-dish. This is
to humidify the dish and reduce antibody solution evaporation
during inking steps (Fig. 2b).
3. Cut stamps with the desired pattern into 5 × 5 mm2 size with
a razorblade.
4. With tweezers, pick up a stamp, wash it with IPA by vigorous
dipping motion, and dry it with N2 gas. Repeat this step with
all stamps to remove any h-PDMS debris from the stamp
surface.
5. On the above petri-dish lined with a piece of parafilm, carefully
place 30 μL drops of antibody solutions (detailed in Subheading
3.6) about 1 in. apart (Fig. 2b).
6. Place stamps gently on the antibody drops (Fig. 2c).
7. Cover the dish with a lid and protect dish from light. Leave the
stamps on antibody solutions for 40 min–1 h. This procedure
is to transfer antibodies to the surfaces of the PDMS stamps
(i.e., inking).
8. Clean bench top with 70% alcohol and prepare the following
items:
●● Weights described in Subheading 2.6.
●● Glass coverslips, cleaned as described above Subheading
3.3.
●● Tweezers.
●● Rinsing solutions, 1× PBS and DIW in 50 mL falcon tubes.
●● A new petri-dish lined with parafilm.
●● PDMS rings.
●● ICAM-1 solution.
●● N2 gas.
●● Timer.
3.8 Stamping—SEG 1. Place glass coverslips on the cleaned bench top.

Pattern (Focusing 2. Ink two stamps.
on Patterns for Mouse ●● A stamp containing arrays of 2-μm diameter dots with
Cells)
anti-CD3 ink (see Subheading 3.6).
●● A stamp containing clusters of 1-μm diameter dots with
anti-CD28 ink (see Subheading 3.6).
3. Pick up an anti-CD3 inked stamp from the petri-dish and wash
in 1× PBS and then DIW by vigorous dipping several times in
respective rinsing solution.
4. Blow dry with N2 gently to remove excess liquid on a stamp.
5. Bring the stamp gently on a coverslip (Fig. 2d).
6. VERY gently push the stamp toward the coverslip using twee-
zers to form contact. Do NOT push hard (see Note 3)!
7. Mark with a permanent pen near the corners of the stamp to
mark the stamp location. Do not touch the stamp (Fig. 2d).
8. Gently place a weight on the top of the stamp (Fig. 2d).

Maintain weight for 5 min.
9. Gently remove the weight from the stamp. With tweezers,
remove the stamp from the coverslip.
10. Pick up an anti-CD28 inked stamp from the dish and wash in
respective rinsing solutions.
11. Gently blow dry with N2 gas to remove excess liquid from the
stamp.
12. Gently place the anti-CD28 inked stamp on the coverslip
where anti-CD3 has been stamped. It is desirable to stamp
anti-CD28 slightly tilted (~ 15°) from anti-CD3 stamped area.
13. VERY gently push the stamp toward the coverslip to form con-
tact. Do NOT push hard!
14. Gently place a weight on the top of the stamp. Maintain weight
for 5 min.
15. Gently remove the weight from the stamp. With tweezers,
remove the stamp from the coverslip.
16. Place a PDMS ring around the stamped area. Confirm the con-
tact of the ring to the coverslip by pushing the ring with twee-
zers (Fig. 2e).
17. Place the coverslip on a new parafilm-lined petri-dish set aside
in advance. Place wet Kimwipes in the dish edge to humidify
the dish.
18. Add 80 μL ICAM-1 to the well. Cover the dish and leave it for
1–2 h while protecting the dish from light.
3.9 Stamping—COL Creating these patterns follows the same procedure used for creat-
Pattern ing the SEG pattern (Subheading 3.8), but using only the first
stamp, inked with the COL ink solution (Subheading 3.6). Steps
10–15 of Subheading 3.8 are omitted.
3.10 Stamping— An important control condition is the patterning of anti-CD3

CD3-Only Patterning alone, omitting anti-CD28 and thus revealing the role of present-
ing costimulatory signals to the cell. For this pattern, follow the
procedure presented for the SEG pattern (Subheading 3.8), but
omit the CD28-inked stamp (steps 10–15 of that section).
4 Seeding Cells and Immunostaining
4.1 Seeding T cells can be isolated from mouse lymph nodes, spleen, or human
blood by fluorescent-activated cell sorting (FACS) method. If FACS
sorter is not available, target cells can be isolated by enrichment kits.
Follow directions from isolation kit manufacturers (e.g., untouched
mouse CD4+ T cells using Thermofisher 11415D, mouse

CD4+CD25+ regulatory T cells using Miltenyi 130–091-041).
1. When cells are ready to be plated, remove ICAM-1 from the
wells and rinse with 1× PBS and culture media.
2. Seed approximately 150 thousand cells per well. Prepare cell
solution of about two million cells per mL and plate 80 μL of
cell solution per well (see Note 4).
3. Incubate the cells on patterns at 37 °C and 5% CO2 for desired
amount of time (e.g., for studying phospho-Lck or PKC-θ
recruitment and localization, 15–30 min of incubation time is
desirable. For IL-2 secretion, minimum of 6-hour incubation
is required [11]).
4.2 Immunostaining 1. Fix samples with 4% paraformaldehyde (PFA) when incubation

time is up. Gently remove half of the media and add 80 μL of
4% PFA along the rim of the PDMS rings. Wait for 15 min.
2. Permeabilize the cells with 0.1% Triton-X. Similar to fixing,
remove half of the PFA and add 80 μL of 0.1% Triton-X. Wait
for 10 min.
3. Wash with 1× PBS. With a similar fashion, gently remove half
of the liquid and add PBS. Repeat this PBS washing with 1 mL
of PBS per well.
4. In order to ensure specific antibody binding, block additional
sites using 4% BSA for 1–2 h by replacing PBS with 4% BSA.
5. Prepare the primary antibody (such as anti-p-Lck, anti-PKC-θ,
etc.) by diluting it in 4% BSA. Follow the instructions per the
antibody for dilution factor. In general, expect a 1:50–1:200
dilution for primary antibodies. Remove half of the 4% BSA
from wells and add 80 μL of the primary antibody solution.
Let it sit for 1 h.
6. Wash with PBS multiple times and replace with 4% BSA.
7. [Optional] If using a secondary antibody (such as Donkey
anti-Rabbit IgG Alexa 488 conjugated), prepare it in 4%
BSA. In general, expect a higher dilution factor than primaries
(1:200–1:500). Remove half of the 4% BSA from wells and
add 80 μL of the secondary antibody solution. Let it sit for
30 min.
8. Wash with PBS.
9. Fix with 4% PFA again for 10 min. This second fix ensures a
better stain.
10. Wash with PBS.
11. Seal the petri-dish with parafilm and keep it 4 °C until analysis.
It is recommended to image within 2–3 days.
12. [Optional] For prolonged storage of samples, samples can be

mounted with mounting media. Gently remove PBS from the
wells and remove PDMS rings. Apply 10–20 μL of mounting
media to a glass slide, invert the sample coverslip and seal the
edges with nail polish.
5 Discussion and Conclusion
Microcontact printing provides a robust approach for understand-

ing how the microscale architecture of the IS drives cell function.
Compared to alternative, emerging strategies for creating multi-
component patterns, microcontact printing requires less use of
specialized equipment and tooling, and is well suited for creating
multiple substrates as is needed for studies of cell function.
Limitations of this approach include the precision to which each
pattern can be aligned. Each printing step is well suited for pattern-
ing a single molecule. While approaches for mechanically aligning
patterns have been described [25], proper registration at the scale
of micrometers, as is needed to properly describe IS layout, remains
elusive. A second limitation of microcontact printing as described
here is that proteins are adsorbed on the surface and not covalently
attached. T cells in our studies were not able to significantly
remodel these surfaces or remove proteins, but other cells are more
aggressive. For such cells covalent attachment can be effected by
stamping proteins onto amino-modified surfaces, followed by
crosslinking or other forms of linkage. The counterpart of the pro-
tein patterns being stable enough for work with T cells is that
microcontact printed proteins lack the lateral mobility that is asso-
ciated with cell membranes and provided by alternative systems
such as supported lipid bilayer. The emergence of techniques for
patterning these bilayers provides a complementary tool for inves-
tigating IS structure [35, 36].
In summary, microcontact printing, as part of a set of tech-
niques for micro−/nano-engineering surfaces, provides a powerful
tool for investigating spatial organization of the IS. In comparison
to approaches to perturb IS structure by modifying the APC and
proteins presented on these cells, surface patterning avoids unin-
tended consequences of such modification. In addition, these pat-
terning techniques allow exploration of geometries and conditions
not achievable in the natural cell-cell interface.
6 Notes
1. Alternatively, a thin layer can be achieved without a spin-coater.

Pour a dollop (1–2 mL) of the h-PDMS mixture onto the mid-
dle of the master and use a piece of folded parafilm to evenly
spread the h-PDMS mixture. Take care not to leave thick

streaks of the PDMS directly on the top of the pattern as this
will result in an uneven thickness after curing and the pattern
will not be stamped properly in that area.
2. Cured h-PDMS is brittle. Do not bend too much. Severe
bending can cause cracks and debris. Cut the stamps of desired
size with a razorblade.
3. You will be able to tell if the stamp is in contact with the cov-
erslip by change in color at the interface.
4. If you are dealing with rare cells such as regulatory T cells or
patients’ cells, you can make smaller PDMS rings (i.e., inner
diameter of 5 mm) to seed small number of cells. However,
keep in mind that evaporation can be an issue in small volume
wells. For extremely rare cells, microfluidic isolation chambers
can be an alternative option [28].
Acknowledgments
This work was supported by the National Institutes of Health,

R01AI088377 and U24AI118669.
References
1. Grakoui A, Bromley SK, Sumen C, Davis MM, 7. Krummel MF, Macara I (2006) Maintenance
Shaw AS, Allen PM, Dustin ML (1999) The and modulation of T cell polarity. Nat Immunol
immunological synapse: a molecular machine 7(11):1143–1149
controlling T cell activation. Science 8. Kupfer A (2006) Signaling in the Immunological
285(5425):221–227 Synapse: Defining the Optimal Size. Immunity
2. Monks CR, Freiberg BA, Kupfer H, Sciaky N, 25(1):11–13
Kupfer A (1998) Three-dimensional segregation 9. Davis DM (2006) Intrigue at the immune syn-
of supramolecular activation clusters in T cells. apse. Sci Am 294(2):48–56
Nature 395(6697):82–86. doi:10.1038/25764 10. Doh J, Irvine DJ (2006) Immunological syn-
3. Bromley SK, Burack WR, Johnson KG, apse arrays: Patterned protein surfaces that
Somersalo K, Sims TN, Sumen C, Davis MM, modulate immunological synapse structure
Shaw AS, Allen PM, Dustin ML (2001) The formation in T cells. PNAS 103(15):5700–
immunological synapse. Annu Rev Immunol 5705. doi:10.1073/pnas.0509404103
19:375–396 11. Shen K, Thomas VK, Dustin ML, Kam LC
4. Davis MM, Krogsgaard M, Huppa JB, Sumen (2008) Micropatterning of costimulatory
C, Purbhoo MA, Irvine DJ, Wu LC, Ehrlich L ligands enhances CD4+ T cell function. Proc
(2003) Dynamics of Cell Surface Molecules Natl Acad Sci 105(22):7791–7796.
During T Cell Recognition. Annu Rev doi:10.1073/pnas.0710295105
Biochem 72(1):717–742 12. Kumar A, Whitesides GM (1993) Features of
5. Groves JT, Dustin ML (2003) Supported pla- gold having micrometer to centimeter dimen-
nar bilayers in studies on immune cell adhesion sions can be formed through a combination of
and communication. J Immunol Methods stamping with an elastomeric stamp and an
278(1–2):19–32 alkanethiol "Ink" followed by chemical etch-
6. Friedl P, den Boer AT, Gunzer M (2005) ing. Appl Phys Lett 63:2002
Tuning immune responses: diversity and adap- 13. Mrksich M, Chen CS, Xia Y, Dike LE, Ingber
tation of the immunological synapse. Nat Rev DE, Whitesides GM (1996) Controlling cell
Immunol 5(7):532 attachment on contoured surfaces with self-
assembled monolayers of alkanethiolates on gold. egy for microcontact printing. Neurochem Res
Proc Natl Acad Sci U S A 93(20):10775–10778 28(11):1639–1648
14. Xia Y, Kim E, Mrksich M, Whitesides GM 25. Shi P, Shen K, Kam L (2007) Local presen-
(1996) Microcontact printing of alkanethiols tation of L1 and N-cadherin in multicompo-
on copper and its application in microfabrica- nent, microscale patterns differentially direct
tion. Chem Mater 8(3):601–603. doi:10.1021/ neuron function in vitro. Dev Neurobiol
cm950464+ 67(13):1765–1776
15. Mrksich M, Dike LE, Tien J, Ingber DE, 26. Fuertes Marraco SA, Baumgaertner P, Legat A,
Whitesides GM (1997) Using microcontact Rufer N, Speiser DE (2012) A stepwise proto-
printing to pattern the attachment of mamma- col to coat aAPC beads prevents out-
lian cells to self-assembled monolayers of alka- competition of anti-CD3 mAb and consequent
nethiolates on transparent films of gold and experimental artefacts. J Immunol Methods
silver. Exp Cell Res 235(2):305–313 385(1–2):90–95. doi:10.1016/j.jim.2012.07.
16. James CD, Davis RC, Kam L, Craighead HG, 017S0022-1759(12)00228-1 [pii]
Isaacson M, Turner JN, Shain W (1998) 27. Bashour KT, Tsai J, Shen K, Lee JH, Sun E,
Patterned protein layers on solid substrates by Milone MC, Dustin ML, Kam LC (2014)
thin stamp Microcontact Printing. Langmuir Crosstalk between CD3 and CD28 is spatially
14:741–744 modulated by protein lateral mobility. Mol Cell
17. St. John PM, Davis R, Cady N, Czajka J, Batt Biol 34(6):955–964
CA, Craighead HG (1998) Diffraction-based 28. Lee JH, Dustin ML, Kam LC (2015) A micro-
cell detection using a microcontact printed anti- fluidic platform reveals differential response of
body grating. Anal Chem 70(6):1108–1111 regulatory T cells to micropatterned costimula-
18. James CD, Davis R, Meyer M, Turner A, tion arrays. Integr Biol (Camb) 7(11)
Turner S, Withers G, Kam L, Banker G, 29. Chen CS, Mrksich M, Huang S, Whitesides GM,
Craighead H, Isaacson M, Turner J, Shain W Ingber DE (1997) Geometric control of cell
(2000) Aligned microcontact printing of life and death. Science 276(5317):1425–1428
micrometer-scale poly-L-lysine structures for 30. Desai RA, Khan MK, Gopal SB, Chen CS
controlled growth of cultured neurons on pla- (2011) Subcellular spatial segregation of integ-
nar microelectrode arrays. IEEE Trans Biomed rin subtypes by patterned multicomponent sur-
Eng 47(1):17–21 faces. Integr Biol (Camb) 3(5):560–567.
19. Kung LA, Kam L, Hovis JS, Boxer SG doi:10.1039/c0ib00129e
(2000) Patterning hybrid surfaces of pro- 31. Tan JL, Tien J, Chen CS (2002) Microcontact
teins and supported lipid bilayers. Langmuir printing of proteins on mixed self-assembled
16(17):6773–6776 monolayers. Langmuir 18:519–523
20. Hovis JS, Boxer SG (2001) Patterning and 32. Bietsch A, Michel B (2000) Conformal contact
composition arrays of supported lipid bilay- and pattern stability of stamps used for soft
ers by microcontact printing. Langmuir lithography. J Appl Phys 88(7):4310–4318
17:3400–3405 33. Lee TW, Mitrofanov O, Hsu JWP (2005)
21. Tsai J, Kam L (2009) Rigidity-dependent cross Pattern-transfer fidelity in Soft lithography: the
talk between integrin and cadherin signaling. role of pattern density and aspect ratio. Adv
Biophys J 96(6):L39–L41 Funct Mater 15(10):1683–1688
22. Kam L, Shain W, Turner JN, Bizios R (2001) 34. Odom TW, Love JC, Wolfe DB, Paul KE,
Axonal outgrowth of hippocampal neurons on Whitesides GM (2002) Improved pattern
micro-scale networks of polylysine-conjugated transfer in soft lithography using composite
laminin. Biomaterials 22(10):1049–1054 stamps. Langmuir 18(13):5314–5320
23. Shi P, Nedelec S, Wichterle H, Kam LC (2010) 35. Mossman KD, Campi G, Groves JT, Dustin
Combined microfluidics/protein patterning ML (2005) Altered TCR signaling from geo-
platform for pharmacological interrogation of metrically repatterned immunological synapses.
axon pathfinding. Lab Chip 10(8):1005–1010. Science 310(5751):1191–1193
doi:10.1039/b922143c 36. Shen K, Tsai J, Shi P, Kam LC (2009) Self-
24. Oliva AA Jr, James CD, Kingman CE, aligned supported lipid bilayers for patterning
Craighead HG, Banker GA (2003) Patterning the cell-substrate interface. J Am Chem Soc
axonal guidance molecules using a novel strat- 131(37):13204–13205
Chapter 18
Spatial Control of Biological Ligands on Surfaces

Applied to T Cell Activation
Haogang Cai, David Depoil, James Muller, Michael P. Sheetz,
Michael L. Dustin, and Shalom J. Wind
Abstract
In this chapter, we present techniques, based on molecular-scale nanofabrication and selective self-assembly,
for the presentation of biomolecules of interest (ligands, receptors, etc.) on a surface with precise spatial
control and arbitrary geometry at the single-molecule level. Metallic nanodot arrays are created on glass
coverslips and are then used as anchors for the immobilization of biological ligands via thiol linking chemistry.
The nanodot size is controlled by both lithography and metallization. The reagent concentration in self-
assembly can be adjusted to ensure single-molecule occupancy for a given dot size. The surrounding glass
is backfilled by a protein-repellent layer to prevent nonspecific adsorption. Moreover, bifunctional surfaces
are created, whereby a second ligand is presented on the background, which is frequently a requirement
for simulating complex cellular functions involving more than one key ligand. This platform serves as a
novel and powerful tool for molecular and cellular biology, e.g., to study the fundamental mechanisms of
receptor-mediated signaling.
Key words Single-molecule, Bifunctional biomimetic surfaces, Nanofabrication, Self-assembly,

Fluorescence microscopy
1 Introduction
Single-molecule techniques have matured into powerful and popular

tools to probe the complex behavior of biomolecules with unparal-
leled resolution and precision, revealing features and phenomena
that are otherwise lost by the averaging inherent in ensemble
experiments [1, 2]. Most of the studies are carried out on mole-
cules immobilized a surface, which have sufficient fluorescence sig-
nal time traces compared with freely diffusing molecules. In most
single-molecule experiments to date, the surface- immobilized
molecules are usually randomly distributed by various approaches
including physical adsorption [3–6], vesicle encapsulation [7–10]
and covalent binding [10] on either bovine serum albumin (BSA)
[11, 12], or poly(ethylene glycol) (PEG) surfaces [13, 14].
307
308 Haogang Cai et al.
Advances in nanotechnology are opening new o pportunities in

biological studies, in particular by adding spatial control to the
placement of biological ligands on surfaces [15, 16]. Various nano-
structures, such as nanogrids [17], nanoapertures [18], zero mode
waveguides (ZMWs) [19], and DNA curtains [20], have been
used to confine a small number of molecules in orderly, predeter-
mined locations.
Recently, arrays of metallic nanoparticles or nanodots have
been used as immobilization anchors [21–23], as such particles
can be formed at a size that is close to or below that of the mol-
ecule of interest, thereby achieving precise spatial control at the
single-molecule level [24–26]. Combined with single-molecule
detection techniques, this platform has the potential to push bio-
sensing nanoarrays to its ultimate performance, in terms of sensi-
tivity, specificity, throughput, etc. [19, 27–29]. More importantly,
it is a novel tool to investigate the effect of molecular geometric
organization and mechanical force on receptor-mediated cellular
functions, including adhesion, spreading, differentiation, apop-
tosis, and tissue maintenance [30–39], offering unique insight
into fundamental mechanisms of mechanobiology [40, 41], with
potential applications in nanomedicine as well [42]. In particular,
the geometric effects of T cell receptor (TCR) organization on
the immune response [26, 43–46] have translational implications
toward adoptive immunotherapies for cancer [47] and other
diseases.
In this chapter, three different fabrication approaches based on
electron beam (e-beam) lithography (EBL) or nanoimprint lithog-
raphy (NIL) are described and compared. Immobilizing monova-
lent and biotinylated anti-CD3 antibodies on the metallic nanodots
through thiol chemistry achieves precise spatial control of TCR
engagement [48]. The dot size and self-assembly reagent concen-
tration can be adjusted to ensure single-molecule occupancy, which
is crucial because T cell response is sensitive to stoichiometry [26].
A PEG layer or a supported lipid bilayer (SLB) is applied to passiv-
ate the surrounding glass surface, in order to prevent nonspecific
adsorption of the ligands. Furthermore, by introducing a second
ligand through an orthogonal linking chemistry, the surrounding
passivation layer can also be functionalized as an active background,
either static (PEG) [49–52] or mobile (SLB) [37]. Intercellular
adhesion molecule-1 (ICAM-1) is used as the second ligand, as it
plays an important role, along with the TCR ligand, in the forma-
tion of the immunological synapse (IS). In this way, a bifunctional
surface is able to better mimic the antigen presenting cell (APC)
[26, 45, 46]. We present three types of schemes for mono- or
bifunctional surfaces with PEG or SLB, all verified by total internal
reflection fluorescence (TIRF) microscopy.
Nanopatterning of T Cell Ligands 309
2 Materials
2.1 Fabrication Cleanliness is essential for molecular-scale nanofabrication. All

work should be performed in a cleanroom, class 10,000 or better,
with the following instruments: spin coater; EBL tool (electron
beam lithography system); NIL tool (nanoimprinter); molecular
vapor deposition (MVD) coater; water bath sonicator with chilled-
water cooling; e-beam physical vapor deposition (EBPVD) tool
(e-beam evaporator) with a tilt holder; reactive-ion etching (RIE)
tool; and hot plate. The resist film thickness is calibrated using
either an ellipsometer or reflectometer. The fabrication process is
monitored via inspection by both scanning electron microscopy
(SEM) and atomic force microscopy (AFM). All chemicals are at
least “Certified ACS” grade.
1. Glass coverslips, no. 1.5. For immune synapse studies we have
used 20 × 24 mm coverslips to be compatible with Ibidi μ-Slide
VI 0.4 and our specific EBL tool.
2. Silicon wafers, doped, with thermal oxide.
3. Alkaline glass cleaning solution.
4. Sulfuric acid (H2SO4).
5. Hydrogen peroxide (H2O2), 30% concentration, stored at
4 °C.
6. Deionized (DI) water (18 MΩ∙cm).
7. Ethanol, 200 proof.
8. Polymethylmethacrylate (PMMA) resist, molecular weight
(MW): 35 K, 495 K.
9. Anisole, 99%.
10. Syringe filters, Polytetrafluoroethylene (PTFE), 0.2 μm.
11. AquaSAVE, conductive polymer (Mitsubishi Rayon).
12. Hydrogen silsesquioxane (HSQ), 6%, stored at 4 °C.
13. Methyl isobutyl ketone (MIBK).
14. PMMA developer, MIBK/IPA 1:3.
15. Isopropanol (IPA), 99%.
16. HSQ developer.
17. Anti-adhesion coating agent for NIL mold release.
18. Ti, pellets.
19. AuPd, 60/40 alloy, pellets.
20. Acetone.
21. Remover PG (MicroChem).
2.2 Functionalization The samples are treated in a plasma cleaner prior to functionaliza-
tion. Through the lens TIRF microscopy is performed with a 100×
1.49 NA objective and a back illuminated EMCCD camera.
1. HS-C11-EG6-Biotin and HS-C11-EG3-OH, stored at −20 °C.
2. Ethanol, anhydrous, > 99.5% (200 proof).
3. mPEG-silane, MW 5000, stored at −20 °C.
4. Silane-PEG-NHS, MW5000, stored at −20 °C.
5. NTA-l-lysine.
6. 4-methylmorpholine (4MM).
7. Toluene, anhydrous, 99.8%.
8. Methanol, anhydrous.
9. Acetic acid, glacial.
10. Triethylamine (TEA).
11. Glass syringes, metal needles for the anhydrous solvents.
12. 6-well plates.
13. Parafilm.
14. Glass jars with PTFE caps, 30 mL.
15. Coverslip mini-rack, Teflon.
16. A chamber for replaceable coverslips.
17. A 6-channel slide attachment with a self-adhesive underside
(μ-Slide VI 0.4, Ibidi).
18. Phosphate buffer saline (PBS), DPBS 1×.
19. HEPES buffered saline (HBS) (see Recipe in [53]).
20. DOPC, 850375C (Avanti Lipids).
21. DGS-NTA(Ni), 790404C (Avanti Lipids).
22. Casein.
23. Human serum albumin (HSA), stored at 4 °C.
24. Bovine serum albumin (BSA), stored at 4 °C.
25. Streptavidin, stored at −20 °C.

26. Mono-biotinylated UCHT1 Fab′, labeled by Alexa-568,
stored at −80 °C.
27. 12× His-tagged ICAM-1, labeled by Alexa-405, stored at −80 °C.
3 Methods
3.1 Fabrication Various techniques have been employed to create metallic nanoar-
of Ordered Metallic rays; these can be divided into two major strategies. Generally
Nanodot Arrays speaking, the bottom-up strategies, such as block-copolymer
micelle nanolithography (BCML) [21, 32–36, 43–45], and
self-assembled nanosphere lithography (NSL) [54], have low cost

and high throughput but are limited to a specific homogeneous
geometric arrangement (typically hexagonal close packed arrays
extending over macroscopic dimensions). Top-down strategies,
which may include photolithography [55], dip-pen nanolithography
(DPN) [25], and EBL/NIL [26, 29–31, 46], enable precise spatial
control and arbitrary, heterogeneous geometries on the same sur-
face, which is particularly advantageous in cellular studies.
Here, we describe three approaches based on EBL or NIL:
A. EBL (PMMA bilayer), B. EBL (PMMA single layer) + hard
mask, C. nil + hard mask (along with D. EBL of NIL mold), as
illustrated in Fig. 1. Either a resist bilayer or an additional hard
mask [56] is used to create a negative, or retrograde slope in the
resist, which is necessary for a successful lift-off of metal deposited
by physical vapor deposition. The processes (A–D) are described
together, step by step (parentheses in subheadings indicating
involved processes), including inspection by SEM/AFM in Fig. 2.
3.1.1 Substrate Glass coverslips for single-molecule nanoarrays and Si wafers for
Preparation (A–D) EBL dose test and NIL mold are cut to appropriate size, so that
they will fit the lithography and/or microscopy sample holders.
Dicing (A–D)
Cleaning (A–C) New Si wafers procured from commercial sources are generally
clean enough for nanofabrication. On the other hand, glass cover-
slips need to be cleaned thoroughly. Any contamination that is not
removed in this process could result in defects in both fabrication
and functionalization. We use a two-step process with both deter-
gent and piranha solution for the cleaning (see Note 1).
1. Dilute alkaline cleaning solution (e.g., 7×, Helmanex III, etc.)
to working strength with DI water.
2. Immerse coverslips in the diluted cleaning solution and heat to
boiling temperature. Keep for 30 min on a hot plate.
3. Remove coverslips after cooling down.
4. Rinse with DI water for 10 min.
5. Prepare piranha solution (3:1 H2SO4:H2O2).
6. Immerse coverslips in the piranha solution for 5 min.
8. Rinse with ethanol.
9. Blow dry with a stream of inert gas (Ar or N2), hereafter simply
referred to as “blow dry” (see Note 2).
3.1.2 Resist Cleanliness is especially important during the resist application

Deposition (A–D) (see Note 3). All the spinning parameters are for reference only; these
should be calibrated to achieve the target thickness (see Note 4).
Fig. 1 Schematic diagram of fabrication processes

Fig. 2 90-nm-spaced hexagonal nanoarrays. Approach A (EBL of PMMA bilayer): SEM of the nanoarray (a) before
and (b) after lift-off, (c) AFM of the annealed nanoarray. Approach C (NIL + hard mask): (d) SEM of Ti hard mask.
The inset shows the AuPd deposition and undercuts in PMMA. (e) SEM of the nanoarray after lift-off. The inset
shows the original HSQ pillars on a NIL mold. (f) AFM of the annealed nanoarray
PMMA Bilayer for EBL (A) In a resist bilayer, the bottom layer is of lower MW, while the
top layer is of higher MW. Due to their different sensitivities to
exposure dose, the top, higher MW layer will develop with a nar-
rower opening than the bottom, lower MW layer under the same
exposure dose, forming an overhang. The resulting negative slope
is crucial for the subsequent lift-off.
1. Spin lower MW (e.g., 35K, mr-I PMMA35k-100 nm, Micro
Resist Technology, diluted 2:1 PMMA:anisole) PMMA at
4500 rpm and 4500 rpm/s for 45 s.
2. Bake at 180 °C for 5 min on a hot plate. Hereafter, “on a hot
plate” is omitted.
3. Spin higher MW (e.g., 495 K, A2 concentration, Microchem)
PMMA at 4500 rpm and 4500 rpm/s for 45 s (see Note 5).
4. Bake at 180 °C for 10 min. This process should result in a
PMMA bilayer with a thickness of ~ 35 nm each.
5. Spin AquaSAVE at 3000 rpm and 300 rpm/s for 45 s (see
Note 6).
PMMA Single Layer 1. Spin PMMA (495 K A2) at 4500 rpm and 4500 rpm for 45 s.
for EBL (B) 2. Bake at 180 °C for 15 min. The resist thickness should be ~
60 nm.
3. Spin AquaSAVE (see Subheading “PMMA Bilayer for EBL (A)”).
PMMA for NIL (C) 1. Spin PMMA (35K, see Subheading “PMMA Bilayer for EBL
(A)”) at 3000 rpm for 45 s.
2. Bake at 180 °C for 5 min. The resist thickness should be ~
50 nm.
HSQ on Si Chip (D) 1. Dilute the original 6% HSQ (e.g., XR-1541, Dow Corning) to
2% with MIBK (see Note 7).
2. Spin HSQ at 6000 rpm and 3000 rpm/s for 1 min. The thick-
ness should be ~ 25–30 nm.
3.1.3 Lithography (A–D) The lithography feature size is typically in the range of 15–20 nm.
An ultrathin metal disk in this size range can be transformed into a
spherical nanoparticle with sub-10 nm diameter by thermal anneal-
ing. For EBL at this scale, exposure tests are necessary to optimize
both the required dose and the system status, such as e-beam focus-
ing and alignment, subfield stitching, uniformity, etc. (see Note 8).
Cold development with ultrasonic agitation is also helpful to ensure
high contrast that yields high resolution [57, 58]. The feature size is
very sensitive to the developing conditions (time, temperature, etc.);
these should be strictly controlled to ensure reproducible results (see
Note 9). On the other hand, the resolution of NIL is only limited
by the feature size on the mold [59], so the process uniformity (from
sample to sample) is easier to control.
EBL of PMMA (A, B) 1. Expose resist-coated samples with predetermined dose in an

e-beam writer.
2. Rinse exposed samples with DI water to remove the AquaSAVE
discharge layer, and then blow dry.
3. Place a beaker of developer (MIBK/IPA 1:3) in a water bath
sonicator. Prepare a beaker of IPA. Both the developer and IPA
have been stored at 4 °C.
4. Immerse samples in the developer solution for 1 min with son-
ication at 4 °C.
5. Remove samples and immediately immerse in cold IPA for 15 s
to halt development.
6. Blow dry.
EBL of HSQ (D) 1. Expose resist-coated samples with predetermined dose in an

e-beam writer.
2. Prepare a beaker of HSQ developer (e.g., MF-CD-26,
Microchem).
3. Immerse samples in the developer solution for 4 min.
4. Remove samples and immediately immerse in DI water to halt
development.
5. Rinse with DI water, and then blow dry (see Note 10).
After lithography, the samples are thermally cured at 540 °C

for 1 h (similar to the thermal anneal in Subheading 3.1.8) to
improve the mechanical properties, and then coated by a mold
release agent (e.g., NXT-110-A, Nanonex) in an MVD coater.
Now the samples are ready for use as NIL molds in the following
Subheading “NIL of PMMA (C)” (Fig. 2e inset).
NIL of PMMA (C) 1. Treat the resist-coated samples by fluorination plasma with
C4F8 flow 100 standard cubic centimeter per minute (sccm),
RF power 100 W, at 40 mTorr for 30 s. This anti-adhesion
treatment reduces the surface energy and facilitates the mold
separation.
2. Blow the surfaces of both fluorinated sample and mold with inert
gas, in order to remove possible particle contamination. Then
load the two face to face in a nanoimprinter (see Note 11).
3. Run thermal NIL at 500 psi and 180 °C for 5 min.
4. Separate the mold from the sample. Try to minimize lateral
movement during separation, which causes shear stress on the
HSQ pillars (see Note 12).
3.1.4 Hard Mask PMMA bilayers patterned by EBL have negative slopes for lift-off,
Deposition (B, C) as discussed above. Alternatively, an additional hard mask is applied
to PMMA single layers patterned by either EBL or NIL for the
same purpose. EBPVD is used for both the hard mask and metal
deposition, because it has smaller grain size and more precise thick-
ness control compared with other methods (e.g., thermal evapora-
tion). The difference is that the hard mask is evaporated at an
angle, so that it covers only the top surface and upper edges of the
patterned openings in the resist. The angle and thickness is for
reference only; these should be adjusted so that the opening size is
reduced to 10–15 nm (see Note 13). An example is shown in Fig. 2d,
where the openings are not perfectly round due to the granular
structure of the Ti hard mask.
1. Load lithography patterned samples on a tilt holder (see
Note 14).
2. Adjust the tilt holder so that there is a 30° angle between the
sample surface and metal vapor flux.
3. Deposit 12 nm Ti in the e-beam evaporator. The actual thick-
ness of the hard mask should be approximately 12 × sin
30° = 6 nm.
3.1.5 Descum (B, C) The descum process not only removes the residual resist, but also
forms an undercut below the hard mask, which further facilitates
the lift-off. The insets of Figs. 1 and 2d show the SEM cross-section
and top view of the resist undercuts respectively. The descum rate
depends on the opening size, which is much slower but more
difficult to calibrate than for a planar resist. The descum time

should be more than enough to etch all the residue in the Z direc-
tion, but the undercuts in the XY plane need to be well controlled,
especially for dense arrays. Therefore, good directionality of the
etching is important.
1. Clean the etcher chamber with O2 plasma, in order to remove
any residual gas that might be reactive to the metal hard mask.
2. Etch the metal-coated samples with O2 flow 80 sccm, RF
power 60 W, at 40 mTorr. Under this condition, the etch rate
for planar PMMA is ~3 nm/s. The etch time is 2 min for NIL
samples (to remove ~ 25 nm thick residue), and 1 min for EBL
(just to create undercuts).
3.1.6 Metal A thin film of AuPd (see Note 15) is deposited by normal EBPVD,
deposition (A–C) with the thickness adjusted in coordination with the lithographi-
cally determined opening size, in order to control the nanodot
size. The film covers the top surface of the resist (Fig. 2a), while
thin nanodisks form on the substrate through the openings (insets
of Figs. 1 and 2d).
1. Deposit 0.5 nm Ti as adhesion layer. This is necessary for the
adhesion of AuPd on glass, especially during the thermal anneal
(see Note 16).
2. Deposit 2 nm AuPd on Top of the Ti.
3.1.7 Lift-Off (A–C) The remaining resist, together with the metal film on the top of
the resist (which is not connected to the nanodots thanks to the
negative slope), is removed in a lift-off solution. Acetone is usually
sufficient for hard mask samples with larger undercuts (B, C). On
the other hand, a stronger solution, Remover PG, is used for
PMMA bilayer samples (A). The former has a smaller dot size than
the latter, because the hard mask further reduces the opening size
at the same resolution of lithography (Fig. 2b, e), which can be
clearly seen by comparing with the NIL mold (inset of Fig. 2e).
1. Immerse the metal-coated samples in a beaker of lift-off solu-
tion and heat to 80 °C for 30 min on a hot plate.
2. After cooling down, seal the beaker with Parafilm to avoid
evaporation, and let sit overnight.
3. Remove samples from the solution full of stripped metal film
pieces. Immediately immerse in a fresh lift-off solution at
80 °C for 30 min, to reduce the possibility of re-deposition of
removed resist.
4. Inspect if there are residual metal films, especially in the pat-
tern regions (see Note 17).
5. Remove samples and immediately immerse in IPA.
6. Rinse with IPA, and then blow dry.
3.1.8 Thermal Samples should be cleaned thoroughly in the lift-off process,

Anneal (A–C) because any resist residue will be difficult to remove after the ther-
mal anneal. A temperature of 540 °C is enough for the granular
nanodots to agglomerate into uniform spheres due to the melting-
point depression at the nanoscale. Little is known about oxidation
of the metal at this scale, but annealing in ambient air does not
affect the subsequent thiolation. A higher temperature and longer
anneal time can be used to further reduce the dot size by evapora-
tion, although higher temperatures may cause the nanodots to
migrate from their original positions. Given the same metal thick-
ness, the annealed nanodots are smaller on hard mask samples
(B, C: ~ 5 nm) than on PMMA bilayer samples (A: ~ 7.5 nm), as
shown in Fig. 2c, f (see Note 18).
1. Place clean samples on a hot plate.
2. Slowly increase the temperature to 540 °C, and then anneal
for 2 h.
3. Remove samples after cooling down.
Direct-write EBL is versatile in that it does not require a new
mold for the creation of new patterns, because the pattern is created
in software. On the other hand, NIL has a higher throughput and
lower cost in creating replicas of a given mold, especially good for
dense patterns whose writing is time-consuming in EBL. The lift-off
based on the PMMA bilayer is a direct and simple approach, but
challenging when the pattern is very dense. On the other hand, the
hard mask approach not only improves the lift-off yield, but also
further reduces the feature size. The three approaches are compared
in Table 1. The choice of process should be dictated by the specific
applications, with consideration of instrument availability.
3.2 Functionalization Artificial APC surfaces are used as an example to demonstrate the
spatial control of biological ligands. A biotinylated UCHT1 Fab′
(a single binding ligand of CD3ε, a component of the TCR) is
immobilized on the nanodots through thiol and biotin-streptavidin
Table 1
Fabrication approaches comparison
Approaches Lithography Merits Applications Lift-off Merits Applications

A EBL No need of mold New pattern Bilayer No need of Sparse pattern, ≥
or mask test, sparse angle 7.5 nm dots
pattern EBPVD
B
C NIL Throughput, cost, Replication of a Hard High yield Dense pattern,
process given pattern, mask ≥ 2nm dots
uniformity dense pattern
Fig. 3 (a) Schematic diagram of Scheme A with PEG. (b) The AFM of 60-nm-spaced hexagonal arrays and SEM
of the hard mask (better contrast than the nanoarray itself). (c) The fluorescence image of UCHT1 Fab′. (d)
Schematic diagram of Scheme B with PEG-NTA(Ni). Fluorescence image of (e) UCHT1 Fab′ and (f) ICAM-1
chemistry [26, 46]. This is prepared from UCHT1 Fab′2 by reduc-

tion and reaction with a biotin-maleimide that reacts with free sulf-
hydryl groups in the C-terminal linker of the cleaved heavy chain.
Either PEG or a SLB is used for the background passivation, where
a second ligand, ICAM-1 (which binds with leukocyte function
associated antigen-1 (LFA-1)), can be introduced by an orthogo-
nal chemistry based on interaction between nitrilotriacetic acid
with Ni(II) (NTA(Ni)) and a poly-His-tag at the C-terminus of
the recombinant ICAM-1 extracellular domain. Three types of
functionalization schemes are presented here: A. PEG, B. PEG-
NTA(Ni), and C. SLB (with lipids containing NTA(Ni)).
3.2.1 Monofunctional As illustrated in Fig. 3a, the dot size and reagent concentration are
with Static Background: adjusted so that there is a single UCHT1 Fab′ molecule bound to
PEG each nanodot, on average [26]. The surrounding background is
passivated by a self-assembled monolayer (SAM) of PEG-silane on
glass to prevent nonspecific adsorption (see Note 19). This mono-
functional scheme is demonstrated on nanoarrays with 7.5 nm dot
size (60-nm-spaced hexagonal arrays as an example in Fig. 3b).
After functionalization, the TIRF image shows a 200 × 200 μm2
square of nanoarray (Fig. 3c) (see Note 20).
Thiolation All three schemes share the same protocol of thiolation, which is
described here and omitted for the rest.
1. Prepare piranha solution (see Subheading “Cleaning (A–C)”).
2. Prepare a 1 mM mixture (1:1) of HS-C11-EG6-Biotin and
HS-C11-EG3-OH in 1.5 mL anhydrous ethanol (see Note 21).
3. Immerse samples in 1.5 h-aged piranha solution for 3 min
(see Note 22).
5. Rinse with ethanol, and then blow dry.
6. Place the dried samples in a plasma cleaner at 18 W for 5 min.
7. Remove samples and immediately immerse in the alkylthiol
mixture solution.
8. Seal the container with Parafilm, cover with Al foil, and incubate
on a shaker, hereafter simply referred to as “incubate on a
shaker,” for 18 h (overnight).
PEGylation 1. Prepare a solution of 2 mg PEG-silane in 25 mL of anhydrous

toluene with 30 μL acetic acid as a catalyst (see Note 23).
2. Remove samples from the thiolation solution, and immerse in
fresh ethanol.
3. Rinse with ethanol and then blow dry.
4. Place the dried samples in the PEG solution and incubate on a
shaker for 48 h.
Biological Ligands 1. Prepare a solution of 10 μg/mL streptavidin in 1.5 mL PBS

with 1 mg/mL BSA.
2. Remove samples from the PEG solution, and rinse with acetone
and then ethanol, blow dry.
3. Immerse samples in the streptavidin solution, and incubate on
a shaker for 1 h.
4. Remove samples from the streptavidin solution, and rinse
thoroughly with PBS.
5. Incubate samples in a well of fresh PBS on a shaker for
30 min.
6. Prepare a solution of 2 μg/mL biotinylated UCHT1 Fab′ in
1.5 mL PBS with 1 mg/mL BSA.
7. Remove samples from the PBS solution, rinse with PBS.
8. Incubate samples in the biotinylated UCHT1 Fab′ solution on
a shaker for 1 h.
9. Remove samples from the Fab′ solution, and rinse thoroughly
with PBS.
10. Incubate samples in a well of fresh PBS on a shaker for 30 min.
3.2.2 Bifunctional T cells plated on surfaces functionalized by Scheme A have poor

with Static Background: adhesion and therefore provide insufficient data [45]. This is
PEG-NTA(Ni) because PEG with a chain-length (MW 5000 in this case) long
enough to prevent nonspecific adsorption of proteins is also cell
repellent [60, 61]. One way to solve this dilemma is to introduce
ICAM-1 on the PEG, so that it repels proteins while supporting
cell adhesion (Scheme B, Fig. 2d). The TIRF image of UCHT1
Fab′ nanoarrays (Fig. 2e) is similar to Scheme A (Fig. 2b) (see Note
24). The ICAM-1 has the same pattern (Fig. 2f) because it was
imaged after permeabilization, which removed the PEG in unpat-
terned regions. This contrast clearly shows that ICAM-1 molecules
are backfilled among the nanodots. The ICAM-1 density is esti-
mated below 50/μm2, by comparing its intensity with a SLB con-
taining ICAM-1 of known density, assuming a linear relationship
between the fluorescence intensity and molecule density.
In scheme B, the thiolation follows the same protocol as
Scheme A (see Subheading “Thiolation”). Meanwhile, the silane-
PEG-NTA solution is prepared as follows.
Synthesis of PEG-NTA 1. Prepare a solution of 5 mg silane-PEG-NHS in 0.125 mL

anhydrous methanol.
2. Prepare a solution of 1.57 mg NTA-l-lysine in 0.5 mL anhy-
drous methanol, with 15 μL 4MM (see Note 25).
3. Add 0.125 mL NTA-l-lysine solution to the silane-PEG-NHS
solution, incubate for 18 h. Silane-PEG-NTA is synthesized in
the mixed solution.
PEGylation 1. Add the silane-PEG-NTA solution in 25 mL of anhydrous

toluene with 250 μL TEA as a catalyst (see Note 26).
2. Remove samples from the thiolation solution, and immerse in
fresh ethanol.
3. Rinse samples with ethanol and then blow dry.
4. Place the dried samples in the silane-PEG-NTA solution, and
incubate on a shaker for 48 h.
Biological Ligands 1. Prepare a solution of 1 mM NaOH in DI water. Prepare a

solution of 40 mM NiSO4 aqueous solutions.
2. Remove samples from the PEG solution. Rinse with acetone
and then ethanol. Blow dry.
3. Incubate samples in the NaOH solution for 5 min.
4. Remove samples from the NaOH solution, and rinse with DI
water.
5. Incubate samples in the NiSO4 solution for 1 h to adsorb
Ni(II) to the surface NTA groups.
6. Prepare a solution of 1 μg/mL His-tagged ICAM-1 in 1.5 mL

PBS.
7. Remove samples from the NiSO4 solution. Rinse with DI water
and then PBS.
8. Immerse samples in the ICAM-1 solution. Incubate on a
shaker for 2 h.
9. Prepare a solution of 10 μg/mL streptavidin in 1.5 mL PBS
with 1 mg/mL BSA.
10. Remove samples from the ICAM-1 solution, and rinse with
PBS.
11. Immerse samples in the streptavidin solution. Incubate on a
shaker for 30 min.
12. Remove samples from the streptavidin solution, and rinse
thoroughly with PBS.
14. Prepare a solution of 2 μg/mL biotinylated UCHT1 Fab′ in
1.5 mL PBS with 1 mg/mL BSA.
15. Remove samples from the PBS solution, rinse with PBS.
16. Incubate samples in the biotinylated UCHT1 Fab′ solution on
a shaker for 1 h.
17. Remove samples from the Fab′ solution, and rinse thoroughly
with PBS.
3.2.3 Bifunctional T cell adhesion is improved by Scheme B, compared with Scheme

with Mobile Background: A. However, the ICAM-1 density is still insufficient (<50/μm2) to
SLB provide good cellular adhesion, so the TCR has to serve a dual role
of antigen recognition and adhesion receptor. This is why the T cell
activation requirements probed by nanoarrays [26, 43–45] are gen-
erally much higher than by APCs [62] or artificial APC surfaces
[17]. In order to exclude the requirement by TCR-mediated adhe-
sion, the ICAM-1 density should be increased to approximately
200 /μm2. This can be done by improving the yield of silane-PEG-
NTA synthesis. Another approach is to replace the PEG by a SLB,
where the ICAM-1 density can be simply adjusted by modulating
the concentration of lipids containing NTA(Ni) (Fig. 4a). Cluster
arrays with the same inter-dot spacing of 60 nm are shown in Fig.
4b (see Note 27), which were backfilled by a SLB (see Note 28).
Field-stop aperture fluorescence recovery after photobleaching
(FRAP) of ICAM-1 in the SLB with embedded nanoarrays is shown
in Fig. 3c–e, which confirms the bilayer continuity and mobility.
We will briefly describe one method for making antigen pre-
senting surfaces that allow the ICAM-1 to be mobile and the
UCHT1 Fab′ to be immobilized on the nanoarray. More detailed
Fig. 4 (a) Schematic diagram of Scheme C with SLB. (b) The fluorescence image of UCHT1 Fab′ and ICAM-1.
The insets show the AFM and SEM of a cluster with 60 nm inter-dot spacing. (c) A bleached central spot in the
SLB. (d) Fluorescence recovery after 1 min. (e) Intensity profiles of the FRAP
descriptions of the planar bilayer system and other interactions of

this process can be found in Ref. [53, 63]. The thiolation follows
the same protocol as Scheme A (see Subheading “Thiolation”).
Then the SLB is formed as follows (see Note 29).
SLB 1. Remove samples from the thiolation solution, and immerse in

fresh ethanol.
2. Prepare the flow cell buffer of HBS with 1% HSA
(HBS-HSA).
3. Thaw a vial of 5% Casein at 37 °C.
4. Prepare a solution of small unilamellar vesicles (SUVs) con-
taining 12.5% DGS-NTA(Ni) and 87.5% DOPC.
5. Rinse samples with ethanol and then blow dry.
6. Immediately glue the dried samples on the bottom of an Ibidi
sticky-slide (VI 0.4). Make sure the pattern region is under-
neath the channel and facing up. Seal the glued samples
properly.
7. Immediately add 30 μL of the SUV solution, and incubate for
5 min. The solution volume is for each channel of the Ibidi
sticky-slide hereafter.
8. Wash out SUVs with 500 μL HBS-HSA.
9. Block with 250 μL casein solution for 1 h.
Biological Ligands 1. Wash out excess casein with 1 mL HBS-HSA.

2. Prepare a solution of 5 μg/mL streptavidin in HBS-HSA.
3. Add 250 μL streptavidin solution, and incubate for 30 min.
Table 2
Functionalization schemes comparison
Schemes Mobility Functionality ICAM-1 density (/μm2) Preparation time (h)

A Static Passive 0 72
B Static Active < 50 72
C Mobile Active 200 24
4. Wash with 1 mL HBS-HSA.

5. Prepare a solution of 4 μg/mL biotinylated UCHT1 Fab′ and
0.625 μg/mL His-tagged ICAM-1 in HBS-HSA.
6. Add 250 μL solution of biotinylated UCHT1 Fab′ and ICAM-
1, and incubate for 30 min.
7. Wash with 667 μL HBS-HSA slowly for three times.
Overall, the Scheme C of SLB with sufficient and mobile ICAM-1
is best among the three at mimicking the surface of the APC, as
shown by the comparison in Table 2. It also reduces the preparation
time significantly, as the PEGylation is time-consuming. Scheme B
can be used when a static bifunctional surface is preferred. Scheme A
is suitable for the studies where only one type of ligand is involved,
and the nonspecific adsorption has to be minimized.
4 Notes
1. The second step involving a piranha solution is helpful but

optional, because cleaning with only detergent solution is usu-
ally sufficient. However, if the cleaned coverslips are not used
immediately, a quick cleaning in piranha will be necessary
because the coverslips accumulate contamination during stor-
age in air.
2. Use caution when handling (especially blowing) coverslips.
Grab a corner and blow toward the tweezers, so that the cov-
erslip does not move or fall. Also, avoid high stress (when
mounting with clips and screws) or acute temperature change
(during thermal anneal).
3. Sometimes, there is contamination or aggregates in the resist
solution. These can be removed by a filter with 0.2 μm pore
size before deposition. Any sample with visible coating defects
should be discarded, or reused only after resist removal and
thorough cleaning. It is critical to remove any resist from the
backside of the sample (e.g., by dabbing with a solvent-soaked
cleanroom cloth) before baking. Because it might cause a tilted
surface or contaminate the holder in the e-beam writer.
4. The resist solution may change with time, due to solvent

evaporation, so a higher spin speed may be needed to achieve
the same film thickness.
5. Start spinning immediately after dropping resist on the sample.
This is especially important for the top layer, whose solvent
permeates the bottom layer, resulting in an uneven coating.
Also, note that the spin curve is different on a resist layer (see
Subheading “PMMA Bilayer for EBL (A)”) from that on a
glass surface (see Subheading “PMMA Single Layer for EBL
(B)”).
6. Dab droplets of AquaSAVE at the corners, so that the samples
are completely dry before loading in an e-beam writer.
7. Water vapor could cause HSQ to cross-link in solution. Stored
at 4 °C in a refrigerator, the containers should only be opened
when warmed up to room temperature; otherwise water vapor
in the air will condense on the solution and cause contamina-
tion. Aliquoting also helps to minimize the exposure to air.
8. The exposure tests can be done on Si chips instead of glass
coverslips. Because the Si wafer we use is doped and conduc-
tive, it enables better imaging in SEM, without charging issues.
Usually, the SEM inspection is performed after Ti hard mask
angle deposition or AuPd normal deposition, which not only
provides information on the correct exposure dose but also
serves as a process test to optimize metallization and lift-off.
9. Resist developing conditions should be strictly consistent, with
attention to the following details: First, when developing mul-
tiple samples at the same time, not more than two samples are
recommended to minimize the time difference in IPA. The
development is halted in IPA, but not stopped completely,
which could cause a difference in feature size. Second, the
water bath temperature could rise after sonication. Wait until it
cools down back to 4 °C for the next development. Finally, the
developer and IPA can be reused, but should be changed peri-
odically (developing rate could be slowed down after reusing
too many times) and always stored at 4 °C.
10. For an aspect ratio of HSQ pillars close to 1 (e.g., diameter
15–20 nm, height 25–30 nm), pillar collapse after drying in
water is usually not a problem. But in case it does occur,
releasing in boiling ethanol instead of water can reduce the
surface tension and solve this problem. Never let the sample
dry during transfer between solutions. This is important
whenever a liquid-air interface is involved, such as development
and lift-off.
11. Sample cleanliness before lithography is especially important
for NIL. A dust particle trapped between the mold and
imprinted sample will result in a defective area much larger
than the original particle size. The defective area depends on

the particle height and the mechanical flexibility of both mold
and sample.
12. The detachment of HSQ pillars during mold separation is the
major factor limiting the mold durability. Efforts have been
made to address this issue. First, the mold is thermally cured
(see Subheading “EBL of HSQ (D)”) to improve the mechani-
cal properties of HSQ pillars. Under e-beam exposure, the
HSQ molecules are cross-linked from cage form to network
form, which become resistant in the developer (i.e., negative
tone). Additional high-temperature thermal curing in O2 dis-
associates the Si-H bonds completely, resulting in a chemical
composition similar to SiO2 [64]. Second, anti-adhesion treat-
ments are applied to both the mold (MVD coating) and the
resist-coated glass coverslips (fluorination plasma). Third, lat-
eral movement should be avoided to minimize the shear stress
applied on HSQ pillars during mold separation. Finally, it is
found that dense patterns have better durability than sparse
ones probably due to a lower shear stress distributed on indi-
vidual pillars. HSQ pillars with a larger diameter also have bet-
ter durability than smaller ones (see Note 24).
13. The hard mask by angle EBPVD covers only the top surface of
resist, which protects the unexposed part during descum and
therefore widens the process window. Moreover, it forms an
overhang around the edge of an opening, which not only
serves as a negative slope for lift-off, but also reduces the open-
ing size. The deposition thickness and angle can be adjusted to
control the size reduction, which may be compromised with
the uniformity of feature shape and size. Because when the
hard mask is thicker, the metal grain size is larger, resulting in
more irregular and granular nanodots. It is not recommended
to reduce the opening size below 10 nm. Because the unifor-
mity of openings is poor (due to the high granularity of the
hard mask), some openings may even be closed, resulting in
failure of pattern transfer.
14. Rotation is not necessary for the tilt holder, but will result in
openings with greater roundness and uniformity. It is impor-
tant to ensure the samples are fixed securely during rotation.
15. AuPd (60/40) is used instead of pure Au for the following
reasons. First, stable thiol SAMs form on not only Au, but also
Pd [65] and AuPd [23]. Second, AuPd, an alloy, has a smaller
grain size than Au, which promotes the agglomeration into a
single particle at each site during thermal anneal. Finally, AuPd
nanodots contribute less to fluorescence quenching than Au,
which is advantageous in fluorescence microscopy [26].
16. Ti has a higher melting point than AuPd. During thermal
anneal, the top layer of AuPd agglomerates to a sphere, which
is still confined on the underlying layer of Ti. Without the

Ti adhesion layer, the AuPd dots migrate after anneal.
17. If there are still residual metal films, the lift-off process can be
repeated for a longer time.
18. The height profile measured by AFM reflects the lateral size
due to the spherical shape of annealed nanodots [56].
19. The PEG passivation can be improved by the following
approaches. First, strict anhydrous conditions must be main-
tained, to minimize the silanol cross-linking after hydrolysis in
solution. Second, BSA can be added to block defects where
target biomolecules can be adsorbed. Recently, we found that
a 15 nm HSQ film, spin-coated and thermally cured prior to
nanoarray fabrication, improved the PEG passivation. It covers
all impurities on common glass coverslip surface, and therefore
prevents the formation of PEG defects [26]. Besides, the func-
tionalization sequence has been optimized and should not be
reversed. PEGylation before thiolation results in more nonspe-
cific adsorption, probably because the repulsion to alkylthiol is
more difficult than streptavidin due to the molecular shape/size.
Streptavidin is used instead of avidin for the same reason [66].
20. Under a magnification of 100×, the field size is 80.6 μm
(512 pixels), so a 3 × 3 field stitching is needed to image a
whole pattern of 200 × 200 μm2. Due to the nonuniform
illumination in microscopy, the signal is lower at the edges
especially corners of each field, which should be excluded for
measurements.
21. It was found that the molecular occupancy can be interpreted
as a packing problem for nanodots of a given size. It is also
linear with the mole fraction of biotin-alkylthiol in the mixed
SAMs, which is easily adjusted [26]. For example, a mole frac-
tion of 100%, 50%, 25% results in single-molecule occupancy
for a dot size of 4 nm [25], 7.5 nm (this work), 12.5 nm,
respectively.
22. Fresh piranha solution etches away the AuPd nanodots, while
aged solution is much less aggressive [23]. Still, the immersion
time should be strictly controlled to ensure that the nanodots
are intact.
23. To maintain the anhydrous condition, the PEG-silane solution
should be prepared in a cold environment (such as a cold room
at 4 °C). The glassware, syringes, and needles are dried in an
oven at 70 °C to remove any adsorbed moisture.
24. The only difference between the two images is the rounded
corners of the square in Fig. 2c. This is because the nanoarray
was fabricated by NIL and the mold was used for many
imprints. Due to the proximity effect of EBL, the HSQ pillars
in the corners have a smaller diameter than in the center, so

they are easier to break or detach. On the other hand, the
majority of the nanoarray remains intact. This is also why the
HSQ pillars are typically 15–20 nm in diameter, although a
smaller feature size can be reached by EBL.
25. 4MM is an organic base used to increase the solution pH, to
improve the solubility of NTA-l-lysine. Meanwhile, the pH
should be in the range of 7.2–8.5 to prevent the hydrolysis of
the NHS ester. The final concentration of 4 MM in the mixed
methanol solution is 1.5% (v/v), whose pH is ≤ 8.5. This is
also why NTA-l-lysine solution is added to silane-PEG-NHS
solution, rather than the opposite.
26. Besides acetic acid, TEA can also be used as a catalyst for
PEGylation, although at a different pH value [25, 67]. The
silane-PEG-NTA solution with 4MM has a pH > 7, so TEA is
used here instead of acetic acid.
27. The cluster arrays (Fig. 4b) have the same inter-dot spacing of
60 nm, but a much lower global density (50 /μm2) than hex-
agonal arrays (321/μm2, Fig. 3b). This demonstrates a better
capability of spatial control than other techniques such as
BCML, which cannot decouple the geometric effects of local
inter-dot spacing and global density. This is particularly impor-
tant because the clustering of transmembrane receptors is
observed in many cellular functions, such as adhesion and
immune response.
28. The dot size in this work is ~ 7.5 nm, in the range between 1.2 and
22 nm, which means that the nanodots are closely surrounded
by the SLB instead of being covered by it [68, 69].
29. The lipids and liposomes can be damaged by O2, so it is crucial
to minimize the exposure in air. Always blow a stream of inert
gas on the top of the lipids before storage. Make sure no
bubbles get in the Ibidi channel when adding solution or
washing. Keep the slide leveled and covered during handling
and incubation.
Acknowledgments
The authors thank Dr. M. Palma for guidance on nanoarray func-

tionalization and Dr. Silvia Curado for coordinating our collabora-
tion. This work was supported primarily by the National Science
Foundation (NSF) under award no. CMMI-1300590 and by the
National Institutes of Health (NIH) Common Fund Nanomedicine
program grant PN2 EY016586. The authors are grateful to the
Columbia Nano Initiative for providing cleanroom and other facil-
ities used in this work.
References
1. Deniz AA, Mukhopadhyay S, Lemke EA 12. Hua BY, Han KY, Zhou RB, Kim HJ, Shi XH,
(2008) Single-molecule biophysics: at the Abeysirigunawardena SC, Jain A, Singh D,
interface of biology, physics and chemistry. J R Aggarwal V, Woodson SA, Ha T (2014) An
Soc Interface 5(18):15–45. doi:10.1098/ improved surface passivation method for
rsif.2007.1021 single-
molecule studies. Nat Methods
2. Tinoco I, Gonzalez RL (2011) Biological 11(12):1233. doi:10.1038/Nmeth.3143
mechanisms, one molecule at a time. Genes 13. Kuzmenkina EV, Heyes CD, Nienhaus GU
Dev 25(12):1205–1231. doi:10.1101/ (2005) Single-molecule Forster resonance
Gad.2050011 energy transfer study of protein dynamics
3. Peterson EM, Harris JM (2010) Quantitative under denaturing conditions. Proc Natl Acad
detection of single molecules in fluorescence Sci U S A 102(43):15471–15476.
microscopy images. Anal Chem 82(1):189– doi:10.1073/pnas.0507728102
196. doi:10.1021/ac901710t 14. Morimatsu M, Mekhdjian AH, Adhikari AS,
4. Hanley DC, Harris JM (2001) Quantitative Dunn AR (2013) Molecular tension sensors
dosing of surfaces with fluorescent molecules: report forces generated by single integrin mol-
Characterization of fractional monolayer cov- ecules in living cells. Nano Lett 13(9):3985–
erages by counting single molecules. Anal 3989. doi:10.1021/nl4005145
Chem 73(21):5030–5037. doi:10.1021/ 15. Whitesides GM (2003) The ‘right’ size in
ac010572h nanobiotechnology. Nat Biotechnol
5. Lacy WB, Olson LG, Harris JM (1999) 21(10):1161–1165. doi:10.1038/nbt872
Quantitative SERS measurements on dielectric- 16. Torres AJ, Wu M, Holowka D, Baird B (2008)
overcoated silver-island films by solution depo- Nanobiotechnology and cell biology: micro-
sition control of surface concentrations. Anal and nanofabricated surfaces to investigate
Chem 71(13):2564–2570. doi:10.1021/ receptor-mediated signaling. Annu Rev
ac981024f Biophys 37:265–288. doi:10.1146/annurev.
6. Ha T, Enderle T, Ogletree DF, Chemla DS, biophys.36.040306.132651
Selvin PR, Weiss S (1996) Probing the interac- 17. Manz BN, Jackson BL, Petit RS, Dustin ML,
tion between two single molecules: Groves J (2011) T-cell triggering thresholds
Fluorescence resonance energy transfer are modulated by the number of antigen within
between a single donor and a single acceptor. individual T-cell receptor clusters. Proc Natl
Proc Natl Acad Sci U S A 93(13):6264–6268. Acad Sci U S A 108(22):9088–9094.
doi:10.1073/pnas.93.13.6264 doi:10.1073/pnas.1018771108/-/
7. Boukobza E, Sonnenfeld A, Haran G (2001) DCSupplemental
Immobilization in surface-tethered lipid vesi- 18. Kinz-Thompson CD, Palma M, Pulukkunat
cles as a new tool for single biomolecule spec- DK, Chenet D, Hone J, Wind SJ, Gonzalez RL
troscopy. J Phys Chem B 105(48):12165–12170. (2013) Robustly passivated, gold nanoaperture
doi:10.1021/jp012016x arrays for single-molecule fluorescence micros-
8. Rhoades E, Gussakovsky E, Haran G (2003) copy. ACS Nano 7(9):8158–8166.
Watching proteins fold one molecule at a time. doi:10.1021/Nn403447s
Proc Natl Acad Sci U S A 100(6):3197–3202. 19. Eid J, Fehr A, Gray J, Luong K, Lyle J, Otto G,
doi:10.1073/pnas.2628068100 Peluso P, Rank D, Baybayan P, Bettman B, Bibillo
9. Heider EC, Peterson EM, Barhoum M, A, Bjornson K, Chaudhuri B, Christians F, Cicero
Gericke KH, Harris JM (2011) Quantitative R, Clark S, Dalal R, Dewinter A, Dixon J, Foquet
Fluorescence Microscopy To Determine M, Gaertner A, Hardenbol P, Heiner C, Hester
Molecular Occupancy of Phospholipid Vesicles. K, Holden D, Kearns G, Kong XX, Kuse R,
Anal Chem 83(13):5128–5136. doi:10.1021/ Lacroix Y, Lin S, Lundquist P, Ma CC, Marks P,
ac200129n Maxham M, Murphy D, Park I, Pham T, Phillips
10. Roy R, Hohng S, Ha T (2008) A practical M, Roy J, Sebra R, Shen G, Sorenson J, Tomaney
guide to single-molecule FRET. Nat Methods A, Travers K, Trulson M, Vieceli J, Wegener J,
5(6):507–516. doi:10.1038/Nmeth.1208 Wu D, Yang A, Zaccarin D, Zhao P, Zhong F,
Korlach J, Turner S (2009) Real-time DNA
11. Ha T, Zhuang XW, Kim HD, Orr JW,
sequencing from single polymerase molecules.
Williamson JR, Chu S (1999) Ligand- Science 323(5910):133–138. doi:10.1126/
induced conformational changes observed in science.1162986
single RNA molecules. Proc Natl Acad Sci
U S A 96(16):9077–9082. doi:10.1073/ 20. Fazio T, Visnapuu ML, Wind S, Greene EC
pnas.96.16.9077 (2008) DNA curtains and nanoscale curtain
rods: high-throughput tools for single mole- Nanolithographic control of the spatial organi-
cule imaging. Langmuir 24(18):10524–10531. zation of cellular adhesion receptors at the
doi:10.1021/la801762h single-molecule level. Nano Lett 11(3):1306–
21. Aydin D, Schwieder M, Louban I, Knoppe S, 1312. doi:10.1021/nl104378f
Ulmer J, Haas TL, Walczak H, Spatz JP (2009) 32. Arnold M, Cavalcanti-Adam EA, Glass R,
Micro-nanostructured protein arrays: a tool for Blummel J, Eck W, Kantlehner M, Kessler H,
geometrically controlled ligand presentation. Spatz JP (2004) Activation of integrin function
Small 5(9):1014–1018. doi:10.1002/ by nanopatterned adhesive interfaces.
smll.200801219 Chemphyschem 5(3):383–388. doi:10.1002/
22. Groll J, Albrecht K, Gasteier P, Riethmueller S, cphc.200301014
Ziener U, Moeller M (2005) Nanostructured 33. Cavalcanti-Adam EA, Volberg T, Micoulet A,
ordering of fluorescent markers and single pro- Kessler H, Geiger B, Spatz JP (2007) Cell
teins on substrates. Chembiochem 6(10):1782– spreading and focal adhesion dynamics are
1787. doi:10.1002/cbic.200500041 regulated by spacing of integrin ligands.
23. Cherniavskaya O, Chen CJ, Heller E, Sun E, Biophys J 92(8):2964–2974. doi:10.1529/
Provezano J, Kam L, Hone J, Sheetz MP, Wind biophysj.106.089730
SJ (2005) Fabrication and surface chemistry of 34. Huang JH, Grater SV, Corbellinl F, Rinck S,
nanoscale bioarrays designed for the study of Bock E, Kemkemer R, Kessler H, Ding JD,
cytoskeletal protein binding interactions and Spatz JP (2009) Impact of order and disorder in
their effect on cell motility. J Vac Sci Technol B RGD nanopatterns on cell adhesion. Nano Lett
23(6):2972. doi:10.1116/1.2132332 9(3):1111–1116. doi:10.1021/Nl803548b
24. Wolfram T, Belz F, Schoen T, Spatz JP (2007) 35. Jaehrling S, Thelen K, Wolfram T, Pollerberg
Site-specific presentation of single recombinant GE (2009) Nanopatterns biofunctionalized
proteins in defined nanoarrays. Biointerphases with cell adhesion molecule DM-GRASP
2(1):44–48. doi:10.1116/1.2713991 offered as cell substrate: spacing determines
25. Chai JA, Wong LS, Giam L, Mirkin CA (2011) attachment and differentiation of neurons.
Single-molecule protein arrays enabled by Nano Lett 9(12):4115–4121. doi:10.1021/
scanning probe block copolymer lithography. nl9023325
Proc Natl Acad Sci U S A 108(49):19521– 36. Ranzinger J, Krippner-Heidenreich A, Haraszti
19525. doi:10.1073/pnas.1116099108 T, Bock E, Tepperink J, Spatz JP, Scheurich P
26. Cai H, Wolfenson H, Depoil D, Dustin ML, (2009) Nanoscale arrangement of apoptotic
Sheetz MP, Wind SJ (2016) Molecular occu- ligands reveals a demand for a minimal lateral
pancy of nanodot arrays. ACS Nano 10(4):4173– distance for efficient death receptor activation.
4183. doi:10.1021/acsnano.5b07425 Nano Lett 9(12):4240–4245. doi:10.1021/
27. Rosi NL, Mirkin CA (2005) Nanostructures in nl902429b
biodiagnostics. Chem Rev 105(4):1547–1562. 37. Lohmuller T, Triffo S, O’Donoghue GP, Xu
doi:10.1021/Cr030067f Q, Coyle MP, Groves JT (2011) Supported
28. Wingren C, Borrebaeck CAK (2007) Progress membranes embedded with fixed arrays of gold
in miniaturization of protein arrays - a step nanoparticles. Nano Lett 11(11):4912–4918.
closer to high-density nanoarrays. Drug Discov doi:10.1021/nl202847t
Today 12(19–20):813–819. doi:10.1016/j. 38. Lohmuller T, Xu Q, Groves JT (2013)
drudis.2007.08.003 Nanoscale obstacle arrays frustrate transport of
29. Palma M, Abramson JJ, Gorodetsky AA, Penzo EphA2-ephrin-A1 clusters in cancer cell lines.
E, Gonzalez RL, Sheetz MP, Nuckolls C, Nano Lett 13(7):3059–3064. doi:10.1021/
Hone J, Wind SJ (2011) Selective biomolecu- nl400874v
lar nanoarrays for parallel single-molecule 39. Liu Y, Medda R, Liu Z, Galior K, Yehl K, Spatz
investigations. J Am Chem Soc 133(20):7656– JP, Cavalcanti-Adam EA, Salaita K (2014)
7659. doi:10.1021/Ja201031g Nanoparticle tension probes patterned at the
30. Schvartzman M, Nguyen K, Palma M, nanoscale: impact of integrin clustering on
Abramson J, Sable J, Hone J, Sheetz MP, Wind force transmission. Nano Lett 14(10):5539–
SJ (2009) Fabrication of nanoscale bioarrays 5546. doi:10.1021/nl501912g
for the study of cytoskeletal protein binding 40. Vogel V, Sheetz M (2006) Local force and
interactions using nanoimprint lithography. geometry sensing regulate cell functions. Nat
J Vac Sci Technol B 27(1):61–65. Rev Mol Cell Biol 7(4):265–275. doi:10.1038/
doi:10.1116/1.3043472 Nrm1890
31. Schvartzman M, Palma M, Sable J, Abramson 41. Geiger B, Spatz JP, Bershadsky AD (2009)
J, Hu X, Sheetz MP, Wind SJ (2011) Environmental sensing through focal adhesions.
Nat Rev Mol Cell Biol 10(1):21–33. 52. Sweetman MJ, Shearer CJ, Shapter JG,
doi:10.1038/nrm2593 Voelcker NH (2011) Dual silane surface func-
42. Moghimi SM, Hunter AC, Murray JC (2005) tionalization for the selective attachment of
Nanomedicine: current status and future pros- human neuronal cells to porous silicon.
pects. FASEB J 19(3):311–330. doi:10.1096/ Langmuir 27(15):9497–9503. doi:10.1021/
fj.04-2747rev la201760w
43. Deeg J, Axmann M, Matic J, Liapis A, Depoil 53. Dustin ML, Starr T, Varma R, Thomas VK
D, Afrose J, Curado S, Dustin ML, Spatz JP (2007) Supported planar bilayers for study of
(2013) T cell activation is determined by the the immunological synapse. In: Current proto-
number of presented antigens. Nano Lett cols in immunology. JohnWiley & Sons,
13(11):5619–5626. doi:10.1021/nl403266t New York, pp 18.13.11–18.13.35
44. Matic J, Deeg J, Scheffold A, Goldstein I, 54. Haynes CL, Van Duyne RP (2001) Nanosphere
Spatz JP (2013) Fine tuning and efficient T cell lithography: a versatile nanofabrication tool for
activation with stimulatory aCD3 nanoarrays. studies of size-dependent nanoparticle optics.
Nano Lett 13(11):5090–5097. doi:10.1021/ J Phys Chem B 105(24):5599–5611.
Nl4022623 doi:10.1021/jp010657m
45. Delcassian D, Depoil D, Rudnicka D, Liu ML, 55. Sun SQ, Mendes P, Critchley K, Diegoli S,
Davis DM, Dustin ML, Dunlop IE (2013) Hanwell M, Evans SD, Leggett GJ, Preece JA,
Nanoscale ligand spacing influences receptor Richardson TH (2006) Fabrication of gold
triggering in T cells and NK cells. Nano Lett micro- and nanostructures by photolitho-
13(11):5608–5614. doi:10.1021/Nl403252x graphic exposure of thiol-stabilized gold
46. Cai H, Depoil D, Palma M, Sheetz MP, Dustin nanoparticles. Nano Lett 6(3):345–350.
ML, Wind SJ (2013) Bifunctional nanoarrays doi:10.1021/Nl052130h
for probing the immune response at the single- 56. Schvartzman M, Wind SJ (2009) Robust pat-
molecule level. J Vac Sci Technol B Nanotechnol tern transfer of nanoimprinted features for sub-
Microelectron 31(6):6F902. 5-nm fabrication. Nano Lett 9(10):3629–3634.
doi:10.1116/1.4823764 doi:10.1021/nl9018512
47. Kalos M, Levine BL, Porter DL, Katz S, Grupp 57. Hu WC, Sarveswaran K, Lieberman M,
SA, Bagg A, June CH (2011) T cells with chi- Bernstein GH (2004) Sub-10 nm electron
meric antigen receptors have potent antitumor beam lithography using cold development of
effects and can establish memory in patients poly(methylmethacrylate). J Vac Sci Technol B
with advanced leukemia. Sci Transl Med 3(95). 22(4):1711–1716. doi:10.1116/1.1763897
doi:10.1126/scitranslmed.3002842 ARTN 58. Chen W, Ahmed H (1993) Fabrication of
95ra73 5–7 Nm wide etched lines in silicon using 100
48. Love JC, Estroff LA, Kriebel JK, Nuzzo RG, kev electron-beam lithography and polymeth-
Whitesides GM (2005) Self-assembled mono- ylmethacrylate resist. Appl Phys Lett
layers of thiolates on metals as a form of nano- 62(13):1499–1501. doi:10.1063/1.109609
technology. Chem Rev 105(4):1103–1169. 59. Chou SY, Krauss PR, Zhang W, Guo L,
doi:10.1021/Cr0300789 Zhuang L (1997) Sub-10 nm imprint lithog-
49. Schenk FC, Boehm H, Spatz JP, Wegner SV raphy and applications. J Vac Sci Technol B
(2014) Dual-functionalized nanostructured 15(6):2897–2904
biointerfaces by click chemistry. Langmuir 60. Blummel J, Perschmann N, Aydin D,
30(23):6897–6905. doi:10.1021/La500766t Drinjakovic J, Surrey T, Lopez-Garcia M,
50. Falconnet D, Koenig A, Assi T, Textor M Kessler H, Spatz JP (2007) Protein repellent
(2004) A combined photolithographic and properties of covalently attached PEG coat-
molecular-assembly approach to produce func- ings on nanostructured SiO2-based interfaces.
tional micropatterns for applications in the bio- Biomaterials 28(32):4739–4747.
sciences. Adv Funct Mater 14(8):749–756. doi:10.1016/j.biomaterials.2007.07.038
doi:10.1002/adfm.200305182 61. Zhu B, Eurell T, Gunawan R, Leckband D
51. Zhen G, Falconnet D, Kuennemann E, Vörös (2001) Chain-length dependence of the pro-
J, Spencer ND, Textor M, Zürcher S (2006) tein and cell resistance of oligo(ethylene
Nitrilotriacetic acid functionalized graft copo- glycol)-terminated self-assembled monolayers
lymers: a polymeric interface for selective and on gold. J Biomed Mater Res 56(3):406–416.
reversible binding of histidine-tagged proteins. d o i : 1 0 . 1 0 0 2 / 1 0 9 7 –
Adv Funct Mater 16(2):243–251. 4 6 3 6 ( 2 0 0 1 0 9 0 5 ) 5 6 : 3 < 4 0 6 : : A i d -
doi:10.1002/adfm.200500232 Jbm1110>3.0.Co;2-R
62. Irvine DJ, Purbhoo MA, Krogsgaard M, Davis J (2011) Controlled confinement of DNA at
MM (2002) Direct observation of ligand recog- the nanoscale: nanofabrication and surface
nition by T cells. Nature 419(6909):845–849 bio-functionalization. In: Zuccheri G, Samorì
63. Crites TJ, Maddox M, Padhan K, Muller J, B (eds) DNA nanotechnology: methods and
Eigsti C, Varma R (2015) Supported lipid protocols. Humana Press, Totowa, NJ,
bilayer technology for the study of cellular pp 169–185. doi:10.1007/978-1-61779-
interfaces. Curr Protoc Cell Biol 68:24 5.1–24 142-0_12
531. doi:10.1002/0471143030.cb2405s68
67. White LD, Tripp CP (2000) Reaction of
64. Liou HC, Pretzer J (1998) Effect of curing (3-aminopropyl)dimethylethoxysilane with
temperature on the mechanical properties of amine catalysts on silica surfaces. J Colloid
hydrogen silsesquioxane thin films. Thin Solid Interface Sci 232(2):400–407. doi:10.1006/
Films 335(1–2):186–191. doi:10.1016/ jcis.2000.7224
S0040–6090(98)00881–5
68. Roiter Y, Ornatska M, Rammohan AR,
65. Love JC, Wolfe DB, Haasch R, Chabinyc ML, Balakrishnan J, Heine DR, Minko S (2008)
Paul KE, Whitesides GM, Nuzzo RG (2003) Interaction of nanoparticles with lipid mem-
Formation and structure of self-assembled brane. Nano Lett 8(3):941–944
monolayers of alkanethiolates on palladium.
69. Roiter Y, Ornatska M, Rammohan AR,
J Am Chem Soc 125(9):2597–2609. Balakrishnan J, Heine DR, Minko S (2009)
doi:10.1021/Ja028692 Interaction of lipid membrane with nanostruc-

66. Palma M, Abramson JJ, Gorodetsky AA, tured surfaces. Langmuir 25(11):6287–6299.
Nuckolls C, Sheetz MP, Wind SJ, Hone doi:10.1021/la900119a
Chapter 19
Probing Synaptic Biomechanics Using Micropillar Arrays

Weiyang Jin, Charles T. Black, Lance C. Kam, and Morgan Huse
Abstract
Recent insights into the importance of mechanosensing and force transmission at the immune synapse
have spurred increased interest in the mechanical properties of leukocyte cell-cell interactions. In this chap-
ter, we describe an imaging-based strategy for measuring cellular forces that utilizes optically transparent
arrays of flexible micropillars. This approach has several distinct advantages over standard traction force
microscopy, and we anticipate that it will prove very useful for investigators who wish not only to quantify
ligand-induced forces with high spatiotemporal resolution but also to place those forces within the context
of a broader cell biological response.
Key words T cell, Mechanobiology, Polydimethylsiloxane, Traction force microscopy, Silicon etch-
ing, Micropillar, Signal transduction, Cytoskeleton
1 Introduction
A substantial fraction of intercellular communication in the

immune system occurs within transient, highly dynamic cell-cell
contacts broadly known as immune synapses. Communication at
these specialized interfaces is mediated both by membrane-
anchored molecules and by secreted proteins that are released
directly into the intercellular space. In recent decades, advances in
molecular biology have enabled the identification and characteriza-
tion of many of these factors in addition to their cognate receptors,
providing a basic understanding of the chemical recognition events
that drive information transfer. We also know a fair amount about
the cytosolic molecules that transduce receptor engagement at
immune synapses into intracellular signaling. It is important to
keep in mind, however, that these nanometer scale chemical pro-
cesses do not occur in isolation, but rather inside roiling cell-cell
contacts that are subject to continuous, micron-scale cytoskeletal
remodeling. Actin and microtubule-dependent movements can
impart considerable mechanical forces on molecules within immune
synapses, and it is becoming increasingly clear that these forces
regulate both the magnitude and the scope of intercellular
333
334 Weiyang Jin et al.
communication. It is now well established that integrins are mech-

anosensitive receptors that associate intracellularly with the actin
cytoskeleton and require actin-based pulling for their activation
and signaling [1, 2]. Other cell surface immunoreceptors, includ-
ing B cell receptors and T cell receptors, have also been shown to
display mechanosensitive properties [3–5], implying a critical role
for cellular mechanics in lymphocyte activation. Furthermore,
recent work with cytotoxic T cells has revealed that interfacial force
can enhance the potency of a secreted protein by altering the
responsiveness of synaptically engaged target cells to that protein
[6]. Hence, force exertion by immune cells is emerging as a key
aspect of their communicative potential.
To better understand the role of cellular mechanics in the
immune system, biophysical methods must be brought to bear that
enable high-resolution quantification of force exertion. In this
chapter, we describe one such method: image analysis of isolated T
lymphocytes on arrays of pliable micropillars. Uniform hexagonal
arrays are fabricated from polydimethylsiloxane (PDMS), a deform-
able, optically transparent elastomer that is easily coated with anti-
bodies and other proteins. On arrays coated with TCR ligand (e.g.,
anti-CD3 antibody or cognate peptide-major histocompatibility
complex (pMHC)) and costimulatory proteins (e.g., B7 or
ICAM1), T cells form radially symmetric, synapse-like interfaces
and begin to bend the pillars within minutes of initial contact [6,
7]. Because all pillars have the same dimensions and composition,
the force exerted by a cell on a pillar of interest is easily calculated
from the size of the pillar deflection.
The micropillar method provides enhanced spatial resolution
relative to traditional traction force microscopy, which makes use
of deformable gels embedded with small beads that act as reporters
for gel distortion [8]. Beads are distributed randomly in gel-based
systems, creating sampling gaps on the surface. By contrast, micro-
pillar arrays provide uniform coverage with pillars as closely spaced
as 1 μm from center to center [9]. Bead-based traction force
microscopy also presents a difficult data analysis problem because
force applied locally to the gel distorts not only that position but
also the surrounding area. Decoupling the effects of direct force
exertion from regional deformation, analogous to deconvolution
in microscopy, is numerically challenging. By contrast, each pillar
in a micropillar array moves independently of its neighbors, simpli-
fying both the calculation of forces and their positional assignment.
Micropillar technology was originally pioneered by Chris Chen
and colleagues for use in adherent cell systems [10]. We and others
have recently extended this approach to examine immune cells [6,
7, 11]. It has proven to be particularly useful for lymphocytes, such
as T cells, which form small, dynamic synapses on the order of 100
μm2. Importantly, incorporating other fluorescent probes (e.g., for
cytoskeletal markers and secretory events) into this experimental
Measuring Synaptic Force Exertion 335
framework is straightforward, which enables visual correlation of

force exertion with other cell biological processes.
The following protocols focus on pillar fabrication and live cell
imaging. Other methods relevant to these studies, including cell
culture and the preparation of recombinant proteins, are fairly
standard and may be found elsewhere. The process comprises four
steps, which are described individually in part 3. First, microfabri-
cation techniques are used to etch a hexagonal micropillar array
onto a silicon wafer. This silicon “master” is then used to prepare a
negative PDMS mold that contains the inverse pattern (i.e., pits
instead of pillars). Next, micropillar arrays made out of PDMS are
generated by stamping these negative molds on glass coverslips.
The micropillars are then coated with fluorescent streptavidin and
stimulatory proteins. Finally, cells are added to the arrays and
imaged live by fluorescence microscopy. The pillar deflections
observed in the resulting videos reflect the spatial distribution of
forces exerted against the array. Both the deflections and the asso-
ciated forces can be derived using standard data processing tools.
Although the protocols detailed below apply specifically to the
analysis of T cells, the micropillar approach can easily be adapted to
examine other immune cell types. As such, we believe it will be one
of the key technical components of the emerging field of leukocyte
mechanobiology.
2 Materials
2.1 Silicon Masters 1. P-type silicon wafers, 4-in. diameter, single side polished
(University Wafer, Inc.).
2. Oxygen Plasma System (March Plasma Model 1701F).
3. Spincoater (Brewer Science CEE100).
4. ZEP520A electron beam resist (Zeon Chemicals).
5. Anisole, ACS grade.
6. Hotplate (Brewer Science).
7. Electron Beam Lithography System (JEOL JBX6300).
8. Amyl acetate, ACS grade.
9. N2 gas.
10. Thin Film Deposition System (Kurt J. Lesker PVD75).
11. Chromium (Kurt J. Lesker).
12. Methyl-2-pyrrolidone, ACS grade.
13. Isopropanol, ACS grade.
14. Bath sonicator (Fisher).

15.
Inductively Coupled Plasma Etching Tool (Oxford
PlasmaLab100).
16. SF6 and O2 gas.

17. Cr etchant (type 1020, nitric acid/ceric ammonium nitrate
aqueous solution, Transene, Inc.).
2.2 Polydi- 1. VersaLaser Lasercut system.

methylsiloxane 2. 60 × 15 mm petri dish.
Micropillars
3. 100 × 15 mm petri dish.
4. Razorblades.
5. Microscope slide.
6. 22 × 22 mm glass coverslips, No. 0 thickness.
7. Ceramic coverslip rack (Thomas Scientific 8542E40).
8. 7× cleaning solution (MP Biomedicals ICN7667094).
9. Distilled, deionized water.
10. 250 mL Beaker with magnetic stirbar.
11. Digital stirring hotplate.
12. N2 gas.
13. Benchtop Furnace (Barnstead Thermolyne).
14. PDMS elastomer (Sylgard 184).
15. 50 mL conical tubes.
16. Tabletop centrifuge.
17. Aluminum foil.

18. Vacuum dessicator chamber (Bel-Art Vacuum Dessicator
999320237).
19. Oven (Fisher Scientific Isotemp Incubator).
20. Tweezers.
21. Plasma cleaner (Harrick Plasma Cleaner/Sterilizer PDC-32G).
22. Trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma 448931).
23. Vacuum container (Cole Parmer 700 mL Desi-Vac Container).
24. Alexa Fluor-conjugated streptavidin (488, 568, or 647).
25. Fluorescently conjugated (Alexa 488 or Alexa 647) anti-CD45
Fab fragment (clone 104).
26. Metal weights, 20–25 g, 4.5–5 g.
2.3 Stimulatory Stimulatory proteins are immobilized on pillar arrays via

Proteins streptavidin-biotin linkage. We initially used commercially available
biotinylated antibodies against CD3 and CD28 [7]. More recently,
we have used purified, biotinylated pMHC and murine ICAM1, a
ligand for the αLβ2 integrin LFA1. Class I and class II MHC pro-
teins are expressed in E. coli and refolded in the presence of
cognate peptide using established techniques (excellent protocols
are provided by the NIH Tetramer Core Facility). ICAM1 is gen-

erated by baculoviral infection of insect cells and purified as
described previously [12].
2.4 T cells We have used PDMS micropillar arrays to measure force exertion
by CD4+ and CD8+ T cells derived from both human and murine
sources. Polyclonal human T cells were purified by negative selec-
tion using standard magnetic bead technology to ~90% purity.
Murine T cells were derived from TCR transgenic animals and
either examined immediately as naive T cells or after in vitro dif-
ferentiation into armed CD4+ helper cells or CD8+ killer cells as
described [13].
2.5 Imaging We use an inverted Olympus IX-71 fluorescence microscope

Equipment equipped with a 100×/1.45 NA Plan Apochromat objective
(Olympus). 488, 568, and 647 nm illumination channels are avail-
able for visualization of lymphocytes and pillars. Environmental
control is achieved through an objective heater (Bioptechs) and a
stagetop incubator (Pathology Devices). Time-lapse acquisition is
controlled using Metamorph software.
3 Methods
3.1 Silicon Masters 1. First, clean a 4-in. diameter, single-side polished p-type silicon
wafer in oxygen plasma at 20 W RF power and 100 mT O2 for
5 min (Fig. 1a).
2. Use a spincoater to spread a thin layer of ZEP520A electron
beam resist (1:1 in anisole) onto the cleaned wafer. We typi-
cally deposit ~5 mL of resist in the center of the wafer and then
spin at 3000 rpm for 60 s. After spincoating, bake the wafer on
a hotplate at 180 °C for 90 s to remove residual solvent. The
resulting film of resist should be ~150 nm thick (Fig. 1b).
3. Pattern hexagonal arrays of dots onto the coated wafer by elec-
tron beam lithography. We use a JEOL JBX6300 system oper-
ating at 100 kV and 15 nA current (Fig. 1c). The electron-beam
dose is 350 C/cm2. The dot arrays typically cover a region of
1 mm2 and are composed of dots with diameters of 0.5, 1.0, or
1.5 μm, with dot separations of 1.0, 1.5, or 2.5 μm
center-to-center.
4. After electron-beam exposure, develop the pattern latent
images by immersing samples in amyl acetate for 90 s, fol-
lowed by rinsing in isopropanol and drying in stream of clean
N2 (Fig. 1d) (see Note 1).
5. To create a physical mask for etching pillars into the silicon
substrate, deposit a 30 nm layer of Cr onto the wafer by
(a) (b) (c) (d)

Starting wafer Resist added Electron beam Amyl acetate
patterning treatment
Cr SF6 + O2
(e) (f) (g) (h)

Chromium Resist Plasma Chromium
deposition removal etching removal
Fig. 1 Diagram schematizing the microfabrication of silicon masters. The silicon

wafer is shown in gray, the electron beam resist in purple, and the chromium in
yellow-green
electron-
beam evaporation using a Thin Film Deposition
System. We typically perform the deposition at ~10−6 Torr at a
rate of ~0.1 nm/s (Fig. 1e).
6. Strip the remaining electron-beam resist by immersing the
substrate in n-methyl-2-pyrrolidone at 80 °C for between 60
and 120 min, which also removes all Cr deposited on the top
resist surface (Fig. 1f). Rinse the sample in isopropanol for
1 min under ultrasonic agitation. The resulting surface should
be covered in arrays of micron scale Cr dots (Fig. 2a), which
will serve as a protective mask for sculpting pillars into the
underlying silicon by plasma etching.
7. Use plasma etching to sculpt the pillars. We carry out etching
at cryogenic temperatures (−100 °C) using a mixture of SF6
and O2, which preferentially removes silicon from the wafer
relative to the Cr mask (Fig. 1g). We use an Oxford PlasmaLab
100 inductively coupled plasma tool operating at 15 W RF
power, 800 W ICP power, and 12 mTorr. Using a 40 sccm:11
sccm SF6∶O2 we typically achieve a near vertical pillar etch pro-
file (Fig. 2b) and an etch rate of approximately 25 nm/s. An
initial 15 s high power breakthrough step (40 W RF power,
800 W ICP power, and 12 mTorr) with the same gas mixture
may be necessary prior to the main etch to remove silicon
oxide from the surface and fully initiate the silicon etch in all
exposed areas (Fig. 2b) (see Note 2).
8. After the silicon dry etching, remove the remaining Cr by
immersing the sample in Cr etchant for 1 min at 40 °C and
rinsing in isopropanol. Then, blow the sample dry in a stream
of N2 (Fig. 1h).
Fig. 2 Microfabrication images. (a) SEM image of the chromium dot pattern after
n-methyl-2-pyrrolidone wash. (b) SEM images of representative micropillar
arrays at low (top) and high (bottom) magnification
9. After drying, the master needs to be silanized to prevent PDMS

adhesion to the silicon. Carefully rinse the master with isopro-
panol, and blow dry with N2 gas. Be careful to not physically
touch the patterned area. Then, treat the master with air
plasma, pattern side up, for 2 min, to activate the surface.
10. Place the master into a vacuum desiccator chamber with 100
μL of Trichloro(1H,1H,2H,2H-perfluorooctyl)silane and
incubate under vacuum for 20 min (see Note 3). Store the
master in a clean and dry environment. It can also be stored
under cured PDMS.
3.2 Negative Molds Negative mold preparation is schematized in Fig. 3.

1. Prepare ~15 g PDMS using a 10:1 mixture (by mass) of Sylgard
base:curing agent. Measure components by weight into a 50
mL tube using a digital scale. Stir PDMS mixture well.
2. Centrifuge PDMS mixture for 3 min at ~1000 × g to remove
air bubbles caused by stirring.
3. Fully line a 60 × 15 mm petri dish with aluminum foil.
4. Place the silicon master in the foil-lined petri dish and pour the
PDMS mixture over the master until it is approximately 5 mm
thick.
Silicon Negative
masters molds
Master submerged Silanized

in PDMS Negative mold negative mold
Silicon master
Mold
After protein coating Submerged pillars Pillars
Fig. 3 Diagram schematizing the preparation of stimulatory PDMS micropillars. Pictures of silicon masters and
negative PDMS molds are shown above for reference
5. Degas the dish in a vacuum desiccator chamber for 1 h to

remove any extra air bubbles.
6. Gently blow N2 gas over the PDMS to remove any bubbles still
trapped under the PDMS surface.
7. Place the aluminum dish in an oven set at 65 °C for 8 h up to
overnight to fully cure the PDMS.
8. Peel the aluminum foil off the PDMS.
9. Cut away excess PDMS from the bottom of the master using a
razorblade.
10. Gently peel the master off the PDMS by slightly bending the
PDMS mold and lifting off the master using a pair of twee-
zers. The patterned area should appear birefringent under
white light.
11. Cut the PDMS using a razorblade along the edges of the
indentation left by the master. If desired, trim excess PDMS,
leaving behind a border around the pattern for ease of han-
dling. Generally, for a 1 × 1 mm pattern, we aim for a negative
mold with a face size of 7 × 7 mm.
12. Treat the negative mold with oxygen or air plasma, pattern side
up, for 2 min to activate the PDMS surface.
13. Place the negative mold into a vacuum container with 5 μL of
Trichloro(1H,1H,2H,2H-perfluorooctyl)silane. Fully pump
the vacuum container and expose the negative mold to silane
vapor for 8 h up to overnight (see Note 3).
3.3 Micropillar 1. Lasercut a 12 mm diameter circle onto a 60 × 15 mm petri

Arrays dish. This circle will be called a well in the rest of the
protocol.
3.3.1 Prepare
in Advance—Dish Wells 2. Level the edges of the well with a razor, making sure the out-
side bottom of the dish is completely flat for imaging.
3.3.2 Micropillars PDMS micropillar preparation is schematized in Fig. 3.

1. Prepare ~ 10 g PDMS using a 10:1 mixture (by mass) of Sylgard
base:curing agent. Measure components by weight into a 50 mL
tube using a digital scale. Stir PDMS mixture well.
2. Centrifuge PDMS mixture for 3 min at ~ 1000 x g to remove
air bubbles caused by stirring.
3. Using a 200 μL pipette tip or gel-loading tip, gently spread
~10 ul PDMS mixture onto the negative mold so that the pat-
tern is fully covered. Be sure not to touch the pattern directly
with the pipette tip (see Note 4).
4. Place negative mold, pattern side up, onto a petri dish, and
place dish into a vacuum dessicator chamber. Degas for 1 h to
remove any extra air bubbles and to fully fill the negative mold.
5. During the desiccation step, use the edge of a microscope slide
to spread some of the PDMS mixture on the outside bottom
of the lasercut dish, making sure the PDMS covers the area
around the well.
6. Using a pair of tweezers, place a cleaned 22 × 22 mm thickness
#0 glass coverslip (see support protocol) on the outside bot-
tom of the petri dish, making sure to center the coverslip over
the well. Gently push down on the coverslip with the twee-
zers to spread the PDMS mixture homogenously underneath
the coverslip.
7. Place petri dish into an oven set at 65 °C, coverslip side facing
up. Place a 20–25 g weight on the top of the center of the
coverslip.
8. Bake coverslip and dish for 1 h at 65 °C.
9. Remove weight and place dish, coverslip side down, onto a
clean benchtop. Allow the dish to cool to room temperature
before proceeding.
10. Using a pair of tweezers, invert the negative mold, pattern side
down, onto the coverslip in a clean single motion without
introducing bubbles in the PDMS. Gently press down with
tweezers on the negative mold to spread out the PDMS
between the mold and the coverslip.
11. Place dish into a 65 °C oven. Place a 4.5–5 g weight on the top
of the center of the negative mold.
12. Bake the dish at 65 °C for 8 h to overnight.

13.
Allow the dish to cool to room temperature before
proceeding.
3.4 Pillar 1. Pour 100% ethanol into the dish until the negative molds are
Functionalization submersed. High aspect ratio pillars (height/diameter greater
than 3) must be peeled under a liquid with low surface tension
(e.g., 100% ethanol) to prevent pillar collapse.
2. Using a pair of tweezers, gently peel the negative molds off the
coverslip in a single motion (see Note 5).
3. Exchange the 100% ethanol in the dish with 1× PBS by simul-
taneously removing the solution in the dish and adding in 1×
PBS. In this step, and in all subsequent steps, make sure the
pillars are continuously submerged in liquid, as they will col-
lapse if exposed to air. For washes or buffer exchanges, we use
one pipette to remove liquid as we add liquid using another
pipette.
4. Remove all the 1× PBS in the dish apart from the 1× PBS
inside the well.
5. Replace the 1× PBS in the well with 20 μg/mL of AlexaFluor-
conjugated streptavidin in 1× PBS. Incubate in the dark for 1
h at room temperature.
6. Thoroughly wash the pattern with 1× PBS (~5 exchanges of
buffer).
7. Replace the 1× PBS in the well with 20 μg/mL biotinylated
protein (e.g., 10 μg/mL biotinylated pMHC and 10 μg/mL
biotinylated ICAM1) in 1× PBS. Incubate in the dark for 1 h
at room temperature.
8. Thoroughly wash the pattern as in step 6.
3.5 Live Cell Imaging 1. An hour before imaging, preheat the objective and the imag-
ing stage with an objective heater and a stage top incubator.
2. Replace the 1× PBS in the well containing the functionalized
pillars with ~100 μL of complete imaging medium (we use
RPMI containing 10% fetal bovine serum, 2 mM glutamine, 1
mM sodium pyruvate, 1× nonessential amino acids mix, and
10 mM Hepes pH 7.5, but without phenol red).
3. Cover the dish and place it in an incubator until the cells are
ready.
4. Preincubate ~1 × 106 T cells in medium containing 1 μg/mL
fluorescently conjugated anti-CD45 Fab fragment (Alexa Fluor
488 or 647) for 20 min at room temperature.
5. Wash cells three times with imaging medium to remove excess
Fab fragment.
Fig. 4 T cells on stimulatory micropillar arrays. (a) Fluorescence image of a

murine CD8+ cytotoxic T cell bound to micropillars coated with stimulatory pMHC
and ICAM1. Yellow asterisks mark substantial pillar deflections. (b) SEM image of
a murine CD4+ T cell bound to micropillars. In both experiments, the center-to-
center distance between pillars was 2 μm
6. Place the well on the microscope stage.

7. Visualize the pillars by imaging fluorescently labeled streptavi-
din (e.g., Alexa Fluor 568). Focus on the pillar tops, which
appear as circular dots of fluorescence (as opposed to rings of
fluorescence, which denote the pillar stalks).
8. Add 1 × 105 T cells in ~30 μL, bringing the final volume to
~130 μL.
9. Start imaging immediately. We typically record images of both
the pillar tops (Alexa Fluor 568) and the T cells (Alexa Fluor
488 or 647) every 15 s for a total of 30 min (Fig. 4). Cell num-
ber should be adjusted in subsequent runs to maximize the
number of isolated T cells in the imaging field.
3.6 Image Analysis Pillar deflections are determined by tracking the xy positions of the
pillar tops over time. To generate pillar tracks, we use a particle-
3.6.1 Pillar Tracking
tracking package written for Matlab by Daniel Blair and Eric
Dufresne (http://site.physics.georgetown.edu/matlab/), which
was adapted from IDL code written by David Grier, John Crocker,
and Eric Weeks. A spatial bandpass filter is first applied to all images
to reduce noise. Particles in each image are then identified based
on their brightness and expected size. Finally, the particle images
are stitched into trajectories based on the expected values of maxi-
mum particle displacement between frames (see Note 6). Pillars are
assigned as being in contact with the T cell in question if their
coordinates overlap with the T cell envelope (derived from images
of the fluorescent anti-CD45 Fab).
3.6.2 Force Calculations The force associated with each micropillar deflection can be calcu-
lated from the deflection length, the pillar dimensions, and the
pillar composition using the following bending formula
F = k bend d = (3p ED 4 / 64L3 )d
where δ is the pillar displacement, E is the elastic modulus of the

material, D is the diameter of the pillar, and L is the pillar height
[14]. We use home-made Matlab scripts to do this. In general,
forces exerted on pillars beneath T cells are compared to the forces
experienced by “background” pillars that are not contacted. The
particle-tracking data also enable analysis of the spatial distribution
of force exertion within the synapse and its directionality over time.
For example, during synapse growth we generally observe out-
wardly directed radial forces concentrated in the periphery of the
contact. As synapse size stabilizes, these forces reverse polarity,
resulting in centripetal “squeezing” of the pillar array [6, 7].
3.7 Support Protocol 1. Place 22 × 22 mm thickness #0 glass coverslips on a ceramic

rack.
3.7.1 Clean Coverslips
(For Pillar Stamping) (a) Coverslip thickness should be minimized to minimize the
distance between the microscope objective and the pillar
top. The distance between the pillar tops and the objective
needs to be less than the working distance of the
objective.
2. Make enough 1× detergent with DI water to submerge cover-
slips in a 250 mL beaker.
3. Gently place a magnetic stir bar and then the rack into the
beaker.
4. Place beaker on a hotplate stirrer and boil the detergent at
80–90 °C with a spin speed of 200 rpm for approximately
40 min until the detergent becomes clear. Maintain conditions
for 30 min.
5. Wash coverslips by submerging the rack and coverslips in DI
water ten times. Replace the DI water and repeat this step
twice.
6. Gently blow N2 gas over the coverslips to remove most of the
DI water on the coverslips.
7. Place the rack and the coverslips in a clean dry beaker. Cover
beaker top with aluminum foil and place beaker in a furnace.
8. Heat furnace to 450 °C and maintain this temperature for 8 h
or overnight.
9. Let coverslips cool to room temperature before use.
10. Store coverslips in a clean and dry environment. Coverslips
should be used within 2 weeks of cleaning.
4 Notes
1. The amyl acetate wash selectively removes the irradiated areas

of electron beam resist.
2. This etching protocol recipe can be used to generate silicon
pillars several microns tall.
3. The speed of silanization will vary with the amount of silane,
the vacuum chamber volume, the vacuum strength, and the
incubation time. Optimization of these factors should be per-
formed to achieve thorough silane coverage and prevent over-
silanization. If over-silanization occurs, the PDMS will have a
thin, semi-transparent coating visible to the eye.
4. Spread only enough PDMS to just fully cover the pattern. Too
much PDMS may result in increased distance between the
microscope objective and the pillars, leading to difficulties
focusing on the pillar tops.
5. Negative molds can be reused as long as there is no damage to
the pattern. Use an inverted light microscope to check pattern
fidelity.
6. The automated particle-tracking process is quite robust, and in
our experience only becomes problematic when pillar deflec-
tions are so large that they force pillar tops together, confound-
ing their assignment as separate particles. In cases like these,
manual tracking is often required to determine a pillar track.
Acknowledgments
We thank A. Gondarenko for assistance with micropillar fabrica-

tion and members of the Huse and Kam labs for advice. This work
is supported by the US National Institutes of Health (R01-AI087644
to M.H., R01-AI088377 to L.C.K., and PN2-EY016586 to
L.C.K.), the Geoffrey Beene Cancer Research Center (M.H.), the
Starr Cancer Consortium (M.H.), and the Leukemia and
Lymphoma Society (M.H.). This research used resources of the
Center for Functional Nanomaterials, which is a U.S. DOE Office
of Science Facility, at Brookhaven National Laboratory under
Contract No. DE-SC0012704.
References
1. Comrie WA, Babich A, Burkhardt JK (2015) 3. Wan Z, Chen X, Chen H, Ji Q, Chen Y, Wang J,
F-actin flow drives affinity maturation and spa- Cao Y, Wang F, Lou J, Tang Z, Liu W (2015)
tial organization of LFA-1 at the immunologi- The activation of IgM- or isotype-switched IgG-
cal synapse. J Cell Biol 208:475–491 and IgE-BCR exhibits distinct mechanical force
2. Friedland JC, Lee MH, Boettiger D (2009) sensitivity and threshold. Elife 4:e06925
Mechanically activated integrin switch controls 4. Wan Z, Zhang S, Fan Y, Liu K, Du F, Davey
alpha5beta1 function. Science 323:642–644 AM, Zhang H, Han W, Xiong C, Liu W
(2013) B cell activation is regulated by the local contractions on submicrometer pillars.

stiffness properties of the substrate presenting Proc Natl Acad Sci U S A 109:5328–5333
the antigens. J Immunol 190:4661–4675 10. Tan JL, Tien J, Pirone DM, Gray DS,
5. Liu B, Chen W, Evavold BD, Zhu C (2014) Bhadriraju K, Chen CS (2003) Cells lying on a
Accumulation of dynamic catch bonds between bed of microneedles: an approach to isolate
TCR and agonist peptide-MHC triggers T cell mechanical force. Proc Natl Acad Sci U S A
signaling. Cell 157:357–368 100:1484–1489
6. Basu R, Whitlock BM, Husson J, Le Floc’h A, 11. Baker DW, Liu X, Weng H, Luo C, Tang L
Jin W, Oyler-Yaniv A, Dotiwala F, Giannone (2011) Fibroblast/fibrocyte: surface interac-
G, Hivroz C, Biais N, Lieberman J, Kam LC, tion dictates tissue reactions to micropillar
Huse M (2016) Cytotoxic T cells use mechan- implants. Biomacromolecules 12:997–1005
ical force to potentiate target cell killing. Cell 12. Abeyweera TP, Merino E, Huse M (2011)
165:100–110 Inhibitory signaling blocks activating receptor
7. Bashour KT, Gondarenko A, Chen H, Shen K, clustering and induces cytoskeletal retraction
Liu X, Huse M, Hone JC, Kam LC (2014) in natural killer cells. J Cell Biol 192:
CD28 and CD3 have complementary roles in 675–690
T-cell traction forces. Proc Natl Acad Sci U S 13. Quann EJ, Merino E, Furuta T, Huse M
A 111:2241–2246 (2009) Localized diacylglycerol drives the
8. Dembo M, Wang Y-L (1999) Stresses at the polarization of the microtubule-organizing
cell-to-substrate interface during locomotion center in T cells. Nat Immunol 10:627–635
of fibroblasts. Biophys J 76:2307–2316 14. Schoen I, Hu W, Klotzsch E, Vogel V (2010)
9. Ghassemi S, Meacci G, Liu S, Gondarenko Probing cellular traction forces by micropillar
AA, Mathur A, Roca-Cusachs P, Sheetz MP, arrays: contribution of substrate warping to
Hone J (2012) Cells test substrate rigidity by pillar deflection. Nano Lett 10:1823–1830
Chapter 20
Microchannels for the Study of T Cell Immunological

Synapses and Kinapses
Hélène D. Moreau, Philippe Bousso, and Ana-Maria Lennon-Duménil
Abstract
T Cells can form very stable (synapses) or very transient and migratory (kinapses) contacts with antigen-
presenting cells. Here, we describe how microchannels can be used to conveniently study the distinct
dynamics of T cells during antigen recognition. Microchannels provide a controlled confined environment
that promotes T cell migration and recapitulates kinapse and synapse behaviors when coated with appro-
priate pMHC molecules. We also depict the advantages of this in vitro approach for addressing mechanistic
issues and for analysis.
Key words Microchannels, Synapse/kinapse, Recombinant pMHC, Antigen affinity, Imaging
1 Introduction
In the last decade, microchannels have become an invaluable tool

to study cell migration in vitro [1, 2]. Indeed, they recapitulate the
confinement cells encounter in vivo and promote cell migration.
They also offer a very controlled experimental environment, allowing
the use of various coatings to simulate extracellular matrix or
adjacent cells as well as various inhibitors to investigate the mecha-
nisms underlying cell migration. Finally, by standardizing cell shape
and migration in one dimension, they strongly facilitate analysis
and extraction of quantifiable migration parameters.
Microchannels have been used to study steady-state migration
of T cells [3]. Using channels of different sizes as well as different
coatings, the authors showed that confinement had a bigger impact
on T cell migration than adhesion molecules such as integrin
ligands. Later on, we set up the use of microchannels for the study
of T cell dynamics during antigen recognition. To this end, we
used 3 μm × 6 μm channels (in which T cell speed is similar to the
one observed in vivo), coated with pMHCs of varying affinities for
the OT-1 transgenic TCR [4, 5], including an irrelevant pMHC as
a control. We demonstrated that T cell dynamics in microchannels
347
348 Hélène D. Moreau et al.
recapitulate many features of T cell dynamics in vivo during anti-

gen recognition. Specifically, we found that microchannels coated
with low-affinity ligands promoted partial T cell deceleration and a
scanning behavior reminiscent of kinapses. In contrast, high-
affinity pMHC favored a more stable T cell arrest, resembling
immunological synapses. We further used this experimental setup
to address the molecular mechanisms controlling the formation of
kinapses and synapses. In particular we evaluated the role of extra-
cellular calcium using calcium-depleted medium as well as the role
of TCR signaling using a specific inhibitor.
Here, we describe the methodology to use microchannels for
the study of T cell immune synapses and kinapses. This review will
include (1) how to prepare microchannel chips for the study of T
cell synapses and kinapses; (2) how to pre-activate, load, and image
the T cells in the chips; and finally (3) how to analyze the gener-
ated movies.
2 Materials
2.1 Microchannel 1. Epoxy mold of straight channels, 3 μm × 6 μm (for pre-

Fabrication activated T cells). This mold needs to be prepared by a special-
ized microfluidic lab and can be asked for to the Lennon Lab.
2. PDMS RTV-615 (Neyco).
3. Vacuum chamber.
4. Plasma cleaner (Harrick Plasma).
5. Oven at 65 °C.
6. Compressed filtered air pistol.
7. Absolute ethanol.
8. Glass bottom petri dishes (Fluorodish WPI FD35-100).
9. Self-healing cutting mat (Harris Uni-Core).
10. Biopsy punches with plunger 2.0 or 3.0 mm (Harris Uni-Core
or Miltex).
11. Scalpels.
12. Tweezers.
2.2 Solution 1. Recombinant pMHC of varying affinity for the TCR can be
of Recombinant pMHC prepared as described previously [6]. Prepare 10 μg/mL
solutions.
2. pMHC can be supplemented with adhesion or co-stimulation
molecules.
2.3 Preparation 1. Anti-CD3/CD28 beads (Dynal).

of T Cells 2. Recombinant IL2 (Roche).
3. CD4 or CD8 negative magnetic isolation kit (Dynal or Miltenyi).
Synapses and Kinapses in Microchannels 349
4. Complete RPMI: RPMI 1640 + 2 mM L-alanyl-L-glutamine,

10% SVF, 50 U/mL penicillin, 50 µg/mL streptomycin, 1
mM sodium pyruvate, 10 mM HEPES, 0.1% β-mercaptoethanol.
5. Complete RPMI without phenol red for imaging.
3 Methods
3.1 Preparation 1. Prepare 10% PDMS mix. For 20 g of PDMS (ten chips in aver-
of Microchannels age), weight 2 g of curing agent and 18 g of PDMS in a small
weighting cup. Mix vigorously with a pipet tip for example. Be
careful to mix the entire curing agent by scraping the bottom
of the weighting cup. The PDMS should be full of bubbles.
2. Pour PDMS in the epoxy mold and place it in the vacuum
chamber for about 2 h, until all the bubbles disappear.
Alternatively, pop up the last bubbles with compressed air (see
Note 1). Cook the PDMS for 1 h at 65 °C (see Note 2).
3. Carefully cut the chips out of the epoxy mold with a scalpel
(Fig. 1a). Keep them always facing up and avoid touching the
channel face with your fingers. With a biopsy punch, make
holes in the chip where indicated by the design (Fig. 1b).
4. Clean the PDMS chips from any dust with tape twice. Put
them in absolute ethanol and sonicate for 30 s. Dry them care-
fully with the compressed air pistol (see Note 3).
5. Put on the plasma cleaner plate the chips (face up) as well as
the bottom of the Petri dishes (Fig. 1c). Plasma treat for 30 s
to 1 min (see plasma cleaner manufacturer’s instructions). Take
the plate out of the plasma cleaner. To assemble your micro-
channels, take the PDMS chip carefully by its sides with twee-
zers, and put it upside down in the Petri dish (Fig. 1d). It
should stick immediately. If not, gently press the PDMS with
tweezers on the sides of the chip (avoid pressing on the chan-
nels since it could deform them). Put the Petri dish lid back on.
Cook for 1 h at 65 °C.
6. Prepare the coating solutions. 10 μL of solution are required
for each hole (i.e., 60 μL per chip). In order to image synapses
and kinapses, we used recombinant pMHC of varying affinities
for the transgenic TCR as well as an irrelevant pMHC as a
steady-state control (see Note 4). Put your assembled chips
(lids off) in the plasma cleaner for 10 min with vacuum on but
plasma off, and then plasma treat for 30 s. Immediately after,
pipet 10 μL of coating solution in each hole of the chips. You
should see the liquid progressing rapidly in the microchannels
until it fills them completely (see Note 5). Coat for 1 h at room
temperature (see Note 6).
Fig. 1 Microchannel preparation. (a) Epoxy mold. (b) Microchannel chip with design (bottom left), before (top)
and after (bottom right) making the holes. (c) Plasma treatment of the chips and Petri dishes. (d) Assembly of
the microchannel chip (left) and assembled chip (right)
7. Rinse carefully the channels. In order to do this, cover the

chips with PBS. Aspirate all PBS out (including in the holes).
Repeat three times, and then repeat three times with complete
RPMI without phenol red. Cover the chips with medium and
incubate at 37 °C for 30 min. If using inhibitors (or a specific
medium, e.g., calcium-depleted medium), rinse with medium
+ inhibitor and incubate for at least 2 h before putting the cells
in (see Notes 7 and 8).
3.2 Pre-activation, 1. Two to three days before the experiment, pre-activate the T
Loading of T Cells, cells (see Note 9). Collect lymph nodes and spleen and mash
and Imaging them through a 70 μm cell strainer. Purify TCR-transgenic T
cells with the appropriate (CD4 or CD8) negative selection
kit, following exactly manufacturer’s instructions (see Note
10). Cultivate the cells at 106 cells/mL of complete RMPI
with anti- CD3/CD28 beads (bead/cell ratio of 1:4) and
recombinant IL2 (25 U/mL) for 2–3 days. If used at day 3,
cells should be split at day 2.
2. On the day of the experiment, remove the beads from the T

cell suspension with the Dynal magnet. Centrifugate and resus-
pend at 2 × 105 cells/10 μL of complete RPMI without phenol
red (see Note 11). If using inhibitor (or specific medium),
resuspend cells directly in inhibitor (see Note 12).
3. Empty the chips of medium (including holes) as during rins-
ing. Add 10 μL of cells (i.e., 2 × 105 cells) in each hole. Add 1
mL of medium around the PDMS so that the holes containing
the cells do not dry. Incubate at 37 °C for 15–30 min to let the
cells sediment. Add as much medium as needed to completely
cover the PDMS.
4. Image overnight every 30 s to 4 min at 10–40× magnification
depending on the aim of the experiment (see Note 13). Always
image the same positions with respect to the holes, or write
down carefully the positions to be able to detect any position
effect.
3.3 Quantification 1. Kymographs can be made easily with ImageJ. Crop the movie
to select one channel, and then use the “make montage” func-
tion of Image J (Fig. 2a).
2. Tracking can be made using Imaris (see Note 14). This method
gives the classical parameters such as cell speed, persistence, or
arrest coefficient. Cell speed, persistence, and arrest coefficient
can be used to classify cell behavior in “migrating,” “kinapse-
like,” and “synapse-like” (see Note 15).
3. Tracking of subcellular compartment can also be performed
using Imaris (Fig. 2b). This allows calculating relative posi-
tions of organelles from the coordinates.
4. Distribution of a fluorescent protein can be analyzed simply
with ImageJ by realizing “line scans” along the middle axis of
the channel (Fig. 2c) (see Note 16).
4 Notes
1. If you don’t have a vacuum chamber, you can leave the PDMS
to solidify overnight at room temperature. The bubbles will
disappear as efficiently.
2. Once cooked, the PDMS chips can be kept in the epoxy for a
few weeks. Avoid keeping the epoxy mold empty of PDMS to
limit dust accumulation.
3. The drying step is very important since if there is any ethanol
left, it can be released from the PDMS during imaging and be
toxic for the cells. This is why it is always preferable to use
absolute ethanol and not 70% ethanol (that dries less easily).
A
Kinapse Synapse B Steady state Kinapse
Time
Time
MTOC nucleus
Time
Fluo intensity
Fig. 2 Analysis of T cells in microchannel. (a) Kymograph corresponding to a cell forming a kinapse (left) or
a synapse (right), generated with ImageJ. Scale bar, 100 μm. (b) Tracking of subcellular compartments with
Imaris: original images (top) and tracking result (bottom). Scale bar, 20 μm. (c) Analysis of fluorescence
along the axis of the cell (here LAT-GFP) using ImageJ. Representative image (bottom) and linescans of fluo-
rescence intensity along the central axis of the channel. Reproduced from Moreau et al. 2015 [5] with
permission of PNAS
4. We use OVA variant pMHC for the OT-1 TCR [4, 5]. Varying
the dose of pMHC and/or adding costimulation or adhesion
molecules may impact the dynamics of the T cells.
5. Coating right after cooking (when the chips are still warm)
makes it more efficient.
6. If needed, the procedure can be paused here: cover the chips
with PBS and keep them overnight at 4 °C.
7. It is also possible to load the channels with cells first and add
the inhibitor once the cells are already in the channels. In this
case, it is better to use “short channels”; the geometry of
which favors rapid diffusion and in which inhibitor concentra-
tion should equilibrate after about 1 h (after which imaging
can be started).
8. Procedure can be paused here: the chips can be stored in com-
plete medium overnight at 37 °C.
9. Pre-activated T cells enter more easily the channels because
they are more motile than naïve T cells, which tend not to
enter channels, probably because they are not polarized at
steady state on 2D surfaces. Finding conditions favoring naïve
T cell polarization and/or migration (such as chemokines)
could enable to perform the experiment with naïve T cells. As
naïve T cells are smaller than pre-activated T cells, it might be
more appropriate to use channels with smaller section.
10. If using a TCR transgenic mouse on a Rag knockout back-

ground (e.g., OT-1 Rag1−/− mice), purification with a kit is
not necessary as all cells collected from mashing the lymph
nodes are the TCR-transgenic T cells of interest. Just collect
the lymph nodes and mash them through a 70 μm cell strainer.
Then put them in culture with anti-CD3/anti-CD28 beads
and IL-2 as indicated in the main text.
11. Using medium without phenol red increases the quality of the
movies by limiting fluorescence interference.
12. Some inhibitors may require preloading of the cells before the
actual experiment.
13. The magnification as well as the timing of imaging really

depends on the aim of the experiment. Magnification of 10× is
enough to characterize cell migration, but higher magnifica-
tion enables to describe intracellular compartments. Timing of
imaging needs to be balanced between the aim of the experi-
ment, the number of conditions, and the speed of the
microscope.
14. Using fluorescent cells for the experiment makes the tracking
with Imaris much easier. Using the “surface-tracking” function
is preferable to “spot tracking” since it also provides cell mor-
phology and fluorescence information. Tracking can also be
performed using Matlab to generate kymographs and analyze
them (Matlab code can be requested from the Lennon Lab).
15. Cells can also be stained with different dyes to follow cell com-
partments (e.g., Hoechst for nucleus) or retrovirally trans-
duced to express fluorescent proteins of interest [7]. Loading
with certain dyes (such as calcium dyes) is tricky, as some dyes
tend to leak from the cells as they are pumped from the cells
by active transport.
16. If the fluorescent protein is not distributed equally between
the central axis and the sides of the cell, it is preferable to use
the “plot profile” function of ImageJ on the whole image. A
recent macro has been developed in the Lennon team to per-
form mapping of fluorescent proteins [8]. This macro can be
asked for to the Lennon Lab.
Acknowledgment
This work was supported by INSERM, Institut Curie, ANR-10-

IDEX-0001-02 PSL*, ANR-11-LABX-0043 and ERC grant
Strapacemi to AML-D, INSERM, Institut Pasteur and ERC start-
ing grant LymphocyteContacts to P.B., and Association pour la
Recherche sur le Cancer (ARC-PDF20140601095) to H.D.M.
References
1. Heuze ML, Vargas P, Chabaud M, Le Berre M, strength regulates antigen-mediated T-cell

Liu YJ, Collin O, Solanes P, Voituriez R, Piel deceleration by distinct mechanisms to pro-
M, Lennon-Dumenil AM (2013) Migration of mote local exploration or arrest. Proc Natl Acad
dendritic cells: physical principles, molecular Sci U S A 112(39):12151–12156
mechanisms, and functional implications. 6. Bousso P, Casrouge A, Altman JD, Haury M,
Immunol Rev 256(1):240–254 Kanellopoulos J, Abastado JP, Kourilsky P
2. Vargas P, Terriac E, Lennon-Dumenil AM, Piel (1998) Individual variations in the murine T
M (2014) Study of cell migration in microfabri- cell response to a specific peptide reflect vari-
cated channels. J Vis Exp 84:e51099 ability in naive repertoires. Immunity
3. Jacobelli J, Friedman RS, Conti MA, Lennon- 9(2):169–178
Dumenil AM, Piel M, Sorensen CM, Adelstein 7. Azar GA, Lemaitre F, Robey EA, Bousso P
RS, Krummel MF (2010) Confinement- (2010) Subcellular dynamics of T cell immuno-
optimized three-dimensional T cell amoeboid logical synapses and kinapses in lymph nodes.
motility is modulated via myosin IIA-regulated Proc Natl Acad Sci U S A 107(8):3675–3680
adhesions. Nat Immunol 11(10):953–961 8. Vargas P, Maiuri P, Bretou M, Saez PJ, Pierobon
4. Moreau HD, Lemaitre F, Terriac E, Azar G, P, Maurin M, Chabaud M, Lankar D, Obino D,
Piel M, Lennon-Dumenil AM, Bousso P (2012) Terriac E, Raab M, Thiam HR, Brocker T,
Dynamic in situ cytometry uncovers T cell Kitchen-Goosen SM, Alberts AS, Sunareni P,
receptor signaling during immunological syn- Xia S, Li R, Voituriez R, Piel M, Lennon-
apses and kinapses in vivo. Immunity Dumenil AM (2016) Innate control of actin
37(2):351–363 nucleation determines two distinct migration
5. Moreau HD, Lemaitre F, Garrod KR, Garcia Z, behaviours in dendritic cells. Nat Cell Biol
Lennon-Dumenil AM, Bousso P (2015) Signal 18(1):43–53
Chapter 21
Purification of LAT-Containing Membranes from Resting

and Activated T Lymphocytes
Claire Hivroz, Paola Larghi, Mabel Jouve, and Laurence Ardouin
Abstract
In T lymphocytes, the immune synapse is an active zone of vesicular traffic. Directional transport of vesicu-
lar receptors and signaling molecules from or to the immune synapse has been shown to play an important
role in T-cell receptor (TCR) signal transduction. However, how vesicular trafficking is regulating the
activation of T cells is still a burning question, and the characterization of these intracellular compartments
remains the first step to understand this process. We describe herein a protocol, which combines a separa-
tion of membranes on flotation gradient with an affinity purification of Strep-tagged fusion transmembrane
proteins with Strep-Tactin® resin, allowing the purification of membranes containing the Strep-tagged
molecule of interest. By keeping the membranes intact, this protocol leads to the purification of molecules
physically associated with the Strep-tagged protein as well as of molecules present in the same membrane
compartment: transmembrane proteins, proteins strongly associated with the membranes, and luminal
proteins. The example shown herein is the purification of membrane compartment prepared from T lym-
phocytes expressing LAT fused to a Strep-tag.
Key words T Lymphocyte, Immune synapse, Flotation gradient, Strep-tag-Strep-tactin affinity

purification, LAT
1 Introduction
How the activation of T cells is regulated remains a fundamental

question for understanding the adaptive immune response. It has
been 17 years now that the immune synapse has been described by
both A. Kupfer’s and M.L. Dustin’s groups [1, 2]. Since then a
growing number of papers have investigated the spatiotemporal
organization of this structure as well as the mechanisms involved in
its formation and its role in T-cell activation. The application of
microscopy techniques revealed that different receptors expressed
by T lymphocytes and involved in T-cell activation localize to
defined supramolecular activation clusters (SMACs) in the contact
zone between the T lymphocyte and the antigen-presenting cell
(APC), with antigen receptors distributing to the center of this
355
356 Claire Hivroz et al.
region (cSMAC), and adhesion molecules accumulating at the

periphery (pSMAC) [1, 2]. Development of high-resolution
microscopy then allowed for a better description of the dynamics
of these contact regions, showing that receptors and signaling mol-
ecules aggregate in dynamic microclusters, which form at the
periphery of the immune synapse and migrate toward its center [3,
4]. These microscopy-based analyses also allowed the characteriza-
tion of the polarized trafficking and secretion of vesicular compo-
nents at the immune synapse [5–7].
Since these early studies, it has become evident that the
immune synapse is an active zone for directional exocytosis, endo-
cytosis, receptor recycling, and more generally vesicular traffic
[8–11]. Yet, how vesicular trafficking is involved in T-cell activa-
tion is still only partially understood.
Several groups including ours have shown that intracellular
vesicles containing signaling molecules are actively involved in sig-
nal transduction processes in T lymphocytes, providing unique
platforms for specific signaling complexes to assemble or be acti-
vated [12–15]. These intracellular vesicles, which localize close to
or interact with the membrane at the immune synapse, may regu-
late T-cell activation by transporting cargo to or from specific sites
of the immune synapse. They may also directly control the T lym-
phocyte signaling process by acting as mobile platforms that com-
partmentalize and organize signaling inside the cell. Along this
line, it is worth noting that location and intracellular trafficking
regulate the signaling induced by many receptors such as EGFR
[16], TGFβR [17], TLR [18], and BCR [19]. Intracellular loca-
tion of signaling also matters in T lymphocytes. Indeed, differen-
tial intracellular locations of signaling molecules have been shown
to translate to different functional outcomes [20–22].
Finally, there is emerging evidence that different pools of sub-
synaptic vesicles containing signaling molecules exist [14, 23, 24].
Yet, their relative role and content are unknown. The mechanisms
controlling the traffic of these different vesicular compartments
and their interaction with the immune synapse are still a matter of
debate [13–15, 25, 26].
To better understand T-cell signaling at the immune synapse,
it thus seems important to develop tools that allow the purification
of these different intracellular compartments and to analyze their
content.
We herein describe a protocol of purification of membrane
compartments containing the adaptor molecule LAT. This mole-
cule, which was cloned at the same time as the first characterization
of the immune synapse [27, 28], plays a key role in T-cell activation.
It is present both at the plasma membrane and in vesicles [29]. The
traffic of the vesicular pool of LAT to the immune synapse requires
the SNARE protein VAMP7 [15]. Our protocol is based on a first
step of separation of different intracellular organelles by flotation
Purification of LAT Containing Membranes 357
gradient and a second step of affinity purification of Strep-tagged

fusion proteins (Strep-tag® is a nine-amino acid peptide with
intrinsic streptavidin-binding activity) on resins coated with Strep-
Tactin® (an engineered streptavidin that binds Strep-tag®) [30].
Because the purification is done in the absence of detergent, it not
only allows the purification of molecules that are physically associ-
ated with the tagged-bait, as usually done for interactome studies
[31, 32] but also the purification of molecules that are present in
the same membranes/vesicles being integral or lumenal proteins of
the compartments (see Fig. 1 of Chapter 19). We think that this
protocol, presented here with the example of LAT, can be used to
purify membrane compartments containing other tagged trans-
membrane proteins.
2 Materials
Prepare all solutions using ultrapure water. Solutions are stored at

room temperature unless stated otherwise.
2.1 Cell Disruption 1. Cells: LAT-deficient JCAM2.5 Jurkat T cells [33] expressing
and Gradient the mouse LAT-Strep-tag® protein [31].
Preparation 2. Homogenization buffer: 0.25 M sucrose, 10 mM Tris–HCl
pH 7.4, 1 mM EDTA.
3. Iodixanol dilution buffer: 0.25 M sucrose, 60 mM Tris–HCl
pH 7.4, 6 mM EDTA.
4. cOmplete™, EDTA-free protease inhibitor tablets from Roche
Life Science, dissolve one tablet in 2 mL of water according to
the manufacturer’s instructions to obtain a 25× concentrated
stock solution (store at −20 °C).
5. Halt™ Phosphatase Inhibitor cocktail from Thermo Fisher
Scientific. Solution is diluted 100× in appropriate solution
(store at 4 °C).
6. OptiPrep™ (Axis-Shield) density gradient medium is a solu-
tion of 60% iodixanol in water with a density of 1.32 g/mL.
8. RPMI-1640.
9. 2 mL glass Dounce homogenizer.
10. Needles 25 G × 5/8″ and 2 mL syringes.
11. SW 55 Ti Rotor, Swinging Bucket (Beckman-Coulter).
12. Ultra-clear centrifuge tubes 5 mL (Beckman-Coulter).
13. RIPA lysis buffer: 25 mM Tris–HCl pH 7.4, 1% NP-40, 0.5%
Na-deoxycholate, 0.1% SDS, 150 mM NaCl. Store at 4°C
2.2 Immuno 1. Strep-Tactin® Sepharose® resin (IBA) 50% suspension (50%

precipitation suspension in 100 mM Tris–HCl pH 8.0, 1 mM EDTA, 150
and Elution mM NaCl) (Fig. 1b). Store at 4°C
2. Sepharose® 50% suspension (IBA) (Fig. 1c). Store at 4°C
3. IBA washing buffer: 100 mM Tris–HCl pH 8.0, 150 mM
NaCl, 1 mM EDTA.
4. Biotin elution buffer: IBA washing buffer supplemented with
2 mM of d-biotin (IBA). Store at 4 °C.
2.3 Western Blot 1. SDS-PAGE running buffer: 25 mM Tris, 0.192 M glycine,

0.1% SDS.
2. Western blot transfer buffer: 25 mM Tris, 0.192 M glycine,
20% methanol.
3. Tris-buffered saline (140 mM NaCl, 25 mM Tris, 3 mM KCl)
containing 0.05% Tween-20 (TBST).
4. Blocking solution: 5% BSA in TBST. Filter on 0.45 μm and
store at 4 °C.
5. BIORAD Mini-PROTEAN® TGX Stain-free™ Precast gels
4%–15%.
6. 4× Laemmli sample buffer (277.8 mM Tris–HCl pH 6.8,
44.4% (v/v) glycerol, 0.02% bromophenol blue, 5% SDS).
7. Reducing sample buffer 10×: 500 mM dithiothreitol.
8. PVDF transfer membrane (Immuno-Blot®, BIORAD).
9. ECL blotting substrate (Pierce).
2.4 Antibodies See Table 1.
3 Methods
3.1 Cell Stimulation It is very important to avoid any difference of temperature. Place the
centrifuge at room temperature and pre-warm medium at 37 °C.
1. Harvest and count the cells expressing the transmembrane
Strep-tag® protein of interest (see Note 1).
2. Wash the cells twice in RPMI medium to remove the FCS con-
tained in culture medium: fill a 50 mL tube with RPMI pre-
warmed at 37 °C and spin for 5 min at 300 × g. Remove the
supernatant, resuspend the pellet, and spin one more time.
3. For the stimulation, resuspend the cells in RPMI at 100 × 106
mL−1 and then transfer 1 mL of this cell suspension in an
Eppendorf tube (2 mL). Do as many tubes as needed. Leave the
cells at 37 °C for 5 min (in a water bath) without any stimula-
tion. Meanwhile, prepare the appropriate dilutions of anti-CD3
and anti-CD28 antibodies (in RPMI). Anti-CD3 Ab is used at
12.5 μg/mL and anti-CD28 Ab at 25 μg/mL (see Note 2).
A LAT
B
Strep-Tag
StrepTactin sepharose
LAT interacting
protein
C
Non interacting proteins
Fraction to be purified Membrane vesicle “nude” sepharose
D E
Membrane bound to StrepTactin sepharose Unbound membranes
F G
Sepharose
Biotin
Elution Eluted membranes
Fig. 1 Experimental approach followed to purify membranes containing a chimeric LAT protein fused to a
Strep-tag®. (a) Membranes containing the chimeric LAT are recovered by flotation gradient (in the example
presented herein fraction 3 is used, see Fig. 2). (b–d) Specific purification of membranes containing chimeric
LAT is obtained with Sepharose coated with Strep-Tactin. Nonspecific binding is obtained with Sepharose
alone (c: “nude” Sepharose). (e) Membranes that do not contain the chimeric LAT are not retained on
Sepharose-Strep-Tactin (unbound membranes). (f) Membranes containing the chimeric LAT are eluted with an
excess of biotin. (g) Molecules recovered in this fraction contain proteins interacting with LAT, membrane
associated or transmembrane proteins present in the same membranes as LAT, and luminal proteins from
LAT-bearing vesicles
Table 1
List of antibodies used in the protocol
Catalog Final concentration

Company Antigen recognized number Clone Host or dilutiona
Western blot antibodies
Santa Cruz CD3ζ SC-1239 6B10.2 Mouse 0.2 μg/mL
In Vitrogen Transferrin receptor 136800 H68.4 Mouse 0.5 μg/mL
Genetex Lamp2a 63319 Rabbit 1/1000
E-Biosciences CD45 14-0459-82 H130 Mouse 0.5 μg/mL
ABCAM GM130 Ab52649 EP892Y Rabbit 0.11 μg/mL
ABCAM TGN46 Ab16052 Rabbit 0.5 μg/mL
ThermoFisher Scientific Mitofilin PA5-30419 Rabbit 2 μg/mL
Upstate LAT 06-807 Rabbit 1 μg/mL
Technology-Millipore
Enzo GP96 ADI-SPA-850 9G-10 Rat 1 μg/mL
T.Galli’s lab VAMP7 Rabbit 1/200
Cell Signaling P-PLCγ1(Tyr783) CS#2821 Rabbit 1/1000
BD Biosciences P-CD3ζ (Tyr142) BD #558402 K25- Mouse 0.5 μg/mL
407.69
Upstate P-LAT (Tyr191) 07-278 Rabbit 1 μg/mL
Technology-Millipore
Stimulating antibodies
Biolegend CD28 302923 CD28.2 Mouse 12.5 μg/mL
E-Biosciences CD3ε 16-0037-85 OKT3 Mouse 25 μg/mL
HRP-conjugated antibodies
Jackson HRP-anti-mouse Ig 115-035-146 Goat 0.08 μg/mL
Immunoresearch
Jackson HRP-anti-rat Ig 112-035-143 Goat 0.08 μg/mL
Immunoresearch
Jackson HRP-anti-rabbit Ig 111-035-144 Goat 0.08 μg/mL
Immunoresearch
Dilution is given when antibody concentration is not known
a
4. Add antibodies to the cells directly in Eppendorf tubes and

mix by gently pipetting up and down without vortexing the
cells. Incubate at 37 °C (in water bath) for the requested
period of time (see Note 20 ).
5. To stop the activation, add 1 mL of ice-cold PBS to the cells
and spin immediately for 5 min at 250 × g at 4 °C.
3.2 Cell Disruption It is important to keep samples on ice as much as possible. To avoid
any contamination between samples, rinse extensively the Dounce
homogenizer with sterile water and do the last rinse with homog-
enization buffer.
1. Resuspend the cells pellet obtained in Subheading 3.1, step 5
in 1.5 mL of ice-cold homogenization buffer supplemented
with both protease and phosphatase inhibitors (see Note 3).
2. At this step, take an aliquot of the cell suspension (100 μL),
and resuspend in RIPA lysis buffer containing protease and
phosphatase inhibitors. This is your total lysate control (see
Note 4).
3. Transfer the cell suspension into the Dounce homogenizer and
apply 25 strokes with the pestle to induce cell breakage.
4. Transfer the suspension into a new Eppendorf tube (2 mL).
5. Homogenize by 15 passages through a 25 GA needle fitted
onto a 2 mL syringe (see Note 5).
6. Centrifuge 3 min at 900 × g at 4 °C; discard the pellet contain-
ing nuclei and unbroken cells and keep the supernatant which
contains cell membranes.
3.3 Membrane 1. Transfer the supernatant in a 5 mL ultracentrifuge tube and

Centrifugation centrifuge at 65,000 × g for 1 h in a SW55Ti rotor (see Note 6).
2. Transfer the supernatant in a new Eppendorf tube; it contains
the cytosolic proteins.
3. Resuspend the pellet in 1.2 mL of homogenization buffer sup-
plemented with both protease and phosphatase inhibitors.
4. Homogenize by five passages through a 25 GA needle fitted
onto a 2 mL syringe.
3.4 Flotation 1. For each gradient, prepare immediately before use 1.3 mL of a
Gradient 20% iodixanol solution (mix 1 vol of Optiprep™ with 2 vol-
umes of iodixanol dilution buffer) and 1.2 mL of a 10% iodixa-
nol solution (mix 1 volume of Optiprep™ with 5 volumes of
iodixanol dilution buffer) (see Note 7).
2. Mix the volume of membranes suspension obtained in
Subheading 3.3, step 4 with the same volume of Optiprep™ to
reach a final concentration of 30% iodixanol (dilution 1:2 from
the 60% original iodixanol solution). Place this sample at the
bottom of an ultracentrifuge tube.
3. Slowly overlay on the top of the 30% iodixanol solution 1.3
mL of 20% iodixanol solution and then overlay 1.2 mL of the
10% iodixanol solution. It is important to avoid mixing the
gradient layers (see Note 8).
4. Centrifuge the gradient at 350,000 × g for 3 h at 4 °C (SW55Ti

rotor) with low acceleration (4) and no brake (see Note 9).
5. After centrifugation, collect carefully ten fractions of 490 μL
from the top using a smooth 1 mL pipet (see Note 10).
3.5 Western Blot 1. For each fraction transfer 49 μL in a new Eppendorf tube and
Analysis mix with 4.9 μL of 10× reducing sample and 17 μL of 4×
of the Different Laemmli buffer.
Fractions 2. Heat the samples at 95 °C for 5 min. Make sure to briefly spin
if condensation was formed in the tube.
3. Load 15 μL of each sample on a 4–15% SDS-PAGE gel and
proceed with electrophoresis till the dye front has reached the
bottom of the gel (see Note 11).
4. Transfer on PVDF membrane.
5. Incubate the membrane with TBST-5% BSA for 1 h at RT
under gentle agitation.
6. Incubate the membrane with primary antibodies OV/N at 4
°C under gentle agitation. For references see Subheading 2.4.
7. Wash three times for 15 min with TBST.
8. Incubate with the appropriate HRP-conjugated secondary
antibodies for 1 h under gentle agitation.
9. Wash four times for 15 min with TBST.
10. Proceed with chemiluminescence detection of proteins (see
Fig. 2a).
3.6 Precipitation 1. Vortex carefully the Strep-Tactin® Sepharose® IBA resin (here-
and Elution after called resin) ( see Note 19 and Note 20 ). Transfer the
of the Membranes required quantity in an Eppendorf tube (see Notes 12 and 13).
Containing the Strep- 2. Wash the resin three times with IBA washing buffer: add 1 mL
tag® Protein of IBA washing buffer to the resin. Mix by inverting the tube
several times. Centrifuge at 1000 × g for 1 min. Remove the
supernatant. Repeat the centrifugation and washing steps two
more times (see Note 14).
3. Resuspend the resin in IBA washing buffer supplemented with
protease and phosphatase inhibitors. The volume of resuspension
is equal to the Strep-tag® protein sample volume (see Note 15).
4. Add the sample containing membranes with Strep-tag® protein
(prepared in Subheading 3.4) to the resin and mix by inverting
the tube (see Note 16).
5. Incubate for 90 min at 4 °C on a rotating wheel.
6. Centrifuge at 1000 × g for 1 min. Carefully transfer the super-
natant to a new Eppendorf tube. This corresponds to the
“unbound membranes” (see Fig. 1e).
A
Fraction n° 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Mitofilin TfR
LAMP2 CD45
GM130 Vamp7
TGN46 LAT
gp96 CD3ζ
100nm
Fig. 2 (a) Western blot analysis of the different fractions collected after flotation gradient. The presence of the
different intracellular organelles is followed using antibodies for specific markers: mitofilin for mitochondria;
LAMP2 for lysosomes; GM130 for cis-Golgi; TGN46 for trans-Golgi; gp96 for endoplasmic reticulum; VAMP7
for vesicular compartment and CD45, TfR, and CD3ζ; and LAT for plasma membrane and endocytic compart-
ments. (b) Electron microscopy images showing immunogold labeling for LAT on the vesicles present in frac-
tion 3. The size of the vesicles is between 50 and 300 nm
7. Wash the resin/membranes pellet five times as in Subheading

3.6, step 2, but add phosphatase and protease inhibitors to the
IBA washing buffer.
8. After the last wash, carefully remove the supernatant and add
120 μL of elution buffer containing 2 mM d-biotin to the
pellet containing membranes bound Strep-tag® protein cou-
pled to the resin (Fig. 1f). Incubate for 25 minutes on a
rotating wheel at 4°C..
9. Centrifuge at 1000 × g for 1 min. Remove the supernatant: it
corresponds to eluted membranes containing the Strep-tag®
proteins (Fig. 1g) (see Note 17).
10. For each elution sample, transfer 20 μL in an Eppendorf tube
and mix with 2 μL of 10× reducing sample buffer and 10 μL
of 4× Laemmli buffer. Proceed to immunoblotting as in
Subheading 3.5.
11. Following the last elution step, resuspend the resin in 500 μL of
IBA washing buffer and transfer 50 μL in an Eppendorf tube;
mix with 5 μL of reducing sample buffer and 18 μL of 4×
Laemmli buffer. This will be the control of the efficacy of elution
(material staying on Strep-Tactin® resin after elution). Proceed
to immunoblotting as in Subheading 3.5 (see Note 18).
Unstimulated +anti-CD3/CD28
Fraction n° 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
P-PLCγ1
P-CD3ζ
P-LAT
Total LAT
Fig. 3 Western blot analysis of the phosphorylated signaling molecules in the different fractions. JCAM2.5 T
cells expressing a chimeric LAT fused to Strep-Tactin® were either left unstimulated or stimulated for 15 min
with anti-CD3 and anti-CD28 antibodies, and the ten fractions obtained after flotation gradient were immunob-
lotted and probed with either the phospho-specific antibodies anti-phospho-PLCγ1 (P-PLCγ1), anti-phospho-
CD3ζ (P-CD3ζ), and anti-phospho-LAT (P-LAT) or with a total anti-LAT antibody
4 Notes
1. This protocol was optimized to purify membranes containing

LAT from JCAM2.5 cells (LAT-deficient Jurkat cells) expressing
the mouse chimeric LAT coupled to a Strep-tag® [30, 31, 33].
2. For each gradient, use between 100 and 400 × 106 cells. If
more than 500 × 106 cells are used, the mechanical lysis will be
less efficient and will reduce both the gradient separation and
the protein yield.
3. Use 150–200 × 106 cells/1.5 mL. Increase the volume of
homogenization buffer in proportion.
4. The total lysate can be immunoblotted for specific phospho-
proteins in order to ensure that the stimulation was successful.
A proper activation of the cells can also be detected in the dif-
ferent fractions following the flotation gradient (see Fig. 3).
5. These two sequential steps of homogenization should break
between 60% and 80% of the cells without damaging nuclei.
Cells breakage can be checked by Trypan blue staining under a
microscope.
6.
Ultracentrifuge settings: acceleration maximum, brake
maximum.
7. Carefully mix the content of the Optiprep™ bottle before use.
Keep the Optiprep™ solution at room temperature, sterile and
protected from light.
8. Carefully equilibrate the tubes between layers to ensure that
the centrifuge will be balanced during centrifugation.
9. Without brake, it will take about 1 h for the rotor to stop.
Unstimulated +anti-CD3/CD28
Elution n° 1 2 3 1 2 3 1 2 3 1 2 3
LAT
CD3ζ
Sepharose: “nude” StrepTactin “nude” StrepTactin
Fig. 4 Western blot analysis of eluted proteins. Fraction 3 obtained from unstimu-
lated or anti-CD3/CD28-stimulated JCAM2.5 T cells expressing a chimeric LAT
fused to Strep-Tactin® (see Fig. 3) was subjected to purification with either
“nude” Sepharose (nonspecific binding) or Strep-Tactin® Sepharose (specific
binding). Membranes were then eluted with d-biotin three times (elution n° 1, 2,
3) and probed with anti-LAT and anti-CD3ζ antibodies. Two bands are revealed
with the anti-CD3ζ, the upper band corresponding to the phosphorylated form of
the protein
10. The density of each fraction can be measured using a refrac-

tometer (Carl Zeiss). After centrifugation, the gradient results
in a linear increase in density and can be assessed to ensure the
reproducibility of the experiments.
11. Each lane contains 1/40 of total fraction proteins (see Fig. 2a).
12. The Sepharose suspension has a tendency to stick to plastic
pipettes and tips. Moreover 200 μL tips are too tight to let the
resin go through, so cut off the end of the tip to ensure repro-
ducible pipetting.
13. Resin binding capacity (from IBA data sheet): 1 mL of sedi-
mented resin (corresponding to 2 mL of a 50% suspension) is
able to purify 50–100 nmol recombinant Strep-tag® protein
(up to 3 mg in the case of a 30 kDa protein).
14. You can make a mark on the Eppendorf tube to visualize the
top of the pellet formed by the resin. Indeed, the resin is whit-
ish and the pellet can be difficult to see.
15. In our experiments, the volume of fraction sample used for
purification is 400 μL, so we resuspend the resin in 400 μL of
IBA washing buffer supplemented with proteases and phos-
phatases inhibitors.
16. In the example described herein, fraction 3 was precipitated
because it contains LAT and the vesicular SNARE protein
VAMP7 that was shown to control LAT trafficking [15] (see
Fig. 2a) and thus constitute an enriched pool of the LAT-
containing vesicles transported to the immune synapse.
17. Elution steps can be repeated to improve the yield of protein
bait recovery (see Fig. 4). Each application will require
preliminary settings to determine recovery yields of precipita-
tion and of elution.
18. To control for the efficacy of the precipitation and elution

steps, run side by side on a gel samples from fraction 3 (before
purification, Subheading 3.5, step 1), unbound membranes
(Subheading 3.6, step 6), elution (membranes containing the
Strep-tag® proteins, Subheading 3.6, step 10), and resin after
elution (non-eluted proteins, Subheading 3.6, step 11). Use
an antibody corresponding to the Strep-tag® protein and quan-
tify its amount in each sample. The comparison of the amount
of Strep-tag® protein found in fraction 3 and in “unbound
membranes” will give the efficiency of purification, whereas
the comparison between the eluted material and material still
bound to resin after elution will give the efficiency of elution.
19. As a control for the specificity of the purification, you can run the
experiment in the exact same conditions but using a “nude”
Sepharose, which does not contain the Strep-Tactin® (see Fig. 1c).
Proteins that will bind in a nonspecific manner on the resin will
constitute the background of the experiment.
20. Also, negative control can come from nonactivated conditions
or using cells that do not express the Strep-tag® protein.
Acknowledgments
We would like to thank R. Roncagalli for providing the LAT-Strep-

tag® construct and for helpful discussion. We are grateful to
S. Dogniaux, J.M. Carpier, M. Saitakis, and A. Zucchetti for their
kind support.This work was supported by ANR-10-IDEX-0001-02
PSL*, ANR-11-LABX-0043, ANR-13- BSV2-0018 “Neuro
ImmunoSynapse”), and Fondation pour la Recherche Médicale
(FRM, FRM DEQ20140329513).
References
1. Monks CR, Freiberg BA, Kupfer H, Sciaky N, 4. Campi G, Varma R, Dustin ML (2005) Actin
Kupfer A (1998) Three-dimensional segrega- and agonist MHC-peptide complex-dependent
tion of supramolecular activation clusters in T T cell receptor microclusters as scaffolds for
cells. Nature 395(6697):82–86 signaling. J Exp Med 202(8):1031–1036
2. Grakoui A, Bromley SK, Sumen C, Davis MM, 5. Kupfer A, Mosmann TR, Kupfer H (1991)
Shaw AS, Allen PM, Dustin ML (1999) The Polarized expression of cytokines in cell conju-
immunological synapse: a molecular machine gates of helper T cells and splenic B cells. Proc
controlling T cell activation [see comments]. Natl Acad Sci U S A 88(3):775–779
Science 285(5425):221–227 6. Kupfer A, Singer SJ, Dennert G (1986) On the
3. Yokosuka T, Sakata-Sogawa K, Kobayashi W, mechanism of unidirectional killing in mixtures
Hiroshima M, Hashimoto-Tane A, Tokunaga of two cytotoxic T lymphocytes. Unidirectional
M, Dustin ML, Saito T (2005) Newly gener- polarization of cytoplasmic organelles and the
ated T cell receptor microclusters initiate and membrane-associated cytoskeleton in the effec-
sustain T cell activation by recruitment of tor cell. J Exp Med 163(3):489–498
Zap70 and SLP-76. Nat Immunol 6(12): 7. Chemin K, Bohineust A, Dogniaux S, Tourret
1253–1262 M, Guegan S, Miro F, Hivroz C (2012)
Cytokine secretion by CD4+ T cells at the pathways from intracellular compartments. Nat
immunological synapse requires Cdc42- Immunol 12(11):1119–1126
dependent local actin remodeling but not 20. Mor A, Campi G, Du G, Zheng Y, Foster DA,
microtubule organizing center polarity. Dustin ML, Philips MR (2007) The lympho-
J Immunol 189(5):2159–2168 cyte function-associated antigen-1 receptor
8. Griffiths GM, Tsun A, Stinchcombe JCThe costimulates plasma membrane Ras via phos-
immunological synapse: a focal point for endo- pholipase D2. Nat Cell Biol 9(6):713–719
cytosis and exocytosis. J Cell Biol 189(3): 21. Daniels MA, Teixeiro E, Gill J, Hausmann B,
399–406 Roubaty D, Holmberg K, Werlen G, Hollander
9. Rossy J, Pageon SV, Davis DM, Gaus K (2013) GA, Gascoigne NR, Palmer E (2006) Thymic
Super-resolution microscopy of the immuno- selection threshold defined by compartmental-
logical synapse. Curr Opin Immunol 25(3): ization of Ras/MAPK signalling. Nature
307–312 444(7120):724–729
10. Purbhoo MA (2013) The function of sub- 22. Perez de Castro I, Bivona TG, Philips MR,
synaptic vesicles during T-cell activation. Pellicer A (2004) Ras activation in Jurkat T
Immunol Rev 251(1):36–48 cells following low-grade stimulation of the
11. Onnis A, Finetti F, Baldari CT (2016) Vesicular T-cell receptor is specific to N-Ras and occurs
trafficking to the immune synapse: how to only on the Golgi apparatus. Mol Cell Biol
assemble receptor-tailored pathways from a 24(8):3485–3496
basic building set. Front Immunol 7:50 23. Choudhuri K, Llodra J, Roth EW, Tsai J,
12. Purbhoo MA, Liu H, Oddos S, Owen DM, Gordo S, Wucherpfennig KW, Kam LC, Stokes
Neil MA, Pageon SV, French PM, Rudd CE, DL, Dustin ML (2014) Polarized release of
Davis DM (2010) Dynamics of subsynaptic T-cell-receptor-enriched microvesicles at the
vesicles and surface microclusters at the immu- immunological synapse. Nature 507(7490):
nological synapse. Sci Signal 3(121):ra36 118–123
13. Williamson DJ, Owen DM, Rossy J, Magenau 2 4. Finetti F, Onnis A, Baldari CT (2015)
A, Wehrmann M, Gooding JJ, Gaus K (2011) Regulation of vesicular traffic at the T cell
Pre-existing clusters of the adaptor Lat do not immune synapse: lessons from the primary cil-
participate in early T cell signaling events. Nat ium. Traffic 16(3):241–249
Immunol 12(7):655–662 25. Lillemeier BF, Mortelmaier MA, Forstner MB,
14. Soares H, Henriques R, Sachse M, Ventimiglia Huppa JB, Groves JT, Davis MM (2010) TCR
L, Alonso MA, Zimmer C, Thoulouze MI, and Lat are expressed on separate protein
Alcover A (2013) Regulated vesicle fusion gen- islands on T cell membranes and concatenate
erates signaling nanoterritories that control T during activation. Nat Immunol 11(1):90–96
cell activation at the immunological synapse. 26. Balagopalan L, Barr VA, Kortum RL, Park AK,
J Exp Med 210(11):2415–2433 Samelson LE (2013) Cutting edge: cell surface
15. Larghi P, Williamson DJ, Carpier JM, Dogniaux linker for activation of T cells is recruited to
S, Chemin K, Bohineust A, Danglot L, Gaus K, microclusters and is active in signaling.
Galli T, Hivroz C (2013) VAMP7 controls T J Immunol 190(8):3849–3853
cell activation by regulating the recruitment 27. Weber JR, Orstavik S, Torgersen KM, Danbolt
and phosphorylation of vesicular Lat at TCR- NC, Berg SF, Ryan JC, Tasken K, Imboden JB,
activation sites. Nat Immunol 14(7):723–731 Vaage JT (1998) Molecular cloning of the
16. Vieira AV, Lamaze C, Schmid SL (1996) cDNA encoding pp36, a tyrosine- phos
Control of EGF receptor signaling by clathrin- phorylated adaptor protein selectively expressed
mediated endocytosis. Science 274(5295): by T cells and natural killer cells. J Exp Med
2086–2089 187(7):1157–1161
17. Di Guglielmo GM, Le Roy C, Goodfellow AF, 28. Zhang W, Sloan-Lancaster J, Kitchen J, Trible
Wrana JL (2003) Distinct endocytic pathways RP, Samelson LE (1998) LAT: the ZAP-70
regulate TGF-beta receptor signalling and tyrosine kinase substrate that links T cell recep-
turnover. Nat Cell Biol 5(5):410–421 tor to cellular activation. Cell 92(1):83–92
18. McGettrick AF, O’Neill LA (2010) Localisation 2 9. Bonello G, Blanchard N, Montoya MC, Aguado
and trafficking of Toll-like receptors: an impor- E, Langlet C, He HT, Nunez-Cruz S, Malissen
tant mode of regulation. Curr Opin Immunol M, Sanchez-Madrid F, Olive D, Hivroz C,
22(1):20–27 Collette Y (2004) Dynamic recruitment of the
19. Chaturvedi A, Martz R, Dorward D, Waisberg adaptor protein LAT: LAT exists in two distinct
M, Pierce SK (2011) Endocytosed BCRs intracellular pools and controls its own recruit-
sequentially regulate MAPK and Akt signaling ment. J Cell Sci 117(Pt 7):1009–1016
30. Junttila MR, Saarinen S, Schmidt T, Kast J, Lat adaptor-independent TCR signaling hub.
Westermarck J (2005) Single-step Strep-tag Nat Immunol 15(4):384–392
purification for the isolation and identification 32. Paster W, Bruger AM, Katsch K, Gregoire C,
of protein complexes from mammalian cells. Roncagalli R, Fu G, Gascoigne NR, Nika K,
Proteomics 5(5):1199–1203 Cohnen A, Feller SM, Simister PC, Molder
31. Roncagalli R, Hauri S, Fiore F, Liang Y, Chen KC, Cordoba SP, Dushek O, Malissen B, Acuto
Z, Sansoni A, Kanduri K, Joly R, Malzac A, O (2015) A THEMIS:SHP1 complex pro-
Lahdesmaki H, Lahesmaa R, Yamasaki S, Saito motes T-cell survival. EMBO J 34(3):393–409
T, Malissen M, Aebersold R, Gstaiger M, 33. Finco TS, Kadlecek T, Zhang W, Samelson LE,
Malissen B (2014) Quantitative proteomics Weiss A (1998) LAT is required for TCR-
analysis of signalosome dynamics in primary T mediated activation of PLCgamma1 and the
cells identifies the surface receptor CD6 as a Ras pathway. Immunity 9(5):617–626
Chapter 22
Quantitative Phosphoproteomic Analysis of T-Cell Receptor

Signaling
Nagib Ahsan and Arthur R. Salomon
Abstract
TCR signaling critically depends on protein phosphorylation across many proteins. Localization of each
phosphorylation event relative to the T-cell receptor (TCR) and canonical T-cell signaling proteins will
provide clues about the structure of TCR signaling networks. Quantitative phosphoproteomic analysis by
mass spectrometry provides a wide-scale view of cellular phosphorylation networks. However, analysis of
phosphorylation by mass spectrometry is still challenging due to the relative low abundance of phosphory-
lated proteins relative to all proteins and the extraordinary diversity of phosphorylation sites across the
proteome. Highly selective enrichment of phosphorylated peptides is essential to provide the most com-
prehensive view of the phosphoproteome. Optimization of phosphopeptide enrichment methods coupled
with highly sensitive mass spectrometry workflows significantly improves the sequencing depth of the
phosphoproteome to over 10,000 unique phosphorylation sites from complex cell lysates. Here we
describe a step-by-step method for phosphoproteomic analysis that has achieved widespread success for
identification of serine, threonine, and tyrosine phosphorylation. Reproducible quantification of relative
phosphopeptide abundance is provided by intensity-based label-free quantitation. An ideal set of mass
spectrometry analysis parameters is also provided that optimize the yield of identified sites. We also provide
guidelines for the bioinformatic analysis of this type of data to assess the quality of the data and to comply
with proteomic data reporting requirements.
Key words Immunoaffinity purification, Label-free quantitation, Phosphoproteomics, T-Cell signaling,

Tyrosine phosphorylation, Mass spectrometry
1 Introduction
Among the many posttranslational modifications (PTMs) that

occur in cells, protein phosphorylation is a critically important
modification found in T-cell receptor (TCR) signaling.
Determination that a protein is phosphorylated is not enough to
understand the precise role that the phosphorylation plays in cellular
signaling. Therefore, identification and quantitation of the exact
position of the phosphorylation site are essential. A single protein
can be targeted by multiple kinases and thus phosphorylated in mul-
tiple sites with different dynamics [1]. In the past, identification of
369
370 Nagib Ahsan and Arthur R. Salomon
multiple phosphorylation sites on target proteins was a very labori-

ous task involving in vivo or in vitro labeling with radioactive phos-
phate, separation of labeled peptides by column or thin-layer
chromatography, and Edman sequencing. However, the recent
and rapid advancement of highly sensitive and accessible mass spec-
trometry-based techniques makes this an invaluable addition to the
toolbox. Analysis of intact purified proteins (top-down proteomics)
or complex mixtures of phosphopeptides from a tryptic digest of
cellular lysates (bottom-up proteomics) can provide direct infor-
mation about the position and relative abundance of phosphoryla-
tion sites within a protein sequence. In particular, the ability of
bottom-up proteomics to provide identification and quantification
of tens of thousands of phosphopeptide sequences from a cell lysate
at high confidence from a single 180 min LC/MS analysis enables
new approaches for the exploration of signaling pathways. Analysis
of phosphopeptides by mass spectrometry is substantially more dif-
ficult compared to detection of unphosphorylated peptides due to
a number of physiochemical properties of the modified peptides.
For example, small phosphopeptides are very hydrophilic and may
not be retained well on reversed-phase matrices typically used in
LC-MS [2]. Furthermore, some phosphorylation sites are often
extremely difficult to detect due to the low abundance of the phos-
phopeptide and low stoichiometry of phosphorylation in a cellular
context of abundant background proteins. Lastly, the presence of
the phosphorylation site on a peptide decreases its ionization effi-
ciency when detected in positive ion mode, making detection of
phosphopeptides even more difficult. The phosphopeptides must
be enriched to achieve optimal sequencing depths of the
phosphoproteome.
Many methods have been developed to enrich the total phos-
phoproteome from the rest of the proteome including Fe(III) and
Ga(III) immobilized metal affinity chromatography (IMAC) [3]
or metal oxide affinity chromatography (MOAC) based on enrich-
ment with TiO2 beads [4] or titanium (IV)-functionalized soluble
nanopolymer (PolyMAC-Ti) [5]. Sequencing depths in excess of
10,000 unique phosphorylation sites can be expected from com-
plex cell lysates using MOAC. Of critical importance in maximiza-
tion of sequencing depth and quantitative reproducibility is the use
of sub 2 μm reversed-phase beads in the LC/MS acquisition of
data. The use of these particles is accompanied by higher backpres-
sures and typically requires ultrahigh-performance liquid chroma-
tography (UHPLC) to achieve sufficient flow at the electrospray
tip. Here we describe an effective protocol for TiO2 enrichment of
phosphopeptides that has achieved widespread success.
Reversible tyrosine phosphorylation which has been estimated
to represent less than 1% of all human phosphorylation events [6,
7] plays a key role in many aspects of regulating several essential
molecular mechanisms and processes including gene transcription,
Quantitative Phosphoproteomic Analysis of TCR Signaling 371
cell growth, cell cycle, differentiation, and motility in mammalian

cells [8]. Therefore, identification of tyrosine-phosphorylated resi-
dues and quantification of the relative phosphorylation level are
critical for understanding their contribution to signaling networks
and, consequently, to pathological processes [9]. Although a num-
ber of methods have been adopted for enrichment and analysis of
tyrosine phosphorylation [9–15], quantitative proteomic analysis
of tyrosine phosphorylation by mass spectrometry is still challeng-
ing, due to the low occurrence of this posttranslational modifica-
tion compared to serine and threonine phosphorylation in
mammalian cells [11].
Immunoaffinity purification using a pan-specific phosphotyro-
sine antibody is commonly used to enrich tyrosine-phosphorylated
peptides. Recently a comparison between the two most commonly
used antibodies showed that the P-Tyr-100 anti-phosphotyrosine
antibody performs superiorly when compared to 4G10 antibody
for label-free phosphotyrosine-based phosphoproteomics [15].
This study also indicated that optimization of phosphotyrosine
peptide capture protocol coupled with mass spectrometry methods
can potentially enhance the identification of phosphotyrosine-
containing peptides. Here we provide an optimized protocol that
can be used to identify and quantify phosphotyrosine-containing
peptides from human T cells. The optimized protocol is able to
detect and quantify 934 unique tyrosine phosphorylation sites
from Jurkat T cells using the Orbitrap Velos mass spectrometer
[16] and 1557 unique tyrosine phosphorylation sites on the Q
Exactive mass spectrometer (manuscript in preparation).
The measurement of statistically significant quantitative
changes in the phosphoproteome from cells lacking T-cell signal-
ing proteins provides critical information on the prospective role of
each phosphorylation site in T-cell signaling [16–20]. Bottom-up
proteomics is a powerful approach for quantitation of relative
changes in peptide and protein abundance across different cellular
states or treatments. Both label-free and label-based strategies can
be employed to quantitate relative peptide abundance [19].
Currently, the two main strategies for label-free quantitation are
spectral counting and signal intensity of detected peptides [19]. In
the spectral counting approach, the rate at which a peptide precur-
sor ion is selected for fragmentation in a mass spectrometer is cor-
related to its abundance. Spectral counts can then be averaged into
a protein abundance index. This approach is not appropriate for
phosphoproteomic analysis due to the necessary phosphopeptide
enrichment steps which would skew the spectral counts according
to the number of phosphorylation sites on each protein which may
not correlate to the protein’s abundance. In the signal intensity
quantitation approach, the selected ion chromatogram for each
peptide is calculated from an LC-MS/MS run, and the peak areas
are integrated over the chromatographic time scale. Retention
time alignment of replicate analyses and accurate mass greatly facil-

itates the comprehensive quantitation and statistical analysis of
each phosphopeptide.
Label-based approaches rely on the assumption that an isotope-
labeled peptide is chemically identical to its native counterpart and
thus behaves identically during chromatographic and MS analysis.
In stable isotope labeling by amino acids in cell culture (SILAC),
stable isotopes of amino acids are metabolically incorporated into
the proteomes of cells, which enables mixing of the experimental
groups being compared and, consequently, minimization of errors
in quantitation that can occur through sample handling [21]. MS/
MS-based quantitation can also be performed by various isotope-
labeling techniques, such as isobaric tags for relative and absolute
quantitation (iTRAQ) [22] and tandem mass tags (TMT) [23].
In the protocols described here, we focus on utilization of
intensity-based label-free quantitation because it combines a high
degree of quantitative reproducibility and a large dynamic range
with less limitation on the number of comparisons between cellular
states and requires the least amount of method development due
to the lack of necessity to optimize the labeling parameters.
2 Materials
For preparation of solutions and buffers, LC-MS grade water is

used. Solutions and buffers are freshly prepared at room tempera-
ture and used immediately. All sample preparation procedures
involving intact proteins are performed at 4 °C unless otherwise
specified.
2.1 Reagents 1. Protein sample: 1 × 108 cells or 10 mg protein.

and Solutions for Cell 2. Cell lysis buffer: 9 M urea, 1 mM Na orthovanadate, 2.5 mM
Lysis and in Solution sodium pyrophosphate, 1 mM β-glycerophosphate in 20 mM
Trypsin Digestion HEPES, pH 8.0 (see Note 1).
3. 45 mM dithiothreitol and 100 mM iodoacetamide.
4. Trypsin, sequencing grade, modified, TPCK treated, affinity
purified (Cat, V5113, Promega, NJ, USA).
2.2 Reagents 1. Solvent A containing 0.1% of trifluoroacetic acid (TFA).

and Solutions for C18 2. Solvent B containing 0.1% of TFA in 40% acetonitrile.
Column (Sep-Pack)
3. Sep-Pack light C18 cartridges (WAT020515, Waters, MA,
Purification
USA) and extraction manifold apparatus (WAT200677, Waters).
4. 10 mL disposable syringe.
2.3 Reagents 1. TiO2 buffer containing 0.1% formic acid in 30% acetonitrile.
and Solutions for TiO2 2. Solution A containing 40% TFA in 50% acetonitrile.
Enrichment
3. Solution B containing 6% TFA, 25% lactic acid in 67.5%

acetonitrile.
4. Buffer A containing 25% solution A in acetonitrile.
5. Buffer B containing 75% of Buffer A and 25% of solution B.
6. Buffer C containing 26.7% ammonium hydroxide.
7. Buffer D containing 26.7% ammonium hydroxide in 40%
acetonitrile.
8. Buffer E containing 50% glacial acetic acid.
9. Titansphere Phos-TiO Kit (GL Sciences, Japan).

10.
A stock of 100 pmol/μL of a custom-synthesized
phosphoserine-containing peptide FQpSEEQQQTEDELQDK
(see Note 2).
2.4 Reagents 1. IAP buffer containing 10 mM sodium phosphate, 50 mM

and Solutions NaCl in 50 mM MOPS, pH 7.2.
for Immune Affinity 2. Solvent A and B (see in Subheading 2.2).
Purification (IAP)
3. Protein G agarose (Roche, USA).
4. Phospho-tyrosine mouse mAb (P-Tyr-100) (Cell Signaling,
MA, USA).
5. A stock of 100 pmol/μL of a custom-synthesized pTyr peptide
LIEDAEpYTAK (see Note 2).
6. A stock of 100 pmol/μL of a custom-synthesized human
angiotensin II peptide DRVpYIHPF (see Note 2).
7. PBS buffer.
2.5 Reagents 1. HPLC Solution A: 0.1 M acetic acid in H2O.

and Solutions 2. HPLC Solution B: 0.1 M acetic acid in acetonitrile (99.8%,
for LC-MS/MS HPLC grade).
3 Methods
3.1 Protein 1. Lyse the cells with 2 mL of ice-cold cell lysis buffer and
Extraction and Trypsin incubate on ice for 20 min (see Note 1).
Digestion 2. Vortex the cell lysates vigorously for 1 min.
of the Samples
3. Sonicate the cell lysates on ice using a microtip sonicator (Q55
Sonicator, Qsonica, USA) with six 5-s bursts (sonication setup
amplification 25%).
4. Centrifuge the cell lysates at 20,000 × g for 15 min at 15 °C.
5. Collect the supernatant for BCA protein quantification and
subsequent trypsin digestion.
6. Proteins are reduced with DTT (final concentration = 4.5

mM) for 30 min at 60 °C.
7. Complete digestion is facilitated by alkylation of cysteine
residues. For protein alkylation, add iodoacetamide (final
concentration = 10 mM) for 20 min at room temperature in
the dark.
8. Dilute cell lysate fivefold with 100 mM ammonium bicarbon-
ate, pH 8.9.
9. Add affinity purified, TPCK-treated trypsin to protein in ratio
1:100 (w/w) trypsin/total protein and incubate overnight at
37 °C.
3.2 Peptide 1. Add TFA solution to each of the trypsin-digested sample to a

Desalting Using C18 final concentration of 1% TFA.
Column 2. Centrifuge the acidified peptide solutions for 5 min at 1800 × g.
3. Transfer the peptide-containing supernatant into a 15 mL con-
ical tube.
4. Set up the C18 Sep-Pak column and the 10 mL syringe to the
extraction manifold reservoir. Equilibrate C18 Sep-Pak col-
umn by washing with 5 mL of 100% acetonitrile and then wash
twice with 3.5 mL of Solvent A (see Subheading 2.2).
5. Load the acidified cleared digested samples to the Sep-Pak col-
umn, and turn on the vacuum slowly (1–2 psi) to pass the solu-
tions through the column.
6. Wash the column three times with 5 mL of Solvent A (see
Subheading 2.2).
7. Elute peptides from column with 10 mL of elution buffer
Solvent B (see Subheading 2.2), and collect the peptide-con-
taining solution into a fresh 15 mL conical tube.
8. Split the eluents (10:90) for phosphopeptide enrichment by
TiO2 (~1 × 107 cell equivalents) and enrichment of phosphoty-
rosine peptide by immune affinity purification (~9 × 107 cell
equivalents).
9. Lyophilize the digested, frozen peptide samples overnight (see
Note 3).
3.3 Phosphopeptide 1. Spin down the trypsin-digested dried peptide sample at 1800
Enrichment × g for 5 min at room temperature, and reconstitute the dried
3.3.1 Global peptide samples in 100 μL TiO2 buffer from the 1 × 107 cell
Phosphopeptide equivalent sample (see Note 4).
Enrichment by TiO2 2. Centrifuge the samples at 12, 000 × g for 5 min at 15 °C, and
collect the clear supernatant into a new Eppendorf tube.
3. Add 500 fmol of synthetic pSer and angiotensin II standard

peptides to every 100 μg of peptide sample (1 × 106 cell equiv-
alents) (see Note 5).
4. Set a TiO2 tips within the centrifugal adaptor (provided by the
vendor) and set it on top of the 2 mL waste fluid tube.
5. Wash tip with 20 μL of Buffer A (see Subheading 2.3) to the
top of the TiO2 phospoTips, and centrifuge at 3000 × g for
2 min at room temperature.
6. Wash tip with 20 μL of Buffer B (see Subheading 2.3) to the
TiO2 phospoTips and centrifuge at 3000 × g for 2 min at room
temperature and discard all the liquid from the waste fluid
tube.
7. Dilute the 110 μL of peptide solution from step 3 with 150
μL of Buffer B and then add to the TiO2 phospoTips (see Note
6) and centrifuge the samples at 1000 × g for 10 min at room
temperature (see Note 6).
8. Wash the TiO2 phospoTips by adding 20 μL of Buffer B fol-
lowed by a centrifugation 3000 × g for 2 min at room
temperature.
9. Wash the TiO2 phospoTips three times by adding 20 μL of
Buffer A followed by a centrifugation 3000 × g for 2 min at
room temperature.
10. Transfer the TiO2 phospoTips with the centrifugal adaptor to
a 1.7 mL recovery collection tube provided in kit, and elute
phosphopeptides with 50 μL of Buffer C (see Subheading 2.3),
and centrifuge at 1000 × g for 5 min at room temperature.
11. Elute phosphopeptides again with 50 μL of Buffer D (see

Subheading 2.3) and centrifuge at 1000 × g for 5 min at room
temperature into the same collection tube.
12. Acidify the eluted phosphopeptides with 5 μL Buffer E (see
Subheading 2.3).
13. Dry the samples in the SpeedVac for 2 h and store the dried
peptides at −80 °C until analysis on the mass spectrometer.
3.3.2 Phosphotyrosine 1. Wash PTMScan P-Tyr-1000 Kit beads twice with 1.0 mL cold
Peptide Enrichment PBS buffer (see section 2.4) (see Note 8). After each wash step,
with IAP (see Note 7) centrifuge at 1500 × g for 2 min at 4 ºC and carefully discard
Antibody Preparation the supernatant.
2. Wash the beads three times with 1.0 mL cold IAP buffer (see
Subheading 2.4). After each wash step, centrifuge at 1500 × g
for 2 min at 4 °C and carefully discard the supernatant.
3. Store the beads on ice for subsequent use.
Elution of Phosphotyrosine 1. Briefly centrifuge the lyophilized peptide at 1500 × g for 5 min
Peptides at room temperature.
2. Reconstitute the 9 × 107 cell equivalent dried peptide samples
(see Subheading 3.2, step 10) in 1 mL of IAP buffer (see
Subheading 2.4) and keep on ice for 5 min (see Note 10).
3. Remove any particulate material by centrifugation 1500 × g
for 5 min at 4 °C.
4. Add 1 pmol of synthetic pTyr peptide standard (see Subheading
2.4) to the peptide solution.
5. Transfer the peptide solution to the bead slurry (see Subheading
“Antibody Coupling”) and incubate for 2 h on a Barnstead/
Thermolyne LABQUAKE rotator (8 rpm) at 4 °C.
6. Centrifuge the mixture at 1500 × g for 2 min at 4 °C and dis-
card the supernatant (see Note 11).
7. Wash the beads three times with 1 mL IAP buffer and remove
the supernatant by centrifugation at 1500 × g for 2 min at 4
°C (see Note 12).
8. Wash the beads with 1 mL of ice-cold water, mix by inverting
tube five times, and remove the supernatant by centrifugation
at 1500 × g for 2 min at 4 °C (see Note 13).
9. Elute the tyrosine-phosphorylated peptides with 55 μL of
0.15% TFA for 10 min at 22 °C followed by collection of the
eluent by centrifugation of the mixture at 1500 × g for 2 min
at 4 °C into a new collection tube (see Note 14).
10. Elute the peptides a second time with 45 μL of 0.15% TFA
and collect the eluent in a different collection tube (as above).
11. Wet a ZipTip with 50 μL of Solvent B (see the recipe in

Subheading 2.2) (see Note 15).
12. Equilibrate the tip with 50 μL Solvent A twice (see the recipe
in Subheading 2.2) (see Note 16).
13. Pipette the first eluted phosphopeptide sample aliquot from
step 9 (55 μL) with micropipette into the ZipTip by r epeatedly
pipetting the solution ten times and then expelling the liquid
into the original tube.
14. Pipette the second phosphopeptide aliquot (45 μL) from step
10 onto the same ZipTip according to procedure in step 13.
15. Wash the tip twice with 50 μL Solvent A (see Note 17).
16. Elute the peptide with 10 μL of Solvent B (see Note 18).
17. Dry the peptides using a SpeedVac for 30 min at 22 °C (see
Note 19).
3.4 LC-MS/MS LC-MS/MS can be performed as described previously [16]. Many

Analysis proteomic core facilities will be able to provide basic capabilities to
collect this type of data according to the parameters specified here.
Tryptic peptides can be analyzed by a fully automated phospho-

proteomic technology platform developed in the Salomon lab
called HTAPP [24] and Peptide Depot [25] or using commercial
software such as Mascot [26], PEAKS 7.5 (http://www.bioinfor.
com) [27], and Scaffold PTM 3.0 (http://www.proteomesoft-
ware.com/products/ptm/). The nanoLC-MS/MS experiments
are performed with an Agilent 1200 Series Quaternary HPLC sys-
tem (Agilent Technologies, Santa Clara, CA) connected to a Q
Exactive Plus mass spectrometer (Thermo Fisher Scientific,
Waltham, MA). For the TiO2 analysis, reconstitute the lyophilized
phosphopeptides from Subheading 3.3.1, step 13 that were
derived from 1 × 107 cell equivalents in 100 μL HPLC Solution A
(see Subheading 2.5), and inject 10 μL for each analysis using
LC-MS/MS. For the phosphotyrosine IAP analysis, reconstitute
the dried peptide from Subheading 3.3.2, step 17 with 10 μL of
HPLC Solution A (see Subheading 2.5), and inject 5 μL for each
analysis. Prior to LC-MS/MS analysis of either TiO2 and phospho-
tyrosine IAP samples, add 500 fmol of human phospho-angiotensin
II peptide as an internal standard.
For LC-MS/MS analysis, the peptides are separated through a
linear reversed-phase gradient from 0% to 40% HPLC Solution B
(see Subheading 2.5) over 60 min with a total 90 min run time. For
optimal sensitivity, the electrospray analytical column can be fabri-
cated with self-pack PicoFrit columns (New Objective Inc., MA,
USA) packed with 3 μm ReproSil-Pur 120 C18 reversed-phase
particles (Dr. Maisch GmbH, Germany). The use of 1.9 μm parti-
cles will improve sequencing depth and quantitative reproducibility
dramatically but may necessitate the use of UHPLC. The electro-
spray ion source is operated at 2.0 kv in a split flow configuration,
as described previously [28]. The Q Exactive Plus is operated in a
data-dependent mode using a top 9 data- dependent method.
Survey full scan MS spectra (m/z 400–1800) are acquired at a res-
olution of 70,000 with an AGC target value of 3 × 106 ions or a
maximum ion injection time of 200 ms. Peptide fragmentation is
performed via higher-energy collision dissociation (HCD) with the
energy set at 28 NCE. The MS/MS spectra are acquired at a reso-
lution of 17,500, with a targeted value of 2 × 104 ions or a maxi-
mum integration time of 200 ms. The underfill ratio, which
specifies the minimum percentage of the target value likely to be
reached at maximum fill time, is defined as 1.0%. The ion selec-
tion abundance threshold is set at 8.0 × 102 with charge state
exclusion of unassigned and z = 1, or 6–8, 8> ions, and dynamic
exclusion time of 30 s.
3.5 Data Analysis Peptide-spectrum matching of MS/MS spectra is performed

against a human-specific database (UniProt, complete proteome
set) using Mascot v. 2.4 (Matrix Science, Ltd., London W1 U
7GB, UK) [26]. A concatenated database containing an equal
number of “target” and reversed “decoy” sequences is employed

to estimate the false discovery rate (FDR) [29]. Msconvert from
ProteoWizard [30], using default parameters and with the
MS2Deisotope filter on, is used to create peak lists for Mascot.
Mascot database searches are performed with the following param-
eters: trypsin enzyme cleavage specificity, two possible missed
cleavages, 7 ppm mass tolerance for precursor ions, and 20 mmu
mass tolerance for fragment ions. Search parameters permit vari-
able modification of methionine oxidation (+15.9949 Da), static
modification of carbamidomethylation (+57.0215 Da) on cysteine,
and variable modification of phosphorylation (+79.9663 Da) on
serine, threonine, and tyrosine residues. The resulting peptide-
spectrum matches (PSMs) are reduced to sets of unique PSMs by
eliminating lower scoring duplicates. Peptide assignments from the
database search are filtered down to 1% false discovery rate (FDR)
by Mascot MOWSE score thresholding. To validate the position of
the phosphorylation sites, the Ascore algorithm [31] can be applied
to all data, and the reported phosphorylation site position should
reflect the top Ascore prediction. The use of other algorithms may
be acceptable, but some sort of probabilistic score should be
reported for each phosphopeptide to estimate the likelihood that
the phosphorylation site position is confidently determined.
3.6 Quantitation Here we provide a detailed description of the quantitative analysis

of Relative parameters used in the Salomon lab HTAPP [24] and Peptide
Phosphopeptide Depot [25] software which may be adapted to commercially avail-
Abundance able software available in many proteomic core facilities. Label-free
relative quantification of phosphopeptide abundance is performed
via calculation of selected ion chromatogram (SIC) peak areas.
Retention time alignment of individual replicate analyses can be
performed as previously described [32]. Peak areas can be calculated
by inspection of SICs using in-house software programmed in R
3.0 based on the Scripps Center for Metabolomics’ XCMS package
[33]. This approach performs multiple passes through XCMS’s
central wavelet transformation algorithm (implemented in the
centWave function) over increasingly narrower ranges of peak
widths and used the following parameters: mass window of 10
ppm, minimum peak widths ranging from 2 to 20 s, maximum
peak width of 80 seconds, signal-to-noise threshold of 10, and
detection of peak limits via descent on the non-transformed data
enabled. For cases when centWave did not identify an MS peak, we
use the getPeaks function available in XCMS to integrate in a pre-
defined region surrounding the maximum intensity signal of the
SIC. SIC peak areas are determined for every phosphopeptide that
is identified by MS/MS. In the case of a missing MS/MS for a
particular peptide, in a particular replicate, the SIC peak area is
calculated according to the peptide’s isolated mass and the reten-
tion time calculated from retention time alignment. A minimum
SIC peak area equivalent to the typical spectral noise level of 1000
is required of all data reported for label-free quantitation. All indi-

vidual SIC peak areas are normalized to the peak area of the exog-
enously spiked phosphopeptide FQpSEEQQQTEDELQDK or
LIEDApYTAK added prior to phosphopeptide enrichment and
reversed-phase elution into the mass spectrometer. The p-values
are calculated from at least three replicate analyses. To overcome
the possibility of missing data caused by the stochastic peak selec-
tion in data-dependent scanning in the mass spectrometer, the col-
lection of at least five biological replicate analyses is strongly
recommended. To select phosphopeptides that show a statistically
significant change in abundance between stimulated cells and con-
trol, two-tailed unpaired student’s t tests and q-values for multiple
hypothesis tests can be calculated based on the determined p-values
using the R package QVALUE as previously described [34, 35].
Other methods such as ANOVA with post hoc Tukey HSD [36]
may be used to correct for multiple hypothesis. Some form of mul-
tiple hypothesis correction is required for publication of this type
of data. Determination of the significantly altered set of phospho-
peptides can be facilitated by selection of an FDR and fold change
threshold for the quantitative data visualized with a volcano plot.
The proteomic community favors a 1% or lower FDR threshold on
peptide sequence assignment and a 5% or lower FDR threshold on
the quantitative data.
4 Notes
1. It is extremely important to freshly prepare and use the urea

solution within 4 h to prevent artefactual carbamylation of
peptides caused by the decomposition of the urea.
2. Synthetic peptides should be dissolved in 0.1 M acetic acid.
3. To achieve optimal phosphotyrosine identification yields, the
frozen purified peptides should be lyophilized.
4. Mix well by pipetting. Do not vortex.
5. Add 5 μL of each standard from a 100 fmol/μL stocks.
6. Mix well in the tips by pipetting. Discard the waste fluid from
the waste fluid tube, and then centrifuge again at 1000 × g for
10 min at room temperature to remove all of the buffers from
the tips.
7. The phosphotyrosine enrichment protocol is adapted from the
p-Tyr-100 manufacturer’s suggested protocol (http://www.
cellsignal.com)
8. Cut 1 cm of the 200 μL tips when pick up the beads slurry.
9. After adding PBS, invert the tube to avoid bead precipitation.
Supernatant should be removed with micropipette. To avoid
the loss of bead, do not completely suck the soup.
10. Do not vortex or mixing by pipet. Dissolve peptide by hand by

gentle shaking (or use a VWR model 200 rocker, speed
2–3 rpm or similar rocker) for 30 min at room temperature.
Use sonication water bath to aid the dissolve (~10 s).
11. Remove supernatant using 250 μL tips.
12. Mix by inverting tube five times. Remove supernatant with 1
mL micropipette (aspirate the supernatant as much as possi-
ble, using 1 mL, 250 μL, and 20 μL tips, respectively).
Minimize the loss of beads during aspiration.
13. In this step, unbuffered LC-grade water should be used.

Remove the supernatant completely with insulin syringe plac-
ing the angled edge of the needle against the plastic and insert-
ing the needle into the beads while keeping the angled opening
against the plastic. The insulin syringe is only mandatory when
all the liquid needs to be removed from the beads.
14. Tap the bottom of the tube with fingers for several times (gen-
tly touch using the palm side of the finger. Do not vortex. Use
an insulin syringe to collect the eluted peptides from the beads.
15. Spin down the supernatant 1500 × g, 2 min, 4 °C. Split the
solution into two tubes. (Because ZipTip can only hold 50 μL,
yet for the whole 100 μL sample, use the same ZipTip.) Spin
down the solution.
16. Cut a 250 μL tip to fit ZipTip; the pipette setting could be
higher than 50 μL to make the aspiration and dispensing eas-
ier. Expelling the solution to a lint-free tissue (Kimwipes),
briefly touch the liquid at the end to the tissue.
17. Pipette the Solvent A with a micropipette. Expel to a lint-free
tissue. After the last wash, dab the tip well, but not to excess,
on the lint-free tissue.
18. Use a 20 μL pipette which exactly fits the ZipTip without an
adapter. Pipette 10 μL of Solvent B into a new microcentri-
fuge tube. Draw the solvent into the tip and expel. Repeat this
for 8–10 times. During this process, the tip is always in the
solution. Expel the entire volume into a sample vial.
19. Leave the cap on the tube, but use needle to make a hole.
Acknowledgments
The authors wish to acknowledge the financial support from NIH

grant R01 AI083636 and NIH grant P30 GM110759. In addi-
tion, this research is based in part upon work conducted using the
Rhode Island NSF/EPSCoR Proteomics Share Resource Facility,
which is supported in part by the National Science Foundation
EPSCoR Grant No. 1004057, National Institutes of Health Grant
No. 1S10RR027027, a Rhode Island Science and Technology

Advisory Council grant, and the Division of Biology and Medicine,
Brown University.
Conflict of Interest: The authors declare no conflict of interest.
References
1. Ahsan N, Huang Y, Tovar-Mendez A, Swatek urification and stable isotope dimethyl label-
p
KN, Zhang J, Miernyk JA, Xu D, Thelen JJ ing. Mol Cell Proteomics 9:84–99
(2013) A versatile mass spectrometry-based 11. Di Palma S, Zoumaro-Djayoon A, Peng M,
method to both identify kinase client- Post H, Preisinger C, Munoz J, Heck AJ
relationships and characterize signaling net- (2013) Finding the same needles in the hay-
work topology. J Proteome Res 12:937–948 stack? A comparison of phosphotyrosine pep-
2. Larsen MR, Graham ME, Robinson PJ, tides enriched by immuno-affinity precipitation
Roepstorff P (2004) Improved detection of and metal-based affinity chromatography.
hydrophilic phosphopeptides using graphite J Proteomics 91:331–337
powder microcolumns and mass spectrometry: 12. Rikova K, Guo A, Zeng Q, Possemato A, Yu J,
evidence for in vivo doubly phosphorylated Haack H, Nardone J, Lee K, Reeves C, Li Y,
dynamin I and dynamin III. Mol Cell Hu Y, Tan Z, Stokes M, Sullivan L, Mitchell J,
Proteomics 3:456–465 Wetzel R, Macneill J, Ren JM, Yuan J, Bakalarski
3. Steen, H., Stensballe, A., Jensen, O.N. (2007) CE, Villen J, Kornhauser JM, Smith B, Li D,
Phosphopeptide purification by IMAC with Zhou X, Gygi SP, Gu TL, Polakiewicz RD,
Fe(III) and Ga(III). CSH Protoc 2007, pdb Rush J, Comb MJ (2007) Global survey of
prot4607 phosphotyrosine signaling identifies oncogenic
4. Thingholm TE, Jorgensen TJ, Jensen ON, kinases in lung cancer. Cell 131:1190–1203
Larsen MR (2006) Highly selective enrich- 13. Rush J, Moritz A, Lee KA, Guo A, Goss VL,
ment of phosphorylated peptides using tita- Spek EJ, Zhang H, Zha XM, Polakiewicz RD,
nium dioxide. Nat Protoc 1:1929–1935 Comb MJ (2005) Immunoaffinity profiling of
5. Jayasundera KB, Iliuk AB, Nguyen A, Higgins tyrosine phosphorylation in cancer cells. Nat
R, Geahlen RL, Tao WA (2014) Global phos- Biotechnol 23:94–101
phoproteomics of activated B cells using com- 14. Salomon AR, Ficarro SB, Brill LM, Brinker A,
plementary metal ion functionalized soluble Phung QT, Ericson C, Sauer K, Brock A, Horn
nanopolymers. Anal Chem 86:6363–6371 DM, Schultz PG, Peters EC (2003) Profiling
6. Hunter T, Sefton BM (1980) Transforming of tyrosine phosphorylation pathways in human
gene product of Rous sarcoma virus phosphor- cells using mass spectrometry. Proc Natl Acad
ylates tyrosine. Proc Natl Acad Sci U S A Sci U S A 100:443–448
77:1311–1315 15. van der Mijn JC, Labots M, Piersma SR, Pham
7. Lemeer S, Heck AJ (2009) The phosphopro- TV, Knol JC, Broxterman HJ, Verheul HM,
teomics data explosion. Curr Opin Chem Biol Jimenez CR (2015) Evaluation of different
13:414–420 phospho-tyrosine antibodies for label-free phos-
8. Hunter, T. (1998) The Croonian Lecture phoproteomics. J Proteomics 127:259–263
1997. The phosphorylation of proteins on 16. Ji Q, Ding Y, Salomon AR (2015) SRC homol-
tyrosine: its role in cell growth and disease. ogy 2 domain-containing leukocyte phospho-
Philos Trans R Soc Lond B Biol Sci 353, protein of 76 kDa (SLP-76) N-terminal
583–605. tyrosine residues regulate a dynamic signaling
9. Zoumaro-Djayoon AD, Heck AJ, Munoz equilibrium involving feedback of proximal
J (2012) Targeted analysis of tyrosine phos- T-cell receptor (TCR) signaling. Mol Cell
phorylation by immuno-affinity enrichment of Proteomics 14:30–40
tyrosine phosphorylated peptides prior to mass 17. Helou YA, Nguyen V, Beik SP, Salomon AR
spectrometric analysis. Methods 56:268–274 (2013) ERK positive feedback regulates a
10. Boersema PJ, Foong LY, Ding VM, Lemeer S, widespread network of tyrosine phosphoryla-
van Breukelen B, Philp R, Boekhorst J, Snel B, tion sites across canonical T cell signaling and
den Hertog J, Choo AB, Heck AJ (2010) actin cytoskeletal proteins in Jurkat T cells.
In-depth qualitative and quantitative profiling PLoS One 8:e69641
of tyrosine phosphorylation using a combina- 18. Helou, Y.A., Petrashen, A.P., Salomon, A.R.
tion of phosphopeptide immunoaffinity (2015) Vav1 regulates T-cell activation through
a feedback mechanism and crosstalk between ing by tandem mass spectrometry. Rapid
the T-cell receptor and CD28. J Proteome Commun Mass Spectrom 17:2337–2342
Res 14:2963-75 28. Licklider LJ, Thoreen CC, Peng J, Gygi SP
19. Helou YA, Salomon AR (2015) Protein net- (2002) Automation of nanoscale microcapil-
works and activation of lymphocytes. Curr lary liquid chromatography-tandem mass spec-
Opin Immunol 33:78–85 trometry with a vented column. Anal Chem
20. Sjolin-Goodfellow H, Frushicheva MP, Ji Q, 74:3076–3083
Cheng DA, Kadlecek TA, Cantor AJ, Kuriyan 29. Elias JE, Gygi SP (2007) Target-decoy search
J, Chakraborty AK, Salomon AR, Weiss A strategy for increased confidence in large-scale
(2015) The catalytic activity of the kinase ZAP- protein identifications by mass spectrometry.
70 mediates basal signaling and negative feed- Nat Methods 4:207–214
back of the T cell receptor pathway. Sci Signal 30. Chambers MC, Maclean B, Burke R, Amodei
8:ra49 D, Ruderman DL, Neumann S, Gatto L,
21. Ong SE, Blagoev B, Kratchmarova I, Kristensen Fischer B, Pratt B, Egertson J, Hoff K, Kessner
DB, Steen H, Pandey A, Mann M (2002) D, Tasman N, Shulman N, Frewen B, Baker
Stable isotope labeling by amino acids in cell TA, Brusniak MY, Paulse C, Creasy D, Flashner
culture, SILAC, as a simple and accurate L, Kani K, Moulding C, Seymour SL, Nuwaysir
approach to expression proteomics. Mol Cell LM, Lefebvre B, Kuhlmann F, Roark J, Rainer
Proteomics 1:376–386 P, Detlev S, Hemenway T, Huhmer A,
22. Ross PL, Huang YN, Marchese JN, Williamson Langridge J, Connolly B, Chadick T, Holly K,
B, Parker K, Hattan S, Khainovski N, Pillai S, Eckels J, Deutsch EW, Moritz RL, Katz JE,
Dey S, Daniels S, Purkayastha S, Juhasz P, Agus DB, MacCoss M, Tabb DL, Mallick P
Martin S, Bartlet-Jones M, He F, Jacobson A, (2012) A cross-platform toolkit for mass spec-
Pappin DJ (2004) Multiplexed protein quanti- trometry and proteomics. Nat Biotechnol
tation in Saccharomyces cerevisiae using amine- 30:918–920
reactive isobaric tagging reagents. Mol Cell 31. Beausoleil SA, Villen J, Gerber SA, Rush J,
Proteomics 3:1154–1169 Gygi SP (2006) A probability-based approach
23. Thompson A, Schafer J, Kuhn K, Kienle S, for high-throughput protein phosphorylation
Schwarz J, Schmidt G, Neumann T, Johnstone analysis and site localization. Nat Biotechnol
R, Mohammed AK, Hamon C (2003) Tandem 24:1285–1292
mass tags: a novel quantification strategy for 32. Demirkan G, Yu K, Boylan JM, Salomon AR,
comparative analysis of complex protein mix- Gruppuso PA (2011) Phosphoproteomic pro-
tures by MS/MS. Anal Chem 75:1895–1904 filing of in vivo signaling in liver by the mam-
24. Yu K, Salomon AR (2010) HTAPP: high- malian target of rapamycin complex 1
throughput autonomous proteomic pipeline. (mTORC1). PLoS One 6:e21729
Proteomics 10:2113–2122 33. Smith CA, Want EJ, O'Maille G, Abagyan R,
25. Yu K, Salomon AR (2009) PeptideDepot: flex- Siuzdak G (2006) XCMS: processing mass
ible relational database for visual analysis of spectrometry data for metabolite profiling
quantitative proteomic data and integration of using nonlinear peak alignment, matching, and
existing protein information. Proteomics identification. Anal Chem 78:779–787
9:5350–5358 34. Storey JD (2003) The positive false discovery
26. Perkins DN, Pappin DJ, Creasy DM, Cottrell rate: a Bayesian interpretation and the q-value.
JS (1999) Probability-based protein identifica- Ann Stat 31:2013–2035
tion by searching sequence databases using 35. Storey JD, Tibshirani R (2003) Statistical sig-
mass spectrometry data. Electrophoresis nificance for genomewide studies. Proc Natl
20:3551–3567 Acad Sci U S A 100:9440–9445
27. Ma B, Zhang K, Hendrie C, Liang C, Li M, 36. Oberg AL, Mahoney DW (2012) Statistical
Doherty-Kirby A, Lajoie G (2003) PEAKS: methods for quantitative mass spectrometry
powerful software for peptide de novo sequenc- proteomic experiments with labeling. BMC
Bioinformatics 13(Suppl 16):7
Chapter 23
Imaging Asymmetric T Cell Division

Mirren Charnley and Sarah M. Russell
Abstract
Asymmetric cell division (ACD) controls cell fate decisions in model organisms such as Drosophila and C.
elegans and has recently emerged as a mediator of T cell fate and hematopoiesis. The most appropriate
methods for assessing ACD in T cells are still evolving. Here we describe the methods currently applied to
monitor and measure ACD of developing and activated T cells. We provide an overview of approaches for
capturing cells in the process of cytokinesis in vivo, ex vivo, or during in vitro culture. We provide methods
for in vitro fixed immunofluorescent staining and for time-lapse analysis. We provide an overview of the
different approaches for quantification of ACD of lymphocytes, discuss the pitfalls and concerns in inter-
pretation of these analyses, and provide detailed methods for the quantification of ACD in our group.
Key words Asymmetric cell division, T Cells, Lymphocytes, Cell polarity, Fluorescence microscopy,
Image quantification, Thymocyte, Cell fate determinants
1 Introduction
A fundamental question in cell biology is how a single cell can reli-

ably produce progeny of different cell types and how the relative
proportions of these different progenies are controlled. Asymmetric
cell division (ACD) enables the differential programming of diver-
gent cell fates in the progeny of a single cell. ACD refers to molec-
ular asymmetry (particularly of fate determinants) that is controlled
by external cues (not stochastic) and has the potential to differen-
tially influence cell fate [1].
ACD was first identified in model organisms such as the worm
and fly, and it is now clear that ACD orchestrates fundamental
aspects of stem cell biology, cell differentiation, and cancer [2–4].
In the last decade, it has become apparent that the development
and diversification of T cells also occur through ACD. To date,
ACD has been shown to occur in CD8+ [5], CD4+ [6] and devel-
oping T cells, specifically DN3a thymocytes, during the β-selection
checkpoint [7]. Correlative evidence further suggests that ACD
influences fate choices, both during development and upon
383
384 Mirren Charnley and Sarah M. Russell
a ctivation to differentiate to effector and memory T cells [5, 8–

14]. This has prompted investigations into the mechanisms by
which ACD is controlled, the molecular and cellular consequences
of ACD, and its role in immunity and leukemogenesis. A central
requirement of these studies is the capacity to measure ACD.
1.1 Imaging ACD: The spatial nature of the ACD process means that approaches to
The Context assess ACD almost invariably require imaging. A critical aspect of
studies of ACD is the role of the microenvironment of the dividing
cell. In general, a polarity cue, such as neighboring cells or the adhe-
sive environment in the cell niche, triggers asymmetric segregation
of molecules during cell division to culminate in two daughter cells
with different fate potentials. In small model organisms, in situ
imaging is readily used to assess ACD, but the difficulties of in vivo
imaging at the resolution and sensitivity needed to discern molecular
asymmetry in a dividing cell, and the very small number of cells
undergoing division at any given time, mean such experiments have
not yet been performed in mammals. Imaging ex vivo (for instance,
extracting cells from lymph nodes at the time of division, fixing, and
immunostaining [5, 9, 10, 12]) requires disengagement of the T cell
from its polarity cue, potentially altering polarity (see Note 1).
Alternative efforts have generally focused on providing the
polarity cue in vitro, which raises the obvious concern of physiologi-
cal representation, but allows for control and observation of the
molecular cue [6, 7, 11, 12]. An in vitro approach allows for suffi-
cient numbers of dividing cells to be analyzed for robust statistical
quantification and avoids the potential artifacts associated with tissue
extraction. In vitro imaging is also compatible with time-lapse analy-
sis [6, 7, 12], which provides opportunities to explore the functional
outcome of an ACD. For both fixed and time-lapse imaging, the
polarity cue can be provided either in the form of a cell (such as the
dendritic cell-presenting antigen to a mature T cell [11] or a stromal
cell interacting with a developing T cell [7]) or the specific molecules
that engage the mature T cell (such as immobilized anti-CD3 and
recombinant ICAM1-Fc fusion protein [6]) or the developing T cell
(ligands for notch and CXCR4). Time-lapse imaging requires fluo-
rescent labeling of individual proteins in living cells, which can be
achieved by genetic modification, or culture in the presence of small
molecule labels or fluorescent antibodies to surface proteins. We
describe in vitro approaches involving co-culture with cells as the
polarizing cue for fixed imaging (Subheading 3.1), expression of
fluorescently tagged proteins in developing T cells by retroviral
transduction (Subheading 3.2), and co-culture for imaging of ACD
using time-lapse microscopy (Subheading 3.3).
1.2 Quantification Early studies of ACD tended to focus on extreme scenarios in which
of ACD all cells underwent a stereotypical ACD that absolutely dictated
subsequent fate of the daughter cells, and the ACD often involved
Monitoring Asymmetric Cell Division in T Cells 385
extensive asymmetry, such a several fold difference in the size of the

two daughters [15]. More recent examples of ACD can involve
much more subtle behavior and might not necessarily impact upon
all the cells [16]. For these scenarios, careful experimentation and
quantification is required to determine the extent of ACD. Efforts
to determine the most reliable means of quantifying ACD in T cells
are ongoing, but recent publications have highlighted increasingly
sophisticated and valuable means of improving the reliability of
both the experimentation and the analysis. Particularly challenging
is the need to capture sufficient examples of this rare and transient
event for statistically robust analysis. One means of doing this is to
increase the throughput by performing imaging flow cytometry
[17], although this has the disadvantage of disengaging the divid-
ing T cell from its microenvironment. With continual improve-
ments in automation of imaging analysis methods, it should soon
be relatively straightforward to perform robust quantification of T
cell ACD in both controlled in vitro environments and in situ. We
describe and discuss current approaches to maximize reliability of
the quantification (Subheading 3.4).
2 Materials
2.1 Media 1. T Cell medium: RPMI medium 1640 with glutamine (1 mM),
fetal calf serum (10% v/v), sodium pyruvate (1 nM), and non-
essential amino acids (100 nM).
2. Dendritic cell medium: RPMI medium 1640 supplemented
with granulocyte macrophage colony-stimulating factor
(10 ng/mL), IL-4 (5 ng/mL), glutamine (1 mM), and fetal
calf serum (10% v/v).
3. OP9-DL1 medium: Minimal Essential Medium alpha modifi-
cation supplemented with fetal calf serum (20% v/v), gluta-
mine (1 mM), and penicillin/streptomycin (100 ng/mL).
4. Stem cell medium: Anne Kelso media with fetal calf serum
(10% v/v), β-mercaptoethanol (50 μM), and l-asparagine
(1 mM) and immediately prior to use supplemented with fetal
calf serum (20% v/v), IL-3 (2 ng/mL), IL-6 (2 ng/mL), and
stem cell factor (10 ng/mL).
5. Thymocyte medium: Minimal Essential Medium alpha modifi-
cation supplemented with fetal calf serum (10% v/v), gluta-
mine (1 mM), β-mercaptoethanol (50 μM), sodium pyruvate
(1 nM), HEPES (10 mM), penicillin/streptomycin (100 ng/
mL), mouse interleukin 7 (IL-7, 1 ng/mL), and mouse FMS-
like tyrosine kinase 3 (Flt-3, 5 ng/mL).
6. Phoenix cell medium: Dulbecco’s Minimal Essential Medium
supplemented with fetal calf serum (10% v/v) and l-glutamine
(1 mM).
2.2 Reagents 1. Fixative buffer (PEMED): Pipes (100 mM), magnesium sul-

for Fixing and Perme fate (MgSO4, 5 mM), ethylene glycol tetraacetic acid (EGTA,
abilizing Samples for 10 mM), and dithiothreitol (DTT, 5 mM), pH 7. Add parafor-
Immunofluorescence maldehyde (3.7% w/v) and heat to 65 °C until dissolved and
freeze at −20 °C until required.
2. Permeabilization solution: Triton X-100 (0.1% v/v), bovine
serum albumin in PBS (BSA, 0.5% w/v), and freeze aliquots at
−20 °C. Ensure both reagents are completely defrosted and
prewarmed to 37 °C before use.
3. Staining buffer: 1% BSA in PBS.
2.3 Transfection 1. Retronectin: 15 μg/mL in PBS.

Reagents 2. Calcium chloride (2 M).
3. HEPES-buffered saline solution (×2): Hepes (50 mM), sodium
chloride (NaCl, 280 mM), disodium hydrogen phosphate
anhydrous (Na2HPO4, 1.5 mM), pH 7.
4. pMSCV retroviral construct.
2.4 Primary Cells 1. T cells: Naive OT-1 CD8+ T cells purified from spleens of
C57BL/6 mice at 8–12 weeks of age using MACS negative
selection (see Note 2).
2. Dendritic cells: Bone marrow cells from hind limbs of C57BL/6
mice cultured in GM-CSF for 6 days to generate immature den-
dritic cells (CD11c+, CD86low, and MHC-IIlow) for use as antigen
presenting cells (APC).
3. Thymocytes: E13.5–E14.5 mouse fetal liver cells or bone mar-
row cells isolated from C57Bl/6 mice, rested overnight in stem
cell media, and seeded onto OP9-DL1 stromal cells at a 1:1 ratio
for differentiation of fetal liver cells into thymocytes. Thymocytes
are harvested via forceful pipetting and co-cultured on freshly
seeded OP9-DL1 stromal cells every 3–8 days (see Note 2).
3 Methods
For discussion on the merits and disadvantages of ex vivo analyses,

see Note 1. For discussion on merits and disadvantages of fixed
staining versus time-lapse microscopy to assess asymmetry in divid-
ing cells, see Note 3.
3.1 Immunolabeling 1. Prepare T cell–APC conjugates using dendritic cells for mature
of Proteins to Assess T cells and stromal cells for developing T cells.
Their Symmetry/ Dendritic cell–T cell conjugates: seed 4 × 105 dendritic cells
Asymmetry in Fixed onto 8-well chamber slides and allow to adhere overnight.
Dividing Cells Incubate with 1 mM SIINFEKL for 1 h at 37 °C; wash twice
and subsequently seed 8 × 104 naive T cells for 32–36 h (see

Note 3) before washing the sample to remove non-adherent
cells.
OP9-DL1-DN3a thymocyte conjugates: seed 2 × 104 OP9-DL1
stromal cells per well onto 8-well chamber slide, and allow to
adhere overnight. Add 2 × 104 thymocytes and co-culture
overnight.
2. Fix cells for 10 min at room temperature using fixative reagent,
wash once with PBS, and permeabilize for 5 min at room
temperature.
3. Label cells with primary antibodies for test proteins (see Note
3) and tubulin, wash once with PBS and label with the appro-
priate Alexa Fluor-conjugated secondary antibodies for detec-
tion, wash once, and mount. All antibody staining solutions
should be prepared in staining buffer.
4. Examine the slides using a confocal microscope and a 60× oil
immersion objective. Three-dimensional images of the cells are
acquired with an optical distance of 0.5 μm between slices and
maximum intensity projections of Z sections spanning the
entire cell used for all analyses.
3.2 Expression 1. Seed Phoenix E cells at 1 × 106 cells per 90 mm petri dish and
of Fluorescent allow to adhere overnight.
Proteins for Time- 2. Change media on Phoenix cells 3 h prior to transfection.
Lapse Imaging
3. Prepare two tubes, in tube 1 add 500 μL 2×HeBS per sample and
3.2.1 Calcium Phosphate in tube 2 combine 61 μL 2 M CaCl2, 15 μg of the pMSCV retro-
Transfection and Virus viral construct and water to a final volume of 500 μL per sample.
Production in Phoenix E 4. Add tube 2 to tube 1 while bubbling vigorously, and incubate
Cells at room temperature for 20 min.
5. Add to Phoenix E cells and incubate for 48 h with a media
change after 24 h.
6. Harvest viral supernatant 48 h after transfection, and add to
6-well plates that have been pre-coated with RetroNectin
(15 μg/mL) and blocked in 2% BSA. After addition of the viral
supernatant, centrifuge plates at 2000 × g for 1 h, and incubate
for 1 h at 37 °C.
3.2.2 Generation 1. Harvest 1–4 × 106 thymocytes (days 4–7 after the start of co-
of Transduced Thymocytes culture on OP9-DL1 cells), seed onto plates coated with
RetroNectin and recombinant retrovirus (5 x 10 thymocytes
per wells), and centrifuge for 1 h at 1200 × g.
2. Rest cells in thymocyte medium enriched with IL-7 (20 ng/mL),
Flt-3 (5 ng/mL), and stem cell factor (10 ng/mL) for 24 h, har-
vest thymocytes, and seed onto fresh OP9-DL1 cells for 48 h.
3. Sort for retrovirally transduced cells by flow cytometry using
GFP or Cherry fluorescence (see Note 4).
3.3 Time-Lapse 1. Sample preparation for live cell imaging (see Note 5).
Microscopy Dendritic cell–T cell conjugates: to achieve a one-to-one ratio
of T cells at the appropriate density in sufficient numbers of
grids in an array of 125 μm cell paddocks (see Note 5), seed
2 × 105 dendritic cells onto an appropriate dish for imaging
and allow to adhere overnight. Incubate with 1 mM SIINFEKL
for 1 h at 37 °C, wash at least two times, and subsequently seed
2 × 105 naive T cells* and immediately commence imaging.
OP9-DL1-DN3a thymocytes: seed 2 × 104 OP9-DL1 stromal
cells into 125 μm cell paddocks and allow to adhere overnight
before adding 2 × 104 thymocytes* (see Note 5, transduced
with fluorophores and purified as in Subheading 3.2) * note
that these numbers will vary depending upon the type of imag-
ing chamber, so should be empirically determined.
2. Antibodies directed against surface markers can be added to
the media prior to imaging (see Note 3).
3. Images are captured on a confocal microscope fitted with a
temperature-controlled chamber maintained at 37 °C and 5%
CO2 using a 40× air objective. Multiple stage positions can be
imaged every 2 min, with five-slice z-stacks of 2–3 μm thick-
ness taken at every time point.
4. For longer-term imaging (greater than 2–4 h, see Note 6), we
use a spinning disk confocal system fitted with an inverted
microscope and temperature-controlled chamber maintained
at 37 °C and 5% CO2. Images are acquired using a 40× air
objective, and multiple stage positions are recorded every
3 min, with eight-slice z-stacks of 1 μm thickness.
3.4 Quantification 1. Blind the samples to prevent operator basis.

of ACD 2. Identify individual mitotic cells by the presence of a mitotic
3.4.1 Subjective Scoring spindle (stained using antibodies to alpha-tubulin).
of Dividing Cells 3. For each dividing cell, switch to the channel of the test protein,
and assign the cell division as symmetric if the fluorescence is
evenly distributed between the two daughter cells and asym-
metric if the fluorescence is greater in one of the daughter cells
than the other (see Note 7). Asymmetric divisions can be fur-
ther classified as proximal or distal based upon proximity of the
increased fluorescence to the polarity cue. In some instances, it
is helpful to score polarity in each dividing cell from 1 to 3 to
indicate the extent of polarization and to assign positive values
if the fluorescence is greater in the daughter cell closer to the
polarity cue (proximal ACD) and negative values if the fluores-
cence in the daughter cell distal to the polarity cue is greater.
3.4.2 Quantification To quantify the polarization of fluorescent markers in the nascent

of Fluorescence Intensity daughter cells, first the stacks are projected into a 2D plane, and then a
to Determine Asymmetry region of interest (ROI) is drawn around the two halves of the dividing
cell or the cell boundary of each daughter cell one frame after division.
1. Measure the total fluorescence of each daughter cell.

2. Next apply a “polarization ratio” (PR) equation, PR = (ΣH1
− ΣH2)/(ΣH1 + ΣH2), where the difference in total intensity
between Daughter 1 (ΣH1) and Daughter 2 (ΣH2) is divided
by the sum of intensities in both Daughter 1 (H1) and
Daughter 2 (H2). A value close to 0 indicates the fluorescent
marker is symmetrically distributed, while a value approaching
1 or −1 indicates asymmetrically distributed fluorescence. If
H1 is always allocated to the daughter cell closest to the polar-
ity cue, positive numbers will indicate proximal polarization
and negative numbers distal polarization.
3. In some instances, it can be useful to artificially assign each divi-
sion as either symmetric cell division (SCD) or ACD by arbi-
trarily assigning a cutoff value, with ratios above this arbitrary
value considered asymmetric (see Note 8). It is important to
recognize that we do not yet know what degree of polarization
will determine differential fate of the two daughters (and this
will no doubt differ for each fate determinant, see Note 9), so
the designation of SCD or ACD should be considered purely as
a tool for comparison between populations and conditions.
4. The measurements described above can be performed using
standard image analysis packages such as MetaMorph or
ImageJ. Alternatively, dedicated toolboxes (such as TACTICS
[18, 19]) allow for more automated, high-throughput, and
robust estimations of the degree of polarization (see Notes 7–9).
4 Notes
1. Ex vivo capture of cells.

One issue with ex vivo analysis is that, to capture sufficient
dividing cells for analysis, cells are frequently treated with the
actin inhibitor, cytochalasin B. It was recently shown that this
treatment can cause an increase in the measured polarization
during ACD [20], raising concerns about analyses using cyto-
chalasin B. Extraction of cells from lymph nodes might also
result in biased populations, such as extracting migratory lym-
phocytes more efficiently than T cells engaged with antigen-
presenting cells [21, 22]. These issues and recent studies showing
that cell reconstitution can alter the biology of lymphocytes [23]
highlight that neither in vitro, in vivo, or ex vivo analyses are per-
fect, and a combination of all these approaches will be required
to be certain of context in which T cell ACD occurs.
2. Primary cells.
OT-1 CD8+ cells: We, and others, use transgenic TCR mice
as a source of T cells of a defined specificity. The OT-1 transgenic
mouse is one of the most widely used; in this model the CD8+ T
cells express a TCR specific for the SIINFEKL peptide of oval-
bumin presented on MHC [24].
Thymocytes: To generate thymocytes in vitro, we culture
mouse fetal liver cells on OP9-DL1 stromal cells. This in vitro
co-culture system recapitulates almost all aspects of thymocyte
development and has been extensively used to study develop-
ment [25–28].
3. Fixed versus time-lapse microscopy.
We use two complementary techniques to determine the
presence of ACD in T cells, fixed immunolabeled samples with
antibodies, and time-lapse microscopy of cells expressing fluo-
rescent fusion proteins. Using these techniques, we and others
have been able to demonstrate that a number of polarity and
cell fate proteins are polarized during division (e.g., Numb,
α-Adaptin, aPKC, Par3, Scribble, Dlg, PKCϴ), while other
markers are symmetrically distributed (e.g., CD25, LFA-1,
LAT) [7, 11]. Both fixed and time-lapse techniques have
inherent advantages and disadvantages. In time-lapse micros-
copy, it is possible to study the development and activation of
individual T cells in real time, vastly increasing the knowledge
gained. For instance, it is possible to watch how the asymme-
try of the fluorescent protein changes during the cell cycle and
track how it is maintained in the daughter cells over time.
However, the use of transduction to introduce a protein cou-
pled to a fluorophore has the potential for artifacts related to
overexpression and/or different trafficking caused by the
fusion to the fluorophore [7]. A further consideration is that
fluorophores, such as GFP and Cherry, fold and fluoresce at
different rates and are differentially affected by acidic environ-
ments in the cell [29, 30]. Thus, the choice of fluorophore can
affect how the protein is visualized in the cell. An alternative
way to incorporate a marker is to add fluorophore-conjugated
antibodies for surface markers on the T cells directly to the
media. This circumvents the issue of protein overexpression;
however, the antibody needs to be tested prior to use to ensure
it does not affect differentiation. To reduce the likelihood of
altering differentiation, the antibody concentration needs to
be as low as possible which means the choice of antibody can
be critical and care needs to be taken that the antibody is not
depleted from the media or bleached during the course of the
experiment. However, this strategy for the incorporation of a
marker can be highly beneficial, particularly with T cells, which
are extremely difficult to transduce. Although the expression
of fluorescent markers of polarity can be difficult, a key advan-
tage of time-lapse imaging is the validation of polarity mea-
surements taken at the time of division. If we see asymmetry at
division and asymmetric inheritance of the fluorophore in the

two daughter cells, this is the most convincing evidence that
the ACD is bona fide and not an artifact of the imaging pro-
cess (see [7] for an example).
Conversely, in immunolabeled samples, antibodies are used to
detect the endogenous protein within the cell, circumventing
potential issues with protein overexpression. However, only a
small percentage of cells will be dividing at the time the sample
was fixed, leading to low cell numbers. In particular, with naive T
cells, the optimal time at which most cells are captured in the
process of dividing can vary slightly from experiment to experi-
ment. To increase the chances of capturing sufficient cells in mito-
sis, cells can be fixed at more than one time point (for instance,
32–36 h after addition of the T cells), and the sample with the
highest number of dividing cells is selected for analysis.
In contrast to time-lapse imaging, where we can precisely
select the stage of division to be analyzed and can assess and
compare multiple stages during division, in fixed staining it is
not possible to examine previous and subsequent frames to
ensure that the cell is undergoing mitosis and/or the conjoined
cells are daughters of a recent division. Staging of cell division
can be achieved by staining for alpha-tubulin and DNA dyes;
typically, we select cells in the late stages of mitosis where alpha-
tubulin staining of the midbody is a convincing indicator that
the cells are undergoing cytokinesis. It can be surprisingly
problematic to distinguish dividing cells from juxtaposed cells,
but alpha-tubulin staining to identify the mitotic spindle, in
combination with cell shape, can partially alleviate this issue.
Even though the mitotic spindle is a particularly striking struc-
ture and typically easy to recognize, care is still required par-
ticularly, for example, when the spindle is perpendicular to the
viewer. Thus, combining fixed and time-lapse approaches can
make analysis of the occurrence of ACD more robust.
4. Fluorescence-activated cell sorting to sort for specific thymocyte
subsets.
Thymocyte subsets can be purified by staining for the cell
surface markers CD44 and CD25 to discriminate between
DN1 and DN4 subsets, CD28 to discriminate between DN3a
and DN3b, and CD4 and CD8 to discriminate between the
DN, DP, and SP populations.
5. Containment for live cell imaging.
For live cell imaging, the choice of culture dish is critical,
i.e., the base is of the correct thickness for the objective and
can contain sufficient media to prevent evaporation during
imaging. We typically use a glass bottom 35 mm culture dish
(MatTek, Ashland, MA) or a glass bottom eight-chamber slide
(IBIDI).
Another important consideration is how to contain the cells

within the same field of view. This is particularly crucial when
attempting to track immune cells, which are highly motile. To
address this issue we use cell paddocks, which consists of an array
of microgrids fabricated in polydimethylsiloxane (PDMS) [31].
These types of containment structures can be readily fabricated
in standard microfabrication facilities or can be purchased (e.g.,
from Microsurfaces Pty Ltd., Melbourne, Australia). The size of
the base of the microgrids can be tailored from 70 to 250 μm to
ensure that they fit appropriately within the field of view of the
microscope, while the walls of 60 μm high are sufficient to corral
the cells and restrict their lateral movement. These cell paddocks
allow the cells to be imaged over at least 7 days with no impact
upon viability and proliferation [31].
6. Keeping cells alive and healthy during time-lapse imaging.
The time-lapse imaging process itself can be detrimental to
the cell health, especially during long-term experiments, so
care must be taken to minimize light exposure, thus avoiding
phototoxicity and maintaining cell viability. Thus, live cell
imaging dictates that low light exposures are used to maintain
cell health. It is important to note that phototoxicity affects
the ability of the cell to divide and the timing of cell division.
One way to assess whether the cells are unduly affected by the
imaging process is to monitor the time that it takes for the cell
to divide. In our experience, confocal imaging is effective for
imaging over minutes to hours, but imaging for multiple gen-
erations requires the lower-intensity laser excitation and faster
capture rate of spinning disk microscopes. This trade-off
between light exposure and cell health can compromise image
quality. To combat this, it might be necessary to apply filtering
to the images to remove noise, correct or remove heteroge-
neous backgrounds, or smooth the image for accurate seg-
mentation and quantification of the fluorescence within the
cell. Of course, any further image processing must be carefully
tested to ensure that it yields a dynamic range of polarity indi-
ces that appropriately reflects ACD and symmetric divisions.
7. Quantifying ACD.
The study of ACD has led to a number of different ways
to determine and measure the asymmetry in a system. One
approach is based on subjective scoring where dividing cells
are visually ascribed as symmetric or asymmetric [7].
Limitations of this approach are its subjectivity, and the fact
that any information on the range of differences between the
daughter cells is lost, i.e., if a pair of daughter cells is assigned
as asymmetric, it is not known if this due to a small or large
difference in the level of the biomarker. Objective quantifica-
tion is needed to ensure that the marker is polarized, but the
rapidity of this approach means that it can be extremely use-

ful for comparing populations.
Alternatively, the fluorescence intensity of the biomarker is
measured to produce a ratio of the relative levels in the two
daughter cells and used to assign a division as symmetric or asym-
metric based on an arbitrary value. For instance, in the initial
identification of T cell ACD, if enrichment in one nascent daugh-
ter cell was greater than 1.5-fold more than the other daughter
cell, the division was considered polarized [6]. Again, informa-
tion on the spread in asymmetry is lost, and the assignment of
asymmetry is based on an arbitrary value. Measurement of this
ratio is highly dependent upon the total fluorescence in the cell
(e.g., low levels of fluorescence will frequently yield very high
ratios that do not reflect a genuine ACD), so a more reliable
ratio is to measure the difference in fluorescence over the sum of
fluorescence [7, 11, 12, 32] as described in Subheading 3.4.2.
8. Controls to ensure reliability of ACD measurements.
We incorporate several controls both to weed out aberrant
data and to calibrate measurements.
Incorporation of a diffuse protein.
An important consideration is that if the dividing cell is not
completely flat on the imaging surface, differences in the focal
plane will result in artificially high ratios, however they are
measured. To control for this, a protein, such as GFP or cherry,
can also be incorporated into the immune cell alongside the
marker under test. Both GFP and Cherry (not coupled to any
protein) should be uniformly distributed in the cells and con-
sequently the quantification of the fluorescence intensity is
expected to give rise to a low PR. Reporters such as Cherry-
tubulin, which should also be equally distributed between the
two daughters, can also serve as a control for non-even illumi-
nation. In fixed imaging, costaining for a marker that is not
polarized serves the same purpose. This then indicates whether
a high PR of the test protein arises from genuine polarization
or from non-even illumination of the dividing cell.
Measuring asymmetry along the major and minor axis of the
dividing cell.
Another way to ensure that quantification of ACD is robust
and to circumvent issues associated with imaging artifacts is to
assess the asymmetry along the major and minor axis of the
dividing cell [32]. Typically, the ROI is drawn around the cell
boundary and bisects the dividing cell into two regions at the
cleavage plane (along the minor axis). The additional step in the
analysis requires bisecting the dividing cells along the mitotic
spindle (Fig. 1). In general, ACD should yield asymmetry along
the major axis, but not along the minor axis. In contrast, asym-
metry that is caused by artifacts in the image acquisition or pro-
Fig. 1 Axial subdivision of the dividing cell pair for polarization analysis. Two axes can be used to subdivide the
mitotic cell and determine the distribution of intracellular fluorescence. The major axis is derived from the
longest diameter of an ellipse that overlaps the cell. The minor axis is perpendicular to the major axis and cor-
responds to the cleavage plane between the two daughter cells (as indicated by the dotted black lines).
ACD is reflected by a high PR value when the cell is bisected along the major axis (ie. divided by the minor
axis), but the PR when the cell is bisected along the minor axis (ie. divided by the major axis) is generally low.
Examples are shown of symmetric versus asymmetric cells with diffuse versus punctuate distributions of the
biomarker. See [32] for further detail
cessing might be evident across both axes. Thus, the inclusion

of a diffuse control protein and analysis along both axes of the
cells enables us to identify dividing cells with artificially ascribed
polarization and remove these from the analysis.
One major advantage of time-lapse imaging is that the
daughters of a dividing cell can be tracked over time. This is
critical for indicating whether a measurement of asymmetry at
division results in a differential inheritance of the protein in
the daughters, meeting the functional definition of ACD. The
asymmetric inheritance of fluorescence in the daughters over
several timeframes is therefore the most compelling control to
verify measurements in the dividing cell.
Determining the appropriate threshold to remove background
fluorescence.
In addition to enabling the identification of artificially high
ratios, these controls (nonpolarized marker and analysis along
both axis) can be used to determine the appropriate filters and
thresholding levels that need to be applied to the image prior
to analysis. Such image processing can be necessary to achieve
a dynamic range of polarization ratios that also appropriately
reflects ACD and symmetric divisions. In our group, to calcu-
late the threshold, we first determine the difference between
the maximum fluorescence intensity in the ROIs and the aver-
age fluorescence intensity of the background. The threshold
level (0–100%) is determined as a percentage of this difference
and added to the minimum intensity of the ROI to produce
the threshold value which is then applied to the image (32).
The integrated fluorescence intensity, at each threshold level,

of the two ROIs is then used to calculate the PR. Subsequently,
the PR values for the diffuse control protein and the test pro-
tein are plotted as function of the threshold level; we typically
plot this data in the form of a heat map (Fig. 2a). From this
plot, the appropriate threshold level can be determined; spe-
cifically this is a threshold level that does not artificially intro-
duce asymmetric events in the diffuse protein control. To
confirm this analysis, the PR values determined by bisecting
the cell along the major axis are also plotted as a function of
threshold. The threshold level is deemed as appropriate if low,
i.e., nonpolarized, PR values are observed.
Determining the cutoff between SCD and ACD.
Often it is useful to be able to assign a cutoff in the polariza-
tion ratios and ascribe the division as asymmetric versus sym-
metric, particularly when comparing between different
subpopulations or experimental setups. Previously, this has
been based on a rather arbitrary value, and currently it is
unclear what is a biologically relevant difference in polarity (see
Note 9). However, the inclusion of the diffuse marker and
analysis along both axes enables us to identify a baseline level
of asymmetry that can be helpful in assigning a cutoff between
ACD and SCD. The PR values, at the appropriate threshold
level, for the control protein for each pair of dividing cells is
plotted against the PR values for the test protein in the form
of a dot plot (Fig. 2b). The spread of the PR values for the
control protein is then used to assign the cutoff between ACD
and SCD; specifically, the level of the cutoff should result in
few to none of the control protein divisions being assigned as
asymmetric. In addition, the PR values across the major and
minor axes can be plotted against each other and used to con-
firm the cutoff and aberrant data determined with the diffuse
protein control. Accordingly, we use both the inclusion of a
diffuse control protein and the quantification of the PR along
the major axis of the dividing cell to objectively ascribe the
threshold level and ensure appropriate quantification.
Diffuse versus punctate biomarkers.
One important consideration is that not all proteins are
evenly distributed within the cell, instead many proteins cluster
or are segregated into different areas of the cell. This can lead
to divisions being falsely assigned as asymmetric. To explore
this problem, we have stimulated dividing cells with varying
numbers of clusters per hemisphere and assessed for ACD [32].
At low cluster numbers (less than 8–10 clusters per cell), the
PR values became aberrantly high, indicating that particular
care needs to be exerted when the protein is clustered.
Fig. 2 (a) Plotting PR values of control versus test protein to determine settings for comparison and to assess data
quality. PR values of control (diffuse) and test protein are calculated for incremental increases in threshold level
(0–100%) and plotted as a heat map. The black–white scale indicates the PR values; the black regions and
dark gray regions indicate low PR values, corresponding to symmetric divisions, while white and light gray
regions indicate higher PRs and asymmetric divisions. The spread of the PR values of the diffuse control protein is
used to determine the appropriate threshold level that avoids artificially high PR values in the diffuse control protein
but provides good dynamic range of PR in the test protein. In this example an appropriate threshold is 20% (as
indicated by the white dotted line). (b) Plotting PR values of control versus test protein to determine a cutoff
value with which to artificially ascribe divisions as symmetric or asymmetric. PR values of control (diffuse) and test
protein are calculated and plotted in dot plot format. The spread of the PR values of the diffuse control protein is
used to designate the boundary between asymmetric and symmetric divisions. The position of the boundary should
be set at a PR value that results in the designation of the PR values of the diffuse protein as symmetric. In the
example dot plot, the boundary PR value would be set at 0.2. Thus, dividing cells with a PR over 0.2 are assigned
as asymmetric, highlighted by the blue box. Potential outliers (indicating poor-quality data) with high PR values
for the diffuse control protein can also be identified, such as the dividing cells highlighted with the red oval
9. What is a meaningful degree of polarity?

Many researchers, ourselves included, rely on a cutoff value
to assign a difference in polarization between two daughter
cells as symmetric or asymmetric. However, we do not know
what degree of asymmetry is required to drive asymmetric cell
fate, making the consequences of measured asymmetry diffi-

cult to predict. It is likely that ACD does not dictate an abso-
lute bifurcation in T cell fate, but might, with other factors,
influence the propensity of each daughter to adopt different
cell fates [33, 34]. The level of asymmetry required to cause
differential cell fates might also vary between different protein
biomarkers. To define the level of meaningful polarity, it will
be necessary for researchers to watch individual immune cells
undergoing ACD and track the daughter cells to determine if
and when this correlates with differences in cell fate. With the
development of new methods that allow for more controlled
estimations of the degree of polarization to improve the accu-
racy of quantification and cell tracking over a number of gen-
erations [7, 32], we are hopefully inching closer to this goal.
Acknowledgments
We would like to acknowledge members of the Russell lab (past

and present) for the development of the methods described here.
References
1. Knoblich JA (2010) Asymmetric cell division: 7. Pham K, Shimoni R, Charnley M, Ludford-
recent developments and their implications for Menting MJ, Hawkins ED, Ramsbottom K,
tumour biology. Nat Rev Mol Cell Biol Oliaro J, Izon D, Ting SB, Reynolds J, Lythe
11:849–860 G, Molina-Paris C, Melichar H, Robey E,
2. Baena-López LA, Baonza A, García-Bellido A Humbert PO, Gu M, Russell SM (2015)
(2005) The orientation of cell divisions deter- Asymmetric cell division during T cell develop-
mines the shape of Drosophila organs. Curr ment controls downstream fate. J Cell Biol
Biol 15:1640–1644 210:933–950
3. Pereira G, Yamashita YM (2011) Fly meets 8. Arsenio J, Kakaradov B, Metz PJ, Kim SH, Yeo
yeast: checking the correct orientation of cell GW, Chang JT (2014) Early specification of
division. Trends Cell Biol 21:526–533 CD8+ T lymphocyte fates during adaptive
4. Yamashita YM (2010) Cell adhesion in regula- immunity revealed by single-cell gene-
tion of asymmetric stem cell division. Curr expression analyses. Nat Immunol 15:365–372
Opin Cell Biol 22:605–610 9. Metz PJ, Arsenio J, Kakaradov B, Kim SH,
5. Chang JT, Palanivel VR, Kinjyo I, Schambach Remedios KA, Oakley K, Akimoto K, Ohno S,
F, Intlekofer AM, Banerjee A, Longworth SA, Yeo GW, Chang JT (2015) Regulation of
Vinup KE, Mrass P, Oliaro J, Killeen N, Orange asymmetric division and CD8+ T lymphocyte
JS, Russell SM, Weninger W, Reiner SL (2007) fate specification by protein kinase Czeta and
Asymmetric T lymphocyte division in the initi- Protein Kinase Clambda/iota. J Immunol
ation of adaptive immune responses. Science 194:2249–2259
315:1687–1691 10. Metz PJ, Lopez J, Kim SH, Akimoto K, Ohno
6. Chang JT, Ciocca ML, Kinjyo I, Palanivel VR, S, Chang JT (2016) Regulation of asymmetric
McClurkin CE, Dejong CS, Mooney EC, Kim division by atypical protein kinase C influences
JS, Steinel NC, Oliaro J, Yin CC, Florea BI, early specification of CD8(+) T lymphocyte
Overkleeft HS, Berg LJ, Russell SM, Koretzky fates. Sci Rep 6:19182
GA, Jordan MS, Reiner SL (2011) Asymmetric 11. Oliaro J, Van Ham V, Sacirbegovic F, Pasam A,
proteasome segregation as a mechanism for Bomzon Z, Pham K, Ludford-Menting MJ,
unequal partitioning of the transcription factor Waterhouse NJ, Bots M, Hawkins ED, Watt
T-bet during T lymphocyte division. Immunity SV, Cluse LA, Clarke CJ, Izon DJ, Chang JT,
34:492–504 Thompson N, Gu M, Johnstone RW, Smyth
MJ, Humbert PO, Reiner SL, Russell SM occurring early after antigen presentation but
(2010) Asymmetric cell division of T cells upon before clonal expansion is impacted by Toll-like
antigen presentation uses multiple conserved receptor stimulation. J Immunol 172:248–259
mechanisms. J Immunol 185:367–375 23. Busch K, Klapproth K, Barile M, Flossdorf M,
12. Verbist KC, Guy CS, Milasta S, Liedmann S, Holland-Letz T, Schlenner SM, Reth M, Hofer
Kaminski MM, Wang R, Green DR (2016) T, Rodewald HR (2015) Fundamental
Metabolic maintenance of cell asymmetry fol- properties of unperturbed haematopoiesis

lowing division in activated T lymphocytes. from stem cells in vivo. Nature 518:542–546
Nature 532:389–393 24. Hogquist, K.A., Jameson, S.C., Heath, W.R.,
13. King CG, Koehli S, Hausmann B, Schmaler M, Howard, J.L., Bevan, M.J., Carbone, F.R.
Zehn D, Palmer E (2012) T cell affinity regu- (1994) T Cell receptor antagonist peptides
lates asymmetric division, effector cell differen- induce positive selection. Cell 76:17–27
tiation, and tissue pathology. Immunity 25. Janas ML, Varano G, Gudmundsson K, Noda
37:709–720 M, Nagasawa T, Turner M (2010) Thymic
14. Ramsbottom KM, Sacirbegovic F, Hawkins development beyond beta-selection requires
ED, Kallies A, Belz GT, Van Ham V, Haynes phosphatidylinositol 3-kinase activation by
NM, Durrant MJ, Humbert PO, Russell SM, CXCR4. J Exp Med 207:247–261
Oliaro J (2016) Lethal giant larvae-1 deficiency 26. Mohtashami M, Shah DK, Nakase H, Kianizad
enhances the CD8(+) effector T-cell response K, Petrie HT, Zuniga-Pflucker JC (2010)
to antigen challenge in vivo. Immunol Cell Direct comparison of DLL1-and DLL4-
Biol 94:306–311 mediated notch activation levels shows differ-
15. Rose LS, Kemphues KJ (1998) Early pattern- ential lymphomyeloid lineage commitment
ing of the C. elegans embryo. Annu Rev Genet outcomes. J Immunol 185:867–876
32:521–545 27. Schmitt TM, Zuniga-Pflucker JC (2002)
16. Pham K, Sacirbegovic F, Russell SM (2014) Induction of T cell development from hemato-
Polarized cells, polarized views: asymmetric cell poietic progenitor cells by delta-like-1 in vitro.
division in hematopoietic cells. Front Immunol Immunity 17:749–756
5:26 28. Van de Walle I, Waegemans E, De Medts J, De
17. Filby A, Perucha E, Summers H, Rees P, Chana Smet G, De Smedt M, Snauwaert S,
P, Heck S, Lord GM, Davies D (2011) An Vandekerckhove B, Kerre T, Leclercq G, Plum
imaging flow cytometric method for measuring J, Gridley T, Wang T, Koch U, Radtke F,
cell division history and molecular symmetry Taghon T (2013) Specific Notch receptor–
during mitosis. Cytometry A 79:496–506 ligand interactions control human TCR-αβ/γδ
18. Pham K, Shimoni R, Ludford-Menting MJ, development by inducing differential Notch
Nowell CJ, Lobachevsky P, Bomzon Z, Gu M, signal strength. J Exp Med 210:683–697
Speed TP, McGlade CJ, Russell SM (2013) 29. Couturier L, Trylinski M, Mazouni K, Darnet
Divergent lymphocyte signalling revealed by a L, Schweisguth F (2014) A fluorescent tagging
powerful new tool for analysis of time-lapse approach in Drosophila reveals late endosomal
microscopy. Immunol Cell Biol 91:70–81 trafficking of Notch and Sanpodo. J Cell Biol
19. Shimoni R, Pham K, Yassin M, Gu M, Russell 207:351–363
SM (2013) TACTICS, an interactive platform 30. Short B (2014) Red and green traffic signals.
for customized high-content bioimaging analy- J Cell Biol 207:319
sis. Bioinformatics 29:817–818 31. Day D, Pham K, Ludford-Menting MJ, Oliaro
20. Hawkins ED, Oliaro J, Kallies A, Belz GT, Filby J, Izon D, Russell SM, Gu M (2009) A method
A, Hogan T, Haynes N, Ramsbottom KM, Van for prolonged imaging of motile lymphocytes.
Ham V, Kinwell T, Seddon B, Davies D, Immunol Cell Biol 87:154–158
Tarlinton D, Lew AM, Humbert PO, Russell 32. Shimoni R, Pham K, Yassin M, Ludford-
SM (2013) Regulation of asymmetric cell divi- Menting MJ, Gu M, Russell SM (2014)
sion and polarity by Scribble is not required for Normalized polarization ratios for the analysis
humoral immunity. Nat Commun 4:1801 of cell polarity. PLoS One 9:e99885
21. Steinert EM, Schenkel JM, Fraser KA, Beura 33. Yassin M, Russell SM (2016) Polarity and
LK, Manlove LS, Igyarto BZ, Southern PJ, asymmetric cell division in the control of lym-
Masopust D (2015) Quantifying memory CD8 phocyte fate decisions and function. Curr Opin
T cells reveals regionalization of immunosur- Immunol 39:143–149
veillance. Cell 161:737–749 34. Buchholz, V.R., Schumacher, T.N., Busch,
22. Maxwell JR, Rossi RJ, McSorley SJ, Vella AT D.H. (2016) T Cell fate at the single-cell level.
(2004) T cell clonal conditioning: a phase Annu Rev Immunol 34:65–92
Chapter 24
Ultrastructure of Immune Synapses

Jaime Llodrá
Abstract
The immunological synapse is a critical event for immune response development. The use of planar
supported bilayers as surrogate antigen-presenting cells is a useful tool to study this phenomenon. Here
we describe electron microscopy methods and approaches to expand our knowledge of the events taking
place during the initial phases of T cell activation after antigen recognition at the nanometer scale.
Key words Immunological synapse, Electron microscopy, Tomography, Correlative microscopy
1 Introduction
Improvements in instrumentation and software have led to the

practical implementation of electron tomography with consequent
gains in resolution and data quality. Furthermore, other techniques
such as focused ion beam-scanning electron microscopy have
enabled the analysis of large volumes although at low resolutions,
thus complementing the data generated by electron tomography.
The use of supported planar lipid bilayers was a milestone in the
study of the immunological synapse [1]. It allows for the construc-
tion of a model membrane that mimics the plasma membrane of
antigen-presenting cells. Its simplicity underlies its power since the
user can modulate the type and density of ligands on the surface
which are contacted by T lymphocytes. Use of fluorescently labeled
probes made it possible to highlight the behavior of specific mol-
ecules during immunological synapse formation, and total internal
reflection microscopy served to highlight events at the contact
interface [2]. While these advances were instrumental in furthering
our knowledge about the immunological synapse, there were no
clear overall high-resolution pictures of this interaction. Here, we
show that chemically fixed samples of CD4+ T lymphocytes on
supported lipid bilayers can be prepared for conventional electron
microscopy/tomography, and in this way all the events taking
place at the contact site could be analyzed [3, 4]. Depending on
399
400 Jaime Llodrá
the type of processing for electron microscopy, it is possible to do

correlative microscopy, thus, for the first time, giving a high-
resolution picture of the events taking place in CD4+ T lympho-
cytes during their activation.
2 Materials
Prepare all solutions using ultrapure water (MilliQ reverse osmosis

water purificator, 18 MΩ at 25 °C) and electron microscopy-grade
reagents. Prepare and store all reagents at room temperature. Due to
the toxicity of some reagents, for electron microscopy sample process-
ing, it is advisable to pay attention when handling these materials and
to follow all the institutional guidelines regarding waste disposal.
2.1 Sample Fixation 1. Cacodylate (0.2 M): weigh 8.56 g of sodium cacodylate (for
and Resin Embedding the trihydrate form), dilute in 180 ml water, and adjust to
pH 7.4 with 2 N HCl. Make up to 200 ml with water.
2.1.1 Fixation Buffers
2. Cacodylate (0.1 M): dilute 0.2 M cacodylate buffer to double
the volume with water.
3. Electron microscopy: 1% glutaraldehyde, 3% paraformalde-
hyde, and 0.3% tannic acid in 0.1 M cacodylate buffer. Add
0.75 ml 20% PFA, 0.5 ml 10% glutaraldehyde, and 0.015 g
tannic acid to 2.5 mlq of 0.2 M cacodylate buffer pH 7.4.
Make up to 5 ml with water.
2.1.2 Post-fixation 1. 1% osmium tetroxide: add 1.25 ml of 4% aqueous osmium

Solutions tetroxide solution to 2.5 ml of 0.2 M cacodylate buffer pH 7.4.
Make up to 5 ml with water.
2. 1% uranyl acetate: weigh 0.05 g of uranyl acetate and dissolve
in 5 ml of water.
2.1.3 Other Components 1. 47 mm polystyrene petri dishes.

2. Fine curved-tip tweezers.
3. Flat embedding capsules.
4. Anhydrous ethanol.
5. LX112 embedding kit.
6. Liquid nitrogen.
7. Orbital shaker.
8. Pioloform for electron microscopy grid coating.
9. Copper slot grids.
10. 100 mesh copper grids.
11. Coping hand saw.
Ultrastructure of Immune Synapses 401
12. Flat embedding molds.

13. BEEM embedding capsules.
14. SerialEM for control of microscope and image acquisition
(available for download at http://bio3d.colorado.edu/
SerialEM/).
15. IMOD for tomogram processing, segmentation, stitching, and
visualization (available for download at http://bio3d.colo-
rado.edu/imod/).
16. Photoshop (Adobe Systems, Inc. San Jose, CA, USA).
3 Methods
3.1 Processing 1. Medium to high cell coverage on the supported lipid bilayer is
for Electron essential for having a good number of cells for analysis. Fix
Microscopy interacting CD4+ T lymphocytes with 5 ml of electron micros-
and Embedding copy fixation buffer for 1 h RT by using the inlet port of the
FCS Bioptechs flow cell chamber.
2. Carefully wash the fixative (see Note 1) with 20 ml of 0.1 M
cacodylate buffer pH 7.4 by flushing it through the inlet port
of the FCS Bioptechs flow cell chamber.
3. Disassemble the FCS Bioptechs flow cell chamber, and with
the help of the fine curved-tip tweezers, carefully recover the
coverslip with the fixed lymphocytes on the supported lipid
bilayer (Fig. 1). From this point it is very important to always
have the coverslip with the fixed sample facing the operator,
paying attention and being careful when transferring it to the
different reagents. It is also very important to avoid drying the
sample on the coverslip—always be fast when transferring the
coverslip between solutions.
4. Transfer the coverslip to a petri dish containing the osmium
tetroxide solution, and incubate for 1 h RT.
5. Wash three times for 5 min each with 0.1 M cacodylate buffer
pH 7.4.
6. Wash for 5 min RT with ultrapure water.
7. Incubate for 1 h in 1% uranyl acetate solution RT. At this point
the process can be stopped by leaving the petri dish containing
the coverslip in a refrigerator overnight at 4 °C.
8. Wash three times for 5 min each with ultrapure water.
9. Incubate for 30 min in 9:1 (water:ethanol) solution.
402 Jaime Llodrá
Fig. 1 Different steps involved in the processing for electron microscopy of CD4+
T lymphocytes interacting with a supported lipid bilayer built on a glass
coverslip

15. Incubate for 30 min in ethanol.
16. Prepare LX112 resin following the recommendations of the
manufacturer.
17. Incubate for 45 min in 3:1 (ethanol:resin) solution.
20. Incubate for 1 h in resin.
21. Incubate overnight in resin.

22. Next day, incubate in freshly made resin for 1 h RT.
23. Fill a flat BEEM capsule with resin. Invert it and very carefully
place it on top of the area containing the fixed cells, while at
the same time being fast enough to avoid spilling too much
resin. Depending on the cell density, this area can be recog-
nized as a 1–2 mm diameter white spot on the surface of the
coverslip. In case this area is difficult to find, an inverted light
microscope with a phase contrast module can be used to pin-
point the location of the fixed cells on the coverslip. Since the
coverslip is coated by a thin resin layer, it is advisable to main-
tain the inverted BEEM capsule in place by using a set of fish-
ing weights placed on the sides of the capsule. This is done to
prevent the sliding of the capsule, thus keeping the inverted
BEEM capsule directly on top of the fixed cells.
24. Polymerize resin for 3 days at 60 °C.
25. After polymerization, detach the capsule from the coverslip by
adding a little of liquid nitrogen. At this point, the researcher
can try to release the capsule from the glass with his fingers; if
it does not come out, then add more liquid nitrogen. The face
of the polymerized resin block will contain the cells and the
supporting lipid bilayer that were previously on the surface of
the coverslip. If the detaching step is not done carefully, the
researcher can risk damaging the face of the polymerized resin
in the capsule, and in some instances, small glass fragments can
remain attached to the block face. These small fragments can
damage the knife during sectioning. In this case the face should
be dipped again in liquid nitrogen until it is clear and no frag-
ments are present.
At this time the organization of the lymphocyte interacting with
the supported lipid bilayer can be analyzed by cutting cross sections
of the cells on the block face. Alternatively, if the researcher wants to
analyze the full contact area for correlation purposes, he can do this
by cutting en face sections from the block face. For the first approach,
it is necessary to re-embed the cells into a new resin block.
3.2 Resin 1. Using the binocular scope, and focusing on the block face,
Re-embedding carefully check the surface until the fixed cells are found; they
should form a clear circular mark on the block face which cor-
responds to the shape of the supported lipid bilayer.
2. Mark the borders of the sample with a pencil and isolate this
region from the rest of the block with the help of a coping
hand saw.
3. Mark the sides and the bottom of the isolated region with a
pencil so it can easily be found after re-embedding.
404 Jaime Llodrá
Fig. 2 Schematic view of the trough of a diamond knife with floating resin sec-
tions. When viewed with the binoculars of the ultramicrotome, a “pearls on a
string” pattern can be seen that runs parallel to the long axis of the section
4. Carefully place this small fragment in a resin mold and re-

embed in freshly made LX112 resin with the sample side facing
up and polymerize in an oven at 60 °C for 3 days.
5. After polymerization the re-embedded sample is ready to cut.
Find the area where the re-embedded block is, and with the
help of a razor blade, start trimming the excess resin until the
edge of the surface with the fixed cells is found.
6. Cut 150 nm thick resin sections and retrieve them by gently
applying a 100 mesh collodion-coated copper electron micros-
copy grid on the water surface (see Note 2 and Fig. 2).
7. Wait for sections to dry on grids.
8. Stain sections with aqueous 1% uranyl acetate for 4 min.
9. Wash five times for 2 min each with double-distilled water.
10. Stain sections with Sato lead for 4 min.
11. Wash five times for 2 min each with double-distilled water (see
Note 3).
3.3 Correlative For this approach a coverslip with a printed locator grid is needed.
Microscopy The grid can be printed on the coverslip for the Bioptechs FCS by
a method described elsewhere [3]. More recently, a new coverslip
for correlative light and electron microscopy has been developed
between FEI and Ibidi, but it remains to be seen if a proper sup-
ported lipid bilayer can be built using this new type of cell culture
system. Use of supported lipid bilayers in combination with total
internal reflection fluorescence microscopy (TIRFM) to study the
immunological synapse has enabled the analysis of events taking
place at the contact surface up to a height of 200 nm above the
substrate. By careful en face sectioning of the resin-embedded
lymphocytes, it is possible to correlate the structures seen in the
electron microscope with the events recorded in the TIRFM mode.
1. Allow the cells to start interacting with the supported lipid
bilayer built over the area with the printed locator grid.
2. Track formation of the immunological synapse by TIRFM.
3. Once the desired time point is reached, fix cells with 1% para-
formaldehyde in 0.1 M cacodylate buffer pH 7.4.
4. Photograph the cells in the fluorescence, interference reflec-
tion microscopy, and bright field modes to make sure that the
fluorescence registered corresponds to the signals derived from
the region at, or close to, the contact interface (see Note 4).
5. After recording the grid and the cells, process the sample accord-
ing to the previous method until the point where the researcher
ends up with the cells and the grid on the face of a block built
inside of a BEEM capsule (step 10 from previous method).
6. Immobilize the resin block into the specimen holder of the
ultramicrotome.
7. With the help of a razor blade, carefully carve a pyramid on the
resin block. Its face with the shape of a trapezoid will contain
the grid and the interacting cells.
8. Clean the blade of the diamond knife with a cleaning polysty-
rene rod dipped in pure ethanol.
9. Immobilize the resin block in the specimen arm of the ultrami-
crotome and set up the cutting window by moving the trimmed
block face a few millimeters above the knife edge and select the
upper option of the window area of the control panel. Now
move the block face a few millimeters below the knife edge and
select the lower option (see Note 5).
10. Check the knife edge for horizontal and vertical alignment
with respect to the block face. For this procedure, use only
the bottom light source of the ultramicrotome; it will illumi-
nate the block face. Move the knife to the block face; as it
gets closer, it should cast a shadow on the block face that
appears as a thin white line. A thin horizontal shadow signals
perfect alignment. If that is not the case then rotate the knife
stage until the shadow becomes horizontal. For vertical
alignment move the block vertically along the arc of the
specimen holder until the shadow of the knife is the same
thickness when moving the block face up and down, mean-
ing that the block face is evenly separated from the knife
during the cutting cycle.
11. Since the TIRF signal will be correlated, only the first two or
three 50 nm thick en face sections will be collected on the
collodion-coated electron microscopy grids (Fig. 3).
12. Stain collected sections according to previous protocol.
13. Coat grids with a fine layer of evaporated carbon to increase
stability under the electron beam.
14. Take pictures of the cells (Fig. 3).
15. Look for the correlative fluorescence image based on the cell
shape and taking the correlation grid as a reference (see Note 6).
Fig. 3 (a) En face resin section showing the squares of the grid printed on a glass coverslip. The red frame
points to the same area as that marked in (b). (b) Same grid as in (a) but showing CD4+ T lymphocytes inter-
acting with the supported lipid bilayer built over the area containing the printed grid. The red frame marks the
same area as in (a). (c) Higher magnification view of the red frame in (b) showing CD4+ T lymphocytes at
the immunological synapse forming stage. T Cell receptor is labeled in red, and cytoplasm is labeled with a
probe for actin. (d) and (e) Electron and fluorescence microscopy views, respectively, of the same T lympho-
cyte. (f) Overlay of electron and fluorescence microscopy images
16. Crop the fluorescence image of the cell of interest and increase
the image size until it is comparable to the size of the cell in the
electron microscopy image. To do this, the pixel size of the
fluorescence microscopy image has to be divided by the pixel
size of the electron microscopy image. Use the resulting factor
to scale the fluorescence microscopy image so both images
have a similar size for correlation.
17. Using Photoshop, overlay the fluorescence image on top of
the electron microscopy image. Use the transform command
to rotate and translate the image until an accurate fit is found
(Fig. 3).
4 Notes
1. Use of tannic acid is essential to avoid the collapse of the lym-

phocyte over the supported lipid bilayer, thus preserving the
spacing and allowing for contact surface analysis (see ref. 4).
2. By using the binoculars of the ultramicrotome to look at the
sections floating on the water surface, it is possible to see a
“pearls on a string” figure which runs along the horizontal axis
of the section (Fig. 2). The pearls correspond to the T lympho-
cytes while the string corresponds to the supported lipid bilayer.
3. At this point, sections are ready to be analyzed with the elec-
tron microscope. Additional stability of the sections under the
electron beam can be achieved by applying a thin layer of car-
bon on top of the stained sections.
4. It is important to also record the printed grid so it can be used
to properly locate the cells once the resin sections are analyzed
with the electron microscope.
5. At this stage, the operator should take care to have the block
face at a safe distance from the knife edge to avoid accidental
collisions.
6. In order to make an accurate correlation, it is advisable to use
a cytoplasmic fluorescent label that could highlight the cell
edges with as much detail as possible.
References
1. Dustin ML, Starr T, Varma R, Thomas VK DL, Dustin ML (2014) Polarized release of
(2007) Supported planar bilayers for study of T-cell-
receptor-enriched microvesicles at the
the immunological synapse. Curr Protoc immunological synapse. Nature 507:118–123
Immunol 18:13 4. Milstein O, Tseng SY, Starr T, Llodra J, Nans
2. Grakoui A, Bromley SK, Sumen C, Davis MM, A, Liu M, Wild MK, van der Merwe PA, Stokes
Shaw AS, Allen PM, Dustin ML (1999) The DL, Reisner Y, Dustin ML (2008) Nanoscale
immunological synapse: a molecular machine con- increases in CD2-CD48-mediated intermem-
trolling T cell activation. Science 285:221–227 brane spacing decrease adhesion and reorga-
3. Choudhuri K, Llodra J, Roth EW, Tsai J, nize the immunological synapse. J Biol Chem
Gordo S, Wucherpfennig KW, Kam LC, Stokes 283:34414–34422
Chapter 25
Systems Imaging of the Immune Synapse

Rachel Ambler, Xiangtao Ruan, Robert F. Murphy,
and Christoph Wülfing
Abstract
Three-dimensional live cell imaging of the interaction of T cells with antigen-presenting cells (APCs) visu-
alizes the subcellular distributions of signaling intermediates during T cell activation at thousands of
resolved positions within a cell. These information-rich maps of local protein concentrations are a valuable
resource in understanding T cell signaling. Here, we describe a protocol for the efficient acquisition of
such imaging data and their computational processing to create four-dimensional maps of local concentra-
tions. This protocol allows quantitative analysis of T cell signaling as it occurs inside live cells with resolu-
tion in time and space across thousands of cells.
Key words Live cell imaging, Computational image analysis, Spatiotemporal distributions,
Immunological synapse, T cell activation
1 Introduction
T cell activation is regulated by the complex interactions of dozens

of signaling intermediates. One of the great, general challenges in
current biology is to understand how dozens of proteins function as
an integrated system. As a key resource to elucidate such complexity,
signaling intermediates are not evenly distributed through an acti-
vating T cell but enrich at particular locations at distinct times [1, 2].
Concurrent enrichment of two proteins increases their interaction
probability and thus the local efficiency of the signaling step medi-
ated by their interaction. Thus at the system scale, uneven signaling
distributions govern the information flow through signaling net-
works [3–6]. Supporting the importance of subcellular location for
function in T cell activation, loss-of-function mutations of various
signaling intermediates consistently yield diminished localization
Rachel Ambler and Xiangtao Ruan are shared first authors.

Robert F. Murphy and Christoph Wülfing are shared senior authors.
409
410 Rachel Ambler et al.
[7–11]. In addition, through the imaging of large numbers of sig-

naling intermediates, subcellular structures mediating signal integra-
tion have emerged [12, 13]. Thus a quantitative determination of
subcellular signaling distributions of a large number of T cell signal-
ing intermediates is a powerful resource to discover functional sig-
naling connectivity. Here we describe the methodology for such
characterization across thousands of cells. Three-dimensional live
cell imaging of the interaction of T cells with antigen-presenting
cells (APCs) visualizes the subcellular distributions of signaling
intermediates at thousands of resolved positions within a cell, i.e., it
yields detailed three-dimensional maps of local concentrations. We
describe a protocol for efficient acquisition of such data. Voxel-by-
voxel quantification of such imaging data provides full access to the
three-dimensional maps of local concentrations. We describe a com-
putational analysis routine to generate maps that are aligned across
all T cells to be analyzed. Such maps form a powerful foundation for
the computational analysis of signaling connectivity and the model-
ing of T cell signal transduction.
2 Materials
2.1 Retroviral 1. Phoenix-E (φNX-E) cell line (obtained from Nolan laboratory,
Transduction Reagents Stanford University).
2. φNX-E incomplete medium: DMEM with 4.5% glucose, 110
mg/L sodium pyruvate and l-glutamine, 10% FBS, 100 U/mL
penicillin, 100 μg/mL streptomycin, MEM nonessential amino
acids.
3. φNX-E complete medium: As item 2, in Subheading 2.1, sup-
plemented with 300 μg/mL hygromycin and 1 μg/mL diph-
theria toxin.
4. Corning Primaria tissue culture plates 35 mm or equivalent.
5. 0.02% EDTA solution in PBS.
6. Chloroquine diphosphate: 4.1 mg/mL chloroquine diphos-

phate solution in ddH2O and sterile filtered (e.g., Sigma).
7. 2× Hepes buffered saline: 16 g/L NaCl, 0.74 g/L KCl, 0.27
g/L NA2HPO4, 2 g/L D-glucose, 10 g/L HEPES. Dissolve in
ultrapure water, adjust pH to 7.05, and sterile filter.
8. Calcium chloride: 2 M solution of calcium chloride in

ddH2O. Sterile filtered.
9. Protamine sulfate: dissolved in PBS at 8 mg/mL.
2.2 T Cell Culture 1. Complete medium: RPMI with l-glutamine, 10% FBS, 100 U/
Components mL penicillin, 100 μg/mL streptomycin, 50 μM
2-mercaptoethanol.
Systems Imaging of the Immune Synapse 411
2. Interleukin-2 medium: complete medium supplemented with

50 U/mL rhIL-2.
3. Agonist peptide for the TCR transgene (moth cytochrome C
for the 5C.C7 TCR): 5 mM solution in ultrapure water.
2.3 Cell Sorting 1. Imaging buffer: PBS without calcium and magnesium, 10%

and Imaging FBS, 1 mM calcium chloride, 500 μM magnesium chloride.
Components 2. Glass-bottomed 348-well imaging plate (Brooks Life Science
Systems) or equivalent.
3 Methods
3.1 Generation 1. Stock plates of φNX-E cells are maintained in φNX-E complete
of Retrovirus medium in a 37 °C, 6% CO2 incubator. When splitting cells,
aspirate medium and add 1 mL 0.02% EDTA per plate. Incubate
cells at 37 °C for approximately 3 min (see Note 1). Following
incubation, tap each plate firmly with a finger to fully loosen the
cells. Use 5 mL DMEM to collect cells from the plate and trans-
fer to a Falcon tube. Centrifuge for 3 min at 200 × g, resuspend
in 1 mL of φNX-E medium, and count cells using a hemocy-
tometer. Plates 600,000 cells into a fresh Falcon Primaria tissue
culture plate in a total volume of 6 mL medium. φNX-E com-
plete medium should be used for general maintenance (see Note
2), whereas incomplete medium should be used when cells will
undergo transduction.
2. On day 3 of culture in incomplete medium, φNX-E cells are
ready for transduction via calcium phosphate precipitation.
Make a fresh solution of chloroquine diphosphate and add 25
μL to each plate. Swirl gently to mix and return cells to
incubator.
3. For each plasmid, label two 15 mL Falcon tubes as tube A and
tube B. In tube A mix 0.5 mL 2× HBS with 2 μL 1 M NaOH
(see Note 3). In tube B mix 0.5 mL sterile ddH20 with 10 μg
plasmid DNA. Add 62 μL 2 M CaCl2 drop by drop to tube B.
4. Using a 1 mL tissue culture pipette, bubble air vigorously

through tube A while adding the contents of tube B in a drop-
by-drop fashion. Let sit at room temperature for a minute.
5. Gently add the entirety of this solution to the φNX-E cells,
swirling gently to aid distribution. Return cells to the incubator
overnight.
6. The following day check for the presence of the calcium
phosphate precipitate as black puncta much smaller than
the φNX-E cells. Gently aspirate the medium from the
φNX-E cells and replace with 3.2 mL incomplete medium
(see Note 4). Incubate cells, without any medium changes

or splitting, for approximately another 48 h. At this time
the medium will contain retrovirus carrying the DNA con-
struct of interest.
3.2 Culture of T Cells 1. Cull a T cell receptor transgenic mouse over 6 weeks of age
from Lymph Nodes using a schedule one technique (see Note 5).
2. Dissect out lymph nodes (see Note 6). Collect them into a tube
containing 5 mL of complete medium.
3. Place a 40 μm cell strainer over a 50 mL Falcon tube. Pour the
lymph nodes into the strainer and gently disrupt them using the
plunger of a 1 mL syringe. Wash the strainer with 5 mL com-
plete medium, and repeat the disruption process until the
medium passing through the strainer appears clear. Centrifuge
the Falcon tube for 3 min at 200 × g.
4. Discard supernatant and thoroughly resuspend the cell pellet in
1 mL complete medium. Count live cells and add complete
medium to the cells to bring a density to 4 × 106 cells/mL.
5. In a 24-well plate, add 0.65 μL of 5 mM agonist peptide to the
bottom of each well you wish to use. Add 1 mL of cell suspen-
sion to each well, so that each contains 4 × 106 cells. Incubate
at 37 °C, 6% CO2 overnight.
3.3 Retroviral 1. After T cells have been activated overnight, they are ready to be
Transduction of T Cells retrovirally transduced. Collect each well into a separate 15 mL
Falcon tube. If more than one well of cells will be transduced
with the same construct, they may be pooled.
2. Collect the medium from the φNX-E cell plate into a separate
15 mL Falcon tube. Gently add 1 mL PBS to the φNX-E cells
to prevent dehydration and set aside. Centrifuge both sets of
tubes at 200 × g for 3 min.
3. Discard the T cell supernatant. Take 2 mL φNX-E cell medium
supernatant and use this to resuspend the T cells. Add 2 μL of
8 mg/mL protamine sulfate to the bottom of a 24-well plate.
Add the 2 mL of cell suspension to this well.
4. Centrifuge the 24-well plate at 200 × g, 32 °C, for 2 hours (see
Note 7).
5. In parallel on a wide-field fluorescent microscope, use a 10×
objective lens to focus on the φNX-E cells left in the plate.
With an appropriate laser, determine whether the φNX-E cells
are expressing the fluorescent protein construct. If not, it is
likely that the calcium phosphate precipitation has failed (see
Note 8).
6. After centrifugation aspirate the medium from each well of T
cells, being careful not to disrupt the cells at the bottom of the
well. Resuspend T cells in 2 mL IL-2 medium.
7. Incubate cells for a further 3–4 days to allow expression of the

transduced protein and to expand cell numbers. On a daily
basis, check the cell density. Split each well into two when the
cells become fully confluent and maintain in IL-2 medium.
3.4 Fluorescent- 1. On the third or fourth day of culture following transduction,

Activated Cell Sorting collect the T cells into a Falcon tube, and centrifuge at 200 × g
of Successfully for 3 min. Discard the supernatant and resuspend the cell pellet
Transduced T Cells in150μL imaging buffer per well of cells.
2. Prepare a collection tube by adding 1.5 mL IL-2 medium to a
6 mL FACS tube.
3. During FACS, use the following gating strategy to select for
positively transduced lymphocytes (Fig. 1):
Select live lymphocytes based on forward scatter versus
(a)
side scatter.
Select for singlets based on trigger pulse width versus side
(b)
scatter.
Create a sorting gate to select for GFP-positive cells, which
(c)
range from 1 to 1.5 log shifts brighter than the negatives
(Fig. 1) (see Note 9).
4. Collect positive cells in multiples of 4 × 104 cells (see Note 10).
5. Return sorted cells to a 24-well plate in the incubator until
ready for imaging. Image the same day as sorting to minimize
loss of fluorescence (see Note 11).
3.5 Live Cell Imaging 1. For the 5C.C7 TCR transgene, CH27 cells are used as APCs
of T Cell/APC Couple and cultured in complete medium (see Note 12). To peptide
Formation load APCs, collect approximately 1 × 106 cells into 1 mL of
complete medium and add to a 24-well plate. Add 2 μL of 5
mM MCC peptide and mix. Incubate for a minimum of 4 h.
Fig. 1 Gating strategy to sort GFP-positive live lymphocytes. CL4 T cells were retrovirally transduced to express
chronophin-GFP. (1) The side scatter (SSC) versus forward scatter (FCS) gate to select live cells is given in red.
(2) The trigger pulse width versus forward scatter (FCS) gate to select singlets is given in green. (3) A GFP-
positive sorting gate (green gate) is created around the GFP-positive population which lies between 1 and 1.5
log shifts above the top of the GFP-negative population (purple gate)
2. Meanwhile, ensure a spinning disk microscope fitted with an

environmental chamber and a 40× oil objective lens seated on a
piezo motor is fully preheated to 37 °C (see Note 13). Also
preheat a glass-bottomed 348-well imaging plate.
3. Collect the sorted T cells into a microcentrifuge tube. Collect
300 μL of the peptide-loaded APCs into a separate microcentri-
fuge tube. Centrifuge both at 300 × g for 5 min in a tabletop
centrifuge.
4. Carefully remove the supernatant from the T cell pellet.

Resuspend the pellet in 5 μL of preheated 37 °C imaging buffer
per 4 × 104 cells.
5. Repeat with the APCs, resuspending in 50 μL of preheated
imaging buffer.
6. Add 50 μL imaging buffer to an empty well on the imaging
plate. Add 5 μL of T cells to this well and fit onto the micro-
scope. Allow the plate temperature to equilibrate for 5 min.
7. Meanwhile set up the microscope so that a z-stack consisting of
21 images 1 μm apart is taken at three time points per minute
in the fluorescent channel. A single differential interference
contrast (DIC) reference image should be taken mid-stack at
each time point. Images should be taken for 15 min (see Note
14). We use 2 × 2 binning to improve the signal-to-noise ratio
of the imaging data (see Note 15).
8. Once the plate has equilibrated, ensure the T cells are in focus.
Firmly flick the microcentrifuge tube containing the APCs to
generate a single cell suspension. Gently add 5 μL APCs to the
well so that the layer of T cells is not disturbed. Immediately
begin monitoring the well through the eyepiece, and wait for
the APCs to begin settling among the T cells. Select a field of
view with a good distribution of T cells and APCs and begin
imaging (see Note 16).
9. Once imaging is completed, separate the images into two dif-
ferent channels: DIC and fluorophore. Split the fluorescent
channel into separate folders for each time point, with each
folder containing the 21 z-stack images. Export all files as TIFFs
(see Note 17).
3.6 Annotation 1. For each movie, prepare an annotation file identifying the posi-
of Cell Couples tions of the synapse between each T cell and APC couple as
and Synapse Positions described below. First create a z-stack maximum projection for
each time point, and then stack all the maximum projections so
they can be scrolled through as a function of time. This can be
accomplished with any of a number of basic image processing
programs, such as ImageJ.
2. Remove any autoscaling from the images (see Note 18). Apply
a pseudocolor lookup table to the fluorescence data to make
differences in sensor intensity more readily apparent.
3. Use the DIC reference images to identify cell couples. Scroll

through the time series until a T cell is observed forming an
immune synapse with an APC, defined as a broad membrane
interface between the T cells and APCs. Fluorescence images
should only be used to confirm the identity of cells as either T
cells or APCs (see Note 19). Start at the frame where initial
contact between the T cell and APC is made and count forward
a further two frames. Within these three frames, select the one
where the membrane interface first reaches its widest point.
Classify this frame as time point 0 or the earliest point at which
the immune synapse is fully formed. For as many frames as
desired, record the coordinates for the two ends of the immune
synapse and the frame number so that the cell couple can be
found again for the automated analysis below (Fig. 2).
4. These coordinates should be saved into a spreadsheet either
directly from the image-processing package or manually. The
spreadsheet should have a header row as shown in Fig. 3 and
then sets of rows with one set per cell couple and one row
within the set for each time point (frame) for that couple. The
sets should be separated by a blank row. In the row for each
frame, put:
In column 1, the name of the image file (depending on
(a)
how the images are acquired, there may be one file per
time point or multiple time points in a single file). The
filename may include an absolute path or a path relative to
the location containing the annotation file.
In column 2, the number of the channel within that file
(b)
that contains the GFP fluorescence for that time point (if
each time point is in a separate file, this would typically be
channel 1; if multiple time points are in the same file, this
would typically be the frame number).
In columns 3–6, the X coordinate of the left end point,
(c)
the Y coordinate of the left end point, the X coordinate of
Fig. 2 Cell coupling and interface annotation. The coupling of an ezrin-GFP-transduced 5C.C7 T cell with a
CH27 APC in the presence of 10 μM MCC agonist peptide is shown as a green transparent overlay of a maxi-
mum projection of the three-dimensional ezrin-GFP fluorescence data onto DIC bright-field images. Time is
given in minutes relative to tight cell coupling. Black lines indicate the T cell/APC interface, and the coordinates
of the two ends of these lines are recorded for use in the computational analysis
Fig. 3 Format of T cell/APC couple annotations for entry into the computational image analysis pipeline. A
representative spreadsheet is given
the right end point, and the Y coordinate of the right end
point for the synapse in that time point.
In column 7, the time difference for that frame relative to
(d)
time point 0 (Fig. 3).
5. Save the spreadsheet as a CSV (comma-separated value) file
with a name that is specific (matched) to the movie file name.
3.7 Building Models 1. Install MATLAB if it is not already installed. Check http://cel-
of Four-Dimensional lorganizer.org to find out which versions of MATLAB are cur-
Protein Distributions rently supported by CellOrganizer. Note that CellOrganizer is
with CellOrganizer only supported for Mac OS and Linux systems.
(See Note 20) 2. Download the latest version of CellOrganizer from http://cel-
lorganizer.org/Downloads, and extract it to some desired
directory (i.e., a CellOrganizer directory (folder) in your home
directory). The instructions in this chapter apply to
CellOrganizer 2.6. You should check the documentation for
future versions to see if any relevant changes have been made.
3. Launch MATLAB, and use the “cd” command to change your
current working directory to the directory in which you put
CellOrganizer. Type “setup” to initialize the environment for
CellOrganizer.
4. If you want to test that the installation has occurred correctly,
type “demo3Dtcell_train” which will process a small set of
movies downloaded from the CellOrganizer website.

5. Type “copydemo(“demo3Dtcell_train”)” and when prompted
enter a name for the script it will create (e.g., “cofilin_analysis_
May1”).
6. Type “edit” followed by a space and the name of the file you
just created. Change the specification of the annotation file to
match the movies that you want to analyze. For example, if
you have only one annotation file called abc.csv, change
“synapse_location=‘annotation/*.csv’” to “synapse_
location=‘abc.csv’”. If you have more than one annotation file,
they are in directory “mymovies,” and if you want to process all
of them, change “synapse_location=‘annovation/*.csv’” to
“synapse_location=‘mymovies/*.csv’”.
7. Change the path for where the results should be saved to the
desired location by changing the value of “results_location”.
8. By default, the output models will have the same name as the
base name of the first annotation file with the number of the
time point appended to it. For example, if the first annotation
file is “abc.csv”, the model for the first time point will be “abc_
reltime_1.mat”. Also by default a model will be created for each
time point in the annotation file. If a different name for the
model is desired, set “options.model_prefix” to that name. If
creation of models for only a subset of the time points is desired,
edit “options.timepoints_to_include” to specify which time
points to include (as a MATLAB vector). For example, to ana-
lyze time points 1 to 7, set it to “[1:7]”, or to analyze just time
points 6 and 10, set it to “[6, 10]”.
9. Save the script and run it (i.e., click the Run button) and wait
for the script to finish the analysis.
3.8 Analysis Models 1. To generate a figure showing slices through the 3D map for
and Creating Figures each time point, use the function “ShowTcellModelMap”. To
(See Note 21) do so, include the full or relative path of the model file as the
argument of the function, e.g., “ShowTcellModelMap(‘/path/
to/model/your_model_reltime_1.mat’)” (a full path) or
“ShowTcellModelMap(‘mymodels/your_model_reltime_1.
mat’)” (a path relative to the current working directory).
2. To generate a movie showing protein intensity change through
time points, use the function “GenerateTcellMovie”. The func-
tion can show up to three proteins in different colors, and it can
therefore take up to three arguments. As before, each argument
is a path to one of the models, e.g., “GenerateTcellMovie(‘models/
protein1/*.mat’,‘models/protein2/*.mat)” would create a
two-color movie.
3. To compare the intensities of each voxel for different models,
use the function “CompareTcellModels”. The function needs
two paths, one for each model, e.g., “CompareTcellModels(‘/
path/to/models/model1.mat’, ‘/path/to/models/model2.
mat’)”.
4. To show how much of the protein is enriched in the synapse
region at various time points, use the function
“ShowTcellEnrichment”. The input argument of the function
is the filenames of the models to be compared, i.e.,
“ShowTcellEnrichment(‘/path/to/model/*.mat’)”.
4 Notes
1. The use of EDTA alone rather than the usual Trypsin/EDTA

combination and limiting the time of EDTA on the φNX-E
cells optimizes φNX-E cell health.
2. We find that maintaining the parental φNX-E stock under
continuous selection ensures consistently high viral titers over
months. Nevertheless, the stock is replaced from freezes about
every 6–9 months.
3. For each batch of HBS, the exact amount of NaOH to be
added is determined in a dose-response series of NaOH. The
NaOH amount giving the smallest visible calcium phosphate
precipitate is chosen.
4. By reducing the medium volume from 6 to 3.2 ml, the con-
centration of virus in the supernatant is effectively doubled.
5. Our most widely used TCR transgene is the 5C.C7
TCR. However, this protocol has worked equally well with a
number of other TCR transgenic mouse strains, including
DO11.10, OTII, AND, Tg4, P14, CL4, and HY [14–18].
6. We commonly use combined inguinal, axillary, submandibu-
lar, and mesenteric lymph nodes. Spleen can be used instead
yet requires red blood cell lysis.
7. The centrifuge is heated by air friction. Therefore, pre-spin
until 32 °C is reached.
8. The LLMV long terminal repeat functions as a very strong
promoter in the φNX-E cells, and resulting viral mRNA is effi-
ciently translated. The presence of imaging sensor in the
φNX-E cells thus is an indirect yet convenient readout of the
amount of viral mRNA produced.
9. The 1–1.5 log range above cellular autofluorescence back-
ground corresponds to the minimal fluorescence level detect-
able by sensitive microscope systems, a GFP concentration of
2.5 μM [18]. The use of minimal sensor concentrations opti-
mizes physiological relevance of the imaging data by minimiz-
ing interference of the imaging sensor with the signaling event
to be studied. Using this sorting strategy across a range of
actin regulators, the combined concentration of endogenous
protein plus its retrovirally expressed GFP-tagged version as
an imaging sensor was commonly within the 5–95 percentile
of the endogenous protein concentration [19].
10. Transduction efficiencies vary as a function of (I) the TCR
transgene and (II) the protein to be expressed. (I) The more
vigorous the initial T cell proliferation in tissue culture is, the
higher the transduction efficiency becomes. Therefore MHCI-
restricted CD8 TCR transgenes generally yield higher efficien-
cies that can easily exceed 50%. (II) The expression of a sensor
is tightly linked to the endogenous concentration of the pro-
tein it is derived from. Thus, actin-GFP expresses substantially
better than, e.g., Itk-GFP. Sensors based on synthetic con-
structs or isolated protein domains, such as the F-tractin pep-
tide of the PLCδ-PH domain, generally express well.
Transduction efficiencies as low as 0.01% can be used for sub-
sequent imaging.
11. We find that homogeneously high sensor expression at the
time of sorting becomes variable already on the time scale of
days.
12. We have used a wide variety of APCs, from transfected CHO
cells, various B and T cell lymphoma lines, and dendritic cells
to tumor target cells. In general, non-adherent cells allow for
easier detection of cell coupling as the T cell/APC interface
forms perpendicular to the cover slip and thus is readily detect-
able in the DIC bright-field images.
13. We have used magnifications from 20× to 100×. Higher mag-
nifications trade better resolution for dimmer images and
lower number of cells and thus cell couples in the field.
Therefore, the majority of our experiments are done with 40×
magnification. For cell coupling efficiencies below 10%, even
lower magnification is desirable to capture an adequate num-
ber of cell couples.
14. Biochemically detectable T cell signaling activity in T cell/
APC couples peaks in the first 5 min of cell coupling [20,
21]. NFAT and NFκB translocate into the nucleus within
about 3 and 7 min, respectively [13, 18]. Therefore, 15 min
imaging times captures the peak of T cell signaling activity.
Nevertheless, we have imaged as long as 16 h. Any imaging
beyond 30 min will require measures to minimize buffer
evaporation.
15. Binning of camera pixels trades resolution in x and y against
signal-to-noise ratio. With the chosen 2 × 2 binning on the
40× objective, resolution is roughly equivalent in all three
dimensions.
16. In response to strong stimulus, e.g., a high concentration of
agonist peptide in the presence of costimulation, cell coupling
can be virtually instantaneous. As cell couples that form before
the onset of data acquisition cannot be accurately timed rela-
tive to the time of cell couple formation and thus have to be
excluded from analysis, a rapid decision on the imaging field is
essential.
17. Data can be saved in any format. However, archiving images as
TIFF files ensure great compatibility with various analysis
packages and ready exchange with colleagues.
18. Autoscaling, the process where the dimmest pixel in an image

is set to black and the brightest one is set to white, is helpful
in image acquisition to get an immediate first impression of
the data. However in image analysis, in particular in the com-
parison of images across a time series, we display all images on
the same scale, as thus apparent differences in brightness
reflect actual differences in fluorescence intensity rather than
potential differences in scaling.
19. We don’t use fluorescence data in the determination of cell
coupling. This would set up a circular argument where fluo-
rescence data would be used to determine time of cell cou-
pling to then be analyzed relative to this time.
20. When imaging data are acquired with a 40× objective and 2 ×
2 binning, each T cell is represented by >5000 voxels. This
paragraph describes how to generate voxel-by-voxel-resolved
models of these distributions. Importantly, the MATLAB
script creates these models in a cell shape-normalized fashion
such that each position in one T cell model can directly be
compared to an equivalent position in any other model [19].
21. Once cell shape-normalized models of the three-dimensional
protein distributions have been generated, they can be com-
putationally processed in whatever form desired. In this para-
graph, we list four examples for visualization and quantitative
analysis.
Acknowledgments
The original research upon which these protocols are based was
supported in part by the National Institutes of Health grant P41
GM103712, by National Science Foundation grants MCB1121793
and MCB1121919, and by the European Research Council grant
PCIG11-GA-2012-321554.
References
1. Monks CR, Kupfer H, Tamir I, Barlow A, 4. Angermann BR, Klauschen F, Garcia AD,
Kupfer A (1997) Selective modulation of pro- Prustel T, Zhang F, Germain RN, Meier-
tein kinase C-theta during T-cell activation. Schellersheim M (2012) Computational mod-
Nature 385(6611):83–86 eling of cellular signaling processes embedded
2. Grakoui A, Bromley SK, Sumen C, Davis MM, into dynamic spatial contexts. Nat Methods
Shaw AS, Allen PM, Dustin ML (1999) The 9(3):283–289. doi:10.1038/nmeth.1861
immunological synapse: a molecular machinery 5. Purvis JE, Lahav G (2013) Encoding and
controlling T cell activation. Science decoding cellular information through signal-
285:221–226 ing dynamics. Cell 152(5):945–956.
3. Schmick M, Bastiaens PI (2014) The interde- doi:10.1016/j.cell.2013.02.005
pendence of membrane shape and cellular sig- 6. Mast FD, Ratushny AV, Aitchison JD (2014)
nal processing. Cell 156(6):1132–1138. Systems cell biology. J Cell Biol 206(6):695–
doi:10.1016/j.cell.2014.02.007 706. doi:10.1083/jcb.201405027
7. DeFord-Watts LM, Dougall DS, Belkaya S, lamellal actin network. PLoS One
Johnson BA, Eitson JL, Roybal KT, Barylko B, 10(8):e0133299. doi:10.1371/journal.pone.
Albanesi JP, Wulfing C, van Oers NS (2011) 0133299
The CD3 zeta subunit contains a 14. Purtic B, Pitcher LA, van Oers NS, Wülfing C
phosphoinositide- binding motif that is (2005) T cell receptor (TCR) clustering in the
required for the stable accumulation of immunological synapse integrates TCR and
TCR-CD3 complex at the immunological syn- costimulatory signaling in selected T cells. Proc
apse. J Immunol 186(12):6839–6847. Natl Acad Sci U S A 102(8):2904–2909
doi:10.4049/jimmunol.1002721 15. Sinai P, Dozmorov IM, Song R, Schwartzberg
8. Ksionda O, Saveliev A, Kochl R, Rapley J, PL, Wakeland EK, Wulfing C (2014) T/B-cell
Faroudi M, Smith-Garvin JE, Wulfing C, interactions are more transient in response to
Rittinger K, Carter T, Tybulewicz VL (2012) weak stimuli in SLE-prone mice. Eur
Mechanism and function of Vav1 localisation J Immunol 44(12):3522–3531. doi:10.1002/
in TCR signalling. J Cell Sci 125(Pt 22):5302– eji.201444602
5314. doi:10.1242/jcs.105148 16. Sinai P, Nguyen C, Schatzle JD, Wülfing C
9. Liang Y, Cucchetti M, Roncagalli R, Yokosuka (2010) Transience in polarization of cytolytic
T, Malzac A, Bertosio E, Imbert J, Nijman IJ, effectors is required for efficient killing and
Suchanek M, Saito T, Wulfing C, Malissen B, controlled by Cdc42. Proc Natl Acad Sci U S A
Malissen M (2013) The lymphoid lineage- 107(26):11912–11917. doi:10.1073/
specific actin-uncapping protein Rltpr is essen- pnas.0913422107
tial for costimulation via CD28 and the 17. Singleton KL, Parvaze N, Dama KR, Chen KS,
development of regulatory T cells. Nat Jennings P, Purtic B, Sjaastad MD, Gilpin C,
Immunol 14(8):858–866. doi:10.1038/ Davis MM, Wülfing C (2006) A large T cell
ni.2634 invagination with CD2 enrichment resets
10. Paster W, Brockmeyer C, Fu G, Simister PC, de receptor engagement in the immunological
Wet B, Martinez-Riano A, Hoerter JA, Feller synapse. J Immunol 177(7):4402–4413
SM, Wulfing C, Gascoigne NR, Acuto O 18. Singleton KL, Roybal KT, Sun Y, Fu G,
(2013) GRB2-mediated recruitment of Gascoigne NR, van Oers NS, Wülfing C (2009)
THEMIS to LAT is essential for thymocyte Spatiotemporal patterning during T cell activa-
development. J Immunol 190(7):3749–3756. tion is highly diverse. Sci Signal 2(65):ra15
doi:10.4049/jimmunol.1203389 19. Roybal KT, Buck TE, Ruan X, Cho BH, Clark
11. Singleton KL, Gosh M, Dandekar RD, DJ, Ambler R, Tunbridge HM, Zhang J,
Au-Yeung BB, Ksionda O, Tybulewicz VL, Verkade P, Wulfing C, Murphy RF (2016)
Altman A, Fowell DJ, Wülfing C (2011) Itk Computational spatiotemporal analysis identi-
controls the spatiotemporal organization of T fies WAVE2 and cofilin as joint regulators of
cell activation. Sci Signal 4(193):ra66. costimulation-mediated T cell actin dynamics.
doi:10.1126/scisignal.2001821 Sci Signal 9(424):rs3. doi:10.1126/scisignal.
12. Roybal KT, Sinai P, Verkade P, Murphy RF, aad4149
Wülfing C (2013) The actin-driven spatiotem- 20. Negulescu PA, Krasieva TB, Khan A,
poral organization of T-cell signaling at the sys- Kerschbaum HH, Cahalan MD (1996) Polarity
tem scale. Immunol Rev 256(1):133–147. of T cell shape, motility, and sensitivity to anti-
doi:10.1111/imr.12103 gen. Immunity 4(5):421–430
13. Roybal KT, Mace EM, Mantell JM, Verkade P, 21. Mustelin T, Tasken K (2003) Positive and neg-
Orange JS, Wulfing C (2015) Early signaling in ative regulation of T-cell activation through
primary T cells activated by antigen presenting kinases and phosphatases. Biochem J 371(Pt
cells is associated with a deep and transient 1):15–27
Chapter 26
Comprehensive Analysis of Immunological Synapse

Phenotypes Using Supported Lipid Bilayers
Salvatore Valvo, Viveka Mayya, Elena Seraia, Jehan Afrose,
Hila Novak-Kotzer, Daniel Ebner, and Michael L. Dustin
Abstract
Supported lipid bilayers (SLB) formed on glass substrates have been a useful tool for study of immune cell
signaling since the early 1980s. The mobility of lipid-anchored proteins in the system, first described for
antibodies binding to synthetic phospholipid head groups, allows for the measurement of two-dimensional
binding reactions and signaling processes in a single imaging plane over time or for fixed samples. The
fragility of SLB and the challenges of building and validating individual substrates limit most experimenters
to ~10 samples per day, perhaps increasing this few-fold when examining fixed samples. Successful experi-
ments might then require further days to fully analyze. We present methods for automation of many steps
in SLB formation, imaging in 96-well glass bottom plates, and analysis that enables >100-fold increase in
throughput for fixed samples and wide-field fluorescence. This increased throughput will allow better
coverage of relevant parameters and more comprehensive analysis of aspects of the immunological synapse
that are well reconstituted by SLB.
Key words Immunological synapse, Supported lipid bilayers, High-throughput screening, Image
analysis, Costimulation, Signaling
1 Introduction
Many steps in immune cell communication require direct cell-cell

contact for molecular recognition of infection [1]. Directed secre-
tion into the interface between the cells then elaborates on this
communication while maintaining the specificity of the initial rec-
ognition [2]. These specialized cell-cell junctions have been referred
to as immunological synapses based on in vitro prototypes [3–5].
The in vitro prototypes share features [6]: (1) cells are separated by
a short, variable intermembrane gap, (2) the interaction is mediated
by bona fide adhesion molecules, (3) the junction is stable in the
*Salvatore Valvo and Viveka Mayya are shared first authors. Daniel Ebner and Michael L Dustin are shared
last authors.
423
424 Salvatore Valvo et al.
time frame needed for function, and (4) there is polarized secretion.
The historical prototypes for the immunological synapse are cyto-
toxic T cells, natural killer cells, and helper T cells [7–11]. However,
the underlying principles are being applied to other immune system
cells that use immunotyrosine activation motif (ITAM)-bearing
receptors [12–14].
The efforts to more precisely define immunological synapses
emerged with data from Kupfer and colleagues that revealed for-
mation of a central cluster of TCR and PKC-θ surrounded by a
ring of LFA-1 and talin that was associated with specific recogni-
tion by helper T cells [15, 16]. Kupfer referred to these compo-
nents of the junction as supramolecular activation clusters
(SMACs) [15], whereas it was proposed in parallel that this pat-
tern represents a functional “immunological synapse” and could
be defined as such [5]. Remarkably, studies with supported lipid
bilayers (SLB) containing ligands for TCR and LFA-1 recapitu-
late this pattern [17], and adding CD80 to the substrate fully
reconstitutes strong PKC-θ recruitment to the central TCR clus-
ter [18, 19]. The advantage of SLB is that they immediately
offered an approach to high-resolution imaging of the dynamics
of immunological synapse maturation and, when combined with
total internal reflection fluorescence microscopy, offered an excel-
lent way to study recruitment of molecules from the cytoplasm to
the synapse [20–22]. Compared to alternative substrates such as
anti-CD3 and anti-CD28 in the presence of complete media, the
SLB better recapitulated patterns observed with live APC [23].
Limitations of SLB include that all receptors are laterally mobile,
and thus some functions that require modulation of ligand mobil-
ity and positioning by APCs cannot be recapitulated [24–26].
The axial rigidity is a limitation in some respects [27] but may
also account for the ability of SLB to provide mechanical feed-
back when cell pushing and pulling forces have an axial vector
[28, 29]. Thus, SLB have continued to be useful in analysis of
immunological synapses, and developing higher-throughput
approaches to generation and imaging would facilitate progress.
The potential of the SLB to be used in a higher-throughput
format has been limited by the challenge of cleaning glass sub-
strates that are integrated into 96-well plates and fragility of the
bilayers, which cannot be exposed to air/water interfaces due to
their reliance on the hydrophobic effect. Recently, liquid-han-
dling workstations have been utilized to form SLB, add pro-
teins, apply cells, and perform imaging [30]. We describe here
our version of this approach. While there are still aspects of this
approach that can be optimized further, we have started to use
this approach to screen libraries of small molecules for effects
on the immunological synapse and report here on a fully func-
tional platform.
Comprehensive Analysis of Immunological Synapse Phenotypes Using Supported Lipid… 425
2 Materials
2.1 SLB Plates 1. 96-well plates with 0.17 mm thickness glass bottoms. We have
used Brooks MGB096-1-2-LG-L plates after screening sam-
ples from five different sources (see Note 1).
2. 96-well deep-bottom plates. These come in a variety of sizes
with volumes from 0.5 to 2 ml. They are used to prepare mas-
ter plates for the liquid-handling workstation.
2.2 SLB Plate 1. Hellmanex III—make a 1% solution in 50% isopropanol.

Cleaning 2. Freshly prepared 3 N NaOH.
3. Ultrapure water.
4. Plate washer in sterile/laminar flow work space that is kept
clean and protein-protein/lipid-free so as not to contaminate
cleaned SLB plate.
5. Centrifuge with swinging bucket rotor and matched plate
carriers.
2.3 Phospholipids, We obtain phospholipids from Avanti Polar Lipids.

Buffers, and Extruder
1. 1,2-Dioleoyl(∆90-cis)-sn-glycero-3-phosphocholine (DOPC)
in chloroform.
2. 1,2-Dioleoyl-sn-glycero-3-[(N-(5-amino-1-carboxypentyl)
iminodiacetic acid)succinyl] (nickel salt) (DOGS-NTA) in
chloroform.
3. 1,2-Dioleoyl-sn-glycero-3-phosphoethanolamine-N-(cap bio-
tinyl) (sodium salt) (DOPE-cap-biotin) in chloroform.
4. Extruder set—0.75″ diameter, 100 nm pore diameter, 50×
polycarbonate membrane, comes with 2 ×1 ml heavy-duty glass
syringes. We use a unit from Avestin, whereas Chapter 2 uses a
similar system from Avanti with a 50 nm pore size filter.
5. Nitrogen and argon gas cylinders, regulators, flexible tubing,
glass Pasteur pipettes, a stand with two clamps, and a beaker
for warm water for making lipid films in a chemical fume hood
(due to evaporation of chloroform).
6. Lyophilizer.
7. Tris-buffered saline (TS) 25 mM 2-amino-2-hydroxymethyl-
propane-1,3-diol; 150 mM NaCl, pH 8 at 24 °C.
2.4 Proteins We use lentiviral expression for recombinant CD80 (see Chapter
and Related Solutions 28). We produce UCHT1 F(ab′)2 from commercial UCHT1 IgG
using pepsin (see Note 2).
1. Alexa 405-ICAM-1-His12 (see Note 3).
2. Alexa 568-UCHT1 Fab′-Cys-maleimide-biotin ~1 μM—store

frozen at 4 °C.
3. CD80-(Cys-mal-Alexa488)-His12 ~2 μM—store frozen at
−80 °C (see Note 4).
4. Human serum albumin, injection grade—20% w/v.
5. Bovine serum albumin—fraction V powder.
6. Phosphate-buffered saline (PBS)—155 mM NaCl, 2.7 mM
Na2HPO4, 1.54 mM KH2PO4, pH 7.2.
7. HEPES-buffered saline (HBS)—20 mM HEPES, 137 mM
NaCl, 1.7 mM KCl, 0.7 mM Na2PO4, 5 mM glucose, 2 mM
MgCl2, 1 mM CaCl2, pH 7.2.
8. Normal donkey serum (e.g., Merck Millipore).
9. Alexa Fluor® 647 AffiniPure Fab Fragment donkey anti-goat
IgG (H+L) (e.g., Southern Biotechnology).
10. Goat anti-PKCθ polyclonal antibody.
2.5 Cells 1. We obtain peripheral blood mononuclear cells through a non-

clinical component request to the UK National Health Service
Blood and Transplant unit in Headington, UK (see Note 5).
2. RosetteSep human CD4 T-cell purification kit (Stemcell
Technologies) or equivalent “untouched” purification strategy.
3. Anti-CD3 and anti-CD28 coated magnetic beads (Dynal) or
equivalent for T-cell stimulation.
4. Complete medium—RPMI-1640, 2 mM l-alanyl-l-glutamine,
10% fetal bovine serum (FBS), 50 U/ml penicillin and 50 μg/ml
streptomycin, 1 mM sodium pyruvate, 10 mM 2-[4-(2-hydroxy-
ethyl)piperazin-1-yl]ethanesulfonic acid (HEPES).
5. 2× PHEM buffer—to make 500 ml: weigh out 18.14 g
1,4-piperazinediethanesulfonic acid (PIPES), 6.5 g HEPES,
3.8 g EGTA, 0.99 g MgSO4, pH to 7.4 with 10 M NaOH. PIPES
will not go into solution until pH approaches 7; bring volume
to 500 ml, 0.2 μm filter, and store at 4 °C.
6. Electron-microscope-grade formaldehyde (8–16%). Dilute
with 2× PHEM buffer to prepare 4% formaldehyde.
2.6 Cellular High- ES and DE are members of the Target Discovery Institute of
Throughput Screening Oxford and are providing cellular high-throughput screening
expertise. If a similar facility is not available to you, it is possible to
obtain a liquid-handling workstation for use in a lab setting. Since
different specific programming will be involved on different work-
stations, we have included a detailed description of the pro-
grammed parameters, but not detailed code, which can be made
available on request for individuals with the same workstations.
1. Liquid-handling workstation. We are currently working with a

PerkinElmer “Janus” liquid-handling workstation utilizing a
P200 MDT dispensing head.
2. High-content imaging system. We are currently working with
a GE InCell 6000 with a 40× NA 0.75 objective.
3. Microtiter plate shaker(s). Small radius with potential for high
rpm to mix contents of microtiter wells without generating
bubbles or bringing meniscus into contact with SLB.
4. Humidified 37 °C incubator with water tray to allow partial
immersion of the plate to rapidly warm the sample plate to 37 °C.
2.7 Analysis 1. “Tool for Integrative Analysis of Motility” (TIAM) was

Software designed to track T-cell motility based on bright-field digital
images and then make measurements from the same cell based
on other imaging modes (reflected light, fluorescence, etc.)
that were interleaved with the bright-field images [31]. We
have taken advantage of TIAM’s ability to identify cells based
on bright-field images to perform feature extraction on fixed
cells in a high-throughput setting. This involved changes to
TIAM that will be released on github.com as TIAM-HT.
2. Analysis software used by TDI to determine Z and B statistics.
3. We have described the action carried out by the Janus and
InCell 6000 systems in some detail but have not provided the
specific instructions in this chapter. These can be made avail-
able on request from DE.
3 Procedures
3.1 Cleaning Glass 1. Add 100 μl of 1% Hellmanex III in 50% isopropanol to wells of
Bottom Plates (SLB the Brooks 96-well plate and incubate overnight, covered at
Plate A, B, etc.) room temperature.
2. Wash wells with 10 ml of ultrapure water per well using the
plate washer.
3. Add 100 μl of 3 N NaOH and incubate for 1 h at room
temperature.
4. Wash well with 10 ml of ultrapure water.
5. Dry SLB plate(s) by centrifuging upside down 3500 × g for
1 min. Complete drying with clean, dry nitrogen gas.
3.2 Master Plate 1. Phospholipids: small unilamellar vesicles (SUVs) are prepared
Preparation at a stock phospholipid concentration of 4 mM by an appro-
priate method (see Note 6) and diluted to a working concen-
tration of 0.4 mM in TS for supported lipid bilayer formation.
The stock SUVs formed by extrusion are (a) 100% DOPC,
(b) 2.5% DOGS-NTA(Ni):75% DOPC, and (c) 0.2% DOPE-

cap-biotin:99.8% DOPC. These stocks are mixed in different
ratios to prepare bilayers with specific composition assuming
linearity, which has been verified in many prior studies.
Prepare 150 μl/well/SLB plate of an SUV mixture with aver-
age composition 2.5% NTA, 0.0083% CapBio, and 97.5%
DOPC and add to deep-well Master Plate 1 (Fig. 1).
2. Proteins: Once the bilayers are formed, the standard wash
solution is HBS with 0.1% BSA. This solution can be kept in a
reservoir for washing with steps taken to avoid contamination
with reagents. A standard bilayer for formation of immuno-
logical synapse has the final composition 200 molecules/μm2
Alexa405-ICAM-1-His12, 100 molecules/μm2 CD80-(Cys-mal-
Alexa488)-His12, and 30 molecules/μm2 Alexa568-UCHT1-
Fab’-mal-biotin-streptavidin. The concentrations of each protein
needed to achieve these densities are determined empirically
using bilayer coated glass beads and flow cytometry with cali-
bration standards (see Chapter 4). A rule of thumb is that, for
His12-tagged proteins, ~10 nM for 30 min should give ~100
Master Plate 1 LHW

SUVs
Master Plate 2
Streptavidin
Waste
Master Plate 3
ICAM-1/CD80 HBS-0.1%BSA
SLB Plate 1,2…
/UCHT1
Master Plate 4
Media Well
1 mm
HCS
Master Plate 5
T cells SLB
Staining
reagents…
TIAM
Fig. 1 Schematic of high-throughput screen on SLB. The master plates can be set up using the most expedient
approach including manual pipetting or liquid-handling workstations. A reservoir with HBS/HSA and a waste
are set up for washing steps in addition to the series of master plates. A detail of the pipette tip positioning
during dispensing is shown
molecules/μm2. To generate surfaces with monodispersed

UCHT1, we adjust the percentage of DOPE-cap-biotin to
generate ~100 molecules/μm2 of streptavidin and then add
sufficient UCHT1-Fab′-mal-biotin to obtain 30 molecules/μm2.
3. Currently, we apply the streptavidin in HBS 0.1% BSA in one
step from Master Plate 2 and the His-tagged proteins and
UCHT1-mal-biotin in HBS 0.1% BSA in a second step from
Master Plate 3. Master Plate 4 can also be set up with 250 μl/
well/SLB plate of complete medium and equilibrate in 5%
CO2 incubator at 37 °C.
3.3 Cell Preparation 1. In order to perform a library screen with human T cells from a
single donor, we obtain peripheral blood mononuclear cell
samples from the UK National Health Service, isolate ~6 × 107
CD4+ T cells using RosetteSep protocol, activate them with
1:1 anti-CD3/anti-CD28-coated magnetic beads for 72 h in
250 ml flasks, remove the beads with a magnet, culture at ~106
cells/ml for another 5 days with 100 U/ml IL-2 added every
2 days in complete media, and then viably freeze the cells at
107 cells per aliquot. This enables a chemical library screen to
be completed with cells from one donor to reduce variability.
2. Cells are then thawed and put in culture with 25 U/ml IL-2
24 h prior to the screen. On the morning of the screen, the
viable cells are enriched by centrifugation over a Ficoll-
Hypaque cushion (d = 1.077) on which the viable cells float
and can be collected and washed three times with completed
media. If the cells will be treated with compounds, this can be
started prior to the formation of the SLB as this can be done
for several hours in the complete media, which we designate
Master Plate 5 in the screening sequence (Fig. 1). Master Plate
5 contains 5 × 105 cells/ml in complete media with 150 μl/
well/SLB. Plate will be set up and kept in the 37 °C CO2 incu-
bator while the SLB plate is assembled.
3.4 SLB Formation 1. The Janus liquid-handling workstation uses disposable pipette
on Liquid-Handling tips on 96 wells simultaneously. We can program the volume
Workstation (up to 200 μl), the rate, and the vertical and lateral position.
Bilayer formation is initiated by transferring 100 μl of the SUV
suspension from Master Plate 1 to each SLB plate at the center
of the well 1 mm from the bottom at a rate of 20 μl/s. Incubate
for 20 min at room temperature.
2. Add 200 μl of HBS 0.1% BSA to center of the well 1 mm from
the bottom at a rate of 20 μl/s, and remove 200 μl from the
same location at the same rate. Repeat seven times to achieve
an exchange factor of over 2000.
3. Add 2 × 200 μl of HBS 2% BSA and 100 μM NiSO4 to center
of the well 1 mm from the bottom at a rate of 20 μl/s, incubate
for 20 min at room temperature, and then remove 2 × 200 μl
from the center of the well 1 mm from the bottom at a rate of
20 μl/s.
same location at the same rate. Repeat seven times.
5. Transfer 100 μl from Master Plate 2 to each SLB plate, shake
the SLB plates at 300 rpm for 3 min, incubate for 20 min at
room temperature, and then remove 100 μl from the center of
the well 1 mm from the bottom at a rate of 20 μl/s.
7. Transfer 100 μl from Master Plate 3, shake for 3 min at
300 RPM and incubate for 30 min at room temperature, and
remove from the center of the well 1 mm from the bottom at
a rate of 20 μl/s.
same location at the same rate. Repeat seven times. This last
step is performed with pre-warmed buffer in order to mini-
mize the warming time once the cells are in the plate.
9. Add 200 μl of pre-warmed complete medium from Master
Plate 4 to each SLB plate at the center of the well 1 mm from
same location at the same rate. Repeat once more to change
the well contents to 90% complete medium.
3.5 Immunological 1. Put Master Plate 5 with the T cells on the workstation and
Synapse Formation resuspend the cells by positioning the pipette tips 1 mm from
and Immuno the bottom of the plate and draw in 200 μl at a rate of 100 μl/s
fluorescence Staining and expel the same 200 μl at 100 μl/s. Repeat this five times to
generate a single-cell suspension.
2. Transfer 100 μl from Master Plate 5 to the SLB plate at a rate
of 20 μl/min 1 mm from the bottom of the plate. If two SLB
plates are processed, then the addition of the cells is staggered
to allow 10 min between start times for the two plates, which
allows time to precisely apply the fixative at the end of 15-min
incubation.
3. Centrifuge the SLB plate(s) at 80 × g for 1 min with a plate
carrier.
4. Incubate for 15 min at 37 °C partly submerged in a 37 °C
equilibrated water trough in 5% CO2 atmosphere.
5. Blot the water from the SLB plate and place on workstation
and remove 100 μl of media from the SLB plate and discard.
6. Remove 100 µl from the center of the well 1 mm from the
bottom at 20 µl/s.

same location at the same rate.
8. Add 100 μl of 4% formaldehyde in PHEM buffer at 1 mm
above the bottom of the well and 20 μl/s. In order to ensure
mixing, the workstation further performs ten cycles of remov-
ing 50 μl and returning it at a rate of 20 μl/s without moving
from its position 1 mm above the bottom (pipetting up and
down). The SLB plates are then incubated for 10 min at room
temperature to complete fixation and then remove 100 μl.
10. Add 100 μl of 0.2% Triton X-100 to the SLB plate(s) at 1 mm
above the bottom of the well and 20 μl/s. In order to ensure
mixing, the workstation further performs ten cycles of remov-
ing 50 μl and returning it at a rate of 20 μl/s without moving
from its position 1 mm above the bottom (pipetting up and
down). Incubate for 2.5 min and then remove 100 μl.
12. Add 100 μl of 10% donkey serum in HBS/0.1% BSA to the
SLB plate(s), shake for 3 min at 300 rpm, incubate for 1 h at
room temperature, and remove 100 μl.
13. Add 100 μl of 1:2000 goat anti-PKC-θ in 10% donkey serum in
HBS/0.1% BSA to the SLB plate(s), shake for 3 min at 300
rpm, incubate for 1 h at room temperature, and remove 100 μl.
15. Add 100 μl of 1:1000 Alexa 647 donkey anti-goat IgG in 10%
donkey serum in HBS/0.1% BSA to the SLB plate(s), shake
for 3 min at 300 rpm, incubate for 1 h at room temperature,
and remove 100 μl.
3.6 Imaging 1. Images were captures with the InCell 6000 using a 40× 0.75
NA objective.
2. The high-content imaging system was programmed to capture
bright-field, blue, green, orange, and red fluorescence signals from
16 fields in the central 1/9th of the well (Fig. 2) (see Note 7).
3. Image files were acquired on a local hard drive and then moved
to a server to provide access to the analysis computer.
Fig. 2 Selection of imaging field with current procedure. With our current bilayer building process, the central
ninth of the well has consistent cSMAC formation (Field 6), whereas the surrounding regions show UCHT1
accumulation but failure to form cSMACs (Field 1). We are currently restricting data collection to the central
ninth of the well (blue-shaded area) but are investigating the cause of this problem with the goal of being
about to use 100% of the well area
3.7 High-Content 1. Install MATLAB if it is not yet installed. Download TIAM_

Image Analysis Using HT (from https://github.com/uvmayya/TIAM_HT), and
TIAM_HT extract it under the MATLAB directory under “Documents”
or “My Documents” (see Note 8). Refer to the user guide and
other supporting documentation under the “tiam_HT/doc/”
folder for detailed instructions and information on the analysis
package. As in the case of original TIAM package [31], TIAM_
HT relies on identifying cells from transmitted light images, be
it bright-field, differential interference contrast (DIC), or
phase-contrast (Fig. 3). The locations of centroids of identified
cells are then used to perform local segmentation of cropped
sections in the fluorescence images, followed by measurement
of a total of 22 features and properties of T-cell synapses.
2. Have TIFF images belonging to a 96-well plate acquired on the
InCell platform in a folder with a suggestive name (for illustra-
tion the name is “plate1”). The TIFF image file names indicate
the well name, the field name, the mode of the image (if bright
field), and the fluorescence channels based on excitation and
emission filters used, as in the following example: “A - 1(fld 2
wv UV - DAPI).” Using ImageJ, create a folder with a subset
of bright-field images cropped to be 512-by-512 pixels in size
and change to 8-bit type, if necessary (see Note 9). This subset
is used for optimizing the parameters for detection of cells.
3. The parameters for detection can be arrived at using the Java-
based graphic user interface (GUI) (see Note 9). To launch the
GUI, navigate to “tiam_HT/src” and type “getParams-
FromGUImain” in the command window of the MATLAB
interface. Follow the prompts from the GUI to load the
transmitted light images, and then follow the displayed images
of the intermediate stages along with the GUI prompts to
Fig. 3 Workflow of TIAM_HT for high-content analysis. Cells in the bright-field images are detected based on
Canny edge detection followed by identification of circle-like patterns in the edges using circular Hough trans-
form. The identified cell centroids (denoted in red in the bright-field image) are then chosen one by one for
local segmentation of cropped sections (shown as a yellow box in the bright-field image) of all fluorescence
channels. Different segmentation algorithms are implemented depending on the fluorescence channel as the
desired information varies. Further, potential adjoining cells in the cropped sections are rejected based on
multiple criteria. The results of the segmentation are shown here in red outlines in cropped sections. Note that
the outlines are placed outside of the boundary of segmented regions. Segmented regions are then used to
calculate several features that quantify various properties of a T-cell synapse. These features are stored with
a unique ID for the detected cell along with the well and field information
adjust the various input parameters that determine operations

of the Canny edge detection and circular Hough transform
algorithms. Illustrative images of the intermediate stages given
in the user guide are useful in choosing these parameters. The
numeric values of input parameters for detection that were
optimized through the GUI routine are returned to the dis-
play panel of the command window. Make a note of the param-
eter names and their values.
4. For high-content analysis of entire datasets, a batch submission
file is used (see Note 10). Example batch submission files
named “batchHTsubmit_*.m” are provided under “tiamHT/
src/”. Copy a submission file to the location containing the

folder of acquired images (which was named “plate1” in step 2
this section 3.7). Parameters or variables in the batch submis-
sion file whose values need to be set are essentially self-explan-
atory. Additional comments are provided in the submission file
itself (start with “%” and appear in green when opened in the
“Editor” panel of MATLAB) and in the user guide that serve
as description of what these variables mean and how they
should be set. If necessary, replace the values of parameters for
detection of cells with those obtained from the GUI-based
optimization routine. After making all the necessary changes
and saving the “batchHTsubmit_*.m” file, execute the analysis
by typing “batchHTsubmit_*” in the command prompt and
hitting the <Enter> key. The status messages indicate the cur-
rent stage of analysis (see Note 11).
5. All output from the analysis is stored in the “tiamHT/ws” direc-
tory. All the result files and folders from the analysis have “exp-
Name” in their name, as designated in the batchHTsubmit_*.m
file that was executed. “<expName>_results.mat” has all the
detected objects and all the associated properties in the MATLAB
“structure” data format. There is also “<expName>_perWell”
folder wherein this information is categorized in separate files as
per the well in which the object was detected. These files also
contain the image and mask arrays of each detected object.
“<expName>_report.mat” contains the count statistics and
mean values for each well. Move all the result files to the location
that contains the folder of acquired images (which was named
“plate1” in step 2 of section 3.7).
6. In the MATLAB interface, open and load the “<expName>_
report.mat.” Copy the table and paste it in a MS Excel work-
sheet. Data from well #10, 11, and 12 come before data from
well #2 as the ordering in MATLAB is based on alphanumeric
sort of the well names. Move these three rows below the row
corresponding to well #9. This needs to happen for every
column of the well (i.e., A, B, C through to H). Plot some of
the count statistics and synapse features to check if treatment
with any of the compounds resulted appreciable deviation
from typical values (see Note 12). Name the Excel sheet
“<expName>_96wellReport.xlsx.”
3.8 Identifying “hit” 1.

Integrate compound information into the Excel sheet
Compounds from the “<expName>_96wellReport.xlsx.”.
Measured Properties 2. Statistical analysis is carried out for 22 features and synapse
of T-Cell Synapses properties using the open-source software package CellHTS2
(Web-based version is http://web-cellhts2.dkfz.de/cellHTS-
java/cellHTS2/) [32].
4 Notes
1. Criteria for plate selection. We screened ~6 plate suppliers using

washing procedures provided by the Mayor Lab (Bangalore,
India). These involve soaking in 5 N NaOH, rinsing with
ultrapure water, and then soaking with an alkaline glassware
detergent in 50% isopropanol followed by extensive washing
with ultrapure water. The plates were then dried prior to add-
ing SUVs to form bilayers. The quality metrics were uniformity
minimal spots and holes (<1% area), lipid mobility (100%, >1
μm2/s), and protein mobility (>90%, ~1 μm2/s). If the speci-
fied plates are not available, these criteria may be applied to
identify suitable plates. It also may be possible to fabricate suit-
able plates by bonding polydimethylsiloxane (PMDS) to bot-
tomless thermoplastic plates [33]. SLB will form on PDMS
substrates [34].
2. UCHT1 Fab’. UCHT1 IgG was obtained from Bio X Cell
(https://bxcell.com). We use this antibody because we could
not develop a protocol to make F(ab’)2 from the more com-
monly used OKT3. To make F(ab′)2 from UCHT1, we pre-
pared a solution of 1–5 mg/ml UCHT1 in digestion buffer (10
mM sodium citrate, 25 mM NaCl, pH 4). Activate the pepsin-
agarose (50 μl slurry/mg of UCHT1) by incubating with acti-
vating buffer (50 mM glycine, 50 mM NaCl, pH 3) and shaking
at 37 °C for 10 min. Sediment the resin and remove the activat-
ing buffer and add the antibody to the resin and mix overnight
at 37 °C. Spin down the resin and save the supernatant contain-
ing the F(ab′)2 and residual IgG. Buffer exchange to TS pH 8
and pass over a clean protein A column. The flow-through con-
tains the F(ab′)2. Anion exchange chromatography may also be
used for this step. Confirm F(ab′)2 purity by non-reducing
PAGE, and determine the protein concentration by OD280
(1.4 ml/mg). Concentrate the F(ab′)2 to 1–10 mg/ml in reac-
tion buffer (PBS, 1 mM EDTA, pH 7.2) PBS, and add
2-mercaptoethylamine∙HCl (2-MEA) at a final concentration of
50 mM. Incubate for 90 min at 37 °C, cool to room tempera-
ture, and separate the 2-MEA and non-reduced F(ab′)2 using a
HiLoad 16/600 Superdex 75 pg pre-equilibrated in reaction
buffer at 4 °C. Elute the Fab′ in reaction buffer. React the
pooled Fab′ with a 40-fold molar excess of maleimide-PEG2-
biotin. Dialyze overnight against two changes of PBS, concen-
trate the UCHT1 Fab′-Cys-maleimide-biotin, and determine
the concentration by OD280 with a spectrophotometer.
3. ICAM-1 species choice. We use mouse ICAM-1 in these systems
because it interacts with both mouse and human LFA-1 similarly
[35]. In contrast, human ICAM-1 interacts with human LFA-1
but not mouse LFA-1 [35]. Therefore, for a lab that works with
both human and mouse T cells, mouse ICAM-1 is a more versa-

tile reagent. There are potentially useful epitopes in domains 4
and 5 of human ICAM-1 so in some cases we have made chime-
ric molecules with two N-terminal immunoglobulin-like domains
of mouse ICAM-1 and three C-terminal Ig-like domains of
human ICAM-1[17]. Current mICAM-1-His12 is purified from
S2 cells as follows. Mouse ICAM-1-expressing S2 cells are
induced at 10–20 millions/ml while in log phase growth with
0.5 mM CuSO4 (final concentration). After the induction, the
cells are allowed to express protein 3–4 days, and then the cells
were centrifuged down to collect the supernatant. The superna-
tant (5–10 L) is concentrated 10–20 fold using a Vivaflow 200
tangential crossflow concentrator. Mix the concentrated super-
natant with an equal volume of equilibration/wash buffer (1×
PBS, 300 mM NaCl, 10 mM imidazole; pH 7.4). Add equili-
brated His Pur™ cobalt resin (binding capacity of 10 mg of pure
His-tagged protein per milliliter of resin or equivalent) to the
supernatant and stir gently overnight at 4 °C. Wash the resin with
five-resin bed volumes of equilibration/wash buffer. Elute pro-
teins with five-resin bed volumes of elution buffer (1× PBS,
300 mM NaCl, 150 mM imidazole; pH 7.4). Measure the con-
centration by OD280 with a spectrophotometer, and check the
purity of protein by reducing PAGE. Pool the appropriate frac-
tions and purify further using a HiLoad 16/600 Superdex 75 pg
gel filtration column. Collect and concentrate appropriate pro-
tein fractions after a reduced PAGE analysis.
4. Free cysteine tags. Labeling of CD80 on lysines quantitatively
inactivates them. Adding a free cysteine into the C-terminal
His12 tag provides a site for modification by maleimide-dye con-
jugates with 50–100% labeling efficiency and full retention of
activity. We have found that this is a versatile approach to label-
ing many proteins that are sensitive to labeling on lysines.
Following is the current maleimide-dye labeling protocol.
CD80-Cys-His12 is buffer exchanged to PBS, pH 7.4, and con-
centrated to between 0.5 and 1 mg/ml using a centrifugal
ultrafiltration system. Lyophilized Alexa fluor 488 C5 maleimide
is added directly to the protein solution at tenfold molar excess.
Mix the dye and protein solution well by vortexing, and incu-
bate at room temperature for 1 h. The protein is separated from
unbound dye by centrifugal ultrafiltration (10 K MWCO). The
dye per protein ratio is calculated from using optical density of
the dye (peak absorbance) and protein (280 nm) absorbance.
5. These materials are from human donors and must be treated as
potentially infectious. Murine T cells or cell lines like Jurkat can also
be utilized for immunological synapse formation experiments.
6. Small unilamellar vesicles (SUVs). SUVs can be prepared by a
variety of methods. We utilize extrusion (see protocol details in
Chapter 2). Other acceptable methods include detergent dialy-

sis (see Chapter 4), freeze-thaw (see Chapter 5), and sonication
(see Chapter 14). In all cases, the solutions at 0.2–5 mM phos-
pholipid should be clear. Any significant turbidity indicates
multilamellar vesicles, which don’t form continuous bilayers.
7. Selection of imaging fields. We have noted that the quality of
the UCHT1 clustering deteriorates near the well edges so con-
fine the imaging fields to the central 1/9th. Given that immu-
nological synapse formation is very efficient (near 100%), we
obtain more sufficient images in the 16 fields for statistical
power. We are experimenting with approaches to make the
entire well area equally active in supporting synapse formation,
which would allow analysis of lower-frequency events, such as
antigen-specific cells in polyclonal populations.
8. MATLAB package. TIAM_HT is a package for high-content
image analysis based on MATLAB. It is compatible with any
version of MATLAB from R2009a to R2014a. Basic opera-
tional knowledge of MATLAB is essential to make use of
TIAM_HT, which can be quickly gathered via any one of
numerous succinct educational resources on the World Wide
Web. Coding knowledge in MATLAB is not essential.
Operational knowledge of ImageJ (https://imagej.nih.gov/
ij/) is also helpful. The step-by-step instructions in this chapter
are tailored for running TIAM_HT on the Windows platform.
However, the package should run on Mac OS and Linux plat-
forms if file paths are properly modified in the MATLAB scripts
and functions of TIAM_HT. Refer to the user guide under
“tiam_HT/doc/” for locations of these file paths. The user
guide also provides information on handling image file names
generated by acquisition platforms other than InCell from
GE. The user guide provides further information on various
image-processing parameters in respective scripts and func-
tions that impact the measured features and properties. It also
describes the measured features and properties that are used to
identify the “hits,” i.e., the small molecule compounds that
significantly alter the synapse properties of T cells.
9. Image scaling. The GUI routine only handles images of size
512-by-512 and of 8-bit depth. Since the original images are
cropped rather than resized, the “imageScale” parameter
remains the same for the actual high-content analysis. Reducing
the bit depth to 8 from 12 or 16 may necessitate increasing the
“edgeValue” for the actual high-content analysis, but we typi-
cally haven’t felt the need to do so. The optimized detection
parameter values are also saved under the variable “params”
along with centroids of detected cells in each image of the sub-
set under the variable “statscell” in an output file named “*_
params.mat” (* is the “name of analysis” provided via GUI) in
the “tiam_HT/ws” folder. These centroids could be manually

verified in ImageJ for randomly selected cases. Further, out-
lines of detected cells in transmitted light images can be
obtained during high-content analysis (see Note 10) to ascer-
tain that cells are being detected with desired sensitivity and
specificity.
10. It is best to test out DIC outlining and fluorescence segmenta-
tion on a small subset (say, all fields from a well) of files corre-
sponding to all the channels (Fig. 4). For this, one needs to
Fig. 4 Visual assessment of quality of outlining of cells in bright field and segmentation of cells in fluorescence
channels. TIAM_HT provides an option for generating cropped sections of each cell in each of the channels and
also for storing binary images of boundaries of each cell in each of the channels. These can be assembled as
a montage in ImageJ to visually assess the quality of detection and outlining of cells in bright field and of
segmentation in fluorescence channels and also to make sure that synapse properties are as per expectation
the case of negative and positive controls. A small section of the montage is shown here for illustration: (a)
While all cells are detected, there is scope for improvement in outlining. Further, some circle-like imperfections
are also being picked as cells, which are false positives. (b) While well-formed cSMAC is segmented accu-
rately, the whole cell is identified as cSMAC when the central accumulation is not appreciable. (c) The current
implementation of segmentation for pSMAC is not capable of identifying the central exclusion of ICAM1 in most
cases. (d) Overlay of pSMAC and cSMAC shows well-formed synapses in central area of the wells. (e) and (f)
Segmentation of accumulated CD80 and immunostained PKCθ is good. It is to be noted that the montage or
the individual stored images can also be used for visual verification of the identified outliers after the high-
content analysis
make sure that the lines corresponding to the call of the

“imwrite” function are not commented out in functions
“tiamHTbatchScript.m” and “getOutline.m” (i.e., remove the
beginning character of “%”). Refer to the user guide for loca-
tions of these in the respective functions. For fluorescence
images, the outlines need to be overlaid using ImageJ to appre-
ciate the quality of segmentation.
11. Computation time. Type <ctrl>c if the analysis needs to be
stopped as per intention. The calculations typically take 6–12 h
on recent generation of Intel Xeon multi-core processors when
100,000–150,000 cells in total are detected; this is without a
direct call for parallel processing in MATLAB. At least 16 GB
of RAM is necessary as images from all fields of the entire
96-well plate for a particular channel have to be available on
the RAM. Once the analysis is finished, the command line (>>)
will reappear. Type “clc” to clear all the status messages.
12. Image montage. For further verification of substantial devia-
tion in a well due to drug treatment, invoke the function
“getCellImages(<wellName>, <expName>)” in the location
where result files are stored. This generates an additional folder
with the name “<expName>_cellImg,” under which a sub-
folder with the <wellName> is created, which in turn stores
images and outlines of all cells from that well in each channel.
Overlaid montages of these can be created in ImageJ to visually
compare synapse properties of drug-treated cells with those of
untreated cells.
Acknowledgments
The authors thank M. Santos and S. Davis for sharing methods for
protein production using the lentiviral system. Wellcome Trust
grant 100262/Z/12/Z and the Kennedy Trust for Rheumatology
Research supported this work.
References
1. Zinkernagel RM, Doherty PC (1974) 4. Paul WE, Seder RA (1994) Lymphocyte

Immunological surveillance against altered self responses and cytokines. Cell 76:241–251
components by sensitised T lymphocytes in 5. Dustin ML, Olszowy MW, Holdorf AD, Li J,
lymphocytic choriomeningitis. Nature Bromley S, Desai N, Widder P, Rosenberger F,
251(5475):547–548 van der Merwe PA, Allen PM, Shaw AS (1998)
2. Poo WJ, Conrad L, Janeway CA Jr (1988) A novel adapter protein orchestrates receptor
Receptor-directed focusing of lymphokine release patterning and cytoskeletal polarity in T cell
by helper T cells. Nature 332(6162):378–380 contacts. Cell 94:667–677
3. Norcross MA (1984) A synaptic basis for 6. Dustin ML, Colman DR (2002) Neural and
T-lymphocyte activation. Ann Immunol (Paris) immunological synaptic relations. Science
135D(2):113–134 298(5594):785–789
7. Geiger B, Rosen D, Berke G (1982) Spatial 19. Tseng SY, Waite JC, Liu M, Vardhana S, Dustin
relationships of microtubule-organizing cen- ML (2008) T cell-dendritic cell immunological
ters and the contact area of cytotoxic T lym- synapses contain TCR-dependent CD28-CD80
phocytes and target cells. J Cell Biol clusters that recruit protein kinase Ctheta.
95(1):137–143 J Immunol 181(7):4852–4863
8. Carpen O, Virtanen I, Saksela E (1982) 20. Campi G, Varma R, Dustin ML (2005) Actin
Ultrastructure of human natural killer cells: and agonist MHC-peptide complex-dependent
nature of the cytolytic contacts in relation to cel- T cell receptor microclusters as scaffolds for
lular secretion. J Immunol 128(6):2691–2697 signaling. J Exp Med 202(8):1031–1036
9. Schmidt RE, Caulfield JP, Michon J, Hein A, 21. Yokosuka T, Sakata-Sogawa K, Kobayashi W,
Kamada MM, MacDermott RP, Stevens RL, Hiroshima M, Hashimoto-Tane A, Tokunaga
Ritz J (1988) T11/CD2 activation of cloned M, Dustin ML, Saito T (2005) Newly generated
human natural killer cells results in increased T cell receptor microclusters initiate and sustain
conjugate formation and exocytosis of cytolytic T cell activation by recruitment of Zap70 and
granules. J Immunol 140(3):991–1002 SLP-76. Nat Immunol 6:1253–1262
10. Kupfer A, Singer SJ (1989) Cell biology of cyto- 22. Varma R, Campi G, Yokosuka T, Saito T,
toxic and helper T cell functions: immunofluo- Dustin ML (2006) T cell receptor-proximal
rescence microscopic studies of single cells and signals are sustained in peripheral microclusters
cell couples. Annu Rev Immunol 7:309–337 and terminated in the central supramolecular
11. Stinchcombe JC, Bossi G, Booth S, Griffiths activation cluster. Immunity 25(1):117–127
GM (2001) The immunological synapse of 23. Bunnell SC, Hong DI, Kardon JR, Yamazaki
CTL contains a secretory domain and mem- T, McGlade CJ, Barr VA, Samelson LE (2002)
brane bridges. Immunity 15(5):751–761 T cell receptor ligation induces the formation
12. Batista FD, Iber D, Neuberger MS (2001) B of dynamically regulated signaling assemblies.
cells acquire antigen from target cells after syn- J Cell Biol 158(7):1263–1275
apse formation. Nature 411(6836):489–494 24. Al-Alwan MM, Rowden G, Lee TD, West KA
13. Carroll-Portillo A, Spendier K, Pfeiffer J, (2001) The dendritic cell cytoskeleton is criti-
Griffiths G, Li H, Lidke KA, Oliver JM, Lidke cal for the formation of the immunological syn-
DS, Thomas JL, Wilson BS, Timlin JA (2010) apse. J Immunol 166(3):1452–1456
Formation of a mast cell synapse: Fc epsilon RI 25. Al-Alwan MM, Rowden G, Lee TD, West KA
membrane dynamics upon binding mobile or (2001) Fascin is involved in the antigen presen-
immobilized ligands on surfaces. J Immunol tation activity of mature dendritic cells.
184(3):1328–1338 J Immunol 166(1):338–345
14. Goodridge HS, Reyes CN, Becker CA, 26. Comrie WA, Li S, Boyle S, Burkhardt JK
Katsumoto TR, Ma J, Wolf AJ, Bose N, Chan (2015) The dendritic cell cytoskeleton pro-
AS, Magee AS, Danielson ME, Weiss A, motes T cell adhesion and activation by con-
Vasilakos JP, Underhill DM (2011) Activation straining ICAM-1 mobility. J Cell Biol 208(4):
of the innate immune receptor Dectin-1 upon 457–473
formation of a ‘phagocytic synapse’. Nature 27. Natkanski E, Lee WY, Mistry B, Casal A,
472(7344):471–475 Molloy JE, Tolar P (2013) B cells use mechani-
15. Monks CR, Freiberg BA, Kupfer H, Sciaky N, cal energy to discriminate antigen affinities.
Kupfer A (1998) Three-dimensional segrega- Science 340(6140):1587–1590
tion of supramolecular activation clusters in T 28. Kumari S, Depoil D, Martinelli R, Judokusumo
cells. Nature 395(6697):82–86 E, Carmona G, Gertler FB, Kam LC, Carman
16. Monks CR, Kupfer H, Tamir I, Barlow A, CV, Burkhardt JK, Irvine DJ, Dustin ML
Kupfer A (1997) Selective modulation of pro- (2015) Actin foci facilitate activation of the
tein kinase C-theta during T-cell activation. phospholipase C-gamma in primary T lympho-
Nature 385(6611):83–86 cytes via the WASP pathway. eLife 4. 10.7554/
17. Grakoui A, Bromley SK, Sumen C, Davis MM, eLife.04953
Shaw AS, Allen PM, Dustin ML (1999) The 29. Kumari S, Vardhana S, Cammer M, Curado S,
immunological synapse: a molecular machine Santos L, Sheetz MP, Dustin ML (2012) T
controlling T cell activation. Science Lymphocyte myosin IIA is required for matu-
285(5425):221–227 ration of the immunological synapse. Front
18. Yokosuka T, Kobayashi W, Sakata-Sogawa K, Immunol 3:230. doi:10.3389/fimmu.2012.
Takamatsu M, Hashimoto-Tane A, Dustin 00230
ML, Tokunaga M, Saito T (2008) 30. Nair PM, Ngu H, Torres E, Marsters S,
Spatiotemporal regulation of T cell costimula- Lawrence DA, Stephan JP, Komuves L,
tion by TCR-CD28 microclusters and protein Ashkenazi A (2015) Membrane display and
kinase C theta translocation. Immunity functional analysis of juxtacrine ligand-receptor
29(4):589–601 signaling. Biotechniques 59(4):231–238 240
31. Mayya V, Neiswanger W, Medina R, Wiggins form using micro- and nanotopography.
CH, Dustin ML (2015) Integrative analysis of Nano Lett. doi:10.1021/acs.nanolett.
T cell motility from multi-channel microscopy 5b04364
data using TIAM. J Immunol Methods
34.
Torres AJ, Contento RL, Gordo S,
416:84–93. doi:10.1016/j.jim.2014.11.004 Wucherpfennig KW, Love JC (2013)
32. Goktug AN, Chai SC, Chen T (2013) Data Functional single-cell analysis of T-cell activa-
analysis approaches in high throughput screen- tion by supported lipid bilayer-tethered ligands
ing. In: El-Shemy HA (ed) Drug discovery. on arrays of nanowells. Lab Chip 13(1):90–99.
InTech, 2013. Rijeka, Croatia doi:10.1039/c2lc40869d
3 3. Hu J, Gondarenko AA, Dang AP, Bashour 35.
Johnston SC, Dustin ML, Hibbs ML,
KT, O’Connor RS, Lee S, Liapis A, Springer TA (1990) On the species specific-
Ghassemi S, Milone MC, Sheetz MP, Dustin ity of the interaction of LFA-1 with intercel-
ML, Kam LC, Hone JC (2016) High- lular adhesion molecules. J Immunol 145(4):
throughput mechanobiology screening plat- 1181–1187
Chapter 27
Studying Immunoreceptor Signaling in Human T Cells

Using Electroporation of In Vitro Transcribed mRNA
Omkar Kawalekar, Carl H. June, and Michael C. Milone
Abstract
The recognition bestowed upon T lymphocytes as key mediators of cellular immunity has been further
attested by recent successful clinical studies using genetically modified T cells. With an ever-growing inter-
est in the application of T cells to treat human malignancies, studying the molecular mechanisms of T cell
activation, signaling, and function has become imperative. This, therefore, calls for the development of
new easy-to-use and accurate models to investigate the biological phenomena that begin at the synaptic
levels of T cell and antigen interactions to the ultimate exhaustion and death of the T cell. Here, we
describe an approach to transiently express a chimeric molecule on the cell surface that permits activation
and expansion of T cells, thereby providing a model to study T cell signaling.
Key words Immunotherapy, Antigen, Synthetic biology, Receptors, Costimulation
1 Introduction
Deciphering the detailed cascade of events that initiate the earliest

biochemical events ultimately leading to T cell activation has come
a long way in the past few decades [1, 2]. More recently, the advent
of synthetic biology has enabled easy genetic manipulation of T
cells permitting the expression of chimeric molecules to enhance
functions of T cells. One such method of genetic engineering is the
engraftment of chimeric antigen receptors (CARs) on the T cell
surface. CARs are synthetic molecules that contain a single-chain
variable fragment (scFv) obtained from the variable chains of a
monoclonal antibody with desired specificity, which is fused to
intracellular domains that provide T cell activation and costimula-
tory signals [3]. CARs allow T cells to recognize predetermined
targets, which would otherwise escape immune recognition,
thereby making this technology an attractive tool in the combating
various cancers and infections [4]. Despite extensive clinical inves-
tigation of genetically engineered T cells, the intricate signaling
mechanisms that occur downstream of such chimeric molecules is
443
444 Omkar Kawalekar et al.
highly understudied. One major obstacle that has limited this

research is the lack of an accurate model to study the antigen recep-
tor synapse as well as signaling downstream of it.
Current in vitro models for activation and expansion to study
T cell signaling and function pose several limitations. Until about
a decade ago, mitogenic lectins such as phytohemagglutinin (PHA)
and concanavalin A (Con A) were being used for stimulation and
expansion of polyclonal T cell population [5]. These mitogenic
molecules bind to glycoproteins on the cell surface. Another such
stimulant is phorbol 12-myristate 13-acetate (PMA), which acti-
vates T cells by direct stimulation of protein kinase C [5]. To
achieve T cell receptor (TCR) complex-specific stimulation, anti-
bodies specific to such molecules, including CD2, CD3, CD28,
and CD45, have been used. These antibodies provide the required
costimulatory signal to trigger complete activation and prolifera-
tion of T cells in culture [6]. The field has since progressed to
immobilizing these antibodies to accessory cells, beads, or a solid
surface for robust expansion of T lymphocytes [7].
Requirement of a functional TCR, reliance on commercial
vendors for production, procurement and application of TCR-
antagonizing antibodies, and the additional costs of acquiring two
different antibodies (primary and secondary stimulants) for com-
plete T cell activation all contribute to the myriad of limitations
and drawback of these methods. Prolonged stimulation with such
antibodies could provide “excess” activation signal, which, in naïve
T cells, for example, has been shown to be detrimental [8, 9].
There is, thus, a clear need for an improved model to study T cell
signaling.
Described below is refined method that can be employed for
the activation and/or expansion of immune cells. Briefly, the tech-
nique involves transient expression of a CAR molecule of the T cell
surface and subsequent activation via a ligand specific to the CAR
molecule. The transient mode of gene delivery allows CAR expres-
sion on over 95% of the cells, thereby allowing activation of almost
the entire cell population.
2 Materials
2.1 Production 1. The IVT mRNA encoding the CAR can be manufactured
of In Vitro Transcribed using a polymerase chain reaction (PCR)-generated template.
(IVT) mRNA This template is the DNA sequence of the CAR of interest
obtained from any appropriate source such as plasmid DNA,
cDNA, or synthetic DNA sequence.
2. The template must contain appropriate promoters and a cor-
responding RNA polymerase. For example, to use the T7
mScript™ RNA system (Catalog no. C-MSC11610, Cellscript,
Studying T Cell Signaling Using mRNA Transfection 445
WI, USA) requires the T7 bacteriophage promoter (TAATAC

GACTCACTATAG) upstream of the double-stranded DNA
template. Other RNA production kits using different promoter
systems, such as SP6 and T3, are also available and can be used
for synthesis of mRNA to be used for this protocol.
3. Follow the manufacturer’s instruction for mRNA production.
Purify IVT mRNA products using an RNeasy Mini Kit (Qiagen
Inc., Valencia, CA, USA) as per manufacturer’s specifications.
Elute purified mRNA in sterile RNAse-free water at a concen-
tration of 1 mg/ml. Prepare aliquots of 10 μl each in RNAse-
free tubes and store at −80 °C until further use.
2.2 Components 1. Electroporation apparatus(See Note 1).

for Electroporation 2. 2 mm gap electroporation cuvette top fit electroporation
apparatus.
2.3 Components 1. Opti-MEM I: Reduced serum medium (Catalog no. 31985,

for Cell Culture Gibco, Grand Island, NY, USA).
and CAR Expression 2. R10: RPMI 1640 medium supplemented with 10% fetal calf
Analysis serum.
3. One T25 culture flask for every 1x107 cells electroporated.
4. FACS tubes: 5 ml round-bottom tubes for flow cytometry
analysis.
5. FACS buffer: Phosphate buffered saline with 1% fetal calf
serum.
6. CAR detection antibodies: Biotin-labeled polyclonal anti-
mouse F(ab)2 or any other antibody that would detect the
CAR scFv. If it is not pre-conjugated with a fluorescent dye,
then use phycoerythrin-labeled streptavidin.
2.4 Antigens 1. Antigen: Purified antigenic protein or an anti-idiotype specific

and Coating to the scFv of the CAR. This cognate antigenic molecule
of Stimulation Beads should specifically bind to and stimulate the distinctive
sequence of the scFv.
2. Stimulation beads: Cognate antigen needs to be coupled with
magnetic tosylactivated beads, such as Dynabeads® M-450
(Catalog no. 14103, Life Technologies, Grand Island, NY,
USA). The coupling procedure needs to be followed as per
the manufacturer’s instructions. Briefly, the coupling is per-
formed overnight by co-incubation of the antigenic molecule
with the Dynabeads at a high pH (8.5–9.5) and at 37 °C. The
coated beads should be stored at 4 °C at a desired concentra-
tion in the bead-storage buffer as specified in the manufac-
turer’s protocol. Suggested concentration for long-term bead
storage is 3 × 107 beads/ml.
3 Methods (see Notes 8 and 9)
Carry out all procedures at room temperature and in sterile condi-

tions unless otherwise specified.
3.1 Electroporation 1. Obtain live T cells from any source (human peripheral blood,
of mRNA into T human umbilical cord blood, etc.) and count cells while ensur-
Cells (see Note 3) ing good cell viability.
2. Centrifuge cells at 300 × g for 5 min at 4 °C. Carefully discard
supernatant and resuspend cell pellet in fresh Opti-MEM media.
3. Centrifuge again and repeat wash steps for a total of three
washes.
4. Count and resuspend cells in fresh Opti-MEM media at 1 ×
108 cells/ml. For each electroporation, aliquot 1 × 107 cells in
a 100 μl of Opti-MEM. Keep cells on ice until use.
5. Pre-configure the electroporator by setting the voltage to 500
V and time to 1000 μs. Prewarm R10 to 37 °C and add 10 ml
of the media to a T25 flask.
6. In a separate tube, combine 10 μg of RNA (stock concentra-
tion of 1 mg/ml) with the 100 μl aliquot of cells (see Note 2).
Uniformly mix by gentle pipetting. Immediately empty the
entire content into a 2 mm cuvette.
7. Place the cuvette into the electroporator cassette, tighten the
electrodes around the metal plates of the cuvette, and initiate
the electric pulse.
8. Immediately transfer the contents of the cuvette into the T25
flask containing R10. Rinse the cuvette once with fresh R10 to
maximize recovery of electroporated cells.
9. Place the cells in a 37 °C CO2 incubator until further use.
3.2 Surface 1. Allow cells to rest for at least 3–4 h before analyzing surface
Detection of CAR expression.
on Electroporated T 2. Count and collect an aliquot of about 150,000 cells in a FACS
Cells (see Note 4) tube in a total of 3 ml. Add additional FACS buffer if needed.
3. Centrifuge cells at 300 × g for 5 min at 4 °C, discard superna-
tant, and carefully resuspend cell pellet in 3 ml FACS buffer.
Centrifuge the tube again and repeat this wash step one more
time with fresh FACS buffer.
4. Resuspend cell pellet in 10 μg of primary antibody diluted in a
total of 100 μl FACS buffer. Incubate on ice for 45 min.
5. After the incubation period, add 3 ml FACS buffer and centri-
fuge the tube to wash off unbound antibody. Repeat this wash
one more time with fresh FACS buffer.
6. If the primary antibody was pre-conjugated to a fluorescent dye,
skip to step 8. If using a non-conjugated primary antibody,
resuspend cell pellet in 1 μg of secondary antibody diluted in a

total of 100 μl FACS buffer and incubate on ice for 15 min.
7. Following the incubation, repeat washes twice as performed
earlier.
8. Finally, resuspend the cells in a desired volume and analyze
samples on a flow cytometer as shown in Fig. 1.
3.3 CAR T Cell 1. After verifying CAR expression and cell viability, collect the
Stimulation (see desired number of cells to be stimulated. Add R10 if required
Note 7) to bring the final cell concentration of 0.8–1×106 cells/ml (see
Note 5).
2. Typical bead to cell ratio for optimal stimulation is 3:1. This
ratio may vary based on the affinity and activation threshold of
the scFv used in the CAR. Calculate the total number of beads
required for the desired number of CAR-positive T cells, and
collect it in an appropriately sized tube.
3. Wash off any bead-storage buffer by applying the beads against
a magnet and rinsing the beads with fresh R10. At least three
rinses are recommended.
4. Finally, add the beads to the cells.
5. Culture the cells in a 37 °C CO2 incubator for desired time
periods. For long-term cultures, certain cell types may require
exogenous supply of growth cytokines.
Fig. 1 CAR surface expression. CAR expression on T cell surface as measured at different time points post
gene transfer. Cells electroporated without any mRNA (mock) serve as a staining control
4 Notes
1. RNA introduction into target cells can be carried out using

commercially available electroporation instruments, including,
but not limited to, ECM830 Square Wave Electroporator
(Harvard Apparatus BTX, MA, USA), Amaxa Nucleofector-II
(Amaxa Biosystems, Cologne, Germany), Gene Pulser Xcell
(Biorad, Denver, CO, USA), or Multiporator (Eppendorf,
Hamburg, Germany).
2. The level of CAR expressed on the surface can be titrated by
varying the amount of mRNA used in the gene transfer proto-
col (Fig. 2).
3. RNA transfection can also be carried out using other methods
of gene transfer, including, but not limited to, lipofection,
polymerase encapsulation, peptide-mediated transfection, or
gene guns.
4. Transfection efficiency and expression of CAR mRNA can be
measured by any other method including Northern analysis,
Western blot, or quantitative real-time PCR.
5. In vitro culture of certain cell types may require culture media
supplemented with cytokines such as IL2, IL7, IL15, etc.
6. For short-term signaling analysis, stimulate cells for desired
time periods with antigen-coated beads and collect and lyse
cell pellet for Western blot analysis (Fig. 3). Alternatively,
Fig. 2 Titration of CAR densities. Surface expression of electroporated CAR mRNA showing gradual increase of
mean fluorescence intensities with corresponding increase in mRNA amounts
Fig. 3 CAR-specific signals induced in CAR T cells (see Note 6). Phosphorylation

of a distal signaling protein (Erk) following stimulation with an anti-idiotype
against the CAR scFv at specified time points. T cells electroporated without any
mRNA (mock) serve as a stimulation control
Fig. 4 Expansion profile of CAR T cells. CD19 28ζ CAR T cell growth recorded
post stimulation with an anti-idiotype against the anti-CD19 scFv and cultured in
the presence of IL7 and IL15. T Cells not expressing CARs (mock) serve as a
stimulation control
signaling events in stimulated cells can be monitored by flow

cytometry-based methods.
7. This method can be utilized for long-term expansion of T cells
in culture. Figure 4 shows a sample growth curve of CD8+ T
cells electroporated with a CD19-BBz CAR cultured with
anti-idiotype beads against CD19 and in the presence of 10
ng/ml of IL7 and IL15 cytokines each.
8. This protocol can be extended to study signaling and perform

in vitro expansion of other T cell subset including CD4+ T
cells, naïve T cells, T regulatory cells, Th-17 cells, as well as
anergized T cells and stem cells.
9. This protocol can also be applied to other lymphocytes includ-
ing, but not limited to, NK, NKT, and B cells.
References
1. Norcross MA (1984) A synaptic basis for 6. Frauwirth KA, Thompson CB (2002)

T-lymphocyte activation. Ann Immunol (Paris) Activation and inhibition of lymphocytes by
135D(2):113–134 costimulation. J Clin Invest 109(3):295–299.
2. Smith-Garvin JE, Koretzky GA, Jordan MS doi:10.1172/JCI14941
(2009) T cell activation. Annu Rev Immunol 7. Trickett A, Kwan YL (2003) T cell stimula-
27:591–619. doi:10.1146/annurev.immunol. tion and expansion using anti-CD3/CD28
021908.132706 beads. J Immunol Methods 275(1–2):
3. Gross G, Waks T, Eshhar Z (1989) Expression 251–255
of immunoglobulin-T-cell receptor chimeric 8. Noel C, Florquin S, Goldman M, Braun MY
molecules as functional receptors with (2001) Chronic exposure to superantigen
antibody-type specificity. Proc Natl Acad Sci U induces regulatory CD4(+) T cells with IL-10-
S A 86(24):10024–10028 mediated suppressive activity. Int Immunol
4. June CH, Levine BL (2015) T cell engineering 13(4):431–439
as therapy for cancer and HIV: our synthetic 9. Collette Y, Benziane A, Razanajaona D, Olive
future. Philos Trans R Soc Lond B Biol Sci D (1998) Distinct regulation of T-cell death by
370(1680). doi:10.1098/rstb.2014.0374 CD28 depending on both its aggregation and
5. Kay JE (1991) Mechanisms of T lymphocyte T-cell receptor triggering: a role for Fas-FasL.
activation. Immunol Lett 29(1–2):51–54 Blood 92(4):1350–1363
Chapter 28
A Protein Expression Toolkit for Studying Signaling

in T Cells
Ana Mafalda Santos, Jiandong Huo, Deborah Hatherley,
Mami Chirifu, and Simon J. Davis
Abstract
Innate and adaptive cellular immunity is dependent on interactions of cell surface receptors that initiate
signaling, resulting in the formation of the immunological synapse and targeted delivery of effector func-
tions. There has been considerable interest over the past 30 years in methods for isolating the extracellular
regions of these receptors and components of the cytoplasmic signaling networks. This chapter describes
our current protein expression toolkit used for structural studies of signaling proteins and the functional
reconstitution of model cell surfaces, which comprises both bacterial and mammalian cell-based protein
expression methodologies.
Key words Inclusion bodies, Protein folding, Glycosylation, Affinity, Chromatography, Receptors,
Adhesion
1 Introduction
It is now well established that differentiated T-cell behavior is con-

trolled for the most part by events occurring at their surfaces, trig-
gered by receptor interactions of soluble effectors (e.g. cytokines)
or of adhesion molecules and signaling receptors with their cell-
bound ligands. The emerging view of T-cell signaling is that it
depends especially upon the local reorganization and interactions
of a fairly complex ensemble of surface proteins. Although the very
earliest events are still disputed [1], it has been proposed that sig-
naling could begin, within seconds, with the local, physical separa-
tion of antagonistic signaling proteins [2, 3], followed, within
minutes at most, by the accumulation of receptors into sub
μm-scale protein assemblies called microclusters considered widely
to promulgate signaling [4, 5]. What is very clear is that, on the
minutes-to-hours timescale, signaling in T cells culminates in the
formation of the now familiar, μm-scale, bull’s-eye-like structure
called the “immunological synapse” [6, 7].
451
452 Ana Mafalda Santos et al.
These developments mean that, rather than attacking the

signaling problem using classical enzymological and protein–protein
interaction-based assays (although the value of properly quantita-
tive approaches cannot be overestimated [8]), the modern approach
to understanding signaling in T cells is mostly an imaging-based
one. The application of super-resolution fluorescence-based imag-
ing to the quantitative analysis of early signaling events at reconsti-
tuted artificial and model cell surfaces, backed up by structural and
other biophysical data and coupled with, perhaps, somewhat more
qualitative observations at bona fide cell–cell contacts, is especially
promising.
The pace of discovery will of course depend on the ease with
which these model systems can be manipulated (most conveniently
transformed cell lines treated cautiously) and the quality of the
other reagents needed (most frequently soluble proteins for inser-
tion into supported bilayers or for binding studies or structural
work). Reliable approaches for expressing green fluorescent
protein-tagged constructs in or on almost any cell have been in
existence for almost two decades [9] and expression systems for
producing soluble proteins in, e.g., Chinese hamster ovary cells for
more than three [10]. The advent of CRISPR-based gene editing
[11] now means that the interpretation of data need not be obfus-
cated by the expression of endogenous proteins, and the develop-
ment of tools for the specific fluorescent dye labeling of proteins in
situ, e.g. with HaloTag® technology [12], has significantly increased
the flexibility of protein labeling and the brightness of the associ-
ated adducts.
Here, we present the suite of tools in current use in our labora-
tory for making the soluble proteins and fluorescent protein-
expressing cell lines needed for studying signaling in T cells.
Whereas in the past the production of high-quality, soluble pro-
teins involved months of work, lentivirus-based expression, in
addition to allowing the facile labeling of signaling receptors and
proteins in situ, can yield tens of milligrams of secreted protein in
just a few weeks. Lentiviral expression is therefore now at the heart
of our activities, but for completeness, we describe our in-house
methods for bacterial expression (given their semi-high-throughput
nature) and for Chinese hamster ovary-based heterologous expres-
sion because of the prodigious amounts of very high-quality anti-
bodies and other proteins these cells are capable of producing.
2 Materials
2.1 Protein 1. Luria–Bertani (LB) medium.

Production in Bacterial 2. LB Broth with agar tablet microbial growth medium.
Systems
3. Ampicillin.
A Protein Expression Tool-Kit 453
4. IPTG.
5. Six-well cell culture plate.
6. Petri dishes.
7. Falcon sterile conical centrifuge tubes, 15 and 50 ml.
8. Plasmid Miniprep Kit (e.g., PureLink® HiPure, Thermo Fisher
Scientific).
9. Rosetta™ 2(DE3) pLysS Competent Cells (Novagen).
10. Chloramphenicol.
11. Tryptone.
12. Yeast extract.
13. Magnesium sulfate.
14. Sodium selenite.
15. Sodium hydroxide.
16. Sodium phosphate dibasic.
17. Potassium phosphate monobasic.
18. Ammonium chloride.
19. Sodium sulfate.
20. Glucose.
21. α-Lactose monohydrate.
22. Chemicals for making up 1000× trace metals (see Note 1).
23. Tween 20 detergent.
24. EDTA-free protease-inhibitor cocktail tablet (e.g., cOmplete™,
Roche).
25. DNase I.
26. Trizma® base for making Tris buffer.
27. Sodium chloride.
28. Imidazole.
29. Minisart® syringe filters, 0.22 and 0.45 μm pore size.
30. 1 ml HisTrap FF column suitable for Akta™ (GE Healthcare).
Centrifugal ultrafiltration units (e.g., Amicon®, Merck
31.
Millipore Ltd.).
32. N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid (HEPES)
sodium salt.
33. Sodium azide.
34. His-tagged 3C protease.
35. Tris(2-carboxyethyl)phosphine (TCEP).
36. Dithiothreitol (DTT).
37. Triton X100 detergent.
38. Guanidine hydrochloride.
39. Urea.
40. Ethylenediaminetetraacetic acid (EDTA).
41. 2-(N-morpholino)ethanesulfonic acid (MES).
42. Bis–Tris Propane (BTP).
43. Polyethylene glycol (different MWs, see Table in Subheading
3.3.2).
44. Non-detergent sulfobetaines (NDSB).
45. Brij58 detergent.
46. l-Arginine.
47. Metal ions for refolding screens (see Table in Subheading 3.3.2).
2.2 Protein 1. DNA: pEE14 vector (Lonza), pHR, pMDG, and p8.91 (these
Production lentiviral vectors were a kind gift from Prof AJ Thrasher at
in Mammalian UCL, London, UK1).
Systems 2. Plasmid Miniprep Kit (e.g., PureLink® HiPure, Thermo Fisher
Scientific).
3. Human embryonic kidney 293T cells (HEK 293T, ATCC no.
CRL-1573).
4. Chinese hamster ovary KI cells (CHO-KI, ATCC no. CCL-61).
5. Dulbecco’s modified Eagle’s (DMEM) medium supplemented
with 10% fetal bovine serum (FBS) (v/v), 2 mM L-glutamine,
1 mM sodium pyruvate, and penicillin–streptomycin–
neomycin.
6. Roswell Park Memorial Institute 1640 (RPMI-1640) medium
supplemented with 10% FBS (v/v), 2 mM l-glutamine, 1 mM
sodium pyruvate and penicillin–streptomycin–neomycin.
7. Dulbecco’s modified Eagle’s (DMEM) medium without
l-glutamine required for the CHO-K1 GS expression system.
8. Dialyzed Fetal bovine serum (FBS).

9. 50× nucleoside and amino acid supplement (see Note 2).
10. Trypsin 10× solution.
11. l-Methionine sulfoximine (MSX).
12. Nunc T25 cm2, T75 cm2, T175 cm2 flasks and 6-, 96-, and
24-well flat-bottom plates.
13. 25 kDa branched form of polyethylenimine (PEI).
14. GeneJuice® transfection reagent (Merck & Co., Inc.) or similar
transfection reagent.
15. 10-layer Nunc™ Cell Factory™ with a growing surface of
6320 cm2.
16. Nunc™ Cell Factory system start-up kit (Fisher, cat. 170769)
and 1 L Pyrex aspirator bottle are used with the Nunc™ Cell
Factory™.
17. 1.0 M sodium butyrate.

18. Gas cylinder containing 10% CO2/air mixture.
19. Trizma hydrochloride solution, 1 M, pH 8.
20. Minisart® syringe filters, 0.22 and 0.45 μm pore size.
21. HEPES sodium salt.
22. Sodium azide (NaN3).
23. Ni-NTA agarose beads for protein purification.
24. Disposable chromatography column for 2 ml bed volume and
5 ml reservoir.
25. Imidazole, purity ≥ 99.5%.
26. Flow cytometer analyzer with 405, 488, and 640 nm lasers.
27. 37% formaldehyde.
28. Instant Blue™ or other Coomassie® Blue-based staining
solution.

29.
Centrifugal ultrafiltration units (e.g., Amicon®, Merck
Millipore Ltd.).
3 Methods
3.1 Construct Design For general cloning methods, readers are directed to other vol-
Considerations umes in this series. For producing soluble proteins, DNA con-
structs are generated that end immediately prior to sequence
encoding the transmembrane region. The use of native signal pep-
tide sequences is also preferable to trying to guess the site of signal
peptide cleavage. The 5′ sequence (ctc aag cag gcc acc) immedi-
ately upstream of the rat CD4 initiating methionine has proved to
be a reliable proxy for the Kozak sequence. Truncating proteins
between extracellular domains and modules is generally to be
avoided except when required, e.g., for structural work or when
single domains need to be isolated for binding studies. In the latter
case, experience indicates that modification of the gene of interest,
e.g., variation of sequence length at the domain boundaries, is gen-
erally a better approach for obtaining well-expressed proteins than
trying many different fusion tags, e.g., thioredoxin, MsyB, MBP
(maltose-binding protein), and trigger factor. Expression using
these fusion partners not only tends to result in formation of pre-
cipitates upon, e.g., 3C protease cleavage, but often very low yields
of tag-free product. For proteins that will be inserted into bilayers,
the “double-hexahistidine tag” method of Khan et al. [13] is rec-
ommended as proteins tagged with this sequence dissociate very
slowly following attachment to nickel-bearing lipids (<20% disso-
ciation in 2 h).
3.2 Expression Bacterial expression can be used to produce large amounts of pro-
of Soluble Proteins tein, either in the form of soluble cytoplasmic or periplasmically
Using Bacterial secreted proteins or as inclusion bodies whereupon the protein of
Expression interest will need to be refolded. However, not all proteins can be
expressed in bacteria, and the proteins that are eventually produced
generally lack posttranslation modifications such as glycosylation.
Extra care needs to be taken over confirming that the protein is
well folded (stoichiometric monoclonal antibody binding and gel
filtration are the best tests for this). The excellent “pET” vector
series (http://www.merckmillipore.com) covers each of the differ-
ent routes to expression in bacteria, also allowing the addition of
numerous tags to aid expression and/or purification.
3.2.1 Protein Expression 1. The sequence-verified expression plasmids are used to trans-
and Preparation of Cell form Rosetta™ 2(DE3) pLysS Competent Cells (Novagen),
Pellets and Cell Lysate according to the manufacturer’s instructions. Briefly, the ali-
quoted competent cells are thawed on ice. One microliter of
plasmid (~100 ng/μl) is used to transform a 35 μl competent
cell aliquot, followed by incubation on ice for 30 min. The
cells are then treated with heat-shock for 30 s at 42 °C in water
bath. Thereafter, the cells are kept on ice for 2 min, followed
by addition of 350 μl LB medium to each tube.
2. The cells are incubated for 1 h at 37 °C in an incubator.
Transformed cells are transferred onto and spread over the LB
agar plate supplemented with 50 μg/ml ampicillin (selection
for the plasmid) and 34 μg/ml chloramphenicol (selection for
the Rosetta™ cells). The plates are allowed to dry off before
turning upside down and incubated overnight at 37 °C.
3. Individual colonies are used to inoculate 10 ml LB medium
supplemented with the appropriate antibiotics in 50 ml sterile
disposable tubes with crew caps that are loosened and secured
with tape to allow air circulation while maintaining sterility.
The bacterial culture is shaken at 225 rpm overnight at 37 °C.
4. One milliliter of each overnight culture is used to inoculate
40 ml of pre-warmed LB medium in a T75 Corning® cell cul-
ture flask shaken at 225 rpm for 2–3 h at 37 °C until turbidity
is seen; this comprises the pre-culture.
(a) For IPTG induction of expression, the entire pre-culture
is used to inoculate pre-warmed 1 L autoclaved LB
medium, in a 2 L flask, supplemented with appropriate
antibiotics. The diluted cultures are grown at 37 °C with
shaking at 225 rpm to an OD600 of ~0.6 before addition of
ITPG to a final concentration of 1 mM and reducing the
temperature to 20 °C. The flasks are then incubated for a
further 22 h with shaking at 225 rpm.
(b) Alternatively, all the pre-culture is used to inoculate 1 L
pre-warmed autoclaved auto-induction medium (see Note 1),
in a 2 L flask, supplemented with appropriate antibiotics.

The diluted cultures are grown at 37 °C with shaking at
225 rpm for ~4 h until the cultures begin to look moder-
ately cloudy, before reducing the temperature to 20 °C
and shaking for a further 20 h.
5. The cultures are collected in a disposable conical centrifuge
bottle and harvested by centrifugation at 3200 × g for 20 min
in a swinging bucket rotor, and pelleted cells are frozen and
stored at −80 °C.
6. The cells are defrosted and fully resuspended in resuspension
buffer (see Note 3), followed by lysis of the cells under condi-
tions of high pressure using, e.g., a Constant Systems TS series
cell disruptor (Constant Systems Ltd.). Briefly, in the case of
the Constant Systems unit, the system is firstly washed with a
two-chamber volume of deionized water at 10 kpsi prior to
loading with the sample. The sample is then processed at 30
kpsi and collected into a suitable flask (the connection between
the outlet and the flask opening is sealed to prevent spread of
bioaerosols).
7. The cell lysate is transferred to 30 ml centrifuge tubes with
centrifugation at 15,000 × g for 30 min in a fixed angle rotor
at 4 °C.
8. The supernatant is filtered through 0.45 μm Minisart® Syringe
Filters to remove residual impurities.
3.2.2 Automated 1.
The ÄKTA Express protein purification system (GE
Histidine-Tagged Soluble Healthcare), which operates at 4 °C, is used for automated
Protein Purification protein purification. A 16/60 HiLoad Superdex 75 column
(or Superdex 200 column, depending on the size of the pro-
tein of interest) is used to perform gel filtration chromatogra-
phy. The column is pre-equilibrated in Gel Filtration Buffer.
2. The system is firstly washed with Wash Buffer (50 mM Tris pH
7.5, 150 mM NaCl, 20 mM imidazole) and the clear lysate and
then applied onto a 1 ml HisTrap FF column attached to the
ÄKTA Express system.
3. The HisTrap FF column is washed with 30 ml Wash Buffer,
and the protein is eluted with 7.5 ml Elution Buffer (50 mM
Tris pH 7.5, 150 mM NaCl, 500 mM imidazole). The eluate
is automatically and subsequently loaded onto the gel filtration
column.
4. Fractions eluting from the gel filtration column containing
proteins are analyzed on SDS-PAGE gels and visualized with
Instant Blue™ protein stain. Protein fractions are collected and
concentrated with Amicon® centrifugal ultrafiltration units of
appropriate molecular weight cutoff values, according to the
manufacturer’s instructions, until the desired level of concen-
tration is achieved.
5. Prior to use, the preservative in the membrane of the concen-

trator is removed by passing Gel Filtration Buffer (10 mM
HEPES pH 7.5, 150 mM NaCl, 0.005% Sodium azide)
through the membrane by centrifugation for 2 min at 4 °C.
The concentrated protein aliquots are frozen in liquid nitrogen
and stored at −80 °C.
6. N-terminal fusion tags (e.g., thioredoxin), where applicable,
are removed by overnight incubation of the protein with His-
tagged 3C protease, according to the manufacturer’s instruc-
tions, supplemented with 1 mM TCEP or DTT, at 4 °C. The
3C protease and any uncleaved protein are removed by passing
the sample through a 1 ml HisTrap FF column.
3.3 Preparation Some proteins become denatured and accumulate in inclusion

and Refolding bodies within bacteria. If this is the case, then a modified culture
of Proteins and lysate preparation is needed to collect the inclusions, which
from Inclusion Bodies would be in discarded fractions above. The inclusions are then
washed, solubilized under strongly denaturing conditions, and
then refolded.
3.3.1 Bacterial Culture 1. One transformed colony is used to inoculate 10 ml LB medium
for Denatured Inclusion supplemented with the appropriate antibiotics in a 50 ml dis-
Bodies posable tube above. The bacterial culture is shaken at 225 rpm
for overnight at 37 °C and used to prepare a 40 ml pre-culture
(see Subheading 3.2.1).
2. The pre-culture is used to inoculate pre-warmed 1 L auto-
claved LB medium, in a 2 L flask, supplemented with
appropriate antibiotics. The diluted cultures are grown at
37 °C with shaking at 225 rpm to an OD600 of ~0.6 before
addition of ITPG to a final concentration of 1 mM.
3. The culture is allowed to grow 37 °C for a further 4 h (or
more) with shaking at 225 rpm.
4. The cell pellet is collected and processed as described in
Subheading 3.2.1, steps 5 and 6.
5. The cell lysate is transferred to 30 ml NALGENE centrifuge
tubes and centrifuged at 15,000 × g for 30 min at 4 °C; the
supernatant is discarded.
6. The pellet is resuspended in 20 ml IBR Buffer A (see Note 4)
by grinding with a 5 ml pipette, pipetting up and down, and
vortexing. The inclusion bodies are then pelleted at 15,000 ×
g for 10 min at 4 °C. The sample is processed on ice, and this
washing step is repeated another three times or until the pellet
becomes white.
7. The pellet is finally resuspended in 20 ml IBR Buffer B (50 mM
Tris pH 8, 100 mM NaCl), followed by spinning at 15,000 × g
for 10 min at 4 °C, which removes any detergent residues.
8. The purified inclusion bodies are then dissolved in 20 ml of

denaturing buffer (see Note 5).
9. The sample is left on ice on a shaker for at least an hour to
allow the pellet to fully dissolve.
10. The sample is then centrifuged at 15,000 × g for 20 min at 4 °C.
11. The supernatant is carefully harvested by decanting, owing to
the fact that the color of the pellet is very similar to the
supernatant.
12. The concentration of the inclusion body solution is estimated at
A280, and the samples are stored as 10 mg/ml aliquot at −20 °C.
3.3.2 The Refolding is achieved by removing the chaotrope via buffer

Refolding Screen exchange. Since there is no “universal” refolding buffer, i.e., the
composition of the necessary refolding buffer is strongly protein
dependent, a refolding screen consisting of many different condi-
tions is usually recommended. Refolding screens can be pur-
chased commercially or prepared in-house. Examples of
components varied in screening buffers are listed in the table
below. In order to test several components and compositions
simultaneously and generate an optimal set of experiments
wherein exploration of the variables is maximized, a fractional
factorial approach using software called SAmBA [14] can be
applied to generate a screen with 96 different buffer conditions in
a microtiter plate format.
Reducing
Buffer pH Ionic Amphiphilic Surfactants agent
(50 mM) strength components (100 mM) (10 mM) Additives
(1) MES— (0) Null (0) Null (0) Null (0) Null (0) Null
pH 6
(2) Tris— (1) NaCl— (1) Glycerol (1) NDSB-195 (1) TCEP (1) l-Arginine—800
pH 7 200 mM 20% (v/v) mM
(3) Tris— (2) NaCl— (2) bPEG-200 (2) NDSB-201 (2) Glucose—500
pH 8 350 mM 0.05% (w/v) mM
(4) aBTP— (3) NaCl— (3) PEG-1500 (3) NDSB-221 (3) EDTA—
pH 9 500 mM 0.05% (w/v) 10 mM
(4) PEG-4000 (4) NDSB-256 (4) cCocktail metal
0.05% (w/v) ions
(5) PEG-10000 (5) Brij58
0.05% (w/v) (0.5%)
a
BTP: Bis–Tris Propane
b
PEG: Polyethylene glycol
c
Cocktail metal ions: 2 mM of CaCl2, MgCl2, ZnCl2, NiCl2, CuSO4, CoSO4
1. The screening plates are stored at −20 °C and defrosted prior

to use.
2. 200 μl of each screening buffer is dispensed into the corre-
sponding well of a microtiter plate. 20 μl of 10 mg/ml inclu-
sion body solution is added to the screening plate using a
multichannel pipette and shaken with an orbital microtiter
plate shaker at 350 rpm for 5 min.
3. Evaluation of the buffer conditions is undertaken by measuring
the light absorbance at 350 nm. Conditions that are potentially
suitable for refolding are characterized by having an OD < 0.05.
3.3.3 Refolding by Rapid 1. DTT is added to an aliquot of denatured protein to a final

Dilution concentration of 10 mM and left on ice for an hour to ensure
that all the disulfides are disrupted.
2. 20 mg of the inclusion body mix is added to 200 ml of pre-
chilled refolding buffer in a 500 ml flask using a P200 pipette
over a period of about 30 min (see Note 6).
3. When addition of the sample is completed, the solution is
transferred to a magnetic stirrer, with the flea spinning at low
speed, and kept at 4 °C overnight.
4. The refolding mix is then centrifuged at 3500 × g for 20 min at
4 °C, after which the supernatant is harvested and concentrated
to ~7 ml (the maximum volume that can be loaded into the
ÄKTA at any one time) using Amicon® Centriprep concentra-
tors or similar large-capacity centrifugal ultrafiltration devices.
5. The sample is then purified by size exclusion chromatography
in the HPLC column and processed as described in Subheading
3.2.2, steps 4 and 5.
3.3.4 Refolding 1. 3 M GuHCl solution or 4 M urea solution (see Note 7) is
by Dialysis added to the solubilized inclusion bodies, the final concentra-
tion of the protein being around 0.5 mg/ml (assuming pro-
tein size is in the region of 25 kDa)
2. 1 M DTT is added to the protein solution to reach a final
concentration of 5 mM. The solution is poured into a beaker,
covered with foil (because DTT is light sensitive), and incu-
bated at 25 °C for 1 h.
3. Dialysis is performed in the cold room. The principle of the
dialysis method is that the concentration of the GuHCl or urea
solution is gradually decreased. The length of each dialysis
cycle is of 12–24 h which is also dependent on the refolding
buffer volume (see Note 8), as follows:
GuHCl concentration: 1.5 M to 0.5 M to 0 M to 0 M
Urea concentration: 2.0 M to 0.5 M to 0 M to 0 M
4. After dialysis, the protein sample is filtered using a 0.45 μm
filter and is kept at 4 °C until purification.
3.4 Stable The Lentivirus most widely used for heterologous protein expres-
Expression of Soluble sion [9] is based on HIV-I, which is from the Lentiviruses sub-
and Transmembrane class of retroviruses. The lentiviral system allows infection of both
Proteins Using dividing and nondividing cells and has been very useful in achiev-
Lentiviruses ing expression in cell lines previously very difficult to transfect
using traditional approaches, such as introduction of plasmids by
calcium phosphate or electroporation methods. The virus RNA is
actively transported to the nucleus where it is reverse-transcribed
into DNA, which is then integrated into the genome of the target
cell. Although, in what follows, the stable expression approach is
described at length, the lentiviral expression system can be also
used for transient gene expression, most commonly in HEK
293T cells.
The HIV genome of this lentivirus has been engineered into
three different vectors: pHRSIN-CSGW encodes the gene of inter-
est—goi—with the HIV-1 LTR; p8.91 encodes gag, pol, rev, and
tat; and pMD-G encodes the VSV-G envelope protein. The meth-
ods used are described below, and these should be carried out in a
biosafety cabinet with appropriate training and facilities.
3.4.1 Lentiviral 1. Day 0—Adherent HEK 293T cells are plated out into a six-
Transduction to Produce well plate, 6 × 105 cells/well in 2 ml of complete medium
Cells Stably Expressing (DMEM supplemented with 10% FBS, 2 mM l-glutamine,
a Protein Intracellularly or and penicillin–streptomycin–neomycin).
a Receptor at the Cell 2. Day 1—Cells should be at ~ 90–95% confluency. Plasmid DNA
Surface is prepared in advance using PureLink® HiPure Plasmid
Miniprep Kit (or similar low endotoxin methods), and only
preparations with absorbance ratios (A260nm:A280nm) of 1.8–2.0
are used for transfection.
(a) For each transfection (per well), 0.5 μg of DNA for each
of the three plasmids (pMD-G, p8.91, and pHR + goi) is
used, giving a total of 1.5 μg of DNA in a final volume of
20 μl.
(b) To 100 μl of DMEM with no supplements in a small tube,
4.5 μl of GeneJuice® reagent (or equivalent transfection
reagent) is added dropwise; the solution is mixed by
shaking.
(c) The 104.5 μl GeneJuice®/DMEM mix is added to the
DNA solution.
(d) The DNA/GeneJuice®/DMEM mix is incubated at room
temperature for 15–30 min to allow DNA–GeneJuice®
complexes to form.
(e) During the incubation time, the medium from the six-well
plate is carefully removed and replaced with new complete
medium (see Note 9).
(f) After incubation, the 124.5 μl of DNA/GeneJuice® com-

plexes is added dropwise to the cells.
(g) The six-well plate is returned to the incubator and incu-
bated at 37 °C, 5% CO2.
3. Day 4—72 h post transfection.
(a) The tissue culture supernatant containing the viral parti-
cles is harvested and placed into a 50 ml sterile conical-
bottom centrifuge tube.
(b) The tissue culture supernatant is spun at 2000 × g for
5 min and then filtered through 0.45 μM Sartorius filter to
remove potential cell debris.
(c) In a T25 flask, the cells to be infected are prepared; 1 ×
106 suspension cells are used in 4 ml of complete medium
(RPMI supplemented with 10% FBS, 2 mM l-glutamine,
1 mM sodium pyruvate and penicillin–streptomycin).
(d) For the infection, the tissue culture supernatant contain-
ing the viral particles is added to the T25 with cells to be
infected (see Note 10).
4. Day 9—5 days postinfection. For confirming the expression
levels for the g.o.i.-encoded receptors or intracellular mole-
cules, cells are washed and resuspended in PBS containing
0.05% w/v sodium azide (NaN3).
(a) For membrane proteins, staining is performed by incuba-
tion of 0.5–1 × 106 cells per condition with a mAb (20
μg/ml) for 45 min on ice, followed by two washes with
PBS/0.05% NaN3. Staining is repeated with an appropri-
ate secondary antibody if needed. Cells are resuspended in
PBS/0.05% NaN3/2% formaldehyde.
(b) For intracellular proteins, 0.5–1 × 106 cells per condition
are resuspended in 300 μl of PBS/0.05% NaN3/1% form-
aldehyde and left at room temperature for exactly 15 min
followed by a wash with PBS/0.05% NaN3. Staining is
performed with the appropriate antibody diluted in
PBS/0.05% NaN3/0.5% Saponin/5% FBS (see Note 11)
for an hour on ice. Two washes are performed with
PBS/0.05% NaN3/0.5% Saponin/5% FBS. Staining is
repeated with an appropriate secondary antibody if
needed. Cells are resuspended in PBS/0.05% NaN3.
(c) When the protein of interest is cloned and expressed as a
SNAP-tag® (New England Biolabs) or HaloTag®
(Promega) fusion protein, labeling is done with a fusion-
tag substrate of choice, most typically a dye molecule (see
Note 12). The substrate of choice is diluted in complete
medium at the concentration suggested by the manufac-
turer, added to 1 × 106 cells, and incubated at 37 °C, 5%
CO2 for 30 min. The cells are washed with complete

medium twice, after which 1 ml of complete medium is
added to the cells, and cells are again incubated at 37 °C,
5% CO2 for 30 min to completely wash out potential
unbound substrate. Finally, two washes are performed
with PBS/0.05% NaN3. Cells are resuspended in
PBS/0.05% NaN3/2% formaldehyde.
(d) If the protein of interest is already being expressed as a
fusion protein tagged with a fluorescent protein, 0.5–1 ×
106 cells per condition are washed twice with PBS/0.05%
NaN3 and resuspended in PBS/0.05% NaN3/2%
formaldehyde.
(e) Analysis is performed by flow cytometry using an appro-
priate analyzer.
3.4.2 Stable Expression 1. Day 0—Adherent HEK 293T cells are plated out into a T75
of Secreted Proteins flask, 4 × 106 cells in 20 ml of complete medium (DMEM sup-
plemented with 2% FBS and 2 mM l-glutamine).
2. Day 1—Cells should be at ~90–95% confluency. Plasmid DNA
is prepared in advance using PureLink® HiPure Plasmid
Miniprep Kit (or similar low endotoxin methods), and only
preparations with absorbance ratios (A260nm:A280nm) of 1.8/2.0
are used for transfection. All reagents (media, PEI, and DNA)
should be at room temperature.
(a) For each T75 transfection, 6.5 μg of each one of the three
plasmids (pMD-G, p8.91, and pHR + goi) is used, giving
a final total of 19.5 μg of DNA.
(b) To 2.5 ml of DMEM with no supplements in a 50 ml
Falcon tube, plasmid DNA (19.5 μg/flask) and 40 μl/
flask of PEI (stock concentration: 1 mg/ml) are added
sequentially and mixed by shaking.
(c) The DMEM/DNA/PEI mix is incubated at room tem-
perature for 15–30 min to allow DNA–PEI complexes to
form.
(d) An extra 8 ml of complete medium is added to the 50 ml
Falcon tube with the transfection mixture.
(e) The medium from the cells in the T75 flask is removed
and replaced with the transfection mixture; the flask is
returned to the incubator.
(a) The medium containing the viral particles (tissue culture
supernatant, TCS) is harvested from the T75 flask and fil-
tered through a 0.45 μm Sartorius filter.
(b) The tissue culture supernatant is stored at 4 °C.
(c) 10 ml of fresh complete medium is added to the infected

cells.
(d) Stock HEK 293T or CHO-KI adherent cells are split into
T175 flasks, 8 × 106 cells per flask in 40 ml of complete
medium.
(a) The second volume of medium from the transfected cells
is harvested and spun at 2000 × g for 5 min and then fil-
tered through 0.45 μm Sartorius filter to remove potential
cell debris.
(b) Medium from the T175 flask seeded on day 3 is discarded
and replaced with 25 ml of fresh medium.
(c) The tissue culture supernatant from day 2 and 3 is pooled
and used to infect cells in the T175 flask.
(d) The day following infection, the virus-containing superna-
tant is removed, and the cells are split into two T175 flasks
using 1% FCS-containing medium (included to allow bet-
ter purification).
(e) The initially transfected cells used to produce the virus
particles can also be split into two T175 flasks for protein
production and harvesting.
5. After an additional 4–5 days, the supernatant containing the
secreted protein is harvested (see Note 13), and fresh medium
added. Harvest every 2–4 days depending on expression levels.
6. The harvested tissue culture supernatant is centrifuged at
3000 × g for 10 min to discard any cell debris contaminant.
The supernatant is kept at 4 °C until purification (see Note 14).
7. See Note 15 for variations of the method helpful for weakly
expressing proteins.
3.5 The CHO GS Gene The following method describes the generation of stable Chinese
Expression System™ hamster ovary-K1 (CHO-K1) cell lines expressing soluble recom-
(Lonza) for Expressing binant protein using the glutamine synthetase (GS)-based expres-
Soluble Recombinant sion system available from Lonza (GS Gene Expression System™;
Protein www.lonza.com). CHO-K1 cells are normally grown in medium
supplemented with glutamine; however, if glutamine is absent, the
cells upregulate their endogenous GS gene. l-Methionine sulfoxi-
mine (MSX) competitively inhibits GS and is therefore lethal to the
cells under these conditions. The pEE14 expression vector con-
tains the human cytomegalovirus promoter that controls the
expression of inserted sequences and a GS minigene (under con-
trol of the SV40 late promoter). Only those cells transfected with
enough extra copies of the vector will survive in the presence of
MSX when glutamine is absent. The cell line thus selected stably
expresses the protein of interest because the pEE14 vector inte-
grates into the host genome.
All the following procedures are carried out in a standard tissue

culture hood.
1. Prior to transfection, CHO-K1 cells are cultured in DMEM
supplemented with 10% (v/v) FBS, 2 mM l-glutamine, 1 mM
sodium pyruvate, 1× nucleoside and amino acid supplement (see
Note 2), and 50 U/ml penicillin and 50 μg/ml streptomycin.
2. Day 0—The required number of T75 flasks are seeded with
2 × 106 CHO-K1 cells in 12 ml medium. An extra flask is
included for a mock transfection, i.e., a control without DNA.
3. Day 1—Plasmid DNA is prepared in advance, and only prepa-
rations with absorbance ratios (A260 nm:A280nm) of 1.8/2.0 are
used for transfection.
(a) For each transfection, 0.5 ml of DMEM (no supplements)
is added to a sterile 1.5 ml Eppendorf tube. 30 μl of
GeneJuice® (or equivalent transfection reagent) is then
added to the DMEM followed by 10 μg of pEE14 con-
struct DNA. This is gently mixed and incubated at room
temperature for 20 minutes. For the mock transfection,
DNA is substituted with ddH2O.
(b) The transfection mix is added dropwise to the T75 flasks
and the cells allowed to recover overnight.
(a) The cells from each T75 flask are trypsinized and resus-
pended in 5 ml of fresh DMEM supplemented as before
but with 10% (v/v) dialyzed FBS but without l-glutamine.
(b) The cells are counted and 1 × 106 cells seeded per flat-
bottom 96-well plate by pipetting 1 × 106 cells into 20 ml
medium containing MSX and dispensing 200 μl of the
cell/medium per well. The medium should contain MSX
at 30, 35, and 40 μM and the plates prepared in duplicate.
A mock plate is also set up for each concentration of
MSX. MSX has a tendency to precipitate after freezing so
after thawing the 100 mM MSX stock, vortex the tube
until it is back in solution (see Note 16).
(c) Feed the cells 1 week after plating by removing 100 μl of
medium per well and replacing it with 100 μl of fresh
medium containing MSX at the appropriate concentration.
5. Selecting transfected clones.
(a) Clones start to appear ~16 days post transfection. The TCS
is assayed for protein expression as it begins to turn yellow,
usually between days 16 and 21. The number and size of
clones per well should be recorded so that only single
clones are selected after screening for protein expression.
For testing expression, 100 μl of TCS is removed and
replaced with fresh medium containing MSX, and the TCS
then assayed for soluble protein using Western blot or

enzyme-linked immunosorbent assay. Mock TCS is
included as a negative control.
(b) Single clones expressing recombinant protein are isolated
by removing TCS from the appropriate well and adding
50 μl trypsin to the cells. Following incubation for a few
minutes, the cells are moved to 1 ml of fresh medium in a
24-well tissue culture plate. To each well in the 96-well
plate, 200 μl of fresh medium is added, so that the residual
cells can serve as a backup.
(c) Once confluent, the 1 ml cultures can be reassayed and
expanded first into a T25 flask (5 ml medium) and then a
T75 flask (15 ml).
6. Large-scale production of recombinant soluble protein
(a) Cells from a T75 flask are expanded into a T175 flask
(40 ml medium). The cells are removed from the conflu-
ent T175 flask with trypsin, spun and resuspended in a
500 ml DMEM bottle supplemented as before but with
the FBS reduced to 5% v/v. (The FBS does not need to
have been d ialyzed at this stage, as the CHO-K1 cell line
expressing recombinant protein will likely be stable, and
it will not in any case be passaged for so long that the
introduced gene will be lost.) The cell suspension is trans-
ferred into the cell factory as per manufacturer’s instruc-
tions (see Note 17), which is then gassed with a 10%
CO2/air mixture for approx. 90 s. The volume of medium
between the ten layers is distributed across the ten layers
by placing the cell factory on its side. The cell factories are
incubated either in a 10% CO2 37 °C incubator or in a 37
°C warm room.
(b) Once the cell factory is confluent, usually 3–5 days after
seeding, sodium butyrate is added to 2 mM final concen-
tration. Sodium butyrate inhibits cell growth and increases
transcription resulting in increased protein production.
The cell factory is re-gassed with 10% CO2/air mixture for
approximately one minute. At this stage, the medium may
be exchanged for fresh medium containing a smaller
amount of FBS or serum-free medium if serum affects the
purification of the recombinant protein. However, protein
expression levels may be lower in serum-free conditions.
(c) Cell factories are left undisturbed until cells stop produc-
ing protein, usually ~3–4 weeks by which time the cells are
exhausted and dying. Because FBS contains natural prote-
ase inhibitors, the cultures can be left until the medium
turns purple (i.e., there is no longer any respiration).
(d) The TCS is then decanted, centrifuged at 15,000 × g for
30 min, and filtered (0.45 μm) to remove cell debris.
Sodium azide is added to a final concentration of 0.02%

(w/v). Expression levels are assayed and recorded after
which the TCS can be stored at 4 °C indefinitely.
3.6 Hexahistidine- 1. The tissue culture supernatant kept at 4 °C (Subheadings

Tagged Glycoprotein 3.4.2, step 6 and 3.5, step 6d) is diluted with two volumes of
Purification phosphate- buffered saline (PBS)/0.05% NaN3. The flask is
gently swirled to mix. It needs to be confirmed that the pH of
the mixture is 8 (and adjusted if needed). The mixture is placed
in a conical flask.
2. A slurry of Ni-NTA agarose beads is prepared in PBS and
added to the supernatant-PBS mixture. In principle, 5 ml of
Ni-NTA agarose beads is sufficient for 1 L of the supernatant
harvested.
3. The mixture is gently stirred at 120 rpm at 4 °C (cold room)
for 4 h or overnight.
4. After the incubation with the beads, the conical flask is taken
to the bench where the purification will proceed. The flask is
left for 2 h in order for the beads to settle down.
5. The tissue culture supernatant is carefully removed without
disturbing the beads (see Note 18). All tissue culture superna-
tant is kept as it might still contain protein that did not bind to
the beads in the first incubation.
6. The remaining mixture of TCS plus beads is then poured into
an Econo-Column for collection of the beads (again all the
“flow through” should be kept).
7. The beads are washed with 20 column volumes of PBS/0.5 M
NaCl/pH 8.
8. A pre-elution of the beads starts with 10 mM imidazole (see
Notes 19 and 20); about 20 fractions of 1 ml are collected. It
is important that all fractions are collected as occasionally the
protein elutes earlier than expected. The column flow is
stopped and the optical density (OD280nm) of the 10 mM frac-
tions is measured, ideally 10 mM imidazole fractions should
be acquired until the OD drops to a background level (see
Note 21). All the fractions are kept for analysis by SDS-
polyacrylamide gel electrophoresis (SDS-PAGE).
9. Another ten fractions pre-eluting with 20 mM imidazole are
collected, followed by another ten fractions using 30 mM
imidazole. This extra step can be added to the purification to
try to improve the purity of the protein. Save the fractions and
run on SDS-PAGE gel (after evaluating the gels it can be
decided whether this step is needed or not in subsequent
extractions).
10. Final elution is done with 250 mM imidazole; 1 ml fractions
are collected. The elution will be complete when the absor-
bance reading of the fraction reduces to the background level.

Keep all the fractions for SDS-PAGE analysis.
11. Fractions containing proteins are analyzed on SDS-PAGE gels
and visualized with Instant Blue™.
12. The fractions containing protein of the expected size are
pooled and concentrated with Amicon® centrifugal ultrafiltra-
tion units with appropriate molecular weight cutoffs, accord-
ing to the manufacturer’s instructions.
13. Final purification is achieved by Sephacryl S-200 (or S-75
depending on the size of the protein) gel filtration in HBS buf-
fer (10 mM HEPES, pH 7.4, 150 mM NaCl, 0.05% w/v
sodium azide).
14. Following gel filtration, fractions are analyzed on SDS-PAGE
gels. Fractions containing the protein of interest at >95% purity
are pooled, concentrated with Amicon® Ultra centrifugal filter
units to the desired concentration and stored at 4 °C for the
short term (1–2 days) and −20 °C for the longer term.
4 Notes
1. Each liter of auto-induction medium has the following

ingredients:
Component Amount
Tryptone 12 g
Yeast extract 24 g
Deionized H2O 900 ml
25× Solution M 40 ml
25× Solution 5052 40 ml
1000× trace metals 200 μl
1 M Magnesium sulfate (MgSO4) 2 ml
20 mM Sodium selenite (Na2SeO3)—stored at 4 °C 20 μl
2 M Sodium hydroxide (NaOH) 20 ml
25× Solution M—autoclaved.
Component Final concentration

Sodium phosphate dibasic (Na2HPO4) 0.5 M
Potassium phosphate monobasic (KH2PO4) 0.5 M
Ammonium chloride (NH4Cl) 1.25 M
Sodium sulfate (Na2SO4) 0.25 M
25× Solution 5052—autoclaved.

Glycerol 12.5%
Glucose 1.25%
α-lactose monohydrate 5%
1000x trace metals—stored at 4 °C.

Iron(III) chloride (FeCl3) 50 mM
Calcium chloride (CaCl2) 20 mM
Manganese(II) chloride (MnCl2) 10 mM
Zinc sulfate (ZnSO4) 10 mM
Cobalt(II) chloride (CoCl2) 2 mM
Copper(II) chloride (CuCl2) 2 mM
Nickel(II) chloride (NiCl2) 2 mM
Sodium molybdate (Na2MoO4) 2 mM
Boric acid (H3BO3) 2 mM
Hydrogen chloride (HCl) 60 mM
2. The nucleosides and amino acids (l-asparagine and l-gluta-

mate) can be made up as a 50× supplement by dissolving
70 mg each of adenosine, cytosine, guanosine, and uridine,
24 mg of thymidine, and 20 mg each of asparagine and gluta-
mate in 200 ml ddH2O. The supplements can be difficult to
dissolve unless heated to 37 °C for 30 min with stirring. Filter
sterilize and store at −20 °C in 10 ml aliquots.
3. 100 ml of resuspension buffer is generally used to resuspend
cell pellets from 1 L cultures. Resuspension can be performed
in a small beaker with a stirring bar, preferably in a cold room.
Protease inhibitors are not usually required for the preparation
of inclusion bodies.
Resuspension buffer.

Tris pH 7.5 50 mM
NaCl 150 mM
Imidazole 20 mM
Tween 20 0.2% (v/v)
(continued)
(continued)
EDTA-free protease-inhibitor cocktail tablet 1 tablet per 100 ml
DNase I 0.01 mg/ml
4. IBR Buffer Aa

Tris pH 8.0 50 mM
NaCl 150 mM
Sodium azide 0.1%
Triton X100 0.5% (v/v)
a
IBR Buffer B is 50 mM Tris pH 8.0, 100 mM NaCl
5. Denaturating buffer

Tris pH 8.0 50 mM
NaCl 150 mM
6 M guanidine hydrochloride 20 mM
(GuHCl) or 8 M Urea
EDTA 10 mM
DTT 10 mM
6. Sufficient time must be left between additions of the inclusion

bodies so that the inclusion bodies are well dissolved in the
buffer before addition of another drop. The refolding buffer is
kept on ice on a magnetic stirrer, with the flea spinning at as
high speed a speed as possible.
7. 3 M GuHCl solution or 4 M urea solution is prepared by mix-
ing 6 M GuHCl solution or 8 M urea solution with resuspen-
sion buffer.
8. The volume of dialysis buffer should be about 13–15-fold
larger than the sample volume.
9. When adding the new medium to the packaging cells (HEK
293T), make sure it is added to the side of each well, so that
the cells are not displaced. If the cells to be infected are to be
grown in RPMI at this point, change the medium to complete
RPMI.
10. A titration of the amount of virus added might be required, as
the expression will be protein dependent. It is worth starting
with 2 ml of virus stock.
11. Higher dilutions of antibody than normally used for surface

staining might be required.
12. A variety of substrates can be chosen from different companies,
and the final concentration of each substrate per condition
used will greatly depend on the substrate chosen. Final con-
centrations can range from 1 to 10 μM.
13. At this stage, it is advisable to do a Western blot of the super-
natant containing the protein of interest to assess the expres-
sion level.
14. Most proteins will retain their integrity for a long time under
these conditions but be aware that this could be protein
dependent.
15. It is important to note that expression levels can vary a lot from
protein to protein. We have, in some instances, improved yields
of weakly expressing proteins obtained using the lentiviral
approach by infecting CHO-K1 cells rather than HEK 293T
cells. This is partly because CHO-K1 cells tend to last longer
in culture and also because for some proteins the expression
levels seem to be cell dependent. Also, large-scale lentiviral
infection of HEK 293T or CHO-K1 cells in Nunc™ Cell
Factories is possible. For very weakly expressing proteins (less
than 0.5 mg/liter), the CHO GS Gene Expression System™
should be used to generate stable long-term cultures secreting
the protein of interest.
16. 100 mM MSX stocks in DMEM (without l-glutamine) can be
prepared, filter sterilized (0.22 μm filter), aliquoted, and stored
at −20 °C.
17. There are several large-scale tissue culture systems available for
growing adherent mammalian cell cultures for protein produc-
tion. One method uses a 10-layer Nunc™ Cell Factories. The
Nunc™ Cell Factory is also available as an EasyFill™ version
allowing users to pour the cell suspension directly into the cell
factory without the need for any accessories such as an aspira-
tor bottle. The 10-layer Corning® CellSTACK is a similar
product with a surface area of 6360 cm2 and like the Nunc™
EasyFill™ has a wide port opening for easy filling directly from
a medium bottle. The DMEM used in cell factories contains
3.7 g/L sodium bicarbonate and requires 5–10% CO2 to
ensure optimum culture pH. Hence, it is important to monitor
the culture pH especially when using a walk-in warm room to
culture cells where CO2 levels are not elevated. Gas the cul-
tures with a 10% CO2/air mixture and replace one or both
vented caps with non-vented caps (taping the vents over is also
effective).
18. The volume of TCS that can be removed will depend on the
total volume; the most important matter is that the beads in
the bottom of the flask should not be disturbed.
19. An imidazole stock solution of 1 M can be prepared by dissolving

6.81 g imidazole in 100 ml of PBS pH 8. This can then be diluted
in PBS pH 8 to obtain 10, 20, 30, and 250 mM solutions.
20. This should remove bovine serum albumin (BSA) and nonspe-
cific proteins that are bound to the beads.
21. The absorbance of the 10 mM imidazole-eluted fractions is
expected to increase and subsequently decrease to back-
ground levels. If it is difficult to reach background levels, the
protein of interest will likely be slowly eluting in the 10 mM
imidazole wash.
References
1. van der Merwe PA, Dushek O (2011) tribution of cell-surface molecules during T cell
Mechanisms for T cell receptor triggering. Nat antigen recognition. Semin Immunol 12:5–21
Rev Immunol 11:47–55 8. Hui E, Vale RD (2014) In vitro membrane
2. Davis SJ, van der Merwe PA (1996) The struc- reconstitution of the T-cell receptor proximal
ture and ligand interactions of CD2: implica- signaling network. Nat Struct Mol Biol
tions for T-cell function. Immunol Today 21:133–142
17:177–187 9. Zufferey R, Nagy D, Mandel RJ, Naldini L,
3. Chang VT, Fernandes RA, Ganzinger KA, Lee Trono D (1997) Multiply attenuated lentiviral
SF, Siebold C, McColl J, Jönsson P, Palayret vector achieves efficient gene delivery in vivo.
M, Harlos K, Coles CH, Jones EY, Lui Y, Nat Biotechnol 15:871–875
Huang E, Gilbert RJ, Klenerman D, Aricescu 10. Kaufman RJ, Sharp PA (1982) Amplification
AR, Davis SJ (2016) Initiation of T cell signal- and expression of sequences cotransfected with
ing by CD45 segregation at ‘close contacts’. a modular dihydrofolate reductase complemen-
Nat Immunol 17:574–582 tary dna gene. J Mol Biol 159:601–621
4. Yokosuka T, Sakata-Sogawa K, Kobayashi W, 11. Dominguez AA, Lim WA, Qi LS (2016)
Hiroshima M, Hashimoto-Tane A, Tokunaga Beyond editing: repurposing CRISPR-Cas9 for
M, Dustin ML, Saito T (2005) Newly generated precision genome regulation and interrogation.
T cell receptor microclusters initiate and sustain Nat Rev Mol Cell Biol 17:5–15
T cell activation by recruitment of Zap70 and 12. N Peterson S, Kwon K (2012) The HaloTag:
SLP-76. Nat Immunol 6:1253–1262 improving soluble expression and applications
5. Varma R, Campi G, Yokosuka T, Saito T, in protein functional analysis. Curr Chem
Dustin ML (2006) T cell receptor-proximal Genomics 6:8–17.
signals are sustained in peripheral microclusters 13. Khan F, He M, Taussig MJ (2006) Double-
and terminated in the central supramolecular hexahistidine tag with high-affinity binding for
activation cluster. Immunity 25:117–127 protein immobilization, purification, and
6. Grakoui A, Bromley SK, Sumen C, Davis MM, detection on ni-nitrilotriacetic acid surfaces.
Shaw AS, Allen PM, Dustin ML (1999) The Anal Chem 78:3072–3079
immunological synapse: a molecular machine 14. Audic S, Lopez F, Claverie J, Poirot O, Abergel
controlling T cell activation. Science 285: C (1997) SAmBA: an interactive software for
221–227 optimizing the design of biological macromol-
7. van der Merwe PA, Davis SJ, Shaw AS, Dustin ecules crystallization experiments. Proteins 29:
ML (2000) Cytoskeletal polarization and redis- 252–257
Chapter 29
Imaging the Effector CD8 Synapse

Gordon L. Frazer, Yukako Asano, and Gillian M. Griffiths
Abstract
Here, we describe 4D imaging of effector CD8+ T cells as they conjugate and kill live targets in vitro and
analyze the polarization dynamics of intracellular compartments to this cell-cell interface.
Key words Nucleofection, Time-lapse confocal imaging, Imaris, 4D object-based image analysis
1 Introduction
The immune synapse of effector CD8+ T cells or cytotoxic T lym-

phocytes (CTLs) is crucial to the eradication of intracellular patho-
gens and tumor cells [1–3]. With the advent of genetic techniques
and high-speed confocal microscopy, it is now relatively straight-
forward to mark proteins with fluorescent tags and investigate
their 3D localization during synapse formation [4–6]. Unlike fixed
population-based approaches, individual CTL-target interactions
may be followed in real time, allowing the dissection of key stages
in the killing process from initial interaction all the way through to
detachment of the CTL and death of the target [7]. While this
provides an excellent overall picture of a protein or organelle’s spa-
tiotemporal dynamics, in as near to an in vivo way as possible, it
does have serious limitations. First, the expression of exogenous
DNA is often a problem with immune cells and correct localiza-
tion, as well as the effects of overexpression on cell behavior, must
always be verified. Second, the process is slow. To capture a CTL-
target interaction, one relies on luck and skill to find an appropri-
ately fluorescent CTL about to attack, that this CTL attacks during
the film while remaining in frame and that it does not interact with
other CTLs or targets during the attack. Data capture may then be
followed by processor intensive rendering and analysis to quantify
the depth of information each interaction contains, before finding
some way to represent this in a meaningful manner. It should also
be noted that conjugated cells rarely sit still and that the synapse
473
474 Gordon L. Frazer et al.
itself is a dynamic interface, as such it rarely lies directly in one

plane and often aligns to the less well-resolved z-plane.
We describe here our technique for time-lapse imaging of the
transgenic OT-I [8] CTL system attacking the thymoma target
EL-4 loaded with the OVA antigenic peptide. This technique may
be easily altered to allow for other systems such as anti-allogeneic-
stimulated CTL or alternative targets with better markers of cell
death. Such alterations are highlighted in the notes.
2 Materials
2.1 Equipment 1. Amaxa nucleofector. We use mouse T cell transfection kits with
the nucleofector from Amaxa and obtain transfection efficien-
cies between 40% and 60%, varying between DNA constructs.
Target cells are stably transformed with membrane marker pro-
teins using retroviral vectors.
2. Spinning disk confocal microscope. We use an inverted micro-
scope with an Olympus 60× silicone oil objective lens, incu-
bator chamber, spinning disk, and iXon Ultra 888 camera
(Fig. 1).
3. Image analysis computer. The example object-based image
analysis described below requires a license for Bitplane Imaris
and its associated cell module (see Note 1).
Fig. 1 Confocal spinning disk microscope. (1) Inverted microscope with an Olympus 60× silicone oil objective
lens. (2) Electronic stage with incubator chamber. (3) Yokogawa spinning disk. (4) iXon Ultra 888 camera
Imaging CD8 Synapses 475
2.2 Consumables 1. 35 mm No. 1.5 glass-bottom culture dishes with 14 mm inset
(MatTek).
2. Amaxa nucleofection kit for mouse T cells (human T cell kits
are also available).
2.3 Media All media should be stored at 4 °C, preferably in the dark.
Components marked with * should be kept at 4 °C, while **
should be stored in aliquots at −20 °C and thawed before mixing.
1. CTL medium (CTLM): Roswell Park Memorial Institute
(RPMI) 1640 medium* supplemented with 10% fetal bovine
serum (FBS)**, 1 mM l-glutamine**, 1 mM sodium pyru-
vate*, 100 U/ml penicillin with 0.1 mg/ml streptomycin*, 50
μM 2-mercaptoethanol**, and recombinant murine interleu-
kin 2** (PeproTech). Please note, this medium has a life span
of 2 weeks from preparation.
2. Target medium (TM): Dulbecco’s modified eagle medium*
(DMEM) supplemented with end concentrations of 10%
FBS**, 1 mM l-glutamine**, and 100 U/ml penicillin with
0.1 mg/ml streptomycin*.
3. Serum-free medium (SFM): DMEM*.
4. Imaging buffer (IB): Phenol red-free RPMI 1640* with
10%FBS**, 1 mM l-glutamine**, 1 mM sodium pyruvate
(GIBCO)*, and 25 mM HEPES.
2.4 Cells (See 1. Fluorescent target cells at ~exponential growth stage (see
Note 2) Note 3)—2 × 105 cells per dish.
2. Activated CD8+ CTL from day 5 to 8 postactivation—5 × 106
per three dishes (roughly 2 h of imaging).
2.5 Target OVA257–264 SIINFEKL peptide or anti-CD3 antibody (we use ham-
Presentation ster anti-mouse clone 145-2c11 or mouse antihuman clone
Components UCHT1 (RUO) both from BD Pharmingen).
2.6 Components 1. Intercellular adhesion molecule (ICAM)-1 (R&D Systems).

for Coating Dishes 2. Dulbecco’s phosphate buffered saline (PBS).
3 Methods
3.1 Preparation One 1. Aliquot IB (~12 ml/dish) and SFM (~15 ml/batch of 2–3
Day before Imaging dishes) into T-25 flasks and leave to equilibrate in an incubator
at 37 °C 8%CO2 overnight.
2. Coat 35 mm glass-bottom culture dishes with 1 μg/ml murine
ICAM-1 by applying 250 μl/dish of 1 μg/ml ICAM-1 in PBS
and leaving overnight at 4 °C (see Note 4).
3. Ensure there will be enough healthy target cells for the follow-
ing day.
4. Nucleofect CTL 24 h in advance of microscopy following the
instructions for the Amaxa nucleofection kit. The details for
the murine CTL kit are outlined in brief below (see Note 5).
(a) Add 10 μl nucleofection medium component B to 1 ml
nucleofection medium in a 12-well plate and warm at 37 °C,
8% CO2 for 30 min.
(b) Take 5 × 106 CTL and wash 2× with PBS.
(c) Add 2.5–10 μg total DNA constructs in <10 μl volume to
a sterile 1.5 ml Eppendorf tube (see Note 6).
(d) Resuspend CTL in 100 μl nucleofection solution, apply
to the DNA, mix, and transfer to a nucleofection cuvette
(see Note 7).
(e) Nucleofect with program X-001 mouse CD8+ T cell (see
Note 8).
(f) Immediately transfer to the pre-warmed medium (5a) in
the 12-well plate.
(g) Return to the incubator for 2–4 h.
(h) Top up to 3 ml with pre-warmed CTLM, spread evenly
among six wells of a 12-well plate, and then top these up
to 3.5 ml with pre-warmed CTLM, and return to the
incubator until used for imaging.
3.2 Day of Imaging 5. Start up the microscope including appropriate heating and
CO2 chambers, and ensure it is ready and functional for the day.
6. Pre-warm the ICAM-1-coated glass-bottom dishes in the incu-
bator at 37 °C 8% CO2.
7. Take 5 ml of targets, and centrifuge for 5 min at 1200 rpm in
Beckman Coulter SX4400 rotor (277 × g).
8. Resuspend in 1 ml TM with 1 μM SIINFEKL peptide (see
Note 9).
9. Incubate at 37 °C for 1 h, with gentle resuspension every
15 min.
10. Add 9 ml pre-warmed TM or SFM (see Note 10) and take
sample for counting.
11. Centrifuge 5 min 1200 rpm (SX4400 rotor) (277 × g).
12. While centrifuging:
(a) Count and calculate the number of target cells in the tube
and therefore the volume for resuspension at 0.7 × 106
cells/ml (see Note 11).
(b) Gently wash unbound ICAM-1 off of the preheated imag-
ing dishes with PBS three times.
13. Resuspend targets at 0.7 × 106/ml in SFM or TM (see Note 10)

and apply 250 μl to each dish.
14. Transfer to incubator for 5 min (see Note 12).
15. Gently rinse off unbound targets by application of 1.5 ml IB to
the dish (see Note 13).
16. Return dishes to the incubator until ready for use on the
microscope.
17. Take ~106 nucleofected CTL (1–2 wells from 12-well plate)
into a total volume of 10 ml IB and centrifuge at 1000 rpm
(SX4400 rotor) (193 × g) for 5 min.
18. During centrifugation of CTL, ensure dish is correctly set up
on the microscope and the settings are ready for imaging (see
Note 14).
19. Pour off the supernatant from the CTL keeping ~150 μl on
the pellet.
20. Pipette the cells in this volume up and down, while keeping
warmly cupped in the hand, and apply to the dish on the
microscope in a gentle dropwise manner (see Note 15).
21. Search for appropriate CTL-target interactions and begin
imaging.
22. Repeat steps 17–21 until dishes with targets are used, and
then repeat from step 7 to 21 until all ICAM-1-coated dishes
are used.
23. Export the data and visualize/analyze with software such as
Imaris (see Notes 16 and 17) (Fig. 2).
3.3 Analysis 24. Convert microscopy data to the Imaris file-type and save to a
Using Imaris local drive (see Note 16).
25. Crop to the cell of interest (see Note 16).
26. Generate “surfaces” of the CTL and target cell (Fig. 3).
27. Use the target cell “surface” to generate a mask of the target
cell channel (Fig. 4).
28. Use the CTL surface to mask this newly generated surface
(Fig. 5).
29. Use the “cell” module to search this new channel for “nuclei”
(Fig. 6).

30.
Detect organelles/intracellular compartments as “spots”
(Fig. 7).
31. Complete the cell module with or without tracking (Fig. 8).
32. Export statistics. Use “closest nucleus distance” for IS polar-
ization dynamics.
33. Repeat for next cell.
Fig. 2 Example time-lapse data. Maximum intensity projection of EL4 expressing mTagBFP2 (blue), presenting
OVA to an OT-I CTL nucleofected with Lifeact-EGFP (binds f-actin [10]) (green), microtubule end-binding pro-
tein 3 (EB3)-iRFP670 [11] (white), and expressing Gzm-B-TdTomato (granzyme B, protease found in cytolytic
granules [12]) (red) from an endogenous promoter
4 Notes
1. Programs such as Imaris require a lot of computer processing

power for 4D data sets. See http://www.bitplane.com/sys-
temrequirements.aspx. Generally, the local drive speed con-
trols how long it takes to load a file and takes over from the
random-access memory (RAM) when it is filled; the RAM tries
to hold the data for calculation and the graphics card controls
data representation, while the processor does the modeling
calculations.
2. Any CTL model may be used here providing both CTL and
target are fluorescent by the day of imaging.
3. Many target cell lines are susceptible to nucleofection
(Subheading 3.1, step 4) and may be repeatedly sorted for
stable expression. Alternatively, dyes such as PKH26 and CFSE
may be successfully employed to color target cells before imag-
ing [9]. Common cell lines used in our lab are EL4 which
carry the H2-Kb MHC for the OT-I system and P815 for both
anti-H2-Dd allogeneic CTL and redirected lysis where the Fc
receptor (FcR) expression on these cells allows capture of anti-
CD3 antibody for CTL activation.
4. The concentration of ICAM-1 we use is optimized to allow the
CTL to move across the dish without being activated or adhere
too tightly.
5. Other CTL nucleofection systems predominately vary in the
naming of the media and solutions but follow the same basic
protocol. Alternatively, CTL lines stably expressing fluorescent
markers of interest may be used.
6. It is important to consider fluorescent construct combinations
carefully. Modeling works best when each structure of interest
can be captured in a separate channel. The analysis below uses
points of high signal intensity as seed points from which to
grow the cell model, so any overlap or bleed-through may
result in the target or CTL models merging with one another
preventing discrimination.
Fig. 3 Modeling the CTL and target as surfaces. (a) Preview of target surface at the threshold value selection
stage. This should be chosen so as to generate a solid surface at the synapse without excessive dilation into
the CTL, as this will define the surface to which distances will be calculated. (b) Finished target surface. (c)
CTL surface preview at the threshold value selection stage. As with the target threshold, aim for a solid syn-
apse without expanding too far into the target. (d) CTL surface generated. (e) Both target and CTL surfaces
generated
Fig. 3 (continued)
Fig. 3 (continued)
7. Remove as much supernatant as possible before resuspension.

Remaining PBS can lead to sparking during the nucleofection
process. We find the nucleofection solution can be harmful to
cells with time. It is therefore advisable to keep CTL in this for
as short a time as possible, and we often split the process into
batches of four before resuspending the next set of CTL in the
nucleofection solution.
8. Or whichever is appropriate for the system used.
9. For redirected lysis systems where it is necessary to stimulate
conjugate formation with anti-CD3 antibody presentation,
replace peptide loading (steps 8 and 9) instead with antibody
at 0.5 μg/ml at step 8 and step 9 may be shortened to around
15 min. Please note this requires FcR-expressing target cells
such as P815.
10. Some target cells adhere to the coverslip readily in IB, while
others (such as EL4) adhere better when plated in SFM.
11. If larger targets are being used, it may be necessary to reduce
the concentration to achieve isolated target cells.
12. The length of time required for cells to adhere to the dish will
vary and optimal timing needs to be established for each cell
type.
Fig. 4 Masking the target cell. (a) Start of masking process to leave just the target volume with values >0. (b)
Masked target cell channel generated
Fig. 5 Generating the “synapse” channel. (a) Start of the masking process to set all pixels not at the CTL-target
boundary to 0. (b) The generated synapse channel in blue
Fig. 6 Defining the synapse channel as the “nucleus” of the cell module
Fig. 7 Defining the intracellular structures of interest. The algorithm is based on the “spots” algorithm and
locates approximate spheres of intensity within a chosen channel. The EB3-iRFP670 channel has been chosen
here to demonstrate that weak signals may be modeled when the background is low
Fig. 8 The completed cell model
~30°
Fig. 9 Gently washing excess cells off of the glass
13. This is easiest done by gently tilting the dish and applying the
IB close to the edge of the dish above the center (Fig. 9).
14. Beware the time it takes for some objective lens oils to adjust
to 37 °C. If the oil temperature is not equilibrated to the
microscope, the change in temperature may alter the focal
height across the imaging session.
15. CTLs have a tendency to clump, and it is essential to separate
them to ensure what is imaged is not a co-attack of a nonfluo-
rescent “ghost” cell.
16. The data generated from this time-lapse 3D imaging can be
considerably larger than most fixed or 2D experiments, with
individual data files reaching 20–40 Gb. This poses considerable

strain on data transfer and storage systems as well as later analysis
where linear file size increases can lead to exponential growths in
processing time. It is therefore highly recommended to opti-
mize all computer systems where possible and preprocess data
appropriately before more intense analysis.
17. There are many free packages available that may be forced to
do similar analysis to Imaris, but this is a very intuitive program
for navigating 4D data, and its modeling functions allow user
input to manually distinguish objects where algorithms alone
cannot.
References
1. Silverstein AM (2001) The lymphocyte in 8. Hogquist KA, Jameson SC, Heath WR,
immunology: from James B. Murphy to James Howard JL, Bevan MJ, Carbone FR (1994) T
L. Gowans. Nat Immunol 2:569–571 cell receptor antagonist peptides induce posi-
2. Masopust D, Vezys V, Wherry EJ, Ahmed R tive selection. Cell 76:17–27
(2007) A brief history of CD8 T cells. Eur 9. Progatzky F, Dallman MJ, Lo Celso C (2013)
J Immunol 37(Suppl 1):S103–S110 From seeing to believing: labelling strategies
3. de la Roche M, Asano Y, Griffiths GM (2016) for in vivo cell-tracking experiments. Interface
Origins of the cytolytic synapse. Nat Rev Focus 3:20130001
Immunol 16(7):421–432 10. Riedl J, Crevenna AH, Kessenbrock K, Yu JH,
4. Schermelleh L, Heintzmann R, Leonhardt H Neukirchen D, Bista M, Bradke F, Jenne D,
(2010) A guide to super-resolution fluores- Holak TA, Werb Z, Sixt M, Wedlich-Soldner R
cence microscopy. J Cell Biol 190:165–175 (2008) Lifeact: a versatile marker to visualize
5. Zhang J, Campbell RE, Ting AY, Tsien RY F-actin. Nat Methods 5:605–607
(2002) Creating new fluorescent probes for cell 11. Nakagawa H, Koyama K, Murata Y, Morito M,
biology. Nat Rev Mol Cell Biol 3:906–918 Akiyama T, Nakamura Y (2000) EB3, a novel
6. Miyawaki A, Niino Y (2015) Molecular spies for member of the EB1 family preferentially
bioimaging—fluorescent protein-based probes. expressed in the central nervous system, binds
Mol Cell 58:632–643 to a CNS-specific APC homologue. Oncogene
19:210–216
7. Ritter AT, Asano Y, Stinchcombe JC,
Dieckmann NM, Chen BC, Gawden-Bone C, 12. Mouchacca P, Schmitt-Verhulst AM, Boyer C
van Engelenburg S, Legant W, Gao L, Davidson (2013) Visualization of cytolytic T cell differ-
MW, Betzig E, Lippincott-Schwartz J, Griffiths entiation and granule exocytosis with T cells
GM (2015) Actin depletion initiates events from mice expressing active fluorescent gran-
leading to granule secretion at the immuno- zyme B. PLoS One 8:e67239
logical synapse. Immunity 42:864–876
Chapter 30
The Mast Cell Antibody-Dependent Degranulatory Synapse

Salvatore Valitutti, Régis Joulia, and Eric Espinosa
Abstract
Mast cells are key effector cells in inflammation that can be activated by specific antigens via IgE or IgG
binding on their FcR. Aggregation of mast cell Fc receptors by cell-bound antigens induces mast cell polar-
ized degranulation toward the stimulatory cell, a process named antibody-dependent degranulatory syn-
apse (ADDS). This polarized degranulation allows mast cells to expose bioactive material embedded in the
granule matrix toward the antibody-targeted cell and is accompanied by the formation of a signaling area
at the cell–cell contact site. In this chapter, we describe (1) how to stimulate mast cells with cell-bound
antigens and (2) how to monitor ADDS formation and to investigate ADDS characteristics by confocal
microscopy.
Key words Mast cell, Degranulatory synapse, Polarized degranulation, Avidin, Confocal microscopy
1 Introduction
Mast cells are unique hematopoietic cells that reside in virtually all
tissues and notably near blood vessels and nerve endings [1]. Mast
cell cytoplasm is filled with secretory granules where a vast array of
mediators are stocked (e.g., histamine, tryptase, chymase, tumor
necrosis factor). Those mediators are embedded in a matrix rich in
heparin and can be swiftly released by the exocytosis of the granule
content, a process called degranulation [2]. Mast cell degranula-
tion is classically triggered in vitro by using soluble stimuli and is
measured using bulk assays that quantify mediators (such as hista-
mine or β-hexosaminidase) released in the supernatant.
Nevertheless, mast cells are expected to be mainly stimulated
in vivo by particulate antigens or IgG-opsonized cells. In a recent
work, we showed that when human mast cells are stimulated by
cell-bound antigens, they exhibit polarized degranulation toward
the stimulatory cell. We named this phenomenon the antibody-
dependent degranulatory synapse (ADDS) [3]. The method
described herein allows to m onitor ADDS by time-lapse micros-
copy or by confocal microscopy on fixed preparations.
487
488 Salvatore Valitutti et al.
The method relies on the properties of avidin, a tetrameric

protein that selectively binds to heparin contained in mast cell
granules [4]. In our recent work, we showed that this probe can
be employed to monitor degranulation of live mast cells by time-
lapse microscopy, when added to the culture medium [3]. During
the degranulation process, granule matrix is exteriorized and
immediately bound by the fluorochrome-labeled avidin present
in the medium. This approach allowed us to show that a fraction
of the exteriorized granules remains attached for hours on the
mast cell surface during the degranulation process and is polar-
ized toward the stimulatory cell [3]. These observations prompted
us to propose the use of fluorochrome-labeled avidin to stain
degranulated mast cells either in live cells or following cell fixation.
In particular, the experimental procedures here described allow
to define the spatiotemporal characteristics of polarized mast cell
degranulation.
Human mast cells express two stimulatory Fc receptors: the
high-affinity IgE receptor FcεRI that binds monomeric IgE and
the low-affinity IgG receptor FcRγIIA that binds IgG immune
complexes [5]. Aggregation of those receptors by either antigens
or cross-linking antibodies induces mast cell degranulation. ADDS
can be achieved either in IgE-sensitized mast cells stimulated with
specific cell-bound antigens or in mast cells interacting with IgG-
opsonized cells [3]. We showed that the ADDS is accompanied by
a partial cortical actin clearance at the degranulation site in the
absence of microtubule-organizing center (MTOC) polarization.
The present method chapter is formed of two main parts. In a first
part, we describe methods for mast cell stimulation with cell-bound
antigens. In a second part, we describe how to detect granule
exposure in parallel with actin and tubulin cytoskeleton staining in
mast cells undergoing ADDS.
2 Materials
2.1 Cells To take advantage of avidin properties to monitor mast cell degran-
ulation, connective tissue-type mast cells must be used as their
granule content is rich in heparin. The method described here is
suited for primary human mast cell cultures derived from periph-
eral blood CD34+ progenitor cells (referred to as hMC) [6–8]
(see Note 1). This method can be adapted to peritoneal cell-derived
mast cells (PCMC) [9, 10].
1. Mast cells: Generated from CD34+ cells isolated from PBMCs.
2. B cells: Epstein–Barr virus (EBV)-transformed lymphoblastoid
cell line JY cells. JY cells [11] are routinely passaged in RPMI-
1640 10% FCS (see Note 2).
Monitoring Mast Cell ADDS 489
2.2 Mast Cell 1. Microscope slides: Use Teflon-printed diagnostic glass slides
Polarized Stimulation (ten wells, Fig. 1). Hydrophobic printed slides offer shallow
wells allowing to perform cell stimulation followed by immu-
nofluorescence staining. Slides are washed carefully with 70%
ethanol before use.
2. Phosphate-buffered salt without calcium and magnesium
(PBS).
3. Poly-D-lysine hydrobromide diluted 1:80 v/v with distilled
water.
4. Avidin sulforhodamine 101.
5. Tyrode’s buffer: 137 mM NaCl, 2.7 mM KCl, 0.4 mM
NaH2PO4, 5.5 mM glucose, 1.6 mM CaCl2, 1 mM MgCl2; pH
7.2, 0.1% BSA.
6. Humanized anti-CD20 IgE (InvivoGen, San Diego, CA).
7. CellTracker™ Blue CMAC (Molecular Probes).
2.3 Immuno 1. Paraformaldehyde 4% (dilute in PBS PFA 16%, Electron

fluorescence Microscopy Sciences, Hatfield, PA).
and Confocal 2. Permeabilization buffer: PBS 0.1% Saponin, 1% BSA (Sigma-
Microscopy Aldrich, St. Louis, MO) (see Note 3).
3. Anti-α-tubulin (clone DM1A, Sigma-Aldrich, St. Louis, MO).
4. Alexa Fluor 488-labeled phalloidin (Molecular Probes, Inc.,
Eugene, OR).
5. Goat anti-mouse IgG1-Alexa Fluor 647 (Molecular Probes,
Inc., Eugene, OR).
6. Mounting medium: DABCO solution consisting of the fol-
lowing: 90% glycerol, 10% PBS, and 2.5% (wt/vol) 1,4-
diazabicyclo(2,2,2)octane, pH 8.6 (Sigma-Aldrich, St. Louis,
MO).
Fig. 1 Teflon-printed diagnostic slide with well specifications

3 Methods
Polarized degranulation is monitored by sulforhodamine

101-labeled avidin (Av-SRho) present in the medium before cell
stimulation and fixation (see Note 4). The method also describes
actin and tubulin staining upon ADDS formation.
3.1 Primary Mast 1. Grow CD34+ cells under serum-free conditions using
Cell Culture StemSpan™ medium (STEMCELL Technologies) supple-
mented with recombinant human IL-6 (50 ng/mL), human
IL-3 (10 ng/mL), and human SCF (50 ng/mL) for 1 week.
2. Grow cells in IMDM GlutaMAX I, sodium pyruvate,
2-
mercaptoethanol, 0.5% BSA, insulin–transferrin–selenium,
ciprofloxacin (10 μg/mL), IL-6 (50 ng/mL), and SCF (50
ng/mL) for 8 weeks.
3. Verify cell purity phenotypically (Tryptase+, CD117+, FcεRI+)
and functionally (β-hexosaminidase release in response to
FcεRI crosslinking) before use for experiments. Only primary
cell lines showing more than 95% CD117+/FcεRI+ cells should
be used (see Note 1).
3.2 Mast Cell 1. Sensitize or not 1 × 106 hMC in 1 mL culture medium in a 5

Stimulation with Cell- mL FACS tube with 1 μg/mL anti-CD20 human IgE for 1 h
Bound Antigens at 37 °C (see Note 5).
2. Prepare 1 mL RPMI-1640 (without serum) with 1 μM
CellTracker™ Blue CMAC (Working Solution).
3. Spin down 2 × 105 JY cell (250 × g, 5 min) in a 15 mL centri-
fuge tube, discard the supernatant, and resuspend the cell
pellet in the working solution. Incubate for 30 min at 37 °C
5% CO2.
4. Centrifuge the cells and remove the CellTracker™ Working
Solution. Wash with 10 mL complete RPMI-1640, then wash
with 10 mL Tyrode’s buffer, and adjust the cell suspension to
1 × 106/mL in Tyrode’s buffer.
5. Prepare a suspension of 1 × 105 CellTracker™ Blue CMAC-
labeled JY cells in 100 μL Tyrode’s buffer and pre-warm the
cells 15 min at 37 °C.
6. Coat eight wells of a microscope slide with 50 μL of poly-d-
lysine for 15 min at 37 °C.
7. Rinse the wells twice with PBS (Siphon the PBS off with a
vacuum pump and add 50 μL PBS).
8. Pre-warm the slide at 37 °C with the 50 μL PBS left.
9. Spin down sensitized and non-sensitized hMCs in a 15 mL
tube for 5 min at 200 × g and resuspend the cell pellet in 100
μL 37 °C pre-warmed Tyrode’s buffer + Av-SRho 8 μg/mL
(see Note 6).
10. Siphon the PBS off from the eight wells, and add gently 25 μL
hMCs per well (sensitized hMCs in the upper row and non-
sensitized hMCs in the lower row, see Fig. 1) plus 25 μL of
pre-warmed JY cells, and then incubate for 30 min at 37 °C in
the incubator (5% CO2) (see Note 7).
3.3 Staining For each well of the slide:

Procedure
1. Siphon carefully the incubation buffer off with a vacuum pump
from each well, and add gently 50 μL of fixative solution and
incubate for 15 min at 37 °C (see Note 8).
2. Siphon carefully the fixative solution off and wash the cells
twice with 50 μL PBS.
3. Add 50 μL of permeabilizing buffer and incubate for 15 min at
37 °C (see Note 9).
4. Prepare the primary Ab solution and isotype-matched control
in permeabilizing buffer: 5 μg/mL anti-tubulin mAb plus
Alexa Fluor 488-labeled phalloidin (1 U/mL) for the six test
wells and 5 μg/mL mouse IgG1 isotype control for the two
isotype control wells.
5. Siphon carefully the permeabilizing buffer off, and add 50 μL
of the primary Ab solution into the six corresponding wells
and 50 μL isotype control solution, and incubate 60 min at 37
°C in an H2O-saturated chamber.
6. Prepare secondary Ab solution: mix secondary antibody
(Alexa647-goat anti-mouse IgG1, 2 μg/mL) in 400 μL per-
meabilizing buffer.
7. Wash carefully the wells with 50 μL permeabilizing buffer.
Repeat three times allowing a 2–3 min between washings to
enable any nonspecifically adherent immunoglobulins to dif-
fuse away.
8. Add 50 μL secondary Ab solution to each well, and incubate
60 min at 37 °C in an H2O-saturated chamber.
9. Wash twice for 2–3 min with 50 μL permeabilizing buffer and
then wash with 50 μL PBS.
10. Mount coverslip using mounting medium.
11. Examine the samples using a confocal microscope with a 63×
objective (1.4 oil), electronic zoom 3 (Fig. 2). In the case of 3-D
images, acquire z-stacks with an interval of 0.2–0.4 μm (Fig. 3).
12. Use MetaMorph software to quantify the fluorescence inten-
sity at the cell–cell contact site versus the other part of the cell
(Fig. 2).
13. Scoring of the slides should be performed in a blinded fash-
ion by evaluating for each condition at least 50 hMC/B cell
conjugates in randomly selected fields from at least three
Fig. 2 Visualization of ADDS in mast cells and quantification of F-actin clearance and MTOC polarization. (a)
Anti-CD20 IgE-sensitized hMCs were incubated with CD20+ B cells (cyan) in Tyrode’s buffer plus Av.SRho for
30 min at 37 °C. Cells were fixed, permeabilized, and stained with phalloidin (blue) and α-tubulin (green). Cells
were analyzed using a confocal microscope. Representative conjugates are shown. Results are from one
representative experiment out of three. (b) Quantification of phalloidin integrated fluorescence intensity (IFI) at
the distal and synaptic areas (see scheme) (n = 59 conjugates). Paired t test ***P < 0.001. (c) Measurement
of the distance of the MTOC from the synapse. Plots show the distance of the MTOC from the synapse divided
by the hMC diameter (d/D) for each conjugate (see scheme). Experiments were performed using either non-
sensitized hMC (− anti-CD20 IgE, n = 62 or sensitized hMC (+ anti-CD20 IgE, n = 64). Unpaired t test ns, P >
0.05. Bars, 5 μm. Data in b and c are from [3]
independent experiments. To evaluate the polarization of

granules toward the degranulatory synapse, unprocessed
images are analyzed using the region measurements func-
tion of the MetaMorph software (Universal Imaging,
Downingtown, PA). To evaluate precisely the localization
of actin at the degranulatory synapse, two regions are ana-
lyzed: one containing the degranulatory synapse (1/20 of
the hMC, synaptic area) and the other containing the
opposed region (1/20 of the hMC, distal area; see scheme
Fig. 2b). The integrated fluorescence intensities (IFI) calcu-
lated by the software in both regions are compared.
Distances between MTOC and the center of the synaptic
area are measured using either the MetaMorph software or
the microscope software (see scheme Fig. 2c).
Fig. 3 3-D reconstruction of a mast cell/B cell conjugate. Mast cells were conjugated with B cells (cyan) in the
presence of avin-SRho (red). After fixation and permeabilization, cells were stained with an anti-α-tubulin mAb
(green). Z-stacks were acquired with an interval of 0.4 μm
4 Notes
1. Human mast cells can be also derived from CD34+ progenitors

cells isolated from bone marrow or cord blood. Alternatively
human mast cells can be isolated from human skin [12].
2. JY cells are currently used in our laboratory but are not neces-
sary. Every B cell line expressing CD20 on its surface is suited
to serve as stimulatory cells.
3. The permeabilization step allows to permeabilize cells and to
block nonspecific binding by using 1% BSA (alternatively, PBS
5% goat serum or 5% human serum can be used). Saponin is
used here to permeabilize the cells. Saponin-based permeabi-
lization buffer must be prepared extemporaneously and added
to all the staining steps as saponin-mediated permeabilization
is reversible. Alternatively, Triton X-100 0.1% can be used to
permeabilize the cells; in this case, Triton X-100 is not
included during the staining procedures as permeabilization is
irreversible.
4. Alternatively, polarized degranulation can be monitored by
time-lapse confocal microscopy in live cells. Anti-CD20-IgE-
sensitized hMCs are settled on poly-d-lysine-coated Lab-Tek™
Chambered Coverglass (Nunc) in Tyrode’s buffer supple-
mented with 8 μg/mL avidin sulforhodamine 101 and warmed
at 37 °C in the chamber of the microscope. Microscope settings
are performed on resting mast cells during the pre-warming
period, and acquisition is started. B cells are very carefully
added into the monitored well and acquisition in resumed.

Fluorescence is acquired every 2–5 s using a confocal micro-
scope equipped with an environmental chamber (37 °C and
5% CO2) using a 63x Plan-Apochromat objective (1.4 oil).
Conjugate formation takes place within ~10 min after B cell
addition.
5. Alternatively, sensitization can be done by incubating cells for
16 h with 1 μg/mL anti-CD20 human IgE.
6. Fluorescent avidin is added in the incubation medium to stain
exposed granules during the degranulation process. It can be
added either in mast cells suspension or in the B cells
suspension.
7. It is mandatory to perform mast cell activation at 37 °C since
a drop in temperature before or during stimulation precludes
degranulation. To this end, pre-warming procedures are very
important.
8. The steps of siphoning solutions off and adding media must be
carried out very gently and cautiously to avoid cells detaching
from the slide.
9. It should be kept in mind that, when staining the tubulin cyto-
skeleton, the fixation procedure should be performed at 37 °C
in order to avoid tubulin depolymerization.
10. It should be kept in mind that avidin is used in the in the pres-
ent protocol for its heparin-binding capacity; however, avidin
also binds biotin with high affinity. As a consequence, biotinyl-
ated antibodies cannot be used in the staining procedure due
to the possible unspecific binding to avidin.
References
1. Abraham SN, St John AL (2010) Mast cell- 5. Jonsson F, Daeron M (2012) Mast cells and
orchestrated immunity to pathogens. Nat Rev company. Front Immunol 3:16. doi:10.3389/
Immunol 10(6):440–452 fimmu.2012.00016
2. Dvorak AM, Massey W, Warner J, Kissell S, 6. Andersen HB, Holm M, Hetland TE, Dahl C,
Kagey-Sobotka A, Lichtenstein LM (1991) Junker S, Schiotz PO, Hoffmann HJ (2008)
IgE-mediated anaphylactic degranulation of Comparison of short term in vitro cultured
isolated human skin mast cells. Blood 77(3): human mast cells from different progenitors -
569–578 Peripheral blood-derived progenitors generate
3. Joulia R, Gaudenzio N, Rodrigues M, Lopez J, highly mature and functional mast cells.
Blanchard N, Valitutti S, Espinosa E (2015) J Immunol Methods 336(2):166–174
Mast cells form antibody-dependent degranu- 7. Holm M, Andersen HB, Hetland TE, Dahl C,
latory synapse for dedicated secretion and Hoffmann HJ, Junker S, Schiotz PO (2008)
defence. Nat Commun 6:6174. doi:10.1038/ Seven week culture of functional human mast
ncomms7174 cells from buffy coat preparations. J Immunol
4. Tharp MD, Seelig LL Jr, Tigelaar RE, Methods 336(2):213–221
Bergstresser PR (1985) Conjugated avidin 8. Bandara G, Metcalfe DD, Kirshenbaum AS
binds to mast cell granules. J Histochem (2015) Growth of human mast cells from bone
Cytochem 33(1):27–32 marrow and peripheral blood-derived CD34(+)
pluripotent hematopoietic cells. Methods Mol eration at the T helper cell/mast cell immuno-
Biol 1220:155–162. doi:10.1007/978-1- logical synapse. Blood 114(24):4979–4988
4939-1568-2_10 11. Valitutti S, Muller S, Salio M, Lanzavecchia A
9. Malbec O, Roget K, Schiffer C, Iannascoli B, (1997) Degradation of T cell receptor (TCR)-
Dumas AR, Arock M, Daeron M (2007) CD3-zeta complexes after antigenic stimula-
Peritoneal cell-derived mast cells: an in vitro tion. J Exp Med 185(10):1859–1864
model of mature serosal-type mouse mast cells. 12. Kulka M, Metcalfe DD (2010) Isolation of tis-
J Immunol 178(10):6465–6475 sue mast cells. Curr Protoc Immunol. John
10. Gaudenzio N, Espagnolle N, Mars LT, Liblau E. Coligan et al. (eds). Chapter 7:Unit 7.25.
R, Valitutti S, Espinosa E (2009) Cell-cell coop- doi:10.1002/0471142735.im0725s90
Chapter 31
Measurement of Lytic Granule Convergence After

Formation of an NK Cell Immunological Synapse
Hsiang-Ting Hsu, Alexandre F. Carisey, and Jordan S. Orange
Abstract
Natural killer (NK) cells contain specialized lysosome-related organelles termed lytic granules allowing
them to mediate cytotoxicity against tumorigenic or virally infected target cells. NK cells polarize their lytic
granules toward a target cell via the microtubule-organizing center (MTOC). Prior to that, however, lytic
granules converge to the MTOC along microtubules utilizing minus-end-directed microtubule motors.
Herein we describe how to visualize and quantify lytic granule convergence using confocal microscopy to
gain quantitative insight into NK cell cytotoxicity and its regulation.
Key words Immunological synapse, MTOC, Lytic granules, Convergence, Confocal microscopy,
Mathematical algorithm
1 Introduction
The interface between an NK cell and its target, where the NK cell
obtains activating and inhibitory inputs through its germline-
encoded receptors, is a form of an immunological synapse (IS). IS
formation is an essential prerequisite for target cell contact-
dependent NK cell function, most notably cytotoxicity [1, 2]. The
formation of a mature lytic NK cell IS (NKIS) is a complex process
which requires numerous steps to promote cellular function
(reviewed in 1, 2). One of the characteristic features of lytic NKIS
formation is polarization of the microtubule-organizing center
(MTOC) to the IS along with the lytic granules followed by exo-
cytosis of the lytic granule contents onto the target cell to trigger
target cell death. By quantitatively studying the dynamics of the
NKIS, the advancement of lytic granules toward their directed
secretion can be utilized as a precision indicator of commitment to
and effectiveness of NK cell cytotoxicity. However, this demands
an in-depth understanding of the process, access to high-resolution
microscopy, and application of unbiased quantitative analyses of
images paired with appropriate statistical evaluations. Performed
497
498 Hsiang-Ting Hsu et al.
rigorously, the NKIS can be measured and linked to effectiveness

in cytotoxicity and other related NK cell functions.
In developing methods for studying the NKIS, we have focused
on the use of confocal microscopy, which provides the advantage of
measuring events within a given optical section in a single cell with-
out interference of signals from elsewhere in the cell. This increases
the specificity of the data, particularly when the NK cell is still con-
jugated to its target. Using this approach, optical sections taken
along a vertical axis of the conjugate make it possible to capture the
entire biology in three dimensions (3D). Additionally, confocal
microscopy allows for the imaging of living cells to create four-
dimensional sequences (3D imaging over time). Live cell imaging
allows individual NK cells in conjugates to be evaluated throughout
the entire range of IS formation and function including lytic granule
convergence, MTOC polarization, granule exocytosis, killing, and
detachment. Quantitative algorithms have been developed for pre-
cise measurement of each of these steps. Because of the large amount
of information that can be collected from a single live cell imaging
sequence, confocal microscopy remains one of the most effective
and affordable means of obtaining highly quantitative information
regarding the kinetics and dynamics of NKIS formation.
Herein we focus specifically on methods for measuring one
important feature of the NKIS: the positioning of lytic granules rela-
tive to the MTOC. Specifically their characteristic behavior of con-
verging lytic granules to the MTOC happens rapidly after NKIS
formation [3, 4]. We refer to this process as granule convergence
[5]. The same process of convergence has recently also been identi-
fied in cytotoxic T cells [6] and is thus likely a general feature of
certain types of activation in cytotoxic lymphocytes. In NK cells
engaged in the destruction of a susceptible target cell, the granules
move along individual microtubules toward the MTOC (in a minus-
end-directed manner) using dynein motors [3]. They aggregate
around the MTOC and then polarize along with the MTOC toward
the IS. Once polarized, they are capable of degranulation upon the
presence of an appropriate signal [5, 7, 8]. Aside from the activation
of motor function, numerous adaptors and signals have been defined
that are needed for granule convergence [4, 5, 7, 9–11]. The ability
to precisely quantify this process is essential in measuring the effec-
tiveness of signaling for and preparedness of lytic granules for
directed secretion [5]. A generalizable protocol for precisely mea-
suring lytic granule positioning and convergence is provided.
2 Materials
1. Culture medium: RPMI 1640, supplemented with 10% fetal

bovine serum (FBS), 10 mM HEPES, 1× penicillin/strepto-
mycin, 100 μM MEM nonessential amino acids, 1 mM MEM
sodium pyruvate, and 2 mM l-glutamine (R10). Phenol red-
Analytical Measurement of Lytic Granule Convergence in NK Cells 499
free RPMI is used for live cell imaging. Myelocult H5100

(StemCell Technologies), supplemented with 1× penicillin/
streptomycin and 100 U/ml of IL-2. All cells were cultured in
R-10 medium except for the NK-92 cells which were cultured
with the Myelocult medium.
2. Human immortalized NK cell line NK-92 (ATCC Cat # CRL-
2407) as an example cytotoxic cell utilized here.
3. K562 (ATCC Cat # CCL-243), an erythroleukemic cell line,
was used as target cells for NK-92 cells.
4. Poly-l-lysine- or silane-coated glass slides (Electron Microscopy
Sciences Cat # 63410-01 or 63411-01, respectively).
5. Coverslips 22 mm square #1.5.
6. Eight-well culture slides (Fisher Scientific Cat # 354108).
7. Hydrophobic ink pen.
8. Fixative and permeabilization agent (Cytofix/Cytoperm, BD)
containing 0.1% (v/v) Triton X-100.
10. PBS containing 1% bovine serum albumin (BSA).
11. PBS containing 1% BSA and 0.1% saponin.
12. Rabbit anti-α-tubulin antibody (Abcam Cat # ab18251).
13. Goat anti-rabbit IgG (H+L) secondary antibody, Alexa Fluor
647 conjugate (Thermo Fisher Cat # A21245).
14. Alexa Fluor 488 anti-human perforin antibody, clone dG9
(Biolegend Cat # 308108).
15. Mounting medium: Vectashield (Vector laboratories Cat #
H-1000) or ProLong Gold antifade mountant (Thermo Fisher
Cat # P36930).
16. Eight-well chambered coverglass (Thermo Scientific Cat #
155411).
17. LysoTracker Red DND-99 (Thermo Fisher Cat # L7528).
18. SiR-tubulin (Spirochrome Cat # SC002). Dissolve the content
of the vial in anhydrous DMSO to make a 1 mM stock solu-
tion. Store in the dark at −20 °C.
19. Verapamil. Dissolve the content of the vial in pure ethanol to
prepare a 10 mM stock solution. Store at 4 °C.
20. CellTrace CFSE cell proliferation kit (Thermo Fisher Cat #
C34554). Dissolve the content of the vial in DMSO to make a
5 mM stock solution. Store in the dark at −20 °C.
21. Clear nail polish.
22. General lab equipment:
(a) Tissue culture biocontainment hood, incubator for cell culture.
(b) Hemocytometer or other suitable device for cell enumeration.
(c) Tweezers.
(d) Leica laser scanning confocal microscope SP8 (see Note 1)
or equivalent.
3 Methods
Carry out all procedures at room temperature unless otherwise

specified.
3.1 Visualizing Lytic 1. Culture NK cells and target cells in appropriate culture medium
Granules (see Subheading 2) at a low density, preferably under 5 × 105/
and Microtubules ml, to enable optimal cellular function (see Note 2). As an
in Fixed Cells alternative ex vivo NK cells freshly isolated by negative selec-
tion using the human NK cell isolation kit (Miltenyi Biotec) or
3.1.1 Preparation of NK
the RosetteSep human NK cell isolation reagent (StemCell
and Target Cells for Fixed
Technologies) may be used.
Cell Imaging
2. Wash NK and target cells once with RPMI medium prior to use.
3. Resuspend NK and target cells with culture medium at the
density of 1 × 106/ml and 2 × 106/ml, respectively, to enable
a 1:2 effector to target ratio.
4. Take 100 μl of NK and target cell suspension and mix them in
a 15 ml polypropylene conical tube (see Note 3).
5. Allow conjugates to form at 37 °C for approximately 10 min.
Note that lytic granule positioning can be measured in non-
conjugated NK cells as well as NK cells that are activated by
another means such as by soluble stimulants (such as cyto-
kines), antibody coated on glass, or planar lipid bilayers (see
Note 4) [12]. The conjugate protocol is offered here as an
example.
3.1.2 Slide Preparation 1. On poly-l-lysine- or silane-coated slides, use a hydrophobic

pen to draw an enclosed section. Squared regions with sides of
1 cm should hold up to 200 μl of liquid (see Note 5).
2. Prepare a humidified covered chamber containing moistened
paper towels for slide incubation during the staining process.
3. Keep the glass slides in the humidified chamber at 37 °C to
pre-warm the slides enabling optimal binding efficiency
of cells.
4. Prepare buffers for the staining process. Fixative and permea-
bilization (fix and perm) buffer: 0.1% Triton X-100 in
Cytofix/Cytoperm buffer. Staining/washing buffer: 1% BSA
and 0.1% saponin in PBS (for the evaluation of living cells, see
Subheading 3.2).
5. Prepare antibodies at desired concentration in the staining/
washing buffer (see Note 6).
6. Following the conjugation process (see Subheading 3.1.1),

gently resuspend the conjugates by pipetting slowly until no
cell aggregates are observed.
7. Transfer the conjugate suspension onto the slides within the
hydrophobic enriched region.
8. Allow cells to adhere to the slides in the humidified chamber
for approximately 20 min at 37 °C.
9. Use a 20–200 μl pipette, gently lay the tip against the ink, and
point the tip toward the constrained liquid to carefully with-
draw liquid from the hydrophobic enriched region. Be sure
not to touch the cells or disrupt the hydrophobic ring. The
cells should now appear as a thin cloudy layer on the slides.
10. Add 150 μl of fix and perm buffer slowly to the hydrophobic
enriched area and incubate for 20 min at room temperature.
Be careful not to disrupt the cells on the slides.
11. Withdraw buffer from the slides as per step 9.
12. Add 200 μl of staining/washing buffer to the ink-enclosed
area to rinse off excess fix and perm buffer.
13. Repeat the washing steps 11 and 12 twice.
14. Aliquot 150 μl of primary antibody solution to the staining
region and incubate in the humidified chamber for 30–60 min.
If using a fluorophore directly conjugated antibody, skip to
step 18.
15. Repeat steps 11–13 to remove residual primary antibody.
16. Add 150 μl of fluorophore-conjugated secondary antibody
solution to the hydrophobic ink region and incubate in the
humidified chamber for 30–60 min.
17. Repeat steps 11–13 to wash off the residual secondary
antibody.
18. Add directly conjugated fluorescent antibody that recognizes
other molecules of interest and incubate for 30–60 min in the
humidified chamber.
19. Repeat steps 11–13.
20. Perform steps 14–19 until all molecules of interest are stained
subsequently.
21. Add one drop of mounting medium to each staining region
(see Note 7).
22. Place a coverslip on the slide using forceps. Gently press on
the coverslip to remove excess mounting medium and dry
with tissue paper.
23. Keep slides in the dark overnight with the coverslip facing
upward.
24. Seal all sides of the coverslip with nail polish.
3.1.3 Imaging 1. Initialize the confocal microscope system according to the

manufacturer’s instructions, and allow the lasers to reach stable
working temperature and output power. The results of the
experiment can be acquired using any of the two most com-
mon setups to achieve confocal imaging: either a confocal laser
scanning microscope (CLSM) or a confocal spinning disk
microscope (CSDM). One typical CLSM system is detailed in
Note 1, but this protocol is applicable to any confocal system
with appropriate laser lines (for CLSM) or laser line and excita-
tion/emission filter sets (for CSDM).
2. Select an apochromatic, high-magnification objective (60× or
higher) with a high numerical aperture (above 1.2) and add a
drop of immersion oil.
3. Secure the sample on the stage and adjust the focus to be able
to clearly identify cells using the eyepiece and bright-field illu-
mination for transmitted light.
4. Adjust the specific illumination parameters for the acquisition
of images of the best resolution possible while maintaining the
photobleaching of the dye to a minimum and removing all
potential chromatic spillover between the fluorescent dyes
imaged:
(a) Confocal scan head settings: rate of 600 Hz for an inter-
mediate image size of 512 × 512 pixels (see Note 8) and a
line averaging value of 6–8 passages. The pinhole should
be set to 1 airy unit to maintain high confocality and Z
resolution precision.
(b) Excitation laser lines are set to coincide with the maximal
excitation wavelength of the fluorescent dye used in the
experiment (see Note 9).
(c) The spectral gates for each detector for each fluorescent
reporter are set to collect photons at wavelength overlap-
ping with the maximum emission peak of the fluorescent
dye (see Note 9).
(d) For conventional PMT (photomultiplier tube) detectors,
the adjustment gain is set between 600 and 800 V. For the
high-sensitivity and low-noise GaAsP (gallium arsenide
phosphide) detectors, the common range is 100–250% (see
Note 10).
5. Adjust the Z position of the sample and, if 3D imaging is per-
formed, define the bottom and the top limits of the volume
encompassing the entire cell or conjugate. Set the step size
according to the recommendation of the acquisition software.
6. Start the acquisition and proceed similarly for at least 20 cells.
7. Analyze the preliminary results as described below before per-
forming a power analysis test to evaluate the required sample
size. Acquire additional data as necessary.
3.2 Visualizing Lytic 1. Pre-coat the chambered coverglass with a monoclonal antibody
Granules at the concentration of 5 μg/ml in PBS that will specifically
and Microtubules recognize the target cells and not the NK cells for the purpose
in Live Cells of immobilizing the target cell (see Note 11, Fig. 1). Allow the
3.2.1 Preparation
antibody to adhere to the coverglass at 37 °C for at least 30 min
of the Imaging Chambered
or at 4 °C overnight. Alternatively, if target cells are not used,
Coverglass
pre-coat the chambered coverglass with antibodies that will
engage the adhesion and/or activation receptors on the sur-
face of NK cells to induce formation of an “artificial” IS as an
activating surface (see Note 4, Fig. 1).
2. Rinse the antibody-coated chambered coverglass three times
with PBS prior to use.
3.2.2 Preparation of NK 1. Prepare NK and target cells separately, washing each once with
and Target Cells for Live RPMI 1640 by spinning at 225 × g for 5 min.
Cell Imaging 2. Add 0.5 μl of SiR-tubulin and 1 μl of verapamil stock solution
to 1 × 105 NK cells in 1 ml of R-10 medium (500 nM and 10
μM, respectively) and incubate for 1 h at 37 °C before con-
tinuing to step 3. Alternatively, a cell expressing a fluorescent
protein-
conjugated microtubule biosensor may be utilized
(while there are many possibilities, examples include FP-α-
tubulin and FP-MAP4) in which case the entirety of this step
would be skipped.
Fig. 1 Diagram of different approaches for analyzing lytic granule convergence in NK cells. The left panel
shows the diffusely distributed lytic granules in the cytoplasm before the formation of the immune synapse
(IS). The right panel demonstrates how lytic granules are tightly clustered around the MTOC after IS formation.
Within each panel, the left-hand side depicts formation of an artificial IS between an NK cell and the glass
surface coated with activating antibodies against the adhesion and activation receptors for NK cells. On the
right-hand side of each panel, it shows IS formation between an NK cell and its susceptible target carrying the
ligands that trigger NK cell activation
3. Add 10 μl of LysoTracker Red DND-99 to 1 × 105 cells NK

cells in 1 ml of R-10 medium (10 μM) and incubate for another
30 min at 37 °C. During incubation, proceed to step 4 for
target cell preparation.
4. Add 1 μl of CFSE stock solution to 2 × 105 target cells in 1 ml
of R-10 (5 μM) and incubate for 5 min at 37 °C.
5. Following incubation, wash NK and target cells three times
with R-10 by spinning at 225 × g for 5 min.
6. In separate tubes, resuspend NK and target cells with dye-free
R-10 medium supplemented with verapamil (10 μM) at the
density of 1 × 105/ml and 2 × 105/ml, respectively, to enable
a 1:2 effector to target ratio.
7. Aliquot 100 μl of target cell suspension to the pre-coated and
PBS-rinsed well of the eight-well chambered coverglass. Allow
the target cells to adhere to the coverglass for approximately
10 min at 37 °C.
8. Add 100 μl of NK cell suspension to the target cells after hav-
ing transferred the sample on the microscope stage.
3.2.3 Imaging 1. To ensure the stability of the sample during live imaging and
avoid axial drifting, it is absolutely paramount to equilibrate
the temperature of the microscope chamber with the sample.
2. Adjust the confocal scanning head settings and illumination
settings according to Subheading 3.1.3. In order to reduce the
phototoxicity, some parameters should be set to conservative
values if available as follows:
(a) Set the laser scanning mode to resonant in order to achieve
faster laser scanning rates (14,000 Hz).
(b) To compensate for the faster rate, increase the line averag-
ing to 16–32 passages per line. It is not recommended to
use any of the accumulation functions as they result in loss
of time resolution.
(c) In practice, increasing the Z-spacing between the optical
sections to 1 μm might be considered as a fair compromise
for reducing photodamaging of the cell (and in light of
the mean size of lytic granules and spatial resolution of a
confocal microscope).
3. Using the eyepiece and using transmitted light, identify iso-
lated cells or conjugates (depending on your experiment).
Using fluorescence and the camera- or photomultiplier tube-
captured preview image, set the Z position of the sample to
define the bottom and the top limits of the volume encompass-
ing the entire object of interest.
4. Acquire one single stack and review your illumination setting
accordingly.
5. Visually identify five objects of interest and start the acquisition

of the time lapse with 5-min interval between each time point,
for a total duration of 2 h (see Note 12).
6. Repeat the experiment to collect at least 20 cells and perform
a power analysis test as described in Subheading 3.1.3, step 7.
3.3 Measuring Lytic The degree of lytic granule convergence is demonstrated by the
Granule Convergence mean distance of individual granules to the MTOC in an NK cell.
to the MTOC in Fixed The more scattered the granules are within an NK cell, the larger
Cells the numerical value becomes. The values may vary among NK cell
lines and primary NK cells due to their size differences. Therefore,
it is critical to always include control groups to enable objective
comparisons. To calculate granule to MTOC distance, it is required
to obtain coordinates of the centroids of the lytic granules and the
MTOC. For this purpose, multiple software platforms are available
to the scientific community. While the use of other platforms is also
viable, we offer below the analysis workflow for two currently avail-
able software platforms: Volocity (Perkin Elmer) and Fiji [13, 14].
3.3.1 Volocity 1. Create a new library in Volocity. Drag the image files onto
Library to open the data set.
2. Select Measurements tab to initiate the features for image
analysis.
3. Go to Properties under Edit. Set the calibrated X, Y, and Z
pixel dimensions (measured) according to the microscope
used for image acquisition.
4. Create a measurement protocol to determine the coordinates
of each and every lytic granule in individual NK cells:
(a) Drag and drop the Find Objects onto the Tasks panel.
Select the appropriate channel in which perforin is to be
identified. Replace the default name Population 1 with
Lytic Granules. Click on the gear icon in the protocol to
access the dialog for intensity thresholding. Select
Threshold using Intensity and drag the red bar to include
voxels that define perforin and remove noise (see Note
13). Lastly, set the Minimum object size as 0.03 μm3 (see
Note 14).
(b) Drag and drop the Clip to ROIs into the Find Objects
protocol.
(c) Drag and drop the Separate Touching Objects into the
existing protocol. Set Object size guide at 0.01 μm2 to
separate lytic granules in close proximity.
(d) Drag and drop Compartmentalize into the Tasks panel
and select to Divide Lytic Granules Between ROIs. This
separates the lytic granules within different ROIs (NK
cells).
(e) Click Measure at the bottom of the protocol and select

Centroid.
(f) Select Save Protocol… from the Measurements menu.
Name and save the protocol as Lytic Granules to allow
rapid application for image analyses.
5. Create a second measurement protocol to determine the coor-
dinates of the MTOC in NK cells:
(a) Drag and drop the Find Objects onto the Tasks panel.
Select the appropriate channel in which α-tubulin or peri-
centrin is to be identified and name the population MTOC.
(b) Click on the gear icon in the protocol to access the dialog
for intensity thresholding. Set the Minimum object size
as 0.1 μm3. Select Threshold using Intensity and drag the
red bar to include voxels that define one single MTOC in
each NK cell.
(c) Drag and drop the Clip to ROIs into the Find Objects
protocol.
(d) Click Measure at the bottom of the protocol and select
Centroid.
(e) Select Save Protocol… from the Measurements menu.
Name and save the protocol as MTOC.
6. Click on the image to be analyzed in the library.
7. Use Freehand or Circle tool to draw individual regions that
include the NK cells in the conjugates that are to be analyzed.
Each enclosed region is accounted as one ROI (NK cell), sepa-
rated from the others in the same image field and thus will be
analyzed individually. Be sure to exclude any part of the target
cells from the marked regions.
8. Click on the saved protocols Lytic Granules and MTOC under
Saved protocols in the Tasks panel to apply all settings to the
image under analysis.
9. Select Make Measurement Item… from the Measurements
menu. Create a new file or add the results to an existing mea-
surement table.
10. Repeat steps 6–9 until all desired images are analyzed.
11. Click on the measurement item created in the library. Go to
Columns under Raw to select the data you wish to be dis-
played (e.g. Item Name, Group ID, Centroid X and Y).
12. Select Lytic Granules from the dropdown menu of Display.
13. Save the X and Y coordinates for Lytic Granules to an Excel
sheet.
14. Repeat steps 12–13 for MTOC measurements.
15. Go to Subheading 3.3.3 for calculations and data plotting.
3.3.2 Fiji 1. Open your raw images by dragging their icon onto the main
task bar.
2. If more than a cell is present in the frame, use the Freehand
tool to circle one and select Clear Outside to restrict all the
channels to one cell only.
3. If running this analysis for the first time, select the desired
measurements to be captured in the Set Measurements menu.
Tick the options for Area, Centroid , Center of Mass, and Stack
Position.
4. This analysis needs to be done for each channel sequentially
(Fig. 2a).
5. Adjust the brightness of each fluorescence channel using the
Brightness/Contrast tool to be able to distinguish the subtle
details surrounding the objects of interest.
6. Using the Threshold panel, turn on the checkbox for Dark
Background and, using the sliders, highlight the population
of pixels that form the object of interest based on the fluores-
cence intensity displayed in the pixel intensity histogram. To
measure the localization of the MTOC from a microtubule
straining (α-tubulin instead of pericentrin), make sure to only
include the brightest portion of your signal to limit the detec-
tion to one object per cell.
7. Optionally, if a high level of background is present in the
image, it is possible to reduce it using a Rolling Ball algo-
rithm found under Process > Subtract Background… and
set to a radius of two to four times the size of the object under
analysis to avoid removing any object of interest (usually a 20
pixel wide radius is a safe choice).
8. Click Apply to transform the fluorescence image into a binary
image. If the granules are contiguous or overlapping, a
Watershed segmentation algorithm can be applied and is
found under the Process > Binary menu.
9. Go to Analyze > Analyze Particles… menu and set a filter for
the granule size from 0.05 μm2 to 2 μm2 or 0.1 μm2 to 1 μm2
for the MTOC. Select the following options: Display Results,
Show Outlines, and Exclude on Edges. Turn on the check-
box Clear Results to not add this set of measurement to the
previously analyzed image.
10. Visually inspect the accuracy of the detection using the out-
lines drawn on the image after detection (Fig. 2a). If too many
objects are missed, return to step 4 and adjust the values of
the threshold until a satisfactory coverage is obtained.
11. Save the spreadsheet containing all the positions measured in
the current image and proceed to the next channel or image.
The position of each granule is under the column X, Y, and Z
if applicable.
Fig. 2 Images extracted from the analysis workflow to illustrate the different critical steps in a sequential way
(a): adjustment of brightness and contrast, analysis performed on each channel independently, segmentation of
the objects, and finally, measurement of the distance between the centroid of each granule and the centroid of
the MTOC (assumed to be the brightest object within the α-tubulin staining shown here). Examples of the confo-
cal images (b) and plots (c, d) displaying the measurement of lytic granule convergence on an activating surface
3.3.3 Calculations 1. Measure lytic granule convergence using the MTOC to gran-
and Data Plotting ule distance (MGD) formula, where x, y, and xi, yi are centroid
coordinates for the MTOC and the individual lytic granules,
respectively. The MGD measures the shortest length between
centroid of granules and MTOC as depicted in Fig. 2a. “n”
indicates the number of granules, in a given frame, selected by
Volocity or Fiji based on the intensity threshold used.
(∑i =1 (x − xi )2 + (y − y i )2 )
n
MGD =
n
2. For images involving 3D reconstructions of multiple z-axis
planes, the centroid coordinates for MTOC and lytic granules
should be measured in the same way as described above using
the formula provided below:
(∑i =1 (x − xi )2 + (y − y i )2 + (z − zi )2 )
n
MGD =
n
3. Calculate the physical distance of each and every granule from
the MTOC as described in step 1 or step 2 (2D or 3D). Take
the mean of all calculated distances in individual NK cells and
plot them as single dots. In the dot plots, x-axis contains dif-
ferent sample populations (e.g., control vs. patient, vehicle vs.
drug treated; Fig. 2b), and y-axis shows the mean lytic granule
to MTOC distances (Fig. 2c). Each dot represents how dis-
persedly lytic granules are localized within one NK cell conju-
gated with a target cell. The more converged the lytic granules
are, the smaller the number will become, ranging from close to
zero (all granules converged tightly around the MTOC) to
approximately the radius of the NK cells (all granule at the
periphery of the cell).
3.4 Measuring Lytic Fixed cell confocal microscopy allows for high-resolution images
Granule Convergence and the convenience of collecting sample sizes that are large
to the MTOC/Granule enough to achieve a desired power with biologically relevant dif-
Centroid in Live Cells ferences. Meanwhile, live cell confocal microscopy provides tem-
poral resolution that is critical in exploring the dynamics in the
biological processes involved in NK cell cytotoxicity such as
synapse maturation. For instance, whether a genetic defect or drug
treatment affects lytic granule convergence in NK cells can be
Fig. 2 (continued) (c) and in effector-target cell conjugates (d). NK cells are labeled with SiR-tubulin and
LysoTracker Red to identify the positions of the MTOC and lytic granules. The dot plot shows the degree of lytic
granule convergence in NK cells using fixed cell confocal microscopy. Each dot indicates the average granule to
MTOC distance within a single NK cell (c). The line plot shows the average granule to MTOC distance over time
using live cell confocal microscopy. Each line represents one NK cell. In the control group, the gradual decrease
in distances indicates the real-time converging process of lytic granules toward the MTOC (d)
rapidly determined by fixed cell imaging analysis, while live cell

imaging can further resolve the potential temporal delay or the
instantaneous velocity and directionality of lytic granule move-
ments. Therefore, fixed and live cell imaging analyses are both
complementary and equally important in studying NK cell
cytotoxicity.
3.4.1 Volocity Considerations for measuring lytic granule convergence in live cell
imaging are fairly similar to those outlined for fixed cells in
Subheading 3.3.1 with several variations:
1. After opening the data file, use the Rectangle or Freehand
tool to draw and encircle one NK-target cell conjugate that is
to be analyzed first.
2. Go to Actions and select Crop to Selection. A new image file
will now appear underneath the original file.
3. Perform the convergence analyses as per steps 2–19 in
Subheading 3.3.1 with a few modifications:
(a) L ytic granules and the MTOC are defined by the fluores-
cent channels encompassing LysoTracker Red and SiR-
tubulin, respectively.
(b) Step 4: Instead of thresholding using intensity, select
Threshold using SD to adjust for the potential decay of
fluorescent intensity over time in time-lapse experiments
(see Note 13).
(c) Step 7: After marking the region, which confines the NK
cell in the initial time frame, go through the following
time frames to ensure that the enclosed area will contain all
the lytic granules within the NK cell. Be sure to exclude
the target cell or any other NK cells in close proximity at
all time points. If the NK-target cell conjugate underwent
substantive movement over time, mark the area that con-
fines the NK cell individually for each time frame.
(d) Step 11: When making a measurement item, select Selected
time point(s) or All time points depending on step 7.
4. If SiR-tubulin staining was not performed, centroid of the
MTOC can be replaced by centroid of all the lytic granules to
determine the variance in relative positioning of lytic granules:
(a) Click on the measurement item containing the coordinates
of lytic granules.
(b) Select all the data within a single time frame.
(c) Go to Raw and select the Join Objects.
(d) The coordinates of the centroid of all granules will appear
as a new object in the measurement table. This can then be
used to determine the distance from each granule.
3.4.2 Fiji Measuring lytic granule convergence in time lapses instead of still
images is virtually identical, and the procedure is identical to the
one described above in Subheading 3.3.2. The measurements are
performed on all time points present in the current image sequence.
The reference to the time point in the final spreadsheet is given by
the Stack Position column. Only two points need particular atten-
tion as follows:
(a) The NK cell must remain inside the region of interest during
the entire duration of the time lapse. Expand slightly the ROI
drawn on the first frame to accommodate all its positions in
the subsequent frames.
(b) If only the staining for the lytic granules was successful, the
position of the centroid of the MTOC over time can be
replaced by the center of mass of all the granules detected in
the binary image (after step 6 in Subheading 3.3.2): select the
option Analyze > Measure… The positions will be registered
in a table under the columns XM, YM, and ZM if applicable.
3.4.3 Calculations Perform measurement of lytic granule convergence as described in

and Data Plotting Subheading 3.3.3 for each time frame and plot the temporal data
from each NK cell as one single line. In the line plots, x-axis shows
all time points during the live imaging process, and y-axis indicates
mean granule to MTOC distance (Fig. 2d).
4 Notes
1. The microscope setup described here can be replaced with any

other confocal system equipped with similar capacities in term
of excitation laser lines and detection: Leica SP8 inverted laser
scanning confocal microscope (Leica microsystems) with Leica
APO PL 63× objective (NA = 1.47) or another high numerical
aperture oil immersion objective. The excitation laser is a
white light laser configured to provide discrete illumination
from 470 to 670 nm and controlled LAS AF software for
image acquisition (supplied with the microscope). Emitted
light is collected by PMT or HyD detectors, following selec-
tive optical plan transmission via a variable pinhole. The micro-
scope stand is enclosed in a heated chamber for live
experiments.
2. To enable optimal cellular function and activity, it is recom-
mended to split the cells or supplement the culture with fresh
culture medium the day before experiments.
3. To optimize conjugation efficiency by increasing the chances
of cell-cell encounter, it is recommended to reduce the vol-
ume of NK-target cell mixture, preferably ranging from 100
to 200 μl with a total number of 1–2 × 106 cells. It is also

recommended to have a 1:2 or even 1:3 effector to target
ratio.
4. To activate NK-92 cells or human primary NK cells, pre-coat
the chambered coverglass or glass slides with anti-CD18
(clone TS1/18) and anti-NKp30 (Bio Legend Cat # 325204).
For antibodies, to activate YTS cells, pre-coat the glass with
anti-CD18 and anti-CD28 antibodies. In each case there are
numerous other alternatives that can be utilized. All coating
antibodies should be diluted to a concentration of 5 μg/ml
with PBS and allowed to adhere to the coverglass for at least
30 min at 37 °C or overnight at 4 °C.
5. Culture slides with removable chambers can be used as an
alternative measure. This is especially useful when the number
of effector cells is limited (e.g., patient primary NK cells).
6. Antibody titration is essential to ensure efficient fluorescent
detection of a molecule of interest while preventing unspecific
staining that will contribute to a high fluorescent background
and a reduced signal-to-noise ratio. Guidelines for antibody
titration are as follows:
(a) Each primary antibody should be titrated individually and
compared to an isotype control.
(b) Each secondary antibody should be titrated using a fixed con-
centration of a primary antibody with known effectiveness.
(c) Utilize a wide range of antibody concentrations to deter-
mine the ideal concentration that provides efficient fluo-
rescence at reasonable microscopic settings and good
signal-to-noise ratio.
(d) The signal-to-noise ratio can be measured for each anti-
body dilution by calculating the mean fluorescence inten-
sity (MFI) of ten objects (at least) and divide it by the MFI
of the same number of similarly sized regions of the back-
ground. Ideally, this ratio should be above 3.
7. Several commercially available mounting media can be used to
preserve the samples for storage and to reduce photobleach-
ing. ProLong Gold is a hard-setting antifade mounting solu-
tion specifically designed for the long-term storage of samples
stained by Alexa Fluor dyes. Its optical properties evolve over
time, and the sample should only be imaged after 20 h mini-
mal to allow the increase of refractive index from 1.42 to 1.45.
Vectashield is another excellent choice which has better optical
properties immediately after use (higher refractive index of
1.45) and, as a non-hardening medium, helps to preserve
three-dimensional structures.
8. All the recent line scanning confocal microscopes are equipped
with resonant scanners to allow acquisition at very high
frequency (above 10,000 Hz). Although this is a substantive

advantage for fast live acquisition while minimizing photo-
damage, it comes at a cost of an increase in noise. Therefore, a
higher value for line averaging needs to be used to average out
the electronic noise in the system. Routinely, the regular scan-
ning method is preferred for fixed sample imaging where reso-
lution is more important than speed and sample preservation.
9. Point scanning confocal microscopes allow the user to pre-
cisely set the spectral gates for each detector, unlike the spin-
ning disk confocal systems which are based on a set of
predetermined chromatic filters and beam splitters. Although
the former are more flexible than the latter, they require pre-
cise knowledge of the optimal absorption and emission wave-
lengths of each dye used in the experiment. This information
is typically provided with the consumables and must be care-
fully studied to organize the collection of emitted photons
between the detectors in different sequences. Usually, the
spectral gates are all set to collect photons at wavelengths
matching the maximum emission peak of each dye and then
split into different sequential illumination sequences to sup-
press spillover between dyes with overlapping spectra. Finally,
the segregation of the different signals must be confirmed by
acquiring all wavelengths while disabling each laser line suc-
cessively. The result should be a disappearance of the signal
specific to the disabled excitation wavelength with little to no
change in the intensity of the other wavelengths acquired.
10. The gain value of the detectors should be very carefully adjusted.
Typical PMT display a linear conversion range between 600 and
800 V. If it needs to be lowered below 600 V, then the laser line
intensity must be lowered instead; conversely it needs to be
increased if the gain setting reaches above 800 V. The newest
generation of detectors (GaAsP) is not so sensitive to noise and
displays a linear amplification across a much broader range (50–
300% from our experience). Consider lowering the laser line
intensity if the amplification is below 100%.
11. To capture and immobilize the K562 and 721.221 cells on the
chambered coverglass for increased stability and hence higher
live imaging quality, pre-coat the wells with antibodies against
CD58 and CD48 antibodies, respectively.
12. The length of acquisition depends on the aim of the experi-
ment. Image for 30–60 min if only observation of conver-
gence/polarization is required and image for 1–2 h if death of
the target cell is also of interest.
13. To allow for unbiased comparisons, it is essential to apply the
same threshold setting for images within one experiment and
ideally for images from both the control and experimental
groups. It is also recommended to acquire all images on the
same day. To apply thresholding using intensity (fixed cell) or

SD (live cell) in Volocity, drag the red bar along the x-axis of
the histogram of intensities until Volocity picks up all the sig-
nals that define the molecule of interest in the image. The goal
is to include the desired signals while diminishing unwanted
(but quantitatively determined) background.
14. The size of lytic granules in NK-92 cells, as determined by
super-resolution microscopy, ranges from 100 to 650 nm in
diameter with the majority of them ranging from 250 to
400 nm [15].
Acknowledgments
We would like to thank Dr. Ashley Mentlik-James for her efforts in

developing this protocol in part as part of her Ph.D. thesis. This
work was supported by NIH grants R01AI069746 and
R01AI120989 to J.S.O.
References
1. Orange JS (2008) Formation and function of 7. Zhang M, March ME, Lane WS, Long EO
the lytic NK-cell immunological synapse. Nat (2014) A signaling network stimulated by
Rev Immunol 8:713–725 beta2 integrin promotes the polarization of
2. Mace EM, Dongre P, Hsu HT, Sinha P, James lytic granules in cytotoxic cells. Sci Signal
AM, Mann SS, Forbes LR, Watkin LB, Orange 7:ra96
JS (2014) Cell biological steps and checkpoints 8. Mace EM, Wu WW, Ho T, Mann SS, Hsu HT,
in accessing NK cell cytotoxicity. Immunol Cell Orange JS (2012) NK cell lytic granules are
Biol 92:245–255 highly motile at the immunological synapse
3. Mentlik AN, Sanborn KB, Holzbaur EL, and require F-actin for post-degranulation
Orange JS (2010) Rapid lytic granule conver- persistence. J Immunol 189:4870–4880
gence to the MTOC in natural killer cells is 9. Briercheck EL, Trotta R, Chen L, Hartlage
dependent on dynein but not cytolytic com- AS, Cole JP, Cole TD, Mao C, Banerjee PP,
mitment. Mol Biol Cell 21:2241–2256 Hsu HT, Mace EM, Ciarlariello D, Mundy-
4. James AM, Hsu HT, Dongre P, Uzel G, Mace Bosse BL, Garcia-Cao I, Scoville SD, Yu L,
EM, Banerjee PP, Orange JS (2013) Rapid Pilarski R, Carson WE 3rd, Leone G,
activation receptor- or IL-2-induced lytic gran- Pandolfi PP, Yu J, Orange JS, Caligiuri MA
ule convergence in human natural killer cells (2015) PTEN is a negative regulator of NK
requires Src, but not downstream signaling. cell cytolytic function. J Immunol
Blood 121:2627–2637 194:1832–1840
5. Hsu, H.T., Orange, J.S. (2014) Distinct
10. Ham H, Huynh W, Schoon RA, Vale RD,
integrin-dependent signals define requirements Billadeau DD (2015) HkRP3 is a microtubule-
for lytic granule convergence and polarization binding protein regulating lytic granule clus-
in natural killer cells. Sci Signal 7:pe24. tering and NK cell killing. J Immunol
6. Ritter AT, Asano Y, Stinchcombe JC, 194:3984–3996
Dieckmann NM, Chen BC, Gawden-Bone C,
11. Tuli A, Thiery J, James AM, Michelet X,
van Engelenburg S, Legant W, Gao L, Davidson Sharma M, Garg S, Sanborn KB, Orange JS,
MW, Betzig E, Lippincott-Schwartz J, Griffiths Lieberman J, Brenner MB (2013) Arf-like
GM (2015) Actin depletion initiates events GTPase Arl8b regulates lytic granule polariza-
leading to granule secretion at the immuno- tion and natural killer cell-mediated cytotoxic-
logical synapse. Immunity 42:864–876 ity. Mol Biol Cell 24:3721–3735

12. Bertolet G, Liu D (2016) The planar lipid Rueden C, Saalfeld S, Schmid B, Tinevez JY,
bilayer system serves as a reductionist approach White DJ, Hartenstein V, Eliceiri K, Tomancak
for studying NK cell immunological synapses P, Cardona A (2012) Fiji: an open-source plat-
and their functions. Methods Mol Biol form for biological-image analysis. Nat
1441:151–165 Methods 9:676–682

13. Schneider CA, Rasband WS, Eliceiri KW 15. Rak GD, Mace EM, Banerjee PP, Svitkina T,
(2012) NIH image to ImageJ: 25 years of Orange JS (2011) Natural killer cell lytic gran-
image analysis. Nat Methods 9:671–675 ule secretion occurs through a pervasive actin

14. Schindelin J, Arganda-Carreras I, Frise E, network at the immune synapse. PLoS Biol
Kaynig V, Longair M, Pietzsch T, Preibisch S, 9:e1001151
Chapter 32
Studying the T Cell-Astrocyte Immune Synapse

George P. Cribaro, Elena Saavedra-López, Paola V. Casanova,
Laura Rodríguez, and Carlos Barcia
Abstract
In this chapter, we describe the technical details to visualize and analyze effector immunological synapses
between T cells and astrocytes in the brain with high-resolution confocal imaging. This procedure is criti-
cal for the optimal and even penetration of labeling antibodies within the nerve tissue to obtain accurate
staining and allow a uniform three-dimensional analysis of the T cell-astrocyte interactions. We emphasize
here the comprehensive exploration of the tissue and analysis with confocal microscope as well as the
display of microanatomical details of the three-dimensional reconstruction for interface visualization
(including peripheral and central supramolecular activation clusters, effector molecules, and other
organelles such as microtubule organizing centers (MTOCs) and Golgi apparatus).
Key words T cell, Astrocyte, c-SMAC, p-SMAC, Tissue, 3D reconstruction, Immunofluorescence,

Confocal microscopy
1 Introduction
Cytotoxic T cells (CTLs) establish immunological synapses (ISs)

with effector functions to execute target cells [1]. Particularly, the
IS engagement of CTLs results in the formation of supramolecular
activation clusters (SMACs), including the LFA-1-rich peripheral
cluster (p-SMAC) and, importantly, two segregated domains at the
central SMAC (c-SMAC) containing either T cell receptor (TCR)
molecules or lytic granules [2]. Effector T cells infiltrate the brain
parenchyma searching for antigens to establish bona fide ISs, able
to form Kupfer-type SMACs within the brain tissue [3, 4]. These
events have been described during the establishment of mature IS
of T cells engaging with astrocytes previously infected with adeno-
viruses in rat brains [5], with astrocyte-phenotype glioma cells in
human tissue biopsies [6] and with astrocytoma cells in experimen-
tal models of astroglioma in mouse [7]. In these cases, multiple
fluorescence labeling of the molecules involved in the IS was
required, along with specific markers to identify the target cells.
517
518 George P. Cribaro et al.
Therefore, optimal histology processing and an adapted immuno-

fluorescence protocol are essential steps for successful results.
We describe here the specifics for achieving optimal quality of
confocal imaging of effector CTL ISs within the brain tissue. This
requires a precise sample fixation procedure as well as the particular
specifics of in-tissue immunolabeling. In the case of animal tissue,
optimal brain perfusion with saline or Tyrode is important to
obtain best results (unfixed frozen tissue is not a good option to
identify the microanatomical details of the IS). After obtaining
serial floating brain sections of the areas of interest, a particular
immunofluorescence protocol, different from standard proce-
dures, should be conducted. Regular immunofluorescence staining
protocols are usually performed on very thin sections, around
5–10 μm. This reduced thickness is advantageous for visualizing a
single layer of cells, which allows feasible and rapid quantifications
providing vital information for pathological diagnosis. However,
to study complex and three-dimensional (3D) intercellular interac-
tions, such as IS, this range of tissue section thickness is insuffi-
cient. Thicker sections of 40–60 μm are ideal to visualize T
cell-astrocyte intercellular communications in the 3D space. In the
following sections, we describe the procedure for achieving uni-
form and even staining within brain tissue sections and the imaging
and analysis methods that must be employed to visualize immune
interactions as well as certain details of the effector T cell-astrocyte
IS at the interface.
2 Materials
2.1 Brain Tissue The protocol described here is suited to brain tissues where previ-
Sections ous adaptive immune responses have been experimentally stimu-
lated or are present due to pathological origin, especially when
referring to human samples. Specific antigen CTL responses against
astrocytes can be visualized in several scenarios, including CTL
responses induced experimentally by the intracranial injection of
astrocyte-specific adenoviral vectors [3] and CTL responses against
astrocyte-phenotype glioma cells [6, 7]. To obtain the highest
image quality and resolution, it is extremely important that tissue
blocks are well fixed, structurally preserved, and properly sec-
tioned, avoiding freezing. We recommend as the best option for
sectioning the use of a vibratome at room temperature. If room
temperature is not an option for sectioning (i.e., when using cryo-
stat), whole brains should be soaked previously in 20–30% sucrose
to preserve the microanatomy of the tissue when freezing (if this
cryoprotection step is not done, ice crystals severely damage the
tissue microanatomy). Thereafter, sectioning with cryostat or
microtome can be performed at freezing temperatures (−20 °C).
For long term, serial sections can be stored at 4 °C in PBS with
Studying the T Cell-Astrocyte Immune Synapse 519
0.01% sodium azide or in PBS 20–30% sucrose at −20 °C, to

enhance preservation by avoiding microbial contamination.
2.2 Buffers All solutions are prepared using distilled water and analytical grade
and Solutions reagents. All reagents are prepared and stored at room temperature
unless otherwise indicated:
1. Phosphate-buffered saline (PBS): 0.171 M NaCI (10 g/L),
0.013 M Na2HPO4 (1.8 g/L), 2.5 mM NaH2PO4 (0.3 g/L),
and pH 7.4 (see Note 1).
2. Tris-buffered saline (TBS) + 0.5% (v/v) Triton® X-100: 0.0578 M
C4H11NO3 (Trizma® base) (7 g/L), 0.154 M NaCl (9 g/L),
0.5% (v/v) Triton® X-100 (5 g/L), and pH 7.4 (see Note 2).
3. Blocking solution 1% or 10% horse serum (see Note 3) block-
ing solution (10% HS): TBS-0.5% Triton® X-100 containing
10% horse serum. Add 20 mL of TBS + 0.5% Triton® X-100 to
a 50- or 100-mL beaker. Weigh 0.03 g NaN3 and transfer to
the beaker. Add 3 mL horse serum and fill to the 30-mL mark
with TBS + 0.5% Triton® X-100. Store at 4–8 °C.
4. Blocking solution 2% or 1% horse serum solution (1% HS):
TBS-0.5% Triton® X-100 containing 1% horse serum. Add 20
mL of TBS + 0.5% Triton® X-100 to a 50- or 100-mL beaker.
Weigh 0.03 g NaN3 and transfer to beaker. Add 3 mL 10%
horse serum and fill to the 30-mL mark with TBS + 0.5%
Triton® X-100. Store at 4–8 °C.
5. Citrate buffer: 0.033 N C6H8O7 (2.1 g/L), pH 6.
2.3 Antibodies Brain tissue samples containing T cell-astrocyte interactions should

and Other Specific be processed for fluorescent immunohistochemistry detection with
Solutions primary antibodies. As described at the beginning of Subheading 2,
good examples are brains from adenovirus-injected animals [3] or
brains with experimentally induced or pathological astrocyte-
phenotype tumors [6].
2.3.1 Detection Preferably, the immunofluorescent detection of astrocytes can be per-

of Astrocytes or Other formed with antibodies against Glial fibrillary acidic protein (GFAP),
Astrocyte-Phenotype Cells which are reliable markers for astrocytes in fixed tissue. However,
(i.e., Glioma Cells) other potential markers can be used [8]. Primary antibodies against
vimentin may also be used but with some reservations (see Note 4).
2.3.2 Detection Ideally, the detection of the p-SMAC can be performed with pri-
of SMACs mary antibodies against LFA-1. Alternatively, phalloidin staining
to detect F-actin can be used, as this will also show a p-SMAC pat-
tern in mature ISs. To detect the c-SMAC, antibodies against
TCR/CD3 are the adequate option (CD8 antibodies may also be
valid depending on the immunological scenario), whereas antibod-
ies against effector molecules such as IFN-γ or granzyme-B may
also detect clusters at the c-SMAC in cytotoxic T cells.
2.3.3 Labeling Other T Detection of T cell activation signaling tyrosine kinases such as
Cell-Activating Molecules antibodies against Lck and ZAP-70 could also be useful to visual-
and Structures ize TCR signaling and polarization at the interface. Additionally,
detection of polarized organelles such as Golgi apparatus with anti-
bodies against GM130 and MTOCs with antibodies against
γ-tubulin can help to visualize the polarization of the T cell toward
the target. DAPI (4′,6-diamidino-2-phenylindole) in PBS 1:1000
(1 μL/mL) to detect the cell nuclei. Antifade mounting medium
ProLong® Gold (Invitrogen) is convenient because the medium is
a gel that hardens and remains stable for months. Nail polish or
other sealers are not needed to retain the liquid and keep the cov-
erslip attached.
3 Methods
3.1 Antigen Retrieval Protocols of antigen retrieval are critical steps to facilitate the pen-
etration of the antibodies in thick tissue sections. In our experi-
ence, citrate buffer treatment is the best option, because it results
in uniform staining, in contrast with other treatments such as tryp-
sin or acetone.
This series of steps employs heated citrate buffer for antigen
retrieval:
1. Add 3 mL PBS wash solution per well to a six-well plate (Fig.
1a). The number of wells to be filled depends on the number
of samples to be processed. Each well is sufficient for several
(3 or 4) tissue sections. Place brain tissue samples into sepa-
rate wells of the plate using a fine, soft-bristled paintbrush (see
Note 5). Replace the plate cover and agitate gently (about
200 rpm) on a platform shaker for 10 min. Remove the cover
and then remove gently the PBS with a plastic Pasteur pipette
(connecting a yellow pipette tip to the Pasteur pipette may
help to avoid the suction of the delicate brain sections) and
discard (Fig. 1b). Repeat this process twice for three total
washes (see Note 6).
2. After removing the PBS from the last wash cycle, add 3 mL
heated citrate buffer (60–70 °C) per well to the plate (depend-
ing on the primary antibodies, more severe retrieval can be
achieved at 80 °C). Place on the platform shaker, agitate gently
(0.1 × g) and leave for 20 min. Remove citrate buffer from
each well and discard.
3. Add 3 mL PBS per well, agitate gently for 5 min, and then
remove and discard.
4. Add 3 mL TBS + 0.5% Triton® X-100 per well, agitate gently
for 5 min, and then remove and discard.
a b
c d
e f
Fig. 1 Handling and mounting floating brain sections. (a) Washing gently, placing the solutions in the six-well
plate with a Pasteur pipette. (b) Discarding the solutions carefully with a Pasteur pipette with a small tip. (c)
Labeling slides on frosted glass area. (d) Mounting floating sections on microscope slide. (e) Placing mounting
media solution on the coverslip. (f) Covering slides facing down
3.2 Blocking The next series of steps uses horse serum at different concentra-
tions in TBS-0.5% Triton® X-100 to avoid nonspecific antibody
binding:
1. Add 3 mL blocking solution 1 (10% horse serum) per well with
a plastic Pasteur pipette; place the plate on the platform shaker
and leave for 60 min. Remove blocking solution 1 and discard

appropriately.
2. Add 3 mL blocking solution 2 (1% horse serum) per well, place
the plate on the platform shaker, and leave for 5 min. Remove
blocking solution 2 and discard appropriately.
3. Add 3 mL TBS + 0.5% Triton® X-100 per well, agitate gently
for 5 min, and then remove and discard after the samples have
been transferred to individual glass vials in the following step.
3.3 Primary 1. Add 1 mL primary antibody solution (diluted in 1% HS) at the

and Secondary appropriate concentration (see Notes 7 and 8) per glass vial.
Antibody Incubation Once the sample is transferred to the vial, seal with the plastic
cap, place vial on the platform shaker, repeat for all samples
and leave for 48 h.
2. Remove primary antibody and add 3 mL TBS + 0.5% Triton®
X-100 per vial, agitate gently on the platform shaker for 10
min, and then remove and discard. Repeat twice for three total
washes.
3. Add 1 mL secondary antibody solution (see Notes 9 and 10)
per vial, seal with the plastic cap, place the vial on the platform
shaker, and leave for 4 h. From here on, impede light exposure
by means of aluminum foil as explained in Note 9.
3.4 Nuclei 1. Remove secondary antibody and add 3 mL PBS per vial, agi-
Counterstain tate gently on the platform shaker for 10 min, and then remove
and discard. Repeat twice for three total washes.
2. Add 1 mL DAPI solution at 1:1000 in PBS (see Note 11) per
vial, seal with the plastic cap, place vial on the platform shaker,
and leave for 30 min.
3. Remove DAPI solution from the vial, add 3 mL PBS per vial,
agitate gently on the platform shaker for 10 min, and then
remove and discard. Repeat twice for three total washes but
leave the samples immersed in PBS after the last wash until
they are removed to be mounted on slides in the next step.
Then discard the PBS.
3.5 Mounting 1. While washing, select pre-cleaned glass slides for mounting
and indicate the sample labeling with pencil in the designated
frosted glass area (Fig. 1c).
2. Partially fill a Petri dish with PBS and transfer the sample from
the vial to the dish with a fine, soft-bristled paintbrush. The
PBS left in the six-well plate can now be discarded. Carefully
place the sample on the previously labeled slide using the fine-
tipped brush (Fig. 1d). Leave to dry for 30–40 min, protected
from light. Repeat for all samples.
3. When slides are dry, mount coverslips using antifade reagent

(i.e., Prolong® Gold or similar products) (see Note 12)
(Fig. 1e and f). Leave to dry overnight in the dark.
3.6 Confocal 1. Proceed to scan the samples with confocal microscope (see
Microscope Scanning Note 13). For IS imaging, examine brain sections thoroughly
and 3D Rendering to seek T cell-astrocyte interactions. Confocal scanning of the
region of interest (ROI) should be performed with a 40× or
63× objective and preferably with immersion oil (see Note 14).
2. Once the ROI is detected, proceed to scan, setting up the suf-
ficient z distance to grab both interacting cells within the stack.
Both cells may appear complete within the 3D box (see Note
15). In this tissue block, we may be able to differentiate cells
that are apart (Fig. 2a and b) or cells that are in apposition
(Fig. 2c and d). Orthogonal views may help to confirm the
a Non contacting cells c Contacting cells

x x
y y
0.5 mm layer 0.5 mm layer

z z
b x z d x z
y y
z z
Fig. 2 Graphic representation of confocal three-dimensional stack of images at the ROI. The blue-lined boxes
represent the scanned volume at the x–y plane through the z axis with 0.5-μm optical layers (red line). Two
cellular elements (gray and red spheres) are represented within the stack. Theoretical examples of noncon-
tacting and contacting cells are represented. Orthogonal views of the x–y central optical layer from a and c are
represented in b and d, respectively. Lateral views along the z axis at the level of the crossing planes (black
broken lines) are depicted at the right and bottom of the x–y optical plane. Noncontacting cells represented in
three-dimensional space (a) can be visualized apart in z planes in the orthogonal view (b). Similarly, contacts
between the cells, represented in 3D (c), can be verified at the z planes with the orthogonal view
Fig. 3 Mature T cell-astrocyte immunological synapse. For optimal SMAC detection, immunofluorescence
with primary antibodies against LFA-1 and TCR or CD3 is recommended. Target astrocyte in this case is
detected by antibodies against thymidine kinase (TK), which indicates the transgene expression of experi-
mentally induced viral infection. Top panel shows the following: (1) DAPI nuclear staining (blue), arrow indi-
cating the nuclear polarized notch; (2) LFA-1 (red) revealing a clear flat interface and the arrow points to the
central area of LFA-1 absence; (3) TCR (green) clearly polarized toward the interface, especially to the nucleus
notched area; (4) an overlap of images 2 and 3; (5) (merge 1) merge of images 1, 2, and 3; (6) TK expression
revealing the viral infected astrocyte; and (7) (merge 2) merge of all the channels (blue, red, and green). In
the bottom panel, image a shows a low-power capture of the T cell-astrocyte immunological synapse in rat
brain. Image b depicts a zoom of the image in a indicating the interface plane (yellow broken line) as well as
the view angle of the 3D reconstruction at the interface-clipping plane shown in c and d. Barcia et al. 2006.
Originally published in The Journal of Experimental Medicine. doi:10.1084/jem.20060420 © 2006 Rockefeller
University Press
contact and interactions of the cells. Depending on the desired

resolution and the scanning speed, the dpi and z-section may
be increased at will. However, increasing zoom and resolution
can compromise the strength of the fluorescence signal. A
careful balance between the two factors should be pondered.
An example of mature effector ISs is shown in Fig. 3.
3. Once the confocal scanning is performed, process the image
stack files generated with specific software for 3D visualization
(see Note 16) and proceed to visualize the astrocyte-T cell
interface with a clipping plane (see Note 17).
4 Notes
1. Use diluted HCl or NaOH initially to achieve the required pH.

2. Use very gentle agitation and add Triton® X-100 slowly and
carefully to avoid creating bubbles. The same applies when
topping up solutions containing Triton® X-100. Pour the liq-
uid along the sides of the container.
3. Serum from other species may be used in this step as long as

there is no cross-reaction with the primary and secondary anti-
bodies in the following steps. Horse serum is recommended
since primary and secondary antibodies are rarely made in
horse.
4. Antibodies against vimentin are commonly used as markers of
astrocytic cells or glioma cells; however, vimentin antibodies
also label microglia/macrophages, and the results may be mis-
leading in some aspects.
5. Be extremely careful in handling samples. Move the tip of the
brush slowly in the solution, and twist the brush gently to
dislodge samples if they adhere to the brush. When pipetting
solutions into wells containing samples, aim the flow toward
the side of the well and not directly onto the sample itself.
Remove liquids from the wells very carefully to avoid damag-
ing the delicate samples or sucking them into the pipette.
6. The time during washes can be used to prepare the citrate buf-
fer for the next step. Transfer approximately 20 mL of citrate
buffer to a 50-mL glass Erlenmeyer flask with a small magnetic
stir bar, and place the flask on a heated platform shaker/stirrer.
Insert a thermometer into the flask without it touching the
bottom nor the sides of the flask for a more precise measure-
ment; and wrap a small piece of aluminum foil over the top of
the flask and around the thermometer to prevent evaporation
as the citrate buffer is heated. Adjust the heated platform
shaker/stirrer controls to 60 °C and about 200 rpm (some
particular antigens may need up to 80 °C to be retrieved).
Leave the buffer to heat while washing the samples in PBS.
7. Incubations in primary and secondary antibody solutions are
preferably done in small, plastic-capped glass vials to avoid evap-
oration of the solution. After preparing the primary antibody
solution, transfer 1 mL to each vial. Transfer the sample carefully
from the plate to the vial using a soft-bristled paintbrush.
8. Prepare the primary antibodies in blocking solution 2 (1%
HS). Approximately 1–1.5 mL per sample is sufficient, consid-
ering the volume of the vial and always ensuring that the tissue
sections are completely covered with solution and that no
samples are stuck to the sides of the vial. The optimal dilution
of the antibodies should be tested prior to final experiments;
the commercial technical sheet recommendation could be a
starting point. Antibodies are normally stored frozen in small
aliquots in microcentrifuge tubes; a sufficient amount must be
removed from the freezer and allowed to thaw before prepara-
tion. As an example for a 1:200 concentration of a single pri-
mary antibody for six different samples in a six-well plate, add
5970 μL TBS-0.5% Triton® X-100 containing 1% horse serum
to a 50-mL tube, measure 30 μL of antibody with a suitable
autopipette, and transfer to the tube. Vortex briefly before

adding 1 mL to each sample. If multiple primary antibodies
are intended to use, mix them accordingly with each desired
concentration.
9. All steps starting with the secondary antibody incubation until
the end of the protocol should be done as much as possible
protected from light. The small glass vials used to incubate
the tissue samples with secondary antibodies should be
wrapped with aluminum foil prior to adding antibody solu-
tions and samples.
10. Prepare the secondary antibodies in blocking solution 2.

Approximately 1–1.5 mL per sample is sufficient to cover the
sections. The secondary antibodies should be chosen accord-
ing to the excitation wavelength of the available lasers of the
confocal microscope, generally 405, 488, 594 or 555, and
633 nm. Usually the appropriate working dilution for second-
ary antibodies is 1:1000 (1 μL/mL). However, each user
should establish the optimal concentration according with the
required level of signal. Take sufficient amount of appropriate
secondary antibodies (thawing if it is stored at −20 °C); spin
the microcentrifuge tubes briefly before pipetting to avoid
wasting the liquid in the cap. Following the example for six
samples in the six-well plate, for a single secondary antibody,
add 5994 μL TBS-0.5% Triton® X-100 containing 1% horse
serum to a 50-mL tube, measure 6 μL of the appropriate sec-
ondary antibody with a suitable autopipette, and transfer to
the tube. Vortex briefly before adding to samples.
11. For six samples, put 5994 μL PBS to foil-wrapped 50-mL
tube, add 6 μL of DAPI with a suitable autopipette, and vor-
tex briefly to avoid precipitates before adding to samples. Use
1 mL per vial.
12. Place the coverslip on the work surface and apply antifade
reagent (we recommend Prolong® Gold antifade reagent) in a
T-shaped trace on the coverslip with the horizontal line on the
longest and furthest edge and the vertical line toward the
nearest edge along the midline of the coverslip (Fig. 1e).
Gently lower the slide onto the coverslip with the sample side
facing down (Fig. 1f). This practice prevents the formation of
bubbles. Let the liquid spread evenly as the coverslip settles
into place; if necessary, daub the edges of the slide gently to
remove excess media, and then stand the slides in a container
that can be covered to protect them from light while drying.
13. We recommend using a confocal microscope able to visualize
the spectrum of wavelength ranging from blue to infrared.
This feature allows the proper separation of multiple emission
channels and the correct visualization of cellular structures.
Commercially available confocal microscopes usually come
with a range including 405, 488, 555/594 nm, and

633/639 nm wavelength lasers.
14. Searching for a suitable ROI is critical to achieve the desired
results. To facilitate the interaction search, it is advisable to
keep the markers of the two interacting cell types in the visible
channels (i.e., TCR/CD3 for T cells in the green channel and
GFAP for astrocytes in the red channel). Non-visible channels,
such as infrared (i.e., marked with secondary antibodies labeled
with fluorochromes with excitation wavelengths in the far red
such as 633, 639, 647, or Cy5), should be left for other mol-
ecules of interest. If quantification of intercellular engage-
ments is planned, random and systematic fields should be
scanned throughout the sampling area.
15. A crucial step is the proper confocal microscopy scanning of
the sections and setting up the appropriate z resolution range
to identify the interactions. For good resolution, scanning
1024 × 1024 pixels per inch in the x–y optical plane may be
sufficient. The scanning of the z optical section should be
done at least every 0.5 μm. The microscope then provides a
stack of images that represents a box with x–y–z dimensions.
16. In our experience, we have previously used IllucidaFX (Illucida
LLC, Los Angeles) or Imaris (Bitplane, Zurich) software. 3D
reconstructions of the images are vital to study the T cell-
astrocyte interfaces. Since the IS in the brain tissue may appear
oriented in different directions and at different angles, the
software should allow (1) the free rotation of the image stack
and (2) the visualization of the interface at a preferred clipping
plane (Fig. 4). First, opening the images with the software
should allow searching for the specific z optical plane at the
level of the synapse. If necessary, a specific volume of the stacks
can be cropped. Software like Image J (National Institutes of
Health, Bethesda, Maryland) allows this feature.
17. When opening the 3D view in the rendering software, the user
should clearly visualize the two interacting cells (for this spe-
cific protocol, T cell and astrocyte). In particular, bona fide
mature ISs are more easily found in body-to-body engage-
ments than in body-to-filament interactions. Once the interac-
tion is clearly detected (see graphical example in Fig. 4),
rotation of the image stack should be checked. A free clipping
plane should then be placed at the level of the interface. To do
that, the software should allow the free user movement of the
plane through the multiple directions of the 3D stack (Figs. 4
and 5). Next, the clipping plane should be positioned in the
right orientation, specifically cutting the T cell-astrocyte inter-
face (Figs. 3, 4, and 5). Best results are achieved when inter-
faces are flat, which is representative of a mature IS [9, 10].
SMACs should then be visible at the clipping plane of the
interface (Figs. 3, 4, and 5). Importantly, labeling the cell
a
b
3D view
c d
Fig. 4 Graphic representation of SMAC 3D visualization. (a) Diagram of the representative 3D view of a stack
of confocal images showing a T cell (red) engaging with an astrocyte (white) within the parenchyma. (b) Free
clipping plane placed at the T cell-astrocyte interface (indicated with orange broken line). (c) Visualization of
c-SMAC (green) and p-SMAC (red) in the plane positioned at the interface. (d) Removal of the information in
front of the clipping plane to visualize the SMACs
nuclei can give an anatomical reference outside the limit of the

T cell-astrocyte SMAC-forming interface (Fig. 5). The
interface-clipping plane should not have nucleus fluorescence
labeling. This reference avoids false depictions of p-SMACs, as
the cell membrane borders of the cell, labeled, for example,
with LFA-1, may provide similar ring shapes (Fig. 5). The
software should have a channel tool to visualize fluorescence
channels independently, to identify the microanatomy of the
clusters without other channel information that may give rise
to misinterpretations of the interface structures. The final vol-
ume of the image stack can be modified as a transparency,
eliminating the black background of the images. Usually, ren-
dering tools also allow the implementation of shadows to
make the 3D more understandable and realistic or to change
fluorescence transparencies to visualize underlying markers
and structures (Fig. 6). With the same procedure, other effec-
tor molecules, such as IFN-γ, can be detected at the interface.
In this case, it is important to notice that this central cluster
appears smaller than the TCR c-SMAC (Fig. 6).
Fig. 5 Visualizations of clipping planes of 3D rendering software. a shows three clipping planes (1, 2, and 3
yellow broken lines). Images 1, 2, and 3 show the perpendicular view at these particular three planes, respec-
tively, image 1 being the interface containing SMAC. Image b shows analogous three clipping planes (4, 5, and
6 yellow broken lines). Images 4, 5, and 6 show the perpendicular view at their respective clipping planes.
Boxes on the left represent the spatial orientation of the clipping planes. Barcia et al. 2006. Originally published
in The Journal of Experimental Medicine. doi:10.1084/jem.20060420 © 2006 Rockefeller University Press
Fig. 6 Imaging CTL Kupfer-type T cell-astrocyte immunological synapse in rat brain. In this particular
case, a triple immunostaining in thick sections of fixed brain tissue was performed. Antibodies against
viraly-infected astrocyte (white) expressing TK, LFA-1 (red), and IFN-γ (green) were used, in addition to
DAPI (blue) as a nuclear counterstain. In a, we show an optical plane at the center of the image stack
of a T cell-astrocyte IS. In b, a 3D rendering with shadows and transparencies is depicted. In c, we show
low-power and general 3D view of a T cell-astrocyte IS in the brain performed with confocal microscope
(63× oil objective). A zoom of the synapse represented in a was scanned with higher magnification and
shown in panels d–h in a 0.5-μm optical slide. In d, the nuclei of the two engaged cells (T cell and
astrocyte) are shown. In e, the cluster of IFN-γ is detected at the p-SMAC, which is surrounded by the
LFA-1-rich p-SMAC (red) shown in f. Image g demonstrates the expression of TK in the infected astro-
cyte. In image h a merge of the channels d–g is shown to demonstrate the contact site. Image i shows
the merge of green and red (IFN-γ and LFA-1, respectively) channels, and the white line represents the
tool to measure the fluorescence profile represented in graph j where the maximum peaks of LFA-1 red
fluorescence are detected at the p-SMAC and the maximum peak of IFN-γ green fluorescence is
detected at the c-SMAC region. Next, 3D reconstructions at the interface were made with IllucidaFX
software. Image k shows the view of the interface. White broken lines and arrow represent the view
angle of the interface view at the clipping plane (yellow broken arrow). In image l, the IFN-γ central
cluster at the interface at the clipping plane level is seen at the T cell-astrocyte IS, and LFA-1 peripheral
cluster is seen in m. The merge of the planes is depicted in n. Scale bar in f equals 10 μm. From Barcia
et al., The Journal of Immunology, vol. 180, pp. 1344–1352, 2008. Copyright 2008. The American
Association of Immunologists, Inc
References
1. Grakoui A, Bromley SK, Sumen C, Davis MM, 6. Barcia C Jr, Gomez A, Gallego-Sanchez JM,
Shaw AS, Allen PM, Dustin ML (1999) The Perez-Valles A, Castro MG, Lowenstein PR,
immunological synapse: a molecular machine Barcia C Sr, Herrero MT (2009) Infiltrating
controlling T cell activation. Science CTLs in human glioblastoma establish immu-
285:221–227 nological synapses with tumorigenic cells. Am
2. Davis DM, Dustin ML (2004) What is the J Pathol 175:786–798
importance of the immunological synapse? 7. Yang J, Sanderson NS, Wawrowsky K, Puntel
Trends Immunol 25:323–327 M, Castro MG, Lowenstein PR (2010) Kupfer-
3. Barcia C, Thomas CE, Curtin JF, King GD, type immunological synapse characteristics do
Wawrowsky K, Candolfi M, Xiong WD, Liu C, not predict anti-brain tumor cytolytic T-cell
Kroeger K, Boyer O, Kupiec-Weglinski J, function in vivo. Proc Natl Acad Sci U S A
Klatzmann D, Castro MG, Lowenstein PR 107:4716–4721
(2006) In vivo mature immunological synapses 8. Molofsky AV, Krencik R, Ullian EM, Tsai HH,
forming SMACs mediate clearance of virally Deneen B, Richardson WD, Barres BA,
infected astrocytes from the brain. J Exp Med Rowitch DH (2012) Astrocytes and disease: a
203:2095–2107 neurodevelopmental perspective. Genes Dev
4. Mitxitorena I, Saavedra E, Barcia C (2015) 26:891–907
Kupfer-type immunological synapses in vivo: 9. Richie LI, Ebert PJ, Wu LC, Krummel MF,
Raison D’etre of SMAC. Immunol Cell Biol Owen JJ, Davis MM (2002) Imaging synapse
93:51–56 formation during thymocyte selection: inability
5. Barcia C, Wawrowsky K, Barrett RJ, Liu C, of CD3zeta to form a stable central accumula-
Castro MG, Lowenstein PR (2008) In vivo tion during negative selection. Immunity
polarization of IFN-gamma at Kupfer and non- 16:595–606
Kupfer immunological synapses during the 10. Huppa JB, Davis MM (2003) T-cell-antigen
clearance of virally infected brain cells. recognition and the immunological synapse.
J Immunol 180:1344–1352 Nat Rev Immunol 3:973–983
Chapter 33
Aberrant Immunological Synapses Driven by Leukemic

Antigen-Presenting Cells
Fabienne McClanahan Lucas and John G. Gribben
Abstract
Aberrant immune synapse formation between antigen-presenting and immune effector cells is a central
mediator of immune dysfunction and can be observed across several haematologic malignancies. Here, we
describe the cell preparation, conjugation and immune synapse quantification of B and T cells obtained
from patients with leukaemia and the adaptions required when using cells from murine models of
disease.
Key words Immune synapse, B cells, T cells, NK cells, Leukaemia, Lymphoma, Eμ-TCL1 model
1 Introduction
Immune dysfunction is a major hallmark of haematological malig-

nancies, and several quantitative and qualitative immune defects
have been described [1]. Using chronic lymphocytic leukaemia
(CLL) as a model cancer to investigate the underlying molecular
mechanisms of T-cell dysfunction, we initially demonstrated that
both CD4+ and CD8+ T cells from patients with CLL show mul-
tiple differentially expressed genes when compared to these cells
from age-matched healthy donors [2]. Deregulated genes were
enriched in pathways of T-cell proliferation, differentiation, vesicle
trafficking and actin cytoskeleton formation and included cdc42,
PIK4CB, RAB35 and ARPC1B, all of which are key regulators of
the formation and stabilization of the immune synapse. Very simi-
lar alterations in gene expression could be induced in normal allo-
geneic T cells after direct contact with CLL cells in co-culture,
indicating that the presence of the malignant CLL cell drives the
observed T-cell changes. Subsequently, we demonstrated the inap-
propriate functional response of T cells to antigen-presenting cells
(APCs) due to their inability to effectively regulate actin remodel-
ling and to recruit key cytoskeletal signalling molecules such as
533
534 Fabienne McClanahan Lucas and John G. Gribben
Lck, Cdc42, WASp, filamin-A and dynamin-2, using immune syn-

apse formation assays and confocal microscopy [3].
Consistent with our gene expression studies, these defects
could be induced in normal allogeneic T cells by co-culturing them
in direct cell-cell contact with CLL cells. Impaired actin remodel-
ling was then found to be mediated by activation of the Rho
GTPases RhoA, CdC42 and Rac1 [4], which also led to impaired
LFA-1-mediated T-cell adhesion and migration [5]. Similar
immune synapse formation defects have been observed in other
haematologic malignancies such as diffuse large B-cell lymphoma
and follicular lymphoma [4], as well as in acute leukaemias [6].
Moreover, this is also recapitulated in the Eμ-TCL1 mouse model
of CLL, one of the most established preclinical disease models
using immune competent mice [3, 7]. Importantly, defective
immune synapse has been demonstrated to be an attractive thera-
peutic target that can be repaired by the immunomodulatory drug
lenalidomide in vivo and in vitro [3–5, 8] and by immune check-
point blockade using PD-L1/PD-1 antibodies [4, 9].
This chapter describes the methods used for cell preparation,
conjugation and immune synapse quantification of B and T cells
obtained from patients with leukaemia and the adaptions required
when using cells from the Eμ-TCL1 mouse model. Although the
majority of work has been done using T cells as effector cells, simi-
lar results have been described for natural killer (NK) cells [10]. To
test whether impaired immune synapse formation is a result of
poor APC function of B cells, of impaired T-effector cell function
or of a combination of both, autologous and allogeneic conjuga-
tion experiments can be performed. Culturing malignant B cells
with healthy allogeneic T cells induces immunological synapse
defects that are very similar to the defects in autologous B and T
cells from cancer patients. Using high numbers of malignant B
cells, these defects can be induced by direct cell-cell co-culture in
previously healthy allogeneic T cells within 24 h, with maximal
induction being seen by 48 h independent of the malignant B-cell
number added. After such co-culture, T cells can be functionally
examined in conjugation assays with third-party healthy allogeneic
B cells pulsed with sAg as APCs following the same protocol.
2 Materials
Pre-warm all media to 37 °C. Always protect fluorescent dyes and

antibodies from light by keeping them covered with aluminium
foil, by wrapping tubes containing such reagents in aluminium foil
and by working out of direct light. Once cells are labelled with
fluorescent dye, protect racks/slides from light. All steps are per-
formed at room temperature. Reagents for fixing cells and
immunofluorescent labelling should be stored at −20 °C in appro-
Aberrant Immunological Synapses in Leukaemias/Lymphomas 535
priately sized aliquots. Although it is possible to freeze/thaw them

several times, single freeze/thaw cycles should be preferred.
2.1 General 1. Water bath set to 37 °C.

Laboratory Equipment 2. Benchtop centrifuge (refrigeration not required).
3. General purpose centrifuge fitted with swinging bucket rotor.
4. Vacuum aspiration system.
5. Manual haemocytometer or automated haemocytometer with
trypan blue dye exclusion method (e.g. Beckman Coulter Vi-
CELL series or equivalent). See Note 1 on acceptable cell
viabilities.
6. 1.5 ml polypropylene microfuge tubes.
7. 15 and 50 ml Falcon tubes.
8. CO2 incubator.
2.2 Buffers 1. Full medium: RPMI-1640 medium, 10% fetal calf serum
and Media (FCS), 1% penicillin/streptomycin. If working with mouse
cells, also add 50 μM β-mercaptoethanol. See Note 2 on sterile
versus non-sterile conditions.
2. Serum-free medium: RPMI-1640 medium, 1% penicillin/
streptomycin.
4. Fixative: 3.2% formaldehyde (methanol-free) in PBS. Make by
adding 10 ml Pierce™ 16% formaldehyde (w/v), methanol-
free, to 40 ml PBS. Aliquot and store at −20 °C.
5. Permeabilization: 0.3% Triton in PBS. Make by adding 150 μl
100× Triton to 50 ml PBS (see Note 3 on dissolving 100×
Triton). Aliquot and store at 4 °C.
6. Blocking solution: 0.1% BSA in PBS. Make by adding 1 ml
10× BSA to 99 ml PBS. Aliquot and store at −20 °C.
7. Goat serum buffer: 5% goat serum in PBS. Aliquot and store at
−20 °C.
8. Superantigen (sAg) cocktail consisting of staphylococcal
enterotoxin B from Staphylococcus aureus (SEB) and staphylo-
coccal enterotoxin A from Staphylococcus aureus (SEA). Make
by dissolving powder in sterile H2O for individual SEB and
SEA stocks at 4 mg/ml. Then mix together to obtain cocktail
stock at 2 mg/ml. Aliquot and store at −20 °C (see Note 4 on
safe handling of sAg).
2.3 Cell Separation 1. Cell separation kits to obtain purified CD19+ B cells and
Components CD3+/CD4+/CD8+ T cells, e.g. CD19/CD4/CD8 micro-
beads and/or pan T-cell/CD4/CD8 isolation kits for mag-
netic activated cell sorting (MACS®) from Miltenyi Biotec or
column-less separation kits from other providers. See Notes 5

and 6 on selection of purification strategies and confirmation
of purity.
2. Cell separation buffers, columns and accessories as indicated in
protocol of selected cell purification product. See Note 7 on
manual versus automated cell separation.
2.4 Slides 1. Cytocentrifuge (e.g. StatSpin® CytoFuge 2 or equivalent).

and Accessories 2. Slide chamber units: one unit consist of one-well or three-well
cell concentrator (disposable or autoclavable), matching sili-
cone gaskets, two clips, one backing plate and one poly-l-
coated glass slide. See Note 8 on choice of cell
lysine-
concentrator, Note 9 on assembling a slide chamber unit and
Note 10 on selection of slides.
3. 22 × 32 mm cover glasses.
2.5 Antibodies 1. CellTracker™ Blue CMAC (7-amino-4-chloromethylcoumarin)

and Conjugates fluorescent dye (excitation/emission maxima 353/466 nm):
make 10 mM stock by adding 2385 μl DMSO to room tem-
perature CMAC lyophilized product. Keep in aliquots of about
10 μl and store at −20 °C wrapped in aluminium foil, always
protect from light.
2. Primary antibodies, e.g. Granzyme B, LAMP-1, signalling
molecules, if required, at concentrations/dilutions optimized
for specific experiments.
3. Secondary antibodies, if required.
4. Rhodamine phalloidin to stain F-actin.
5. Mounting medium, such as DAKO fluorescence mounting
medium.
2.6 Microscopy 1. Confocal laser scanning microscope fitted with 63× objective
and Image Processing and Diode 405 (to excite CMAC), Helium Neon 543 (to
excite rhodamine phalloidin) and additional tuneable Argon/
Helium Neon 633 lasers to detect fluorophores selected for
specific experiments.
2. Digital image processing software, such as AxioVision Rel48
image analysis software (Zeiss).
3 Methods
3.1 Purification 1. B and T cells are used at a 1:1 ratio in cell conjugation assays.
of B- and T-Cell Determine total number of cells required for experimental and
Subsets control conditions and account for pipetting errors by adding
two to three additional conditions to calculations. See Note 8
for required cell numbers in one- and three-well concentrator

chambers, and account for cell loss during lymphocyte subset
purification steps. We strongly recommend optimization
experiments to determine cell yield and purity after cell purifi-
cation. See Note 5 on choice of cell separation kits and Note 6
for determination of cell purity.
2. See Note 11 on differences in the handling of human PBMCs
and mouse whole spleen single suspensions, Note 1 on required
cell viabilities and Note 12 on using fresh versus frozen cells.
3. If starting from frozen cells: retrieve cells from liquid nitrogen
storage, thaw in water bath set to 37 °C, slowly add into 10 ml
pre-warmed full medium, and centrifuge at 300 × g for 10 min
at room temperature (if cells are prone to sticking together,
DNase can be added to cells while still in freeze mix). Discard
supernatant and resuspend in 5 or 10 ml full medium (volume
depending on pellet size), count cells and determine viability.
4. If starting from fresh cells or whole mouse organs, follow insti-
tutional protocols and procedures to obtain PBMCs and sin-
gle-cell suspensions.
5. Purify B cells using CD19 microbeads following the recom-
mendations of the supplier. Briefly, for manual separation, cen-
trifuge cells at 300 × g for 10 min at 4 °C. Resuspend in MACS
buffer 90 μl/107 cells. Add 10 μl CD19 microbeads/107 cells.
Mix and incubate at 4 °C. We have found that increasing the
incubation time to 15 min, with gentle shaking or vortexing
the tube every 5 min to ensure homogenous labelling of cells,
results in consistently high purities. Wash with 1–2 ml MACS
buffer/107 cells and centrifuge at 300 × g for 10 min at
4 °C. Discard supernatant and resuspend up to 108 cells in
500 μl MACS buffer.
6. Fit column into magnet and prepare column by rinsing with
3 ml MACS buffer as per manufacturer’s recommendation.
Also refer to the manufacturer’s recommendation for choice of
column size (guided by cell numbers used in experiment).
7. Prepare collection tubes for negative and positive cell fractions:
label tubes clearly, including sample name and fraction. Place
tubes for negative fraction into cooling racks or embed in ice
under column, and add 200 μl MACS buffer.
8. Apply cell suspension to the column, wash three times with 3 ml
MACS buffer each time. When doing the first wash, add the
3 ml MACS buffer to the tube that has been used for magnetic
labelling of cells, wash tube, and apply onto column. Always
wait until the column runs dry before adding more wash.
9. The collected effluent is the negative fraction and contains
non-B cells. These will be further purified by T-cell isolation
kits in autologous conjugation assays. Refer to the manufac-

turer’s recommendation for volumes, incubation times and
required columns.
10. Steady tube labelled with positive fraction in a rack, add 5 ml
MACS buffer to column, immediately remove column from
magnet and insert plunger provided with each column to elute
the CD19 positively labelled B cells. Mix well and count cells.
Take aliquots containing cell numbers required for the experi-
ment, centrifuge at 300 × g for 10 min at 4 °C and resuspend
in full medium. Rest cells in incubator.
11. Proceed to T-cell purification from the CD19 microbeads neg-
ative fraction (see step 9). If using CD3, CD4 or CD8 micro-
beads, cells of interest will be magnetically labelled and have to
be eluded as described for B cells. If using negative selection,
unlabelled cells of interest will be contained in the effluent. Mix
well and count cells. Take aliquots containing cell numbers
required for the experiment, centrifuge at 300 × g for 10 min at
4 °C and resuspend in full medium. Rest cells in incubator.
3.2 Preparation 1. Retrieve purified B cells from incubator and centrifuge at

of Purified B Cells 300 × g for 10 min at room temperature.
2. Discard supernatant, wash cells in 10 ml serum-free medium,
and centrifuge at 300 × g for 10 min at room temperature.
3. While cells are spinning, prepare CMAC dye in serum-free
medium: in tube wrapped in aluminium foil, add 500 μl
serum-free medium containing 1 μl CMAC for each condi-
tion plus one more to account for pipetting errors. Mix well
and keep in the dark until ready to use.
4. Discard supernatant, and resuspend B cells in serum-free
medium at a concentration of 1 × 106 cells per 500 μl (if using
three-well concentrator. If one-well concentrator is used,
resuspend 4 × 106 cells per 500 μl). Mix well and transfer
500 μl into individual 1.5 ml polypropylene microfuge tubes
(one tube per experimental condition required). Make sure
cells are in suspension, then add 500 μl of the prepared CMAC
dye in s erum-free medium while gently vortexing the cells for
a final CMAC concentration of 10 μM/ml. See Note 13 for
CMAC labelling of cells. Cover tightly with aluminium foil
and rest in incubator at 5% CO2 at 37 °C for 30 min.
5. Centrifuge CMAC-labelled B cells in benchtop centrifuge at
1500 × g for 1 min. Align tubes carefully to ensure that pellet
is in the same position for all tubes. Aspirate supernatant using
the aspirator unit. Always check cell pellet size before and after
aspiration.
6. Prepare sAg in full medium: for each condition (plus one
more to account for pipetting errors) mix 1 ml full medium
with 1 μl 2 mg/ml sAg cocktail. Handle with great care and
mix well. See Note 14 on importance of sAg stimulation.
7. Resuspend B-cell pellet in 1 ml of the prepared sAg in full
medium, and mix well. Cover tightly with aluminium foil and
rest in incubator at 5% CO2 at 37 °C for 30 min. Using full
medium will ensure that any residual CMAC binds nonspecifi-
cally to constituents in the serum.
3.3 B-Cell/T-Cell 1. While B cells are incubating: thaw/prepare/bring to room tem-

Conjugation perature reagents required for immunofluorescent labelling, i.e.
fixative, Triton, blocking solution and goat serum buffer.
2. Retrieve T cells from incubator, centrifuge at 300 × g for 10 min
at room temperature. Discard supernatant, and resuspend T
cells in full medium at a concentration of 1 × 106 cells per 75 μl
(if using three-well concentrator. If one-well concentrator is
used, resuspend 4 × 106 cells per 200 μl). Mix well and transfer
75 μl (or 200 μl) into individual 1.5 ml polypropylene microfuge
tubes (one tube per experimental condition required). It is
essential that this is completed with the end of step 3.
3. Retrieve B cells from incubator, centrifuge at 1500 × g for
1 min at room temperature, and then carefully aspirate super-
natant. It is essential that T cells (see step 2) and CMAC
labelled, stimulated B cells are ready for the next step at the
same time.
4. Resuspend B cells in 75 μl full medium (if using three-well
concentrator. If one-well concentrator is used, resuspend
4 × 106 cells per 200 μl).
5. Mix T cell and B cells. For three-well concentrator, combine
75 μl B cells with 75 μl T cells (150 μl total cell mix); for one-
well concentrator, combine 200 μl B cells with 200 μl T cells
(400 μl total cell mix) by taking T cells into 200 μl/1 ml Gilson
pipette (set to 200 μl/500 μl) and transfer to B cell tubes. If
several conditions/patients are being tested, it is essential that
correctly matched B and T cells are combined.
6. Centrifuge tube containing the combined B and T cells at
300 × g for 5 min to allow a cell pellet to form. Incubate cell
pellet at 5% CO2 at 37 °C for 20 min wrapped in aluminium
foil (see Note 15 on incubation times).
7. Prepare slide chamber units so that they are ready for rapid
transfer of cells post conjugation (see Note 9).
8. Resuspend cells very gently to disperse the pellet using a Gilson
pipette p200 set to 200 μl, then transfer to slide chamber unit
(see Note 16 on importance of consistency).
9. Centrifuge slide chamber units at 1000 rpm for 6 min in
StatSpin Cytofuge 2 or equivalent cytocentrifuge.
10. Remove slide chamber units from cytocentrifuge and tilt on

their sides, and carefully aspirate liquid. See Note 17 on han-
dling large numbers of slide chamber units.
3.4 Immuno 1. Use all buffers and reagents at room temperature.

fluorescent Labelling 2. Fix cells for 15 min by adding 200 μl 3.2% formaldehyde
(methanol free)/PBS per chamber (if using three-well concen-
trator. If one-well concentrator is used, add 500 μl). Keep cov-
ered with aluminium foil.
3. Aspirate fixative and wash three times by adding 200 μl (or
500 μl) PBS. Soaking of cells between washes is not required
for any of the washes.
4. Permeabilize cells with 200 μl (or 500 μl) 0.3% Triton in PBS
for 5 min and keep covered with aluminium foil.
5. Aspirate 0.3% Triton and wash three times by adding 200 μl
(or 500 μl) PBS.
6. Block for 10 min by adding 200 μl (or 500 μl) 0.1% BSA/PBS
solution, and keep covered with aluminium foil.
7. If using rhodamine phalloidin only, omit steps 8–10.
8. If using additional antibodies: aspirate blocking solution and
wash three times by adding 200 μl (or 500 μl) PBS. Apply 100 μl
(or 250 μl) primary antibodies per well at previously optimized
concentrations in goat serum buffer and incubate for the
required times at 4 °C, and keep covered with aluminium foil.
9. Aspirate primary antibody and wash three times by adding
200 μl (or 500 μl) PBS.
10. Apply 100 μl (or 250 μl) secondary antibodies per well at pre-
viously optimized concentrations in goat serum buffer and
incubate for the required times at 4 °C. At the same time, add
rhodamine phalloidin at 1:40 dilution to the same secondary
antibody/goat serum buffer mix. Keep covered with alumin-
ium foil.
11. If using rhodamine phalloidin alone, aspirate blocking solution
and wash three times by adding 200 μl (or 500 μl) PBS. Make
rhodamine phalloidin in blocking solution at a 1:40 dilution,
and make enough to have 100 μl (or 250 μl) per well (always
calculate one extra well to account for pipetting errors). Incubate
at 4 °C for 20 min and keep covered with aluminium foil.
12. Bring mounting medium to room temperature.
13. Aspirate rhodamine phalloidine and wash three times by add-
ing 200 μl (or 500 μl) PBS.
14. Disassemble slide chamber unit carefully and remove slide con-
taining the prepared cells. If working with several slide chamber
units, do this one-by-one and mount to avoid cells drying out.
15. Add a drop of mounting medium to a clean coverslip and gen-

tly lay it on top of the glass slide over the area containing the
prepared cells. Press coverslip gently between paper towels and
use tweezers if necessary to remove any air bubbles.
16. Allow slides at least 30 min to dry while protected from light.
Strong nail hardener can be applied to seal the cover glasses.
Slides are then ready for microscopy or can be stored at 4 °C
protected from light.
3.5 Microscopy 1. Confocal microscopy and acquisition/export of data should be

and Image Processing conducted following local institutional guidelines and proto-
cols. It is important to image all areas in each cell preparation,
optimally by an investigator blinded to the examined condi-
tion. We routinely scan each high-power field using a 63×
objective and capture images in all areas containing cell aggre-
gates between B and T cells. B cells can be identified based on
their CMAC labelling.
2. We routinely export images as LSM files and analyse them
using the AxioVision Version 4.8 image analysis software
(Zeiss). The AxioVision outline tool is used to draw around
each synapse between T cells and B cells, and all available inter-
actions are scored in each condition. Optimally, scoring is per-
formed by two independent and blinded investigators. The
synapse area is then reported as the area of T-cell F-actin
immune synapses (μm2) value and is exported into Prism
Version 5 software (GraphPad) for statistical analysis.
4 Notes
1. Acceptable cell viabilities: to obtain suitable numbers of

intact cells and cell/cell conjugates, the cell viability after thaw-
ing/before starting the cell purification should be at least 80%.
Using cells with lower viability results in poor images and very
low numbers of quantifiable B-cell/T-cell conjugates.
2. Requirements of sterile conditions: although the complete
assay can be performed on the bench, we keep cell culture
media sterile and thaw and purify cells under sterile conditions.
This is particularly important when cells are used in short-term
co-culture experiments or are undergoing ex vivo treatment
with specific drugs.
3. Dissolving Triton: To facilitate pipetting of very viscous
Triton, cut the tip of a 1000 ml pipette tip and pipette slowly.
For adequate mixing, vortex the tube and, if necessary, keep on
roller.
4. Handling sAg: Handle with care and avoid contact with skin
and eyes and formation of dust and aerosols! To prevent acute
oral, dermal and inhalation toxicity, follow appropriate health
and safety guidelines.
5. Selection of purification strategies: the choice of cell separation
kit is guided by required cell purities and numbers and whether
unlabelled or labelled cells are needed in downstream experi-
ments. We routinely use column-based cell purification strategies,
where cells of interest are magnetically labelled and separated by
running the sample over a column (alternatively, column-less cell
selection strategies can be applied). The flow-through contains
cells depleted of the labelled cells, and labelled cells retained in the
column are eluded. We are routinely labelling peripheral blood
mononuclear cells (PBMCs) or mouse spleen cells with anti-
human or mouse CD19 microbeads to obtain B cells as APCs.
For selection of T cells, CD3, CD4 and CD8 microbeads can be
used, as well as specific T-cell isolation kits for negative selection
of T cells. As all cell purifications are a multistep process, always
label tubes clearly with sample names and whether cells were
obtained as positive (labelled) or negative (unlabelled) frac-
tions. Working swiftly, avoiding major delays or gaps and keeping
cells and reagents at 4 °C are essential for high cell viabilities and
purities. After selection, we routinely confirm the purities of cell
fractions by flow cytometry (optimally >95%).
6. Confirmation of purity: We retain small aliquots of cells
(<0.5 × 106) from unpurified PBMCs/whole spleen suspen-
sions, purified B and purified T cells. We have often found that
labelling cells with CD19 microbeads prevents binding of
fluorochrome labelled CD19 antibody. To determine the

purity of CD19+ B cells, we therefore stain human cells with
CD20 and murine cells with B220.
7. Manual versus automated cell separation: for large numbers
of samples, we have facilitated the experimental workflow by
using automated cell separation techniques (e.g. autoMACS®
Pro Separator).
8. Choice of cell concentrator: if using a one-well cell concen-
trator, 4 × 106 B cells and 4 × 106 T cells are required. B- and
T-cell numbers can be reduced to 1 × 106 each by using a
three-well cell concentrator. This also allows to have up to
three separate conditions on one slide (minimization of batch
effects) and reduces the total number of slide chamber units to
be handled during one experiment.
9. Assembling a slide chamber unit: make sure to use clean
components that are free of cell contamination from previous
experiments. Label slide on frosted end, and place into backing
plate. Check silicon gasket for wear and tear, then firmly push
into bottom opening of the cell concentrator and place on top
of the backing plate containing the slide. Fixate by adding a

metal clip on either side. It is essential that all slide chamber
units are fully assembled and the cytocentrifuge unit set up by
the end of the B-cell/T-cell conjugation incubation.
10. Selection of slides: using poly-l-lysine-coated glass slides or
other slides that are modified to attract and adhere cells is
essential to obtain assessable cell conjugates. They must not be
replaced by conventional microscope slides.
11. Differences between human and mouse cells: the protocol
described above can be followed for both human and mouse
cells. When working with mouse cells, full medium is supple-
mented with 50 μM β-mercaptoethanol, and after centrifuga-
tion cells are always carefully resuspended with a p1000 Gilson
pipette before adding additional volume.
12. Using fresh versus frozen cells: the assay can be performed
with both fresh and frozen cells, but this should be consistent
within one experiment. Because of the substantial time required
for cell purification and preparation of cell conjugates, we usu-
ally perform the assay on frozen cells. In selected experiments,
we have purified cells from fresh samples and have cryopre-
served B- and T-cell lymphocyte subsets.
13. CMAC labelling of cells: CMAC cell tracker dye is used to label
purified APCs to allow identification of B-cell/T-cell conjugates.
As this is a nonspecific dye, high purities (>95%) are essential. We
have previously resuspended B-cell pellets directly in 1 ml serum-
free medium/1 μl CMAC dye. However, we have found more
consistent labelling of B cells by resuspending them in 500 μl
serum-free medium first and then adding 500 μl serum-free
medium/1 μl CMAC dye while gently vortexing cells.
14. Importance of sAg stimulation: stimulation of B cells with
sAg cocktail prior to conjugation increases the area of F-actin
polymerization in both healthy and neoplastic B cells and facil-
itates the visualization of cytoskeleton defects.
15. Duration of conjugation: we have found similar results for
short (5 min) and long (30 min) conjugation between B and T
cells, but recommend a time-course experiment tailored to
specific experimental questions.
16. Importance of consistency: it is essential that this step is as
consistent as possible, i.e. the same techniques for resuspend-
ing and pipetting cells are used. Optimization experiments
should be conducted to ensure reproducibility and to establish
degree of variability between researchers.
17. Handling large numbers of slide chamber units: if doing
more than four cytofuge chamber experiments, spin, fix and
wash (time each batch separately). Keep all in last PBS wash
until all done.
References
1. Gribben, J.G., Riches, J.C. (2013) diagnosis have abnormal phenotype and geno-
Immunotherapeutic strategies including trans- type and form defective immune synapses with
plantation: eradication of disease. ASH AML blasts. Blood 114:3909–3916
Education Program Book 2013, pp 151–157 7. McClanahan F, Riches JC, Miller S, Day WP,
2. Gorgun G, Holderried TAW, Zahrieh D, Kotsiou E, Neuberg D, Croce CM, Capasso
Neuberg D, Gribben JG (2005) Chronic lym- M, Gribben JG (2015) Mechanisms of PD-L1/
phocytic leukemia cells induce changes in gene PD-1 mediated CD8 T-cell dysfunction in the
expression of CD4 and CD8 T cells. J Clin context of aging-related immune defects in the
Invest 115:1797–1805 Eμ-TCL1 CLL mouse model. Blood
3. Ramsay AG, Johnson AJ, Lee AM, Gorgun G, 126:212–221
Le Dieu R, Blum W, Byrd JC, Gribben JG 8. Shanafelt TD, Ramsay AG, Zent CS, Leis JF,
(2008) Chronic lymphocytic leukemia T cells Tun HW, Call TG, LaPlant B, Bowen D,
show impaired immunological synapse forma- Pettinger A, Jelinek DF, Hanson CA, Kay NE
tion that can be reversed with an immunomod- (2013) Long-term repair of T-cell synapse
ulating drug. J Clin Invest 118:2427–2437 activity in a phase II trial of chemoimmunother-
4. Ramsay AG, Clear AJ, Fatah R, Gribben JG apy followed by lenalidomide consolidation in
(2012) Multiple inhibitory ligands induce previously untreated chronic lymphocytic leu-
impaired T-cell immunologic synapse function kemia (CLL). Blood 121:4137–4141
in chronic lymphocytic leukemia that can be 9. McClanahan F, Hanna B, Miller S, Clear AJ,
blocked with lenalidomide: establishing a Lichter P, Gribben JG, Seiffert M (2015)
reversible immune evasion mechanism in PD-L1 checkpoint blockade prevents immune
human cancer. Blood 120:1412–1421 dysfunction and leukemia development in a
5. Ramsay AG, Evans R, Kiaii S, Svensson L, mouse model of chronic lymphocytic leukemia.
Hogg N, Gribben JG (2013) Chronic lympho- Blood 126:203–211
cytic leukemia cells induce defective LFA-1- 10. Xing D, Ramsay AG, Gribben JG, Decker WK,
directed T-cell motility by altering Rho GTPase Burks JK, Munsell M, Li S, Robinson SN, Yang
signaling that is reversible with lenalidomide. H, Steiner D, Shah N, McMannis JD, Champlin
Blood 121:2704–2714 RE, Hosing C, Zweidler-McKay PA, Shpall EJ,
6. Le Dieu R, Taussig DC, Ramsay AG, Mitter R, Bollard CM (2010) Cord blood natural killer
Miraki-Moud F, Fatah R, Lee AM, Lister TA, cells exhibit impaired lytic immunological syn-
Gribben JG (2009) Peripheral blood T cells in apse formation that is reversed with IL-2
acute myeloid leukemia (AML) patients at ex vivo expansion. J Immunother 33:684–696
Chapter 34
Studying the Immune Synapse in HIV-1 Infection

Iratxe del Río-Iñiguez*, Jérôme Bouchet*, and Andrés Alcover
Abstract
T cells are the main cellular targets of the human immunodeficiency virus 1 (HIV-1). HIV-1 infection
induces pleiotropic effects on the infected T cell that modify the T cell capacity to respond to antigen and
facilitates virus replication. HIV-1 infection subverts the formation and function of the immunological
synapse altering both actin cytoskeleton remodeling and intracellular vesicle traffic. We describe here our
methods to unveil how HIV-1 and in particular its protein Nef modify vesicle traffic to the immunological
synapse, perturbing the synapse activation capacity.
Key words Immunological synapse, Vesicle traffic, Endosomes, Cytoskeleton, T cell activation, TCR
signaling, HIV-1, Nef
1 Introduction
Immunological synapses are characterized by the accumulation

and clustering of TCRs, co-signaling receptors, adhesion mole-
cules, and signaling effectors. In addition, immunological synapses
are characterized by robust and precise rearrangements of the actin
and microtubule cytoskeleton, as well as of intracellular vesicle traf-
fic [1, 2] (see also Chapter 7).
HIV-1 infection induces pleiotropic effects in the infected T
cells. Among these effects, HIV-1 modulates intracellular vesicle
traffic of a variety of plasma membrane proteins, affecting their
subcellular localization, including their cell surface expression. The
HIV-1 viral proteins Nef, Vpr, and Vpu are responsible for many
of these intracellular traffic effects. These proteins target a variety
of cell surface molecules, including the virus receptor and co-
signaling molecule CD4, the major histocompatibility complex
(MHC) molecules classes I and II, CD1a, tetraspanins, NK activa-
tory ligands, and others, reviewed in [3]. In addition, HIV-1 Nef
can modify the intracellular traffic of signaling molecules, like the
*These authors contributed equally to this work.
545
546 Iratxe del Río-Iñiguez et al.
protein tyrosine kinase Lck, the first kinase engaged upon TCR
engagement. Thus, Lck accumulates in recycling endosomes, pre-
venting the formation of immunological synapses capable to effi-
ciently transduce TCR signals. The HIV-1 Nef protein is necessary
and sufficient to induce these effects [4]. In addition, HIV-1 Nef
impedes the traffic of vesicles carrying the signaling adapter LAT to
the immunological synapse, preventing the local generation of sig-
naling complexes [5]. Moreover, it has been reported that Nef-
induced Lck accumulation also brings to the Lck intracellular
compartment active Erk1/2 serine-threonine kinase. Together,
Lck and Erk may enhance IL2 production [6]. Under physiologi-
cal conditions, Lck traffic depends on the transport protein MAL
[7], the Unc119 protein [8], and Rab11 GTPase and its effector
FIP3 (Bouchet et al., submitted). The mechanism by which HIV-1
subverts Lck intracellular traffic remains poorly understood.
We describe here our recent methods aiming to elucidate how
HIV-1 Nef subverts intracellular traffic of signaling molecules and
its effects on T cell activation.
2 Materials
2.1 Cells 1. Jurkat T cell leukemia cells, J77 clone 20 cells, and Raji B cell
lymphoma cells have been previously described [4]. Cells are
cultured in RPMI 1640 + GlutaMAX™ + phenol red medium
(Gibco®) supplemented with 10% fetal calf serum and 10 mM
Hepes. We culture Jurkat and Raji cells at a density average of
0.5–1 × 106 cells/mL, splitting the cultures every 2–3 days.
2. Peripheral blood mononuclear cells from healthy donors are
isolated by centrifugation through Ficoll-Hypaque using
Unisep Maxi tubes (Eurobio, No. U-10) (see Note 1). For
HIV-1 infection assays, PBMCs are cultured at 2 × 106 cells/
mL in RPMI 1640 medium supplemented with 10% FCS, 1%
penicillin-streptomycin, and 5 μg/mL phytohemagglutinin
(PHA) for 2 days. At day 3, PBMCs are washed once in RPMI
1640 medium supplemented with 10% FCS and 1% penicillin-
streptomycin to get rid of PHA and resuspended at 2 × 106
cells/mL in RPMI 1640 medium supplemented with 10% FCS
and 1% penicillin-streptomycin. For transfection assays of pri-
mary cells, CD4+ T cells are further purified using the CD4+ T
cell isolation kit (Miltenyi Biotech, 130-096-533) (see Note 2).
After isolation, they are cultured at 2 × 106 cells/mL in RPMI
1640 medium supplemented with 10% FCS, 1 mM sodium
pyruvate, and 1% MEM nonessential amino acids.
2.2 Microscopy 1. Confocal microscope: LSM 700 confocal microscope (Zeiss)

Materials equipped with a Plan-Apochromat 63× objective and ZEN
software (Zeiss).
Immune Synapse in HIV-1 Infection 547
2. Square glass coverslips 20 × 20 mm. Coverslips are coated with

500 μL poly-l-lysine (0.002% w/v in water) during 20 min at
room temperature, washed with water, and air-dried before use
(see Note 3).
3. Round glass coverslips 12 mm diameter, coated as in item 2,
using 150 μL per round coverslip.
4. Glass slides 76 × 26 mm.
5. ProLong® Gold Antifade mounting medium with DAPI
(Molecular Probes®, Life Technologies™, No. P36935).
2.3 Chemicals 1. Poly-l-lysine MW: 150–300 kDa, 0.1% (w/v) (Sigma-Aldrich®,

and Biological No. P8920). Coating solution 0.002% in water.
Products 2. Paraformaldehyde (Electron Microscopy Sciences, No. 15714,
aqueous stock solution 32%). Paraformaldehyde solution 8%
(w/v) in water was prepared from commercial stock at 32%
(see Note 4).
3. Triton X-100 0.1% (v/v) in phosphate buffer pH 7.5, 150 mM
NaCl (PBS).
4. Methanol.
5. Bovine serum albumin, 1% (w/v) in PBS (PBS-BSA). Store
solution at 4 °C.
6. Staphylococcus enterotoxin E superantigen (SEE, Toxin
Technology Inc.), 10 μg/mL in PBS. Biohazard (see Note 5).
7. Texas red-coupled phalloidin probe from Invitrogen (1/100
dilution).
8. Phytohemagglutinin PHA-P.
9. Primary antibodies for immunofluorescence. Mouse mono-
clonal IgG2b anti-Lck, clone 3A5 (Santa Cruz Biotechnology),
is used at 2 μg/mL. Mouse monoclonal IgG1 anti-CD3ε,
clone UCHT1 (BioLegend Inc), is used at 10 μg/mL. Rabbit
anti-centrin-3, gift of M. Bornens (Institut Curie, France), is
used at 1/400 dilution. Mouse IgG2b anti-β-tubulin, clone
KMX1 (Millipore), is used at 10 μg/mL. Rabbit anti-phospho-
ZAP70 (Y319) (Cell Signaling Technology) is used at 1/100
dilution, and mouse monoclonal IgG2a anti-phospho-TCRζ
(Y142) and clone K25–407.69 (Becton Dickinson) are used
at 5 μg/mL. Anti-HIV-1 JR-CSF Nef monoclonal (6.2) and
anti-HIV-1 SF2 p24 polyclonal antibodies can be obtained
from NIH AIDS Reagent Program and are used at 1/1000
and 1/50 dilution, respectively.
10. Secondary antibodies for immunofluorescence. Highly cross-
adsorbed Cy3-coupled goat anti-mouse IgG2a, anti-mouse
IgG2b, and anti-rabbit (Jackson Immuno Research
Laboratories) are used at 1/100 dilution. FITC-coupled goat
anti-mouse IgG1 (Southern Biotech) is used at 0.7 μg/mL.

FITC-coupled goat anti-rabbit (Jackson Immuno Research
Laboratories) is used at 1/100 dilution. Alexa Fluor
488-coupled goat anti-fluorescein (Molecular Probes) is used
at 1 μg/mL.
11. Antibody used for pseudosynapses: mouse IgG1 anti-CD3ε,
clone UCHT1 (BioLegend), is used at 500 ng/mL.
12. Expression vectors: pCGFP and wild-type or mutated Nef-
hemagglutinin (Nef-HA) and Nef-green fluorescent protein
(Nef-GFP), pCNef-GFP [9, 10].
13. Cell-free stocks of HIV-1 virions are produced by transient
transfection of HEK-293T cells with proviral plasmid as
described in [11]. The concentration of p24 antigen in viral
stocks is measured by enzyme-linked immunosorbent assay
(ELISA) (PerkinElmer Life Sciences) (see Note 6). Cell-free
HIV-1 virions, wild type and ΔNef, are stored at −80 °C in 1
mL aliquots at 2 μg/mL p24 in a BSL-3 facility.
14. Human recombinant IL-2.
2.4 Transfection 1. Neon™ Transfection System and Neon™ Transfection 100 μL

Systems Kit, containing electrolytic buffer E2, resuspension buffer R,
100 μL Neon™ tips, and Neon™ electroporation tubes
(Invitrogen™, Life Technologies™).
2. Nucleofector™ 2b Device and Primary Cell Nucleofector™
Kit, containing Nucleofector™ solution, single-use pipette,
and Amaxa™ 100 μl aluminum electrode cuvettes (Lonza).
3 Methods
3.1 T Cell Cell synapse assays are done 24–48 h after T cell transfection of
Transfection expression vectors.
by Electroporation
3.1.1 Transfection of T Transfection is done with 5–10 × 106 cells. Use 10 μg of DNA
Cell Lines Using plasmid for 10 × 106 cells. Complete with resuspension buffer R to
the Neon™ equilibrate different volumes for the same amount of DNA when
Transfection System preparing Eppendorf tubes with plasmid DNA.
1. Preincubate at 37 °C one flask with 10 mL T cell culture
medium (RPMI 1640 supplemented with 10% fetal calf serum
and 10 mM Hepes) for each transfection. Prepare Eppendorf
tubes with DNA plasmid.
2. Harvest the amount of cells required and centrifuge at 290 × g
at 20 °C for 4 min. Wash twice in PBS and aspirate
supernatant.
3. Insert a new Neon™ tube into the Neon™ pipette station and
fill it with 3 mL electrolytic buffer E2. Turn on Neon™ device,
and select electroporation protocol (voltage, 1400 V; width,
10 ms; pulses, 3).
4. Resuspend the cell pellet in 100 μL resuspension buffer R per
required transfection.
5. Add 100 μL resuspended cells in each Eppendorf tube with
DNA plasmid, and mix gently.
6. Take 100 μL of the cells-plasmid mix using the Neon™ pipette
and a 100 μL tip—avoid air bubbles (change Neon™ tip every
transfection with different DNA). Insert the Neon™ pipette
with the sample vertically into the Neon® tube, and press start.
7. Once electroporation is complete, remove Neon™ pipette and
transfer the sample into the pre-warmed culture flask, and
incubate at 37 °C and 5% CO2 for 24–48 h (see Note 7).
3.1.2 Transfection Transfection is done with 5–10 × 106 cells. Use 5 μg of DNA plas-
of Purified Primary CD4+ T mid for 10 × 106 cells. Complete with resuspension buffer R to
Cells Using Nucleofector™ equilibrate different volumes for the same amount of DNA when
2b Device preparing Eppendorf tubes with plasmid DNA.
1. Preincubate at 37 °C one flask with 1 mL primary CD4+ T cell
culture medium (RPMI 1640 medium supplemented with
10% FCS, 1 mM sodium pyruvate, and 1% MEM nonessential
amino acids) per 10 × 106 cells transfected. Prepare 1.5 mL
Eppendorf tubes with DNA plasmid.
2. Harvest the amount of primary CD4+ T cells required, and
centrifuge at 453 × g at 20 °C for 10 min. Wash twice in PBS.
3. Turn on Nucleofector™ 2b Device and select electroporation
protocol (U014).
4. Add corresponding volume of DNA plasmid to the Amaxa
cuvette per transfection required.
5. Resuspend the cell pellet in 100 μL Amaxa buffer per 10 × 106
CD4+ cells.
6. Take 100 μL of resuspended cells and add them to the cuvette
with DNA plasmid. Mix gently, and avoid air bubbles.
7. Insert cuvette in the Nucleofector™ 2b Device and electropor-
ate the cells using protocol U014.
8. Use single-use pipettes to recover cells and transfer to the pre-
warmed flask with medium. Incubate 10 min at 37 °C.
9. Count cells and resuspend to a final concentration of 2 × 106
cells/mL in RPMI1640 medium supplemented with 10%
FCS, 1 mM sodium pyruvate, and 1% MEM nonessential
amino acids.
3.2 Infection 1. Jurkat T cells (5–10 × 106) are cultured with 2 μg/mL of cell-
free HIV-1 virions during 16 h in RPMI 1640 medium supple-
mented with 10% FCS. Cells are then washed four times in
RPMI 1640 and resuspended in RPMI 1640 medium supple-
mented with 10% FCS and cultured during 3 days.
2. After 2 days of PHA stimulation, 5 × 106 PBMCs are resus-
pended at 2 × 106 cells per mL in a suspension of 2 μg/mL
cell-free HIV-1 virions (Subheading 2.3, item 13), in RPMI
1640 medium supplemented with 10% FCS, during 16 h. Cells
are then washed four times in RPMI 1640 and resuspended in
RPMI 1640 medium supplemented with 10% FCS and 10 U/
mL IL-2 and cultured during 3 days (see Note 8).
3.3 Immunological All the procedures involving active viruses have to be performed in
Synapse Formation a BSL-3 facility by trained personnel.
Between T Cells
1. Pulse antigen-presenting cells with superantigen. Harvest Raji
and Antigen- cells (5 × 106 cells/mL) by centrifuging at 290 × g, 20 °C,
Presenting Cells 4 min. Resuspend in RPMI 1640 (without serum) supple-
mented with 10 μg/mL Staphylococcus enterotoxin superanti-
gens (SEE for Jurkat or a mix of SEA, SEB, SEE for primary T
cells) during 30 min at 37 °C.
2. Immunological synapse formation. Harvest Jurkat J77 clone
20 cells by centrifuging at 290 × g at 20 °C for 4 min or pri-
mary CD4 T cells, by centrifuging at 453 × g at 20 °C for 6
min, and resuspend at 5 × 106 cells/mL RPMI 1640 (without
serum). T cells are incubated with pulsed Raji cells at 1:1 ratio,
at 37 °C during 5, 15, or 30 min in RPMI 1640 medium
(without serum) in 1.5 mL Eppendorf tubes at 37 °C in a
water bath.
3. Plating cells on coverslips. 160 μL of conjugated cells are
plated onto poly-l-lysine-coated square coverslips during
4. Fixation. 160 μL of 8% paraformaldehyde is mildly dropped
onto the coverslips containing cell suspensions (final con-
centration of paraformaldehyde 4%) and incubated for
20 min at room temperature. After incubation, paraformal-
dehyde is removed, and coverslips are washed with PBS-
BSA (see Notes 4 and 10).
bation in PBS-BSA.
6. First antibody preparation. Primary antibody (or mix of pri-
mary antibodies in case of multiple staining) at the recom-
mended dilution is suspended in PBS-BSA, 0.1% Triton X-100.
Anti-HIV-1 p24 or anti-HIV-1 Nef antibodies are included to
distinguish infected cells.
7. Cell immunostaining. 80 μL drops of primary antibody solution

are deposited on Parafilm®, and coverslips with fixed cells are
turned over on drops (see Note 11). Incubate for 1 h at room
washed twice in PBS-BSA by submerging them several times in
a beaker using forceps to handle the coverslips.
pended in PBS-BSA.
10. Staining with second antibodies. 80 μL drops of secondary
antibody solution are deposited on Parafilm®, and coverslips
are turned over on drops, as in step 7. Incubate for 45 min–1
h at room temperature, protected from light.
11. Washes. After secondary antibody incubation, coverslips are
forceps to handle the coverslips. Drain the washing solution by
setting the edge of the coverslip on absorbent paper, in order
to remove excess PBS-BSA before mounting.
ing medium with DAPI are deposited on 76 × 26 mm slides,
and coverslips are turned over on the drops. Let mounted
slides harden overnight at room temperature, protected from
light. Slides may be then stored at 4 °C and used during 2–4
weeks (see Note 12).
13. Microscopy analysis. Cells are observed under a LSM 700 con-
acquired at different increments depending on desired image
analysis: 0.2 μm depth increments for deconvolution and colo-
calization analysis and 1 μm depth increments for fluorescence
intensity analysis are used. Green and red laser excitations are
intercalated to minimize fluorescence spill over different chan-
nels. Image acquisition is done with ZEN software (Zeiss).
3.4 Pseudosynapse 1. Coating coverslips with anti-CD3 antibodies. Prepare 10 μg/

Formation on Anti- mL anti-CD3 (UCHT-1) in PBS. Round poly-l-lysine-coated
CD3-Coated Coverslips coverslips are turned over 50 μL drops of anti-CD3 solution
and incubated 2 h at 37 °C in a humidified chamber or over-
night at 4 °C (see Note 13). Coverslips are washed once in PBS
and saturated with 200 μL RPMI 1640 with serum to prevent
nonspecific binding to poly-l-lysine.
2. Cell resuspension. Jurkat J77 clone 20 cells or primary CD4 T
cells are washed twice by centrifugation at 290 × g, 20 °C,
4 min and resuspended at 2 × 106 cells/mL in RPMI 1640
medium without serum.
3. Plating cells on coverslips. 100 μL of cells are plated onto

anti-CD3-coated coverslips during 3, 5, or 15 min at 37 °C
in a humidified chamber kept at 37 °C during the assay.
4. Fixation. 100 μL of 8% paraformaldehyde is mildly dropped
room temperature. After incubation PFA is removed, and cov-
erslips are washed with PBS-BSA (see Notes 4 and 10).
bation in PBS-BSA.
6. First antibody preparation. Primary antibody (or mix of pri-
mary antibodies in case of multiple staining) at the recom-
mended dilution is suspended in PBS-BSA, 0.1% Triton X-100.
7. Staining of coverslips. 30 μL drops of primary antibody solu-
tion are deposited on Parafilm®, and coverslips with fixed cells
washed twice in PBS-BSA by submerging several times in a
beaker using forceps to handle the coverslips.
pended in PBS supplemented with 1% (w/v) BSA.
10. Staining of coverslips: 30 μL drops of second antibody solution
are deposited on Parafilm®, and coverslips are turned over on
drops, as done in step 7. Incubate for 45 min–1 h, at room
11. Washes. After second antibody incubation, coverslips are
forceps to handle the coverslips. Drain the washing solution by
setting the edge of the coverslip on absorbent paper, in order
to remove excess PBS-BSA before mounting.
ing medium with DAPI are deposited on 76 × 26 mm slides,
and coverslips are turned on the drops. Let mounted slides
harden overnight at room temperature, protected from light.
Slides may be then stored at 4 °C and used during 2–4 weeks
(see Note 12).
13. Microscopy analysis. Cells are observed under a LSM 700
confocal microscope (Zeiss) equipped with an oil-immersion
Plan-Apochromat 63× objective. Z-stack optical sections are
acquired at different increments depending on image analysis
performed after 0.2 μm depth increments for deconvolution
and colocalization analysis and 1 μm depth increments for
fluorescence intensity analysis. Green and red laser excitations
are intercalated to minimize cross talk between the acquired

fluorescence channels. Image acquisition is done with ZEN
software (Zeiss).
3.5 Quantitative 1. Deconvolution of confocal images is used to improve image

Image Analysis rendering, especially for 3D reconstruction and prior to colo-
calization analysis. Deconvolution is performed on Z-stacks of
confocal optical sections obtained at 0.2 μm depth increments,
using Huygens Professional software (Scientific Volume
Imaging).
2. Colocalization analysis is performed with Fiji software (open
platform for scientific image analysis). JaCoP plugin or colocal-
ization threshold option of Fiji is used.
Pearson’s correlation coefficient corresponds to the linear
relationship between intensity of pixels in the two analyzed
channels. Mander’s coefficient is defined as the ratio of the
summed intensities of pixels from one image for which the
intensity in the second channel is above the threshold [12].
Costes automatic threshold allows us to consider only the
pixels in each channel showing statistical correlation, as
explained in Costes et al. [13].
3. Fluorescence intensity analysis. Using Fiji, in a chosen chan-
nel, creates a Z-projection of the image (whole cell or a
number of optical sections corresponding to an intracellular
compartment) and selects a region of interest (i.e., immuno-
logical synapse or an intracellular compartment), depending
on experimental requirements. Then, in the other channel,
measure fluorescence intensity of the selected area of interest
(Fig. 1).
4 Notes
1. Density-gradient centrifugation to isolate peripheral blood

mononuclear cells (PBMC). Ficoll separation: spin Ficoll
tubes 5 min at 453 × g at room temperature. Lay 28 mL of
blood per tube of 15 mL of Ficoll-Hypaque-Lymphoprep
tubes. Centrifuge 30 min at 805 × g at room temperature with
no brake. Aspirate plasma layer containing PBMCs (Fig. 2),
and add it to new empty falcon tubes (15 mL recovered
PBMCs per 50 mL falcon tube). Add 35 mL RPMI no FCS per
15 mL of recovered PBMCs per tube. Spin 10 min at 453 × g
at room temperature. Discard supernatants and pool pellets
together to one tube. Wash twice with 50 RPMI no FCS
(10 min 453 × g). Resuspend pellet in 50 mL of RPMI no
FCS, and count cells using cell counter or trypan blue vital
staining and Malassez chambers.
Fig. 1 Quantification of fluorescence accumulation at the immunological synapse. Immunological synapse of

a Jurkat (J77 cl20) lymphocyte with Raji B lymphocyte as antigen-presenting cell (APC) pulsed with SEE
superantigen. Confocal image was post-treated by deconvolution. A 4 μm Z-projection is shown. (A) Non-
transfected cells. (B) Nef-GFP transfected cells. From left to right: Nef-GFP (green), Lck (red) intracellular, CD3
(far red) surface staining, merge and phase contrast image. Rectangles show the region of interest in which
quantification of the intensity of fluorescence is performed and is referred to the total fluorescence of the cell.
Scale bar, 5 μm
Fig. 2 Scheme of a Ficoll gradient centrifugation tube after centrifugation
2. Purification of CD4+ T cells using MACS CD4 cell isolation kit

II: follow manufacturer’s instructions.
3. Coverslip coating with poly-l-lysine: prepare a 1:50 dilution of
0.01% poly-l-lysine solution in water. Cover a flat surface with
parafilm, and place round or square coverslips over it. Add 500
μL for square coverslips or 150 μL for round coverslips of the
1:50 poly-l-lysine dilution over the coverslips, covering the
surface completely. Incubate 20 min at room temperature and
remove excess. Wash once with water and dry the coverslips
before use.
4. Paraformaldehyde. We recommend purchasing a stock com-
mercial solution, avoiding manipulation of paraformaldehyde
powder, in order to prevent toxicity. Paraformaldehyde solu-
tion manipulation should be carried out under a chemical
hood.
5. Staphylococcus enterotoxins are biohazard and should be
manipulated as such. Inactivation of all solutions and materials
in contact with the toxin should be performed at the end of
experiments using an excess of 0.5% (w/v) sodium hypochlo-
rite for at least 1 h. Dispose as biohazard.
6. All the procedures involving active viruses have to be per-
formed in a BSL-3 facility by trained personnel. For the detec-
tion and quantification of HIV-1 p24 antigen in cell culture
supernatant, follow manufacturer’s instructions.
7. Different plasmids might need different expression times in
order to reach enough protein levels. Expression may be differ-
ent in Jurkat and in primary T cells.
8. Schematic representation of infection time line (Fig. 3). All the
procedures involving active viruses have to be performed in a
BSL-3 facility by trained personnel.
9. Cover with parafilm a flat surface, and place the poly-l-lysine-
coated coverslips over it (poly-l-lysine side of the coverslips
facing up). Add cell suspension over the coverslip, trying to
distribute it over the whole coverslip surface.
Fig. 3 Schematic representation of infection time line

10. Once fixed and washed, round or square coverslips containing

cells may be stored in 6- or 24-well plates, respectively, in an
excess of PBS-BSA, for 1–2 weeks at 4 °C. Staining does not
require to be performed immediately after fixation, except for
protein phosphorylation staining, which requires freshly pre-
pared cells for optimum results.
11. Cover with parafilm a flat surface and label areas for different
staining solutions planned. Deposit 80 μL drops of primary
antibody mix on the parafilm. Take the coverslips with fixed
cells, and drain the excess of PBS-BSA by setting the edge of
the coverslip on absorbent paper. This prevents dilution of
antibody staining solutions. Place coverslips over primary anti-
body drops on parafilm, cells facing the drops. Create a humid-
ified chamber for the incubation by placing wet pieces of paper
around the parafilm, and cover everything with a lid or opaque
surface, in order to protect from light.
12. Once slides have been analyzed, it is recommended to store
them at −20 °C, to prevent loss of fluorescence. While fluoro-
chrome fluorescence tends to be stable, GFP fluorescence is
much faster lost.
13. Anti-CD3 coating of coverslips. Cover with parafilm a flat sur-
face, and deposit 50 μL drops of anti-CD3 antibody mix on
the parafilm. Create a humidified chamber for the incubation
by placing wet pieces of paper around the parafilm, and cover
everything with a lid.
Acknowledgments
This work was supported by grants from Agence Nationale de la

Recherche (ANR, No. 11 BSV3 025 01), Agence National de
Recherche sur le SIDA et les Hepatitis Virales (ANRS), Institut
Pasteur, CNRS, and INSERM. IdRI is a scholar in the Pasteur-
Paris University (PPU) International PhD program and is funded
by the People Programme (Marie Curie Actions) of the European
Union’s Seventh Framework Programme FP7/2007–2013/
under the REA grant agreement n° 317057 HOMIN. J.B. was
supported by ANRS and Roux-Institut Pasteur postdoctoral fel-
lowships. The following reagents were obtained through the
NIH AIDS Reagent Program, AIDS Program, NIAID, and
NIH: Anti-HIV-1 JR-CSF Nef Monoclonal (6.2) from Dr.
K. Krohn and Dr. V. Ovod [14] and anti-HIV-1 SF2 p24 poly-
clonal antibody and human rIL-2 from Dr. M. Gately,
Hoffmann-La Roche Inc.
References
1. Soares H, Lasserre R, Alcover A (2013) 8. Gorska MM, Liang Q, Karim Z, Alam R (2009)
Orchestrating cytoskeleton and intracellular Uncoordinated 119 protein controls trafficking
vesicle traffic to build functional immunologi- of Lck via the Rab11 endosome and is critical
cal synapses. Immunol Rev 256:118–132 for immunological synapse formation.
2. Agüera-Gonzalez S, Bouchet J, Alcover A J Immunol 183:1675–1684
(2015) Immunological Synapse. eLS. John 9. Laguette N, Bregnard C, Bouchet J, Benmerah
Wiley & Sons, Ltd., Chichester. A, Benichou S, Basmaciogullari S (2009) Nef-
doi:10.1002/9780470015902.a0004027. induced CD4 endocytosis in human immuno-
pub2 deficiency virus type 1 host cells: role of p56lck
3. Sugden SM, Bego MG, Pham TN, Cohen EA kinase. J Virol 83:7117–7128
(2016) Remodeling of the host cell plasma 10. Madrid R, Janvier K, Hitchin D, Day J, Coleman
membrane by HIV-1 Nef and Vpu: a strategy S, Noviello C, Bouchet J, Benmerah A, Guatelli
to ensure viral fitness and persistence. Viruses J, Benichou S (2005) Nef-induced alteration of
8:1–30 the early/recycling endosomal compartment
4. Thoulouze MI, Sol-Foulon N, Blanchet F, correlates with enhancement of HIV-1 infectiv-
Dautry-Varsat A, Schwartz O, Alcover A ity. J Biol Chem 280:5032–5044
(2006) Human immunodeficiency virus type-1 11. Craig HM, Pandori MW, Guatelli JC (1998)
infection impairs the formation of the immuno- Interaction of HIV-1 Nef with the cellular
logical synapse. Immunity 24:547–561 dileucine-based sorting pathway is required
5. Abraham L, Bankhead P, Pan X, Engel U, for CD4 down-regulation and optimal viral
Fackler OT (2012) HIV-1 Nef limits commu- infectivity. Proc Natl Acad Sci U S A
nication between linker of activated T cells and 95:11229–11234
SLP-76 to reduce formation of SLP-76- 12. Bolte S, Cordelieres FP (2006) A guided tour
signaling microclusters following TCR stimula- into subcellular colocalization analysis in light
tion. J Immunol 189:1898–1910 microscopy. J Microsc 224:213–232
6. Pan X, Rudolph JM, Abraham L, Habermann 13. Costes SV, Daelemans D, Cho EH, Dobbin Z,
A, Haller C, Krijnse-Locker J, Fackler OT Pavlakis G, Lockett S (2004) Automatic and
(2012) HIV-1 Nef compensates for disorgani- quantitative measurement of protein-protein
zation of the immunological synapse by induc- colocalization in live cells. Biophys J 86:
ing trans-Golgi network-associated Lck 3993–4003
signaling. Blood 119:786–797 14. Ovod V, Lagerstedt A, Ranki A, Gombert FO,
7. Anton O, Batista A, Millan J, Andres-Delgado Spohn R, Tahtinen M, Jung G, Krohn KJ
L, Puertollano R, Correas I, Alonso MA (2008) (1992) Immunological variation and immuno-
An essential role for the MAL protein in target- histochemical localization of HIV-1 Nef dem-
ing Lck to the plasma membrane of human T onstrated with monoclonal antibodies. AIDS
lymphocytes. J Exp Med 205:3201–3213 6:25–34
Chapter 35
In Vivo Imaging of T Cell Immunological Synapses

and Kinapses in Lymph Nodes
Hélène D. Moreau and Philippe Bousso
Abstract
T cells can become activated in lymph nodes following a diverse set of interactions with antigen-presenting
cells. These cellular contacts range from short and dynamic to stable and long-lasting interactions, termed
kinapses and synapses, respectively. Here, we describe a methodology to generate naïve T cells expressing
a fluorescent probe of interest through the generation of bone marrow chimeras and to image T cell
dynamics using intravital two-photon microscopy. In these settings, the formation of kinapses and synapses
can be triggered by the administration of low and high affinity peptides, respectively. Finally, 3D cell track-
ing can help classify distinct T cell behaviors. These approaches should offer new possibilities for dissecting
the process of T cell activation in vivo.
Key words T cell activation, Synapse, Kinapse, In vivo imaging, Antigen affinity
1 Introduction
Naive T cells migrate vigorously in lymph nodes in an apparent

random manner, searching for their cognate antigen [1]. The rec-
ognition of specific pMHC complexes profoundly alters T cell
motility leading to at least two possible behaviors. Strong signal
tends to result in complete T cell arrest and the formation of a
stable contact termed immunological synapse. Such interaction
typically lasts for several hours and is often associated with efficient
T cell priming [2]. Weaker signals favor the formation of a more
dynamic T cell-APC interaction (termed kinapse) during which T
cells maintain a low motility, scanning the APC for a few minutes
before detaching and resuming motility [3–5]. Kinapses can poten-
tially result in T cell activation, in part due to the ability of T cells
to sum-up suboptimal signals received during successive interac-
tions with APCs [6–9].
These different contacts have been visualized with two-photon
imaging in a variety of settings, revealing an unappreciated com-
plexity in the process of T cell activation in vivo [10].
559
560 Hélène D. Moreau and Philippe Bousso
In addition to differences in duration of contact and T cell

velocity, kinapses and synapses may exhibit distinct molecular and
subcellular reorganization as well as distinct signaling properties.
Subcellular and functional imaging of T cell activation in vivo has
been limited to a few studies [11–13] due to the typical low signal-
to-noise ratio for fluorescently tagged proteins and to the difficulty
to introduce fluorescent probes in naive T cells.
We describe here a methodology to generate naive T cells
expressing a fluorescent probe of interest (such as a GFP-fusion
protein relevant to TCR signaling) and to image T cell dynamics
upon antigen recognition in lymph nodes by means of intravital
two-photon imaging. By administrating altered peptide ligands
harboring different affinities for the TCR of interest, it is possible
to favor and study the formation of either kinapses or synapses.
Finally, generated time-lapse movies can be subjected to three-
dimensional cell tracking to help identify freely migrating as well as
kinapse- and synapse-forming T cells.
2 Materials
2.1 Production 1. Plat-E cell line [14].

of Viral Vectors 2. pCl-eco helper plasmid.
3. Retroviral vector. The cDNA of the protein of interest can be
cloned into pMSCV2.2-EGFP to generate a fusion protein as
it was previously done for LAT [12].
4. Lipofectamine 2000 (Invitrogen) or equivalent.
5. Complete DMEM: DMEM + 2 mM l-alanyl-l-glutamine,
10% fetal calf serum (FCS), 50 U/mL penicillin and 50 μg/
mL streptomycin, 10 mM HEPES, 1 mM sodium pyruvate.
6. Plat-E medium: Complete DMEM, puromycin (1 μg/mL),
blasticidin (10 μg/mL).
7. Opti-MEM.
2.2 Generation 1. 5-Fluorouracil (5-FU).

of Bone Marrow 2. BM medium: DMEM + 2 mM l-alanyl-l-glutamine, 15% FCS,
Chimeras 50 U/mL penicillin and 50 μg/mL streptomycin, 10 mM
HEPES, interleukin (IL)-6 (20 ng/mL), IL-3 (10 ng/mL),
stem cell factor (SCF) (20 ng/mL).
3. Polybrene.
4. PBS.
5. 70% ethanol.
2.3 Preparation 1. Complete RPMI: RPMI 1640 + 2 mM l-alanyl-l-glutamine,

of T Cells 10% FCS, 50 U/mL penicillin and 50 μg/mL streptomycin, 1
mM sodium pyruvate, 10 mM HEPES, 0.1% β-mercaptoethanol.
Intravital Imaging of Synapses and Kinapses 561
2. CD8+ T cell negative purification kit (e.g., Dynal or Miltenyi).

3. PBS.
4. PBS + 10% FCS.
5. SNARF-1 carboxylic acid, acetate, succinimidyl ester (e.g.,
Invitrogen).
2.4 Imaging 1. Two-photon laser scanning microscope.

of Synapses 2. Peristaltic pump with heater.
and Kinapses
3. Homemade heating platform (Fig. 1a).
in the Lymph Node
4. Ring-shaped metallic tube assembled with silicone sealant on a
glass coverslip (Fig. 1e and f).
5. Surgical instruments.
6. Anesthetic: 7.65 mL of PBS, 1.6 mL of Imalgène (100 mg/mL
ketamine), 0.75 mL of Rompun (20 mg/mL xylazine). These
are controlled substances and should be used in accordance
with local regulations.
7. Plaster bandages.
Fig. 1 Preparation of the mouse for intravital lymph node imaging. These pictures illustrate some of the
important steps during the preparation of the popliteal lymph node for intravital imaging. (a) Custom-designed
heating platform. (b–d) Preparation of the plaster cast to immobilize the lower hind leg. (e, f) A ring-shaped,
metallic tube glued to a coverslip is placed on the top of the popliteal lymph node. (g) After completing expo-
sure of the lymph node, the heated platform is placed on the microscope stage. Reproduced from Celli and
Bousso 2007 [15] with permission of Springer
8. Surgical tape.
9. Glue.
10. PBS.
11. Heparin sodium.
12. Altered peptide ligands for the model TCR. For the OT-I
TCR, we use the native peptide SIINFEKL (N4) and the
altered peptide ligand SIIQFEKL (Q4) to generate synapses
and kinapses, respectively. We typically inject 50 μg in 100 μL
of PBS.
3 Methods
3.1 Transfection 1. Day -1 around 3 pm: Plate Plat-E cells at 5.5 × 106/100 mm
of Packaging Cell Line Petri dish in 15 mL of Plat-E medium (see Note 1).
and Production 2. Day 0 morning: Transfection. Prepare tube #1: 1.5 mL of
of Viruses Opti-MEM + 60 μL of Lipofectamine 2000 (see Note 2). Tap
to mix the tube. Incubate 5 min at room temperature. Prepare
tube #2: 1.5 mL of Opti-MEM + 30 μg of vector DNA + 5 μg
of pCl-eco helper plasmid. Mix tube #1 and tube #2 after
5 min incubation (within 30 min). Incubate 25 min at room
temperature. During incubation period, wash the Plat-E plate
with PBS and add 5 mL of Opti-MEM. After 25 min incuba-
tion, add tube #1 + tube #2 mixture (3.12 mL total) dropwise
to the plate (see Note 3). Incubate 8 h at 37 °C. Quench with
12 mL of complete DMEM. Incubate overnight at 37 °C.
3. Day 1 morning: Replace media with 8 mL of fresh complete
DMEM and incubate at 37 °C overnight (see Note 4).
4. Day 2 morning: Collect viral supernatant and place on ice.
Replace with 8 mL of fresh complete DMEM and put plate
back at 37 °C. Spin viral supernatant at 4 °C, at 300 × g for
5 min. Pour viral supernatant into clean tube, and place back
on ice until used for bone marrow transduction.
5. Day 3 morning: Repeat Day 2 procedure (see Note 5).
3.2 Generation All work with mice should be done in accordance with local laws
of Bone Marrow and regulations.
Chimeras
1. Day -5 or Day -4: Inject 5 mg/mouse of 5 FU intravenously
(see Notes 6 and 7).
2. Day 0: Harvest bone marrow (BM) from long bones. Dissect
muscle away from the bone (place in PBS on ice during har-
vest). Soak in 70% ethanol for 2–3 min then rinse in PBS. Cut
ends with sterile scissors and flush with 27 ga needle with com-
plete DMEM. Disrupt core by flushing through a 18 ga needle,
pass over cell strainer into 50 mL tube, and pellet at 300 × g for
10 min at 4 °C. Lyse red blood cells and wash with complete

DMEM. Count cells. Resuspend at 1.5–2 × 106 cells/mL in
BM medium. Plate between 15 and 20 million cells per 100 mm
dish (see Note 8). Incubate for 2 days at 37 °C.
3. Day 2 morning: BM transduction #1. Prepare viral cocktail:
mix 2.67 mL of BM medium with 1.33 mL of viral superna-
tant (prepared as described in Subheading 3.1) (see Note 9).
Wash BM from plates; pellet and count cells. Resuspend BM
cells at 4 × 106 cells in 4 mL of viral cocktail/well (six-well
plate) (see Note 10). Add 2 μg/mL of polybrene in each well.
Spin infect BM plates at 1000 × g for 90 min at 32 °C (see
Note 11), and then incubate 4 h at 37 °C. Remove carefully
the viral cocktail, and replace with 4 mL of BM medium.
Incubate overnight at 37 °C.
4. Day 3 morning: BM transduction #2. Repeat the steps of Day 2.
5. Day 4: FACS sorting. Sort the BM cells positive for transgene
expression using non-transduced BM as a control. Count cells
and resuspend in PBS for injection.
6. Day 5: BM transfer. Lethally irradiate the recipient mice (see
Note 12). Inject irradiated mice intravenously with at least
1 × 105 transduced cells.
7. Monitor mice for recovery. Chimeras can be used as a source
of naive T cells 6–8 weeks after BM reconstitution.
3.3 Adoptive T Cell 1. Collect spleen and lymph nodes from a chimeric mouse. Mash
Transfer organs through a 70 μm cell strainer in complete RPMI.
2. Purify CD8+ T cells with negative CD8+ T cell isolation kit
according to manufacturer’s instructions.
3. Optional (see Note 13): cell staining with SNARF. Pellet cells,
resuspend at 2 × 107 cells/mL in PBS + dye (5 μM), incubate
12 min at 37 °C, and wash twice with PBS + 10% FCS (with at
least twice the volume).
4. Resuspend cells in PBS for injection.
5. Inject intravenously between 1 and 10 × 106 cells. Let the cells
home to the lymph node 2–4 h before imaging.
3.4 Preparation 1. Anesthetize the mouse by injecting 100 μL of anesthetic intra-

of the Mouse peritoneally (see Note 14).
for Imaging 2. Optional: Cannulate the tail vein for peptide injection. Heat
the mouse tail (with a heating lamp or by dipping the tail into
warm water ~30–35 °C) to induce vasodilation of the tail vein.
Fill the cannula with heparin solution. Slightly tilt the needle
to insert it parallel to the tail into the vein (see Note 15). Once
the needle is in the vein, fix the cannula with a drop of glue and
surgical tape. Prepare a syringe with antigenic peptide solution,
and stick the needle in the cannula. Fix it with surgical tape
(see Note 16).
3. Expose popliteal LN. For this, lay the mouse on its belly and
immobilize its leg with surgical tape on the footpad. Cut a
square of skin out from the ankle to the middle of the thigh.
The popliteal lymph node is located in the back of the knee,
along a small blood vessel. Gently remove the fat behind the
knee until the lymph node becomes apparent (see Note 17).
To check your preparation, apply a glass coverslip on top of the
leg: the lymph node should stay clearly visible.
4. Transfer the mouse on the heating stage. Humidify a stripe of
plaster and place it below the mouse leg (Fig. 1b). Fold the
plaster to immobilize the leg, and create a flat surface (Fig. 1c
and d) to fix the ring-shaped metallic tube (see Note 18). Let
the plaster solidify slightly. Glue it to the heating stage. Put a
drop of PBS on the lymph node (see Note 19). Add a drop of
glue on the plaster cast to fix the ring coverslip (Fig. 1e and f).
3.5 Live Imaging 1. Install the stage on the microscope (Fig. 1g). Pour water in the
of Synapses ring to immerse the objective.
and Kinapses 2. Start imaging of the region of interest. Depending upon the
in the Lymph Node microscope and objective used the field of view, axial step size,
number of images per stack, and time required to complete an
image stack will vary. Typically, a cycle time of 20–30 s is the
maximum for tracking fast-moving lymphocytes. Record a
30 min to 1 h control movie (steady state).
3. Inject peptide intravenously to induce antigen recognition
(see Note 20). Using peptides of varying affinities enables to
promote the formation of synapse or kinapse (see Note 21).
Pursue imaging for 30 min to 1 h.
3.6 Analysis of T Cell 1. Tracking can be performed with Imaris (Fig. 2). The “surface
Contacts tracking” option should be chosen. Tracks can be corrected
manually at the end of the analysis.
2. Export tracking data. This may include the classical migration
parameters (cell coordinates, speed, straightness), as well as cell
morphology parameters (sphericity) and fluorescence
parameters.
3. Classifying cells according to average speed enables to distin-
guish various typical T cell behaviors in the lymph node: freely
migrating (> 5 μm/min), kinapse formation (<5 μm/min and
>2.5 μm/min), and synapse formation (<2.5 μm/min) (see
Note 22).
4. Surface tracking of cells based on the cell tracker dye signal allows
creating new channels of “masked fluorescence” with Imaris.
Fig. 2 Tracking of T cells during synapse and kinapse formation. (a) Representative two-photon images and
overlaid T cell trajectories (corresponding to 5 min of imaging) in uninjected mice or mice receiving the Q4 or
N4 peptide. (b–d) Average cell speed (μm/min) (b), arrest coefficient (percentage of time during which a cell
exhibited an instantaneous speed <2 μm/min) (c), and straightness (d) are graphed. Reproduced from Moreau
et al. 2015 [16] with permission of PNAS
This strategy enables to remove the background fluorescence

for the protein of interest and only get the signal emanating
from the cells.
4 Notes
1. Plat-E cells [14] are adherent cells. On thawing, they should

be placed in culture in complete DMEM to recover. From the
next day on, they should be cultivated in flasks with Plat-E
medium containing selection antibiotics (puromycin, blastici-
din) and be passed every 2–3 days. When passing, detach with
trypsin. They should be thawed at least 1 week before being
used for transfection.
2. Always use polystyrene FACS tube for preparing the transfec-
tion mix. Do not use polypropylene.
3. To ensure good distribution of the mix, start at the periphery
of the plate and add the mix to the cells drop by drop, turning
clockwise. Repeat the same procedure following a smaller

circle inside until reaching the center. Add the few drops left
randomly.
4. Viral particles half-life is higher at 32 °C, while 37 °C is opti-
mal for Plat-E cells. In our hands, performing the viral produc-
tion at 32 °C or at 37 °C did not substantially modify
subsequent BM transduction efficiency.
5. Transfected Plat-E cells can be discarded or can be harvested
and analyzed by FACS for transfection efficiency.
6. 5-FU eliminate bone marrow differentiated cells enriching for
hematopoietic stem cells.
7. Use approximately ten (6–10 weeks old) mice for each trans-
duction. We use OT-1 mice (Rag competent) as donor mice
as they gave better results than OT-1 Rag−/− mice for
reconstitution.
8. If using less than 15 mice, it is usually fine to plate the BM in
one 100 mm dish. If using more than 20 mice, it is preferable
to plate in two dishes.
9. It is possible to use viral supernatant from frozen stocks, but
the efficiency of transduction is reduced.
10. Keep a small aliquot of BM cells (5 × 105) in 1 mL of BM
medium in a 24-well plate for use as a non-transduced control
during FACS sort.
11. 32 °C is the temperature that increases the most the viral par-
ticles half-life and therefore increases the efficiency of BM spin
infection.
12. For lethal irradiation of mice, follow the guidelines of your
irradiator. Splitting the irradiation in two suboptimal ones (4 h
apart) increases the survival of reconstituted mice.
13. Staining with vital dyes can be used to facilitate analysis, for
example, in order to detect the entire T cell. If using GFP for
the protein of interest, SNARF is an appropriate compatible
dye.
14. To keep the mouse anesthetized, we recommend to reinject
between 30 and 50 μL of anesthetic intramuscularly every
hour after the first injection.
15. To check if the needle is in the vein, try to inject 10 μL of the
heparin solution: it should flow freely without any resistance.
16. It is easier to finish the preparation of the mouse first and only
put the syringe with antigenic peptide at the end (just before
going to the microscope for imaging).
17. Adding a drop of PBS on the leg helps to distinguish the dif-
ferent tissues: fat appears white and shining, muscle slightly
pink, and lymph node white and round.
18. To test that the plaster cast preparation is flat and properly
adjusted to the lymph node, apply a glass slide on top and
check that the lymph node is still clearly visible.
19. The PBS should be in contact with the coverslip for efficient
imaging. When fixing the coverslip, it is of crucial importance
to avoid pressing on the lymph node. Too much pressure will
strongly alter cell motility. The small blood vessel next to the
lymph node should be visible: if it disappears because it col-
lapses, this is an indication that the pressure is too high.
20. If cannulation has been performed (in the tail or in the jugular
vein), injection can be made during imaging. If not, it is also
possible to simply stop image acquisition, inject the peptide
retro-orbitally, and then immediately resume imaging. This
procedure should only take 1–3 min.
21. With OT-1 T cells, injection of high affinity peptide N4 trig-
gers primarily the formation of synapses, while injection of the
low affinity peptide V4 triggers kinapses. Intermediate affinity
peptide Q4 promotes mixed behavior [5].
22. Classification based on speed has been shown to correlate with
other parameters typical of synapses and kinapses such as
straightness [5]. The threshold for synapse has been set to 2.5
μm/min because minor movement of the tissue during imag-
ing can produce an apparent speed of a few μm/min even for
cells that are completely arrested.
Acknowledgment
This work was supported by INSERM, Institut Pasteur, and a

starting grant from the European Research Council
(LymphocyteContact). HDM was funded by the Association pour
la Recherche sur le cancer (ARC).
References
1. Miller MJ, Wei SH, Parker I, Cahalan MD 4. Skokos D, Shakhar G, Varma R, Waite JC, TO
(2002) Two-photon imaging of lymphocyte C, Lindquist RL, Schwickert T, Nussenzweig
motility and antigen response in intact lymph MC, Dustin ML (2007) Peptide-MHC potency
node. Science 296(5574):1869–1873 governs dynamic interactions between T cells
2. Celli S, Lemaitre F, Bousso P (2007) Real-time and dendritic cells in lymph nodes. Nat
manipulation of T cell-dendritic cell interac- Immunol 8(8):835–844
tions in vivo reveals the importance of pro- 5. Moreau HD, Lemaitre F, Terriac E, Azar G,
longed contacts for CD4+ T cell activation. Piel M, Lennon-Dumenil AM, Bousso P
Immunity 27(4):625–634 (2012) Dynamic in situ cytometry uncovers T
3. Dustin ML (2008) T-cell activation through cell receptor signaling during immunological
immunological synapses and kinapses. Immunol synapses and kinapses in vivo. Immunity
Rev 221:77–89 37:351–363
6. Gunzer M, Schafer A, Borgmann S, Grabbe S, signaling dynamics. J Exp Med 207(12):

Zanker KS, Brocker EB, Kampgen E, Friedl P 2733–2749. doi:10.1084/jem.20091201
(2000) Antigen presentation in extracellular 12. Azar GA, Lemaitre F, Robey EA, Bousso P
matrix: interactions of T cells with dendritic (2010) Subcellular dynamics of T cell immuno-
cells are dynamic, short lived, and sequential. logical synapses and kinapses in lymph nodes.
Immunity 13(3):323–332 Proc Natl Acad Sci U S A 107(8):3675–3680
7. Faroudi M, Zaru R, Paulet P, Muller S, Valitutti 13. Garrod KR, Moreau HD, Garcia Z, Lemaitre
S (2003) Cutting edge: T lymphocyte activa- F, Bouvier I, Albert ML, Bousso P (2012)
tion by repeated immunological synapse forma- Dissecting T cell contraction in vivo using a
tion and intermittent signaling. J Immunol genetically encoded reporter of apoptosis. Cell
171(3):1128–1132 Rep 2(5):1438–1447
8. Munitic I, Ryan PE, Ashwell JD (2005) T cells in 14. Morita S, Kojima T, Kitamura T (2000) Plat-E:
G1 provide a memory-like response to second- an efficient and stable system for transient
ary stimulation. J Immunol 174(7):4010–4018 packaging of retroviruses. Gene Ther 7(12):
9. Celli S, Garcia Z, Bousso P (2005) CD4 T cells 1063–1066
integrate signals delivered during successive 15. Celli S, Bousso P (2007) Intravital two-pho-
DC encounters in vivo. J Exp Med 202(9): ton imaging of T-cell priming and tolerance in
1271–1278 the lymph node. Methods Mol Biol 380:
10. Bousso P (2008) T-cell activation by dendritic 355–364
cells in the lymph node: lessons from the mov- 16. Moreau HD, Lemaitre F, Garrod KR, Garcia
ies. Nat Rev Immunol 8(9):675–684 Z, Lennon-Dumenil AM, Bousso P (2015)
11. Friedman RS, Beemiller P, Sorensen CM, Signal strength regulates antigen-mediated
Jacobelli J, Krummel MF (2010) Real-time T-cell deceleration by distinct mechanisms to
analysis of T cell receptors in naive cells in vitro promote local exploration or arrest. Proc Natl
and in vivo reveals flexibility in synapse and Acad Sci U S A 112(39):12151–12156
Chapter 36
Studying Dendritic Cell-T Cell Interactions Under In Vivo

Conditions
Nicholas van Panhuys
Abstract
The interactions between dendritic cells and T cells during the initiation of an adaptive immune response
underpins one of the most important checkpoints in ensuring that the correct immune response is induced,
in order to direct the development of immune tolerance or to mount an immune response against patho-
genic organisms. The advent of two-photon intravital imaging allows us to directly study the interactions
between dendritic cells and T cells as they occur under physiological conditions, greatly improving our
understanding of the dynamic relationship that occurs during T cell activation and expanding our knowl-
edge of how the adaptive immune system is regulated during both priming and memory responses. Here,
we describe a technique for the in vivo analysis of the interactions between either CD4+ or CD8+ T cells
and dendritic cells as they occur in the popliteal lymph nodes of live mice. The adoptive transfer of both
dendritic cells and T cells is utilized here in order to allow investigators to control the time point of interac-
tion being analyzed, the activation state and amount of antigen presented by the dendritic cell, and the
maturation/differentiation state of the interacting T cell.
Key words Dendritic cell, T cell, Two photon, Intravital, Immune synapse, Popliteal lymph node
1 Introduction
1.1 Dynamics The advent of intravital microscopy and its application to the study
of Dendritic Cell-T Cell of the immune system has led to a rapid expansion of our knowl-
Interactions edge of how the immune system operates. In particular, the study
of T cell activation under physiological conditions using intravital
imaging has profoundly altered the way we envisage the interac-
tions between dendritic cells and T cells that drive activation, divi-
sion, and differentiation [1, 2]. Prior to the application of intravital
imaging techniques, investigators were limited to studying the
dynamic behaviors of cells in either 2D (plates or slides) and 3D
systems (collagen or matrigels) or tissue explant models [3]. While
each of these model systems has provided valuable knowledge,
they are unable to fully replicate features of intact lymph nodes
(LN), such as the provision of migrational cues, the ability of cells
569
570 Nicholas van Panhuys
to migrate to/from LN, the lymphatic input/drainage, and the

maintenance of a homeostatic environment that allows imaging to
be conducted for extended periods of time. The imaging of intact
lymph nodes has been made possible by the application of laser
scanning microscopy through the use of either confocal or two-
photon (2P) excitation techniques which allow for the excitation
of fluorophores in highly defined optical sections and the exclusion
of out-of-focus light. The use of single-photon confocal laser scan-
ning microscope (LSM) has proved to be useful in tissues at <100
μm of depth. However, the relatively high energy of the lower
wavelengths of light utilized (400–700 nm) leads to both increased
levels of photobleaching and phototoxicity, limiting the length of
time samples can be analyzed for. Alternatively, as 2P excitation
requires the exposure of fluorophores to two coincident photons
at an almost synchronized point in time, this specifically excites
fluorophores present only at the LSM focal point allowing lower
energy wavelengths (800–1400 nm) to be utilized. Additionally,
this feature ensures that no out-of-focus light is generated, signifi-
cantly reducing the background and allowing for the imaging of
tissues at much greater depths (0–500 μm).
The application of LSM techniques to the study of T cells has
greatly enhanced our appreciation for the extremely dynamic nature
of these cells by allowing us to add a temporal component to our
investigations, extending the analysis of T cells in their resident tis-
sues beyond just the description of morphological characteristics and
anatomical localization. Visualization of T cells in vivo reveals that
they are in an almost constant state of migration as they travel
through the immune system scanning antigen presenting cells
(APCs) for their cognate antigen. CD4+ T cells have been shown to
migrate through peripheral lymph nodes with a mean residence of
12 h, as compared to 21 h for CD8 T cells [4, 5]. As they migrate
through lymph nodes at approximately 10 μm/min, T cells con-
stantly form brief non-cognate interactions with APC, the majority
of which last 2–7 min, followed by a rapid dissociation and migra-
tion away from the APC enabling approximately 5000 T cells to scan
an individual APC per hour [6, 7]. Initial studies analyzing T cell
activation in vivo have revealed that after the subcutaneous delivery
of antigen, LN-resident T cells exhibited an abrupt halt in migration
and formed extended cognate interactions with APC [6]. Later
studies, built on this system by use of the adoptive transfer of
Ag-loaded dendritic cell (DC), revealed a program of activation that
could largely be broken down into three distinct dynamic phases. A
primary phase (~0–6 h) consisting of kinaptic interactions, in which
T cells make transient contact that are nonetheless influenced by
antigen recognition. A secondary phase (~6–18 h) consisting of
long-lived interactions, indicative of the formation of stable immune
synapses. This is followed by a tertiary phase (~18–30 h) consisting
of repetitive intermediate length contacts and is typified by a swarming
Studying Dendritic Cell-T Cell Interactions Under In Vivo Conditions 571
behavior [8, 9]. These interactions, and the transition between

phases, have been shown to be dependent on multiple factors includ-
ing the quality or quantity of the cognate antigen and both the phe-
notype and activation status of the stimulatory APC.
Here, an experimental methodology for the analysis of T cell
and APC interactions is described, which can be utilized to assess
encounters during either the initial priming of naïve T cells or dur-
ing the activation of a memory response. While the basic protocol
remains identical, the time points analyzed can be set by the
researcher to study the phase of response that best suits their
experimental question, i.e., from the time that the T cells are ini-
tially recruited to the LN to 3–4 days post-transfer (or longer if
non-dividing cells or fluorescent protein-labeled cells are being
studied). While the use of Ag-loaded DC is described here, this
protocol can be readily adjusted to study T cell activation in
response to subcutaneously injected Ag or an infectious model sys-
tem. Variations on this basic protocol have also been used to assess
other factors associated with T cell-APC interactions by studying
Ca2+ flux analysis [10], the translocation of transcription factors
[11], and the use of dynamic in situ cytometry [12]. Results can
also potentially be further extended by use of intravital dynamics-
immunosignal correlative microscopy, whereby tissues can be fixed
immediately after imaging and histological features used to realign
the observed dynamic T cells with cells present in static sections
allowing for the further analysis of the cellular phenotype [13].
2 Materials
2.1 Mice All work with mice should be done in accordance with local laws
and regulations regarding ethical treatment of animals.
1. Wild-type or transgenic animals as a source of dendritic cells
(see Note 1).
2. Animals with a transgenic T cell receptor (see Note 2).
3. Wild-type animals as a source of control T cells (see Note 3).
2.2 Preparation 1. Tools for dissection: dissection board, pins, forceps, and
of DCs for Transfer scissors.
2. Spray/squirt bottle containing 70% ethanol.
3. Six-well plate.
4. Complete-RPMI (cRPMI; 2 mM l-glutamine, 50 U/ml peni-
cillin and 50 μg/ml streptomycin, 10% fetal calf serum (FCS),
50 μM 2ME).
5. Liberase DL at 13.1 U/ml.
6. 3 ml syringe and 21 G needle.
7. Cell culture incubator at 37 °C with 5% CO2.

8. 70 or 100 μm nylon mesh cell strainers.
9. 3 ml syringe plunger.
10. 50 ml Falcon tube.
11. Refrigerated centrifuge.
12. MACS buffer (PBS, 1% FCS, 2 mM Ethylenediaminetetraacetic
acid, EDTA).
13. Mouse CD11c positive selection microbeads (Miltenyi).
14. LS separation columns.
15. Magnets and cell sorter stand for manual separation or auto-
matic cell sorting device.
16. 15 ml conical tube.
17. Cell counter and trypan blue.
18. 24-well plate.
19. Adjuvant for DC activation (see Note 4).
20. Peptide for DC loading (see Note 5).
21. PBS.
22. 10 mM CellTracker Blue (e.g., CMF2HC, Thermo Fisher) in
Dimethyl sulfoxide, DMSO.
23. 31 G insulin needle.
2.3 Preparation of T 1. Tools for dissection: dissection board, pins, forceps, and
Cells for Transfer scissors.
3. 50 ml conical tubes.
4. Complete-RPMI (cRPMI; 2 mM l-glutamine, 50 U/ml peni-
cillin and 50 μg/ml streptomycin, 10% fetal calf serum (FCS),
50 μM 2ME).
5. 3 ml syringe plungers.
6. 70 μm or 100 μm nylon mesh cell strainers.
7. Refrigerated centrifuge.
8. ACK lysis buffer.
9. MACS buffer (PBS, 1% FCS, 2 mM EDTA).
10. Mouse CD4 or CD8 positive selection microbeads (Miltenyi
or other preferred brand).
11. LS separation columns.
12. Magnets and cell sorter stand for manual separation or auto-
matic cell sorting device.
13. 15 ml conical tubes.
14. Cell counter and trypan blue.
15. PBS.
16. 10 mM CellTracker Green (CMFDA, Thermo Fisher) in
DMSO or other dye.
17. 10 mM CellTracker red (CMTPX, Thermo Fisher) in DMSO
or other dye.
18. Pluronic.
19. Incomplete RPMI.
20. 31 G insulin needle.
2.4 Popliteal Lymph 1. Tools for dissection: scissors, forceps, microscissors (i.e.,
Node Imaging Vannas Spring Scissors—3 mm cutting edge, Fine Science
Tools), ultrafine forceps (i.e., Dumont #5, Fine Science Tools).
2. Isoflurane anesthesia system.
3. Surgical tape.
4. Dissecting microscope.
5. Heating pad.
7. Electric hair clippers for shaving animals.
8. Depilatory cream for hair removal (i.e., Nair).
9. Cotton buds.
10. Imaging stage (see Fig. 1).
11. Gauze pads.
12. Phosphate buffered saline, PBS warmed to 37 °C.
13. 12 ml syringe for application of PBS.
14. Vetbond glue.
15. Glass coverslip.
16. Vacuum grease.
17. 12 ml syringe.
18. Two-photon laser scanning microscope fitted with epifluores-
cence illumination source, opaque environmental chamber,
and high numerical aperture (NA = 1.0) water immersion
objective (either ×20 or ×25 magnification).
19. NDD light path setup, as per manufacturer recommendation
with 470/30 filter for CMF2HC photon collection, 525/50
for CMFDA photon collection, and 600/75 for CMTPX
collection. Optional, 400/75 filter for second harmonic col-
lection on four channel microscopes.
2.5 Image Analysis 1. Software for analysis of data. Commercial sources include
and Quantification MetaMorph (Molecular Devices), Imaris (Bitplane), and
of Interactions Volocity (PerkinElmer). Open-source software includes pro-
grams such as ImageJ, BioImageXD, Icy, and Fiji.
Fig. 1 Stage design for popliteal lymph node preparation. Use 2 mm ferritic stainless steel for the base of the
stage, non-stainless steel will rust rapidly due to the use of PBS, and ferritic steel will allow the attachments
of small magnets for holding the lymph node preparation in place. Secure 5 mm Perspex top, precut to the
dimensions indicated using strong adhesive glue. For bolts labeled (A), cut a small slot through their diameter
to allow the insertion of a 23 G syringe needle (used to secure the thigh). For bolts labeled (B), use bolts with
small washers to allow the secure attachment of the coverslip holder. Coverslip holder should be made from
aluminum to avoid altering the position of the magnets while imaging and to prevent rusting
3 Methods
3.1 Preparation 1. Euthanize animals according to the ethical guidelines govern-

of DCs for Transfer ing the use of animals in your laboratory (see Note 6).
3.1.1 Spleen Isolation 2. Place animal on dissection board in a supine position, and ster-
ilize the mouse and dissection equipment with ethanol, before
making a left paracostal incision of approximately 2 cm.
3. Locate the spleen and use forceps to gently remove it into a
well of a six-well plate, containing 4 ml of digest media (cRPMI
and 125 μl Liberase DL at 13.1 U/ml, final concentration 0.4
U/ml. Up to five spleens can be digested per well).
4. Carefully inject 3 ml of digest media into each spleen ensuring
that the excess solution is collected back into the well contain-
ing the spleens.
5. Incubate spleens for 20 min in cell culture incubator.
6. Place spleens into a cell strainer atop a 50 ml tube, use 3 ml
plunger to gently dissociate spleen tissue, and wash into 50 ml
tube using cRPMI.
7. Centrifuge at 350 × g for 5 min.
8. Resuspend pellet in 400 μl MACS buffer and 90 μl CD11c
beads per spleen.
9. Incubate for 15 min at 4 °C (fridge). Add 20 times volume of
MACS buffer and re-strain.
10. Wash via centrifugation at 350 × g for 5 min and proceed to

MACS separation as per manufacturer’s instructions.
11. To enhance CD11c cell purity, pass over a second MACS
column.
3.1.2 Antigen Loading 1. Count cells, wash via centrifugation at 350 × g for 5 min, and
and DC Stimulation resuspend at 2 × 106 cell/ml. Plate out at 2 ml/well in 24-well
plates.
2. Add relevant adjuvant for DC activation and peptide for DC
loading at desired concentrations (see Notes 4 and 5).
3. Incubate cells for 4 h in cell culture incubators.
3.1.3 Cell Staining 1. Pool wells and count cells. Wash cells in cold PBS via centrifu-
and Adoptive Transfer gation at 350 × g for 5 min.
of DCs 2. Prepare CellTracker Blue (CMF2HC) stain. From 10 mM
stock CMF2HC to a final concentration of 200 μM in PBS,
add Pluronic at 1:10 of amount of CMF2HC used. Use 1 ml
stain per 2 × 106 cells.
3. Incubate cells with CMF2HC in cell culture incubators for
20 min.
4. Wash cells in cRPMI × 2 via centrifugation at 350 × g for 5 min.
5. Incubate cells in cRPMI in cell culture incubators for 20 min.
6. Wash in PBS via centrifugation at 350 × g for 5 min.
7. Count and wash cells in PBS via centrifugation at 350 × g for
5 min.
8. Resuspend cells at 1–2 × 106 per 25 μl in PBS.
9. Adoptively transfer CMF2HC stained CD11c+ DC into the
right rear footpad of host animals using a 31 G insulin needle.
3.2 Preparation of T 1. Euthanize animals according to the ethical guidelines governing
Cells for Transfer the use of animals in your laboratory (see Notes 2, 3, and 6).
3.2.1 Tissue Isolation 2. Pin animal to dissection board in a supine position; sterilize the
for TCR Tg and Control T mouse and dissection equipment with ethanol before making
Cells requisite incisions required to locate the spleen, mesenteric,
axial, and brachial lymph nodes; and use forceps to gently
remove them into a well of a six-well plate, containing cRPMI.
3. Place lymph nodes and spleens into a cell strainer atop a 50 ml
tube, use 3 ml plunger to gently dissociate tissues, and wash
into the 50 ml tube using cRPMI.
4. Centrifuge at 350 × g for 5 min.
5. Lyse red blood cells in ACK solution or similar lysis solution.
For ACK usage resuspend cells in ACK for 5 min at room
temperature.
6. Wash in MACS buffer by centrifugation at 350 × g for 5 min.

7. Resuspend pellet in 90 μl MACS buffer and 10 μl anti-CD4 or
anti-CD8 beads per 1 × 107 cells.
8. Incubate for 15 min at 4 °C (fridge). Add 20 times volume of
MACS buffer and re-strain.
9. Wash via centrifugation at 350 × g for 5 min and proceed to
MACS separation as per manufacturer’s instructions.
10. To enhance cell purity, pass over a second MACS column.
3.2.2 Cell Staining and 1. Wash cells in cold PBS via centrifugation at 350 × g for 5 min.
Adoptive Transfer of T Cells 2. Prepare CellTracker Red (CMTPX) and CellTracker Green
stain (CMFDA) from 10 mM stocks, CMTPX to a final con-
centration of 1.25 μM (1:8000) in PBS, and add Pluronic at
1:1 of amount of CMTPX used. CMFDA to a final concentra-
tion of 1.0 μM (1:10,000) in PBS, add Pluronic at 1:1 of
amount of CMFDA used. Use 1 ml stain per 4 × 106 cells.
3. Incubate TCR Tg T cells with CMTPX and control T cells
with CMFDA in cell culture incubators for 15 min (or vice
versa).
4. Wash cells in incomplete RPMI × 2 via centrifugation at 350 × g
for 5 min.
5. Incubate cells in incomplete RPMI in cell culture incubators
for 20 min.
6. Wash in PBS via centrifugation at 350 × g for 5 min.
7. Count and wash cells in PBS via centrifugation at 350 × g for
5 min.
8. Resuspend cells at 2 × 106 per 50 μl in PBS.
9. Adoptively co-transfer 2 × 106 TCR Tg T cells and 2 × 106
control T cells into host animals 18 h post-transfer of DCs
(see Note 7).
3.3 Preparation 1. Perform initial induction of anesthesia in a chamber at 2–4%

of Popliteal Lymph isoflurane in 100% oxygen, and then secure anesthetic nose
Node for Intravital cone to the animal’s head with surgical tape and maintain anes-
Imaging thesia with 1–2% isoflurane in 100% oxygen. Monitor respira-
tion rates to maintain deep anesthesia.
2. Use electric hair clippers to shave the hair on the right hind leg
of the animal (Fig. 2a).
3. Use cotton buds to apply depilatory cream to the shaved area
for 5 min, and thoroughly wash area with paper towels and
water to remove hair and excess cream (Fig. 2b).
4. Transfer animal to imaging stage in a prone position, and
secure nose cone to the stage. Use surgical tape and Vetbond
Fig. 2 Surgical preparation of animal for imaging. (a–g) As described in Subheading 3.3 Preparation of Popliteal
Lymph Node for Intravital Imaging
to secure the right footpad to the base of the imaging stage.

Use surgical tape to secure the left leg and tail to the left side
of the stage (Fig. 2c and d).
5. Use a 23 G needle to secure the leg to the stage via holes in
bolts (refer to bolts labeled A in Fig. 1), and place the small
magnets on either side of the leg to stabilize the tissue sur-
rounding the popliteal lymph node (Fig. 2e).
6. Place the animal and stage onto a warming pad located under
a dissecting microscope. Sterilize the skin with ethanol and
make a small incision in the middle of the back of the right
thigh (Fig. 2f).
7. Locate the lymph node in the popliteal fossa by using the ultra-
fine forceps to carefully separate the covering tissue layer by
layer.
8. Place a layer of gauze on top of the magnets on each side of the
leg and moisten using PBS. Gauze pads should be kept moist-
ened during imaging to prevent tissue dehydration and to help
maintain stability of the LN preparation (Fig. 2g).
9. Fill 12 ml syringe barrel with vacuum grease, and apply a thin
layer along all four edges of a coverslip, attach to coverslip
holder, and slot onto the stage. Moisten the area surrounding
the LN and lower the coverslip so that it is in contact with the

LN. Avoid applying too much pressure, as this will inhibit cell
migration (Fig. 2h and i).
3.4 Acquisition 1. Pre-warm the environmental chamber on the two-photon

of Two-Photon microscope to 37 °C.
Intravital Images 2. Add H2O to the top of the reservoir on the coverslip created
with the vacuum grease.
3. Using the epifluorescent light source, locate and focus the
water-dipping lens on the LN.
4. Tune the laser to the appropriate wavelength (Fig. 3) to excite
the fluorophores used to stain the cells, and locate area of
interest to be imaged. Adjust laser intensity to the minimum
power required to illuminate the cells of interest in order to
avoid phototoxicity (see Note 8).
Fig. 3 Experimental design for multicolor two-photon imaging. (a) Normalized emission intensities for the dyes
outlined in the experimental procedures dependent on the excitation wavelength of the two-photon laser used.
In order to maximize the collection of photons from all three dyes, a two-photon excitation wavelength of
770–800 nm should be used. However, it should be noted that as the excitation wavelength decreases, the
background may increase, thus the optimal wavelength should be determined for each preparation. (b)
Comparison of the emission wavelength spectra of the dyes used in the experimental procedure. Dyes are
selected to minimize overlap in spectral emissions. (c) Excitation and emission maximal values for common
dyes and fluorescent proteins used in two-photon imaging
5. Set imaging parameters: Z-stack approximately 40–80 μm with

a distance between Z-layers of 3 or 4 μm. Time between
Z-stacks should be set to 15–45 s, and the time interval should
be minimized by adjusting the scan speed and size of Z-stack
in order to attain data allowing accurate tracking of individual
cells. Adjust digital zoom to an appropriate level to maximize
number of interactions vs. degree of detail (×1.0–2.0).
6. Acquire images. During imaging the animal’s breathing rate,
gauze pads, and reservoir should all be checked every 30 min
to ensure proper anesthesia is maintained, LN is kept moist,
and the reservoir does not dry out.
7. Subsequent to the experimental endpoint, euthanize animals
according to the ethical guidelines governing the use of ani-
mals in your laboratory.
3.5 Analysis 1. Analysis of interaction lengths between DCs and T cells can be
of Intravital Images measured either by direct or indirect methods. Direct methods
and Quantification involve either manual measurement of the length of interac-
of Interactions tions or automated methods [14, 15], which rely on measur-
ing interactions by assessing the colocalization of voxels from
cells with alternately colored stains. For these automated
methods, contact times between cells need to be specifically
validated for each individual data set being quantified.
2. Direct manual measurement: control and TCR Tg cells tracked
using software-driven algorithms.
3. DC present in imaged volume should be mapped and num-
bered for future reference (Fig. 4b).
4. To determine DC interaction history, each DC should be
assessed on a per cell basis to determine whether a specific T
cell is in contact with the DC during each of the time points
imaged (Fig. 4c).
5. To determine T cell interaction history, each T cell should be
assessed on a per cell basis to determine whether a specific DC
is in contact with the T cell during each of the time points
imaged, by following the previously calculated migratory path
of the cell.
6. Indirect methods of measuring alterations in cellular interac-
tion dynamics between populations can be applied by analyz-
ing the migratory paths of cells [16] as determined in 3.2.2.
–– Cellular velocity: when T cells interact with DC, their cel-
lular velocity tends to decrease to <2 μm/s. Therefore on
a population level, a decrease in the mean speeds of T cells
being imaged correlates with an increase in interactions.
–– Arrest coefficient: as cellular velocity <2 μm/s correlates
with a halt in T cell migration, the arrest coefficient is
Fig. 4 Analysis of T-DC interactions. (a) Anatomy of the popliteal lymph node, subcapsular sinus (SNS), interfol-
licular zone (IFC), B cell zones (BCZ), T cell zone/paracortex (TCZ), medullar (Med). B cells CD19—purple. High
endothelial venules CD34—yellow. Medullary sinus’ collagen IV—white. Transferred DC CMf2HC—blue.
Transferred CD4+ T cells CMTPX—red. (b) Reference map of DC present in the interfollicular zone of the pop-
liteal lymph node during intravital imaging. (c) Manual tracking of an interaction between an antigen-specific
CD4+ T cell and an antigen-loaded DC during intravital imaging
calculated from the fraction of time where a cells’ velocity

is <2 μm/s.
–– Confinement ratio: calculated as the ratio of the total dis-
tance traveled by a cell to the displacement from its initial
point of measurement (also, referred to as the straightness
index or meandering index).
4 Notes
1. This protocol describes the isolation of DC from WT animals

and the subsequent staining with a fluorescent dye to allow
visualization by two-photon excitation. However, depending
on the model system being designed, DC expressing a fluo-

rescent protein marker such as the CD11c-YFP strain can be
utilized [17]. Care should be taken to design an experimental
setup whereby fluorescent markers for all cell types being
studied can be optimally illuminated at a single wavelength
(see Fig. 2).
2. In order to study the interactions between T cells and DC, a
TCR Tg strain recognizing its cognate antigen presented by a
corresponding haplotype matched DC must be selected. This
protocol is adaptable to either the study of CD4+ or CD8+ T
cell interactions dependent on the TCR-Ag system utilized.
Common previously studied systems include pigeon cyto-
chrome C peptide (pPCC) for analysis of 5CC7 CD4+ T cell
activation [10] and OTI TCR Tg CD8 T cell activation by
ovalbumin (Ova) peptide [18]. It should also be noted that in
order to analyze cells after they have entered into division,
TCR Tg animals should be crossed onto backgrounds which
express fluorescent protein markers, as cellular division will sig-
nificantly decrease the ability to detect fluorescent dyes.
3. Wild-type control T cells either allogenically or congenically
matched to the host are an essential component of the experi-
mental setup, as they allow for the study of a population of
non-antigen-specific T cells, which do not engage in cognate
interactions with Ag-loaded DC and thereby act as a reference
population. The study of the behavior of this reference popula-
tion allows for the comparison of results obtained between
multiple intravital experiments and critically allows us to deter-
mine whether the lymph node preparation is of a satisfactory
standard by analyzing the mean migratory velocity of the con-
trol population (mean migrational velocity should be 8–12
μm/min).
4. Depending on the experimental hypothesis, an adjuvant may
be required to induce a specific program of DC activation. We
have previously noted that activation of DC in the presence of
adjuvants for 4 h was sufficient to induce a pattern of differen-
tial activation [10]. Alternately, other studies have been con-
ducted where the adjuvant was injected subcutaneously at the
time of adoptive transfer. However, it should be taken into
account that this will additionally lead to the activation of APC
present in the draining LN. It should also be noted that the
in vitro culture of DC required to load the DC with peptide
leads to a spontaneous activation as observed by an increase in
the expression of MHC class II, CD80, and CD86 [10, 18].
5. As noted above, in order to study T cell-DC interactions, DC
must be loaded with a cognate peptide. It has previously been
determined that optimal loading of MHC molecules present
on the surface of DC occurs within 2–4 h. However, as DC

takes at least 12 h to migrate to the draining LN, the level of
peptide presentation will be decreased in comparison to that
observed during in vitro experiments [18].
6. When considering the number of animals to use in order to
obtain the required number of cells for adoptive transfer, we
have found that typically 1–2 × 106 CD11c+ DCs are obtained
per spleen, 5–10 × 106 CD4 or CD8 per WT animal, and 2–4
× 106 CD4 or CD8 per TCR Tg animal after processing.
7. In order to image the initial interactions between T cells and
DC immediately post-transfer, it is recommended to perform
the surgical preparation and then transfer T cells by retro-
orbital injection, as the administration of isoflurane makes it
difficult to perform i.v. injections. If later time points are to be
studied, it is recommended to perform adoptive transfer by i.v.
injection to minimize stress to the animal. Additionally, in
order to synchronize the phase of interactions being studied
and restrict the analysis to the initial cohort of cells, anti-
CD26L Ab can be injected at 2 h post-transfer to inhibit the
recruitment of cells at later time points.
8. It is important to note that significant tissue “settling” may
occur during the first 30 min following the surgical prepara-
tion due to the relaxation of surrounding tissue and acclimati-
zation to the imaging chamber. It can be useful to prepare the
LN in advance and observe the stability of the LN in terms of
X, Y, Z stability during this time period. This is especially criti-
cal if cells are to be observed immediately after adoptive trans-
fer, as T cells will migrate into the LN within 5–10 min
post-transfer and any significant drift in the imaging field will
impede downstream analysis.
Acknowledgments
We would like to thank Menna Clatworthy and Andrea Radtke for

their helpful insights and for critical input during the preparation
of this manuscript.
References
1. Cahalan MD, Parker I (2008) Choreography 3. Germain RN, Robey EA, Cahalan MD (2012)
of cell motility and interaction dynamics A decade of imaging cellular motility and inter-
imaged by two-photon microscopy in lym- action dynamics in the immune system. Science
phoid organs. Annu Rev Immunol 26: 336:1676–1681
585–626 4. von Andrian UH, Mempel TR (2003) Homing
2. van Panhuys N (2016) TCR signal strength and cellular traffic in lymph nodes. Nat Rev
alters T-DC activation and interaction times Immunol 3:867–878
and directs the outcome of differentiation. 5. Mandl JN, Liou R, Klauschen F, Vrisekoop N,
Front Immunol 7:6 Monteiro JP, Yates AJ, Huang AY, Germain
RN (2012) Quantification of lymph node tran- 12. Moreau HD, Lemaitre F, Terriac E, Azar G, Piel
sit times reveals differences in antigen surveil- M, Lennon-Dumenil AM, Bousso P (2012)
lance strategies of naive CD4+ and CD8+ T Dynamic in situ cytometry uncovers T cell recep-
cells. Proc Natl Acad Sci U S A 109: tor signaling during immunological synapses
18036–18041 and kinapses in vivo. Immunity 37:351–363
6. Stoll S, Delon J, Brotz TM, Germain RN 13. Chodaczek G, Papanna V, Zal MA, Zal T
(2002) Dynamic imaging of T cell-dendritic (2012) Body-barrier surveillance by epidermal
cell interactions in lymph nodes. Science gammadelta TCRs. Nat Immunol 13:272–282
296:1873–1876 14. Shulman Z, Gitlin AD, Weinstein JS, Lainez B,
7. Miller MJ, Safrina O, Parker I, Cahalan MD Esplugues E, Flavell RA, Craft JE, Nussenzweig
(2004) Imaging the single cell dynamics of MC (2014) Dynamic signaling by T follicular
CD4+ T cell activation by dendritic cells in helper cells during germinal center B cell selec-
lymph nodes. J Exp Med 200:847–856 tion. Science 345:1058–1062
8. S.E. Henrickson, T.R. Mempel, I.B. Mazo, 15. Klauschen F, Ishii M, Qi H, Bajenoff M, Egen
B. Liu, M.N. Artyomov, H. Zheng, A. Peixoto, JG, Germain RN, Meier-Schellersheim M
M. Flynn, B. Senman, T. Junt, H.C. Wong, (2009) Quantifying cellular interaction dynam-
A.K. Chakraborty, and U.H. von Andrian ics in 3D fluorescence microscopy data. Nat
(2008) In vivo imaging of T cell priming. Sci Protoc 4:1305–1311
Signal 1:pt2 16. Beltman JB, Maree AF, de Boer RJ (2009)
9. Mempel TR, Henrickson SE, Von Andrian UH Analysing immune cell migration. Nat Rev
(2004) T-cell priming by dendritic cells in Immunol 9:789–798
lymph nodes occurs in three distinct phases.
17. Lindquist RL, Shakhar G, Dudziak D,
Nature 427:154–159 Wardemann H, Eisenreich T, Dustin ML,

10. van Panhuys N, Klauschen F, Germain RN Nussenzweig MC (2004) Visualizing den-
(2014) T-cell-receptor-dependent signal inten- dritic cell networks in vivo. Nat Immunol
sity dominantly controls CD4(+) T cell polar- 5:1243–1250
ization in vivo. Immunity 41:63–74 18. Henrickson SE, Mempel TR, Mazo IB, Liu B,

11. F. Marangoni, T.T. Murooka, T. Manzo, Artyomov MN, Zheng H, Peixoto A, Flynn
E.Y.Kim, E. Carrizosa, N.M. Elpek, and MP, Senman B, Junt T, Wong HC, Chakraborty
T.R. Mempel (2013) The transcription factor AK, von Andrian UH (2008) T cell sensing of
NFAT exhibits signal memory during serial T antigen dose governs interactive behavior with
cell interactions with antigen-presenting cells. dendritic cells and sets a threshold for T cell
Immunity activation. Nat Immunol 9:282–291
Index
A C
Acid stripping�� 144, 148, 153 Capacitance measurements�� 158–163, 165–168
Actin��7, 40, 66, 130, 166, 177, 183, 333, Cell
389, 406, 418, 478, 488, 519, 533, 545 B cells��2, 32, 37, 38, 41, 48, 52, 55, 61,
Adhesion�� 1, 7, 31, 62, 89, 129, 150, 77–87, 90, 97, 106, 113, 130, 144, 145, 149–154, 160,
165, 174, 212–213, 231–235, 238, 242, 244, 245, 173, 185, 334, 450, 488, 491–494, 534, 535,
249–251, 253, 255, 272, 308, 339, 347, 356, 423, 537–543, 546, 580
451, 475, 503, 534, 545 cell activation�� 2, 7, 31, 51, 65, 67, 102,
Affinity�� 2, 39, 53, 57, 78, 165, 173, 203, 129, 130, 183, 184, 232, 260, 293, 308–311, 314, 315,
208, 232, 245, 293, 297, 347–349, 357, 370–374, 355, 356, 409, 443, 444, 494, 503, 520, 546, 559, 560,
447, 488, 494, 560, 564, 567 569–571, 581
Agent-based modeling�� 173, 178 cell fate determinants�� 383, 389
Antibodies�� 1, 2, 8, 32, 53, 77, 91, 104, cell polarization�� 352, 394, 489
106, 143, 150–151, 185, 208, 236, 241, 293, 296, 297, cytotoxic T cells�� 129, 334, 343, 424, 498, 519
300, 301, 303, 308, 334, 358, 360, 371, 384, 426, 443, lymphocytes�� 1, 22, 32, 33, 101, 104,
452, 475, 487, 499, 519–520, 534, 547 138, 145, 146, 148–152, 157, 159, 172, 173, 185, 192,
Antigen 334, 337, 356–358, 361–363, 365, 366, 389, 399,
antigen affinity��347–349, 560, 564, 567 401–404, 406, 407, 413, 444, 450, 473, 498, 537,
antigen internalization��77–87 543, 554, 564
antigen presenting cell (APC)��2, 7–9, 31, 32, mast cell�� 2, 158, 185, 487, 488, 490, 491,
36–37, 39–41, 48, 51, 52, 55, 61, 62, 77, 78, 89, 90, 493, 494
94, 95, 97, 102, 103, 106, 113–115, 129, 134, 137, natural killer (NK) cells��2, 160, 172, 283,
138, 144, 150–151, 172, 173, 178, 179, 183, 232, 260, 450, 498–507, 509–513, 534, 545
271, 273, 274, 291–292, 294, 304, 308, 317, 321, 323, red blood cell (RBC)��84, 232, 234–236, 238–245,
355, 386, 389, 410, 413–416, 419, 424, 534–543, 247–251, 254, 255, 418, 563, 575
550–551, 554, 559, 570, 571, 581 T cell activation�� 2, 7, 31, 51, 65, 67, 102,
Antigen receptor��7, 21, 65, 67–75, 89–97, 130, 183, 184, 232, 260, 308–311, 314, 315, 355, 356,
129, 130, 143–149, 152–154, 260–270, 272–278, 409, 443, 444, 520, 546, 559, 569–571, 581
280, 282–288, 292, 348, 369, 519, 520, 527, 560 T cells��1, 2, 7, 31, 51, 65, 77, 89, 90,
B cell antigen receptor�� 77–82, 84–87 97, 101, 143, 144, 149–154, 158, 160, 174, 183, 214,
T cell antigen receptor�� 97, 129, 228 232, 253, 291, 308, 334, 347, 352, 369, 383, 406, 409,
Astrocyte�� 517–519, 522–527 424, 443, 473, 474, 498, 517, 530, 539–541, 543, 545,
Asymmetric cell division (ACD)�� 383–385, 388–395 559, 569–571
Avidin��326, 488–490, 493, 494 T cell signaling�� 7, 177, 356, 371, 410, 419, 444, 451
thymocyte��232, 385–388, 390, 391
B Centrosome�� 32, 41, 45, 144
Bcl10�� 103, 108, 118 Chromatography�� 86, 208, 228, 264, 370–372,
Bifunctional biomimetic surfaces�� 308, 323 435, 455, 457, 460
Biomechanics�� 333, 335–338, 340–345 Computational biology�� 174, 177
Biomembrane force probe (BFP)��233, 235, 237–239, Computational image analysis�� 84, 85, 416
242, 245, 247–252, 255 Co-stimulation�� 51, 58, 292, 293, 348, 352, 419, 444
vol. 1584, DOI 10.1007/978-1-4939-6881-7, © Springer Science+Business Media LLC 2017
585
The Immune Synapse: Methods and Protocols
586 Index

Cytoskeleton Immune/immunological synapse (IS)
actin cytoskeleon�� 8, 130, 334, 533 degranulatory synapse�� 487, 488, 490, 491, 493, 494
microtubule cytoskeleton�� 130, 545 microcluster��3, 51–63, 65–67, 71, 90,
Cytotoxic cells 174, 177, 178, 183, 184, 191, 202, 203, 217, 218,
cytotoxic granules (CG)��129, 157, 158, 162, 224–225, 228, 356, 451
165, 166, 168 Immunoaffinity purification��371
degranulatory synapse�� 487, 488, 490, 491, 493, 494 Immunofluorescence (IF)��33, 34, 36–39, 47,
granule convergence�� 498–507, 509–513 91–92, 131, 386, 430–431, 489, 518, 524, 547
lytic granules�� 497–507, 509–513, 517 Inclusion bodies�� 208, 228, 456, 458–460, 469, 470
polarized degranulation�� 487, 490, 493 Integrins�� 7, 12, 51, 58, 130, 178, 184, 232, 334, 336, 347
D K
3D reconstructions�� 137, 509, 524, 527, 530, 553 Kinapses�� 347–351, 353, 559, 561–564, 566, 567
Kymographs��8, 16–19, 26, 351–353
E
L
Eμ-TCL1 model��534
Label-free quantitation�� 371, 372, 379
F Leukemia��130, 145, 534–543, 546
Flotation gradient�� 356–357, 359, 361–364 Linker for activation of T cells (LAT)�� 66–68, 71, 74,
Flow cytometer�� 93, 118, 121, 145, 241, 447, 455 129, 130, 186, 188, 191–193, 196–197, 199, 352,
Forces�� 3, 171, 172, 174, 176–181, 356–358, 361–363, 365, 366, 390, 546, 560
233–235, 238, 248–253, 255, 256, 297, 308, 309, Lymphoma��17, 32, 90, 130, 145, 345, 419, 534, 546
333–335, 337, 344, 345, 424
M
Force spectroscopy��233
Mass spectrometry�� 370, 371
G Mathematical algorithm��507
Glycosylation��456 Mechanics��172
GPIb��232 Mechanobiology�� 308, 335
Mechanotransduction��240, 246, 265, 272, 333
H Membrane binding��259–270, 272–278, 280, 282–288
Microchannels�� 347–351, 353
High throughput screening��426–428
Microclusters (MC)��3, 51–63, 65–67, 71, 90,
Human immunodeficiency virus (HIV-1)�� 461, 546–550,
174, 177, 178, 183, 184, 191, 202, 203, 217, 218,
552, 553, 556
224–226, 228, 356, 451
I Microcontact printing��292, 293, 296, 297, 304
Micropillar�� 333, 335–338, 340–345
Imaging Micropipettes�� 232–244, 247–249, 254, 376, 379, 380
image analysis��12, 16–20, 34, 43–47, 71, Microscopy
121, 135–139, 262, 270, 334, 343–344, 389, 420, confocal microscopy��8, 15–16, 39, 45–47,
432–434, 437, 474, 505, 536, 541, 551–553, 573 54, 61–63, 80, 90, 92, 93, 95, 96, 108, 117, 118, 124,
image quantification��410 131, 135, 136, 144–146, 148–150, 278, 387, 388,
imaging flow cytometry�� 103, 104, 108, 473, 474, 487, 489, 491–493, 498, 500, 502, 504,
118–123, 125, 385 508–509, 511–513, 523–524, 526, 527, 530, 534,
Imaris��34, 39–41, 43, 45, 47, 54, 62, 351–353, 541, 546, 551, 552
474, 477, 478, 486, 527, 564, 573 correlative microscopy�� 400, 404–407, 571
in vivo imaging�� 4, 32, 384, 559, 561–564, 566, 567 direct stochastic optical resolution microscopy
live cell imaging��8, 11, 15, 36, 51, 52, 59, 62, 335, (dSTORM)�� 185, 187–190, 193–199, 201, 202
342–343, 388, 391, 392, 410, 413–414, 498, electron microscopy�� 1, 69, 131, 202, 295, 309, 363,
503–504, 510 399–405, 407, 426, 489, 499, 547
time-lapse confocal imaging�� 474, 478, 485, 487 fluorescence lifetime imaging microscopy
time-lapse imaging��15, 16, 43–46, 63, 71, (FLIM)��260–270, 272–278, 280, 282–288
72, 74, 92, 94, 269, 270, 285, 337, 384, 386–388, fluorescence microscopy��2, 11, 32, 34, 51, 52,
390–392, 394, 474, 478, 485, 487, 493, 505, 510, 71, 92, 103, 112, 124, 125, 157, 159, 178, 235, 325,
511, 560 335, 337, 404, 406, 407, 424
587
Index

fluorescence resonance energy transfer (FRET) R
microscopy�� 3, 207–229, 260–270, 272–278,
280, 282–288 Receptors�� 3, 7, 31, 51, 65, 77, 101, 129,
photo-activation localization microscopy 143, 165, 172, 183, 212–213, 231, 259, 291, 308, 333,
(PALM)��8, 185, 187–201, 203 355, 369, 406, 412, 424, 443, 451, 478, 488, 497, 517,
single molecule localization microscopy�� 8, 184, 545, 571
186–189, 191–194, 198, 199, 203 Recombinant�� 3, 4, 9, 21, 32, 66, 109, 236,
spinning disk microscopy�� 414, 474, 502 263, 295, 318, 335, 348–350, 365, 384, 387, 425,
super-resolution microscopy��8, 104, 108, 113, 464–467, 475, 490, 548
117, 118, 124, 184, 186–189, 191–194, 198, 199, Reconstitution�� 3, 4, 48, 57–60, 63, 65,
202, 203, 452, 514 67–75, 78, 209–214, 261, 374, 376, 377, 389, 424,
TIRF microscopy�� 42–45, 62, 63, 66, 68, 71, 452, 563, 566
74, 158–163, 165–168, 187, 308, 310 Retroviral expression system��52
traction force microscopy��334 Retroviruses��51–53, 55, 104–105, 109–110,
Microtubule-organizing centre (MTOC)�� 31, 37, 387, 411–412, 461
39–41, 43, 44, 46, 48, 154, 488, 492, 497, 498, 505,
S
508–511, 520
Microtubules��31–45, 47–49, 130, Science history��2–3
136–137, 154, 166, 202, 333, 478, 488, 497, 498, Self-assembly�� 172, 308, 311, 318
500–505, 507, 545 Signal transduction�� 67, 102–104, 183, 185, 356, 410
Mitochondria�� 32, 363 Silicon etching��338
Mobile ligands�� 9, 12–15 Single-molecule�� 3, 8, 184, 186–189, 191–194,
Modeling�� 3, 4, 173, 410 198, 199, 203, 209, 212–216, 220–221, 226–228,
Multivalency��66, 67 234, 235, 304, 307, 308, 311, 326
Soft-lithography��292
N Spatiotemporal distributions��409
Nanofabrication�� 309, 311 Strep-Tag-Strep-Tactin affinity purification�� 357–359,
Nef�� 545–548, 550, 554 362–366
NF-κB��102–125 Supported lipid bilayer (SLB)�� 2, 65–67, 69–75,
Nucleofection��35–36, 48, 475–478, 481 173–175, 179, 208, 209, 212–216, 218, 219, 221,
226–229, 304, 308, 318, 320–323, 399, 401–404,
P 407, 424–427, 429–431, 433–437
Supramolecular activation clusters (SMACs)
Partial differential equation (PDE)��172
c-SMAC��89, 90, 517, 519, 528, 530
Patterns��2, 89, 171, 177, 238, 291, 292, 335,
p-SMAC��89, 517, 519, 528, 530
404, 424, 519, 581
Peptides bound to major histocompatibility complex T
(pMHC) molecules�� 172–179, 207–229,
232, 234, 236, 238, 240–244, 246–249, 252, 274, T cell receptor (TCR)
334, 336, 342, 343, 347–349, 352, 559 TCR-CD3 complex��130, 143–149, 152–154,
Phase separation�� 67, 73, 74 519, 527
Phosphoproteomics�� 370, 372–380 TCR internalization��89–98
Phosphorylation�� 25, 51, 65–68, 71, 74, 90, TCR signaling�� 21, 65, 67–75, 89, 90, 129,
102, 103, 140, 196–197, 364, 365, 369–371, 376, 292, 348, 370, 372–380, 520, 560
378, 556 Tissue�� 15, 24, 80, 81, 104–106, 108,
Planar lipid bilayers�� 8–11, 13–15, 66, 87, 399, 500 110–112, 117, 119, 124, 190, 191, 214, 308, 380, 384,
Plasma membrane sheets��77–87 410, 411, 418, 462–467, 471, 487, 488, 499, 501,
Polarization�� 2, 48, 143, 144, 150–151, 154, 517–520, 523, 525–527, 530, 566, 567, 569–571,
352, 356, 389, 392, 394, 396, 397, 424, 487 574–577, 582
POLKADOTS�� 103–105, 108, 109, 113, 117, 118 Tomography��4, 399
Polydimethylsiloxane (PDMS)�� 295, 334, 336, Transendocytosis��77, 78
392, 435 Trogocytosis��77
Protein folding�� 456, 458–460 Tyrosine phosphorylation�� 370, 371
588 Index

V internalization��84–86
lytic granules�� 498–507, 509–513, 517
Vesicle traffic polarized degranulation�� 487, 490, 493
endocytosis�� 37, 77, 78, 356 polarized recycling��144
endosomes��130, 143, 144, 150–151, 546 transendocytosis��77, 78
exocytosis�� 157, 158, 165–168, 356, 487, 497, 498 trogocytosis��77

Dustin & Baldari, 2017 - The Immune Synapse PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dustin & Baldari, 2017 - The Immune Synapse PDF

Uploaded by

Copyright:

Available Formats

Methods in

Molecular Biology 1584

For further volumes:

Methods and Protocols

ISSN 1064-3745 ISSN 1940-6029 (electronic)

Library of Congress Control Number: 2017931687

© Springer Science+Business Media LLC 2017

Printed on acid-free paper

This Humana Press imprint is published by Springer Nature

Initiation of the T cell-mediated adaptive immune response to pathogens is crucially depen-

Siena, Italy Cosima T. Baldari

1 The Immune Synapse: Past, Present, and Future . . . . . . . . . . . . . . . . . . . . . . . 1

15 Two-Dimensional Analysis of Cross-Junctional Molecular Interaction

30 The Mast Cell Antibody-Dependent Degranulatory Synapse . . . . . . . . . . . . . . 487

Jehan Afrose • University of Oxford, Oxford, UK

John G. Gribben • Queen Mary University of London, London, UK

Muaz Nik Rushdi • Institute of Technology, Atlanta, USA

The Immune Synapse: Past, Present, and Future

Key words Science history, Fluorescence, Affinity, Modeling, Microscopy

Phagocytosis and antibodies were described within a decade of each

monoclonal antibodies [8]. The field’s recognition of directed

This book is composed of 35 chapters (excluding this introductory

details of technical approaches that can be applied to the multiple

nano-fabrication methods, proteomics, asymmetric cell division,

3.3 Biological Chapters 28–36 focus on some compelling biological examples

We thank Éva Culleton-Oltay for assistance with editing the chap-

Analyzing Actin Dynamics at the Immunological Synapse

The formation of the immunological synapse (IS) between a T cell

Much of what is known about actin dynamics at the IS comes

In studies using planar surfaces to analyze protein dynamics at

All solutions should be prepared using reverse osmosis (Milli-Q)

2.1 Generation 1. Coverslips for Sticky-Slides (Ibidi): # 1.5H (170 μm ± 5 μm)

2.1.2 Preparation 1. Bottomless Sticky-Slide VI 0.4 6-channel chambers (recom-

2.1.3 Coating Surfaces 1. Mouse anti-human T cell receptor: OKT3 (recommended

2.1.4 Coating Surfaces 1. 50 ml glass round-bottom flask.

12. Biotinylated OKT3 (0.5 mg/ml). Store at 4 °C.

2.3 Image Analysis 1. Volocity v. 6.3 imaging software (Perkin Elmer).

3.1 Generation 1. Sonicate the glass coverslips in 2% Helmanex detergent for

Coating with Immobile We typically apply OKT3, a monoclonal antibody that reacts

2. Using the multichannel pipette, add 100 μl of 10 μg/ml OKT3

DSPE-­PEG(2000), and 1 μl DGS-NTA into the flask, pipet-

environmental chamber (see Subheading 3.2.2). Ligand mobility

3.2.2 Preparation 1. Set the environmental chamber on the microscope to 37 °C,

Flow rate [nm/s]

2 Jurkat_Kymograph Lines 2 Lines Line 24.08 77 113 93 131 138.37

to calculate actin flow rates. If you are interested in calculating

End Normalized β (radian) = V [μm/s] =

1. Piranha solutions are extremely energetic and may result in

a fresh vial about once a month. Thus, it is important to main-

types. Purified CD4+ T cells are activated with CD3/CD28

9. It is best to prepare chambers just before usage and to use

to have fluorescent dye. For lipid bilayers we have used Texas

Inhibitor Dose Effect on T cell actin

Inhibitor Dose Effect on T cell actin

where each measurement was made. Because the diameter of

1. Kumari S, Curado S, Mayya V, Dustin ML ognition is facilitated by invadosome-like pro-

40. Kumari S, Depoil D, Martinelli R, Judokusumo latrunculin B on mechanical properties of cells.

Analysis of Microtubules and Microtubule-Organizing

The immune synapse (IS) is a cell-cell contact between a T cell and

such as the Golgi Apparatus (GA), the endolysosomal system,

2.2 Reagents 1. Stimulatory antibodies and recombinant proteins: anti-CD3

Recombinant SEE (Staphylococcus aureus enterotoxin E; Toxin

2.3 Media 1. Complete medium: RPMI 1640 supplemented with Glutamine

Siena, Italy Cosima T. Baldari

DSPE-PEG(2000), and 1 μl DGS-NTA into the flask, pipet-

This work was supported by KAKENHI (Grant-in-Aid for Scientific

All buffers and media should be filtered through a 0.22 μm bottle-