You are on page 1of 115


Genehmigt von der Philosophisch-Naturwissenschaftlichen Fakultät

Auf Antrag von

Prof. Dr. Andreas Engel & Prof. Dr. Jean-Louis Rigaud

Basel, den 19. Dezember 2000

Prof. Dr. Andreas D. Zuberbühler

Dekan der Philosophisch
Naturwissenschaftlichen Fakultät

To my parents,
my brothers,
and my friends.


1. AFM and EM in structural biology........................................................................15
1.1. General introduction..................................................................................................15
1.1.1. The driving force to do science...........................................................................15
1.1.2. Membrane proteins............................................................................................15
1.1.3. 2D crystals allow the acquisition of structural information on membrane proteins
in a native-like environment................................................................................16
1.1.4. Atomic force and electron microscopy cover large resolution ranges and together
provide both surface and volume information.....................................................17
1.1.5. Results and Perspectives....................................................................................18
1.1.6. References..........................................................................................................19
1.2. Atomic force microscopy: A powerful tool to observe the assembly and function
of native proteins........................................................................................................21
1.2.1. Abstract..............................................................................................................21
1.2.2. Introduction........................................................................................................21
1.2.3. Conditions for single molecule imaging.............................................................22
1.2.4. Imaging the ion-driven rotor of the ATP synthase..............................................23
1.2.5. Conformational flexibility of proteins................................................................25
1.2.6. The tongue-and-groove interaction of MIP tetramers.........................................26
1.2.7. Imaging the subcomplexes of the GroE chaperonin system: GroEL and GroES27
1.2.8. Observing the assembly of membrane proteins..................................................28
1.2.9. Outlook..............................................................................................................29
1.3. Imaging streptavidin 2D-crystals on biotinylated lipid monolayers at high
resolution with the atomic force microscope...........................................................35
1.3.1. Summary............................................................................................................35
1.3.2. Introduction........................................................................................................35
1.3.3. Materials and Methods.......................................................................................36 Materials................................................................................................36 Hydrophobicity measurement......................................................................36 Crystallization of streptavidin on biotin-lipid monolayer..................................36 Atomic force microscopy (AFM).................................................................37 Transmission electron microscopy (TEM).....................................................37
1.3.4. Results...............................................................................................................37 Hydrophobicity and topography of HOPG.....................................................37 Crystallization of streptavidin on biotin-lipid monolayer..................................38 AFM of streptavidin crystals......................................................................39 TEM of streptavidin crystals.......................................................................40
1.3.5. Discussion.........................................................................................................40
1.3.6. Acknowledgment................................................................................................42
1.3.7. References..........................................................................................................43

2. Application of high resolution AFM.......................................................................47
2.1. High resolution AFM topographs of the Escherichia coli waterchannel
aquaporin Z................................................................................................................47
2.1.1. Abstract..............................................................................................................47
2.1.2. Introduction.......................................................................................................47
2.1.3. Results...............................................................................................................49
2.1.4. Discussion.........................................................................................................50
2.1.5. Materials and methods.......................................................................................54 Reconstitution..........................................................................................54 Trypsin digestion......................................................................................54 Atomic force microscopy............................................................................54 Image processing......................................................................................55
2.1.6. Acknowledgment...............................................................................................55
2.1.7. References.........................................................................................................55
2.2. High resolution AFM topographs of Rubrivivax gelatinosus light-harvesting
complex LH2...............................................................................................................59
2.2.1. Abstract..............................................................................................................59
2.2.2. Introduction.......................................................................................................59
2.2.3. Results...............................................................................................................61
2.2.4. Discussion.........................................................................................................65
2.2.5. Materials and methods.......................................................................................67 Materials.................................................................................................67 Isolation, purification and proteolysis of LH2 complex....................................67 Biochemical and biophysical techniques.........................................................67 Reconstitution and 2D crystallization............................................................67 Atomic force microscopy............................................................................67 Image processing......................................................................................68
2.2.6. Acknowledgment...............................................................................................68
2.2.7. References.........................................................................................................68

3. Combining surface and projection techniques....................................................73
3.1. The aquaporin sidedness revisited...........................................................................73
3.1.1. Summary...........................................................................................................73
3.1.2. Introduction.......................................................................................................73
3.1.3. Results...............................................................................................................74
3.1.4. Discussion.........................................................................................................78
3.1.5. Materials and Methods......................................................................................80 2D crystallization......................................................................................80 Trypsin digestion......................................................................................80 Atomic force microscopy............................................................................80 Freeze-drying & metal-shadowing.................................................................80 Cryo electron microscopy...........................................................................80
3.1.6. Acknowledgment...............................................................................................81
3.1.7. References.........................................................................................................81

4. Structural studies of a membrane transporter...................................................87
4.1. The functional Escherichia coli lactose permease LacY/Cytb562/6His forms
trimers: A 2.8 nm 3D reconstruction and preliminary electron crystallographic
4.1.1. Summary............................................................................................................87
4.1.2. Introduction........................................................................................................87
4.1.3. Results and discussion.......................................................................................89 Protein purification...................................................................................89 Single particle analysis..............................................................................90 Reconstitution and 2D crystallization...........................................................91
4.1.4. Perspectives........................................................................................................93
4.1.5. Material and Methods........................................................................................94 Materials................................................................................................94 Protein Expression and Purification.............................................................94 Reconstitution.........................................................................................94 Electron microscopy and image processing....................................................95
4.1.6. References..........................................................................................................95

5. General discussion and conclusions.......................................................................101

6. Acknowledgment..........................................................................................................107

7. Curriculum vitae..........................................................................................................113
7.1. Education...................................................................................................................113
7.2. Teaching.....................................................................................................................113
7.3. Publications...............................................................................................................113
7.4. Meetings.....................................................................................................................114

1. AFM and EM in
structural biology

1. AFM and EM in structural biology

1.1. General introduction

1.1.1. The driving force to do science result of a genome study. Genomes of
different organisms (E. coli, M. jannaschii, H.
What does it look like? What does it do? It sapiens) were screened for their protein
is curiosity, the wish to know, that induces coding open reading frames and these open
these questions. It is human nature - and it is reading frames were translated into amino
the driving force to do fundamental research acid chains. The peptides were discriminated
in any field ranging from astronomy to by their hydrophobicities. This resulted in
physics of smallest matter. The primary two families of gene products: The
questions of molecular biologists are: What is hydrophilic cytoplasmic proteins, and the
the structure, what is the function, of a membrane proteins, which contain large
biomolecule? Once the structure and function hydrophobic stretches (e.g., representing
of a biomolecule are known, its physiological transmembrane helices). By this relatively
role and interactive mechanisms can be simple approach, it has been demonstrated
brought into context within the framework of that 20 - 30 % of genes (the higher the
the cell and the whole organism. It is our organism, the larger the percentage) code for
ultimate goal to understand the molecular strongly hydrophobic proteins which are
mechanisms of all processes in each cell of most probably integrated into cell
our body, ranging from neurons working in membranes.
our brain, muscle cells allowing us to move, to This intriguing result emphasizes the
the cells of our skin ultimately defining the extreme importance of membrane proteins for
borderlines of ourselves and our environment. living organisms. Membrane proteins connect
the cytoplasm with the extracellular space of
1.1.2. Membrane proteins each living cell, form junctions between living
cells or play an important role in the
Intrinsic or integral membrane proteins are intracellular compartments. Hence, in bacteria,
defined as proteins that penetrate into and, such molecules work in transport, secretion
most often, traverse the lipid bilayer of a and bioenergetic processes. Multicellular
biological membrane. Protein structures, organisms even require active communication
which partition into lipid rather than remain in between their cells. Consequently a large
aqueous solution have specific chemical number of membrane proteins have evolved,
properties. They are rich in exposed working as receptors for intercellular
hydrophobic amino acids and are restricted in trafficking or cellular adhesion and
their secondary structure. A consequence of recognition. Evolution has also created highly
these physico-chemical properties is that an specific channels and transporters, which are
integral membrane protein can only be essential for the survival of biological
brought into aqueous solution when systems; the deletion of many membrane
solubilized in the presence of detergents. proteins is lethal or leads to severe disease.
The challenge of understanding membrane The study of membrane protein structure
proteins and transporters has attracted our is a difficult challenge: Membrane proteins
interest. Figure 1 represents an interesting remain only folded in their active state when

Figure 1. Number of genes of different organisms as a function of the hydrophobicity of their gene product.
Integration over the two peaks corresponding to cytoplasmic and membrane proteins shows that 20 - 30 % of all
genes code for membrane proteins.

their hydrophobic transmembrane domains 1.1.3. 2D crystals allow the acquisition
are embedded in a hydrophobic environment - of structural information on
e.g., a lipid bilayer or a detergent micelle. This membrane proteins in a native-
prerequisite makes the growth of 3D crystals like environment
for structure determination by X-ray
diffraction difficult. Consequently, ~2000 In order to acquire biologically valid
atomic structures of water-soluble proteins information, it is important to study the
are available but only ~20 atomic structures structure of membrane proteins under
of integral membrane proteins. conditions where they remain functional. To
The structures so far determined, divide this end, membrane proteins are reconstituted
membrane proteins into two categories: α- into 2D crystals in the presence of lipids
helical and β-barrel membrane proteins. The which mimic their native membrane
majority of the structures were determined environment within a cell (chapters 2.1, 2.2,
using 3D crystallization and X-ray 3.1, 4.1). Although only 3 membrane protein
diffraction. However, three structures have structures have been solved to atomic
been solved by electron crystallography: Plant resolution (below 4Å) using electron
light-harvesting complex II, crystallography (Kühlbrandt et al., 1994;
bacteriorhodopsin, and human aquaporin 1. Henderson et al., 1990; Kimura et al., 1997;
The electron crystallography was carried out Murata et al., 2000), numerous proteins have
using 2D crystals of protein integrated in been solved to medium resolution (4Å-10Å)
lipid bilayers. in 2D projection or 3D density maps from
electron micrographs. Such medium
resolution maps revealed helix arrangements

or/and structural similarities within the microscope are our tools to investigate the
aquaporin protein family (Stahlberg et al., fascinating microcosm of membrane proteins.
2001). As listed above, a combination of these two
techniques covers a resolution range from
1.1.4. Atomic force and electron micrometers to atomic scale, and yields both
microscopy cover large resolution surface and volume information of proteins in
ranges and together provide both the close to native environment of 2D
surface and volume information crystals.
Electron microscopy of negatively stained
Structural biology encompasses a range of samples allows the efficient acquisition of
techniques, to elucidate structures and structural information. Despite the limited
interactions of biomolecules. The table below resolution of ~15Å, the technique provides a
summarizes some of the advantages and high signal-to-noise ratio (chapters 1.3, 4.1).
disadvantages of the various approaches. By these means, samples of 2D crystals of
The atomic force and the electron membrane proteins can be checked and a first

Table 1. Advantages and disadvantages of techniques applied in structural biology

technique advantages disadvantages
X-ray crystallography -atomic resolution -well ordered 3D
crystals are required
-absence of phases
nuclear magnetic resonance -atomic resolution -requires large amounts
-information about of protein
protein dynamics -requires protein
-problems with large
molecules (>40 kDa)
electron crystallography -resolution range from -requires well ordered
nanometers to atomic 2D crystals
resolution -technically difficult
electron microscopy -study of large -resolution limit ~5Å
complexes -problems with small
-study of interactions molecules (<100 kDa)
between single
atomic force microscopy -investigations under -only surface information
native-like conditions -protein must be
-single molecules can immobilized on a
be addressed substrate
-resolution range from
µm to atomic scale
-time resolved
information can be
light microscopy -whole living cells can -resolution limit λ/2 =
be studied ~200 nm
-resolution range from
mm to nm scale

impression of the crystal quality obtained. functions; a water channel, a light-harvesting
Furthermore, single solubilized particles can protein, and a sugar transporter.
be visualized, and the averaging of untilted The family of aquaporins, are abundant
and tilted images can lead to low resolution channel forming transmembrane proteins
3D reconstructions. responsible for selective water transport. They
Imaging unstained samples under cryo are found in bacteria, plants, fungi and
conditions, allows the features of freeze-dried mammals (Agre et al., 1993). We have
preparations to be resolved to ~ 9 Å (chapter reconstituted purified tetrameric AqpZ, the
3.1), and those of vitruous ice embedded water channel of E. coli responsible for
samples to be resolved to atomic resolution in maintenance of cell turgor during the volume
ideal cases (see above), but projection and 3D expansion of cell division, into densely
maps can be acquired at any intermediate packed vesicles and 2D crystals. High
resolution (chapter 3.1). resolution topographs revealed two distinct
Atomic force microscopy allows protein protein surfaces. To assign the sidedness, 2D
surfaces to be contoured at subnanometer crystals of AqpZ with an N-terminal His-tag
resolution in buffer solution mimicking a were digested using trypsin. The cleavage
physiological environment (chapters 1.2, 2.1, results in a dramatic change of surface
2.2, 3.1) (Müller et al.1999). This offers the appearance of one side, allowing the
possibility of monitoring function related sidedness to be unambiguously assigned
structural conformational changes of single (chapter 2.1). Imaging surfaces using loading
proteins or of proteins within complex forces (50 - 200 pN) on the AFM tip,
functional assemblies (chapters 1.2, 2.1, 2.2). flexibility mapping, and volume calculations
Loops and termini protruding out of the led to an assignment of the large loop C
transmembrane α-helices, can directly be (chapter 2.1). By combining AFM, metal
assessed by the AFM. Consequently, shadowing electron microscopy, freeze-
proteolytic cleavage of termini can be drying cryo-electron microscopy, and
monitored (chapters 2.1, 2.2), leading to trehalose-embedded cryo-electron
unambiguous sidedness assignments. microscopy, the sidedness assignment could
be applied to the whole aquaporin family by
1.1.5. Results and Perspectives an elegant experiment (chapter 3.1): Images
of 2D crystals were taken, where parts of the
AFM of membrane proteins demands crystal were metal-shadowed, while other
most careful sample preparation in order to parts were prevented from metal deposition.
achieve high resolution topographs. In the From such images a direct link between
present work, preparation procedures for both surface and density projection could be
hydrophilic (chapter 1.2) and hydrophobic gained. This experimental approach can be
(chapter 1.3) samples were established. The applied to any membrane protein
two supports reported (chapter 1.2, 1.3) allow reconstituted into 2D crystals (chapter 3.1).
the physisorption of a wide range of Photosynthetic bacteria efficiently convert
biological samples without further fixation, light energy into biochemical energy using a
and therefore allow the immobilization and set of membrane proteins. In an initial step
visualization of proteins in a close to native photons are trapped by light-harvesting
environment. complex (LH2) oligomers, which are
Using atomic force and electron composed of 9 αβ heterodimers and contain
microscopy, we have investigated membrane a total of 27 bacteriochlorophyls and 9
proteins responsible for very different carotenoids. These ring-shaped oligomers

were integrated into 2D crystals in the 1.1.6. References
presence of lipids. Spectra showed the
reconstituted proteins to still contain Agre, P., Preston, G., Smith, B., Jung, J.,
chromophores in their native conformation Raina, S., Moon, C., Guggino, W. &
Nielsen, S. (1993). Aquaporin CHIP: the
(chapter 2.2). High resolution topographs archetypal molecular water channel.
before and after thermolysin cleavage of the American Journal of Physiology 265,
C-terminus of the α-subunit defined the F436-476.
position of this terminus and the sidedness of
Henderson, R., Baldwin, J. M., Ceska, T. A.,
the protein (chapter 2.2). In less pure Zemlin, F., Beckman, E. & Downing, K.
preparations larger rings were also imaged H. (1990). Model for the structure of
and assigned as light-harvesting complex 1 bacteriorhodopsin based on high-
oligomers. The two protein complexes resolution electron cryo-microscopy. J.
Mol. Biol. 213, 899-929.
represent a functional assembly in
photosynthetic membranes (Kühlbrandt, Kimura, Y., Vassylev, D. G., Miyazawa, A.,
1995). The capability of the AFM to image Kidera, A., Matsushima, M., Mitsuoka, K.,
and identify two different protein species in Murata, K., Hirai, T. & Fujiyoshi, Y.
(1997). Surface of bacteriorhodopsin
one membrane opens future perspectives to revealed by high-resolution electron
directly image native membranes under microscopy. Nature 389, 206-211.
native-like conditions (chapter 2.2).
Lactose permease is the best studied Kühlbrandt, W. (1995). Many wheels make
light work. Nature 374, 497-498.
membrane transporter. It catalyzes the
coupled stoichiometric translocation of β- Kühlbrandt, W., Wang, D. N. & Fujiyoshi, Y.
galactosides and H+ across the cytoplasmic (1994). Atomic model of plant light-
membrane. To fulfill their physiological role harvesting complex by electron
crystallography. Nature 367, 614-621.
of transport transporters like this must
undergo strong conformational changes. This Müller, D. J., Fotiadis, D., Scheuring, S.,
conformational variability poses difficulties to Müller, S. A. & Engel, A. (1999).
integrate LacY into 2D crystals (Zhuang et al. Electrostatically balanced subnanometer
imaging of biological specimens by atomic
1999). Well diffracting crystals were force microscopy. Biophys. J. 76, 1101-
multilayered (chapter 4.1). Negative stain 1111.
electron microscopy of solubilized particles in
different orientations with respect to the Murata, K., Mitsuoka, K., Hirai, T., Walz, T.,
Agre, P., Heymann, J. B., Engel, A. &
electron beam resulted in a low resolution 3D Fujiyoshi, Y. (2000). Structural
map (chapter 4.1). This map shows that the determinants of water permeation through
cytochrome b562 engineered into loop 6 aquaporin-1. Nature 407, 599-605.
induces trimerization by hydrophilic
Stahlberg, H., Braun, T., Philippsen, A.,
interactions (chapter 4.1). Future experiments Borgnia, M., Agre, P., Kühlbrandt, W. &
should increase crystal quality and result in a Engel, A. (2001). The 6.9 Å structure of
3D density map from unstained proteins, GlpF: a basis for homology modeling of
which can then be combined with the the glycerol channel from E.coli. J. Struct.
Biol., in press
knowledge acquired by biochemical and
biophysical analyses. Zhuang, J., Prive, G. G., Werner, G. E.,
Ringler, P., Kaback, R. H. & Engel, A.
(1999). Two-dimensional crystallization of
the Escherichia coli lactose permease. J.
Struct. Biol. 125, 63-75.

1.2. Atomic force microscopy: A powerful tool to observe the assembly and
function of native proteins

Simon Scheuringa, Dimitrios Fotiadisa, Clemens Möllera,b,c , Andreas Engela and Daniel
J. Müller a,c

a M.E.Müller Institute for Structural Biology, Biozentrum, University of Basel, CH-4056 Basel,
b Forschungszentrum Jülich, IBI-2: Structural Biology, D-52425 Jülich, Germany
c Max-Planck-Institute of Molecular Cell Biology and Genetics, D-01307 Dresden, Germany

1.2.1. Abstract topographs of single proteins allow
submolecular details to be resolved to a lateral
Here we discuss the experimental resolution of ~ 6 Å and a vertical resolution
approaches that have allowed high resolution of 1 Å (Fotiadis et al., 2000; Müller et al.,
atomic force microscopy (AFM) imaging, 1995; Schabert et al., 1995; Scheuring et al.,
and review results that show AFM to be of 1999b;
great interest for biologists, AFM allows Czajkowsky & Shao, 1998). This
single proteins to be imaged under resolution can be attained no matter whether
physiologically relevant conditions. The the imaged proteins are randomly packed or
exceptional signal-to-noise ratio and assembled into two-dimensional (2D)
resolution of AFM topographs enables the crystalline arrays, which are usually difficult
oligomerization state and characteristic to grow (Czajkowsky & Shao, 1998; Fotiadis
substructures of individual proteins to be et al., 2000; Karrasch et al., 1994; Mou et al.,
resolved. Several examples demonstrate the 1996; Mou et al., 1995; Müller & Engel,
capabilities of AFM to directly observe single 1999; Scheuring et al., 1999b; Seelert et al.,
proteins, and their conformational changes, to 2000). Standard image averaging techniques
study protein-protein interactions and to have proved useful both to interpret the
follow the assembly of membrane proteins. structural appearance of a protein and to
Here we consider the AFM techniques that assess the resolution attained (Schabert &
have allowed high resolution imaging, and Engel, 1994). In such averages common
review results that show AFM to be a structural details of the individual proteins are
powerful method to analyze biological enhanced while variable features are
processes at the level of single molecules. suppressed. However, the variable structural
regions of the protein are clearly identified by
1.2.2. Introduction the simultaneously calculated standard
deviation map (Müller et al., 1998; Schabert
Atomic force microscopy (AFM; (Binnig & Engel, 1994). Interestingly, such flexible
et al., 1986)) has become a complementary structural regions can undergo functional
technique to X-ray crystallography and related conformational changes such as
electron microscopy (EM). Of great interest recently observed by AFM for porin OmpF
for biologists, fragile biological samples can from Escherichia coli (Müller & Engel,
be observed under physiologically relevant 1999). In addition, conformational changes
conditions. As a result of the AFM's have been visualized in the bacterial surface
exceptionally high signal-to-noise ratio, layer of Deinococcus radiodurans (Müller et

al., 1996) and measured on bacterio- sample preparation and imaging conditions of
rhodopsin from Halobacterium salinarum soft biological structures. Operating the
(Haupts et al., 1999; Subramaniam et al., microscope in buffer solution not only allows
1999). The surface loop connecting helices E structural analysis to be carried out under
and F involved in functional conformations of native conditions (Drake et al., 1989), but
bacteriorhodopsin has also been also provides important possibilities to
demonstrated by AFM to reversibly change control the tip-sample interactions (Müller et
its conformation upon changing the force al., 1999b). Stable sample supports which
applied to the stylus (Müller et al., 1995; allow to achieve Angstroem resolution and to
Heymann et al., 1999). protect the piezo from contact with the buffer
These results demonstrate how well suited solution have been described for both
AFM is for the observation single proteins hydrophilic (Hoh et al., 1991; Schabert &
and their native substructures. The fact that Engel, 1994) and hydrophobic supports
intrinsic mechanical properties of these (Scheuring et al., 1999a). To adsorb
structural elements can also be mapped and hydrophilic biological samples onto freshly
their functional related conformational cleaved mica (negative surface charge at pH ≥
changes directly observed allows conclusions 6) sufficient concentrations of monovalent
to be drawn about the function of these and/or divalent cations are required to
regions (Engel & Müller, 2000). In future, compensate repulsive electrostatic forces
single molecule AFM imaging will be (Müller et al., 1997). In contrast to
routinely combined with methods providing hydrophilic samples, hydrophobic surfaces
complementary data for the single proteins on readily adhere to the hydrophobic surface of
a molecular scale (Engel & Müller, 2000; highly oriented pyrolytic graphite (HOPG)
Schmitt et al., 2000; Weiss, 1999; Weiss, (Scheuring et al., 1999a).
2000). Recently the combination of single- Most importantly, fragile biological
molecule AFM imaging and single-molecule samples necessitate the adjustment of the
force spectroscopy has allowed information force applied to the cantilever to < 100 pico-
on surface structure, protein folding (Müller Newtons (pN). Otherwise the AFM stylus
et al., 1999a; Oesterhelt et al., 2000) and will reversibly or irreversibly deform the soft
protein-protein interactions to be acquired biological material. Steric, electrostatic and
simultaneously (Raab et al., 1999; Ros et al., van der Waals interactions are the major
1998). forces acting between AFM stylus and
Here we consider AFM techniques that biological sample when imaging in buffer
have provided high-resolution AFM solution (Butt et al., 1995; Müller & Engel,
topographs of native non-crystalline proteins, 1997; Rotsch & Radmacher, 1997). Force-
and demonstrate the importance of single distance curves, acquired by decreasing the
molecule imaging as a technique to provide separation between the sample and the AFM
novel insights into fundamental mechanisms stylus while measuring the cantilever
of molecular biological processes. deflection, reveal the nature of these forces
(Butt et al., 1995). The long-range
1.2.3. Conditions for single molecule electrostatic double-layer forces (several tens
imaging of nm) depend on the charge density of both
interacting surfaces, the pH and the
Although suitable AFM instrumentation electrolyte composition of the buffer solution.
has been commercially available for many Consequently the electrostatic forces can be
years, significant progress has recently been regulated by adjusting pH and ionic strength
achieved in several laboratories by optimizing of the imaging buffer solution. Since only

Figure 1. Interaction forces between the AFM stylus and a biological specimen. The effective
interaction force is the sum of the applied force, the electrostatic repulsion (long range forces) and the van der Waals
attraction (short range forces). The electrostatic interactions are mainly due to the large global tip, while the short
range forces are mediated by a small local probe, which allows high resolution imaging. Since the electrostatic
repulsion can be regulated by adjusting the ionic strength of the buffer solution the applied force can be
electrostatically 'damped': |Feff| = |F appl + F el + F vdW| < |F appl| (for further details see Müller et al., 1999b).

short-range interactions between tip and microscope must be carefully adjusted, scan
sample (≤ 1 nm) allow acquisition of the frequencies between 4 - 7 Hz and a nominal
structural details evident on high-resolution magnification giving a pixel sampling of 2 – 3
topographs, the long-range electrostatic Å / pixel have proved satisfactory.
forces can be used to minimize the local Unfortunately, molecules from the laboratory
interaction forces (Fig. 1, Müller et al., environment can contaminate the AFM tip.
1999b). Contaminated tips can be cleaned using SDS-
After buffer adjustment, topographs of solution and nanopure water (Scheuring et al.,
protein surfaces revealing details with a lateral 2001), Helmanex (Hoerber et al., 1995), or
resolution between 5 - 8 Å and a vertical ultraviolet radiation (Thomson et al., 1996)
resolution of ~ 1Å can be recorded making their repetitive use possible.
reproducibly (Fotiadis et al., 2000; Müller et
al., 1999c; Scheuring et al., 1999b; Seelert et 1.2.4. Imaging the ion-driven rotor of the
al., 2000). ATP synthase
To avoid organic contamination of both
sample and tip, buffer solutions are best FoF1-ATP synthases use the energy of a
prepared using nanopure water (10-18 transmembrane proton (or Na+ ) gradient to
MΩ/cm). The biological samples should be synthesize the biological energy currency
checked for their purity by SDS gel ATP. The flow of cations appear to drive the
electrophoresis and, if possible, by electron transmembrane rotor of the integral
microscopy. The fluid cell of the AFM must membrane complex FO which is coupled to a
be thoroughly cleaned by repeated ultra- molecular shaft (Kato-Yamada et al., 1998;
sonication in the presence of ethanol and in Noji et al., 1997; Sabbert et al., 1996) that
nanopure water (Müller & Engel, 1997). activates the catalytic F1 complex. In current
To achieve high-resolution the scanning models of the FoF1-ATP synthases, the
speed and feedback parameters of the number of subunits forming the ion driven

Figure 2 . Assembly and stoichiometry o f the proton driven rotor from chloroplast ATP
synthase. a) Reconstituted, densely packed Fo rotors. The topograph was acquired in contact mode AFM at applied
forces of ~ 50 - 100 pN and in buffer solution (10 mM Tris-HCl, pH 7.8, 25 mM MgCl2). Wide and narrow rings
represent the two aqueous surfaces of the membrane spanning rotor. b) Non-symmetrized average after reference free
rotational alignment of 220 wide rotor ends imaged as in a). c) Merging of the angular power spectra of the 220
large rings clearly elucidates a 14-fold rotational symmetry. d) 14-fold symmetrized average of the average displayed
in b).
All topographs exhibit a full gray scale of 20Å and are displayed as reliefs tilted by 5°. Scale bars: 100Å (a) and 20Å
(b and d).

rotor has direct implications for the H+ (Na+ ) / handedness of the 14 subunits more clearly
ATP stoichiometry and for the molecular (Fig. 2d). This stoichiometry is in contrast to
mechanism of ATP synthesis. that reported for the E.coli F o complex which
The high signal-to-noise ratio and was postulated to comprise twelve c-subunits
resolution of the AFM has been exploited to (equal to a subunit III oligomer of chloroplast
image subunits of isolated Fo rotors from ATP synthase), mainly based on cross-
chloroplast ATP synthase (Seelert et al., linking experiments (Jones & Fillingame,
2000). The number of subunits per rotor 1998), genetic engineering (Jones &
could be directly counted in the unprocessed Fillingame, 1998), model building (Dmitriev
AFM topograph (Fig. 2a). In addition, the et al., 1999; Groth & Walker, 1997; Rastogi
reference free (Penczek et al., 1992) average & Girvin, 1999) and biochemical data that
of the 220 rotors showed a clear 14-fold suggest that four protons are required for the
symmetry (Fig. 2b). The angular power synthesis of one ATP. Interestingly, X-ray
spectra of these 220 single rotors were analyses of the yeast FoF1-ATP synthase
merged revealing a clear peak for 14-fold have yielded a decameric rotor (Stock et al.,
rotational symmetry (Fig. 2c; Seelert et al., 1999) while the Na+ driven Fo rotor of the
2000). Consequently the average (Fig. 2b) ATP synthase from Ilyobacter tartaricus was
was 14-fold symmetrized to reveal the found to be undecameric (Stahlberg et al.,

2001). The occurrence of different projection map of the tetramers was calculated
stoichiometries implies that there is a to 8 Å resolution from cryo electron
biological significance which is not as yet micrographs of ice embedded samples
understood (Ferguson, 2000). (Ringler et al., 1999). Topographs with a
resolution of 8 Å resolution have since been
1.2.5. Conformational flexibility of acquired from densely packed vesicles
proteins containing the oligomers and a high amount
of lipid (Fig. 3a; Scheuring et al., 1999b).
Aquaporin Z (AqpZ) is the E.coli water The AqpZ tetramers are clearly laterally
channel responsible for the maintenance of stabilized by their close packing, but
cell turgor during the volume expansion of rotationally misaligned (Fig. 3a). All
cell division (Borgnia et al., 1999; Calamita et tetramers share the crown-like appearance,
al., 1998). After overexpression and although careful comparison of their
purification, tetrameric AqpZ oligomers have polypeptide loops protruding from the bilayer
been reconstituted in the presence of lipids surface reveals their structural individuality
into two-dimensional crystals and densely (Fig. 3a). The aligned averaged topograph
packed vesicles (Ringler et al., 1999). A shows to possess twelve surface protrusions

Figure 3. Modulation of individual AqpZ tetramers by the AFM probe. a) Topograph of a vesicle
densely packed with AqpZ tetramers exposing their extracellular surface. The topograph was recorded in buffer
solution (17 mM Tris-HCl, pH 7.2, 150 mM KCl) using contact mode AFM at an applied force of ~ 80 pN. b)
Four-fold symmetrized average of 289 aligned particles as displayed in a) after translational and angular alignment.
The large peripheral protrusion has been identified as polypeptide loop C connecting two transmembrane α-helices.
c) Minimal force image of ~ 20 individual tetramers. d) The region displayed in c) imaged with a loading force of +
80 pN. e) Computed reconstruction illustrating the force induced conformational change on the extracellular surface
of AqpZ. As the imaging force (left) is increased the large C-loop is progressively displaced from the peripheral
location it otherwise occupies (right; minimal force) (scale bar: 100 Å; full gray scale: 7 Å; Scheuring et al., 1999b).
All topographs exhibit a full gray scale of 7 Å and are displayed as reliefs tilted by 5°. Scale bars: 100 Å (a, c, d and
e) and 20 Å (b).

per tetramer (Fig. 3b). AqpZ has a large C- bacteriorhodopsin (Müller et al., 1999c) and
loop (~ 26 aminoacids), which may exhibit the phi29 connector (Müller et al., 1998) the
pronounced flexibility. Comparison of observed flexibility was of functional
images of AqpZ tetramers recorded at relevance (Brown et al., 1995; Subramanian et
minimal forces (~ 80 pN, Fig. 3c) with those al., 1999; Simpson et al., 2000). In case of
acquired after increasing the stylus loading AqpZ, however, the functional relevance of the
force by + 80 pN (Fig. 3d), shows the four structural flexibility remains to be elucidated.
elongated peripheral protrusions to disappear
under higher loading forces. The computed 1.2.6. The tongue-and-groove interaction
reconstruction of the force induced of MIP tetramers
conformational change of the extracellular
surface of AqpZ demonstrates the effect (Fig. The major intrinsic protein (MIP)
3e). This flexibility and the volume of the expressed in eye lens fiber cells is the
unperturbed elongated peripheral protrusions founding member of the aquaporin family
identify them as the C-loops connecting two (Gorin et al., 1984). Tetrameric MIP
transmembrane α-helices. The use of AFM extracted and purified from sheep lens fiber
images recorded while increasing the applied cells has been reconstituted into double
force to assign flexible protein domains has layered 2D crystals by dialysis of a protein-
been demonstrated previously. For lipid-detergent mixture (Hasler et al., 1998).

Figure 4. Structure and interaction of the major intrinsic protein (MIP). a) Densely packed vesicle
showing the extracellular surface of MIP. The topograph was recorded in buffer solution (20 mM Na-acetate-HCl, pH
5.0, 50 mM NaCl) using contact mode AFM and applied forces of < 100 pN. b) 4-fold symmetrized average of 445
particles imaged as in (a). c) Computer reconstruction of two stacked MIP membranes. To visualize the tongue-to-
groove interaction of their extracellular surfaces the top layer has been displaced by one unit cell to the left and to the
back (Fotiadis et al., 2000).
All topographs exhibit a full gray scale of 14 Å and are displayed as reliefs tilted by 5° (a and b) and by 60° (c). Scale
bars: 100 Å (a), 20 Å (b) and 50 Å (c).

Figure 5. Observing individual chaperonins, GroEL and GroES. a) Topograph of GroEL cylinders
adsorbed to mica. The apical domain of the heptameric GroEL rings positioned at the opening of the barrel are clearly
resolved. b) Seven-fold symmetrized average (n=26) of the particles displayed in (a). c) Topograph of GroES discs
adsorbed to mica. d) Seven-fold symmetrized average (n=54) of the particles displayed in (c). The topographs were
recorded in deionized water after fixation of the sample with 2% glutaraldehyde using contact mode AFM (Mou et al.,
1996). Topographs exhibit a full gray scale of 140 Å (c and b) or of 80 Å (c and d) and are displayed as reliefs tilted
by 5°. Scale bars: 200 Å (a and c) and 50 Å (b and d). (Images by courtesy of Zhifeng Shao, Virginia).

AFM investigations of crystalline sheets and 1.2.7. Imaging the subcomplexes of the
densely packed vesicles elucidated the GroE chaperonin system: GroEL
extracellular (Fig. 4a) and the cytoplasmic and GroES
surface (Fig. 4c, top) at ~ 6 Å resolution
(Fotiadis et al., 2000). Using high forces Imaging the subcomplexes of the GroE
applied to the AFM stylus the double layered chaperonin system: GroEL and GroES
crystals were disected and access to the GroEL is found in the cytoplasm of E. coli
interacting inner surfaces was gained. The where it acts as an ATP dependent molecular
results demonstrate that the extracellular chaperone that ensures the correct folding of
domains of MIP tetramers from two adjacent soluble proteins (Houry et al., 1999). GroEL
layers interact with each other through a consists of 14 identical subunits assembled
'tongue-and-groove' fit (Fig. 4c). This finding into a double ring barrel with a diameter of ~
supports both the proposed dual function of 14 nm. The apical domains of GroEL,
MIP, namely to channel water and to act in exposed at both ends of the cylinder, house
cell-cell adhesion (Benedetti et al., 2000), and hydrophobic aminoacid residues which trap
the result of a recent study of lens fiber cell unfolded or misfolded proteins via hydro-
architecture in mice expressing mutated MIP: phobic interactions (Braig et al., 1994). Mou
lenses that do not integrate functional MIP in et al. (1996) have immobilized GroEL
the plasma membranes have disorganized cylinders head-on onto a mica support and
fiber cells (Shiels et al. 2000). imaged the opening of the cylinder cavity at
submolecular resolution by contact mode
AFM (Fig. 5a, b).
GroES acts as a co-chaperonin together
with GroEL in the folding cycle. The model
postulates that the interaction between GroES

Figure 6. Observing the disassembly of purple membranes. a) Photobleached purple membrane imaged
in buffer solution (10 mM Tris-HCl, pH 7.8, 150 mM KCl) using contact mode AFM at applied forces of ~ 100
pN. Upon exposure to light in the presence of hydroxylamine purple membranes lose their crystallinity. During
disassembly of the purple membrane lattice the bacteriorhodopsin molecules remain assembled into trimers. b)
Three-fold symmetrized average calculated by single particle alignment of 172 disordered trimers as displayed in (a).
c) Adsorption spectra showing the progress of photobleaching in the presence of hydroxylamine. The topographs
exhibit a full gray scale of 8 Å and are displayed as reliefs tilted by 5°. Scale bars: 100 Å (a) and 20 Å (b).

and the apical domain of GroEL is stronger possible to directly visualize the binding and
than the affinity of an unfolded protein to the unbinding of GroES to GroEL in real time, an
latter. Consequently the unfolded protein is important event in the chaperonin mediated
pushed into the cavity (Hartl & Martin, protein folding process (Viani et al., 2000).
1995). To be observed with the AFM, GroES
heptamers were adsorbed onto a mica surface 1.2.8. Observing the assembly of
where they formed a densely packed layer membrane proteins
(Mou et al., 1996). After this, the biological
sample was chemically fixed with Bacteriorhodopsin, a light-activated proton
glutaraldehyde and subsequently imaged by pump of the archae Halobacterium
AFM (Fig. 5c). The calculated average salinarum, shares structural and functional
elucidates the disc-like molecular organization similarities to rhodopsin, a G-protein coupled
of GroES most clearly: a wide ring with a receptor in eukaryotes. In addition to their
diameter of ~ 84 Å which is thought to seven transmembrane α-helices, both proteins
interact with GroEL is topped by a narrower have a retinal as photoactive chromophore.
crown of ~ 45 Å in diameter (Fig. 5d). A Both structures have been solved to atomic
combination of the two experiments by Mou resolution (Belrhali et al., 1999; Luecke et al.,
et al. (1996) was made possible due to the 1999; Pebay-Peyroula et al., 1997; Mitsuoka
development of a fast scanning AFM (Viani et al., 1999; Kimura et al., 1997; Palczewski
et al., 1999). With such an instrument it was et al., 2000). Although the organization of

bacteriorhodopsin in native membranes has require averaging processes. However,
been well studied, little is known for biological processes involve action of single
rhodopsin (Körschen et al., 1999). proteins whose structure cannot be assessed
During biosynthesis the characteristic by structural techniques applying averaging
functional and structural properties of the processes. Insight into molecular interactions
purple membrane are only observed after the and the individual functional states of
formation of bacteriorhodopsin from proteins can only be acquired by the direct
bacterioopsin and retinal (Sumper & observation of single proteins (Weiss, 1999;
Herrmann, 1976a; Sumper & Herrmann, Weiss, 2000). The AFM opens a new way of
1976b). The reversal of this process has been assessing the function-related conformational
monitored by AFM after cleaving the Schiff states of individual proteins at submolecular
base bond between retinal and resolution (Engel & Müller, 2000). The
bacteriorhodopsin in the presence of unique ability of this instrument to directly
hydroxylamine and light (Fig. 6; Möller et al., visualize single proteins in their native
2000). While individual bacterioopsins environment will provide novel insights into
remained stably assembled into trimers, these the interactions between biomolecules and the
trimers lost their orientation relative to the formation of functional assemblies within a
other protein trimers in the membrane (Fig. native membrane.
6a). This disorder of the bacterioopsin trimers Driven by their dynamic clustering, lipids,
increased with the completeness of the Schiff other amphiphilic molecules and membrane
base bond cleavage as can be assessed by the proteins have been found to form rafts that
absorption spectra shown in Fig. 6c. move within the fluid bilayer (Simons &
Interestingly, after washing away the Ikonen, 1997). Such rafts are proposed to
retinaloxime and adding retinal, the Schiff serve as platforms for the attachment of
base bond was re-established. With the re- proteins which control protein sorting and
formation of bacteriorhodopsin, the trimers trafficking through secretory and endocytic
spontaneously re-assembled into membrane pathways (Brown & London, 1998; Simons
patches structurally and spectroscopically & Ikonen, 1997). The assembly of membrane
indistinguishable from native purple proteins into rafts appears to be of significant
membrane. importance during signal transduction. It will
be a challenge to understand the molecular
1.2.9. Outlook mechanisms driving the assembly of rafts. In
this context, the observation of the de- and re-
Proteins are soft and flexible nano assembly of bacteriorhodopsin into fully
machines, which exist in different function active purple membrane can be seen as a step
related conformational states. In nature towards studying the formation of functional
different proteins interact with each other to membrane protein assemblies (Möller et al.,
form functionally relevant complexes. Such 2000). In the future, the ability of AFM to
interactions often induce changes of the image individual proteins will allow more
protein structure (Müller et al., 1999c; complex biological systems to be studied and
Kellenberger, 1968). The understanding of will deliver a novel insight into the biogenesis
these molecular interactions together with the of various native membranes, into the
function related conformational states is of interactions between similar or different
significant importance for biological membrane proteins within a membrane, and
processes. If such interactions are into the formation of supramolecular
synchronized among a class of proteins they complexes.
can be observed using techniques which

1.2.10. Acknowledgement Biophys. J. 69(5), 2103-11.

Butt, H.-J., Jaschke, M. & Ducker, W.
This work was supported by the Swiss (1995). Measuring surface forces in
National Foundation for Scientific Research aqueous solution with the atomic force
(grant 4036 - 44062 to A. E.), the Swiss microscope. Bioelect. Bioenerg. 38, 191-
Priority Project for Micro and Nano System 201.
Technology (MINAST), the EC project for Calamita, G., Kempf, B., Bonhivers, M.,
research on water channels (grant BIO4- Bishai, W. R., Bremer, E. & Agre, P.
CT98-0024 to A. E.), and the Maurice E. (1998). Regulation of the Escherichia coli
Müller Foundation of Switzerland. water channel gene AqpZ. Proc. Natl.
Acad. Sci. USA 95, 3627-3631.
1.2.11. References Czajkowsky, D. M. & Shao, Z. (1998).
Submolecular resolution of single
Benedetti, E. L., Dunia, I., Recouvreur, M., macromolecules with atomic force
Nicolas, P., Kumar, N. M. & Bloemendal, microscopy. FEBS Lett. 430(1-2), 51-54.
H. (2000). Structural organization of gap
junctions as revealed by freeze-fracture Dmitriev, O. Y., Jones, P. C. & Fillingame, R.
and SDS fracture-labeling. Eur. J. Cell H. (1999). Structure of the subunit c
Biol. 79, 575-582. oligomer in the F1Fo ATP synthase:
Model derived from solution structure of
Belrhali, H., Nollert, P., Royant, A., Menzel, the monomer and cross linking in the
C., Rosenbusch, J. P., Landau, E. M. & native enzyme. Proc. Natl. Acad. Sci. USA
Pebay-Peyroula, E. (1999). Protein, lipid 96, 7785-7790.
and water organization in
bacteriorhodopsin crystals: a molecular Drake, B., Prater, C. B., Weisenhorn, A. L.,
view of the purple membrane at 1.9 Å Gould, S. A. C., Albrecht, T. R., Quate, C.
resolution. Struct. Fold. Des. 7(8), 909- F., Cannell, D. S., Hansma, H. G. &
17. Hansma, P. K. (1989). Imaging crystals,
polymers, and processes in water with the
Binnig, G., Quate, C. F. & Gerber, C. (1986). atomic force microscope. Science 243,
Atomic force microscope. Phys. Rev. Lett. 1586-1588.
56(9), 930-933.
Engel, A. & Müller, D. J. (2000). Observing
Borgnia, M., Kozono, D., Calamita, G., single biomolecules at work with the
Maloney, P. C. & Agre, P. (1999). atomic force microscope. Nat. Struct. Biol.
Functional reconstitution and 7(9), 715-718.
characterization of AqpZ, the E. coli water
channel protein. J. Mol. Biol. 29(5), 1169- Ferguson, S. J. (2000). ATP synthase: What
1179. determines the size of a ring? Curr. Biol.
10, R804-R808.
Braig, K., Otwinowski, Z., Hedge, R.,
Boisvert, D. C., Joachimiak, A., Horwich, Fotiadis, D., Hasler, L., Müller, D. J.,
A. L. & Sigler, P. B. (1994). The crystal Stahlberg, H., Kistler, J. & Engel, A.
structure of the bacterial chaperonin (2000). Surface tongue-and-groove
GroEL at 2.8Å. Nature 371, 578-585. contours on lens MIP facilitate cell-to-cell
adherence. J. Mol. Biol. 300(4), 779-789.
Brown, D. A. & London, E. (1998).
Functions of lipid rafts in biological Gorin, M. B., Yancey, S. B., Cline, J., Revel,
membranes. Annu. Rev. Cell. Dev. Biol. J.-P. & Horwitz, J. (1984). The major
14, 111-136. intrinsic protein (MIP) of the bovine lens
fiber membrane: Characterization and
Brown, L. S., Varo, G., Needleman, R. & structure based on cDNA cloning. Cell 39,
Lanyi, J. K. (1995). Functional 49-59.
significance of a protein conformation
change at the cytoplasmic end of helix F Groth, G. & Walker, J. E. (1997). Model of
during the bacteriorhodopsin photocycle. the c-subunit oligomer in the membrane

domain of F-ATPases. FEBS Lett. 410, resolution images of protein surfaces in an
117-123. aqueous environment. Proc. Natl. Acad.
Sci. USA 91, 836-838.
Hartl, F. U. (1996). Molecular chaperones in
cellular protein folding. Nature 381, 571- Kato-Yamada, Y., Noji, H., Yasuda, R.,
580. Kinosita, K. J. & Yoshida, M. (1998).
Direct observation of the rotation of
Hartl, F.-U. & Martin, J. (1995). Molecular epsilon subunit in F1-ATPase. J. Biol.
chaperones in cellular protein folding. Chem . 273, 19375-19377.
Curr. Opin. Struct. Biol. 5, 92-102.
Kellenberger, E. (1968). Studies on the
Hasler, L., Heymann, J. B., Engel, A., Kistler, morphopoiesis of the head of phage T-
J. & Walz, T. (1998). 2D crystallization even. V. The components of the T4 capsid
of membrane proteins: rationales and and of other, capsid-related structures.
examples. J. Struct. Biol. 121, 162-171. Virology 34(3), 549-561.

Hasler, L., Walz, T., Tittmann, P., H., G., Kimura, Y., Vassylev, D. G., Miyazawa, A.,
Kistler, J. & Engel, A. (1998). Purified Kidera, A., Matsushima, M., Mitsuoka, K.,
lens major intrinsic protein (MIP) forms Murata, K., Hirai, T. & Fujiyoshi, Y.
highly ordered tetragonal two-dimensional (1997). Surface of bacteriorhodopsin
arrays by reconstitution. J. Mol. Biol. 279, revealed by high-resolution electron
855-864. microscopy. Nature 389(6647), 206-211.

Haupts, U., Tittor, J. & Oesterhelt, D. (1999). Körschen, H. G., Beyermann, M., Müller, F.,
Closing in on bacteriorhodopsin: progress Heck, M., Vantler, M., Koch, K.-H.,
in understanding the molecule. Annu. Rev. Kellner, R., Wolfrum, U., Bode, C.,
Biophys. Biomol. Struct. 28, 367-399. Hofmann, K. P. & Kaupp, U. B. (1999).
Interaction of glutamic-acid-rich proteins
Heymann, J. B., Müller, D. J., Landau, E., with the cGMP signaling pathway in rod
Rosenbusch, J., Pebay-Peroulla, E., Büldt, photoreceptors. Nature 400, 761-766.
G. & Engel, A. (1999). Charting the
surfaces of purple membrane. J. Struct. Luecke, H., Schobert B., Richter, H.T.,
Biol. 128(3), 243-249. Cartailler, J.P., Lanyi, J.K. (1999).
Structure of bacteriorhodopsin at 1.55 Å
Hoh, J. H., Lal, R., John, S. A., Revel, J.-P. & resolution. J. Mol. Biol. 291(4), 899-911.
Arnsdorf, M. F. (1991). Atomic force
microscopy and dissection of gap Mitsuoka, K., Hirai, T., Murata, K.,
junctions. Science 253, 1405-1408. Miyazawa, A., Kidera, A., Kimura, Y. &
Fujiyoshi, Y. (1999). The structure of
Hörber, J. K. H., Mosbacher, J., Haberle, W., bacteriorhodopsin at 3.0 Å resolution
Ruppersberg, J. P., Sakmann, B. (1995). based on electron crystallography:
A look at membrane patches with the implication of the charge distribution. J.
scanning force microscope. Biophys. J. 68 Mol. Biol. 286(3), 861-82.
(5), 1687-1693.
Möller, C., Büldt, G., Dencher, N., Engel, A.
Houry, W. A., Frishman, D., Eckerskorn, C., & Müller, D. J. (2000). Reversible loss of
Lottspeich, F. & Hartl, F. U. (1999). crystallinity on photobleaching purple
Identification of in vivo substrates of the membrane in presence of hydroxylamine.
chaperonin GroEL. Nature 402(6758), J. Mol. Biol. 301(4), 869-879.
Mou, J., Czajkowsky, D. M., Sheng, S., Ho,
Jones, P. C. & Fillingame, R. H. (1998). R. & Shao, Z. (1996a). High resolution
Genetic fusions of subunit c in the Fo surface structure of E. coli GroES
sector of H+ -transporting ATP synthase. oligomer by atomic force microscopy.
J. Biol. Chem. 273(45), 29701-29705. FEBS Lett. 381, 161-164.

Karrasch, S., Hegerl, R., Hoh, J., Baumeister, Mou, J., Sheng, S., Ho, R. & Shao, Z.
W. & Engel, A. (1994). Atomic force (1996b). Chaperonins GroEL and GroES:
microscopy produces faithful high- views from atomic fore microscopy.

Biophys. J. 71, 2213-2221. Müller, D. J., Sass, H.-J., Müller, S., Büldt,
G. & Engel, A. (1999c). Surface
Mou, J. X., Yang, J. & Shao, Z. F. (1995). structures of native bacteriorhodopsin
Atomic force microscopy of cholera toxin depend on the molecular packing
B-oligomers bound to bilayers of arrangement in the membrane. J. Mol.
biologically relevant lipids. J. Mol. Biol. Biol. 285(5), 1903-1909.
248(3), 507-512.
Müller, D. J., Schabert, F. A., Büldt, G. &
Müller, D. J., Amrein, M. & Engel, A. Engel, A. (1995). Imaging purple
(1997). Adsorption of biological membranes in aqueous solutions at
molecules to a solid support for scanning subnanometer resolution by atomic force
probe microscopy. J. Struct. Biol. 119(2), microscopy. Biophys. J. 68, 1681-1686.
Noji, H., Yasuda, R., Yoshida, M. & Kinosita,
Müller, D. J., Baumeister, W. & Engel, A. K. (1997). Direct observation of the
(1996). Conformational change of the rotation of F1-ATPase. Nature 386, 299-
hexagonally packed intermediate layer of 302.
Deinococcus radiodurans imaged by
atomic force microscopy. J. Bacteriol.
178(11), 3025-3030. Oesterhelt, F., Oesterhelt, D., Pfeiffer, M.,
Engel, A., Gaub, H. E. & Müller, D. J.
Müller, D. J., Baumeister, W. & Engel, A. (2000). Unfolding pathways of individual
(1999a). Controlled unzipping of a bacteriorhodopsins. Science 288(5463),
bacterial surface layer with atomic force 143-146.
microscopy. Proc. Natl. Acad. Sci. USA
96(23), 13170-13174. Palczewski, K., Kumasaka, T., Hori, T.,
Behnke, C. A., Motoshima, H., Fox, B. A.,
Müller, D. J., Büldt, G. & Engel, A. (1995). Le Trong, I., Teller, D. C., Okada, T.,
Force-induced conformational change of Stenkamp, R. E., Yamamoto, M. &
bacteriorhodopsin. J. Mol. Biol. 249(2), Miyano, M. (2000). Crystal structure of
239-243. rhodopsin: A G protein-coupled receptor.
Science 289, 739-745.
Müller, D. J. & Engel, A. (1997). The height
of biomolecules measured with the atomic Pebay-Peyroula, E., Rummel, G.,
force microscope depends on electrostatic Rosenbusch, J. P. & Landau, E. M.
interactions. Biophys. J. 73, 1633-1644. (1997). X-ray structure of
bacteriorhodopsin at 2.5 angstroms from
Müller, D. J. & Engel, A. (1999). pH and microcrystals grown in lipidic cubic
voltage induced structural changes of phases. Science 277(5332), 1676-1681.
porin OmpF explain channel closure. J.
Mol. Biol. 285(4), 1347-1351. Palczewski, K., Kumasaka, T., Hori, T.,
Behnke, C. A., Motoshima, H., Fox, B. A.,
Müller, D. J., Engel, A., Carrascosa, J. & Le Trong, I., Teller, D. C., Okada, T.,
Veléz, M. (1997). The bacteriophage ø29 Stenkamp, R. E., Yamamoto, M. &
head-tail connector imaged at high Miyano, M. (2000). Crystal structure of
resolution with atomic force microscopy rhodopsin: A G protein-coupled receptor.
in buffer solution. EMBO J. 16, 101-107. Science 289(5480), 739-45.

Müller, D. J., Fotiadis, D. & Engel, A. Penczek, P., Radermacher, M. & Frank, J.
(1998). Mapping flexible protein domains (1992). Three-dimensional reconstruction
at subnanometer resolution with the AFM. of single particles embedded in ice.
FEBS Lett. 430(1,2), 105-111. Ultramicroscopy 40, 33-53.

Müller, D. J., Fotiadis, D., Scheuring, S., Raab, A., Han, W., Badt, D., Smith-Gill, S.,
Müller, S. A. & Engel, A. (1999b). Lindsay, S., Schindler, H. & Hinterdorfer,
Electrostatically balanced subnanometer P. (1999). Antibody recognition imaging
imaging of biological specimens by by atomic force microscopy. Nature
atomic force microscopy. Biophys. J. 76, Biotech. 17, 902-905.

Rastogi, V. K. & Girvin, M. E. (1999). Schmitt, L., Ludwig, M., Gaub, H. E. &
Structural changes linked to proton Tampe, R. (2000). A metal-chelating
translocation by subunit c of the ATP microscopy tip as a new toolbox for
synthase. Nature 402(6759), 263-8. single-molecule experiments by atomic
force microscopy. Biophys. J. 78(6),
Ringler, P., Borgnia, M., Stahlberg, H., Agre, 3275-85.
P. & Engel, A. (1999). Structure of the
water channel AqpZ from Escherichia coli Seelert, H., Poetsch, A., Dencher, N. A.,
revealed by electron crystallography. Engel, A., Stahlberg, H. & Müller, D. J.
Journal Molecular Biology 291, 1181- (2000). Proton powered turbine of a plant
1190. motor. Nature 405, 418-419.

Ros, R., Schwesinger, F., Anselmetti, D., Shiels, A., Mackay, D., Bassnett, S., Al-
Kubon, M., Schäfer, R., Plückthun, A. & Ghoul, K &. Kuszak, J. (2000).
Tiefenauer, L. (1998). Antigen binding Disruption of lens fiber cell architecture in
forces of individually addressed single- mice expressing a chimeric AQP0-LTR
chain Fv antibody molecules. Proc. Natl. protein. FASEB 14, 2207-2212
Acad. Sci. 95, 7402-7405.
Simons, K. & Ikonen, E. (1997). Functional
Rotsch, C. & Radmacher, M. (1997). rafts in cell membranes. Nature 387, 569-
Mapping local electrostatic forces with the 572.
atomic force microscope. Langmuir
13(10), 2825-2832. Simpson, A. A., Tao, Y., Leiman, P. G.,
Badasso, M. O., He, Y., Jardine, P. J.,
Sabbert, D., Engelbrecht, S. & Junge, W. Olson, N. H., Morais, M. C., Grimes, S.,
(1996). Intersubunit rotation in active F- Anderson, D. L., Baker, T. S. &
ATPase. Nature 381, 623-625. Rossmann, M. G. (2000). Structure of the
bacteriophage phi29 DNA packaging
Schabert, F. A. & Engel, A. (1994). motor. Nature 408(6813), 745-50.
Reproducible acquisition of Escherichia
coli porin surface topographs by atomic Simons, K. & Ikonen, E. (1997). Functional
force microscopy. Biophys. J. 67(6), rafts in cell membranes. Nature 387, 569-
2394-2403. 572.

Schabert, F. A., Henn, C. & Engel, A. (1995). Stahlberg, H., Müller, D. J., Suda, K.,
Native Escherichia coli OmpF porin Fotiadis, D., Engel, A., Meier, T., Matthey,
surfaces probed by atomic force U. & Dimroth, P. (2001). Bacterial
microscopy. Science 268, 92-94. sodium ATPase has an undecameric rotor.
EMBO Reports , in press.
Scheuring, S., Müller, D. J., Ringler, P.,
Heymann, J. B. & Engel, A. (1999a). Stock, D., Leslie, A. G. & Walker, J. E.
Imaging streptavidin 2D crystals on (1999). Molecular architecture of the
biotinylated lipid monolayers at high rotary motor in ATP synthase. Science
resolution with the atomic force 286(5445), 1700-5.
microscopy. J. Microsc. 193, 28-35.
Subramaniam, S. (1999). The structure of
Scheuring, S., Reiss-Husson, F., Engel, A., bacteriorhodopsin: an emerging
Rigaud, J. L. & Ranck, J. L. (2001). High consensus. Curr Opin Struct Biol 9(4),
resolution AFM topographs of Rubrivivax 462-8.
gelatinosus light-harvesting complex
LH2. EMBO J. in press. Sumper, M. & Herrmann, G. (1976a).
Biogenesis of purple membrane:
Scheuring, S., Ringler, P., Borgnia, M., Regulation of bacteriorhodopsin. FEBS
Stahlberg, H., Müller, D. J., Agre, P. & Lett. 69(1), 149-152.
Engel, A. (1999b). High resolution
topographs of the Escherichia coli Sumper, M. & Herrmann, G. (1976b).
waterchannel aquaporin Z. EMBO J. 18, Biosynthesis of purple membrane:
4981-4987. Control of retinal synthesis by bacterio-
opsin. FEBS Lett. 71(2), 333-336.

Thomson, N. H., Fritz, M., Radmacher, M., Viani, M. B., Schäfer, T. E., Chand, A., Rief,
Cleveland, J. P., Schmidt, C. F. & M., Gaub, H. & Hansma, P. K. (1999).
Hansma, P. K. (1996). Protein tracking Small cantilevers for force spectroscopy
and detection of protein motion using of single molecules. J. Appl. Phys. 86(4),
atomic force microscopy. Biophys. J. 2258-2262.
70(5), 2421-2431.
Weiss, S. (1999). Fluorescence spectroscopy
Viani, M. B., Pietrasanta, L. I., Thompson, J. of single biomolecules. Science
B., Chand, A., Gebeshuber, I. C., Kindt, J. 283(5408), 1676-83.
H., Richter, M., Hansma, H. G. &
Hansma, P. K. (2000). Probing protein- Weiss, S. (2000). Measuring conformational
protein interactions in real time. dynamics of biomolecules by single
Nat.Struct. Biol. 7(8), 644-647. molecule fluorescence spectroscopy. Nat
.Struct. Biol. 7(9), 724-9.

1.3. Imaging streptavidin 2D-crystals on biotinylated lipid monolayers at
high resolution with the atomic force microscope

Simon Scheuring, Daniel J. Müller, Philippe Ringler, J. Bernard Heymann, Andreas

M.E.Müller Institute for Structural Biology, Biozentrum, University of Basel, CH-4056 Basel,

1.3.1. Summary 1996) or silanized glass (Karrasch et al.,
1993) allowing covalent crosslinking to be
Streptavidin crystals were grown on achieved. These supports are suitable for the
biotinylated lipid monolayers at an air/water adsorption of membrane proteins
interface and transferred onto highly oriented reconstituted into lipid bilayers (Jap et al.,
pyrolytic graphite (HOPG). These arrays 1992) and of single particles such as soluble
could be imaged to a resolution below 1 nm proteins (Mou et al., 1996).
with the atomic force microscope. The surface In this report atomically flat HOPG
topographs obtained were compared with (highly oriented pyrolytic graphite) was used
negative stain electron microscopy images as a hydrophobic support to acquire high
and the atomic model determined by X-ray resolution topographs of streptavidin 2D
crystallography. The streptavidin tetramer (60 crystals on biotinylated lipid monolayers
kD) exposes two free biotin binding sites to using the AFM. HOPG is produced by
the buffer solution, while two are occupied by deposition of carbon at high temperature (up
the linkage to the lipid monolayer. Therefore to 3000°C) and under pressure from the gas
the streptavidin 2D crystals can be used as phase (Moore, 1973). The material thus
nanoscale matrices for binding biotinylated obtained consists of crystallites that are well
compounds. Furthermore, this HOPG-based oriented perpendicular to the basal graphite
preparation method provides a general novel planes (Ohler et al., 1997), and can easily be
approach to study the structure of protein cleaved with scotch tape. On a macroscale
arrays assembled on lipid monolayers with HOPG is not flat, but the surface is separated
the AFM. into atomically flat terraces. The flatness of
these plateaus assures that the surface
1.3.2. Introduction features are specimen specific and not due to
irregularities of the substrate itself. The
The atomic force microscope (AFM) graphite, consisting of hexagonally ordered
(Binnig et al., 1986) has become a powerful carbon atoms, presents a nonpolar strongly
tool in structural biology, because topographs hydrophobic surface.
of biomolecules can be acquired under Streptavidin is a tetrameric protein with
physiological conditions at subnanometer four biotin binding sites (Green, 1975). The
resolution. High resolution AFM requires ease with which streptavidin crystallizes in 2D
atomically flat surfaces (Müller et al., 1995). arrays on a biotinylated lipid monolayer
Frequently used supports to immobilize (Darst et al., 1991; Avila-Sakar & Chiu,
biological objects for imaging with the AFM 1996) makes it an ideal model system to
are mica having a polar surface (Müller et al., investigate two-dimensional crystals grown
1997) and functionalized gold (Wagner et al., on lipid monolayers that are transferred to

HOPG. In addition, a comparison can be 1984), washed glass (Karrasch et al., 1993),
made with the transmission electron untreated glass, HOPG (Moore, 1973), and
microscope (TEM) of the same crystals as teflon. Right after deposition of the drops the
well as the available atomic structure from X- diameters were photographed at 6x
ray crystallography (Avila-Sakar & Chiu, magnification using a binocular microscope
1996; Hendrickson et al., 1989). Topographs with a mounted camera, and the diameters
recorded with the AFM compare favorably measured.
with negatively stained samples and the
atomic structure of streptavidin. Crystallization of streptavidin on
biotin-lipid monolayer
1.3.3. Materials and Methods
Highly ordered streptavidin arrays were Materials produced by depositing 15 µl of streptavidin
solution (10 mM Tris-HCl, pH 7.5, 150
Streptavidin and biotin-LC-DPPE (biotin- NaCl) at a concentration of 0.1 mg/ml in a
longchain-dipalmitoyl phosphatidyl ethanol- Teflon well 0.5 mm deep and 4 mm in
amine) were obtained from Pierce Ltd. (Rock- diameter (Fig. 1a). A 0.5 µl drop of the lipid
ford, USA), DOPC (dioleoyl phosphatidyl mixture (0.5 mg/ml Biotin-LC-DPPE :
choline) from Avanti Polar Lipids (Birming- DOPC, 1:4 (mol : mol), in chloroform :
ham, AL), mica from Mica New York (Varick hexane, 1:1 (vol : vol)) was then deposited on
Street, N. Y. 10013), HOPG from Advanced top of the protein solution with a Hamilton
Ceramics Corporation (Cleveland, USA), and syringe (Fig. 1a). Incubation overnight at
Araldit from Ciba-Ceigy (Basel, Switzerland). room temperature allowed the adsorption of
protein to the lipid monolayer and subsequent Hydrophobicity measurement growing of 2D-crystals (Fig. 1b, c).

A 10 µl drop of millipore filtered H 2 O was
deposited on the substrates mica (Bailey,

Figure 1 . Crystallization o f streptavidin on a biotin-lipid containing monolayer and
adsorption to HOPG. a) A 15 µl drop of streptavidin solution (0.1 mg/ml in 10 mM Tris-HCl, pH 7.5, 150
NaCl) was deposited into a teflon well and a 0.5 µl drop of lipid mixture (Biotin-LC-DPPE : DOPC, 1:4) was
spread on the drop to form a monolayer. b) Overnight incubation at room temperature allowed streptavidin binding
to the lipid monolayer and the formation of 2D-streptavidin crystals. c) The monolayer was adsorbed to freshly
cleaved HOPG and d) mounted in the AFM in a drop of scanning buffer (10 mM Tris-HCl, pH 7.5, 150 KCl)

36 Atomic force microscopy (AFM) deposited on the lipid monolayer covering the
protein solution in a teflon well for 1 min.
Highly ordered pyrolytic graphite The grid was removed and blotted with filter
(HOPG) with dimensions of 3 mm x 3 mm x paper, and subsequently washed three times
1 mm was glued with water-insoluble Araldit with double distilled water. The specimen was
epoxy onto a teflon disc (diameter: 11 mm). negatively stained twice for 15 sec with
The teflon disc was glued to a steel disc 0.75% uranyl formate, blotted, and dried in an
(diameter: 10 mm) which was mounted in the air stream. Micrographs were taken in a
AFM (Fig. 1d). Subsequently, the HOPG Hitachi H7000 TEM at low dose conditions
was cleaved with scotch tape, ensuring that the at 50000 x magnification. The negatives were
surfaces separate over the whole area. The digitized with a Leafscan-45 (Leaf Systems
surface was scanned in buffer solution (10 Inc., Cupertino, CA) at a stepsize of 20 µm
mM Tris-HCl, pH 7.5, 150 mM KCl) with a (~0.4 nm at the specimen) and selected areas
Nanoscope III AFM (Digital instruments, were correlation averaged (Saxton &
Santa Barbara, Calif.) equipped with a 120 Baumeister, 1982).
µm scanner (J-scanner). A 120 µm long
cantilever from Digital instruments (k = 0.38 1.3.4. Results
N/m) was used. For recording HOPG
topographs the AFM was operated at minimal Hydrophobicity and topography of
force (<0.2 nN) and 4 Hz scan speed. HOPG
To adsorb 2D streptavidin crystals, the
freshly cleaved HOPG was brought into The hydrophobicity of HOPG was
contact with the monolayer on the surface of compared with other surfaces using the sitting
the drop in the teflon well (Fig. 1). The drop diameter method (Karrasch et al., 1993).
sample was kept wet throughout the In comparison with the well-used substrates,
preparation procedure. A drop of 30 µl mica and glass, HOPG exhibits an increased
scanning buffer (10 mM Tris-HCl, pH 7.75, hydrophobicity (Table 1). As indicated by the
200 mM KCl) was immediately added and difference in drop diameter, the difference in
the specimen was mounted in the atomic force
microscope (AFM). The 120 µm scanner (J- T a b l e 1 . Determination of the hydrophobicity of
various substrates by measuring the diameter of 10 µl
scanner) was used together with oxide nanopure water droplets deposited on each substrate.
sharpened Si3N4 cantilevers from Digital
instruments with a length of 200 µm (k = Substrate Diameter (mm)
0.06 N/m). For imaging the AFM was
mica † 12.3± 1.9 (n=41)
operated at constant force mode applying
minimal forces (<0.2 nN) at a scanning speed glass (etched)* 7.5 ± 0.7 (n=50)
of 4-6 Hz. The images were correlation glass (washed)† 7.4 ± 1.3 (n=66)
averaged with the SEMPER image processing glass (washed)* 7.4 ± 0.7 (n=50)
system (Saxton et al., 1979).
glass (untreated)† 4.9 ± 0.1 (n=60) Transmission electron microscopy glass (silanized)* 4.1 ± 0.2 (n=40)
(TEM) HOPG † 3.7 ± 0.2 (n=54)
teflon † 3.1 ± 0.1 (n=70)
For transmission electron microscopy,
samples were prepared on a copper grid spherical drop 2.67 (V = 4/3πr3)
covered by a parlodion film and a carbon
* (Karrasch et al., 1993)
layer. A 10 day old hydrophobic grid was † This work

Figure 2. a) Height image of HOPG recorded in buffer solution (10 mM Tris-HCl, pH 7.5, 150 mM KCl) at
minimal force and a scan speed of 4 Hz, showing terraces with atomically flat surfaces (scale bar: 400 nm; full gray
scale: 5 nm); Inset: Fourier filtered image of HOPG atoms scanned at minimal force and a scan speed of 12 Hz (scale
bar: 5 Å; full gray scale: 2Å). b) Deflection image of a) (scale bar: 400 nm; full gray scale: 1 nm). c) Height image
of streptavidin 2D crystals on biotinylated lipid monolayers scanned in buffer solution (10 mM Tris-HCl, pH 7.6,
150 mM KCl) using minimal force and 4.3 Hz scan speed, showing the HOPG terraces (arrows) overlayed by
crystalline patches of irregular shape (scale bar: 300 nm; full gray scale: 6 nm); asterisk indicates high plateau. d)
Deflection image of c) (scale bar: 300 nm; full gray scale: 1 nm).

hydrophobicity between HOPG and silanized Crystallization of streptavidin on
glass is significant. Indeed, streptavidin biotin-lipid monolayer
crystals on a biotinylated monolayer could
not be successfully adsorbed to the silanized Of many soluble proteins crystallized on
coverslip, whereas they could be transferred lipid monolayers (Brisson et al., 1994),
to HOPG reproducibly. streptavidin is the best-studied simple system
The surface of freshly cleaved HOPG to test the suitability of HOPG as a substrate
imaged in buffer solution (10 mM Tris-HCl, and develop a protocol for the preparation of
pH 7.5, 150 mM KCl) revealed large smooth such crystals for AFM. A lipid monolayer
terraces (Fig. 2 a). The atomically flat terraces containing biotin lipids was formed on a drop
had dimensions up to 2 µm, providing large of streptavidin solution and incubated to allow
areas over which adsorbed specimens can be the adsorption and crystallization of the
imaged. Scanning a flat area of 200 nm, protein (Fig. 1). The appearance of 2D-
which is about the scan range for imaging crystals was inspected by TEM of negatively
biomolecules at high resolution, a roughness stained samples and AFM. Many crystals of
(rms) of 0.02 nm was measured. The carbon varying size (80-1500 nm) and shape were
atoms of the HOPG could be seen as found by TEM. Crystals imaged by AFM
hexagonally ordered arrays with a lattice were in general smaller (50-400 nm),
constant of 0.23 ± 0.03 nm (Fig. 2 a, inset). suggesting that crystals adsorbed to the
The HOPG substrate mounting protocol HOPG surface may be disrupted by the
described above (Materials and Methods, unevenness of the substrate. By lowering the
Fig.1 d) and used for all measurements in this speed in approaching the HOPG substrate to
work, was thus good enough to achieve the monolayer surface (Fig. 1c), and by
atomic resolution. cleaving the HOPG with the scotch tape at an
angle >120˚, bigger crystals (up to 3 µm)
were found apparently adsorbed to the indicate that the transfer was influenced by
atomically flat surfaces of the terraces (Fig. the unevenness of the substrate.
2c, d). Selection of high terraces for scanning The different crystalline patches typically
contributed to the quality of high resolution exhibit different lattice orientations. Between
imaging (see Fig. 2c). the crystalline patches the protein was not
ordered and not well resolved. Imaging AFM of streptavidin crystals streptavidin crystals at medium magnification
(Fig. 3a, b) allows single proteins missing
The overview AFM micrograph (Fig. 2c, within the crystal lattice to be seen (Fig. 3a,
d) shows many crystalline patches of arrow 1) and proteins floating away from the
streptavidin 2D-crystals over a scan range of crystal edges on the lipid monolayer (Fig. 3a,
2.3 µm. The arrows indicate the edges of arrow 2). The crystal in the center and the left
HOPG terraces that are clearly visible as of Fig. 3a had continuous lattice lines,
discontinuities between areas of crystalline although there were big defects (~100 * ~30
streptavidin. Further, breaks between crystal nm) within the crystal. The crystalline patch
patches on the terraces themselves may on the right top of Fig. 3a reveals a different

Figure 3. a) Height image of streptavidin 2D crystals showing defects of single missing proteins (arrow 1) and
square shaped edges with loosened proteins (arrow 2) scanned in buffer solution (10 Tris-HCl, pH 7.2, 20 mM KCl)
(scale bar: 100 nm; full gray scale: 6 nm). b) Deflection image of a) (scale bar: 100 nm; full gray scale: 0.3 nm). c)
Height image of the streptavidin 2D crystal of the central region of a) at higher magnification (scale bar: 50 nm; full
gray scale: 6 nm), arrows correspond to those in a); Inset: Power spectrum of c) (circle marks spot of 7th order;
resolution: 0.83 nm). d) Average over 7 different AFM images using different tips and scan angles (square indicates
the unit cell with dimensions of a = b = 8.2 ± 0.2 nm, g = 88 ± 2˚; indexed according to Avila-Sakar & Chiu, 1996;
full gray scale: 1 nm).

Figure 4. a) Low dose electron micrograph of a negatively stained streptavidin 2D-crystal (scale bar: 100 nm);
Inset: Power spectrum of a) (circle marks spot of 3rd order; resolution: 2.59 nm). b) Average of picture a) (square
indicates the unit cell with dimensions a = b = 8.2 ± 0.2 nm, g = 90 ± 2˚).

orientation in comparison to the crystal in the compared with low dose electron micrographs
center of the image. Since the appearance of of negatively stained streptavidin crystals
these two crystals was the same tip (Fig. 4). Features up to ~2 nm were resolved
asymmetries can be excluded. At higher (see diffraction pattern, inset Fig. 4a). The
magnification (Fig. 3c) the central region of unit cell containing two tetramers were found
the big streptavidin crystal is imaged at ~0.8 to have dimensions of a = b = 82 ± 2 Å and g
nm lateral resolution (see diffraction pattern, = 90 ± 2 ˚. In the average from EM (Fig. 4b)
inset of Fig. 3c). The high signal to noise each tetramer shows four densities of about
ratio of the AFM allows every single protein equal intensity. These correspond to the two
and missing proteins to be resolved (Fig. 3c, large and two small protrusions in the AFM
arrow 1). image (Fig. 3d).
As it is known from other studies (Avila-
Sakar & Chiu, 1996; Hendrickson et al., 1.3.5. Discussion
1989), the streptavidin crystals showed a
C222 point group symmetry with unit cell In this work we introduce a novel
dimensions of a = b = 82 Å and g = 90 ˚ with preparation method for high resolution AFM
two tetramers per unit cell (Fig. 3d). In AFM of 2D crystals grown by the lipid monolayer
images we found unit cell dimensions of a = method (Uzgiris & Kornberg, 1983). The
b = 84 ± 2 Å and g = 88 ± 2 ˚ (Fig 3c (inset), streptavidin-biotin lipid system was used as
d). The height of the crystalline patches above test sample. The fact that we only take
the lipid monolayer were measured as 4.65 ± advantage of hydrophobic interactions
0.3 nm (n=20). between the fatty acyl chains of the lipid and
the pure carbon of the HOPG, indicates the TEM of streptavidin crystals generality of this approach. Interestingly, the
difference in hydrophobicity (Table 1)
To examine distortions and other effects between silanized glass (drop diameter: 4.1 ±
the adsorption of the streptavidin crystals to 0.2 nm) and HOPG (drop diameter: 3.7 ± 0.2
HOPG introduce, the AFM images were nm) is crucial for the adsorption of a lipid

monolayer prepared as described in this work measuring the rupture force of a single biotin-
(Fig. 1). HOPG is a layered material breaking streptavidin bond. To this end, the AFM tip
into different layers on cleavage. This can be was biotinylated and subsequently
seen in the AFM images as terraces of streptavidin was adsorbed to it. This tip was
varying height and size (up to 2 µm in width) approached to a biotinylated agarose bead and
(Fig. 2a, b). Thus, HOPG appears uneven at retracted. Free biotin binding sites on the
low magnification, but each terrace is streptavidin attached to the tip bound the
atomically flat, as evidenced by the imaging of biotinylated bead and the measured force
individual carbon atoms (Fig. 2a; inset). This curve exhibit multiple peaks separated by 160
also indicates that the substrate is stable ± 20 pN, the break force of a single bond
enough for high resolution AFM. Feedback (Florin et al., 1994).
loop gains of the instrument can therefore be With this new preparation method
set high enough for imaging atoms as well as streptavidin can be imaged with the AFM at
biological samples adsorbed to the substrate. submolecular resolution. Proteins within the
The surface features on the adsorbed 2D crystals have a high lateral stability as
streptavidin crystals are not influenced by they are supported in the crystallographic
abnormalities in the underlying substrate packing arrangement. As a consequence they
support (Fig. 2, 3). Streptavidin 2D crystals are better resolved than those tetramers
are found to preferentially adsorb to the high floating away from a crystal edge. The slight
plateaus of the HOPG substrate (Fig. 2c, d), deviation of the lattice parameters measured
which is probably a result of how the contact on the 2D crystal (a = b = 84 ± 2 Å and g =
between HOPG and the monolayer is 88 ± 2 ˚) from the literature data may either
established during the transfer procedure. be the result of drift and distortion of the
Since streptavidin-biotin is a well known raster scan, or may be caused by the transfer
system, different attempts have been made to of the sample. The height measured on the
achieve molecular resolution in the AFM to streptavidin crystals (4.65 ± 0.3 nm) under
investigate the biotin binding with the protein. appropriate ion conditions (Müller & Engel,
A lipid bilayer was transferred to mica using 1997) compares favorably to the thickness of
the Langmuir-Blodgett (LB) and the the molecule derived from the atomic
Langmuir-Schaefer technique. While the first coordinates (4.3 nm; (Hendrickson et al.,
layer consisted only of DPPE (dipalmitoyl 1989)). It is reasonable to assume that the
phosphatidyl ethanolamine), the second layer LC-part (CH2)6 of the biotin-LC-DPPE,
was a lipid mixture of biotin-DPPE and which reduces the steric hindrance for
DMPE (dimyristoyl phosphatidyl streptavidin binding, is the reason of this
ethanolamine). Streptavidin was added, but difference, indicating a good correspondence
the protein did not form highly ordered arrays between the X-ray data and the AFM height
and molecular resolution could not be analysis.
achieved (Weisenhorn et al., 1992). Another The unit cell parameters (a = b = 82 ± 2 Å
work reports on the adsorption of streptavidin and g = 90 ± 2 ˚) determined by negative
to a support that was biotinylated after stain electron microscopy are similar to those
photoactivation of well defined regions. found by cryo-EM (Avila-Sakar & Chiu,
Although the biotinylated sites showed high 1996). The small difference between the EM-
surface corrugations, only granular features and AFM-derived unit cell dimensions (see
with a diameter of 30 nm could be seen in previous paragraph) may be due to the strong
high magnification images (Mazzola & adsorption to the HOPG, or to the commonly
Fodor, 1995). Furthermore, the biotin- observed slight drift in the AFM. The former
streptavidin interaction was studied by is also manifested in the breakup of the

Figure 5. Composite of the TEM average (left side), the AFM average (right side) and a surface contour model
derived from the atomic coordinates (lstp.pdb, Protein Databank) (middle). Because of the C222 symmetry of the
tetrameric streptavidin molecule, subunits on left top and right bottom are exposed to the AFM tip, while subunits
right top and left bottom are more concealed. The EM projection map resolves four equal density peaks, while the
AFM surface reveals the differences in access to the subunits from one side of the crystal.

crystal patches (Fig. 2c), compared to electron size to that of the streptavidin tetramer or
micrographs where the contiguous crystals smaller may form a regular packing following
are bigger (Fig. 4a). In contrast to the four the lattice of the streptavidin crystal. Larger
equal density peaks in the negative stain proteins would likely arrange in different
electron microscopy average representing a orientations, but because of the high signal-
projection through the negative stain envelope to-noise ratio of the AFM, may be studied as
(Fig. 4b), the AFM data shows a clear single particles.
difference in height and in shape of the About fifteen proteins have been
subunits (Fig. 3d). This difference in the crystallized into two-dimensional arrays on
appearance of the streptavidin tetramer could planar lipid films (Brisson et al., 1994).
be correlated with a surface model derived While some proteins crystallize on a lipid
from the X-ray data. Two subunits (right top monolayer of charged lipids, a specific
and left bottom of the molecule) facing the interaction is an advantage for binding. The
biotinylated lipid monolayer are resolved as use of Nickel-chelating lipids to bind proteins
small protrusions, while the two subunits with a histidine-tag is a promising step to
facing towards the tip and exposing the free establish a general procedure to crystallize
biotin binding sites (left top and right bottom water soluble proteins on lipid monolayers, as
of the molecule) appear as high protrusions demonstrated with a his-tagged reverse
(Fig 3 d). This interesting feature that can be transcriptase (Kubalek et al., 1994).
seen in AFM topographs is outlined in Fig. 5. Combining a general procedure for
The two free biotin pockets of each tetramer crystallization of water soluble proteins on a
on the top surface of the streptavidin crystal lipid monolayer and a general preparation
can be used to bind other biotinylated method for these specimen for the AFM, a
proteins, providing a nanoscale matrix for promising avenue is now available for further
immobilizing proteins. The flatness of the studies on proteins under native conditions.
HOPG surface is essential to minimize
undulations in the crystal, which would also 1.3.6. Acknowledgment
affect the imaging of, bound proteins. The
nature of a protein layer bound to the crystal The work was supported by the Maurice
surface is in principle determined by the size E. Müller foundation of Switzerland, the
and shape of the protein. Proteins of similar Swiss National Foundation for Scientific

Research (grant 4036-44062 to A. Engel), the 2190-2194.
Swiss Priority Project for Micro and Nano Jap, B. K., Zulauf, M., Scheybani, T., Hefti,
System Technology, and the French A., Baumeister, W., Aebi, U., Engel, A.
INSERM. (1992). 2D crystallization: from art to
science. Ultramicroscopy 46, 45-84.
1.3.7. References Karrasch, S., Dolder, M., Schabert, F.,
Ramsden, J. & Engel, A. (1993). covalent
Avila-Sakar, A. J. & Chiu, W. (1996). binding of biological samples to solid
Visualization of ß-sheets and side-chain supports for scanning probe microscopy
clusters in two-dimensional periodic in buffer solution. Biophys. J . 65, 2437-
arrays of streptavidin on phospholipid 2446.
monolayers by electron crystallography.
Biophys. J. 70, 57-68. Kubalek, E. W., Grice, S. F. J. & Brown, P.
O. (1994). Two-dimensional
Bailey, S. W. (1984). Classification and crystallization of histidine-tagged, HIV-1
structures of the micas. Mineralogical reverse transcriptase promoted by a novel
Society of America. nickel-chelating lipid. Journal of Structural
Biology 113, 117-123.
Binnig, G., Quate, C. F. & Gerber, C. (1986).
Atomic force microscope. Phys. Rev. Lett. Mazzola, L. T. & Fodor, S. P. A. (1995).
56, 930-933. Imaging biomolecule arrays by atomic
force microscopy. Biophys. J. 68, 1653-
Brisson, A., Olofsson, A., Ringler, P., 1660.
Schmutz, M. & Stoylova, S. (1994). Two-
dimensional crystallization of proteins on Moore, A. W. (1973). Chemistry and physics
planar lipid films and structure of carbon. Marcel Dekker Inc., New York.
determination by electron crystallography.
Biol. Cell 80, 221-228. Mou, J., Sheng, S., Ho, R. & Shao, Z. (1996).
Chaperonins GroEL and GroES: views
Darst, S. A., Ahlers, M., Meller, P. H., from atomic fore microscopy. Biophys. J.
Kubalek, E. W., Blankenburg, R., Ribi, H. 71, 2213-2221.
O., Ringsdorf, H. & Kornberg, R. D.
(1991). Two-dimensional crystals of Müller, D. J., Amrein, M. & Engel, A. (1997).
streptavidin on biotinylated lipid layers Adsorption of biological molecules to a
and their interactions with biotinylated solid support for scanning probe
macromolecules. Biophys. J. 59, 387-396. microscopy. Journal of Structural Biology
119, 172-188.
Engel, A., Schabert, F., Müller, D. J. & Henn,
C. (1995). Imaging membrane proteins in Müller, D. J. & Engel, A. (1997). The height
their native environment with the atomic of biomolecules measured with the atomic
force microscope. Kluwer, Germany. force microscope depends on electrostatic
interactions. Biophys. J. 73, 1633-1644.
Florin, E.-L., Moy, V. T. & Gaub, H. E.
(1994). Adhesion forces between Müller, D. J., Schabert, F. A., Büldt, G. &
individual ligand-receptor pairs. Science Engel, A. (1995). Imaging purple
264, 415-418. membranes in aqueous solution at
subnanometer resolution by atomic force
Green, M. (1975). Avidin. Adv. Protein chem. microscopy. Biophys. J. 68, 1681-1686.
29, 85-143.
Ohler, M., Baruchel, J., Moore, A. W., Galez,
Hendrickson, W. A., Pähler, A., Smith, J. L., P. & Freund, A. (1997). Direct
Satow, Y., Merritt, E. A. & Phizackerley, observation of mosaic blocks in highly
R. P. (1989). Crystal structure of core oriented pyrolytic graphite. Nuclear
streptavidin determined from Instruments and Methods in Physics
multiwavelength anomalous diffraction of Research B 129, 257-260.
synchrotron radiation. Proceedings of the
National Academy of Sciences, U.S.A. 86, Saxton, W. O. & Baumeister, W. (1982). The

correlation averaging of a regularly Wagner, P., Hegner, M., Kernen, P., Zaugg,
arranged bacterial cell envelope protein. J. F. & Semeza, G. (1996). Covalent
Microsc. 127, 127-138. immobilization of native biomolecules
onto Au(111) via N-hydroxysuccinimide
Saxton, W. O., Pitt, T. J. & Horner, M. ester functionalized selfassembled
(1979). Digital image processing: Semper monolayers for scanning probe
system. Ultramicroscopy 4, 343-354. microscopy. Biophys. J. 70, 2052-2066.

Uzgiris, E. E. & Kornberg, R. D. (1983). Weisenhorn, A. L., Schmitt, F.-J., Knoll, W.
Two-dimensional crystallization technique & Hansma, P. K. (1992). Streptavidin
for imaging macromolecules, with an binding observed with an atomic force
application to antigen-antibody- microscope. Ultramicroscopy 42-44,
complement complexes. Nature 301, 125- 1125-1132.

2. Application of high
resolution AFM

2. Application of high resolution AFM

2.1. High resolution AFM topographs of the Escherichia coli waterchannel
aquaporin Z

Simon Scheuring1, Philippe Ringler 1, Mario Borgnia2, Henning Stahlberg1, Daniel J.
Müller 1, Peter Agre2, Andreas Engel1,3

1M.E.Müller Institute for Structural Biology, Biozentrum, University of Basel, CH-4056 Basel,
2Dept. of Biological Chemistry, John Hopkins University School of Medicine, 725 N Wolfe St.,
Baltimore, MD 21205, USA

2.1.1. Abstract 2.1.2. Introduction

Aquaporins form a large family of Aquaporins are ubiquitous membrane
membrane channels involved in channels in bacteria, fungi, plants and
osmoregulation. Electron crystallography has animals. They are highly specific for water or
shown monomers to consist of six membrane small uncharged hydrophilic solutes and are
spanning a-helices confirming sequence involved in osmoregulation. Hydropathy
based predictions. Surface exposed loops are analysis of the first sequenced members of
the least conserved regions, allowing this family indicated six membrane spans and
differentiation of aquaporins. Atomic force two unusually long loops (Gorin et al., 1984;
microscopy was used to image the surface of Preston & Agre, 1991). Meanwhile more than
aquaporin Z, the water channel of Escherichia one hundred and sixty genes have been
coli. Recombinant protein with an N-terminal sequenced, almost all having the highly
fragment including 10 histidines was isolated conserved NPA motifs within these loops
as tetramer by Ni affinity chromatography, (Heymann & Engel, 1999). The role of the
and reconstituted into two-dimensional unique triplets remains to be established.
crystals with p4212 symmetry. Small Approximately half of these channel proteins
crystalline areas with p4 symmetry were are exclusively water selective and do not
found as well. Imaging both crystal types allow the permeation of small or charged
before and after cleavage of the N-termini, solutes. Other channels facilitate the passage
allowed the cytoplasmic surface to be of small hydrophilic molecules such as
identified; a drastic change of the cytoplasmic glycerol or urea (Ishibashi et al., 1997a).
surface accompanied proteolytic cleavage, Passage of ions through waterchannel
while the extracellular surface morphology proteins have been reported, and recent data
did not change. Flexibility mapping and suggest this to be regulated by pH (Yasui et
volume calculations identified the longest al., 1999).
loop at the extracellular surface. This loop The best characterized waterchannel is
exhibited a reversible force induced aquaporin-1 (AQP1) from human red blood
conformational change. cells (Agre et al., 1993). Three dimensional
(3D) density maps of this protein have been
established to a resolution of 6 Å by three

frame of six a-helices that houses the two
NPA motif carrying loops which fold back
into the membrane to form a highly specific
channel. The water flow per channel was
found to be the same in two dimensional (2D)
crystals of AQP1 as in erythrocyte ghosts, 3
x 109 water molecules per channel and
second (Zeidel et al., 1992; Walz et al.,
In Escherichia coli a waterchannel has
been identified by homology cloning
(Calamita et al., 1995). Sequence analysis of
this bacterial channel, AqpZ, revealed a
significant homology to AQP1. Recombinant
AqpZ bearing a histidine tag has been
produced and shown to be active (Borgnia et
al., 1999). 2D crystals with sizes ranging up
to 5µm have been assembled from this
recombinant AqpZ tetramers by dialysis of a
protein-lipid-detergent mixture (Ringler et al.,
1999). The 3D map of negatively stained
preparations revealed the same packing
arrangement as in AQP1 2D crystals, p4212,
while the 8 Å projection map from vitrified
unstained preparations showed a striking
similarity to the AQP1 and the major intrinsic
protein (MIP) projection maps (Ringler et al.,
Figure 1 . a) AFM topograph of AqpZ 2D
crystals adsorbed to mica (recorded in buffer 1999; Walz et al., 1995; Hasler et al., 1998).
solution: 17 mM Tris-HCl, pH 7.2, 150 mM We have used the atomic force microscope
KCl). Double and multi layered areas can clearly be
distinguished from single layer crystals by their (AFM) (Binnig et al., 1986) to measure the
higher appearance (scale bar: 2 µm; full gray scale: surface topography of AqpZ crystals in a
30 nm). b) Section analysis along the white line in
image a). The 2D crystals show a uniform height of native environment. As previously reported,
57 ± 4 Å (n = 45). Double layered areas appear as this method allows protein surfaces to be
plateaus twice as high as the single layered crystal
sheets (vertical scale bar: 50 Å). c) AFM topograph imaged at subnanometer resolution (Schabert
of reconstituted AqpZ (recorded in buffer solution: et al., 1995; Müller et al., 1995b; Mou et al.,
17 mM Tris-HCl, pH 7.2, 150 mM KCl,). A
densely packed vesicle containing crystalline areas 1996; Scheuring et al., 1999). Topographs of
with p4 symmetry adsorbed onto a crystal sheet AqpZ crystals bearing an N-terminal poly
with p4212 symmetry (scale bar: 100 nm; full gray
scale: 20 nm). Top inset: average of the sheet with histidine tag to allow rapid isolation exhibited
p4212 symmetry (scale bar: 50 Å);. Bottom inset: a floppy protrusion related to the N-terminus
average of the crystalline areas with p4 symmetry
within the densely packed vesicle (scale bar: 50 Å). that could be eliminated by proteolysis. This
allowed the sidedness of AqpZ to be
groups (Walz et al. 1997; Cheng et al., 1997; identified. After cleavage of the histidine tag,
Li et al., 1997). These maps show a bundle of topographs had a lateral resolution of 7 Å and
six highly tilted transmembrane helices that a vertical resolution of 1 Å. The surface
surround a central density formed by the long topography could be related to the loops
loops. These results confirm the hourglass predicted from the sequence by hydropathy
model of Jung et al. (1994), who proposed a analysis.

F i g u r e 2 . a) Aminoacid sequence model of AqpZ showing the six membrane spanning helices derived from
hydropathy analysis. Trypsin cleavage sites are located on arg26 and arg230. b) Silver stained SDS - polyacrylamide
10% (w/v) gel. Columns from left to right: (M) marker: 97.4 kDa, 66.2 kDa, 42.7 kDa, 31.0 kDa; (1) AqpZ
solubilized in 2% OG; (2) AqpZ crystals after overnight trypsin treatment. The AqpZ band is broadened indicating
that a minor part of the protein remained undigested. The two diffuse bands in the low molecular weight region
(below 30 kDa) correspond to trypsin and the cleaved N-terminal fragments; (3) AqpZ-10his crystals. The faint bands
at high molecular weight (~ 200 kDa) in all three lanes arise from specific aggregates.

2.1.3. Results crystalline areas comprising about 30
tetramers arranged with p4 symmetry (Fig.
In the presence of 25 mM MgCl2 the 2D 1c, main frame, bottom) The p4 crystals had
crystals of AqpZ adsorbed to mica without unit cell dimensions of a = b = 72 ± 2 Å and
wrinkles and were thus suitable for high- housed a single tetramer (Fig. 1c, bottom
resolution imaging. To this end, the right).
adsorption buffer was exchanged with the The recombinant AqpZ crystallized has an
recording buffer that was adjusted to N-terminal fragment of 26 aminoacids,
electrostatically balance the van der Waals containing a trypsin cleavage site at arg26 and
forces (Müller, 1999a). The overview in Fig. a 10 his-tag at aminoacid positions 2 to 12
1a demonstrates the flatness of the 2D (Fig. 2a; Borgnia et al., 1999). As a
crystals whose thickness was found to be 57 consequence, a total of 104 aminoacids,
± 4 Å (n = 45; Fig. 1b). including 40 histidines, protruded from the
At higher magnification the square lattice cytoplasmic side of each tetramer. These
became distinct (Fig. 1c, main frame, top). peptides produced a strong signal in the
The unit cell dimensions were a = b = 95 ± 2 AFM, resulting in a 20 ± 2 Å high peak (Fig.
Å, in excellent agreement with results from 1c, 3a), the exact position and appearance of
electron microscopy (Ringler et al., 1999). which depended critically on the force applied
Correlation averaging revealed a unit cell to the stylus, the scan speed and the direction
housing two tetramers in opposite of the scan (compare Fig. 1c and 3a). This
orientations with respect to the membrane extreme flexibility prevented the resolution of
plane (p4212 packing; Fig. 1c, top right). The substructure. To prove that the large
high signal-to-noise ratio of the AFM also protrusion observed indeed arose from the N-
allowed high resolution imaging on densely terminal domain, crystals were treated with
packed vesicles which only exhibited small trypsin (see materials & methods) to cleave

off this flexible end domain as documented in identify the flexible regions of the
Fig. 2b. The digested crystals exhibited a extracellular surface (Müller et al., 1998). As
striking change; instead of an ill defined displayed in Fig. 4f, one region exhibited a
protrusion of 20 Å height, the cytoplasmic pronounced SD, while the rest yielded highly
side now showed four distinct protrusions reproducible heights (SD ≤ 0.2 Å).
each with a height of 3.5 ± 0.4 Å above the Interestingly, this flexible region also
lipid bilayer (Fig. 3b). In contrast, the exhibited the major force induced
extracellular side was neither changed in conformational change (compare left and
shape nor height by the trypsin treatment right tetramer in Fig. 4f).
(Fig. 3c; see Table 1). Sequence based structure prediction
The extracellular side was not sensitive to postulates AqpZ to be a protein consisting of
trypsin digestion, however it underwent a six transmembrane helices connected by three
reversible conformational change when the loops on the extracellular side and two loops
force applied to the tip was increased during on the cytoplasmic side, as well as two
imaging. At minimal force (~ 80 pN) each cytoplasmic termini (see Fig. 2a). In
AqpZ subunit showed three major agreement with this, on imaging at minimal
protrusions, probably related to the loops force, three protrusions were found on the
connecting the membrane spanning helices on extracellular surface of the AqpZ monomer,
the extracellular side. (Fig. 4a, c). Recording a one close to the 4-fold symmetry center, and
second image of the same areas with a force one small and one elongated protrusion at the
increased by + 80 pN, a drastic periphery (Fig. 5a). Volume calculations (see
conformational change was observed materials and methods) on the protrusion
(compare figures 4a and b, and 4c and d). The close to the 4-fold symmetry center resulted
extracellular AqpZ surface reversibly changed in 1278 ± 150 Å3. The small peripheral
its rather circular appearance into a left- protrusion yielded a volume of 984± 134 Å 3,
handed windmill, which still protruded out of while the elongated protrusion had a volume
the membrane by 7 Å (Fig. 4e). This force of 3187 ± 528 Å 3, the larger SD reflecting
induced conformational change was not the flexibility of this region. The single
influenced by the trypsin treatment, as protrusion observed per monomer on the
illustrated by comparing the topographs of digested cytoplasmic surface had a volume of
digested (Fig. 4c and d) and undigested AqpZ 1222 ± 144 Å3 (Fig. 5b). Since the termini
crystals (Fig. 4a and b). The digested are removed (Fig. 2a, b), this protrusion is
cytoplasmic surface did not show the same expected to house loops b and d.
force dependence, the minor force induced
conformational change was barely noticeable 2.1.4. Discussion
(Figs. 4c and d, insets). The standard
deviation (SD) map of 289 densely packed Significant progress in the understanding
single tetramers (such as shown in Fig. 4a) of imaging conditions and the interpretation
recorded at minimal force was calculated to of topographs recorded with the atomic force

Table 1. Heights of surface protrusions of AqpZ before and after trypsin treatment.

undigested undigested digested
p4 crystal p4212 crystal p4212 crystal
Extracellular 7.3 ± 0.9 Å 6.7 ± 1.0 Å 7.0 ± 0.9 Å
Cytoplasmic 18.6 ± 1.8 Å 20 ± 2.0 Å 3.5 ± 0.4 Å

microscope (AFM) has allowed the surface
topography of bacteriorhodopsin to be
correlated with the helix connecting loops to a
lateral resolution of 5 Å (Müller et al.,
1999b). Here we have used this technology to
study the surface of AqpZ, the first bacterial
waterchannel identified (Calamita et al.,
1995). Its overexpression, isolation and 2D
crystallization have recently been described
(Borgnia et al., 1999; Ringler et al., 1999).
2D crystals adsorbed firmly and without
folds or wrinkles to freshly cleaved mica in a
high ionic strength buffer (Müller et al.,
1997). Subsequent change to a buffer
adjusted to compensate for van der Waals
interactions allowed their height to be
measured accurately (Müller & Engel 97).
The result, 57 ± 4 Å, compares favorably with
the height previously reported for AQP1, 58
± 3 Å (Walz et al., 1996).
The p4212 crystals of AqpZ with unit cell
dimensions of a = b = 95 ± 2 Å have similar
lattice parameters to those found for AQP1
(unit cell dimensions: 96 ± 2 Å) (Walz et al.,
1996). However, the p4 crystals of AqpZ
(unit cell dimensions of a = b = 72 ± 2 Å) are
more loosely packed than 2D crystals of MIP
which also exhibit p4 symmetry (unit cell
dimensions: a = b = 64 ± 1 Å; Hasler et al.,
1998). This suggests that more lipid
molecules are interspersed between the AqpZ
tetramers within the small crystalline areas in
the densely packed vesicles than in the highly
ordered MIP crystals.
Figure 3 . a) AFM topograph (recorded in buffer In experiments with undigested AqpZ
solution: 17 mM Tris-HCl, pH 7.2, 150 mM KCl) of crystals the N-terminal tail of 26 aminoacids
an AqpZ-10his 2D crystal with p4212 symmetry
recorded using minimal force (scale bar: 500 Å; full prevented the visualization of substructures
gray scale: 30 Å). Inset: relief view (tilt: 85°) of a 300 on the cytoplasmic surface. The four weakly
Å square, raw data; extracellular surfaces are marked by
circles. b) AFM topograph (recorded in buffer solution: ordered protruding peptides, each containing
17 mM Tris-HCl, pH 7.2, 150 mM KCl) of an AqpZ 10 histidines, have a total mass of ~12kDa,
2D crystal with p4212 symmetry after trypsin
treatment recorded using minimal force (scale bar: 500 and appeared to interact strongly with the
Å; full gray scale: 10 Å). Inset: relief view (tilt: 85°) of silicon nitride tip. Consequently the
a 300 Å square, raw data; extracellular surfaces are
marked by circles. c) 3D reconstruction of the trypsin cytoplasmic surfaces exhibited peaks of
cleavage process (see Fig. 2a) observed on the approximately 20 Å height that were
cytoplasmic surface. On digestion this surface changes
shape and height drastically in the location of the N- influenced by the scan direction (Fig. 1c, 3a).
terminal his-tags (right top: undigested state, left In spite of these large protrusions, the
bottom: digested state). intervening extracellular sides could be
microscope (AFM) has allowed the surface
resolved as tetramers with a central hole and After trypsin treatment both the
four major protrusions. extracellular and the cytoplasmic surfaces

Figure 4. a) AFM topograph of AqpZ-10his densely packed in a vesicle, scanned in buffer solution (17 mM Tris-
HCl, pH 7.2, 150 mM KCl) at minimal force (~ 80 pN) (scale bar: 250 Å; full gray scale: 18 Å). A small
crystalline area with p4 symmetry is outlined. Inset: average of a). The white square indicates the unit cell (a = b =
72 ± 2 Å) which houses 1 tetramer. b) AFM topograph of the same area as a) recorded in buffer solution (17 mM
Tris-HCl, pH 7.2, 150 mM KCl) applying an additional force of +80 pN to the tip during scanning. The outlined
area corresponds to the area marked in a) (scale bar: 250 Å; full gray scale: 18 Å). Inset: average of c). The white
square indicates the unit cell (a = b = 72 ± 2 Å) which houses 1 tetramer. c) AFM topograph of trypsin treated AqpZ
2D crystals with p4212 symmetry recorded in buffer solution (10 mM Tris-HCl, pH 7.5, 150 mM KCl) using
minimal force (~ 80 pN) (scale bar: 250 Å; full gray scale: 20 Å). The outlined area shows a pronounced lattice
distortion. Inset: average of c). The white square indicates the unit cell (a = b = 95 ± 2 Å) which houses 2 tetramers.
d) AFM topograph of the same area as c) recorded in buffer solution (10 mM Tris-HCl, pH 7.5, 150 mM KCl)
applying an additional force of +80 pN to the tip during scanning. Note the same lattice irregularity as in c) (scale
bar: 250 Å; full gray scale: 20 Å). Inset: average of d). The white square indicates the unit cell (a = b = 95 ± 2 Å)
which houses 2 tetramers. e) 3D reconstruction illustrating the effect observed on the extracellular surface when the
imaging force is increased by + 80 pN during scanning. (right top: minimal force, left bottom: + 80 pN). f)
Comparison of the averages of the extracellular surface at minimal force (left), the standard deviation (SD) map
(middle), and the extracellular surface average at a additional force of + 80 pN (right) (full image sizes: 72 Å). The
outlined regions in the in the middle image represent a SD ~ 0.7 Å. These regions correspond to the four elongated
peripheral protrusions in the minimal force average which are strongly displaced in the average gained from images
recorded at + 80 pN.

could be imaged at high resolution on p4212 The reliability of topographic data acquired
crystals (Fig. 3b and 4c). Not only did the has been documented by direct comparison
extracellular side show a consistent surface with results from electron microscopy
appearance before and after trypsin treatment (Karrasch et al., 1994) and X-ray
(compare central particles in the insets of Fig. crystallography (Schabert et al., 1995). In
4a and c), but also its height over the lipid addition, a force induced reversible
bilayer was not affected by the proteolytic conformational change observed on the
cleavage (Table 1). In contrast, only a small cytoplasmic surface of bacteriorhodopsin has
cytoplasmic protrusion remained after trypsin indicated the location of the longest
digestion; the surface appearance changed cytoplasmic loop (Müller et al., 1995a).
from a single large protrusion to a well Finally, structural features having the
defined tetramer protruding by only 3.5 Å strongest variability in the raw data are
over the lipid bilayer (Fig. 3c). Thus removal enhanced in standard deviation maps allowing
of the large flexible cytoplasmic domains by flexible regions of proteins to be identified
trypsin allowed the unambiguous assignment (Müller et al., 1998). All three methods were
of the AqpZ sidedness. This elegant method applied to assess the surface topography of
will be useful to determine the sidedness of AqpZ. Firstly, we have identified two
other recombinant membrane proteins with approximately equal, small protrusions (984
the AFM. Å3, 1278 Å3), and one large, elongated
The visualization and identification of protrusion (3187 Å3) at the extracellular
distinct loops on a protein surface with the surface of AqpZ (Fig. 5). Secondly by
AFM has been shown for bacteriorhodopsin applying an additional force of ~ 80 pN to the
(Müller et al., 1995b; Müller et al., 1999b). tip during scanning, the large protrusion was

Figure 5. Proposed assignment and borderlines between adjacent loops of the AqpZ tetramer.
The x, y and z scalings used for three dimensional integration of the protruding volumes are derived from high
resolution AFM topographs (full image sizes: 95 Å). a) Extracellular surface exposing protrusions 1, 2, and C with
volumes of 984 ± 134 Å3, 1278 ± 150 Å 3, and 3187 ± 528 Å3 respectively. Protrusions 1 and 2 correspond to
loops A and E. The similarity of their volumes prevents an unambiguous assignment. b) Cytoplasmic surface
having only one defined protrusion housing loops B and D with a volume of 1222 ± 144 Å3.

found to undergo a drastic conformational physiological conditions, function related
change (Fig. 4). Because this change was structural changes can be directly assessed
completely reversible, the same area could be (Müller & Engel, 1999). Such experiments
scanned many times, and the change will be relevant to study the recently
monitored repeatedly. Thirdly, the SD map discovered regulation of water channels by
calculated from 289 AqpZ tetramers indicated pH (see Engel et al., accompanying paper).
that the large protrusion was the most flexible
one (Fig. 4). Taken together, this suggests 2.1.5. Materials and methods
this large surface domain located at the
periphery of the tetramer with a total volume Reconstitution
of about 3200 Å3 to be related to loop C of
AqpZ, predicted to comprise 26 aminoacids Large 2D crystals were produced by
(3700 Å3) (Fig. 2a). dialysis as described by Ringler et al. (1999).
In spite of the limitations related to the tip Recombinant AqpZ isolated by Ni-affinity
geometry, an estimate of the volume of chromatography (Borgnia et al., 1999) was
contoured protrusions is possible (Fritzsche solubilized in 2% n-octyl-b-D-glucoside
& Henderson, 1996; Schneider S. W. et al., (OG) at a concentration of 0.5 mg/ml and
1998). We have tested the algorithms used to mixed with dimyristoyl phosphatidyl
delineate protrusions and calculate their choline/palmitoyl oleoyl phosphatidyl choline
volumes on surface topographs of (DMPC/POPC) (1/1) solubilized in 2% OG
bacteriorhodopsin (Müller et al., 1999b) and to a final lipid-to-protein ratio of 0.3. The
porin (Schabert et al., 1995, Müller & Engel, mixture was dialyzed against a detergent free
1999), and found them to produce correct buffer (20 mM citric acid, pH 6.0, 200 mM
volumes within an experimental error of < 20 NaCl, 100 mM MgCl2, 3 mM NaN3, 10%
%. The volumes accordingly calculated for glycerol) for three days. 2D crystals were
the AqpZ extracellular protrusions are close washed by centrifugation and resuspended in
to those expected from the sequence predicted adsorption buffer (see below).
loops, although the small loops A (predicted
as 6 aminoacids, 900 Å3) and E (9 Trypsin digestion
aminoacids, 1300Å3) cannot be
unambiguously assigned. However, the For cleavage of the N-terminal fragment,
position of the two loops remaining on the AqpZ-10his crystals were incubated overnight
cytoplasmic surface after digestion could be at 4oC with trypsin (1 mg/ml). The crystals
defined. Importantly, the unambiguous were then washed twice through
determination of sidedness achieved using the centrifugation in a table centrifuge (Heraeus
AFM together with digestion experiments, is Biofuge A) at 5000 rpm for 3 minutes with
essential for the interpretation of structural subsequent removal of the supernatant and
data obtained from AqpZ crystals by cryo addition of fresh buffer solution. After
electron microscopy. trypsin treatment samples were investigated
In conclusion, structural information on by sodium dodecyl sulfate-polyacrylamide
the surface exposed loops of a membrane gel electrophoresis (SDS-PAGE) using 10%
channel has been acquired with the AFM. (w/v) acrylamide gels.
These data are complementary to those
obtained by electron crystallography which Atomic force microscopy
provides mainly the 3D density map of the
membrane resident part of the protein. Mica prepared as described (Schabert &
Because the AFM is operated under Engel, 1994) was used as support. For each

experiment the mica was freshly cleaved with the AFM.
scotch tape and imaged in 30-50µl of
adsorption buffer (10 mM Tris-HCl, pH 7.5, 2.1.6. Acknowledgment
150 mM KCl, 25 mM MgCl2) to check the
cleavage quality, 3µl of protein crystal The authors would like to thank Dr. S. A.
solution (0.1 mg/ml) were then injected into Müller for proofreading and discussing the
the adsorption buffer drop on the mica manuscript. We also thank D. Fotiadis and L.
surface. After 2h the sample was carefully Hasler for inspiring discussions. The work
rinsed with recording buffer (10 mM Tris- was supported by the Maurice E. Müller
HCl, pH 7.5, 150 mM KCl). The recording foundation of Switzerland, the Swiss National
buffer was optimized to achieve high Foundation for Scientific Research (grant
resolution as described (Müller et al., 1999a). 4036-44062 to A. Engel), the Swiss Priority
Imaging was performed with a commercial Project for Micro and Nano System
Nanoscope III AFM (from Digital Technology, and the French INSERM (to P.
Instruments, Santa Barbara, CA, USA) R.).
equipped with a 120 µm scanner (J-scanner)
and oxide-sharpened Si3N4 cantilevers with a 2.1.7. References
length of 120µm (k=0.1 N/m) (from
Olympus Ltd., Tokyo, Japan). The AFM was Agre, P., Preston, G., Smith, B., Jung, J.,
Raina, S., Moon, C., Guggino, W. &
operated in contact mode applying constantly Nielsen, S. (1993). Aquaporin CHIP: the
minimal forces (<0.2 nN) at a scan frequency archetypal molecular water channel. Am. J.
of 4-6 Hz. The instrument was calibrated Physiol. 265, F436-476.
using layered crystals of MoTe2 as described
Binnig, G., Quate, C. F. & Gerber, C. (1986).
previously (Müller & Engel, 1997). Atomic force microscope. Phys. Rev. Let.
56, 930-933. Image processing
Borgnia, M., Kozono, D., Calamita, G.,
Nielsen, S., Maloney, P. C. & Agre, P.
AFM images of crystalline areas were (1999). Functional reconstitution and
processed by correlation averaging using the characterization of E. coli Aquaporin-Z. J.
SEMPER image processing system (Saxton Mol. Biol., in press.
et al., 1979). Tetramers of disordered regions
Calamita, G., Bishai, W., Preston, G.,
were aligned and averaged by a single particle Guggino, W. & Agre, P. (1995).
averaging protocol, and standard deviation Molecular cloning and characterization of
maps were calculated to assess the AqpZ, a waterchannel from Escherichia
reproducibility of height measurements and to coli. J. Biol. Chem. 270, 29063-29066.
Cheng, A., van Hoek, A. N., Yeager, M.,
identify flexible regions (Müller et al., 1998). Verkman, A. S. & Mitra, A. K. (1997).
To calculate the volumes of distinct Three-dimensional organization of a
protrusions in such an average, a SEMPER human water channel. Nature 387, 627-
routine was used to determine the borderlines 630.
Engel, A., Fujiyoshi, Y. & Agre, P. (1999).
between adjacent elevations. To this end, a 3 x The importance of aquaporin water
3 pixel box was scanned over the image and channel protein structures. The EMBO
all the central pixels having a lower gray value Journal, this issue
than any of the 4 pairs of adjacent pixels were
Fritzsche, W. & Henderson, E. (1996).
identified as borderline. Outlined areas were Volume determination of human
evaluated by integrating all the pixels metaphase chromosomes by scanning
determined within this boundary using unit force microscopy. Scan. Micr. 10, 103-
cell dimensions and heights determined with 108

Gorin, M. B., Yancey, S. B., Cline, J., Revel, bacteriorhodopsin. J. Mol. Biol. 249, 239-
J.-P. & Horwitz, J. (1984). The major 243.
intrinsic protein (MIP) of the bovine lens
fiber membrane: Characterization and Müller, D. J., Schabert, F. A., Büldt, G. &
structure based on cDNA cloning. Cell Engel, A. (1995b). Imaging Purple
39, 49-59. Membranes in aqueous solution at
subnanometer resolution by atomic force
Hasler, L., Walz, T., Tittmann, P., Gross, H., microscopy. Biophys. J. 68, 1681-1686.
Kistler, J. & Engel, A. (1998). Purified
lens major intrinsic protein (MIP) forms Müller, D. J. & Engel, A. (1997). The height
highly ordered tetragonal two-dimensional of biomolecules measured with the atomic
arrays by reconstitution. J. Mol. Biol. 297, force microscope depends on electrostatic
855-864. interactions. Biophys. J. 73, 1633-1644.

Heymann, J. B. & Engel, A. (1999). Müller, D. J. & Engel, A. (1999). pH and
Aquaporins: phylogeny, structure and voltage induced structural changes of
physiology of water channels. News in porin OmpF explain channel closure. J.
Physiological Sciences, in press . Mol. Biol. 285, 1347-1351.

Ishibashi, K., Kuwahara, M., Gu, Y., Müller, D. J., Fotiadis, D. & Engel, A.
Kageyama, Y., Tohsaka, A., Suzuki, F., (1998). Mapping flexible protein domains
Maruma, F. & Sasaki, S. (1997a). Cloning at subnanometer resolution with the AFM.
and functional expression of a new water FEBS Lett. 430, 105-111.
channel abundantly expressed in the testis
permeable to water, glycerol and urea. J. Müller, D. J., Fotiadis, D., Scheuring, S.,
Biol. Chem. 272, 20782-20786. Müller, S. A. & Engel, A. (1999).
Electrostatically balanced subnanometer
Jung, J., Preston, G., Smith, B., Guggino, W. imaging of biological specimens by atomic
& Agre, P. (1994). Molecular structure of force microscopy. Biophys. J. 76, 1101-
the water channel through aquaporin 1111.
CHIP. The hourglass model. J. Biol.
Chem. 269, 14648-14654. Müller, D. J., Sass, H.-J., Müller, S., Büldt, G.
& Engel, A. (1999). Surface structures of
Karrasch, S., Hegerl, R., Hoh, J. H., native bacteriorhodopsin depend on the
Baumeister, W. & Engel, A. (1994). molecular packing arrangement in the
Atomic Force Microscopy Produces membrane. J. Mol. Biol. 285, 1903-1909.
Faithful High-Resolution Images of
Protein Surfaces in an Aqueous Preston, G. M. & Agre, P. (1991). Isolation
Environment. Proc Natl. Cad Sci USA 91, of the cDNA for erythrocyte integral
836-838. membrane protein of 28 kilodaltons:
Member of an ancient channel family.
Li, H., Lee, S. & Jap, B. K. (1997). Molecular Proc. Natl. Acad. Sci. USA 88, 11110-
design of aquaporin-1 water channel as 11114.
revealed by electron crystallography. Nat.
Struct. Biol. 4, 263-265. Ringler, P., Borgnia, M., Stahlberg, H., Agre,
P. & Engel, A. (1999). Structure of the
Mou, J., Czajkowsky, D. M., Sheng, S., Ho, water channel AqpZ from Escherichia coli
R. & Shao, Z. (1996). High resolution revealed by electron crystallography. J.
surface structure of E. coli GroES Mol. Biol., in press.
oligomer by atomic force microscopy.
FEBS Lett. 381, 161-164. Saxton, W. O., Pitt, T. J. & Horner, M.
(1979). Digital Image Processing: Semper
Müller, D. J., Amrein, M. & Engel, A. (1997). System. Ultramicr. 4, 343-354.
Adsorption of biological molecules to a
solid support for scanning probe Schabert, F. A. & Engel, A. (1994).
microscopy. J. Struct. Biol. 119, 172-188. Reproducible acquisition of Escherichia
coli porin surface topographs by atomic
Müller, D. J., Büldt, G. & Engel, A. (1995a). force microscopy. Biophys. J. 67, 2394-
Force-induced conformational change of 2403.

Schabert, F. A., Henn, C. & Engel, A. (1995). dimensional crystals of aquaporin CHIP.
Native Escherichia coli OmpF porin J. Biol. Chem. 269, 1583-1586.
surfaces probed by atomic force
microscopy. Science 268, 92-94. Walz, T., Tittmann, P., Fuchs, K. H., Müller,
D. J., Smith, B. L., Agre, P., Gross, H. &
Scheuring, S., Müller, D. J., Ringler, P., Engel, A. (1996). Surface topographies at
Heymann, J. B. & Engel, A. (1999). subnanometer resolution reveal asymmetry
Imaging streptavidin 2D crystals on and sidedness of aquaporin-1. J. Mol.
biotinylated lipid monolayer at high Biol. 264, 907-918.
resolution with the atomic force
microscope. J. Microsc. 193, 28-35. Walz, T., Typke, D., Smith, B. L., Agre, P. &
Engel, A. (1995). Projection map of
Schneider, S. W., Larmer, J., Henderson, R. aquaporin-1 determined by electron
M. & Oberleithner, H. (1998). Molecular crystallography. Nat. Struct. Biol. 2, 730-
weights of individual proteins correlate 732.
with molecular volumes measured by
atomic force microscopy. Pflugers Arch. Yasui, M., Kwon, T. H., Knepper, M. A.,
435, 362-367. Nielsen, S. and Agre, P. (1999).
Aquaporin-6: An intracellular vesicle water
Walz, T., Hirai, T., Murata, K., Heymann, J. channel protein in renal epithelia. Proc.
B., Mitsuoka, A., Fujiyoshi, Y., Smith, B. Nat. Ac. Sc., U.S.A. 96, 5808-5813.
L., Agre, P. & Engel, A. (1997). The 6Å
three-dimensional structure of aquaporin- Zeidel, M. L., Ambudkar, S. V., Smith, B. L.
1. Nature 387, 624-627. & Agre, P. (1992). Reconstitution of
functional water channels in liposomes
Walz, T., Smith, B., Zeidel, M., Engel, A. & containing purified red cell CHIP28
Agre, P. (1994b). Biologically active two- protein. Biochem. 31, 7436-7440.

2.2. High resolution AFM topographs of Rubrivivax gelatinosus light-
harvesting complex LH2

Simon Scheuring1, Francoise Reiss-Husson2, Andreas Engel1, Jean-Louis Rigaud3 and
Jean-Luc Ranck 3

1M.E. Müller Institute for Microscopy at the Biozentrum, University of Basel, Klingelbergstrasse
70, CH-4056 Basel, Switzerland.
2Centre génétique Moléculaire, UPR-CNRS 2167, 91198 Gif-sur-Yvette Cedex, France
3Institut Curie, UMR-CNRS 168 and LRC-CEA 8, 11 rue Pierre et Marie Curie, 75231 Paris
Cedex 05, France

2.2.1. Abstract assembly and organization of photosynthetic
systems. Absorption of light and its
Light harvesting complexes 2 (LH2) are conversion into chemical energy is performed
antenna proteins in the bacterial by highly organized transmembrane pigment-
photosynthetic apparatus and are built of αβ- protein complexes: the light harvesting
heterodimers containing 3 complexes 2 (LH2) and 1 (LH1), and the
bacteriochlorophylls and 1 carotenoid each. reaction center (RC). In addition to the wealth
We have used atomic force microscopy of information from biochemistry, molecular
(AFM) to investigate reconstituted LH2 from biology and spectroscopy, structural data have
Rubrivivax (Rvi.) gelatinosus, which has a C- advanced our understanding of the single
terminal hydrophobic extension of 21 amino components of the bacterial photosynthetic
acids residues on the α-subunit. High apparatus and have provided a model for its
resolution topographs revealed a nonameric functional mechanism. According to this
organization of the regularly packed model, the light energy is initially trapped by
complexes incorporated into the membrane in the peripheral antenna LH2 complexes and
both orientations. Native LH2 showed one the excitation energy is transferred to LH1
surface which protruded by ~ 14 Å and one complexes which are closely associated with
which protruded by ~ 5 Å from the the RC, forming the so-called core complex.
membrane. Thermolysin cleaved protein with The subsequent photon induced redox
a shorter C-terminus of the α-subunits had a reaction in the RC causes charge separation
height of ~ 9 Å and a different appearance of across the membrane which in turn initiates a
protruding structures, allowing the cyclic electron transport and the formation of
assignment of the periplasmic surface. Minor an electrochemical proton gradient across the
contaminants were imaged as rare large rings membrane (Papiz et al., 1996).
(~ 120 Å diameter) in interaction with LH2. Different types of LH antenna complexes
Their diameter and rotational power spectra have been isolated from various species of
suggested these rings to be hexadecameric purple bacteria and their structures solved at
LH1 complexes. high resolution. The basic motif is a
heterodimer consisting of two small
2.2.2. Introduction polypeptides α and β, which both span the
membrane once as transmembrane α-helices.
Purple photosynthetic bacteria provide an Together they form a heterodimer to which
ideal experimental system for describing the three bacteriochlorophyll a pigments

molecules (Bchl) and one carotenoid are non sphaeroides, both exhibiting a nonameric
covalently bound. Ring-like associations of organization (Montoya et al., 1995; Walz et
the αβ-heterodimers have been reported from al., 1998). The structure of the LH1 complex,
structural analysis of three-dimensional (3D) however, is still unknown at atomic
and two-dimensional (2D) crystals. resolution. An electron crystallographic
McDermott et al. (1995) first solved the projection structure of Rhodospirillum (Rs.)
structure of LH2 from Rhodopseudomonas Rubrum LH1 at 8.5 Å showed a similar ring-
(Rps.) acidophila by X-ray crystallography like structure, in this case consisting of 16
showing an αβ-nonamer shaped as a hollow αβ-heterodimers (Karrasch et al., 1995). The
cylinder, formed by the membrane spanning diameter of the LH1 complex suffices to
helices of the 2 subunits, with the α-subunit accommodate a reaction center within the
inside and β outside. Atomic structures from ring, a model confirmed by analyses of 2D
3D crystals of Rhodospirillum (Rp.) crystals of LH1-RC core complexes from
molischianum LH2 revealed an octameric Rhb. sphaeroides (Stahlberg et al., 1998;
arrangement (Koepke et al., 1996). Electron Walz & Ghosh, 1997).
crystallographic data were also acquired on In spite of the wealth of information
2D crystals of LH2's from Rhodovulum available on the individual components of the
(Rhv.) sulfidophilum and Rhodobacter (Rhb.) photosynthetic apparatus of photosynthetic

Figure 1. a) Topology model of Rvi. gelatinosus LH2 (strain S1; gene sequence accession number AF312921).
The α- (consisting of 71 aminoacids) and the β-polypeptide (consisting of 51 aminoacids) cross the membrane once
each. The boxed regions in the sequence correspond to a-helical stretches in the Rps acidophila LH2 structure,
aminoacids surrounded by circles in dark gray are predicted to be transmembrane (using TMpred5), those surrounded
by circles in light gray are predicted to be α-helical (using GOR4 6). The triangle indicates the thermolysin cleavage
site on the C-terminus of the a-polypeptide. The arrow points towards the center of the nonamer in the membrane
plane. b) Silver stained SDS-polyacrylamide 10 % (w/v) gel. Columns from left to right: (1) marker: 97.4 kDa,
66.2 kDa, 42.7 kDa, 31.0 kDa; (2) Thermolysin (37 kDa), (3) Crystals of native LH2, band at 115 kDa (4),
Crystals of thermolysin treated LH2, band at 82 kDa. c) Absorption spectra of native and thermolysin cleaved LH2
reconstituted into lipid bilayers. The bacteriochlorophyll and carotenoid absorption spectra do not change upon
cleavage of the C-terminus of the a-subunit (native LH2: black line, digested LH2: gray line). The absorption spectra
document the native state of the protein: arrows 1, 2, 3: carotenoid peaks at 460 nm, 488 nm and 517 nm
respectively; arrow 4: Qx peak at 595 nm; arrow 5: Qy peak at 802 nm; arrow 6: Qy peak at 859 nm.

bacteria, their supramolecular organization in extension could not be identified.
the membrane is poorly understood. Models We have used the atomic force microscope
of interaction between LH2, LH1 and the RC (AFM) (Binnig et al., 1986) to measure the
have been established to explain the highly surface topography of LH2 of Rvi.
efficient energy transfer of a photon trapped gelatinosus in a native environment and to
by an LH2 antenna complex to LH1 and locate the C-terminal extension. As previously
finally the RC as reviewed by Kuhlbrandt reported, this method allows protein surfaces
(1995). The structural model for to be imaged at subnanometer resolution
understanding light capture and transfer is a (Schabert et al., 1995; Müller et al., 1995;
close packing of ring-like structures with Scheuring et al., 1999; Fotiadis et al., 2000;
LH2 complexes surrounding the LH1/RC Engel & Müller, 2000). Topographs of LH2
core complex. However, the size of the rings complexes had a lateral resolution of ~ 8 Å
of the different LH’s complexes as well as an and a vertical resolution of ~ 1 Å. A
open organization of LH1 complexes around nonameric organization of the αβ-
RCs allowing an efficient quinone/quinol heterodimers with one strongly (~ 14 Å) and
transfer is still under debate (Jungas et al., one weakly (~ 5 Å) membrane protruding
1999; Francia et al., 1999; Loach, 2000; Frese surface was found. Imaging membranes
et al., 2000). reconstituted with native and digested LH2
Here we present a structural study of the allowed localization of the C-terminal
LH2 complex from Rvi. gelatinosus. As aminoacids of the α-subunit of Rvi.
compared to LH2's from other species, the gelatinosus and consequently the assignment
most distinct feature is related to the sequence of the periplasmic surface. In addition,
of the α-polypeptide chain, which has a C- together with LH2 nonamers (diameter ~ 50
terminal extension that is 21 amino acids Å), occasional large rings (diameter ~ 120 Å)
longer than the C-terminus of α-subunits were imaged. Their dimensions and rotational
from Rps. acidophila. As proposed on the power spectra suggest these rings to be a
basis of its hydrophobicity, this extension minor LH1 contaminant. These results
could be folded into a second transmembrane demonstrate the potential of AFM for the
helix, leading to an α-subunit spanning the assessment of the in vivo photosynthetic
membrane twice in a hair pin structure system.
(Brunisholz et al., 1994). Recent electron
microscopy studies of Rvi. gelatinosus LH2 2.2.3. Results
reconstituted in 2D crystals provided
projection maps of negatively stained and in Rvi. gelatinosus LH2 protein is built by
ice embedded samples revealing a cylindrical the α- and the β-polypeptides consisting of
ring-like assembly of Rvi. gelatinosus LH2 71 and 51 aminoacids. Sequence alignment
complexes with a 9-fold symmetry with inner with LH2 of Rps. acidophila (using
and outer diameters similar to those reported ClustalW4) and hydropathy analysis led to
from X-ray models of the nonameric Rps. the topology model displayed in figure 1a.
acidophila LH2 (Ranck et al., in preparation; The thermolysin cleavage site has been
McDermott et al., 1995). Comparison of the determined by HPLC and mass spectroscopy
projection maps from 2D crystals of native (Ranck et al., in preparation) and is indicated
and truncated LH2, in which the C-terminal by the triangle. The limited proteolysis is
extension has been digested by thermolysin, illustrated by the silver stained gel in Figure
did not allow any extra density to be detected 1b which was obtained after a mild
inside or outside the nonameric ring. solubilization of 2D crystals reconstituted
Therefore, the location of the C-terminal from native and from digested LH2. Native

F i g u r e 2 . Height measurements of LH2 2D crystals adsorbed to mica in buffer (10 mM Tris-HCl, pH 7.2,
150 mM KCl, 25 mM MgCl 2) and imaged under physiological conditions (10 mM Tris-HCl, pH 7.2, 150 mM
KCl). a) AFM topograph of double and multi layered areas can clearly be distinguished from single layer crystals by
their higher appearance (full image size: 4 µm; full gray scale: 30 nm) b) Section analysis along the white line in
image a). The 2D crystals show a uniform height of 64.5 ± 2.8 Å (n = 46) (vertical scale bar: 100 Å). c) AFM
topograph of a LH2 sheet containing particles in up-and-down crystalline packing in the center and crystalline areas
exposing only the lower surface on the edges (full image size: 700 nm; full gray scale: 10 nm). d) Section analysis
along the white line in image c). The two different surface types are clearly visible: While the central region shows a
characteristic section analysis for up-and-down packing with strong surface corrugation, the edge areas appear smooth
and show a height of only 57 Å above the mica support (vertical scale bar: 50 Å). e) Medium magnification image
of a crystalline sheet of LH2 rings. At this magnification the ring-structure of the complexes is already clearly
visible. The crystals show coherence only over small regions and lattice displacements of half a unit cell are frequent.
While the ring-structure of the high side is well visible in the crystalline areas, the lower circles can better be seen
within the crystal defects (bottom right) (full image size: 400 nm; full gray scale: 4 nm). f) Section analysis along
the white line in image e). The height of LH2 rings can directly be measured from such section analysis of raw data
(vertical scale bar: 20 Å).

LH2 complexes run with an apparent Mw of al., in preparation).
115 kDa, while digested complexes run at
82 kDa, reflecting the removal of 20 ClustalW.html
aminoacids from each α-subunit of the 5link:
complex. Sharp bands document the specific TMPRED_form.html
and effective cleavage of the protein. Fig. 1c page=npsa_gor4.html
displays the absorption spectra of the native
LH2 (black line) and truncated LH2 (gray Large 2D arrays of LH2 complexes were
line) after reconstitution into lipid bilayers. produced by detergent removal from a
Both spectra show the characteristic micellar solution containing the LDAO-
absorption bands of carotenoids at 460 nm, purified protein supplemented with egg
488 nm and 517 nm, the Qx Bchl band at phosphatidyl choline and octylthioglucoside,
595 nm, and the Qy bands at 802 nm and a detergent which has been recently reported
859 nm, indicating native Bchls and to significantly increase the size of the
carotenoid binding sites (see also Ranck et reconstituted 2D crystals (Chami et al., in

Table 1. LH2 dimensions measured with the AFM under physiological conditions

native LH2 sample digested LH2 sample

average SD a nb average SD a nb

Overall thickness

lipid bilayer: 41.9 Å ± 2.5 Å 37 41.2 ± 2.5 Å 23

up-and-down 2D-crystal: 64.5 Å ± 2.8 Å 46 60.7 ± 1.6 Å 33

one sided 2D-crystal: 60.4 Å ± 1.4 Å 11 57.4 ± 2.5 Å 81

dimensions of large protrusions:

height (lipid-top)c 13.9 Å ± 1.7 Å 72 8.9 ± 1.4 Å 110

height (ring center-top)d 7.6 Å ± 2.0 Å 73 5.3 ± 0.9 Å 33

diameter of tope 49.0 Å ± 0.5 Å 3* 49.0 ± 0.5 Å 3*

volume / subunit 3297Å3 ± 517 Å 3 # 1765 Å3 ± 353 Å 3 #

dimensions of small protrusions:

height (lipid-top)c 5.6 Å ± 0.8 Å 152 5.4 ± 0.5 Å 34

height (ring center-top)d 5.6 Å ± 1.0 Å 24 5.3 ± 0.8 Å 137

diameter of tope 54.0 Å ± 0.5 Å 3* 54.0 ± 0.5 Å 3*

volume / subunit 1001 Å3 ± 161 Å 3 # 952 Å3 ± 182 Å 3 #

a standard deviation.
b number of measurements.
c height difference between the top of the ring and the lipid surface.
d height difference between the top and the center of the ring.
e ring diameter measured at its top after rotational averaging.
* average of 3 independent AFM topographs at different nominal magnifications scaled with respect to the unit
cell dimensions found in cryo electron microscopy.
# volumes and deviations calculated taking standard errors in height measurements and standard deviation signal
trough single particle averaging into account.

preparation; see also material and methods). Tris-HCl, pH 7.2, 150 mM KCl) to achieve
The use of bio-beads as detergent removing slightly repulsive force curves (data not
agent allowed production of the reconstituted shown) on the reconstituted 2D crystals
membranes in few hours, avoiding any (Müller et al., 1999). Under such conditions
denaturation of the LH2 complexes. the 2D-crystals of both sample types
Using a high ionic strength buffer (10 mM remained attached to the mica for hours. The
Tris-HCl, pH 7.2, 150 mM KCl, 25 mM height measured for the lipid bilayer, 41.9
MgCl2) 2D-crystals of native and predigested Å (native sample), 41.2 Å (digested sample)
Rvi. gelatinosus LH2 could be firmly attached and for the crystals, 64.5 Å (native sample),
to the mica AFM support (Schabert & Engel, 60.7 Å (digested sample) indicated the C-
1994). To acquire high resolution topographs, terminus protrudes from the membrane (see
the buffer was carefully exchanged (10 mM Table 1, Fig. 2a and 2b).
At higher nominal magnification the ring- al., 1995), Rhb. sphaeroides (Walz et al.,
structure of LH2 became distinct, revealing 1998) and Rhv. sulfidophilum (Montoya et
the two different sides of the transmembrane al., 1995). The reconstituted LH2 complexes
complex (Fig. 2e), which are regularly packed (Fig. 3a, b, c, d) protruded by 13.9 ± 1.7 Å
in alternative orientations. Unit cells from one side of the membrane (Table 1, Fig.
containing 2 rings have dimensions of 3a, b), but only by 5.6 ± 0.8 Å on the other
a = 82 Å, b = 133 Å, and γ = 90°. Arrows in side (Table 1; Fig. 3c, d). These lower
figure 2e indicate defects in the crystal lattice protrusions could be best imaged close to the
where lower rings were more clearly visible edges of sheets (see arrows in Fig. 2c), where
than within the up-and-down packing between unidirectionally inserted complexes were
the strongly protruding surfaces. Such areas found to assemble into a hexagonal closest
were also identified at lower magnification packing with unit cell dimensions of
(see arrows in Fig. 2c, Fig. 2d) and allowed a = b = 76 Å, and γ = 60°. The membranes
their height to be measured, 60.4 Å (native reconstituted in presence of digested LH2
sample), 57.4 Å (digested sample), as revealed a very similar overall organization of
summarized in Table 1. the protein (Fig. 3e, f, g, h). The weakly
In agreement with electron microscopy protruding (5.4 ± 0.5 Å) surface showed an
studies (Ranck et al., in preparation) high identical surface appearance compared to the
resolution topographs (Fig. 3a, c, e, g) native sample (compare Fig. 3g, h with Fig.
revealed that Rvi. gelatinosus shares the 3c, d). Relatively large areas with complexes
circular nonameric organization of LH2 exposing their lower surface to the tip allowed
subunits with Rps. acidophila (McDermott et the right handed twist of the protruding

Figure 3. High resolution raw data AFM topographs and corresponding averages (all in 15° tilted representation
and corresponding heights). Averages were gained through a single particle alignment routine an 9-fold symmetrized.
a) Raw data topography of the strongly (~ 14 Å) protruding surface of the native LH2 complex. The 9-fold
symmetry is in the raw data visible (scale bar: 100 Å; full gray scale: 20 Å). b) Average of image a) (scale bar:
20 Å; full gray scale: 14 Å). c) Raw data topography of the weakly (~ 5 Å) protruding surface of the native LH2
complex (scale bar: 100 Å; full gray scale: 20 Å). d) Average of image c) (scale bar: 20 Å; full gray scale: 6 Å).
e) Raw data topography of the strongly (~ 9 Å) protruding surface of the digested LH2 complex (scale bar: 100 Å;
full gray scale: 20 Å). f) Average of image e) (scale bar: 20 Å; full gray scale: 9 Å). The loss of protruding
structure compared to b) localizes the C-terminal position of the a-subunit. g) Raw data topography of the weakly
(~ 5 Å) protruding surface of the digested LH2 complex (scale bar: 100 Å; full gray scale: 20 Å). h) Average of
image g). The proteolysis had no influence on the topography of the weakly protruding surface (scale bar: 20 Å; full
gray scale: 6 Å).

AFM data, the respective diameters were
54.0 Å for the lower ring and 49.0 Å for the
higher ring (see Table 1).
In membranes reconstituted from a LH2
preparation in which the hydroxyapatite
column purification step was omitted (see
materials and methods), a few giant rings
(diameter of ~ 120 Å) were imaged at
submolecular resolution surrounded by
smaller LH2 rings (diameter of ~ 50 Å) (Fig.
4a). Rotational power spectra of 68 LH2
rings and 7 large rings from the same images
are displayed in figure 4b. The merged power
spectra of the LH2 rings showed a strong
signal corresponding to the 9-fold symmetry
(gray line in Fig. 4b). The large rings,
however, showed a much stronger intensity
for 8-fold symmetry than for 9-fold
symmetry, together with weak signals
corresponding to 12 and 16-fold symmetries.
Figure 4. a) Raw data AFM topograph showing a
~ 120 Å ring surrounded by LH2 nonamers revealing 2.2.4. Discussion
distances between the complexes of ~ 10 Å (see
Discussion; scale bar: 100 Å; full gray scale: 20 Å) b)
Rotational power spectra of 7 large rings (black line) During the last few years AFM has
and 68 small rings (gray line) from images as displayed become a powerful tool in membrane protein
in figure a). The small LH2 rings display a clear 9-fold
symmetry. The large rings show a strong amplitude on research. This progress is the result of
the 8-fold and a weak on the 16-fold symmetry, while improved instrumentation as well as
the 9-fold and 18-fold symmetry is less pronounced,
supporting the assignment of the large rings with the optimized recording conditions (Engel &
hexadecameric LH1 rings. Müller, 2000). The AFM allows information
to be acquired on the membrane protruding
structure to be resolved (Fig. 3g, h). The band structure of single proteins. Such information
shift in the gel (Fig. 1b) documents a is difficult to obtain with cryo-electron
complete cleavage of the C-terminus of the α- microscopy where the membrane embedded
subunit. Accordingly, a change in the strongly parts are preserved, but connecting loops and
protruding surface was observed in the protruding termini are often distorted.
digested sample. Its height was reduced to Furthermore heights of membranes and loops
8.9 Å over the lipid bilayer (Fig. 3e, f), and can be measured accurately in buffer solution
the cleaved protrusions appeared broader than with the AFM (Müller et al., 1999a), poorly
the uncleaved protrusions in figure 3a. Thus, ordered single particles can be recognized and
the volume integral over the protruding imaged at high resolution (< 10 Å)
structure changed from ~ 3300 Å3 to (Scheuring et al., 1999; Fotiadis et al., 2000;
~ 1800 Å3 (Table 1), compatible with the loss Seelert et al., 2000), and sidedness
of the C-terminus. assignments can directly be obtained from
As displayed in figure 3, the lower rings raw data (Scheuring et al., 1999).
had a larger diameter than the higher rings. LH2 complexes are rings of αβ-
Taking unit cell dimensions determined by heterodimers, each subunit crossing the
cryo-electron microscopy for scaling the plasma membrane of the photosynthetic

bacteria once. For Rps. acidophila LH2 the surface which is thought to comprise the 35
α-polypeptide contains a perpendicular C-terminal aminoacids of the α-polypeptide
transmembrane alpha-helical segment, which and 6 C-terminal aminoacids of the β-
constitutes the inner surface of the ring, while polypeptide (Fig.1) has a smaller diameter
the β-polypeptide is a transmembrane helix (~ 49 Å) than the cytoplasmic protrusions
that is tilted by ~ 15° and forms the periphery (~ 54 Å) housing 11 N-terminal aminoacids
of the cylinder (McDermott et al., 1995). The of the α-polypeptide and 14 N-terminal
LH2 of Rvi. gelatinosus carries a large C- aminoacids of the β-polypeptide (Table 1).
terminus on the α-polypeptide of about 35 Since the α-subunits face the inner diameter
amino acids (half of the polypeptide), with a of the LH2 rings in Rps. acidophila
hydrophobic extension of 21 amino acids, (McDermott et al., 1995), the more
protruding out of the membrane on the protruding structure is expected to form a
periplasmic side (Fig. 1). narrower cylinder, corroborating the
High resolution AFM topographs of sidedness assignment based on proteolytic
reconstituted arrays of LH2 complexes cleavage.
revealed the nonameric organization of the The high signal-to-noise ratio of AFM
LH2 rings (Fig. 3), in agreement with an topographs allows single non ordered
electron crystallographic study of the same proteins to be imaged at submolecular
2D crystals, which showed the ring-like resolution (Scheuring et al., 1999; Seelert et
nonameric organization of the αβ- al., 2000; Fotiadis et al., 2000). Figure 4a
heterodimers, with outer and inner diameters shows a ring with a diameter of ~ 120 Å
of about 66 Å and 30 Å respectively (Ranck which is in surrounded with LH2 rings with
et al., in preparation). However, these diameters of ~ 50 Å. Contact distances
projection maps did not show major between large and LH2 rings are identical to
differences between digested and undigested those between LH2 rings, ~ 10 Å.
LH2 complexes, preventing the identification Identification of the large ring as minor LH1
of the C-terminal extension. Topographs of contaminants is based on the agreement of its
membranes packed with native and digested diameter with that of LH1 (116 Å, Karrasch et
LH2, revealing a thermolysin induced change al., 1995), and the merged rotational power
of height and surface appearance of the spectrum of seven large rings, revealing a
strongly protruding surface, allowed major peak at 8-fold and minor peaks at 12-
identification of the C-terminus of the a- and 16-fold symmetry. Fusions of two LH2
polypeptide (Fig. 3). Thus, the surface rings can be excluded, as such rings should
protruding ~ 14 Å from the membrane possess a strong peak at 9-fold symmetry and
represent the periplasmic side, while the should occur more frequently.
cytoplasmic surface housing the N-termini In conclusion, we have imaged the surfaces
protrudes by only ~ 5 Å. The volume change of the nonameric complexes from Rvi.
of the periplasmic protrusions from gelatinosus LH2 with the AFM and identified
~ 3300 Å3 to ~ 1800 Å3 (Fig. 3, Table 1) the periplasmic surface with the C-terminus
induced by thermolysin corresponds to a of the a-polypeptide. The high signal-to-noise
calculated change of 10 aminoacids. This is ratio of images acquired with the AFM
much less than the 20 aminoacids cleaved by allowed recognition of different protein
thermolysin (Fig. 1a, b), but can be explained complexes that work together in the native
by the flexibility of the uncleaved C-terminus membrane. As a next step in understanding
protruding far out of the membrane surface the photosynthetic apparatus native
(14 Å). membranes will be directly imaged with the
The strongly protruding periplasmic AFM under physiological conditions. This

will resolve questions related to the Biochemical and biophysical
oligomerization states of different LH techniques
complexes in situ and to the open/closed
configuration of the LH1 rings. Protein concentration was determined from
the absorption at 854 nm using ε = mM -1cm-
2.2.5. Materials and methods 1 = 382 (Sturgis et al., 1995) and values of
12530 Daltons and 10933 Daltons for the Materials molecular weights of the native and the
thermolysin cleaved LH2 respectively (Ranck
All phospholipids were of the highest et al., in preparation). Absorption spectra of
purity and were purchased from Avanti Polar the vesicles were recorded on a Cary 2300
Lipids. N,N- dimethyldodecylamine N-oxide spectrophotometer.
(LDAO, 30 % solution) was from Fluka and Vesicles containing native or cleaved LH2
n-Octyl-b-D thioglucopyranoside (OTG) was were analyzed by SDS gel electrophoresis on
from Sigma. Thermolysin was purchased silver stain 10 % acrylamide gels under
from Boehringer. Bio-Beads SM2 (25 - 50 oxidizing conditions.
mesh) from Bio-Rad were extensively washed
with methanol and water before use as Reconstitution and 2D
described (Levy et al., 1990). All other crystallization
reagents were of analytical grade.
2D crystallization of native and truncated Isolation, purification and LH2 complexes were performed as described
proteolysis of LH2 complex previously (Ranck et al., in preparation;
Chami et al., in preparation). Briefly, purified
The light-harvesting LH2 complex was LH2 complexes were diluted to about 0.5 mg
isolated from photosynthetically grown Rvi. /ml in a buffer containing 10 mM Tris-HCl,
gelatinosus cells (Strain S1) essentially as pH 8.0, 400 mM NaCl, 0.1 % LDAO and
described previously (Jirsakova et al., 1996), supplemented with 20 mM Octyl
with slight modifications (Ranck et al., in thioglucoside. Then egg phosphatidyl choline
preparation). Briefly, solubilization of the was added at lipid-to-protein ratios ranging
broken cells with LDAO was followed by from 0.3 to 0.5 w/w and the micellar solution
two successive chromatographic purification allowed to equilibrate for about one hour in
on DEAE-Sepharose FF and Sepharose CL- the dark at room temperature. Detergent
6B (Pharmacia) columns. Further purification removal was performed through three
was achieved by chromatography on a successive additions of 5 mg SM2 Bio-Beads
hydroxyapatite (Biosepra) column eluted in a for 1 hour each, according to the batch
buffer containing 10 mM Tris-HCl, pH 8.0, procedure previously described by Rigaud et
1 mM EDTA and 0.1 % LDAO. al. (1998). After 3 hours of stirring in
A limited digestion by thermolysin (4 hr presence of polystyrene beads at room
incubation time at 22°C, enzyme/LH2 molar temperature, the reconstituted material was
ratio = 20) was performed on purified LH2 in kept at 4° C for AFM analysis.
LDAO solution, and its effect was analyzed
by denaturing SDS-PAGE and by MALDI- Atomic force microscopy
TOF mass spectroscopy as described
elsewhere (Ranck et al., in preparation). Mica prepared as described by Schabert &
Engel (1994) was used as support and freshly
cleaved before every experiment using scotch

tape. To check the cleavage quality the mica 2.2.6. Acknowledgment
was imaged in 30 - 50 µl of adsorption buffer
(10 mM Tris-HCl, pH 7.2, 150 mM KCl, We thank Kitaru Suda for making the gel
25 mM MgCl2). Subsequently 1 µl of shown in figure 1b, C. Möller and D.J.
protein crystal solution (0.1 mg/ml) was Müller for fruitful discussions on the AFM
injected into the adsorption buffer drop on the technique, and C. Leiniger for her help with
mica surface. After 2 hours the sample was the manuscript. This work was supported by
carefully rinsed with recording buffer the Swiss National Foundation for Scientific
(10 mM Tris-HCl, pH 7.2, 150 mM KCl). Research (grant 4036 - 44062 to A. E.), the
The recording buffer was optimized to Maurice E. Müller Foundation of Switzerland
achieve high resolution as described (Müller and the Centre National de la Recherche
et al., 1999a). Imaging was performed with a Scientifique (programme PCV to FRH).
commercial Nanoscope III AFM (from
Digital Instruments, Santa Barbara, CA, USA) 2.2.7. References
equipped with a 120 µm scanner (J-scanner)
and oxide-sharpened Si3N4 cantilevers with a Binnig, G., Quate, C. F. & Gerber, C. (1986).
length of 120 µm (k = 0.1 N/m; Olympus Atomic force microscope. Phys. Rev. Lett.
56, 930-933.
Ltd., Tokyo, Japan). Sharpest tips adequate
for high resolution imaging were cleaned in Brunisholz, R. A., Suter, F. & Zuber, H.
1 % SDS and subsequently washed 3 times (1994). Structural and spectral
in distilled water. The AFM was operated in characterization of the antenna complexes
of Rhodocyclus gelatinosus. Indication of
contact mode applying forces below < 0.2 nN a hair-pin-like arranged antenna
at a scan frequency of 4 - 6 Hz. The apoprotein with an unusually high alanine
instrument was calibrated using layered content. Eur. J . Biochem. 222, 667-675.
crystals of MoTe2 as described previously
Chami, M., Pehau-Arudet, G., Lambert,O.,
(Müller & Engel, 1997). Ranck, J.L., Levy, D. and Rigaud ,J.L.
(2000). Use of octyl-beta-thiogluco- Image processing pyranoside in 2D crystallization of
membrane proteins. J. Struct. Biol. in
AFM images were calculated as 15° tilted
surface representations using Image SXM Engel, A. & Müller, D. J. (2000). Observing
(Fig. 3a, c, e, g, 4a). Rings were aligned and single biomolecules at work with the
averaged by a single particle averaging atomic force microscopy. Nat. struct. biol.
7, 715-718.
protocol using the SEMPER image
processing system (Fig. 3b, d, f, h; Saxton et Fotiadis, D., Hasler, L., Müller, D. J.,
al., 1979). Stahlberg, H., Kistler, J. & Engel, A.
Volumes in Å3 were directly calculated (2000). Surface tongue-and-groove
contours on lens MIP facilitate cell-to-cell
using x, y and z dimensions of AFM adherence. J. Mol. Biol. 300, 779-789.
topograph averages, aminoacid numbers were Francia, F., Wang, J., Venturoli, G., Melandri,
deduced using 1.35 Å3/Da for average B., Barz, W. & Oesterheld, D. (1999). The
protein density in secondary protein structure, reaction center-LH1 antenna complex of
Rhodobacter sphaeroides contains one
and 110 Da as average molecular weight of PufX molecule, which is involved in
one aminoacid. dimerization of this complex. Biochem. 38,

Frese, R. N., Olsen, J. D., Branvall, R.,
Westerhuis, W. H., Hunter, C. N. & van
Grondelle, R. (2000). The long-range

supraorganization of the bacterial of biomolecules measured with the atomic
photosynthetic unit: A key role for PufX force microscope depends on electrostatic
[see comments]. PNAS 97, 5197-202. interactions. Biophys. J. 73, 1633-1644.

Jirsakova, V. & Reiss-Husson, F. (1994). A Müller, D. J., Fotiadis, D., Scheuring, S.,
specific carotenoid is required for Müller, S. A. & Engel, A. (1999a).
reconstitution of the Rubrivivax Electrostatically balanced subnanometer
gelatinosus B875 light harvesting complex imaging of biological specimens by atomic
from its subunit form B820. FEBS Lett. force microscopy. Biophys. J. 76, 1101-
353, 151-4. 1111.

Jungas, C., Ranck, J. L., Rigaud, J. L., Joliot, Müller, D. J., Sass, H.-J., Mülller, S., Büldt,
P. & Vermeglio, A. (1999). G. & Engel, A. (1999b). Surface
Supramolecular organization of the structures of native bacteriorhodopsin
photosynthetic apparatus of Rhodobacter depend on the molecular packing
sphaeroides. EMBO J. 18, 534-42. arrangement in the membrane. J. Mol.
Biol. 285, 1903-1909.
Karrasch, S., Bullough, P. A. & Ghosh, R.
(1995). The 8.5 A projection map of the Müller, D. J., Schabert, F. A., Büldt, G. &
light-harvesting complex I from Engel, A. (1995). Imaging purple
Rhodospirillum rubrum reveals a ring membranes in aqueous solutions at
composed of 16 subunits. EMBO J. 14, subnanometer resolution by atomic force
631-8. microscopy. Biophys. J. 68, 1681-1686.

Koepke, J., Hu, X., Muenke, C., Schulten, K. Papiz, M.Z., Prince, S.M., Hawthornthwaite-
& Michel, H. (1996). The crystal structure Lawless, A.M., McDermott, G., Freer,
of the light-harvesting complex II (B800- A.A. , Isaacs, N.W., Cogdell, R.J., (1996).
850) from Rhodospirillum molischianum. A model for the photosynthetic apparatus
Structure 4, 581-97. of purple bacteria. Trends in Plant
Science, 1, 198-206
Kuhlbrandt, W. (1995). Many wheels make
light work. Nature 374, 497-498. Ranck, J.L., Ruiz, T., Pehaut-Arnaudet, G.,
Arnoux,B. and Reiss-Husson, F., (2000).
Levy, D., Bluzat, A., Seigneuret, M. & 2D structure of the light-harvesting
Rigaud, J.-L. (1990). A systematic study complex LH2 from Rv. Gelatinosus:
of liposome and proteoliposome comparison of the native complex and of a
reconstitution involving bio-bead-mediated truncated form obtained by limited
triton-X removal. BBA 1025, 179-190. proteolysis. J. Biol. Chem., in
Loach, P. A. (2000). Supramolecular
complexes in photosynthetic bacteria. Rigaud, J.L., Lévy, D., Mosser, G., and
PNAS 97, 5016-5018. Lambert, O., (1998). Detergent removal by
Bio-Beads. Eur. Biophys.J., 27, 305-319.
McDermott, G., Prince, S.M., Freer, A.A.,
Hawthornthwaite-Lawless, A.M., Papiz,
M.Z., Cogdell, R.J., Isaacs, N.W. (1995). Saxton, W. O., Pitt, T. J. & Horner, M.
Crystal structure of an integral membrane (1979). Digital image processing: The
light-harvesting complex from Semper system. Ultramicroscopy 4, 343-
photosynthetic bacteria. Nature 374, 517- 354.
Schabert, F. A. & Engel, A. (1994).
Montoya, G., Cyrklaff, M. & Sinning, I. Reproducible acquisition of Escherichia
(1995). Two-dimensional crystallization coli porin surface topographs by atomic
and preliminary structure analysis of light force microscopy. Biophys. J. 67, 2394-
harvesting II (B800-850) complex from 2403.
the purple bacterium Rhodovulum
sulfidophilum. J. Mol. Biol. 250, 1-10. Schabert, F. A., Henn, C. & Engel, A. (1995).
Native Escherichia coli OmpF porin
Müller, D. J. & Engel, A. (1997). The height surfaces probed by atomic force

microscopy. Science 268, 92-94. Sturgis, J., Jirsakova, V., Reiss-Husson, F.,
Cogdell, R. & Robert, B. (1995). Structure
Scheuring, S., Ringler, P., Borgnia, M., and property of bacteriochlorophyll
Müller, D. J., Agre, P. & Engel, A. (1999). binding site in peripheral light-harvesting
High resolution topographs of the complexes of purple bacteria. Biochem.
Escherichia coli waterchannel AqpZ. 34, 517-523.
EMBO J. 18, 4981-4987
Walz, T. & Ghosh, R. (1997). Two-
Seelert, H., Poetsch, A., Dencher, N., Engel, dimensional crystallization of the light-
A., Stahlberg, H. & Müller, D. J. (2000). harvesting I-reaction centre photounit from
Proton-powered turbine of a plant motor. Rhodospirillum rubrum. J. Mol. Biol. 265,
Nature 405, 418-419. 107-11.

Stahlberg, H., Dubochet, J., Vogel, H. & Walz, T., Jamieson, S. J., Bowers, C. M.,
Ghosh, R. (1998). Are the light-harvesting Bullough, P. A. & Hunter, C. N. (1998).
I complexes from Rhodospirillum rubrum Projection structures of three
arranged around the reaction center in a photosynthetic complexes from
square geometry? J. Mol. Biol. 282, 819- Rhodobacter sphaeroides: LH2 at 6 Å,
31. LH1 and RC-LH1 at 25 A. J. Mol. Biol.
282, 833-45.

3. Combining surface and
projection techniques

3. Combining surface and projection techniques

3.1. The aquaporin sidedness revisited

Simon Scheuring 1, Peter Tittmann2, Henning Stahlberg 1, Philippe Ringler1, Mario
Borgnia 3, Peter Agre3, Heinz Gross2, and Andreas Engel1

1M.E. Müller Institute for Microscopy at the Biozentrum, University of Basel, Klingelbergstr. 70,
CH-4056 Basel, Switzerland
2Institute of Applied Physics, ETH Zürich, CH-8093 Zürich, Switzerland
3Department of Biological Chemistry and Medicine, Johns Hopkins University School of
Medicine, Baltimore MD 21205-21850, USA

3.1.1. Summary allowed an unambiguous alignment of the
surface reliefs to the underlying density
Aquaporins are transmembrane water maps. AqpZ topographs previously
channel proteins, which play important determined by AFM could then be aligned
functions in the osmoregulation and water with projection maps of AqpZ, and finally
balance of micro-organisms, plants, and with human erythrocyte aquaporin-1 (AQP1).
animal tissues. All aquaporins studied to date Thereby features of the AqpZ topography
are thought to be tetrameric assemblies of could be interpreted by direct comparison to
four subunits each containing its own the 6Å three-dimensional structure of AQP1.
aqueous pore. Moreover, the subunits contain We conclude that the sidedness we originally
an internal sequence repeat forming two proposed for aquaporin density maps was
obversely symmetric hemichannels predicted inverted (Walz et al., 1996).
to resemble an hourglass. This unique
arrangement of two highly related protein 3.1.2. Introduction
domains oriented at 180° to each other poses
a significant challenge in the determination of To maintain metabolic processes water
sidedness. Aquaporin Z (AqpZ) from molecules must efficiently permeate the
Escherichia Coli was reconstituted into highly plasma membranes of cells in all living
ordered two-dimensional crystals. They were organisms. Since the diffusion of water
freeze-dried and metal-shadowed to establish molecules through lipid bilayers has an
the relationship between surface structure and activation energy > 10 kcal/mol (Chandy et
underlying protein density by electron al., 1997), the existence of specific water
microscopy. The shadowing of some surfaces pores was postulated more than 4 decades
was prevented by protruding aggregates. ago (Sidel & Solomon, 1957). The first
Thus, images collected from freeze-dried member of this family termed the aquaporins
crystals that exhibited both metal-coated and (Chrispeels & Agre, 1994) and designed by
uncoated regions allowed surface relief evolution to facilitate water transport, was
reconstructions and projection maps to be identified by Preston et al. in 1992.
obtained from the same crystal. Cross Aquaporin sequences share six hydrophobic
correlation peak searches along lattices stretches, which correspond to trans-
crossing metal-coated and uncoated regions membrane helices. Two long conserved loops,

B and E, connect helices 2 and 3, and 5 and 6, had p4212 symmetry (Ringler et al., 1999).
respectively, and accommodate the highly Cryo electron microscopy provided a
conserved NPA motifs (Gorin et al., 1984; projection map at 7 Å resolution exhibiting
Preston & Agre, 1991). These loops fold the characteristic features of AQP1 consistent
back into the membrane, to form the structure with the high sequence homology of the
of the pore (Jung et al., 1994). Permeability aquaporins (Ringler et al., 1999). 2D crystals
studies by stopped flow measurements assembled from AqpZ bearing an N-terminal
indicate flow rates in the range of 109 water fragment of 26 aminoacids containing 10
molecules per channel and per second, and an histidines had the same symmetry and unit
activation energy < 5 kcal/mol (Walz et al., cell dimensions. AFM studies before and
1994b; Zeidel et al., 1992). after removal of this N-terminal fragment with
Aquaporin-1 (AQP1) of human trypsin, allowed the sidedness of AqpZ
erythrocytes (Agre et al., 1993) is structurally surfaces to be unambiguously assigned
the best studied aquaporin. Two-dimensional (Scheuring et al., 1999). The crown-like
(2D) crystals with two tetramers packed in extracellular side possesses three protrusions
opposite orientation into a unit cell with of 7 Å height per monomer, of which the
dimensions of a = b = 96 Å and g = 90° have largest one was identified as loop C, which
been reconstituted in the presence of lipids comprises 26 aminoacids. One protrusion per
(Walz et al., 1994a). These highly ordered 2D monomer was visible on the cytoplasmic
crystals diffracted a 300 kV electron beam to surface, probably resulting from loops B
at least 3.5 Å resolution (Mitsuoka et al., and/or D (Scheuring et al., 1999).
1999). A 3D density map at 6 Å resolution In contrast, the sidedness of AQP1 has
calculated from projections of samples tilted been determined using surface relief
with respect to the electron beam, revealed a reconstruction of metal-shadowed AQP1 2D
right-handed bundle consisting of six crystals before and after digestion with
transmembrane a-helices surrounding a carboxy peptidase Y (Walz et al., 1996).
central density (Walz et al., 1997; Cheng et Although these results promoted a straight-
al., 1997), in agreement with sequence based forward interpretation, the recent 4.5 Å 3D
structure prediction. The central density density map by Mitsuoka et al., (1999)
formed by the long loops B and E has suggested a different assignment of the
recently been resolved as two short helices sidedness. Therefore, we analyzed freeze-
projecting outwards from the center of the dried AqpZ 2D crystals that were partially
monomer which are connected to adjacent metal-shadowed to calculate both surface
helices by loop regions (Mitsuoka et al., reliefs and projection maps of one and the
1999), thus confirming the hour glass model same crystal. In this way the sidedness of the
(Jung et al., 1994). AqpZ projection map could be identified and
The E. coli waterchannel AqpZ has been related to that of AQP1. The results presented
identified by expression cloning (Calamita et here suggest that the sidedness of AQP1
al., 1995) and overexpressed in its native previously reported by Walz et al. (1996) has
system (Borgnia et al., 1999). This bacterial to be revised.
waterchannel maintains cell turgor during the
volume expansion of cell division (Calamita et 3.1.3. Results
al., 1998). Highly ordered 2D crystals have
been grown by dialysis of protein-lipid- Trypsin digested AqpZ 2D crystals
detergent mixtures (Ringler et al., 1999). The adsorbed to glow discharged carbon films
square unit cells with dimensions a = b = 95 were washed in distilled water, and quickly
Å, g = 90°, contained eight monomers and frozen in liquid nitrogen. After freeze-drying

Figure 1. a) Overview image of a freeze-dried and subsequently metal-shadowed crystalline sheet of AqpZ adsorbed
to a carbon-coated electron microscopy grid. The black and white arrows pointing towards each other indicate the
borderline of metal-coated and uncoated areas. The asterisk indicates the aggregate, which prevented parts of the
crystal from metal deposition (compare aggregate and shadow borderline). The white frame defines the position of the
following higher magnification image. The shadowing azimuth is indicated in the top right (scale bar: 1µm). b)
Low dose image of the area outlined in Figure 1a. The metal-shadowed side (top left) is darker than to the uncoated
region (bottom right). The squares (n = 971) and crosses (n = 1444) indicate unit cell positions on the metal-coated
side and on the uncoated side, respectively, which fitted with a displacement tolerance smaller than 0.05 (4.75 Å) to
the square lattice of 95 Å. The shadowing azimuth is given in the top right (scale bar: 100 nm). c) Correlation
average of the 971 unit cells found on the unidirectionally metal-shadowed area. The four-fold symmetry of the
tetramers is lost due to the unidirectional metal deposition. The shadowing azimuth is displayed in the top right (the
white frame indicates the unit cell: 95 x 95 Å). d) Four-fold symmetrized surface reconstruction of the cross
correlation average displayed in Figure 1c (the white frame indicates the unit cell: 95 x 95 Å). Rhombi surrounding
the higher, extracellular surface form a right-handed windmill-like structure. e) Correlation average of the 1444 unit
cells found on the uncoated freeze-dried area. The four-fold symmetry of the tetramers is preserved. Adjacent tetramers
appear with different brightness probably due to crystal - carbon film interaction (white frame indicates unit cell: 95
x 95 Å). f) p4212 symmetrized projection map of the cross correlation average displayed in Figure 1e (the white
frame indicates the unit cell: 95 x 95 Å). The central tetramer is the view from the extracellular side (corresponding
to the central particle in Figure 1d) and is surrounded by rhombi forming a right-handed windmill-like structure.

and deposition of a 5 Å thick heavy-metal flat single layered sheets were selected (see
film (see Walz et al. 1996), the crystals were Figure 1a).
imaged at a temperature of - 180° C at doses Overview images of metal-shadowed
below 5 electrons/Å2. Overview images samples (Figure 1a; arrow top right indicates
(Figure 1a) were taken with the Gatan slow the shadowing direction) were carefully
scan CCD camera at a magnification of searched for the borderlines of metal-coating
x4'000 and directly used to position (indicated by the black and the white arrows
crystalline areas for low dose imaging at high in Figure 1a). Unshadowed areas resulted
magnification. To this end, tightly adsorbed, when aggregates (indicated by the asterisk in

Figure 1a) blocked the metal beam. Areas shows the four fold symmetrized surface
such as that outlined by the square in Figure relief reconstruction of Figure 1c. The pixel
1a were then recorded at a magnification of sampling of 5 Å prevented resolution of fine
x45'000 (5 Å/pixel on the CCD camera, substructures in this average, but the
Figure 1b), yielding a large number of unit orientation of the lipid filled rhombus
cells from the freeze-dried, metal-shadowed between adjacent tetramers, and the height
area and from the uncoated area. Raw data of difference of the tetramers with respect to
shadowed areas appear dark (Figure 1b, top each other is distinct. The projection average
left), while uncoated areas are bright (Figure of 1444 unit cells of the uncoated freeze-dried
1b, bottom right). Such images were Fourier area (Figure 1e, white frame indicates the unit
peak filtered to produce a first reference for a cell) reveals the tetramer organization and
cross correlation peak search on both the orientation within the crystal lattice. Probably
metal-coated and the uncoated areas (Figure as result of interactions between the crystal
1b, the cross correlation peaks on the metal- and the carbon film or residual metal
shadowed area are indicated by squares, on shadowing neighboring particles differ
the uncoated area by crosses). A lattice slightly in brightness. Nevertheless, the
yielding the perpendicular vectors of 95 Å appearance of adjacent tetramers is very
length was fitted to the correlation peaks with similar and the four-fold symmetry is
a tolerance of 0.05 (4.75 Å). Figure 1c essentially preserved. After p4212
displays the cross correlation average of the symmetrization (Figure 1f, white frame
971 unit cells (white frame indicates the unit indicates the unit cell) the rotation of the
cell) found on the metal-coated area in Figure particles with respect to each other and the
1b. Adjacent tetramers have a different orientation of the rhombus-shaped lipid
appearance due to the up and down packing interspace is clearly visible. From Figure 1 we
of the particles within the crystal. As a conclude that the surface protruding most
consequence of unidirectional shadowing from the lipid bilayer (central tetramer in
(arrow top right indicates the shadowing Figure 1d) corresponds to the tetramer
direction) the four fold symmetry is lost. similarly surrounded by rhombi in right-
Figure 1d (white frame indicates the unit cell) handed orientation in the projection map

Figure 2. Correlation between surface topography and electron density projection map (panel sidelenghts 190 Å).
a) Average of high resolution AFM topographies (10° tilted surface representation) of trypsin digested AqpZ 2D
crystals imaged in buffer solution. The crown-like extracellular surface protrudes 7 Å out of the lipid bilayer, while
the cytoplasmic surface only protrudes by 3.5 Å (Scheuring et al. 1999). b) Surface reconstruction (10° tilted surface
representation) of freeze-dried unidirectionally metal-shadowed AqpZ 2D crystals. As indicated, right-hand oriented
rhombus-shaped lipid interspaces surround the extracellular surface. The asterisk indicates the position of protruding
structure in the metal-shadowing surface reconstruction map, which is not present in the topography recorded by the
AFM. c) Average density map of freeze-dried AqpZ 2D crystals. As indicated, right-hand oriented rhombus-shaped
lipid interspaces surround the projection viewed from the extracellular side. The asterisk indicates protein density
which induces a protrusion signal in the surface reconstruction (indicated by asterisk in Figure 2b, see Discussion).
d) Average density map of cryo electron micrographs from trehalose embedded AqpZ 2D crystals (Ringler et al.

(central tetramer in Figure 1f). As previously
reported, the most protruding surface is
extracellular (Scheuring et al., 1999).
Therefore, this central tetramer represents the
projection from the extracellular side.
To obtain higher resolution structural
information images of both metal-coated and
uncoated crystals were recorded at a
magnification of x77'000 (3.1Å/pixel on the
CCD camera). The unit cell dimensions were
found to be a = b = 95 ± 1 Å and g = 90 ± 1°
(n = 25). Surface relief reconstructions from
such images of unidirectionally metal-
shadowed AqpZ crystals had a resolution of
12 Å (Figure 2b, Fuchs et al., 1995; Kistler et
al., 1977; Guckenberger, 1985). The average
surface relief obtained from the metal-
shadowed specimen (Figure 2b) is consistent
with the average resulting from high
resolution AFM topographs (Figure 2a;
Scheuring et al., 1999) in that both exhibit
one strongly and one weakly protruding
tetramer. The two averages also correlate
favorably in the topology along the periphery Figure 3 . Overlay of AqpZ AFM topography
of the higher extracellular surface, the strong recorded in buffer solution on projection maps of
AqpZ and AQP1, both rendered at 7Å resolution (full
indentation in the center of the cytoplasmic frame sizes 95 Å). a) AFM topography of the
surface, and the overall particle organization. extracellular surface of AqpZ with outlined and
annoted protrusions and overall shape (see
However the inner ring of protruding Discussion; Scheuring et al. 1999). b) AFM
structures on the extracellular surface is topograph of the cytoplasmic surface of AqpZ with
outlined protrusions housing non membrane buried
oriented differently in the AFM topograph parts of loop B (see Discussion; Scheuring et al.
and the outer ring protrusions of the 1999). c) Projection map of AqpZ (Ringler et al.
1999) viewed from the extracellular side. Outlines of
cytoplasmic surface of the reconstructed relief corresponding surface protrusions are overlaid: Loop
are more pronounced than the features C is at the periphery and spans a major part of a
monomer. d) Projection map of AqpZ (Ringler et al.
determined with the AFM. 1999) viewed from the cytoplasmic side. Outlines of
A projection map was calculated to 12 Å corresponding surface protrusions are overlaid: non-
membrane-buried parts of loop B (comprising the
resolution from the unshadowed freeze-dried NPA motif, which participates in the water pore)
crystal areas (Figure 2c). This map is very cross the center of one monomer. e) Projection map
of AQP1 (Walz et al. 1995) viewed from the
similar to the cryo electron microscopy extracellular side after an applied clockwise rotation of
projection map with a resolution of 7 Å 15°. Protrusions are outlined as in Figure 3c. f)
Projection map of AQP1 (Walz et al. 1995) viewed
(Figure 2d, Ringler et al. 1999). The tetramers from the cytoplasmic side after an applied
show an inner ring of densities close to their counterclockwise rotation of 15°. Protrusions are
outlined as in Figure 3d.
four-fold symmetry center and an outer ring The maximum rotational alignment cross correlation
of densities along their periphery. The coefficient was obtained when the lowest density
between monomers (indicated by the asterisk) and of
monomers can be distinguished and the asymmetric densities within the monomer (indicated
opposite rotation of adjacent tetramers with by 1 and 2) were superimposed.
respect to the lattice lines leading to rhombus-
shaped lipid interspaces is evident. These features allow the unambiguous alignment of

Figure 4. Surface rendered 3D electron density map of AQP1 at 6 Å resolution (Walz et al. 1997). The right
handed bundle of the 6 transmembrane spanning a-helices is aligned according to the projection maps in Figures 3e
and f. Superposed contours indicate the topographical features of AqpZ 2D crystals measured by AFM in buffer
solution. Helix assignment as proposed by Heymann & Engel (2000). a) View from the extracellular side, with
assigned loops indicated. The small protrusion on the periphery might be part of loop C which connects the end of
helix 3 with the beginning of helix 4 (see Discussion). b) View from the cytoplasmic side. The AFM topograph
shows one protrusion corresponding to parts of loop B which spans the water pore (see Discussion).

the two maps (Figure 2c and 2d). topography contours.
The extracellular and cytoplasmic Figures 4a and 4b display perspective
topologies revealed by AFM experiments views of the 3D map of AQP1 at 6Å
(Scheuring et al 1999) are shown in Figures resolution from the extracellular and the
3a and 3b, respectively. Corresponding cytoplasmic side, respectively. a-helices are
regions of the projection map of AqpZ are represented as gray sausages, the bright parts
shown aligned in Figures 3c and 3d, of which display the ends facing the viewer.
respectively. Finally, Figures 3e and 3f show The map is overlaid by contours of
the projection map of AQP1 aligned with protrusions imaged by AFM under native
respect to that of AqpZ (Figures 3c and 3d). conditions (see Figures 3a and 3b). Flexible
This alignment is compatible with that parts of proteins, such as loops, are mostly
reported by Ringler et al. (1999), but it has averaged out, hence the AFM topography
been improved by considering the 1 % implements additional information on the
difference of the unit cell size. This organization of the non-membrane-stabilized
improvement yielded a correlation coefficient protein structure to the 3D density map.
of 78 % after a clockwise rotation of the
AQP1 tetramer (Figure 3e) by 15°. If, 3.1.4. Discussion
however the AQP1 tetramer shown in Figure
3f is rotationally aligned with AqpZ in Figure AFM investigations of densely packed or
3c, the correlation coefficient is 65 %, regularly arranged membrane protein layers
allowing the unambiguous assignment of the can provide information on their sidedness in
AQP1 projection map in Figure 3e as conjunction with either proteolytic cleavage of
extracellular. To relate protrusions observed a terminal domain (Scheuring et al., 1999), or
by AFM with loops or protruding helices, the specific binding of antibodies (Müller et al.,
projection maps are overlaid by the 1996). The sidedness of AqpZ surfaces has

been unambiguously defined by imaging general similarity of projection maps from
crystals before and after proteolytic cleavage different aquaporins (Engel et al. 2000;
of the cytoplasmic N-terminus identifying the Daniels et al. 1999) this suggests a conserved
7 Å high crown-like surface as extracellular helical packing arrangement. From helical
(Figure 2a; Scheuring et al., 1999). periodicity analysis Heymann & Engel
To link topographical data obtained by (2000) proposed a helix assignment that is
AFM with the projection structure acquired indicated in Figure 4. This assignment is
by electron crystallography, we have analyzed consistent with the findings from fitting
freeze-dried crystals that were partially metal- helical stretches to elongated structures in the
shadowed. As illustrated in Figure 1, surface 4.5 Å 3D density map (de Groot et al. 2000).
reliefs calculated from metal coated areas Therefore, the contoured protrusions of AqpZ
were thus in register with projection maps overlayed on the 3D map of AQP1 are highly
from uncoated areas of the same crystal, relevant. They provide a solid basis to select
allowing the unambiguous assignment of the most likely helix assignment from the two
topography and projection map. This novel possibilities given by de Groot et al. (2000)
approach is of particular advantage for and Heymann & Engel (2000). Figure 4a
aquaporins whose quasi two-fold symmetry indicates that the peripheral protrusion C is
makes the assignment of the sidedness likely to connect helices 3 and 4. This is
difficult (Mitsuoka et al., 1999). compatible with the assignment of this
The experiment described here can be protrusion to loop C based on volume
applied in general provided that large, calculations and flexibility mapping by AFM
coherent 2D crystals are available. It has been (Scheuring et al. 1999). The previously
designed to solve the discrepancy between the unassigned peripheral protrusion on the
sidedness assignment of AQP1 by Walz et al. extracellular surface appears to be the C-
(1996) and the 4.5 Å 3D density map by terminal end of helix 3. Thus, the other
Mitsuoka et al., (1999). The regions assigned protrusion, now labeled A, must represent
to loop B and E in the latter map suggest an loop A, in agreement with the helical
opposite sidedness than that proposed by assignment by Heymann & Engel (2000) and
Walz et al. (1996). Since AFM experiments de Groot et al. (2000). Since loop D is short
allowed the sidedness of AqpZ topographies in all aquaporins and the N- and C-termini are
to be firmly assigned (Scheuring et al. 1999), almost completely removed by trypsin
and because AqpZ projection maps could be digestion of AqpZ 2D crystals (Scheuring et
aligned with those of AQP1 (Figure 3; al. 1999), the protrusion contoured in Figure
Ringler et al. 1999), establishing the link 4b is likely to present the surface of loop B
between relief reconstruction and projection that folds back into the membrane and
map appeared to be a straight-forward connects helices 2 and 3. Position and shape
approach to settle this pertinent question. The of this protrusions further confirm the helical
experimental results documented in Figures 1 assignment shown in Figure 4.
and 2 provide a solid basis to align the Taken together, compelling evidence has
surface topography of AqpZ recorded by been accumulated to justify revision of the
AFM to 7 Å resolution with the projection sidedness assignment proposed by Walz et
map of AQP1 (Figure 3) and hence with the al. (1996). The pertinent question arises as to
3D map of AQP1 (Figure 4). how apparently solid data could have been
A recent sequence alignment study of 160 misinterpreted. A possible explanation is
aquaporin sequences revealed highly related to the observation that carboxy
conserved residues within all helical segments peptidase Y treatment in solution tends to
(Heymann & Engel 2000). Together with the produce disordered crystals and aggregates

(B. Heymann, D. Fotiadis, D. J. Müller, (w/v) acrylamide gels (data not shown, for
unpublished results). Therefore, the few details see Scheuring et al. 1999).
crystals found by Walz et al. (1996) after
decarboxylation may not have been properly Atomic force microscopy
digested. The structural differences however,
may have resulted from lattice disorder and AqpZ 2D crystals were deposited onto
surface contamination by proteolytic freshly cleaved muscovite mica (from Mica
fragments. New York Corp., New York, USA) and
The alignment of the AqpZ topography imaged in buffer solution (10 mM Tris-HCl,
with the AQP1 3D map corroborates the pH 7.2, 150 mM NaCl) at high resolution
topology of AQP1 derived from fitting helical (for details see Scheuring et al. 1999).
stretches using the program ROTTRANS (de
Groot et al. 2000) to the 4.5 Å 3D density Freeze-drying & metal-shadowing
map (Mitsuoka et al. 1999) and the helix
assignment from sequence analysis AqpZ crystals were adsorbed (2 min.) to
(Heymann & Engel 2000). With the glow-discharged (1 min.) carbon-coated 400
sidedness issue resolved, the overall mesh grids. These were washed twice with
architecture of AQP1 is now established and double-distilled water, blotted and plunged
will help tracing the polypeptide in higher into liquid nitrogen. The grids were then
resolution electron crystallographic analysis. freeze-dried and metal-shadowed in the
MIDILAB (Gross et al., 1990), as detailed in
3.1.5. Materials and Methods Walz et al. (1996). After shadowing, grids
were transferred to a specially designed Gatan 2D crystallization cryo holder, and examined. Images were
digitally recorded with a Gatan-694 slow scan
2D crystals were produced by dialysis of CCD camera with a maximal image size of
solubilized AqpZ (1 mg / ml, in 2% octyl-b- 10242 pixels. Images were correlation
D-glucopyranoside (OG)) mixed with 1- averaged using the SEMPER image
palmitoyl-2-oleoyl-sn-glycero-3- processing package (Saxton et al., 1979),
phosphocholine (POPC) / 1,2-dimiristoyl-sn- while surface reconstructions of metal-
glycero-3-phosphocholine (DMPC) 1/1 shadowed crystals were calculated in
(from Avanti Polar Lipids, Inc., USA) in 2% SEMPER or MILAN (Fuchs et al., 1995)
OG (from Anatrace, Inc., USA) at a LPR of image processing packages. The handedness
0.4 against a detergent free buffer (10 mM - determined by the sample orientation in the
citrate, pH 6, 200 mM NaCl, 100 mM MgCl2, microscope, the image acquisition with the
10% glycerol, 0.005% NaN3) (Ringler et al., CCD camera, and the data transfer to various
1999). Crystals were harvested after three processing systems was carefully controlled
days. by coadsorption of amyloid fibrils. Trypsin digestion Cryo electron microscopy

For cleavage of the N-terminal fragment, AqpZ 2D crystals mixed with 3 - 10%
AqpZ-10his crystals were incubated overnight trehalose were adsorbed to carbon-coated
at 4°C with trypsin (1 mg/ml). After trypsin copper electron microscopy grids. The grids
treatment samples were investigated by were blotted and plunged into liquid ethane.
sodium dodecyl sulfate-polyacrylamide gel Vitrified specimens were recorded in a
electrophoresis (SDS-PAGE) using 10% Hitachi H8000 electron microscope with a

LaB6 filament, operated at 200 kV under low Cheng, A., van Hoek, A. N., Yeager, M.,
dose conditions (~ 5 e- / Å2) and processed Verkman, A. S. & Mitra, A. K. (1997).
Three-dimensional organization of a
using the MRC image processing package human water channel. Nature 387, 627-
(for details see Ringler at al 1999). 630.

3.1.6. Acknowledgment Chrispeels, M. & Agre, P. (1994).
Aquaporins: water channel proteins of
plant and animal cells. Tr. Biochemic. Sci.
We thank Dr. Claire Goldsbury who 19, 421-425.
provided the amylin fibers coadsorbed with Daniels, M. J., Chrispeels, M. J. & Yeager,
the AqpZ crystals to overcome sidedness M. (1999). Projection structure of a plant
vacuole membrane aquaporin by electron
problems. We acknowledge fruitful cryo-crystallography. J. Mol. Biol. 294,
discussions with Dr. S. A. Müller who helped 1337-1349
in assembling the manuscript. This work was
supported by the Swiss National Foundation De Groot, B. L., Heymann, J. B., Engel, A.,
Mitsuoka, K., Fujiyoshi, Y. & Grubmüller,
for Scientific Research (grant 4036 - 44062 H. (2000). The fold of human aquaporin
to A. E.), the Swiss Priority Project for Micro 1. . J. Mol. Biol. submitted.
and Nano System Technology (MINAST),
the Maurice E. Müller Foundation of Fotiadis, D., Hasler, L., Müller, D.J.,
Stahlberg, H., Kistler, J. & Engel A.
Switzerland and the National Institutes of (2000). The surface topography of lens
Health (to P. A.). MIP supports dual functions. . J. Mol.
Biol. submitted
3.1.7. References
Fuchs, K., Tittmann, P., Krusche, K. &
Gross, H. (1995). Reconstruction and
Agre, P., Preston, G., Smith, B., Jung, J., representation of surface data from two-
Raina, S., Moon, C., Guggino, W. & dimensional crystalline, biological
Nielsen, S. (1993). Aquaporin CHIP: the macromolecules. Bioimaging 3, 12-24.
archetypal molecular water channel. Am. J.
Physiol. 265, F436-476. Gorin, M. B., Yancey, S. B., Cline, J., Revel,
J.-P. & Horwitz, J. (1984). The major
Borgnia, M., Kozono, D., Calamita, G., intrinsic protein (MIP) of the bovine lens
Maloney, P. C. & Agre, P. (1999). fiber membrane: Characterization and
Functional reconstitution and structure based on cDNA cloning. Cell 39,
characterization of AqpZ, the E. coli water 49-59.
channel protein. J. Mol. Biol. 29, 1169-
1179. Guckenberger, R. (1985). Surface relief
derived from heavy-metal-shadowed
Calamita, G., Bishai, W., Preston, G., specimen-Fourier space techniques
Guggino, W. & Agre, P. (1995). applied to periodic objects.
Molecular cloning and characterization of Ultramicroscopy 16, 357-370.
AqpZ, a waterchannel from Escherichia
coli. J. Biol. Chem. 270, 29063-29066. Heymann, J. B. & Engel, A. (2000).
Structural clues in the sequences of the
Calamita, G., Kempf, B., Bonhivers, M., aquaporins. J. Mol. Biol. 295, 1039-1053.
Bishai, W. R., Bremer, E. & Agre, P.
(1998). Regulation of the Escherichia coli Heymann, J. B., Müller, D. J., Landau, E. M.,
water channel gene AqpZ. Proc. Natl. Rosenbusch, J. P., Pebay-Peyroula, E.,
Acad. Sci. USA 95, 3627-3631. Büldt, G. & Engel, A. (1999). Charting the
surfaces of purple membrane. J. Struct.
Chandy, G., Zampighi, G. A., Kreman, M. & Biol. 128, 243-249.
Hall, J. E. (1997). Comparison of the
water transporting properties of MIP and Jung, J., Preston, G., Smith, B., Guggino, W.
AQP1. J. Membr. Biol. 159, 29-39. & Agre, P. (1994). Molecular structure of
the water channel through aquaporin

CHIP. The hourglass model. J. Biol. Schabert, F. A., Henn, C. & Engel, A. (1995).
Chem. 269, 14648-14654. Native Escherichia coli OmpF porin
surfaces probed by atomic force
Kistler, J., Aebi, U. & Kellenberger, E. microscopy. Science 268, 92-94.
(1977). Freeze-drying and shadowing a
two-dimensional periodic specimen. J. Scheuring, S., Ringler, P., Borgnia, M.,
Ultrastr. Res. 59, 76-86. Stahlberg, H., Agre, P. & Engel, A.
(1999). High resolution AFM topographs
Li, H., Lee, S. & Jap, B. K. (1997). Molecular of the Escherichia Coli waterchannel
design of aquaporin-1 water channel as aquaporin Z. EMBO J. 18, 4981-4987.
revealed by electron crystallography.
Nature Struct. Biol. 4, 263-265. Sidel, V. W. & Solomon, A. K. (1957).
Entrance of water into human red cells
Mitsuoka, K., Murata, K., Walz, T., Hirai, T., under an osmotic pressure gradient. J.
Agre, P., Heymann, J. B., Engel, A. & Gen. Phys. 41, 243-257.
Fujiyoshi, Y. (1999). The structure of
aquaporin-1 at 4.5-Å resolution reveals Walz, T., Hirai, T., Murata, K., Heymann, J.
short a-helices in the center of the B., Mitsuoka, A., Fujiyoshi, Y., Smith, B.
monomer. J. Struct. Biol. 128, 34-43. L., Agre, P. & Engel, A. (1997). The 6 Å
three-dimensional structure of aquaporin-
Mulders, S., Preston, G., Deen, P., Guggino, 1. Nature 387, 624-627.
W., van, O. C. & Agre, P. (1995). Water
channel properties of major intrinsic Walz, T., Smith, B., Agre, P. & Engel, A.
protein of lens. J. Biol. Chem. 270, 9010- (1994a). The three-dimensional structure
9016. of human erythrocyte aquaporin CHIP.
EMBO J. 13, 2985-2993.
Müller, D. J. & Engel, A (1999). pH and
voltage induced structural changes of Walz, T., Smith, B., Zeidel, M., Engel, A. &
porin OmpF explain channel closure. J. Agre, P. (1994b). Biologically active two-
Mol. Biol. 285, 1347-1451. dimensional crystals of aquaporin CHIP.
J. Biol. Chem. 269, 1583-1586.
Müller, D. J., Schoenenberger, C. A., Büldt,
G. & Engel, A. (1996). Immuno-atomic Walz, T., Tittmann, P., Fuchs, K. H., Müller,
force microscopy of purple membrane. D. J., Smith, B. L., Agre, P., Gross, H. &
Biophys. J. 70, 1796-1802. Engel, A. (1996). Surface topographies at
subnanometer resolution reveal asymmetry
Preston, G. M. & Agre, P. (1991). Isolation and sidedness of aquaporin-1. J. Mol.
of the cDNA for erythrocyte integral Biol. 264, 907-918.
membrane protein of 28 kilodaltons:
Member of an ancient channel family. Walz, T., Typke, D., Smith, B. L., Agre, P., &
Proc. Natl. Acad. Sci. USA 88, 11110- Engel, A. (1995). Projection map of
11114. aquaporin-1 determined by electron
crystallography. Nature Struct. Biol. 2,
Preston, G. M., Carroll, T. P., Guggino, W. 730-732.
B. & Agre, P. (1992). Appearance of
water channels in Xenopus Oocytes Zeidel, M. L., Ambudkar, S. V., Smith, B. L.
expressing red cell CHIP28 protein. & Agre, P. (1992). Reconstitution of
Science 256, 385-387. functional water channels in liposomes
containing purified red cell CHIP28
Ringler, P., Borgnia, M., Stahlberg, H., Agre, protein. Biochem. 31, 7436-7440.
P. & Engel, A. (1999). Structure of the
water channel AqpZ from Escherichia coli
revealed by electron crystallography. J.
Mol. Biol. 291, 1181-1190.

4. Structural studies of a
membrane transporter

4. Structural studies of a membrane transporter

4.1. The functional Escherichia coli lactose permease LacY/Cytb562/6His
forms trimers: A 2.8 nm 3D reconstruction and preliminary electron
crystallographic data

4.1.1. Summary large family of secondary transport proteins
from archeae to the mammalian central
A chimeric protein consisting of lactose nervous system that convert free energy
permease with cytochrome b562 in the middle stored in an electrochemical ion gradient into
cytoplasmic loop and 6 His residues at the C a concentration gradient (reviewed in Kaback,
terminus (LacY/L6cytb562/417H6; “red 1976; Kaback, 1983; Poolman & Konings,
permease”) was overexpressed in 1993). The polytopic, hydrophobic membrane
Escherichia coli, and isolated by nickel protein, LacY, has been solubilized, purified to
affinity chromatography after solubilization homogeneity, reconstituted into
with dodecyl-ß,D-maltopyranoside or decyl- proteoliposomes (Viitanen et al., 1986), and
ß,D-maltopyranoside. Red permease was then found to be solely responsible in its
either reconstituted in the presence of monomeric form for ß-galactoside transport
phospholipids yielding densely packed (Sahin-Tóth et al., 1994). All available
vesicles and well-ordered two-dimensional evidence indicates that the protein is
(2D) crystals, or examined as single particles. composed of 12 α-helical membrane-
Single particle analysis of 1188 trimeric spanning segments connected by hydrophilic
particles uniformly adsorbed to the electron loops with the N and C termini on the
microscopy grid reveals a ~ 5.5 nm x ~ 7.7 cytoplasmic face (reviewed in (Kaback, 1996;
nm bean-shaped protein with a central stain- Kaback & Wu, 1997). Hydropathy profiling
filled indentation, at a resolution of ~ 2 nm. A suggests that many membrane transporters
three dimensional (3D) reconstruction could have similar secondary structures, indicating
be calculated at ~ 2.8 nm resolution from that both the basic tertiary structure and
tilted images, showing that the cytochrome mechanism of action of these enzymes have
b562 induces trimerization. Reconstituted probably been conserved. Therefore, studies
samples yielded densely packed vesicles or on bacterial transport proteins, which are
trigonal crystals of trimeric LacY. The best considerably easier to manipulate than their
packed crystals showed diffraction spots up eukaryotic counterparts, have important
to a resolution of 8 Å, however these crystals relevance to transporters in higher order
were multilayered and separation of the layers systems.
by Fourier peak filtering was not possible. Site-directed and Cys-scanning
mutagenesis (Frillingos et al., 1998a;
4.1.2. Introduction Frillingos et al., 1998b) demonstrate that only
6 out of the 417 amino acid residues in
The lactose (lac) permease of Escherichia lactose permease are irreplaceable with
coli catalyzes the coupled stoichiometric respect to active transport. On the other hand,
translocation of ß-galactosides and H+ across site-directed fluorescence and chemical
the cytoplasmic membrane. Encoded by the labeling indicate that widespread
lacY gene, the permease is a paradigm for a conformational changes occur during enzyme

turnover (reviewed in (Frillingos et al., 1998b; direct means (i.e., helix packing) will provide
Kaback & Wu, 1997). About 10% of the important additional insight. Little is known
active Cys-replacement mutants are altered by about the helix tilts, for example. The most
alkylation, and these Cys residues cluster on severe limitation to obtaining structural
helical faces. Moreover, monoclonal antibody information on LacY is the lack of three-
(mAb) 4B1 which binds to the periplasmic dimensional (3D) crystals. Attempts to
loop connecting helices VII and VIII (Sun et produce highly ordered two-dimensional
al., 1996) specifically blocks all reactions (2D) crystals suitable for electron
catalyzed by the permease that involve net H+ crystallography have been reported for LacY
translocation (Carrasco et al., 1984; Frillingos (Li & Tooth, 1987) and for melibiose
& Kaback, 1996). Taken together, the results permease (Rigaud et al., 1997), but these
suggest that the permease is comprised of a crystals were not of sufficient quality for
bundle of loosely packed rigid bodies (i.e., high-resolution structural analysis. Because
transmembrane helices) with few residues that LacY is extremely hydrophobic and probably
are essential for the mechanism and specific highly flexible outside of the bilayer (Le
surface contours that permit sliding motions Coutre et al., 1997), it is notoriously difficult
to occur during transport. Based on a variety to crystallize. To generate a higher proportion
of site-directed approaches which include of polar surface area, a fusion between LacY
second-site suppressor analysis and and cytochrome b562 (red permease) has
site-directed mutagenesis, excimer been constructed and shown to have native
fluorescence, engineered divalent metal lactose transport activity (Privé et al., 1994).
binding sites, chemical
cleavage, electron
paramagnetic resonance,
thiol cross linking and
identification of
discontinuous mAb
epitopes, a helix
packing model has been
formulated (reviewed in
(Frillingos et al., 1998b;
Kaback et al., 1997;
Kaback & Wu, 1997).
Although site-
directed structural
approaches and ex-
tensive mutational ana-
lysis have led to a
proposed helix packing
model for LacY and
mechanism for coupling
between substrate and
H+ translocation (Fril-
lingos et al., 1998b;
Kaback, 1997), a
relatively low-resolution F i g u r e 1 . a) Negative stain electron micrograph of LacY trimers uniformly
adsorbed to the electron microscopy grid (scale bar: 50nm). b) Average of 1188
structure obtained by rotationally aligned, end-on views (scale bar: 5nm).

Zhuang et al. (1999)
found conditions for
reconstitution and 2D
crystallization of red
permease obtained by
overexpression in E.
coli. Single particle
analysis of densely
packed LacY mono-
mers and reconstituted
trimers were reported
to have an appro-
ximately trapeziform
shape 5.4 nm long and
4.1 nm wide on one
side and 5.1 nm wide
on the other with
peripheral protrusions
and a central stain-
filled depression
(Zhuang et al., 1998).
Here we present a
3D reconstruction Figure 2. a) Micrograph of negatively stained of LacY trimers and b) micrograph
calculated from tilted of the same region after tilting the grid to 60°. The particles numbered 1, 2 and 3 in
images of solubilized both projections document how the tilt parameters were determined.
LacY (red permease) trimers at 2.8 nm were harvested at an OD 600 of 2. Crude
resolution. The protein has a bean-like shape membranes containing overexpressed red
with two peripheral density maxima. permease were collected after rupturing the
Sideviews showed that the cytochrome b562 cells in a French press and washed with 6%
emerging from loop 6 of each monomer sodium cholate to remove adsorbed
induces trimerization. Furthermore, densely impurities. LacY/L6cyt b562/417H6 was
packed vesicles and 2D crystals were grown, efficiently extracted by solubilizing the
also revealing trimeric organization of the cholate-washed membranes with 3% dodecyl-
particles. However, the best crystals had a β,D-maltopyranoside (DDM) or 3% decyl-
strong tendency to stack. Diffraction patterns β,D-maltopyranoside (DM) at pH 7.5. The
of trehalose embedded crystals showed spots protein was subsequently purified to near
to a resolution < 8 Å, but the layers could not homogeneity in a single step by metal chelate
be separated by Fourier peak filtering, affinity chromatography on a Ni-NTA
implying that the stacking of the crystalline column. During this step, the DDM or DM
layers is in register. concentration was adjusted to 0.01% or 0.1%
respectively. The fractions containing red
4.1.3. Results and discussion permease were directly used for single
particle analysis or reconstitution Protein purification experiments.

After induction with 0.3 mM IPTG, cells

89 Single particle analysis respect to the electron beam (Fig 2a and b).
Images were recorded at 40'000 x
Red permease solubilized with DDM were magnification to acquire large areas for
directly investigated using negative stain particle alignment. Particles numbered 1, 2
electron microscopy. An image at 50'000 x and 3 indicate 3 corresponding LacY trimers
magnification taken from an untilted in the untilted and the tilted image. The
specimen is displayed in Fig. 1a. The circle rotational orientation of the particles was
indicates a lacY trimer adsorbed to the grid defined in the non-tilted images and
end-on. Such particles were automatically subsequently used for the tilted projection.
selected (see Materials and Methods) and Figure 3 shows the 3D reconstruction at ~
averaged after alignment, resulting in a 2.8nm resolution comprising 850 projection
projection map at ~ 2nm resolution (Fig. 1b). pairs from 8 image pairs as displayed in
The trimer clearly consists of three bean- figure 2. The two end-on views (images 4 and
shaped proteins with dimensions of ~ 5.5 nm 10) are markedly different. One side is
x ~ 7.7 nm each and 2 density maxima capped at the center of the trimer (4), the other
towards the two ends of the monomers. The side shows a strong indentation in the center
map suggests an indentation in the center of of the trimer (10). This suggests that the
each molecule (Fig. 1b). capped side corresponds to the cytoplasmic
To calculate a 3D density map the surface where the three large loops containing
specimen was imaged at 0° and 60° tilt with cytochromes b562 interact to induce trimer

1 2 3 4

5 6 7 8

9 10 11 12

F i g u r e 3 . Surface rendered 3D reconstruction of LacY trimers at a resolution of ~2.8nm calculated from image
pairs like that shown in figure 2. The image top left is a sideview, the following 11 images show the trimer rotated
by 30° around the axis indicated. The cap structure on one side is supposed to arise from the three cytochromes b562
in loop 6 which induce trimerization through hydrophilic interactions (scale bar: 5nm).

formation. Correspondingly, the other surface crystalline layers. The best reconstitution
represents the periplasmic side where only buffer solution contained 10 mM Tris (pH
little surface protruding loops are to be 7.4), 150 mM NaCl, 0.05% NaN3 and 40
expected. mM Mg2+ , corroborating conditions
established by Zhuang et al. (1999). High Reconstitution and 2D protein concentrations (at least 1 mg/ml) were
crystallization a prerequisite for assembling densely packed
vesicles (Fig. 4). In addition, the presence of
Different reconstitution strategies were 1-palmitoyl 2-oleolyl phosphatidylcholine
tested. No matter whether dodecyl ß,D- (POPC) (transition temperature 3 °C) was
maltopyranoside (DDM; critical micelle essential for efficient integration of the
concentration, cmc = 0.007%) or decyl ß,D- protein into the bilayer because of the low
maltopyranoside (DM; cmc = 0.07%) was temperature (12 °C) during the initial stages
used to solubilize the protein, the lipid was of the dialysis. The best crystals were grown
prepared in octyl ß,D-glucopyranoside (OG; using a mixture of POPC and DMPC of 1/1
cmc = 0.6%) or a mixture of OG and DM for (w/w).
the dialyses which were carried out at a lipid- Densely packed vesicles were found
to-protein ratio (LPR) ranging from 0.5 to 1.5 containing trimeric LacY in pseudocrystalline
(w/w). order. The appearance of the particles in the
As previously described, the temperature raw data micrograph (Fig. 4a) and the average
profile for one dialysis cycle profile was produced after single particle alignment (Fig.
used: 12 °C (12 h), 12-
37 °C (24 h), 37 °C (24
h), 37 °C-12 °C (12 h).
While DM was
removed in one dialysis
cycle, two cycles were
required to remove
DDM. Various divalent
cations were added to
the dialysis buffer, but
in most cases this led to
protein aggregation.
Trying to trap LacY in
one conformation by
the addition of lactose
or beta – D – galacto-
pyranosyl – 1 – thio –
beta – D – galacto –
pyranoside (TDG) to
the dialysis buffer did
not improve crystal-
linity. Similarly, neither
glycerol nor DDT
helped to improve
crystallinity or to pre- F i g u r e 4 . a) Negative stain micrograph of densely packed, pseudocrystalline
LacY. The trimeric organization is directly visible (scale bar: 50nm). b) Average of
vent stacking of the 571 rotationally aligned particles from image a) (scale bar: 5nm).

Figure 5. a) Electron micrograph of negatively stained LacY 2D crystals. The stacking of the crystalline layers
leads to a contrast variation depending on whether the layers are in register (center) or out of register (left edge),
respectively (scale bar: 50nm). b) Calculated diffraction pattern of image a). Diffraction spots are sharp and 3rd order
spots are visible corresponding to a resolution of ~2nm. c) Cross correlation average of the right part of the crystal
in image a). The trimeric organization is resolved, handedness information is lost, probably due to overlay of trimers
in both orientations (scale bar: 5nm).

4b) are in very good agreement with the methods document that the crystalline areas
structure of the solubilized protein (Fig. 1). are fairly well ordered. In negatively stained
Again, the monomers are bean-shaped with samples trigonal lattices with dimensions of a
two density maxima at either end. The = b = 9.8 nm, similar to those reported by
monomer reveals a handedness showing a Zhuang et al. (1999) (a = b = 10.4 nm), were
stronger density in the clockwise direction found (Fig. 5). Calculated diffraction patterns
with respect to the trimer symmetry axis, revealed spots up to the 3rd order (Fig. 5b).
consistent with the density organization The average (Fig. 5c) calculated from the
revealed by single particle analysis of the image displayed in figure 5a again shows
solubilized proteins (Fig. 1b). However, the clearly a trimeric organization of the protein.
particle dimensions, ~ 4.5 nm x ~ 6.5 nm are However no handedness can be detected in
smaller than found for the solubilized protein. such maps. This is probably a consequence
There are two explanations for this of stacking of crystalline layers with proteins
discrepancy: Firstly, the solubilized particles integrated in opposite orientations with
are surrounded by detergent molecules, which respect to the membrane planes. As a result,
increase their apparent size. Secondly, only the contrast of such crystals is strong, but
membrane protruding domains of trimers handedness information is lost and the
incorporated in a membrane are stained. particles appear small ~ 4.0 nm x ~ 5.2 nm.
Well ordered crystals were imaged using This explanation is supported by the fact that
negative stain (Fig. 5) or trehalose embedded the contrast of the array displayed in figure 5a
cryo electron microscopy (Fig. 6). Both seems to fade towards the left side of the

Figure 6. a) Cryo electron micrograph of trehalose embedded LacY 2D crystals. The stacking of crystalline layers
induce large distance modulations (scale bar: 50nm) b) Calculated diffraction pattern of image a). Diffraction spots
corresponding to a resolution of ~8Å are visible (indicated by circle).

image; where the stacking of the layers is not Mg2+ and DDM at a concentration close to
in perfect register, inducing a contrast the cmc or into lipid bilayers. At a LPR close
modulation. to 1, these trimers associate into highly
Analogous to the negative stain electron ordered 2D crystal stacks.
microscopy studies, cryo-electron The recently found stability of LacY in
micrographs of trehalose embedded LacY DM allows detergent to be removed both
crystals showed strong long range efficiently and rapidly by dialysis. This
modulations (Moiré fringes), due to their possibility allows a new dialysis conditions to
stacking arrangement (Fig. 6a). Calculated be explored using the dialysis machine or
diffraction patterns (Fig. 6b) show spots at ~ buttons.
8Å resolution (indicated by the circle). The already established crystallization
Because of the in-register-stacking of the protocol (Zhuang et al., 1998) using DDM
crystalline layers, information from the solubilizing LacY offers further
individual crystal layers could not be improvements. In particular, dialysis buffers
separated by Fourier filtering. However, the which mediate destacking of the crystalline
diffraction pattern proves that LacY can be multi layers (Fig 5, 6) should be screened.
integrated into highly ordered crystals Other 2D crystallization techniques must
suitable for high resolution cryo-electron be explored as well, such as controlled
microscopy. dilution of protein-lipid-detergent mixtures
(Remigy, personal communication),
4.1.4. Perspectives crystallization at a functionallized lipid
monolayer (Levy et al. 1999) or the
To obtain further structural information on withdrawal of detergents using biobeads as
the membrane transporter LacY, highly described in Rigaud et al 1998.
ordered 2D crystals must be grown. The
results of the present study support the idea
that the functional LacY/cyt b562/417H6
forms stable trimers in buffers containing

4.1.5. Material and Methods 4h, washed in 6% sodium cholate (Viitanen et
al., 1986), and solubilized with 2% dodecyl- Materials ß,D-maltopyranoside (DDM) or 2% decyl-
ß,D-maltopyranoside (DM). The
Isopropyl-ß,D-thiogalactopyranoside solubilization buffer contained 1M NaCl
(IPTG) was from Boehringer Mannheim and/or 1mM imidazol or 1mM histidine to
GmbH, Germany. Sodium cholate was from prevent the unspecific binding of
Fluka, Switzerland., dodecyl-ß,D-malto- contaminants to the columns. Solubilized red
pyranoside (DDM), decyl-ß,D-malto- permease was purified by nickel affinity
pyranoside (DM) and octyl-ß,D-gluco- chromatography (Loddenkotter et al., 1993).
pyranoside (OG) were from Anatrace, Briefly, the 2% DDM extract from 10 g wet
Switzerland. 1,2-dimyristoyl-sn-glycero-3- weight of cells was added to 1.5 ml of the Ni-
phosphocholine (DMPC) and 1-palmitoyl-2- NTA-agarose slurry in the column, and
oleoyl-sn-glycero-3-phosphocholine (POPC) rotated at 4 °C overnight. The column was
were from Avanti Polar Lipids, Inc., USA. then washed with 50 mM potassium
Nickel-nitrilotriacetic acid (Ni-NTA) resin phosphate (pH 7.5), 200 mM NaCl, 0.01%
was from QIAGEN AG, Switzerland. All DDM or 0.1% DM respectively until the
other chemicals used were of analytical grade absorption baseline was reached. Then the
and purchased from Merck (Schweiz) AG, column was washed with 50 mM potassium
Dietikon, Switzerland. phosphate (pH 7.5), 200 mM NaCl, 0.01%
DDM or 0.1% DM, 10mM Imidazol or Protein Expression and 10mM histidine. Once the baseline had again
Purification been reached, the protein was eluted with the
same buffer containing either 300mM of
Expression of the fusion protein between imidazol or 150mM histidine. Purified
lactose permease and cytochrome b562 has samples were analyzed by sodium dodecyl
been described (Privé et al., 1994). Briefly, sulfate/polyacrylamide gel electrophoresis
the cytochrome b 562 gene was inserted into (SDS-PAGE) and visualized by Coomassie
the XhoI restriction site in the region blue and silver staining. Protein fractions
encoding the middle cytoplasmic loop (L6). contained 2-5 mg/ml.
The cassette version of the lacY gene used
was under control of the lac Reconstitution
promoter/operator. To facilitate purification, 6
consecutive His residues were engineered to Purified red permease solubilized in
the C-terminus yielding the 0.01% DDM or 0.1% DM was mixed with
LacY/L6cytb562/417H6, referred to as "red various amounts of detergent solubilized
permease". E. coli XL-Blue cells transformed phospholipids. The lipid:protein ratios (LPR,
with a plasmid encoding the fusion protein w/w) varied between 0.5 and 1.5. The lipids
were cultivated in LB broth at 37 °C and the used were 1:1 (w/w) mixtures of
growth rate was monitored by optical density DMPC:POPC. They were solubilized in 2%
(OD) measurements at 600 nm. When an OG or a 1:1 (v/v) mixture of 2% OG and
OD 600 of 0.8 was achieved, the cells were 0.2% DM. The protein-lipid mixtures were
induced with 0.3 mM IPTG. Cells were incubated for 6 hours at room temperature,
harvested by centrifugation when OD600 of and then dialyzed against detergent-free
about 2 was reached, and were then ruptured buffer in a temperature controlled dialysis
in a French press cell. Membranes were apparatus (Jap et al., 1992). The best results
collected by centrifugation at 250,000 g for were obtained with 10 mM Tris (pH 7.5), 150

mM NaCl, 0.005% NaN3 and 40 mM correlation peaks at the particle positions,
MgCl2. In a typical dialysis experiment, the irrespective of their angular orientation. Peak
sample was dialyzed for 12 h at 13 °C. The coordinate lists served to extract subframes
temperature was then increased linearly to 37 for subsequent alignment.
°C over the next 24 h, and kept at 37 °C for Micrographs of 2D crystals were selected
24 h. It was subsequently continuously with an optical diffractometer based on the
lowered over a 12 h period to 13 °C. Large alignment of the microscope and the crystal
crystalline sheets were observed after further order. Suitable areas were digitized and their
dialysis for 3 days at 13 °C. The reconstituted crystal quality and unit cell morphology
membranes were stored at 4 °C. assessed by Fourier peak filtration (Aebi et
al., 1973). Electron microscopy and image
processing 4.1.6. References

Solubilized particles and crystalline sheets Aebi, U., Smith, P. R., Dubochet, J., Henry, C.
were adsorbed for 1 min to parlodion carbon- & Kellenberger, E. (1973). A study of the
structure of the T-layer of bacillus brevis.
coated grids rendered hydrophilic by glow- J. Supramol. Struct. 1, 498-522.
discharge in a low pressure of air, washed
with 3 drops of water, and stained with 0.75% Carrasco, N., Viitanen, P., Herzlinger, D. &
uranyl formate. Images were recorded with a Kaback, H. R. (1984). Monoclonal
antibodies against the lac carrier protein
Hitachi H-7000 electron microscope operated from Escherichia coli. 1. Functional
at 100 kV. studies. Biochemistry 23, 3681-3687.
LacY 2D crystals mixed with 3 - 10%
trehalose were adsorbed to carbon-coated Frillingos, S., Gonzalez, A. & Kaback, H.
(1998a). Cysteine-scanning mutagenesis
copper electron microscopy grids. The grids of helix IV and the adjoining loops in the
were blotted and plunged into liquid ethane. lactose permease of Escherichia coli: Glu
Vitrified specimens were recorded in a 126 and Arg 144 are essential.
Hitachi H8000 electron microscope with a Biochemistry 47, 14284-14290.
LaB6 filament, operated at 200 kV under low Frillingos, S. & Kaback, H. R. (1996).
dose conditions (~ 5 e- / Å2). Monoclonal antibody 4B1 alters the pKa
Electron micrographs of densely packed of a carboxylic acid at position 325 (helix
vesicles selected for image processing were X) of the lactose permease of Escherichia
coli. Biochemistry 35, 10166-71.
digitized using a Leafscan-45 scanner (Leaf
systems, Inc., Westborough, MA, USA). Frillingos, S., Sahin-Toth, M., Wu, J. &
All image processing steps described Kaback, H. R. (1998b). Cys-scanning
below were performed using the SEMPER mutagenesis: a novel approach to struc-
ture-function relationships in polytopic
image processing system (Saxton et al., membrane proteins. FASEB J in press.
1979). For single-particle analysis a reference
was established by selecting a particularly Jap, B. K., Zulauf, M., Scheybani, T., Hefti,
well preserved triangularly shaped particles A., Baumeister, W., Aebi, U. & Engel, A.
(1992). 2D crystallization: from art to
(such as that indicated by a circle in Fig. 1a). science. Ultramicroscopy 46, 45-84.
A 20-fold rotational symmetry was applied to
this reference resulting in a circle with Kaback, H. R. (1976). Molecular biology and
dimensions of the LacY triangles (Thuman- energetics of membrane transport. J. Cell
Physiol. 89, 575-593.
Commike & Chiu, 1996) . The cross-
correlation function of such a reference with a Kaback, H. R. (1983). The lac carrier protein
field of triangle-shaped particles revealed in Escherichia coli. J Membr Biol 76, 95-
Kaback, H. R. (1996). The lactose permease Privé, G. G., Verner, G. E., Weitzman, C.,
of Escherichia coli: past, present and Zen, K. H., Eisenberg, D. & Kaback, H.
future. In Handbook of Biological R. (1994). Fusion proteins as tools for
Physics: Transport processes in crystallization: The lactose permease from
Eukaryotic and prokaryotic Organisms. Escherichia coli. Acta Cryst. 50, 375-379.
Vol II. Edited by Konings 2n, Kaback
HR, Lolkema JS. Amsterdam: Elsevier, Rigaud, J. L., Mosser, G., Lacapere, J. J.,
203-227. Olofsson, A., Levy, D. & Ranck, J. L.
(1997). Bio-Beads: An efficient strategy
Kaback, H. R. (1997). A molecular for two-dimensional crystallization of
mechanism for energy coupling in a membrane proteins. J. Struct. Biol. 118,
membrane transport protein, the lactose 226-235.
permease of Escherichia coli. Proc Natl.
Acad. Sci U S A 94, 5539-43. Sahin-Tóth, M., Lawrence, M. C. & Kaback,
H. R. (1994). Properties of permease
Kaback, H. R., Voss, J. & Wu, J. (1997). dimer, a fusion protein containing two
Helix packing in polytopic membrane lactose permease molecules from
proteins: the lactose permease of Escherichia coli. Proc Natl. Acad. Sci U S
Escherichia coli. Curr Opin Struct Biol 7, A 91, 5421-5.
Saxton, W. O., Pitt, J. T. & Horner, M.
Kaback, H. R. & Wu, J. (1997). From (1979). Digital image processing: the
membrane to molecule to the third amino Semper system. Ultramicroscopy 4, 343-
acid from the left with a membrane 354.
transport protein. Q Rev Biophys 30, 333-
64. Sun, J., Wu, J., Carrasco, N. & Kaback, H. R.
(1996). Identification of the epitope for
Le Coutre, J., Narasimhan, L. R., Kumar, C., monoclonal antibody 4B1 which
Patel, N. & Kaback, H. R. (1997). The uncouples lactose and proton translocation
lipid bilayer determines helical tilt angle in the lactose permease of Escherichia coli.
and function in lactose permease of Biochemistry 35, 990-8.
Escherichia coli. Proc. Natl. Acad. Sci. 94,
10167-10171. Thuman-Commike, P. A. & Chiu, W. (1996).
PTOOL: a software package for the
Li, J. & Tooth, P. (1987). Size and shape of selection of particles from electron
the Escherichia coli lactose permease cryomicroscopy spot-scan images. J
measured in filamentous arrays. Struct Biol 116, 41-7.
Biochemistry 26, 4816-4823.
Viitanen, P., Newman, M. J., Foster, D. L.,
Loddenkotter, B., Kammerer, B., Fischer, K. Wilson, T. H. & Kaback, H. R. (1986).
& Flügge, U.-I. (1993). Expression of the Purification, reconstitution, and
functional mature chloroplast triose characterization of the lac permease of
phosphate translocator in yeast internal Escherichia coli. Methods Enzymol. 125,
membrane and purification of the 429-52.
histidine-tagged protein by a single metal-
affinity chromatography step. Proc. Natl. Zhuang, J., Prive, G. G., Werner, G. E.,
Acad. Sci. U S A 90, 2155-2159. Ringler, P., Kaback, R. H. & Engel, A.
(1998). Two-dimensional crystallization of
Poolman, B. & Konings, W. N. (1993). the Escherichia coli lactose permease. J.
Secondary solute transport in bacteria. Struct. Biol. 125, 63-75.
Biochem. Biophys. Acta 1183, 5-39.

5. General discussion
and conclusions

5. General discussion and conclusions

This thesis presents a new preparation achievement of such high resolution also
method for 2D crystals grown on a possible for poorly ordered proteins as
functionalized lipid monolayer for high documented for three membrane and two
resolution atomic force microscopy and new water soluble proteins (Scheuring et al., 1999;
structural data on several physiologically Fotiadis et al., 2000; Seelert et al., 2000; Mou
relevant membrane proteins derived from et al., 1996). Single particle averaging of the
atomic force and electron microscopy molecules yields a representative surface of
investigations. the protein species revealing the
Chapter 1 presents a technical development oligomerization and loop organization.
for high resolution AFM. In order to develop In Chapter 2, the use of high resolution
a preparation technique which allows the AFM to investigate the Escherichia coli water
study of 2D crystals of water soluble proteins channel AqpZ and the Rubrivivax gelatinosus
grown on functionalized lipid monolayers light-harvesting complex LH2 is presented.
with the AFM, 2D crystals of streptavidin This lead to new insights into the organization
were assembled on biotinylated lipid of their loops and termini and assignment of
monolayers and transferred into the AFM their sidedness.
using highly oriented pyrolytic graphite AqpZ was imaged before and after
(HOPG) as sample support. Parallel cleavage of the N-terminal his-tag with
hydrophobicity measurements revealed that trypsin. Proteolysis induced a drastic change
HOPG surfaces are more hydrophobic than of surface appearance and height over the
silanized glass, allowing protein crystals lipid membrane on one side of the protein,
grown on lipid monolayers to be stably which consequently could be assigned as
attached and imaged using AFM. The cytoplasmic. The extracellular surface,
recently developed technique for 2D however showed strong changes of
crystallization of membrane proteins by Levy topographical features when imaged with a
et al. (1999) using functionalized lipid loading force of ~100pN applied to the tip:
monolayers together with the novel the largest protrusion was strongly displaced
preparation technique presented in this thesis by the tip. Together with flexibility mapping
opens a wide range of membrane proteins to and volume calculations this protrusion was
be investigated: All his-tagged proteins may assigned as the loop C.
be immobilized on NiNTA functionalized Similarly, LH2 was imaged under
lipid monolayers and investigated under physiological conditions before and after
physiological conditions. This will give protease treatment. In this case, thermolysin
insight into the structure of membrane was used to specifically cleave the C-terminus
proteins, which have resisted crystallization of the α-subunit. Absorption spectra showed
by other techniques. that both the native and the cleaved proteins
The use of defined scanning buffers which contained all the chromophores. The most
compensate the surface charges of the AFM strongly protruding surface changed both it's
tip and the sample (Müller et al., 1999) has height over the lipid membrane and it's
allowed protein surfaces to be contoured at a surface appearance upon protease treatment.
lateral resolution of ~7Å and a far better This determined the periplasmic surface and
vertical resolution. The extremely high signal- the position of the C-terminus of the α-
to-noise ratio of the AFM makes the subunit. In reconstitutions from less

stringently purified samples, large rings, permease) was overexpressed in it's native
assigned as light-harvesting complexes 1 system and purified using Ni-affinity
were imaged in interaction with LH2 rings, chromatography. The purified protein was
supporting the proposed model of the action imaged in it's solubilized state revealing a
mechanism of the photosynthetic apparatus trimeric organization. A low resolution 3D
(Kühlbrandt, 1995). reconstruction showed that trimerization is
Because of it's high signal-to-noise ratio induced by hydrophilic interactions between
high resolution AFM allows loops and the cytb 562 of each monomer. Such trimers
termini of proteins to be identified and the could be reconstituted into lipid bilayers
different protein species within a membrane resulting in densely packed vesicles and 2D
to be distinguished on the raw data images crystals (Zhuang et al., 1999). Best crystals
without further processing. This opens the pack the trimers in a hexagonal lattice and
perspectives of imaging native membranes diffract to ~8Å, but stacking of crystalline
under close to native conditions in future layers prevented the calculation of a high
studies. resolution projection map. The observed
Chapter 3 presents the combination of cytochrome b562 induced trimerization of red
surface and projection techniques to elucidate permease together with the finding that the
the sidedness of aquaporins. Atomic force protein is stable in decyl-maltoside increases
microscopy studies combined with the range of crystallization conditions, which
proteolysis experiments lead to an can now be investigated to obtain high
unambiguous assignment of surface resolution crystals.
sidedness of membrane proteins (see Chapter Atomic force and electron microscopy
2). Cryo-electron microscopy density maps cover a resolution range from micrometers to
elucidate the intramembraneous parts of such angstroms and provide both surface and
molecules (Walz et al., 1997). In order to volume information that is acquired under
assign the sidedness information from AFM close to native conditions. Membrane proteins
experiments to a density map, large 2D were reconstituted in lipid membranes, which
crystals of AqpZ were freeze dried and metal represent their native environment. Therefore,
shadowed (Gross et al., 1990). Micrographs the results presented in this thesis meet the
were taken where aggregates had blocked the goal to acquire structural information on
deposition of metal on parts of a crystal, biological molecules under conditions, which
hence surface and density information could ensure that the unseparable relationship
be obtained from one crystal. Datasets from between structure and function is intact.
such preparations were used to align the high
resolution AFM topography to the high
resolution cryo-EM density maps of AqpZ Fotiadis, D., Hasler, L., Müller, D. J.,
and AQP1. It was shown that the sidedness Stahlberg, H., Kistler, J. & Engel, A.
(2000). Surface tongue-and-groove
assignment previously published (Walz et al., contours on lens MIP facilitate cell-to-cell
1996) had to be revised. The method adherence. J. Mol. Biol. 300, 779-789.
described can be applied to any membrane
protein reconstituted into 2D crystals Gross, H., Krusche, K., Tittmann, P. (1990).
Recent progress in high-resolution
allowing a unambiguous surface sidedness shadowing for biological transmission
assignment to a density map. electron microscopy. Proceedings of XIIth
Chapter 4 reports on structural studies of International Congress for Electron
the Escherichia coli lactose permease (LacY) Microscopy, 510-511
using electron microscopy techniques. The Kuhlbrandt, W. (1995). Many wheels make
fusion protein LacY/cytb562/C417His6 (red light work. Nature 374, 497-498.

Lévy, D., Mosser, G., Lambert, O., Moeck, G. Seelert, H., Poetsch, A., Dencher, N., Engel,
S., Bald, D. & Rigaud, J. L. (1999). Two- A., Stahlberg, H. & Müller, D. J. (2000).
dimensional crystallization on lipid layer: a Proton-powered turbine of a plant motor.
successful approach for membrane Nature 405, 418-419.
proteins. J. Struct. Biol. 127, 44-52.
Walz, T., Tittmann, P., Fuchs, K. H., Müller,
Mou, J., Sheng, S., Ho, R. & Shao, Z. (1996). D. J., Smith, B. L., Agre, P., Gross, H. &
Chaperonins GroEL and GroES: views Engel, A. (1996). Surface topographies at
from atomic fore microscopy. Biophys. J. subnanometer resolution reveal asymmetry
71, 2213-2221. and sidedness of aquaporin-1. J. Mol.
Biol. 264, 907-918.
Müller, D. J., Fotiadis, D., Scheuring, S.,
Müller, S. A. & Engel, A. (1999). Walz, T., Hirai, T., Murata, K., Heymann, J.
Electrostatically balanced subnanometer B., Mitsuoka, A., Fujiyoshi, Y., Smith, B.
imaging of biological specimens by atomic L., Agre, P. & Engel, A. (1997). The 6Å
force microscopy. Biophys. J. 76, 1101- three-dimensional structure of aquaporin-
1111. 1. Nature 387, 624-627.

Scheuring, S., Ringler, P., Borgnia, M., Zhuang, J., Prive, G. G., Werner, G. E.,
Müller, D. J., Agre, P. & Engel, A. (1999). Ringler, P., Kaback, R. H. & Engel, A.
High resolution topographs of the (1999). Two-dimensional crystallization of
Escherichia coli waterchannel AqpZ. the Escherichia coli lactose permease. J.
EMBO Journal 18, 4981-4987 Struct. Biol. 125, 63-75.

6. Acknowledgment

6. Acknowledgment

I thank Prof. Dr. Andreas Engel. You well, when you face the responsibility for
always showed up with an extremely good leading your own group: may the force be
feeling for the situations I was in during my with you!
phD. After my diploma, when I thought I had
to stop there with science, you pushed me I thank Henning Stahlberg. You entered
almost against the wall, telling me that you the lab just about two years ago and I can not
believed I had a certain talent. Then, tell how much knowledge and character you
sometimes you put me under more pressure, brought into the group. You are a great
sometimes less - sometimes you just smiled scientist and in addition you take the time to
or went running with me. Now, looking back communicate what you know. Most of what I
I have the impression it was always exactly know about computers and image processing
what I needed. Fortunately, you thought me I learned from you, during the time you
after a certain time to remain in a, helped me writing these routines (for volume
scientifically spoken, stable conformation and and probability calculations). Recently you
our work got better and more efficient. also showed me how to do cryo-EM. Your
Maybe I remember mostly your reaction after modest ways and integrity should be honored,
the AFM slipped from this carriage I was and I hope for you that you will succeed in
steering. You examined the surface of the your personal and professional life!
carriage and said: "It's slippery, it could have
happened to anyone.", and it was never a topic I thank Shirley Müller. You shaped up
again. Scientifically you are full of knowledge surely every manuscript which ever left my
and full of ideas. I grabbed as much of your desk. Aware of the importance of
knowledge as I could, and you made me transmission of data into a understandable
combine your ideas and mine. I wish, and I written form, I owe you a lot. Unfortunately I
am sure, that everything you take into your never had a STEM project, so we never
hands in your personal and professional life worked on a subject together, nevertheless
will be conducted by this force I have seen your clear and scientific brain helped me and
you can give! everyone in the group when you stated
something, which was probably true, in a
I thank Daniel J. Müller. It was you who group seminar. It was also you, who listened
motivated me to join the lab, and it was you carefully when I ran around with my
who trained me in AFM. During the next aquaporin sidedness theory. Thanks for that.
years, we had a lot of ups and downs, which You are extremely important for this group,
might be because we are very different or very scientifically exact, and humanly warm
similar in our minds, something, which I have hearted, keeping all us boys around you on
not yet figured out. Now, for some time the ground.
everything is superb for work as for going to
the bars together. Besides the AFM I have I thank Philippe Ringler. It was great
learned from you how to make beautiful working with you on the AqpZ project. I wish
figures, how to talk to people, and how to put for you and your family all the best. You are a
creativity into my work. The next thing good microscopist, it's a pity that you do not
should be, how to become extremely use the microscope at the moment. But maybe
organized. I am convinced that you will do you will come back to microscopy and make

the beautiful Goldringler images. proteins. You have a lot of skills, hopefully
you will use them all to solve the problems
I thank Bernard Heymann. You supported ahead. Good luck.
strongly the writing of the streptavidin story.
Later we mostly disagreed which taught me to I thank Clemens Möller. You are a
hold on to my opinions. sensible colleague, something like the
barometer of the group, able to spread a calm
I thank Dimitrios Fotiadis. Hey Dimiman, and peaceful atmosphere. I hope you will get
as you know, we are not the same. But the more self confidence and stronger pushing
more time we spent together (and now it's yourself and your opinions through. You
more than 8 years!!!), the better we do. We surely bring a great knowledge of physics to
did our diploma and phD thesis together and the group and this in the context of structural
had some good laughs together, and I assume biology makes a great combination. We have
they won't be the last. You are an extremely a very different music taste, which made it
exact scientist. If you put something in your very calm measuring AFM together, since
head, you surely get there. I wish you all the you could not stand my music, similarly in
best, and if we ever might be in labs far apart, the other direction. I wish you all the best on
we'll meet for a beer in Rio-Bar, or a coffee in your way.
the Biozentrum cafeteria, and see how time
passes by. I wish Gabriel Fedrigo all the best for his
phD. You just started a few weeks ago, and
I thank Lorenz Hasler. I always liked your put yourself already deeply into the blackbox
ways, although we do not have "s'Heu uf dr of 2D crystallization. Everyone, who works
gliche Bühni". Everyone around you can be with electron crystallography would like to
sure, that if you say something, you mean it know more what's going on in there,
exactly that way (...and I have there some hopefully your work, supported by the
experience). You were always straight, helpful biophysical background you bring with you,
and supported me with biochemical problems. will elucidate some of the secrets. All the
I wish you all the best for your future in best!
America and hope to see you sooner or later.
Backgammon rocks, and sorry I took your I wish Andreas Schenk and Thomas
pipette so often! Kaufmann all the best for their diploma
thesis. I think you two are extremely
I thank Thomas Braun. You are a honest intelligent and interested in this research and
and modest person. Your computer will become specialists soon. For Thomas I
knowledge is way beyond my understanding hope also that we will have a good time
(I can't promise you here that I will learn together and work successfully. I am
'Latex'). You are daily asking the group for convinced we can make large LacY crystals.
lunch and by this social being together, you
create a great atmosphere in the group. I wish I thank Christine Widmer. You supported
you all the best for the rest of your phD! the difficult LacY project a lot, and I hope we
will do more in the near future. After some
I thank Herve Remigy. You have a great difficulties in the beginning, we have both
sense of humor and made me smile a lot of learned how to calm the waves. I hope we
times during the last years. Now you have put soon get good crystals, consequently using
a lot of energy into the dilution machine, and I more the microscope than the french press. I
hope it will help to crystallize difficult wish you all the best for your personal life

with family and hobbies. I thank Urs Fürstenberger. Fortunately,
you are floormanager and not complicated at
I thank Kitaru Suda. Like a magician you the same time. Thanks to you, I got the
are doing biochemistry. And like a magician chemicals ordered quickly , the packets were
you play the guitar. Your wisdom is broad in sent immediately, the money reimbursed in
science, history and culture. You even know time. Good luck with your new job, now
the Swiss folk songs better than anyone of taking care of the whole Biozentrum!
us. You help everybody. What can I say?! -
We are all lucky having you around. Thank I thank Beat Schumacher. For a few
you again for the millions of little and big months now you have been the floormanager,
helps. and you do your job extremely well.
Everything goes quick, go on so!
I thank Ueli Aebi. It is an extraordinary
institute of which I could use all the facilities I thank Peter Tittmann and Heinz Gross. I
and infrastructure. I know, you are strongly had a great time with you two in Zürich doing
involved in the background organization of the sidedness experiments on AqpZ. I hope
everything here, work I do not see directly, you will use the advantage of your machine
but which leads to this effective and again to get surface and projection
stimulating atmosphere. Thanks too, for information on one crystal. Working with
taking out your credit card and saving my life you is a pleasant thing, there was science
in this hotel in Atlanta. work and good discussions during the coffee
times. Peter, thanks for the hours you spent
I thank Cora-Ann Schönenberger, Claire on the microscope with me.
Goldsbury and Martin Stolz. You amused me
always a lot when passing by on the 3rd Merci, Jean-Louis Rigaud. Depuis qu'on a
floor. Claire, thanks again for the amyloid commence a collaborer, j'avais deja deux fois
fibrils used during the sidedness assignment de temps exelente à Paris, grâce a toi. On avait
experiments. travaile efficace et avec des idées, et on avait
bien rigoler ensemble les soirées apres le
I thank Robert Häring, Robert Wyss and boulot. En plus tu as un humeur uncroyable.
Roland Bürki. Thanks to you, the MIH is Je serais heureux, si on continue a travailler
running. El-Röbi and Rolli take surely 90% ensemble. J'ai bien trouvé avec toi quelqu'un,
of all the computer problems away from us, duquel je peux beaucoup apprendre pour le
meanwhile Mech-Röbi fixes the centrifuges. travail et la vie. Merci pour tous.
Taking this together I assume, if you were not
around I would just be writing the first line. Merci, Daniel Levy. Tu m'avais appris de
Thanks! faire la technique de la crystallisation
monocouche. En plus tu es un très bon
I thank Barbara Merz. Your work on the scientifique, qui prend son temps a expliquer
3 floor makes these lab so efficient. Also, comment faire les manipes. J'espere que les
you always prepare beautiful plates for the gens vont remarquer que la technique que tu
3rd floor parties. avais introduit est forte. Tout le bon pour ta
vie avec ta jeune famille et la science.

7. Curriculum vitae

7. Curriculum vitae
Simon Andreas Scheuring Instructor in the "Blockkurs in Biophysik
born in Basel, Switzerland, on March 8,1973 und Strukturbiologie"

2000 Teacher in the "EMBO Practical
7.1. Education Course of Current Methods in Membrane
Protein Research" at the EMBL, Heidelberg,
1979-1984 Primary school (Basel) Germany.
Instructor in the "Blockkurs in Biophysik
1984-1992 High school (Basel), Matura und Strukturbiologie"
(Typus B)

1992-1997 Diploma in molecular biology, 7.3. Publications
Biozentrum, University Basel.
Thesis topic: "Structure and function of Single proteins observed by atomic force
GroEL and its complexes investigated with microscopy.
the EM and the AFM" (in the group of Prof. Single Molecules, IN PRESS
A. Engel.) Simon Scheuring, Dimitrios Fotiadis,
Clemens Möller, Andreas Engel and Daniel J.
1997-2000 PhD thesis in structural Müller
biology, Biozentrum, University Basel.
Thesis topic: "atomic force and electron Imaging and manipulation of biological
microscopic analysis of membrane channels structures with the AFM.
and transporters" (in the group of Prof. A. Micron, SUBMITTED
Engel.) Dimitrios Fotiadis, Simon Scheuring,
Andreas Engel and Daniel J. Müller

7.2. Teaching High resolution AFM topographs of the
Rubrivivax gelatinosus light-harvesting
1996 Instructor in the "Blockkurs in complex LH2.
Biophysik und Strukturbiologie" The EMBO Journal, IN PRESS
Simon Scheuring, Francoise Reiss-
1997 Instructor in the "EMBO Practical Husson, Andreas Engel, Jean-Louis Rigaud
Course in Scanning Probe Microscopy" at and Jean-Luc Ranck
the Biozentrum, Basel, Switzerland.
Instructor in the "Blockkurs in Biophysik Conformational changes, flexibilities and
und Strukturbiologie" intramolecular forces observed on individual
proteins using AFM.
1998 Teacher in the "Practical Course in Single Molecule, 2000, 1: 115-118
Atomic Force Microscopy in Biology" at the Daniel J. Müller, Dimitrios Fotiadis,
Biozentrum, Basel, Switzerland. Clemens Möller, Simon Scheuring and
Instructor in the "Blockkurs in Biophysik Andreas Engel
und Strukturbiologie"
The aquaporin sidedness revisited.
1999 Tutoring for students of Biology at Journal of Molecular Biology, 2000, 299
the Biozentrum at the University of Basel (5):1271-1278
7.4. Meetings
Simon Scheuring, Peter Tittmann,
Henning Stahlberg, Philippe Ringler, Mario Raster-Sonden-Mikroskopien und
Borgnia, Peter Agre, Heinz Gross and Organische Materialien V. Diskussions-
Andreas Engel tagung, 7 – 9 Oktober 1996, Münster,
Direct observation of postadsorption Präsentation: "Die E. coli Chaperonin
aggregation of antifreeze glycoproteins on GroEL untersucht mit dem EM und dem
silicates. AFM". Simon Scheuring, Daniel J. Müller
Langmuir, 2000, 16 (13):5785-5789 and Andreas Engel.
Ph. Lavalle, A. L. DeVries, C.-C. C.
Cheng, S. Scheuring and J. J. Ramsden Raster-Sonden-Mikroskopien und
Organische Materialien VI. Diskussions-
High resolution AFM topographs of the tagung, 8 - 10 Oktober 1997, Tübingen,
Escherichia coli water channel aquaporin Z. Deutschland.
The EMBO Journal, 1999, 18 (18):4981- Präsentation: "A novel preparation method
4987 for high resolution AFM introduced with 2D-
Simon Scheuring, Philippe Ringler, Mario streptavidin crystals grown on biotinlipid
Borgnia, Henning Stahlberg, Daniel J. Müller, monolayer". Simon Scheuring, Daniel J.
Peter Agre and Andreas Engel Müller and Andreas Engel.

Imaging streptavidin 2D-crystals on Microscopy and Microanalysis '98, July
biotinylated lipid monolayers at high 12-16, Atlanta, Georgia, USA.
resolution with the atomic force microscope. Presentation: "A novel preparation method
Journal of Microscopy, 1999, 193:28-35. for high resolution AFM introduced with 2D-
Simon Scheuring, Daniel J. Müller, streptavidin crystals grown on biotinlipid
Philippe Ringler, J. Bernard Heymann and monolayer". Simon Scheuring, Daniel J.
Andreas Engel Müller, Philippe Ringler, J. Bernard
Heymann and Andreas Engel.
Electrostatically balanced subnanometer
imaging of biological specimens by atomic Raster-Sonden-Mikroskopien und
force microscopy. Organische Materialien VII. Diskussions-
Biophysical Journal, 1999, 78:1101-1111 tagung, 7 - 9 Oktober 1998, Berlin,
Daniel J. Müller, Dimitrios Fotiadis, Deutschland.
Simon Scheuring, Shirley A. Müller and Präsentation: "High resolution AFM
Andreas Engel topographs identify sides, handedness and
loops of the E. coli waterchannel AqpZ".
A novel preparation method for high Simon Scheuring, Mario Borgnia, Daniel J.
resolution AFM introduced with 2D- Müller, Peter Agre and Andreas Engel.
streptavidin crystals grown on biotinlipid
monolayer. Meeting of the EU-Biotech Program:
Microscopy and Microanalysis Meeting "Water and glycerol channels from the MIP
Proceedings, 1998, 4, (2), 312 - 313. family: structure, function, regulation and
Simon Scheuring, Daniel J. Müller, exploitation". 15 - 17 September, 1999,
Philippe Ringler, J. Bernard Heymann and Hamburg, Deutschland.
Andreas Engel Presentation: "The sidedness and loops of
AqpZ". Simon Scheuring, Philippe Ringler,

Mario Borgnia, Henning Stahlberg, Daniel J. 3rd International Conference on
Müller, Peter Agre and Andreas Engel. "Molecular biology and physiology of water
and solute transport". July 1 - 5, 2000,
Raster-Sonden-Mikroskopien und Orga- Göteborg, Sweden.
nische Materialien VIII. Diskussionstagung, Poster: "The Sidedness of Aquaporins
4 - 6 Oktober 1999, Basel, Schweiz. determined by atomic force and transmission
electron microscopy". Simon Scheuring,
Meeting of the EU-Biotech Program: Peter Tittmann, Henning Stahlberg, Philippe
"Water and glycerol channels from the MIP Ringler, Mario Borgnia, Peter Agre, Heinz
family: structure, function, regulation and Gross and Andreas Engel.
exploitation". 15 - 17 March, 2000, Haute-
Nendaz, Schweiz. Rastersondenmikroskopie in Forschung
Presentation: "The surface of AQP2 und Industrie, 10. Oktober 2000, Wuppertal,
investigated with the atomic force Deutschland.
microscope". Simon Scheuring, Lorenz Präsentation: "Neue biologische Appli-
Hasler, Paul Werten, Peter Deen and Andreas kationen in der Rasterkraftmikroskopie".
Engel. Simon Scheuring and Andreas Engel.