Unit of Analytical Chemistry

Department of Environmental Science and Analytical Chemistry

Cellular phototoxicity in live cell

fluorescence microscopy measured by
SWATH mass spectrometry

45 ECTS
Advanced Level Graduate Thesis
in Analytical Chemistry

Sofia Emmanouela Theodorou

2018
ii
 45 ECTS
Advanced Level Graduate Thesis
for Master of Science in Analytical Chemistry

performed at

Aebersold Group
Institute of Molecular Systems Biology (IMSB)
Department of Biology
ETH Zürich

jointly with

Weis Group
Institute of Biochemistry (IBC)
Department of Biology
ETH Zürich

Supervisors: Rodolfo Ciuffa (IMSB), Elisa Dultz (IBC)

Principal Investigators: Ruedi Aebersold, Karsten Weis

Contact in Home University: Ulrika Nilsson, Conny Östman

Examiner: Conny Östman

© ETH Zürich, 2018

iii
«Σα βγεις στον πηγαιμό για την Ιθάκη,
να εύχεσαι νάναι μακρύς ο δρόμος,
γεμάτος περιπέτειες, γεμάτος γνώσεις.
Τους Λαιστρυγόνας και τους Κύκλωπας,
τον θυμωμένο Ποσειδώνα μη φοβάσαι,
τέτοια στον δρόμο σου ποτέ σου δεν θα βρεις,
αν μέν’ η σκέψις σου υψηλή, αν εκλεκτή
συγκίνησις το πνεύμα και το σώμα σου αγγίζει.
Τους Λαιστρυγόνας και τους Κύκλωπας,
τον άγριο Ποσειδώνα δεν θα συναντήσεις,
αν δεν τους κουβανείς μες στην ψυχή σου,
αν η ψυχή σου δεν τους στήνει εμπρός σου.
Να εύχεσαι νάναι μακρύς ο δρόμος.
Πολλά τα καλοκαιρινά πρωιά να είναι
που με τι ευχαρίστησι, με τι χαρά
θα μπαίνεις σε λιμένας πρωτοειδωμένους·
να σταματήσεις σ’ εμπορεία Φοινικικά,
και τες καλές πραγμάτειες ν’ αποκτήσεις,
σεντέφια και κοράλλια, κεχριμπάρια κ’ έβενους,
και ηδονικά μυρωδικά κάθε λογής,
όσο μπορείς πιο άφθονα ηδονικά μυρωδικά·
σε πόλεις Aιγυπτιακές πολλές να πας,
να μάθεις και να μάθεις απ’ τους σπουδασμένους.
Πάντα στον νου σου νάχεις την Ιθάκη.
Το φθάσιμον εκεί είν’ ο προορισμός σου.
Aλλά μη βιάζεις το ταξείδι διόλου.
Καλλίτερα χρόνια πολλά να διαρκέσει·
και γέρος πια ν’ αράξεις στο νησί,
πλούσιος με όσα κέρδισες στον δρόμο,
μη προσδοκώντας πλούτη να σε δώσει η Ιθάκη.
Η Ιθάκη σ’ έδωσε τ’ ωραίο ταξείδι.
Χωρίς αυτήν δεν θάβγαινες στον δρόμο.
Άλλα δεν έχει να σε δώσει πια.
Κι αν πτωχική την βρεις, η Ιθάκη δεν σε γέλασε.
Έτσι σοφός που έγινες, με τόση πείρα,
ήδη θα το κατάλαβες η Ιθάκες τι σημαίνουν.»

iv
Καβάφης Κωνσταντίνος, Από τα Ποιήματα 1897-1933
“As you set out for Ithaka
hope the voyage is a long one,
full of adventure, full of discovery.
Laistrygonians and Cyclops,
angry Poseidon—don’t be afraid of them:
you’ll never find things like that on your way
as long as you keep your thoughts raised high,
as long as a rare excitement
stirs your spirit and your body.
Laistrygonians and Cyclops,
wild Poseidon—you won’t encounter them
unless you bring them along inside your soul,
unless your soul sets them up in front of you.
Hope the voyage is a long one.
May there be many a summer morning when,
with what pleasure, what joy,
you come into harbors seen for the first time;
may you stop at Phoenician trading stations
to buy fine things,
mother of pearl and coral, amber and ebony,
sensual perfume of every kind—
as many sensual perfumes as you can;
and may you visit many Egyptian cities
to gather stores of knowledge from their scholars.
Keep Ithaka always in your mind.
Arriving there is what you are destined for.
But do not hurry the journey at all.
Better if it lasts for years,
so you are old by the time you reach the island,
wealthy with all you have gained on the way,
not expecting Ithaka to make you rich.
Ithaka gave you the marvelous journey.
Without her you would not have set out.
She has nothing left to give you now.
And if you find her poor, Ithaka won’t have fooled you.
Wise as you will have become, so full of experience,
you will have understood by then what these Ithakas mean.”

Cavafy Constantine (1897 – 1933), translated by Edmund Keeley and Philip Sherrard.

v
Στον Λιαντίνη Δημήτρη, που ακόμα και αν δεν είναι εδώ, μου έδωσε δύναμη όταν έπεφτα, να συνεχίσω.

vi
Table of Contents
Abbreviations .......................................................................................................................................... ix
Abstract ..................................................................................................................................................... x
Popular science abstract ......................................................................................................................... xi
Research aim ............................................................................................................................................. 1
1. Introduction ..................................................................................................................................... 1
1.1 Phototoxicity in live cell microscopy.................................................................................... 1
1.2 Proteomics and mass spectrometry ...................................................................................... 2
1.2.1 Ionization sources ........................................................................................................... 3
1.2.2 Mass analyzers ................................................................................................................. 5
1.2.3 Acquisition modes ........................................................................................................ 12
1.3 Model cell ............................................................................................................................... 16
1.4 Challenges of the study......................................................................................................... 16
1.4.1 Exposure conditions..................................................................................................... 16
1.4.2 Miniaturization .............................................................................................................. 17
2. Methods.......................................................................................................................................... 17
2.1 Chemicals................................................................................................................................ 17
2.2 Solvents and solutions .......................................................................................................... 18
2.3 Cells preparation .................................................................................................................... 18
2.4 Light exposure ....................................................................................................................... 18
2.5 Lysis ......................................................................................................................................... 19
2.6 Digestion ................................................................................................................................ 20
2.7 C-18 purification and desalting ........................................................................................... 21
2.8 Fractionation .......................................................................................................................... 21
2.9 Instrumental method ............................................................................................................ 22
2.10 Computational analysis ......................................................................................................... 24
3. Results and Discussion................................................................................................................. 25
3.1 Exposure conditions ............................................................................................................. 25
3.2 Sample preparation optimization ........................................................................................ 26
3.3 Data quality ............................................................................................................................ 31
3.4 Candidates .............................................................................................................................. 33
4. Conclusions.................................................................................................................................... 37
Acknowledgements................................................................................................................................ 37
References ............................................................................................................................................... 38
Appendices.............................................................................................................................................. 43

vii
Fig. 1. Schematic representation of the electrospray ionization process. ........................................ 5
Fig. 2. Diagram of Quadrupole MS adapted from Shimadzu ........................................................... 6
Fig. 3. Schematic representation of an ion trap ................................................................................. 10
Fig. 4. Orbitrap architecture. ................................................................................................................ 11
Fig. 5. Data depended acquisition, scheme explanation. .................................................................. 13
Fig. 6. Data independent acquisition DIA-SWATH, scheme explanation.................................... 16
Fig. 7. Meander scanning method. ...................................................................................................... 19
Fig. 8. Fractionation of the sample...................................................................................................... 22
Fig. 9. Sample preparation workflow .................................................................................................. 26
Fig. 10. Elution profile of yeast samples ............................................................................................ 28
Fig. 11. CV of the response of proteins from glass beads and ultrasonication. ........................... 29
Fig. 12. CV of the response of proteins from FASP, Fractionation and No FASP-No Fract . 30
Fig. 13. CV of the response of peptides from different SWATH settings .................................... 31
Fig. 14. Comparison of the yeast proteome datasets ....................................................................... 33
Fig. 15. Heatmap showing the proteins that differentiate in each sample.. .................................. 35
Fig. 16. Annotation of the proteins reported on Fig. 15, to their function. .................................. 36
Fig. 17. The scheme shows which candidates bind to which enzymes. ......................................... 36

Table 1. MS2 settings were tested for optimization of the method ............................................... 23
Table 2. Results from the optimization of the SWATH analysis.................................................... 31

viii
Abbreviations
UV Ultraviolet
SWATH Sequential Windowed Acquisition Of All Theoretical Fragment Ion Mass Spectra
DDA Data Depended Acquisition
DIA Data Independent Acquisition
SRM Selected Reaction Monitoring
MRM Multiple Reaction Monitoring
MS Mass Spectrometry
ESI Electron Spray Ionization
MALDI Matrix Assisted Laser Desorption/Ionization
TOF Time of Flight
qTOF Quadrupole Time of Flight
LIT Linear Ion Traps
FT-ICR Fourier Transform Ion Cyclotron Resonance
m/z Mass to Charge
FWHM Full Width Half Maximum
AC Alternative Current
DC Direct-Current
RF Radio Frequency
FFT Fast Fourier Transform
LC Liquid Chromatography
nLC Nano Liquid Chromatography
ROS Reactive Oxygen Species
PTMs Post-Translational Modifications
ID Identifiers
ambic Ammonium Bicarbonate
GFP Green Fluorescent Protein
YNB Yeast Nitrogen Base
OD600 Optical Density at 600 nm
LED Light Emitting Diode
ConA Concanavalin A
BCA Bicinchoninic Acid
FASP Filter Aided Sample Preparation
TCEP (2-Carboxyethyl)Phosphine, Hydrochloride
IAA Iodoacetamide
FA Formic Acid
ACN Acetonitrile
NADPH Nicotinamide Adenine Dinucleotide Phosphate
ATP Adenosine Triphosphate
LFQ Label-Free Quantification
FDR False Discovery Rate
PB-MS Proteomics Based-Mass Spectrometry

ix
Abstract
In this research, the first known attempt of measuring phototoxicity by mass spectrometry was
conducted. The two challenges of the study, the exposure conditions of the cells and the
compatibility of minute samples taken from microscopy with mass spectrometry, were both
tackled. The model organism used was the budding yeast Saccharomyces cerevisiae. Optimization of
microscopy while part of the work, is not discussed here and only the final optimized settings
are reported. Three conditions of light exposure were tested: 475 nm, 542 nm and Ultra Violet
(UV) 390 nm. Optimization of the sample preparation for mass spectrometry for minute samples
of low number of yeast cells was conducted. Lysis of low number of yeast cells was an important
challenge and was tackled based on the work by Wenger et al (2008) (Wenger, DePhillips, &
Bracewell, 2008) but was adapted to microvolume. The method was tested against the routine
yeast lysis methods for proteomics, glass beating. The sample preparation was further optimized
with tests on use of filter aided sample purification (FASP) and fractionation, on an LTQ
Orbitrap XL on data depended acquisition mode (DDA). The final study run on an OrbitrapTM
FusionTM LumosTM Tribrid Mass Spectrometer on Sequential Windowed Acquisition of all
THeoretical fragment ion mass spectra (SWATH-MS) acquisition mode, and the SWATH
window and instrumental resolution was optimized prior to the study. Ultrasonication was
selected as lysis method and was found to not enrich or deplete specific protein function and/or
compartments. However, the data from the SWATH-MS had statistical anomalies which are
under investigation. A reason for that can be the use of a DDA library (to deconvolute the
SWATH-MS2 spectra) which was generated with different sample preparation methods.
Preliminary data analysis showed 25 candidates which change their abundance upon exposure
to microscopic light. The candidates map on mainly metabolic pathways and glycolysis which
was previously found to be upregulated in mammalian cells upon exposure of cells to UV light
(McGonigle et al., 2017). However due to the statistical anomaly of the data, these results should
be taken with caution and further evaluation of the data and analysis should be conducted.

x
Popular science abstract
We - humans, and any other organism - are made of tiny building blocks called cells. In turn,
cells are to a large extent made of even tinier blocks called proteins; and these proteins are
nothing but strings of amino acids, like words made of letters. The genetic code, the DNA,
define what words the cells will write. There are so many protein types in a cell in order of
thousands, depending on the cell. When there is a disease, one or more proteins are changing.
Some of these thousand proteins either are produced in excess (so their amount is increasing),
or less (so their amount is decreasing). Our general goal is to find out which of these thousand
proteins increase or decrease when the cell is perturbed - for instance, by changing the
temperature at which they grow or by altering some “sentences” in the DNA. These changes
can be communicated and the proteins that change their levels can potentially become targets of
drugs. The drugs will in turn attack these proteins and regulate their level, keeping them in the
right amount in order for the cells to be healthy. In my work, I studied how one of the most
important and pervasive physical parameter in our everyday’s life - light - changes protein
abundance in the cell. This is the first attempt of its kind and has potentially relevance for human
diseases (e.g. phytotoxic effect), ecological studies (e.g. unicellular organisms) and studies of light
microscopy, a critically important method used in biological laboratories. Most of my work was
dedicated to define a method that scientists can use to study the effect of light on proteins,
similar to a programmer writing a program that people can use to predict the weather or solve
equations. Using this method, I have obtained preliminary results describing how different types
of light change the protein levels of a small unicellular organism called baker’s yeast. The data
requires still extensive investigation and we are excited about the results to come.

xi
Research aim
In this study, we aimed at quantifying protein abundance changes in cells exposed to microscopic
light. Target of the research is to describe at the protein level the well-known but poorly
characterized phenomenon of phototoxicity in microscopy experiments with use of state-of-the-
art mass spectrometry. The study focused on two areas, a) developing an analytical method
compatible with minute microscopic samples and b) the discovery of proteins whose abundance
is changed upon exposure to light.

1. Introduction
1.1 Phototoxicity in live cell microscopy
Live cell imaging microscopy is an integral part of modern cell biology, thus is widely used in
order to answer fundamental questions regarding cellular and tissue dynamics/function
(Frigault, Lacoste, Swift, & Brown, 2009). Technological advances in microscopy have resulted
in an increased number of researches in several fields using this technique. For instance, live
microscopy is lately used for looking into the spatial organization and direct interactions of immune cells
within blood and liquid multi-lineage tissues, a field called pharmacoscopy and is highly evolving
(Vladimer et al., 2017). However, when one uses live microscopy to answer related questions,
has to ensure that the physiological and biological processes that are being examined are not
altered during cell excitation of fluorophore.

Phototoxicity is a phenomenon occurring during live microscopy and refers to how

exogenous light energy may interact with the tissue/cell and alter its structure/function (Tinevez
et al., 2012). Phototoxicity manifests itself in a variety of phenomena, ranging from subtle effects
that can go unnoticed and undetected ("Artifacts of light," 2013; Icha, Weber, Waters, &
Norden, 2017) to cell death (Magidson & Khodjakov, 2013). Table A- 1 in Appendices [as
taken from 5] shows a number of instances of subtle phototoxicity. According to a publication
in Nature Methods ("Artifacts of light," 2013) only 0.4% to 7% of the papers published between
2005 -2013 discuss phototoxicity in relation to microscopy terms. However, the topic has lately
received a particular attention (Kilian et al., 2018; "Phototoxicity revisited," 2018), as it might
compromise experimental results ("Artifacts of light," 2013). An interesting, recent example is
the case of the protein NEMO; it has been shown that the NEMO’s membrane-associated
superstructure was moving towards the nucleus, but in fact it moved as a result of phototoxic

1
effects (Tinevez et al., 2017) . In addition to that, phototoxicity has been also observed in
optogenetics, where protein activity is controlled by blue light (Stockley et al., 2017).

Light exposure can cause generation of reactive oxygen species (ROS) (Icha et al., 2017),
including superoxide radicals, hydroxyl radicals, and hydrogen peroxide. ROS can change the
redox state of the cytoplasm, can cause mutations to DNA upon oxidation and induce direct
DNA damage (Godley et al., 2005; Icha et al., 2017), or oxidize proteins and unsaturated fatty
acids in lipids disturbing their functions (Spikes, 1983). Additionally, it has been found that
fluorescence microscopy can induce mitotic arrest and eventually cell death (Dixit & Cyr, 2003).

There are several methods to estimate phototoxicity, like monitoring the cell rounding
as secondary indicator of apoptosis (De Vos et al., 2009), or the percentage of the cells that show
cell division during the next 20–24 hour after irradiation (Wäldchen, Lehmann, Klein, van de
Linde, & Sauer, 2015), or counting the dead cells after exposure to microscopic light (Dixit &
Cyr, 2003; Zdolsek, Olsson, & Brunk, 1990). Other methods to approach phototoxicity are to
count the number of cells that were able to form a colony after exposure (Wagner et al., 2010),
or the number of cells after a given time and compare that with a phototoxicity threshold value
(Tinevez et al., 2012).

The quantification of phototoxicity becomes of extreme importance. From the existing

methods, it can be seen that it is challenging to measure phototoxicity. The proposed methods
either look at what could be a secondary effect of phototoxicity (e.g. cell rounding) and therefore
would not help elucidating its mechanisms of action; and/or they capture only one single aspect
of the cellular phenomenology (e.g. ROS generation). Thus, it would be beneficial to the
microcopy community to propose a light induced toxicity mechanism based on a sensitive,
reliable and reproducible method of quantification. In this study, we quantify phototoxicity at
the protein level using proteomics-based mass spectrometry (PB-MS) - a method that, to our
knowledge has not been previously applied to tackle the problem - with the ultimate goal to
describe a possible mechanism of cellular phototoxicity induced from microscopic light.

1.2 Proteomics and mass spectrometry

Proteomics is the large-scale study and characterization of the entire protein complement of a cell line,
tissue, or organism (Aebersold & Mann, 2003; Graves & Haystead, 2002). The entire protein
complement known as proteome can differ from cell to cell, over time and as a function of

2
environmental stimuli and stress (Liebler, 2002). The use of mass spectrometry for the complete
characterization of the proteome (de Godoy et al., 2006; Mann, Kulak, Nagaraj, & Cox, 2013)
has increased dramatically in the last years. Mass spectrometry is an extremely sensitive method
that can identify and quantify not only high but importantly, also low abundant proteins making
it the method of choice when it comes to proteomics (de Godoy et al., 2006).

PB-MS finds applicability in the study of protein function, protein localization, on the
description of post-translational modifications (PTMs) like glycosylation, phosphorylation or
proteolysis, on determination of protein-protein interactions as well as on many other biological
areas of study explained in (Graves & Haystead, 2002). The overall aim of the field is to identify,
quantify and locate all the proteins within the cell with the target to create an accurate three
dimensional map of the cell under study (Graves & Haystead, 2002). This is achievable with the
state of the art mass spectrometry and the development of new experimental approaches, which
enable the simultaneous analysis of large number of proteins expressed in a cell (Domon &
Aebersold, 2006).

Proteomics has been applied to a number of biological studies, such as the

characterization of protein modifications to access the effects of environmental exposures
(Liebler, 2002); the quantitative and physical mapping of cellular proteins (physical mapping
refers to the identification of protein complexes which reveals information regarding a protein’s
function within the cell) (Blackstock & Weir, 1999); mapping where the Homo Sapiens’ proteins
are expressed in the different cells and tissues of the human body, a project known as Human
Protein Atlas (Uhlén et al., 2015). Additionally, PB-MS coupled with advanced bioinformatics
approaches is a method that gains attention, for biomarker discovery in early diagnosis (Kohn,
Azad, Annunziata, Dhamoon, & Whiteley, 2007) or for toxicity biomarker discovery in early
stage of drug development (Walgren & Thompson, 2004). It is a powerful tool that its use
and applicability to several questions has sharply raised in the last years (Aebersold & Mann,
2003).

1.2.1 Ionization sources

Ionization sources used for proteomics are the electron spray ionization source (ESI) and Matrix
Assisted Laser Desorption/Ionization (MALDI).

3
Electron Spray Ionization
Electron Ionization is a soft ionization technique (Carlsohn & Nilsson, 2007). The diluted in
polar volatile solvent sample solution is injected in the capillary at low flow rate. The analyte is
charged inside the capillary through electrochemical reactions. High voltage is applied on the tip
of the capillary. The capillary is connected with the counter through a metal wire, which upon
application of high voltage (±3 - ±8 kV) electrons are passing through establishing a circuit
between the capillary and the counter. Separation of positive and negative charged ions take
place at the meniscus of the capillary, which is formed due to the electric field applied. In case
we target positive ionization, positive voltage is applied. The cations will then accumulate on the
surface of the meniscus forcing the liquid to form a cone known as Taylor cone. When the
electrostatic repulsion force between the cations equals the surface tension of the liquid, so when
the charge-to-size ratio reaches the Rayleigh limit, a fine jet spray arising from the Taylor cone,
is generated. This mechanism is known as Coulombic fission (Carlsohn & Nilsson, 2007) (Fig.
1). The fine droplet creation depends on radius of the capillary - the narrower the more efficient
the fission-, on the surface tension of mobile phase and the distance between the capillary tip
and on the electrode counter. The higher the distance, the higher the voltage to be applied
(Carlsohn & Nilsson, 2007). Neutral and solvent molecules are removed by reduction when they
adhere to the 100 V metal plate of the MS inlet and the formed jet spray droplets fly towards the
counter electrode, the inlet to MS.

There is an extensive discussion on literature on how the naked ions in gas phase are
formed (M. Wilm, 2011). Two models have been widely accepted and are discussed here. The
first one is the charge residue model (Dole et al., 1968; M. S. Wilm & Mann, 1994), at which
mobile phase continuously evaporates until complete release of the ion, and is proposed for the
large molecule ion formation. Ionization of multiprotonated proteins is also best described by
the model found in (Carlsohn & Nilsson, 2007). The second model is the evaporation model
(Iribarne & Thomson, 1976; Thomson & Iribarne, 1979), at which the ion moves to the surface
of the droplet and pops out when the field strength overcomes the surface tension, upon
simultaneous evaporation of solvent. The latter one is applied to small molecule ion formation
(Fig. 1).

4
Fig. 1. Schematic representation of the electrospray ionization process. Adopted from (Cech & Enke, 2001).

1.2.2 Mass analyzers

Even though a wide range of mass analyzers is available, only a few are being used in the analysis
of proteins. These are the quadrupoles, the time of flight (TOF), and the ion traps (IT) in the
multiple configurations of linear ion traps (LIT), electrostatic traps (Orbitraps) or magnetic field-
based traps (Fourier transform ion cyclotron resonance (FT-ICR)) [1]. FT-ICR mass analyzer,
even though very important in the proteomics field, are beyond the scope of the present work
and will not be discussed here.

Quadrupoles
The quadrupole mass analyzers are mass filters that consist of four metal or gold-plated quartz
rods placed in parallel (de Hoffmann & Stroobant, 2007; Gushue, 2013; Mellon, 2003) (Fig. 2).
The ions are separated based on the stability of the ion trajectories, which is dependent on the
mass of the ion upon oscillating electric field applied on the rods. Opposite rods are connected
with each other by the application of the potential (U+Vcos(ωt)) to one opposite pair, and by
the application of –(U+Vcos(ωt)) to the other opposite pair, where Vcos(ωt) is an alternative
current (AC) supply in the range of radio frequency (RF), U a direct-current (DC) power supply
and ω the angular frequency. For a given DC and RF only specific ions have stable trajectories

5
and can pass through the quadrupole, the rest will be affected by the DC or RF and will collapse.
Selecting ions can be done by two methods: either by keeping the ω constant (therefore, RF is
constant) and alternate U in such as the ratio of U/ V remains constant, or by keeping U constant
and alternate the ω so that U/V remains constant. In the first case, where U is alternated, heavy
ions will have unstable trajectories due to the effect of defocusing from the DC and will collapse
on the rods. Meanwhile, lighter ions will be stabilized and pass through the quadrupoles, upon
RF application whenever the amplitude of the trajectories inclines towards expansion. In the
second case where RF is alternated, the light ions are discharged by receiving energy from the
field resulting in their oscillation with progressively large amplitude until they will crash in one
of the rods. Therefore, in the first case only light ions will go through while in the second case,
only high masses ions will pass through the field and reach the detector (de Hoffmann &
Stroobant, 2007). A detailed explanation of the quadrupole theory can be found in (de Hoffmann
& Stroobant, 2007).

Fig. 2. Diagram of Quadrupole MS adapted from Shimadzu (www.shimadzu.com on 7th November 2018).

The quadrapoles are used to select a narrow range of m/z or one particular m/z based
on the method described above. This particular selection of ions makes the quadrupoles highly
selective. In full scan mode, they have low sensitivity. However triple quadrupole set ups, have
the highest sensitivity in selected reaction monitoring (SRM) or multiple reaction monitoring
(MRM), where a specific m/z or particular multiple m/z are selected, respectively, and their
fragments are being monitored throughout the analysis. In the case of SRM, duty cycle reach
100 %. SRM and MRM scanning methods are being used frequently for quantitation studies
(Dunn, 2011).

In proteomics, quadrupoles are often mounted in a time of flight analyzer (qTOF)

(Armirotti, Benatti, & Damonte, 2009; Griffin et al., 2001) or an Orbitrap (hybrid quadrupole-

6
Orbitrap) (Michalski et al., 2011; Shalit, Elinger, Savidor, Gabashvili, & Levin, 2015) and are
used to isolate specific m/z range before entering the collision cell.

Time of flight
TOF analyzer was first described by Wiley and McLaren in 1955 (Wiley & McLaren, 1955).
Homogenous electric field with potential V is applied to ions with different masses, and TOF
separates ions based on their velocity upon drifting in a region free of electrical field known as
free ion path or flight tube (de Hoffmann & Stroobant, 2007). The time that an ion takes to
move from the source to the detector in a free ion path determines the mass to charge ratio. The
time is given by:

𝐿 (1)
𝑡=
𝑣
where L is the length of the tube and v the velocity of the ion. The energy in the free-field region
becomes kinetic and the velocity is derived by the mass–energy relationship:

𝑚𝑣 2 (2)
𝐸𝑘 = = 𝑞𝑉 = 𝑧𝑒𝑉 = 𝐸𝑒𝑙
2
where Ek and Eel is the kinetic and electric energy respectively, m is the ion mass and V is the
electrical potential through which an ion having total charge q = ze. Rearranging eq. 2, the ions
v can be calculated the according to:

(3)
2𝑧𝑒𝑉
𝑣= √
𝑚

by replacing v in Eq. 1 with the Eq. 3, given that everything else is held constant in Eq. 4, the
time becomes a function of the square root of m/z.

𝑚 𝐿 (4)
𝑡= √ ∗
𝑧 √2𝑒𝑉
The Eq. 4 shows that the heavier an ion is, the slower it will arrive to the detector. All
equation have been derived from (de Hoffmann & Stroobant, 2007).

The resolving power R of a TOF analyzer, defined as R = m/Δm where Δm is the full
width at half-maximum, is limited. This is due to spatial energy distribution and time-spread of
ions with the same m/z which are caused upon ionization and dissociation of molecules (Wiley
& McLaren, 1955). In order to achieve high resolution, then both have to be minimized. Time-

7
spread of ions is tackled with the use of pulsed delayed extraction when MALDI is used as
ionization source, proposed in the initial publication of Wiley and McLaren [12] but
implemented later (Vestal, Juhasz, & Martin, 1995), while the spatial energy distribution is
resolved with the use of electrostatic mirrors (Mamyrin, Karataev, Shmikk, & Zagulin, 1973). In
delayed extraction, a time lag between generation of ions and their extraction from the source is
employed. During the lag time, the ions move to a new position. Then an extraction potential is
applied and ions with same m/z and low initial kinetic energy will have greater potential while
the same m/z ions with greater initial kinetic energy will be found more away from the surface
at which the potential is applied, therefore will have lower potential energy. In other words, ions
with lower velocity are found near the surface therefore will receive higher potential than the
ones that have greater velocity and are found away from the surface. With correct extraction
time and potential, the identical m/z ions can be focused and arrive to the detector in the same
time.

Another way to increase resolution and sensitivity in TOF analyzers is the extension of
drift tube and the application of reflectron. By increasing the flight tube, an enhancement of tens
of thousands at m/z at <300 m/z can be achieved and a mass accuracy greater than 1 ppm
(Satoh, Tsuno, Iwanaga, & Kammei, 2005).

Additionally, the use of reflectron, which was invented by Mamyrin et al (1973)

(Mamyrin et al., 1973) space-time focuses the ion packets in terms of energy near the detector.
As the ion are travelling from the extracted surface to the detector and a pulsed delayed
extraction is applied, even though they seem to have been focused in terms of space, actually the
packet of ion expands as a result of the difference in the velocities of ions as well as the time of
motion of the ions in the free-field space. Reflectron is focusing the ion packet by equalizing the
time of motion for the ions with same m/z before reaching the detector. To achieve that a gap
with a retarding electric field followed by a uniform electrostatic field are placed in the end of
the drift space known as secondary focusing - reflectron. By choosing the correct parameters,
one can achieve the time of an ion takes to go from the source to the detector to be independent
of energy. As the ions slow down before entering the reflectron, the faster ones will have slower
deceleration and will penetrate deeper into the reflectron, while the slower ones will not
penetrate as deep as the fast ones and will go around earlier than the former. Based on the

8
centripetal force both fast and slow ions will exit the reflectron with the same speed at the same
time arriving together to the detector.

TOF, usually coupled with MALDI, is a rapid instrument with simple design, records a
mass spectrum in microseconds, has theoretical unlimited mass range and with use of delayed
extraction and reflectron one can achieve high sensitivity (Mamyrin et al., 1973), allowing the
simultaneous transmission of all ions (de Hoffmann & Stroobant, 2007). It is used in proteomics,
and SWATH-MS, an upcoming method acquisition in proteomics was originally developed on
such this analyzer (Gillet et al., 2012).

Ion traps
Paul ion traps were invented by Wolfgang Paul and Hans Dehmelt, who shared the Nobel Prize
in Physics in 1989 for this work (Stafford, 2002). It consists of four parallel linear ring electrodes
with four positive end cap electrodes which have a hole for injecting into and ejecting ions out
of the trap (Fig. 3) (Stafford, 2002). The ions are injected from the source through the ion optics
into the trap until it will fill up, for a given time, which is regulated electrostatically by the
“aperture” of the interoctapole lens. When the RF voltage on the ring electrode known as storage
voltage is low, ions with a m/z above a certain level are trapped. The ions which tend to arrange
themselves in order are confined along a line in the central axis of the trap by applying an RF
voltage to the end-cap electrodes. By increasing the amplitude of RF voltage applied to the ring
electrodes, the trajectories of the ions destabilize, and the ions are ejected through the
aforementioned holes, in order of increased m/z, towards the electron multiplier to produce a
mass spectrum (Stafford, 2002). The particular RF voltage that ion needs to have applied on it
in order it to be ejected from the mass analyzer to the detector is called resonance voltage.

9
Fig. 3. Schematic representation of an ion trap as adapted from (Georgiou & Danezis, 2015)

By applying the appropriate voltage and RF amplitudes the ions can be confined in the
center or selected and ejected towards the detector. Selection of parent ion is done by applying
a resonance RF voltage which will force all ions to exit the trap except the ones we are interested
in. To fragment the parent ion with collisional induced dissociation, helium damping gas is
infused inside the trap. The kinetic energy of the ions is increased by applying low amplitude RF
voltage on end caps such as to increase resonance excitation but avoid resonance ejection. As a
result, the ions collide with the helium damping gas and fragment. The same injection method
is used to record the fragments ion spectra. The use of helium gas is not only used to fragment
the selected ion or ions but also is beneficial to confine the ions in the center of the trap by
collision kinetically cooling of the trapped ions. This increases sensitivity and mass resolution of
the instrument (Huber & Hölzl, 2001).

Ion traps can be functioned as mass analyzers as well as ion storage devices (Clarke,
2017). With the use of ions traps, experiments can be conducted in space and time (Huber &
Hölzl, 2001). This means that manipulation and analysis of ions can be done simultaneously in
different regions of the instruments, respectively, and thus occur to perform multiple stages of
mass spectrometry, separated in time. MSn can be performed. Additional advantage is that since
ions are formed or injected into the ion trap ions in a non-continuous or pulsed fashion the
noise is low. However important disadvantages of Paul traps are that they have low accuracy
and insufficient mass resolution, limited linear range and relatively low space/charge capacities
(Perry, Cooks, & Noll, 2008).

10
Orbitrap
To tackle the disadvantages of linear ion traps, commercial Orbitraps (Fig. 4) were introduced
by Makarov in 2000 (Makarov, 2000). The architecture of orbitrap is cylindrical and is based on
Kingdon traps (Kingdon, 1923). It consists of one external electrode which is split in two parts,
one for excitation and one for detection, and one central wire cathode known as central electrode
with flat end caps electrode to surround the trapping volume (Makarov, 2000). A DC voltage is
applied between the central electrode and the outer electrode producing a radial logarithmic
potential, which traps the ions in the radial direction. A quadrupolar potential produced due to
the shape of the outer electrode confines the ions axially. This configuration allows the ions to
undergo harmonic oscillation along z direction, while RF voltage applied to the outer electrodes
excites the ions to start oscillating along z direction. Τhe injected ions with mass m, charge q
and initial velocity v, will have a stable orbit around the central electrode, known as orbital
trapping, only if the potential difference between the wire and the outer electrode is greater than
a specific value derived from equations described in (Perry et al., 2008).

Fig. 4. Orbitrap architecture. Background images were taken from www.massspecpro.com/mass-

analyzers/orbitrap on 18 Oct 2018 and were adapted to create a complete scheme with explanations.

11
 High vacuum conditions (10-8 Torr or less) are generated in the space between the
external and internal electrode which are connected to different power supplies. For increased
sensitivity, ions are injected in short time known as bursts or packets, into the Orbitrap
(Makarov, 2000). Linear ion traps are used to store the ions in low pressure, which are coming
from high pressure ionization source making possible to couple an Orbitrap to a pulsed or
continues ionization source. Another function of quadrupole ion traps is to control the number
of ions entering the Orbitrap. Between the storage trap and the Orbitrap, a set of lenses are
placed to accelerate and focus the ion beam. For fast injection and further increased sensitivity
curved ion traps (Makarov, Denisov, Kholomeev, et al., 2006) are preferred injecting ions at a
position offset from the Orbitrap’s equator (Makarov, Denisov, Lange, & Horning, 2006).

After excitation of the ions, the ions axially oscillate back and forward between the
opposite rods, producing an electric current (image current) which is detected on one of the part
of the outer electrodes and amplified by a differential amplifier. The m/z ratio is derived from
the axial frequency of harmonic ion oscillations which can be determined from using image
current detection and Fast Fourier Transform (FFT) functions (Makarov, 2000). Another way
of detection it is the mass selective instability, which is detection using secondary electron
multiplier. However, this detection mode will not be discussed in this work. Fig. 4 summarizes
all parts of an Orbitrap presented above.

1.2.3 Acquisition modes

For the last years, the data-dependent acquisition (DDA) mode has been the golden standard
for untargeted proteomics experiments (Hu, Noble, & Wolf-Yadlin, 2016) while single or
parallel/multiple reaction monitoring (S-P/MRM) is routinely used for targeted proteomics.
However due to technical advances and improvements in algorithms, a new acquisition mode
has emerged known as data-independent acquisition (DIA) mode. In this work a review of DDA
and DIA modes, which have been used for the project, will be given, while S-P/MRM will not
be presented.

Data Dependent acquisition

Data-dependent acquisition (DDA) and also known as shotgun proteomics, is the most widely
used method for proteomics-based mass spectrometry (Mann, Hendrickson, & Pandey, 2001).
Over the course of chromatographic elution, in each duty cycle, a short full scan (MS1) is
performed over the selected mass range followed by the isolation fragmentation and sequence

12
of the highest abundance peptides (MS2) (Fig. 5). The MS1 is used to monitor peptide intensity
and identify target peptides. The MS2 will only select a fixed number of precursor ions set by
the user (e.g. top 5 peptides) and will be fragmented in order of decreasing intensity. The
instrument detects when a peptide starts eluting and predicting its peak, will perform DDA at
its maximum abundance, in order to produce high quality MS2 spectra.

Dynamic exclusion window is used to make sure that peptides that have been selected
and fragmented are not re-selected within a given time, in order to give space to other peptides
to be selected and recorded as well (Domon & Aebersold, 2006).

During DDA, a set of parameters defined by the user have to be selected and optimized.
Resolving power, ion population (AGC target value), and maximum ion injection time for the
MS1 and MS2, number of microscans per MS/MS scan are some of them. A detailed review and
suggested values for Orbitrap Elite and Orbitrap XL is given by Kalli et al (2013) (Kalli, Smith,
Sweredoski, & Hess, 2013).

Fig. 5. Data depended acquisition, scheme explanation.

Data Independent Acquisition

Data Independent Acquisition (DIA) is less used mode, however significant attention has been
given lately to it (Koopmans, Ho, Smit, & Li, 2018). In the previous mode, only one peptide

13
per time was selected and fragmented and its MS2 spectrum was recorded. In DIA mode, a
predefined m/z window is selected and all the peptides that fall into that window are fragmented.
There are several DIA techniques. The most common is the Sequential Windowed Acquisition
of all THeoretical fragment ion mass spectra (SWATH) (Gillet et al., 2012). In this approach,
(Fig. 6) every 3 seconds, which are defined from the chromatography (black dots during the
elution, the instrument does a full scan, of a scan range defined by the user (e.g. 500 – 900 m/z).
Between the three seconds, the instrument selects and isolates the predefined m/z windows
sequentially, in chronological order (colorful dots), and fragments the peptides within each
window range. One MS2 spectrum per window is recorded containing the fragments of all
isolated peptides. If the user decides to have 10 windows to scan the range of 500 – 900 m/z,
this means that the instrument will select and fragment windows with a step of 40 m/z. After
the full spectra (first black dot) is recorded, the instrument will select the window from 500 –
540 m/z, fragment it and register its fragments, the next moment (purple dot), will select 540 –
580 m/z fragment it and register the fragments, and will continue until 900 m/z, until the next
full scan take place (next black dot). Therefore, in case the analysis has 20 windows, 20 windows
have to fit in in 3 seconds instead of 10, is halving the SWATH duty cycle (the time that it spends
the instrument on analyzing one SWATH window) therefore lowers the sensitivity, since less
ions arrive to the detector. From the other hand, larger isolation window in higher probability
having fragment’s interferences, since more peptides are fragmented. Therefore, resolution,
mass accuracy and specificity decreases with larger SWATH step size (Gillet et al., 2012). So
tuning a SWATH analysis is a compromise between sensitivity, specificity, resolution and
depends on the nature of the sample. For accurate coverage of the complete isotopic pattern of
the parent ion, the SWATH windows should partially overlap (Gillet et al., 2012). The collision
energy of the SWATH window alters ±15 eV to the point where complete fragmentation of
doubly charged parent ion found in the middle of the window is achieved.

DIA is performed on high-resolution MS. The SWATH-type of DIA was originally

developed in qTOF due to the high performance of the analyzer and adequate SWATH duty
cycle that the analyzer can provide (Gillet et al., 2012). However, DIA on trybrid orbitraps
showed to have higher sensitivity (Senko et al., 2013; Sidoli, Fujiwara, & Garcia, 2016). In trybrid
orbitraps like the Orbitrap Fusion™ Lumos™ Tribrid™ Mass Spectrometer, selection and
isolation of m/z windows take place at the quadrupole, then the selected ions will go to multipole
where fragmentation takes place, and the fragments will move to C-Trap and injected into the
14
Orbitrap in order to be analyzed and recorded. During fragmentation of the first window,
selection of the second window is performed. This is repeated throughout the full
chromatographic elution gradient.

MS2 is a highly complexed spectra since contains the fragments of every peptide found
within the window. In order to deconvolute the spectra, a reference library with information
about fragments, precursor peptides and their elution time, is needed as well as state of the art
algorithms to match fragments with precursors. Library is built either with extreme fractionation
of the sample and DDA acquisition or in silico with sophisticated algorithms. Software that have
this option are DIAUmpire and Spectronaut with PulsarX. Additionally, due to the MS2
complexity a lot of effort has been given to the development of sensitive software. DIAUmpire
(Tsou et al., 2015), mapDIA (Teo et al., 2015), OpenSWATH (Röst et al., 2014), are some of the
software that have been introduced for the deconvolution and statistical analysis of DIA-MS2
spectra, and are constantly upgraded and evolving.

DIA is believed to be advantageous over DDA (Gillet et al., 2012). Even though DIA
requires DDA acquisition, a significant difference between the modes, is while DDA is biased
to the most abundant peptides, in DIA mode all peptides, including the lowest abundance are
recorded. Additionally, due to the semi-stochastic selection of most abundant peptides in DDA
mode, it is a challenge to reproducibly quantify low abundant peptides. From the other hand,
DDA has higher specificity than DIA, since it is selecting specific fragments. However,
foreseeable future improvements of the performance of mass spectrometers, promise a DIA
isolation window of only some Da, increasing by that the selection and accuracy of the precursor
and competing the specificity of DDA or other high specificity acquisition modes (i.e. SRM).
Another significant advantage of DIA is that the data can be stored and re-analyzed, as well as
even though is untargeted acquisition mode, one can perform targeted data extraction (Gillet et
al., 2012). It has been shown that DIA has, less missing values, better CV across replicates, and
shorter analysis time than DDA (Gillet et al., 2012). DIA is an unbiased, untargeted, more
sensitive acquisition technique.

In this work, we utilize an LTQ Orbitrap XL operating in DDA mode for the
optimization of the analytical method, and an Orbitrap Fusion™ Lumos™ Tribrid™ operating
in SWATH mode, for the main study. Library was not built for the specific analytical procedure,
as it is advised to do so, but was previously generate by colleagues in the laboratory using the

15
same instrument but using a different protein extraction procedure from the one adopted in this
work.

Fig. 6. Data independent acquisition DIA-SWATH, scheme explanation

1.3 Model cell

Saccharomyces cerevisiae was selected as model cell. Saccharomyces cerevisiae it is generally used by the
proteomics community to establish LC-MS-based approaches (Paulovich et al., 2010). Upon
normal growth conditions a significant number of proteins are expressed. The abundance level
of the expressed proteins varies from fewer than 50 to more than 106 molecules per cell. It has
an extensively characterized proteome with ∼ 6,000 proteins mapped (Picotti et al., 2013). To a
more extent, it is inexpensive and easy to produce (Paulovich et al., 2010).

1.4 Challenges of the study

1.4.1 Exposure conditions
Finding the correct exposure conditions, which would fit with the downstream analysis for
proteomics analysis, was a challenge. The project has entailed a significant optimization of the
light microscopy exposure conditions. This optimization was an important and challenging
process and included parameters such as wavelength, light intensity, imaging chambers, cell

16
density, and others; however, we consider this aspect somewhat beyond the main scope of this
thesis and we will therefore not discuss it in detail.

1.4.2 Miniaturization
The final exposure conditions were set on 60 μL of OD 600 = 0.4, which is a significant low
number of yeast cells. Yeast cell has a rigid wall composed by polysaccharides and
mannoproteins, around the plasma membrane (Cabib et al., 2008), which it is difficult to brake.
A wide range of methods have been developed including, mechanical disruption like sonication,
high pressure homogenization, or chemical lysis (e.g. zymolase), and autolysis, and are applied
for the lysis of usually high volume of yeast cells. However for minute samples in the range of
milliliters, a method was developed by Wenger et al (2008) (Wenger et al., 2008) which make use
of ultrasonication. Therefore, miniaturization for micro-proteomics analysis and in the same
time achieving a significant number of protein IDS, was a great challenge as well. Based on the
routine proteomics sample preparation method [2], a novel lysis method adapted from (Wenger
et al., 2008) but downscaled to 100 μL of yeast cells, is proposed. Additionally, several analytical
procedures included FASP technology and fractionation, were tested to increase the
identification and quantification of protein IDs. Nano LC was used to increase sensitivity and
nano electrospray to permit the loading of low sample volumes (Meher & Chen, 2017), without
compromising the quality of the data.

2. Methods
2.1 Chemicals
Methanol and Acetonitrile were obtained from Fisher Scientific (United Kingdom). Ammonium
bicarbonate was obtained from Sigma Aldrich (Germany). RapiGest SF Surfactant was
purchased from Waters (USA). Urea obtained from Eurobio (France). Tris(2-
carboxyethyl)phosphine (TCEP), Iodoacetamide (IAA) and formic acid (FA) 98-100% purity,
all obtained from Sigma Aldrich (USA). Lys-C and trypsin obtained from Promega (WI, USA).
The CRZ1-GFP strains (strain BY4741 crz1-GFP::HIS3) stems from the yeast-GFP collection,
which was purchased from Invitrogen and comes from (Huh et al., 2003). Yeast Nitrogen Base
(YNB) obtained from Chemie Brunschwig AG (Switzerland), Glucose from Sigma Fluka Aldrich
(Switzerland), concanavalin A (ConA) from Life Technologies (Switzerland). Amino acids and
methionine were generously given by Weis Lab at ETH Zurich. Milli Q water was purified on

17
Milli-Q Integral 3/5/10/15 Systems (Millipore SAS, France) and was used for all the solutions
except otherwise noticed.

2.2 Solvents and solutions

Medium consisted of 50 mL of 10x 20% glucose, 50 mL 10x Yeast Nitrogen Base, 50 mL of 10
x amino acids without methionine, 5 mL of 100 x Methionine, all dissolved in deionized MilliQ
water and filled up with deionized MilliQ water until 500 mL. Lysis buffer was made of 0.9 mL
0.1 M Ammonium Bicarbonate (AmBic), 9 mL 8.9 M Urea in 0.1 M AmBic, and 100 μL of 1
mg/mL Rapigest dissolved in 0.1 M AmBic in order to achieve 8 M final concentration of urea.
Mobile phase A consisted of water:FA (99.9:0.1, v/v) and mobile phase B consisted of ACN:FA
(99.9:0.1, v/v). All chemicals were dissolved in MilliQ water.

2.3 Cells preparation

The cells grew on liquid medium. Yeast cells were streaked out onto an agar plat using a sterile
stick in order to form individual colonies. The plate was inverted and placed in a 30 °C incubator
for 2 days. After colonies were formed the cells were inoculated by dissolving a colony in 2-3
mL and twisted overnight at 25 °C. The next day the optical density at 600 nm (OD600) was
measured on a Denovix DS-11 Series (DeNovix, Wilmington, DE, USA) spectrophotometer
and the cells were dissolved in medium such as after an overnight growth to reach OD600 =
0.38-0.42. The volume of cells to be dissolved so to reach the optimal number of cell after
specific time (target time) of growth was determined according to:

Target OD600
( 𝑡𝑎𝑟𝑔𝑒𝑡 𝑡𝑖𝑚𝑒 ) ∗ 𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑑𝑖𝑙𝑢𝑡𝑒𝑛𝑡
𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑐𝑒𝑙𝑙𝑠 𝑤𝑖𝑡ℎ 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑂𝐷600 = 2 2
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑂𝐷600

Cells grew on 1 mL of medium overnight at a 24-well cell culture plates (Thermo Fisher
Scientific, USA) placed in a Kühner shaker incubator fixed at 30 °C and 150 rpm.

2.4 Light exposure

The cells were exposed using meander as scanning method (Fig. 7) to microscopic light using a
100x oil objective (CFI Plan Apo Lambda 100× Oil, Nikon) on an epi-fluorescence microscope
(Eclipse Ti-E, Nikon), equipped with a 455 nm high power LED light source (LedHUB light
engine, Omicron), a 455–490 nm excitation, and a 505–550 emission width filter set (EGFP ET

18
Filter set, Chroma Technology Corporation). Images were collected every 2 min from the center
of the well using a CCD camera (Zyla sCMOS, Andor). All the measurements were performed
in a 384-well plate (MatriPlate 384Well Plate, Glass Bottom, Brooks).

Two different conditions of exposure to the wavelengths of 475 nm and 542 nm were
tested in the main study along with a positive and negative control. Positive control is defined
as exposure to UV light. Negative control is defined as no exposure at all, but bright field images
were taken every two minutes in order to follow cell growth. Since CRZ1 has been found to
localize to the nucleus upon cell stress caused from light exposure (Bodvard, Jörhov, Blomberg,
Molin, & Käll, 2013), CRZ1 fused to green fluorescent protein (GFP) was used as main reporter
of cell exposure to light. For the negative control, since there was no exposure, cell growth was
used as reporter.

20 µl of 0.1% concanavalin A (ConA) was spotted onto the well and left overnight to
dry. Next day excess of ConA was gently removed and 60 μL of the grown yeast cells with
OD600 = 0.38-0.42 were deposited onto the well. The plate was centrifuged for 1 min at 300 g
and placed on the microscope for exposure. All exposure experiments took place at 30 °C.

Fig. 7. Meander scanning method. The well is scanned from up left to down left with that scanning method. A well
scan takes approx. 2 min. When the microscope finished the scanning, a new cycle starts. In total, 60 cycles were
repeated, which makes a total exposure of 120 min.

2.5 Lysis
The cells were removed from the microscopic plate by pipetting vigorously up and down and
were placed on an Eppendorf tube. They were pelleted with 30 min centrifugation at maximum

19
speed on a Centrifuge 5417 R (Eppendorf, Germany). The medium was gently removed and the
pellet was reconstituted in 120 μL of lysis buffer being ready for cell disruption.

Two methods were tested for the production of yeast cell lysates. The widely used
mechanical glass bead disruption method was tested against acoustic focused ultrasonication. In
a low binding protein 1.5 ml Eppendorf tube, 100 μl of glass beads acid-washed, 425-600 μm,
30-40 U.S. sieve (Sigma-Aldrich St. Louis, Mo, USA) in water were placed, and the water was
gently removed. Then the sample was added and the cells were disrupted by 10 min vortexing
at 4 ºC on a Mini-Bead Beater (BioSpec Produts, Inc.). Following the disruption, the sample was
moved to a new clean low binding protein 1.5 ml Eppendorf tube.

Ultrasonication was performed on a Covaris S 220 focused ultrasonicator (Covaris,

Woburn, MA, USA). The sample was moved to a 120 μL Sovaris vial microTube (Covaris,
Woburn, MA, USA), and the vial was placed in the Covaris ultrasonicator bath. We used the
optimized instrumental settings for 1 mL yeast lysis by Wenger et al (2008) (Wenger et al., 2008)
and adjusted to the micro-volume of 120 μL. The lysis took place in 10 min focused
ultrasonication with peak power 120 W, duty factor 20 and 200 cycles/bursts at 4 ºC – 8 ºC.

For the main study, the cells were not pelleted. Immediately after imaging the medium
was gently removed and 60 μl of the lysis buffer was added to the plate to flash freeze any
activity. The cells were detached from the glass bottom by vigorously pipetting up and down
and were inserted into the Covaris vial. Then 60 μL of lysis buffer were added again to the well
and by pipetting as described above and the last cells that were attached to the bottom were
moved to the same Covaris vial. Final volume of the vial was 120 μL.

2.6 Digestion
Prior to the digestion the protein amount was estimated with bicinchoninic acid (BCA) assay
(Pierce BCA Protein Assay Kit, Thermo Scientific, USA) according to the manufacturer
protocol. Detailed explanation of BCA can be found in (Smith et al., 1985) and (Walker, 1994).
The micro-BCA was used, therefore 10 μL of the sample was used for BCA protein quantitation.

Following the lysis, the remained samples with volume 110 μL were reduced with 5 mM
TCEP dissolved in 0.05 M AmBic, incubating at 400 rpm in a Eppendorf Thermomixer Comfort
(Eppendorf, Germany), for 30 minutes at 37 ºC. The samples were alkylated by 10 mM IAA and
left in darkness for 30 min at RT. Following the reduction and alkylation, the urea concentration

20
was reduced from 8M to 4M with 50mM AmBic and Lys-c in a ratio 1 μg of enzyme in 50 μg of
protein, was spiked into the sample. The digestion took place at 37 ºC incubating in 400 rpm for
2 to 4 hours. Following Lys-C digestion, the sample urea concentration was further reduced to
2 M, and trypsin in 1:100 (enzyme:protein) ratio was spiked into the sample. The proteins were
digested overnight at 37 ºC shaked at 650 rpm.

Beyond this method, FASP technique as described in (Potriquet, Laohaviroj, Bethony,

& Mulvenna, 2017) with cut off 10 kDa was tested in triplicates in order to access whether it
results in higher protein ID identifications or not.

2.7 C-18 purification and desalting

The digestion was stopped with addition of 100% formic acid (FA) to a final concentration of
2%. Following the acidification of the sample, a C18 clean up using StageTips (Yu, Smith, &
Pieper, 2014) centrifuged at 300 g for 1 min each time, was performed as follows: The column
was activated with 100 μL of 100 % ACN and condition with 2 times of 200 μL 0.1% FA. The
sample was loaded and the column was washed 4 times with 100 μL of 0.1% FA in order to
remove salts, detergents and excess of urea. The peptides were eluted with 3 times 50 μL of
ACN:FA (50:0.1 v/v). After elution the solvent was evaporated in Speedvac (Centrifugal
Vacuum Concentrators Labconco, MO, USA) at 30oC for 4 hours. Following evaporation the
peptides were reconstituted in 10 μL ACN:FA (2:0.1 v/v), sonicated for 10 min in a Bransonic
5510 bath sonicator (Emerson, CT, USA) centrifuged at maximum speed for 7 mins, and the
upper part was gently transferred to an LC vial. The samples were snap-freezed in liquid nitrogen
and stored at -80 ºC until analysis.

2.8 Fractionation
Fractionation was tested in triplicates in order to assess whether it results in higher protein ID
identifications or not. Fractionation took place in the last step of C18 purification were the
peptides were eluted in increased order of hydrophobicity with increase order of 100 μL ACN
concentration as shown in Fig. 8. The resulted seven fractions were dried down and
reconstituted as described in C18 purification. Following reconstitution, sonication and
centrifugation, the seven fractions were pooled in one LC vial.

21
Fig. 8. Fractionation of the sample consisted of elution of the peptides based on the peptides’ hydrophobicity with
increasing concentration of organic solvent.

2.9 Instrumental method

Analytical method development was performed on an LTQ Orbitrap XL (Thermo Fisher
Scientific, San Jose, CA, USA) on DDA mode. The final study was run on an Orbitrap Fusion
Lumos (Thermo Fisher Scientific, San Jose, CA, USA) on DIA mode with SWATH acquisition.

DDA LC-MS/MS settings

Separation of peptides was achieved on a C18 (11 cm * 360 μm, homemade pack with 3 μm
particle size) with PicoTip emitter (New Objective, Woburn) column mount on an EASY-nLC
1200 System (Thermo Scientific, San Jose, CA, USA) with cooled auto sampler over an 80
minutes gradient. Mobile phase A consisted of ACN:FA (2:0.1, v/v) and mobile phase B
consisted of ACN:FA (100:0.1 v/v). The gradient started with 7% mobile phase B for 5 minutes,
increased to 35% mobile phase B in 60 minutes. To clean the column the gradient reached 80%
in 2 mins and remained steady for 10 minutes. Injection volume was 4 μL which according to
BCA was 6 μg of protein at a flow rate of 300 nl/min. However, BCA overestimated the protein
content, which will be discuss in the Results and Discussion section.

DDA mode was used, with selection of TOP 10 peptides in positive polarity on FT-FT
detection mode and CID. FT-FT detection mode means that both precursor and fragments,
both are analyzed in the Orbitrap. In order to gain higher speed of the analysis one could use
the LTQ Linear Ion trap to scan the fragments (FT-IT mode), but this would come as a cost of
the mass resolution (Kalli et al., 2013). Scan range of MS1 set at m/z 350 -1600. Resolution at
FWHM was set at 60k, scan time 75 min and isolation width 2 m/z. Capillary temperature was
set at 200 ºC, capillary voltage at 60 V, and collision energy at 32 %. Number of microscans set
to 1 as more microscans requires higher scan cycle time therefore returns lower MS/MS spectra.

22
Automatic gain control for MS and MS/MS was set at 106 and 2x105 ions, respectively.
Maximum ion injection time for MS and MS/MS was set at 500 and 1000 msec, respectively

DIA LC-MS/MS settings

The peptides were separated on a reversed phase LC column C18 (25 cm * 75 μm, 2 μm particle
size, Acclaim PepMap 100) mounted on an EASY-nLC 1200 System (Thermo Scientific, San
Jose, CA, USA) with cooled auto sampler. Mobile phase A consisted of 0.1% FA and mobile
phase B consisted of ACN:FA (80:0.1 v/v). The gradient started with 5% mobile phase B for 5
minutes and went up to 37% mobile phase B in 120 minutes. To clean the column, it went up
to 100 % mobile phase B in 2 minutes and remained for 10 minutes. Injection volume was 8 μL
at a flow rate of 300 nl/min. The resolution, the maximum injection time and the isolation
window were optimized as described below.

DIA SWATH/MS was used. Instrumental settings were collision energy was set at 27%,
with AGC target at 5 x 105 and 1 microscan and the resolution of the MS2 as well as the fill time
was optimized. Fragment scan range was set from m/z 500 to 900 as this is the peptide-dense
region of Saccharomyces cerevisiae in a tryptic digestion (King et al., 2006). Three different settings
were tested, for 30,000; 60,000 and 120,000 resolution with fill times 54 ms, 118 ms, and 246 ms
respectively. The DIA isolation window scheme was set to 40, 20 and 10, for each resolution,
respectively. Table 1 summarizes the resolutions that were tested along with the corresponding
fill times and isolation windows. The settings were compared based on reproducibility -
Coefficient of Variation (CV) - among triplicates and the percentage of missing values in each
group. Data analysis was performed on RStudio.

Table 1. MS2 settings were tested for optimization of the method

500 – 900 m/z Resolution Fill time DIA isolation

window
Setting 1 30,000 54 40
Setting 2 60,000 118 20
Setting 3 120,000 246 10

23
2.10 Computational analysis
DDA
For the optimization of the methods, CV per protein and number of protein IDs were used to
evaluate the DDA results. Protein identification took place using Trans-Proteomic Pipeline
(Deutsch et al., 2010). Parent mass error unit was set to ± 15 ppm, fragment mass error ± 4 Da,
2 missed cleavages, Trypsin digestion, with Carbamidomethylation as fixed modification. The
matching of tandem mass spectra with their peptide sequences was done on Xtandem (Craig &
Beavis, 2003) and Omssa (Geer et al., 2004) search engines. Discovery Rate (FDR) was set at
0.01 cutoff and FDR type iprophet-pepFDR. FASTA-formatted protein sequence lists was taken
from Saccharomyces Genome Database (SGD) was inserted in the search. In the TPP results
1% FDR was introduced. To access the number of protein IDs, the triplet results were merged,
and Bioinformatics & Evolutionary Genomics Venn Diagram web tool
(http://bioinformatics.psb.ugent.be/webtools/Venn/) was used to illustrate the number of
protein IDs per method tested.

The CV of the proteins was calculated by label-free quantification (LFQ). LFQ

intensities per protein were generated in MaxQuant (Cox & Mann, 2008). MaxQuant parameters
were set to default except the following: max missed cleavages to 2, variable modifications were
set to Acetyl N-term, Oxidation and Phospho and fixed modifications to carbamidomethylation,
digestion mode to Trypsin/P, main search peptide tolerance at 4.5 ppm and max number of
modifications per peptide set to 3. An LFQ intensity per protein was generated in MaxQuant,
by summing up all the intensities of the peptides attributed to that protein. LFQ intensities per
protein were analyzed by R to estimate and visualize the CV of the triplicates, and perform
comparison among the methods.

DIA
The deconvolution of the complex DIA mass spectra fragment data, and their matching with
the parent peptide was done on Spectronaut software (Biognosys AB, Schlieren, Switzerland)
with a spectral library given over from colleagues. The library was previously generated in the
laboratory on the same instrument but adopting a different protein extraction procedure from
the one adopted in the current work. DIA files were analyzed using default settings with three
files specified as three replicate experiments.

24
 The Spectronaut generated data were deposit on mapDIA (Teo et al., 2015), were
retention time-based normalization of data as well as imputation and significance testing for
differential expression was performed. The preprocessing and statistical analysis was done on
fragment level. The data were first log2-transformed, and then were centered by the overall
median value in each fragment. The missing values were imputed as 0.9 x the smallest value for
the group in the row (imp = group .9). Control triplicate 1 was excluded from the analysis since
more than 50% of the values were missing. Fragments that were deviating from the median ±2
standard deviations in each sample (SDF) were disposed. The pseudo coefficient of variation
(PSEUDOCV) was set to 0.1 and the median intra-protein correlation cutoff (MIN_CORREL)
at 0.2. The list of minimal number of non-missing value for each group (MIN OBS) was set to
2.

Protein quantification and protein-level differential expression analysis using fragment-

level intensity data was also performed on mapDIA. The minimum number of fragments per
peptide (MIN_FRAG_PER_PEP) as well as the minimum number of the resulting peptides per
protein (MIN_PEP_PER_PROT), both were set 1. Maximum fragments per peptide
(MAX_FRAG_PER_PEP) and peptides per protein (MAX_PEP_PER_PROT) were set at
infinitive. Experimental design was set at Independent-Design which permits differential
expression analysis between groups of individual samples and conditions 475, 542 and UV were
tested against the Control, as well as 475 and 542 against UV.

Visualization (heatmaps) and Gene Ontology (GO) enrichment analysis of the mapDIA
data output analysis were generated and performed respectively, with Prohits (Knight et al.,
2017) with the following settings: FDR filter, 0.01; points to label,10; secondary FDR cutoff,
0.05; with performing of clustering with Euclidean distance and Ward’s type.

3. Results and Discussion

3.1 Exposure conditions
The optimization of exposure conditions, in such a way that the cells do not have time to recover
themselves after exposure and ensure stress, was succeeded and the defined settings were used
for the final experiment. However, the method establishment is not presented and discussed in
this work, as the focus of the work is on the measurement of phototoxicity with mass
spectrometry. The sample produced from the exposed conditions was minute sample of 60 μL

25
OD600 = 0.38, which is equivalent to 120,000 - 480,000 yeast cells. By roughly estimating the
amount of proteins produced from that number of yeast cells following the (Milo, 2013), we
found that protein content is roughly two orders of magnitude less than the amount of proteins
produced from the same number of human cells. Therefore, the samples of this study contained
yeast cells approximately equivalent to 1200 to 4800 human cells.

3.2 Sample preparation optimization

The routine sample preparation for bottom up proteomics is cell lysis and protein extraction,
quantification of protein content with BCA, followed by reduction, alkylation and digestion,
desalting/clean up, sample concentration with evaporation and sonication and performing the
separation of peptides and analysis with nLC-MS/MS (Fig. 9).

Fig. 9. Sample preparation workflow

BCA quantified the protein to 20 μg for glass beads lysis and 50 μg for ultrasonication
lysis. This is an overestimation and can be seen by the elution profile of the peptides (Fig. 10-A
& B). Fig. 10-A shows the chromatographic profile of a reference sample from which 1 μg yeast
protein content was injected into the LTQ Orbitrap XL, for analysis (reference sample was given
over the colleagues and information about its content was communicated orally), while the Fig.
10-B illustrates the chromatographic profile of yeast protein content of the current study’s
samples which according to BCA should contain 6 μg. It can clearly be seen that the sample of
this study has half of the intensity of the reference sample. Even though intensity and
concentration are not absolutely proportional, one can clearly understand that due to the halving
of signal intensity the sample of the current study could not contain 6 μg of protein. Later we
found that the reason why BCA overestimated the protein content is probably due to the assay’s
incompatibility with 8 M Urea. In the compatibility sheet of the assay was found to be written

26
that the assay is compatible with maximum 3 M Urea. Due to this incompatibility, we did not
estimate the protein amount in the sample therefore BCA was not taken into consideration for
the rest of the study.

The optimization steps were tested in triplicates and a bottom up proteomics sample
preparation without use of FASP, or fractionation, was applied to prepare the samples for nLC-
MS/MS. The samples were run on DDA mode in the LTQ Orbitrap XL with chromatography
and MS settings as described in the methods. For optimization of FASP and fractionation,
triplicates were used, and the same instrumental analysis method was applied. The methods were
evaluated upon the coefficient variation of the sum of the intensity response of the peptides that
belong to one protein, among triplicates and the number of protein IDs.

In order to ensure that an efficient yeast lysis was performed in our minute sample, we
tested two different yeast lysis methods. The most widely used for large amount of yeast cells
lysis method (Papanayotou, Sun, Roth, & Davis, 2010), the mechanical glass bead disruption
method was tested against the ultrasonication. Ultrasonication was firstly shown by Wenger et
al (2008) (Wenger et al., 2008), to be effective in the lysis of highly concentrated yeast cells
sample, on milliliter level. In this study we used the optimized settings by (Wenger et al., 2008)
and adjusted to microliter level of low amount of cells.

27
 RT: 0.00 - 75.01
23.79 NL:
100
A 19.76
23.05
24.01
1 μg protein content 7.21E8
TIC MS
95 thsofia_M1
21.72 807_190
90 18.19
25.74
16.76 26.41
85
13.38 13.67
28.72
80 13.08 29.86

75
32.28
70 33.01
11.14
65
8.58 33.38 35.30
60 35.42
Relative Abundance

8.43
55 7.55 36.58 40.42
7.14
40.71
50 6.98
6.80 43.98
45 44.11
47.07
40
49.72 51.25
6.09
35 51.51
5.63
51.73
30
55.72 57.59

25
60.13
20
62.54
62.70
15
63.10
10 64.11
71.02 72.55
0.18 65.13
5 67.02

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Time (min)

RT: 0.00 - 75.08

24.54 NL:
100
B 6 μg protein content (according to BCA) 7.35E8
TIC MS
95 thsofia_M1
808_328
90
22.89
85 22.79

80
17.19
75
39.55
70
39.68

65 16.93
25.60
22.33 26.89
60
7.46 31.47
Relative Abundance

21.90
55 39.90
12.64 21.11
50 16.69 35.32
16.26
45 8.34 37.25
8.81 34.01 61.55
11.49
40 69.90
44.01
35 6.98 44.68 47.93
61.71
46.80 62.53
30 49.69
6.18 69.73
52.66
6.05
25
5.68 54.07
61.39 63.28
20
57.41
54.31
15
63.45
10 64.12 69.56
70.84
5 0.31 0.99

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Time (min)

Fig. 10. Elution profile of yeast samples. A. Elution profile of injection of 1 μg yeast proteins from reference sample
B. Elution profile of injection of 5 μg yeast proteins according to BCA from the study’s samples

The results showed that ultrasonication produced 891 protein IDS against the glass
beating which produced 869 protein IDs. The coefficient of variation for both methods tested

28
was lower than 15% (Fig. 11), with glass beats to have greater spread among triplicates, in the
upper range of CV data points. Ultrasonication due to the fact that produced higher number of
Protein IDs, and the CV of protein intensities among triplicates was less than 15%, which is
considered generally acceptable for DDA mode, was selected for method of choice for the study.
Another important reason that ultrasonication was selected is that it has space for improvement,
since Wenger et al (2008) (Wenger et al., 2008) proved that ultrasonication reached the efficiency
of homogenization pressure, only when enzymatic lysis was used prior to ultrasonication .
Therefore, the choice of ultrasonication leaves space for improvement of the method, and
potentially more IDs to be extracted. Ultrasonication was selected and the continuation of the
workflow optimization was done with use of this lysis method.

Fig. 11. CV of the response of proteins in triplicates from glass beads and ultrasonication. Ultrasonication
performed slightly better.

The lysis buffer we used in this study, contained high concentration of urea in order to
denaturate the proteins and assist in the digestion. However, the digestion efficiency of the
proteases Lys-c and trypsin is at its maximum at half and ¼ of the urea concentration used in
lysis buffer, respectively. This requires dilution of the sample, two and four times respectively.
Since the initial material is limited, we tested FASP technology as a method to remove urea,
subsequently avoiding dilution of the sample. In FASP, a 10 kDa cut-off filter is added on a
tube, and whatever is smaller than 10 kDa will pass through and will be discarded. The removal
of urea is done before digestion with the addition of 40 μL of Ambic, 0.5 M, therefore the

29
dilution of sample is avoided. FASP technology was tested against the normal workflow, without
fractionating the sample, in triplicates.

The results showed the application of FASP produced 531 protein IDs while the typical
workflow, produced 573 protein IDs. The CV of the protein intensities among triplicated was
better with use of FASP technology (Fig. 12). However, both methods were similar, therefore
for convenience we continued with no FASP.

The fractionation of the sample took place, in order to test whether elution of peptides
after clean up based on the hydrophobicity would result in higher number of protein IDs. The
results showed fractionation did not increase the number of protein IDs (n = 505), but had
lower variation than the non-fractionated method (Fig. 12). Because fractionation is laborious
and the initial material for fractionation was too little, since the results were better in protein IDs
for non-fractionated and the CV was still less than 15%, it was selected for the final study. From
the other hand, since the fractions were pooled all together for analysis, this might not be an
accurate optimization step as all fractions should be analyzed independently. Independent
analysis of each fraction could not take place due to limited availability of the instruments.

Fig. 12. CV of the response of proteins in triplicates from FASP, Fractionation and No FASP – No Fractionation.

The DIA isolation window scheme was set to 40, 20 and 10, in order to span 500 - 900
m/z with resolution 30,000; 60,000 and 120,000 respectively. The results showed that resolution
of 60,000 quantified 8016 peptides (Table 2) while missing values were estimated at 0.09% of

30
the dataset. The coefficient variation of the peptide intensities among duplicates was better than
30,000 but worse than 120,000 (Fig. 13). However, because the other two methods, either had
a very spread CV of the intensities or the peptides identified were too low, with many missing
values, resolution 60,000 was the best compromise, therefore was selected for the main study.

Table 2. Results from the optimization of the SWATH resolution, injection time and window

Setting 1 Setting 2 Setting 3

Resolution (FWHM) 30,000 60,000 120,000

Fill time 54 118 246

DIA isolation window (m/z) 40 20 10

No Replicates 3 2 2

Missing values [%] 0 0.09 3.6

Peptides identified 8016 8016 7747

Fig. 13. CV of the response of peptides from different SWATH settings described in Table 1.

3.3 Data quality

Most of the master thesis time was spent on the optimization of the microscopy; the biochemical
extraction of proteins; and the acquisition method. An exhaustive evaluation of the acquired
data still needs to be carried out. What we will be presented in the next two paragraphs are
preliminary analyses. Given their preliminary nature and some unsolved statistical evaluation

31
problems met at this early stage, they must be understood as potentially indicative of the results
a more comprehensive analysis may generate, but by no means as validated or conclusive.

The analysis was carried out in a two-tiered process. First, spectra deconvolution and
peptide identification was carried out using Spectronaut in default settings. Second, peptide
abundance values across different conditions were used as input for the software mapDIA, a
statistical package for the evaluation and determination of differential abundance of analytes in
different conditions.

In order to ensure that the new lysis method we used did not selectively enriched or
depleted selectively any protein, we compared our data with the yeast proteome of Gene
Ontology Consortium, and we checked whether the method is enriching or depleting selectively
proteins which play a role in molecular function, biological process, or localized n a particular
cellular component. We found out that the dataset does not enrich or deplete selectively, protein
class/compartments (Fig. A- 1 in Appendices). This gave us confidence that the introduced lysis
method is indeed on a par with traditionally more widespread ones.

To evaluate the method sensitivity, we compared the dataset with the unified protein
abundance datasets of Saccharomyces cerevisiae proteome (Ho, Baryshnikova, & Brown, 2018). As
expected from the low starting material, we found that the dataset is selectively towards the high
abundance proteins (Fig. 14), while maintain an abundance distribution similar to the full yeast
proteome.

32
Fig. 14. Comparison of the yeast proteome datasets: as taken from (Ho et al., 2018) with the yeast proteome dataset
generated by the current method described in this study. The graph shows the frequency of the appearance of the
of number of proteins per cell. It revealed that our method is bias towards the high abundance proteins which was
expected due to low initial material.

3.4 Candidates
We next performed a statistical analysis using mapDIA and compared the sum of the intensities
of all peptides belonging to a protein found in the exposed conditions to the sum of the
intensities of all peptides belonging to a protein found in the control. We found 25 candidates,
which seemingly change their abundance upon microscopic light exposure. Fig. 15 shows a
heatmap of the 25 candidates fold changes. The red color represents that the cells increase the
quantity of the specific protein (upregulation of the protein expression) in response to the
microscopic light, while blue color represents the downregulation of the protein expression (the
cell decreases the quantity of the specific protein). The vertical column shows the condition at
which the protein is upregulated or downregulated. We observe that the proteins that
upregulated group together, in all the conditions against the control, and vice versa the
downregulated.

Annotation of the proteins to their function revealed that most of the proteins function
on metabolic pathways, which is highly downregulated, and on glycolysis which is upregulated
(Fig. 16). Interestingly, regulation of glycolysis was reported in previous studies (McGonigle et
al., 2017), which was found to be upregulated in mammalian cells upon their exposure to UV
light.

33
 We next asked whether the regulated proteins shared specific structural or biochemical
features. We found that most of the candidates are enzymes binding co-factors. Specifically, was
found that proteins that are mainly upregulated binding in the enzyme nicotinamide adenine
dinucleotide phosphate (NADPH), which acts as a reducing agent during the synthesis of nucleic
acids and lipids. The phosphate group of NADPH can absorb light, and potentially start a
cascade of events, which triggers the increased production of the specific proteins by the cell.
However, this speculation requires further research and experiments to be confirmed. Most of
the candidates that were downregulated bind on adenosine triphosphate (ATP) synthase, which
is used as the main energy source for metabolic functions. In same time, we previously observed
that metabolic pathways are downregulated. Lastly but not least, two of the candidates bind on
thiamine pyrophosphate, a Vitamin B1 involved in carbohydrate and amino acid
metabolism. Fig. 17 reports which proteins bind to which enzyme, respectively, and whether
they are up or downregulated. Upregulation is denoted with red and downregulation is denoted
with blue color.

34
Fig. 15. Heatmap showing the proteins that differentiate in each sample. The red color represents the upregulation
of the protein in response to the microscopic light described on x axis, while blue color represents the
downregulation of the protein expression according to different light wavelength.

35
 Protein Count

Fig. 16. Annotation of the proteins reported on Fig. 15, to their function. Metabolic pathways are mostly affected
by the exposure of cells to microscopic light.

Fig. 17. The scheme shows which candidates bind to which enzymes. Most of the candidates are enzyme binding
co-factors.

36
4. Conclusions
This study is, to our knowledge, the first known attempt to measure cellular phototoxicity from
microscopic light with mass spectrometry. Two challenges of this study - the optimization of the
exposure conditions and the optimization of the downstream analysis, to prepare the samples
for mass spectrometric analysis - were both tackled. A method to analyze with mass
spectrometry, samples directly from live cell microcopy micro-plates, was developed. However,
further optimization of the analytical method to increase yield of lysis and potentially number of
monitored proteins is still needed. Potential improvements to the methods could be the addition
of an enzymatic lysis step prior to ultrasonication; or avoiding dilution of sample with shifting
to another buffer of removal of urea system.

As already mentioned, the analysis of the data generated is still ongoing and some
statistical anomalies have not been solved. A potential reason is that the utilized spectral library
was not built for the specific study but was given over and was based in different lysis method.
To ascertain whether this could indeed be the case, a generation of in silico library is suggested.
However preliminary data are available and are being processed. Preliminary analysis of the data
showed 25 candidates change their abundance upon cell exposure to microscopic light, map
mainly to metabolic pathways, amino acid synthesis and glycolysis, and they are enzyme binding
co-factors. Glycolysis was found previously, to be regulated during mammalian cells exposure
to UV light, and our preliminary results indicate a regulation of glycolysis too. Overall, this thesis
has provided the first proof of principle of an analysis of the impact of light on the proteome by
mass spectrometry. Data analysis is still in progress, and a thoroughly data quality control and
analysis needs to be conducted in order to suggest with confidence a concrete mechanism of
action of phototoxicity.

Acknowledgements
The author of this study would like to thank Professor Ruedi Aebersold from Institute of
Molecular Systems Biology and Professor Karsten Weis from Institute of Biochemistry, both
from the Department of Biology of ETH Zürich, for offering financial support in order to
perform the study. Additionally, I would like to thank Dr. Rodolfo Ciuffa (ETH Zürich) for his
close supervision and Dr. Elisa Dultz (ETH Zürich) for her discussions and feedback in
microscopy. Moreover, would like to thank Professor Ulrika Nilsson, from the Environmental

37
Science and Analytical Chemistry department of Stockholm University, for her intellectual
contribution and feedback.

References
Aebersold, R., & Mann, M. (2003). Mass spectrometry-based proteomics. Nature, 422, 198.
doi:10.1038/nature01511
Armirotti, A., Benatti, U., & Damonte, G. (2009). Top-down proteomics with a quadrupole time-of-flight
mass spectrometer and collision-induced dissociation. Rapid Communications in Mass Spectrometry,
23(5), 661-666. doi:doi:10.1002/rcm.3925
Artifacts of light. (2013). Nature Methods, 10, 1135. doi:10.1038/nmeth.2760
Blackstock, W. P., & Weir, M. P. (1999). Proteomics: quantitative and physical mapping of cellular
proteins. Trends in Biotechnology, 17(3), 121-127. doi:10.1016/S0167-7799(98)01245-1
Bodvard, K., Jörhov, A., Blomberg, A., Molin, M., & Käll, M. (2013). The Yeast Transcription Factor
Crz1 Is Activated by Light in a Ca(2+)/Calcineurin-Dependent and PKA-Independent Manner.
PLOS ONE, 8(1), e53404. doi:10.1371/journal.pone.0053404
Cabib, E., Farkas, V., Kosík, O., Blanco, N., Arroyo, J., & McPhie, P. (2008). Assembly of the yeast cell
wall. Crh1p and Crh2p act as transglycosylases in vivo and in vitro. The Journal of biological chemistry,
283(44), 29859-29872. doi:10.1074/jbc.M804274200
Carlsohn, E., & Nilsson, C. L. (2007). Chapter 12 - Proteomic Techniques for Functional Identification
of Bacterial Adhesins. In C. L. Nilsson (Ed.), Lectins (pp. 299-325). Amsterdam: Elsevier Science
B.V.
Cech, N. B., & Enke, C. G. (2001). Practical implications of some recent studies in electrospray ionization
fundamentals. Mass Spectrometry Reviews, 20(6), 362-387. doi:doi:10.1002/mas.10008
Clarke, W. (2017). Chapter 1 - Mass spectrometry in the clinical laboratory: determining the need and
avoiding pitfalls. In H. Nair & W. Clarke (Eds.), Mass Spectrometry for the Clinical Laboratory (pp. 1-
15). San Diego: Academic Press.
Cox, J., & Mann, M. (2008). MaxQuant enables high peptide identification rates, individualized p.p.b.-
range mass accuracies and proteome-wide protein quantification. Nature Biotechnology, 26, 1367.
doi:10.1038/nbt.1511
Craig, R., & Beavis, R. C. (2003). A method for reducing the time required to match protein sequences
with tandem mass spectra. Rapid Communications in Mass Spectrometry, 17(20), 2310-2316.
doi:doi:10.1002/rcm.1198
de Godoy, L. M. F., Olsen, J. V., de Souza, G. A., Li, G., Mortensen, P., & Mann, M. (2006). Status of
complete proteome analysis by mass spectrometry: SILAC labeled yeast as a model system.
Genome Biol, 7(6), R50. doi:10.1186/gb-2006-7-6-r50
de Hoffmann, E., & Stroobant, V. (2007). Mass Spectrometry: Principles and Applications, 3rd Edition: Wiley-
Interscience.
De Vos, W. H., Hoebe, R. A., Joss, G. H., Haffmans, W., Baatout, S., Van Oostveldt, P., & Manders, E.
M. M. (2009). Controlled light exposure microscopy reveals dynamic telomere microterritories
throughout the cell cycle. Cytometry Part A, 75(5), 428-439. doi:10.1002/cyto.a.20699
Deutsch, E. W., Mendoza, L., Shteynberg, D., Farrah, T., Lam, H., Tasman, N., Sun, Z., Nilsson, E.,
Pratt, B., Prazen, B., Eng, J. K., Martin, D. B., Nesvizhskii, A. I., & Aebersold, R. (2010). A
guided tour of the Trans-Proteomic Pipeline. Proteomics, 10(6), 1150-1159.
doi:10.1002/pmic.200900375
Dixit, R., & Cyr, R. (2003). Cell damage and reactive oxygen species production induced by fluorescence
microscopy: effect on mitosis and guidelines for non-invasive fluorescence microscopy. The Plant
Journal, 36(2), 280-290. doi:doi:10.1046/j.1365-313X.2003.01868.x
Dole, M., Mack, L. L., Hines, R. L., Mobley, R. C., Ferguson, L. D., & Alice, M. B. (1968). Molecular
Beams of Macroions. The Journal of Chemical Physics, 49(5), 2240-2249. doi:10.1063/1.1670391
Domon, B., & Aebersold, R. (2006). Mass Spectrometry and Protein Analysis. Science, 312(5771), 212-
217. doi:10.1126/science.1124619
38
Dunn, W. B. (2011). Chapter two - Mass Spectrometry in Systems Biology: An Introduction. In D.
Jameson, M. Verma, & H. V. Westerhoff (Eds.), Methods in Enzymology (Vol. 500, pp. 15-35):
Academic Press.
Frigault, M. M., Lacoste, J., Swift, J. L., & Brown, C. M. (2009). Live-cell microscopy – tips and tools.
Journal of Cell Science, 122(6), 753-767. doi:10.1242/jcs.033837
Geer, L. Y., Markey, S. P., Kowalak, J. A., Wagner, L., Xu, M., Maynard, D. M., Yang, X., Shi, W., &
Bryant, S. H. (2004). Open Mass Spectrometry Search Algorithm. Journal of Proteome Research, 3(5),
958-964. doi:10.1021/pr0499491
Georgiou, C. A., & Danezis, G. P. (2015). Chapter 3 - Elemental and Isotopic Mass Spectrometry. In Y.
Picó (Ed.), Comprehensive Analytical Chemistry (Vol. 68, pp. 131-243): Elsevier.
Gillet, L. C., Navarro, P., Tate, S., Röst, H., Selevsek, N., Reiter, L., Bonner, R., & Aebersold, R. (2012).
Targeted Data Extraction of the MS/MS Spectra Generated by Data-independent Acquisition:
A New Concept for Consistent and Accurate Proteome Analysis. Molecular &amp; Cellular
Proteomics, 11(6). doi:10.1074/mcp.O111.016717
Godley, B. F., Shamsi, F. A., Liang, F.-Q., Jarrett, S. G., Davies, S., & Boulton, M. (2005). Blue Light
Induces Mitochondrial DNA Damage and Free Radical Production in Epithelial Cells. The Journal
of biological chemistry, 280, 21061-21066. doi:10.1074/jbc.M502194200
Graves, P. R., & Haystead, T. A. J. (2002). Molecular biologist's guide to proteomics. Microbiology and
molecular biology reviews : MMBR, 66(1), 39-63. doi:10.1128/MMBR.66.1.39-63.2002
Griffin, T. J., Gygi, S. P., Rist, B., Aebersold, R., Loboda, A., Jilkine, A., Ens, W., & Standing, K. G.
(2001). Quantitative Proteomic Analysis Using a MALDI Quadrupole Time-of-Flight Mass
Spectrometer. Analytical Chemistry, 73(5), 978-986. doi:10.1021/ac001169y
Gushue, J. N. (2013). Chapter 11 - Principles and Applications of Gas Chromatography Quadrupole
Time-of-Flight Mass Spectrometry. In I. Ferrer & E. M. Thurman (Eds.), Comprehensive Analytical
Chemistry (Vol. 61, pp. 255-270): Elsevier.
Ho, B., Baryshnikova, A., & Brown, G. W. (2018). Unification of Protein Abundance Datasets Yields a
Quantitative <em>Saccharomyces cerevisiae</em> Proteome. Cell Systems, 6(2), 192-205.e193.
doi:10.1016/j.cels.2017.12.004
Hu, A., Noble, W. S., & Wolf-Yadlin, A. (2016). Technical advances in proteomics: new developments
in data-independent acquisition. F1000Research, 5, F1000 Faculty Rev-1419.
doi:10.12688/f1000research.7042.1
Huber, C. G., & Hölzl, G. (2001). Chapter 8 - Hyphenation of Capillary Electrochromatography and
Mass Spectrometry: Instrumental Aspects, Separation Systems, and Applications. In Z. Deyl &
F. Švec (Eds.), Journal of Chromatography Library (Vol. 62, pp. 271-316): Elsevier.
Huh, W.-K., Falvo, J. V., Gerke, L. C., Carroll, A. S., Howson, R. W., Weissman, J. S., & O'Shea, E. K.
(2003). Global analysis of protein localization in budding yeast. Nature, 425, 686.
doi:10.1038/nature02026
Icha, J., Weber, M., Waters, J. C., & Norden, C. (2017). Phototoxicity in live fluorescence microscopy,
and how to avoid it. BioEssays, 39(8), 1700003. doi:doi:10.1002/bies.201700003
Iribarne, J. V., & Thomson, B. A. (1976). On the evaporation of small ions from charged droplets. The
Journal of Chemical Physics, 64(6), 2287-2294. doi:10.1063/1.432536
Kalli, A., Smith, G. T., Sweredoski, M. J., & Hess, S. (2013). Evaluation and Optimization of Mass
Spectrometric Settings during Data-Dependent Acquisition Mode: Focus on LTQ-Orbitrap
Mass Analyzers. Journal of Proteome Research, 12(7), 3071-3086. doi:10.1021/pr3011588
Kilian, N., Goryaynov, A., Lessard, M. D., Hooker, G., Toomre, D., Rothman, J. E., & Bewersdorf, J.
(2018). Assessing photodamage in live-cell STED microscopy. Nature Methods, 15(10), 755-756.
doi:10.1038/s41592-018-0145-5
King, N. L., Deutsch, E. W., Ranish, J. A., Nesvizhskii, A. I., Eddes, J. S., Mallick, P., Eng, J., Desiere,
F., Flory, M., Martin, D. B., Kim, B., Lee, H., Raught, B., & Aebersold, R. (2006). Analysis of
the Saccharomyces cerevisiae proteome with PeptideAtlas. Genome Biology, 7(11), R106.
doi:10.1186/gb-2006-7-11-r106

39
Kingdon, K. H. (1923). A Method for the Neutralization of Electron Space Charge by Positive Ionization
at Very Low Gas Pressures. Physical Review, 21(4), 408-418. doi:10.1103/PhysRev.21.408
Knight, J. D. R., Choi, H., Gupta, G. D., Pelletier, L., Raught, B., Nesvizhskii, A. I., & Gingras, A.-C.
(2017). ProHits-viz: a suite of web tools for visualizing interaction proteomics data. Nature
Methods, 14, 645. doi:10.1038/nmeth.4330
Kohn, E. C., Azad, N., Annunziata, C., Dhamoon, A. S., & Whiteley, G. (2007). Proteomics as a tool for
biomarker discovery. Disease markers, 23(5-6), 411-417. doi:10.1155/2007/967023
Koopmans, F., Ho, J. T. C., Smit, A. B., & Li, K. W. (2018). Comparative Analyses of Data Independent
Acquisition Mass Spectrometric Approaches: DIA, WiSIM‐DIA, and Untargeted DIA.
Proteomics, 18(1), 1700304. doi:10.1002/pmic.201700304
Liebler, D. C. (2002). Proteomic approaches to characterize protein modifications: new tools to study the
effects of environmental exposures. Environmental Health Perspectives, 110(suppl 1), 3-9.
doi:doi:10.1289/ehp.02110s113
Magidson, V., & Khodjakov, A. (2013). Circumventing photodamage in live-cell microscopy. Methods in
cell biology, 114, 10.1016/B1978-1010-1012-407761-407764.400023-407763. doi:10.1016/B978-
0-12-407761-4.00023-3
Makarov, A. (2000). Electrostatic Axially Harmonic Orbital Trapping: A High-Performance Technique
of Mass Analysis. Analytical Chemistry, 72(6), 1156-1162. doi:10.1021/ac991131p
Makarov, A., Denisov, E., Kholomeev, A., Balschun, W., Lange, O., Strupat, K., & Horning, S. (2006).
Performance Evaluation of a Hybrid Linear Ion Trap/Orbitrap Mass Spectrometer. Analytical
Chemistry, 78(7), 2113-2120. doi:10.1021/ac0518811
Makarov, A., Denisov, E., Lange, O., & Horning, S. (2006). Dynamic Range of Mass Accuracy in LTQ
Orbitrap Hybrid Mass Spectrometer. Journal of the American Society for Mass Spectrometry, 17(7), 977-
982. doi:10.1016/j.jasms.2006.03.006
Mamyrin, B. A., Karataev, V. I., Shmikk, D. V., & Zagulin, V. A. (1973). The mass-reflectron, a new
nonmagnetic time-of-flight mass spectrometer with high resolution. Soviet Physics JETP, 37(1),
45.
Mann, M., Hendrickson, R. C., & Pandey, A. (2001). Analysis of Proteins and Proteomes by Mass
Spectrometry. Annual Review of Biochemistry, 70(1), 437-473.
doi:10.1146/annurev.biochem.70.1.437
Mann, M., Kulak, Nils A., Nagaraj, N., & Cox, J. (2013). The Coming Age of Complete, Accurate, and
Ubiquitous Proteomes. Molecular Cell, 49(4), 583-590. doi:10.1016/j.molcel.2013.01.029
McGonigle, T. A., Dwyer, A. R., Greenland, E. L., Scott, N. M., Carter, K. W., Keane, K. N., Newsholme,
P., Goodridge, H. S., Pixley, F. J., & Hart, P. H. (2017). Reticulon-1 and Reduced Migration
toward Chemoattractants by Macrophages Differentiated from the Bone Marrow of Ultraviolet-
Irradiated and Ultraviolet-Chimeric Mice. The Journal of Immunology.
doi:10.4049/jimmunol.1700760
Meher, A. K., & Chen, Y. C. (2017). Electrospray Modifications for Advancing Mass Spectrometric
Analysis. Mass Spectrom (Tokyo), 6(Spec Iss). doi:10.5702/massspectrometry.S0057
Mellon, F. A. (2003). MASS SPECTROMETRY | Principles and Instrumentation. In B. Caballero (Ed.),
Encyclopedia of Food Sciences and Nutrition (Second Edition) (pp. 3739-3749). Oxford: Academic Press.
Michalski, A., Damoc, E., Hauschild, J.-P., Lange, O., Wieghaus, A., Makarov, A., Nagaraj, N., Cox, J.,
Mann, M., & Horning, S. (2011). Mass Spectrometry-based Proteomics Using Q Exactive, a
High-performance Benchtop Quadrupole Orbitrap Mass Spectrometer. Molecular &amp; Cellular
Proteomics, 10(9). doi:10.1074/mcp.M111.011015
Milo, R. (2013). What is the total number of protein molecules per cell volume? A call to rethink some
published values. BioEssays : news and reviews in molecular, cellular and developmental biology, 35(12),
1050-1055. doi:10.1002/bies.201300066
Papanayotou, I., Sun, B., Roth, A. F., & Davis, N. G. (2010). Protein aggregation induced during glass
bead lysis of yeast. Yeast (Chichester, England), 27(10), 801-816. doi:10.1002/yea.1771
Paulovich, A. G., Billheimer, D., Ham, A.-J. L., Vega-Montoto, L., Rudnick, P. A., Tabb, D. L., Wang,
P., Blackman, R. K., Bunk, D. M., Cardasis, H. L., Clauser, K. R., Kinsinger, C. R., Schilling, B.,

40
 Tegeler, T. J., Variyath, A. M., Wang, M., Whiteaker, J. R., Zimmerman, L. J., Fenyo, D., Carr,
S. A., Fisher, S. J., Gibson, B. W., Mesri, M., Neubert, T. A., Regnier, F. E., Rodriguez, H.,
Spiegelman, C., Stein, S. E., Tempst, P., & Liebler, D. C. (2010). Interlaboratory Study
Characterizing a Yeast Performance Standard for Benchmarking LC-MS Platform Performance.
Molecular &amp; Cellular Proteomics, 9(2), 242-254. doi:10.1074/mcp.M900222-MCP200
Perry, R. H., Cooks, R. G., & Noll, R. J. (2008). Orbitrap mass spectrometry: Instrumentation, ion motion
and applications. Mass Spectrometry Reviews, 27(6), 661-699. doi:doi:10.1002/mas.20186
Phototoxicity revisited. (2018). Nature Methods, 15(10), 751-751. doi:10.1038/s41592-018-0170-4
Picotti, P., Clément-Ziza, M., Lam, H., Campbell, D. S., Schmidt, A., Deutsch, E. W., Röst, H., Sun, Z.,
Rinner, O., Reiter, L., Shen, Q., Michaelson, J. J., Frei, A., Alberti, S., Kusebauch, U., Wollscheid,
B., Moritz, R. L., Beyer, A., & Aebersold, R. (2013). A complete mass-spectrometric map of the
yeast proteome applied to quantitative trait analysis. Nature, 494, 266. doi:10.1038/nature11835
Potriquet, J., Laohaviroj, M., Bethony, J. M., & Mulvenna, J. (2017). A modified FASP protocol for high-
throughput preparation of protein samples for mass spectrometry. PLOS ONE, 12(7), e0175967.
doi:10.1371/journal.pone.0175967
Röst, H. L., Rosenberger, G., Navarro, P., Gillet, L., Miladinović, S. M., Schubert, O. T., Wolski, W.,
Collins, B. C., Malmström, J., Malmström, L., & Aebersold, R. (2014). OpenSWATH enables
automated, targeted analysis of data-independent acquisition MS data. Nature Biotechnology, 32,
219. doi:10.1038/nbt.2841
Satoh, T., Tsuno, H., Iwanaga, M., & Kammei, Y. (2005). The Design and Characteristic Features of a
New Time-of-Flight Mass Spectrometer with a Spiral Ion Trajectory. Journal of the American Society
for Mass Spectrometry, 16(12), 1969-1975. doi:10.1016/j.jasms.2005.08.005
Senko, M. W., Remes, P. M., Canterbury, J. D., Mathur, R., Song, Q., Eliuk, S. M., Mullen, C., Earley, L.,
Hardman, M., Blethrow, J. D., Bui, H., Specht, A., Lange, O., Denisov, E., Makarov, A.,
Horning, S., & Zabrouskov, V. (2013). Novel Parallelized Quadrupole/Linear Ion
Trap/Orbitrap Tribrid Mass Spectrometer Improving Proteome Coverage and Peptide
Identification Rates. Analytical Chemistry, 85(24), 11710-11714. doi:10.1021/ac403115c
Shalit, T., Elinger, D., Savidor, A., Gabashvili, A., & Levin, Y. (2015). MS1-Based Label-Free Proteomics
Using a Quadrupole Orbitrap Mass Spectrometer. Journal of Proteome Research, 14(4), 1979-1986.
doi:10.1021/pr501045t
Sidoli, S., Fujiwara, R., & Garcia, B. A. (2016). Multiplexed data independent acquisition (MSX-DIA)
applied by high resolution mass spectrometry improves quantification quality for the analysis of
histone peptides. Proteomics, 16(15-16), 2095-2105. doi:10.1002/pmic.201500527
Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto,
E. K., Goeke, N. M., Olson, B. J., & Klenk, D. C. (1985). Measurement of protein using
bicinchoninic acid. Analytical Biochemistry, 150(1), 76-85. doi:10.1016/0003-2697(85)90442-7
Spikes, J. D. (1983). Photosensitization in Mammalian Cells. In J. A. Parrish, M. L. Kripke, & W. L.
Morison (Eds.), Photoimmunology (pp. 23-49). Boston, MA: Springer US.
Stafford, G. (2002). Ion trap mass spectrometry: A personal perspective. Journal of the American Society for
Mass Spectrometry, 13(6), 589-596. doi:10.1016/s1044-0305(02)00385-9
Stockley, J. H., Evans, K., Matthey, M., Volbracht, K., Agathou, S., Mukanowa, J., Burrone, J., &
Káradóttir, R. T. (2017). Surpassing light-induced cell damage in vitro with novel cell culture
media. Scientific Reports, 7(1), 849. doi:10.1038/s41598-017-00829-x
Teo, G., Kim, S., Tsou, C.-C., Collins, B., Gingras, A.-C., Nesvizhskii, A. I., & Choi, H. (2015). mapDIA:
Preprocessing and statistical analysis of quantitative proteomics data from data independent
acquisition mass spectrometry. Journal of proteomics, 129, 108-120. doi:10.1016/j.jprot.2015.09.013
Thomson, B. A., & Iribarne, J. V. (1979). Field induced ion evaporation from liquid surfaces at
atmospheric pressure. The Journal of Chemical Physics, 71(11), 4451-4463. doi:10.1063/1.438198
Tinevez, J.-Y., Dragavon, J., Baba-Aissa, L., Roux, P., Perret, E., Canivet, A., Galy, V., & Shorte, S. (2012).
Chapter fifteen - A Quantitative Method for Measuring Phototoxicity of a Live Cell Imaging
Microscope. In P. M. Conn (Ed.), Methods in Enzymology (Vol. 506, pp. 291-309): Academic Press.

41
Tinevez, J.-Y., Perry, N., Schindelin, J., Hoopes, G. M., Reynolds, G. D., Laplantine, E., Bednarek, S. Y.,
Shorte, S. L., & Eliceiri, K. W. (2017). TrackMate: An open and extensible platform for single-
particle tracking. Methods, 115, 80-90. doi:10.1016/j.ymeth.2016.09.016
Tsou, C.-C., Avtonomov, D., Larsen, B., Tucholska, M., Choi, H., Gingras, A.-C., & Nesvizhskii, A. I.
(2015). DIA-Umpire: comprehensive computational framework for data-independent
acquisition proteomics. Nature Methods, 12, 258. doi:10.1038/nmeth.3255
Uhlén, M., Fagerberg, L., Hallström, B. M., Lindskog, C., Oksvold, P., Mardinoglu, A., Sivertsson, Å.,
Kampf, C., Sjöstedt, E., Asplund, A., Olsson, I., Edlund, K., Lundberg, E., Navani, S., Szigyarto,
C. A.-K., Odeberg, J., Djureinovic, D., Takanen, J. O., Hober, S., Alm, T., Edqvist, P.-H.,
Berling, H., Tegel, H., Mulder, J., Rockberg, J., Nilsson, P., Schwenk, J. M., Hamsten, M., von
Feilitzen, K., Forsberg, M., Persson, L., Johansson, F., Zwahlen, M., von Heijne, G., Nielsen, J.,
& Pontén, F. (2015). Tissue-based map of the human proteome. Science, 347(6220).
doi:10.1126/science.1260419
Vestal, M. L., Juhasz, P., & Martin, S. A. (1995). Delayed extraction matrix-assisted laser desorption time-
of-flight mass spectrometry. Rapid Communications in Mass Spectrometry, 9(11), 1044-1050.
doi:doi:10.1002/rcm.1290091115
Vladimer, G. I., Snijder, B., Krall, N., Bigenzahn, J. W., Huber, K. V. M., Lardeau, C.-H., Sanjiv, K.,
Ringler, A., Berglund, U. W., Sabler, M., de la Fuente, O. L., Knöbl, P., Kubicek, S., Helleday,
T., Jäger, U., & Superti-Furga, G. (2017). Global survey of the immunomodulatory potential of
common drugs. Nature Chemical Biology, 13, 681. doi:10.1038/nchembio.2360
Wagner, M., Weber, P., Bruns, T., Strauss, W. S. L., Wittig, R., & Schneckenburger, H. (2010). Light dose
is a limiting factor to maintain cell viability in fluorescence microscopy and single molecule
detection. International Journal of Molecular Sciences, 11(3), 956-966. doi:10.3390/ijms11030956
Wäldchen, S., Lehmann, J., Klein, T., van de Linde, S., & Sauer, M. (2015). Light-induced cell damage in
live-cell super-resolution microscopy. Scientific Reports, 5, 15348-15348. doi:10.1038/srep15348
Walgren, J. L., & Thompson, D. C. (2004). Application of proteomic technologies in the drug
development process. Toxicology Letters, 149(1), 377-385. doi: 10.1016/j.toxlet.2003.12.047
Walker, J. M. (1994). The Bicinchoninic Acid (BCA) Assay for Protein Quantitation. In J. M. Walker
(Ed.), Basic Protein and Peptide Protocols (pp. 5-8). Totowa, NJ: Humana Press.
Wenger, M. D., DePhillips, P., & Bracewell, D. G. (2008). A Microscale Yeast Cell Disruption Technique
for Integrated Process Development Strategies. Biotechnoly Progress, 24(3), 606-614.
doi:10.1021/bp070359s
Wiley, W. C., & McLaren, I. H. (1955). Time-of-Flight Mass Spectrometer with Improved Resolution.
Review of Scientific Instruments, 26(2), 1150-1157. doi:10.1063/1.1715212
Wilm, M. (2011). Principles of Electrospray Ionization. Molecular & Cellular Proteomics : MCP, 10(7),
M111.009407. doi:10.1074/mcp.M111.009407
Wilm, M. S., & Mann, M. (1994). Electrospray and Taylor-Cone theory, Dole's beam of macromolecules
at last? International Journal of Mass Spectrometry and Ion Processes, 136(2), 167-180. doi: 10.1016/0168-
1176(94)04024-9
Yu, Y., Smith, M., & Pieper, R. (2014). A spinnable and automatable StageTip for high throughput
peptide desalting and proteomics.
Zdolsek, J. M., Olsson, G. M., & Brunk, U. T. (1990). PHOTOOXIDATIVE DAMAGE TO
LYSOSOMES OF CULTURED MACROPHAGES BY ACRIDINE ORANGE. Photochemistry
and Photobiology, 51(1), 67-76. doi:doi:10.1111/j.1751-1097.1990.tb01685.x

42
Appendices
Table A- 1. Instances of subtle phototoxicity found in literature [as taken from 5]

43
Fig. A- 1. Comparison of the yeast reference proteome as taken from Gene Ontology Consortium with the proteome
analyzed in this study. The method does not enrich or deplete selectively, protein class/compartments.

44
45
46
47
48
 Study time plan

21st March -April May June July August September October

Literature review
Microscopy optimization
Cell lysis optimization
Main study
Data Analysis
Thesis writing

49