Feeling sugar-protein interactions using carbon nanotubes : a molecular recongition force microscopy study

recongition force microscopy study

Klein, D.C.G.

Citation

Klein, D. C. G. (2004, November 11). Feeling sugar-protein interactions using carbon

nanotubes : a molecular recongition force microscopy study. Retrieved from

https://hdl.handle.net/1887/106077

Version:

Publisher's Version

License:

Licence agreement concerning inclusion of doctoral thesis in the

_{Institutional Repository of the University of Leiden}

Downloaded from:

https://hdl.handle.net/1887/106077

The handle

http://hdl.handle.net/1887/106077

holds various files of this Leiden University

dissertation.

Author:

Klein, D.C.G.

Feeling sugar-protein interactions

using carbon nanotubes

A digital version of this thesis can be downloaded from http://www.physics.leidenuniv.nl

Cover:

Front: Artist impression of a functionalized carbon nanotube AFM tip

probing a surface that contains a receptor. Both sample surface and tip are immersed in liquid. The drawing was made by Henriette Jensenius.

Back: from left upper corner to right lower corner:

AFM image taken in liquid of a Chinese hamster ovary cell, picture was taken by Maarten van Es; molecular structure of mannose with a three-carbon linker and an amine group at the end, image was made by Karin Sliedregt-Bol; AFM image of purple membrane in liquid, trimers of bacteriorhodopsin are arranged in a hexagonal lattice;

Pisum sativum lectin immobilized on mica, imaged by AFM in liquid,

force-distance curve taken with a carbon nanotube tip on a carboxyl-SAM in liquid, functionalization cell for AFM tips.

Background: modified mica surface imaged by light microscopy in air.

Feeling sugar-protein interactions

using carbon nanotubes

A molecular recognition

force microscopy study

PROEFSCHRIFT

de graad van Doctor aan de Universiteit Leiden, op gezag van de Rector Magnificus Dr. D.D. Breimer,

hoogleraar in de faculteit der Wiskunde en Natuurwetenschappen en die der Geneeskunde,

volgens besluit van het College voor Promoties te verdedigen op donderdag 11 november 2004

klokke 14.15 uur

door

Dionne Clara Gertrud Klein

PROMOTIECOMMISSIE

Promotores: Prof. dr. J.W.M. Frenken Prof. dr. J.W. Kijne Co-promotor: Dr. T.H. Oosterkamp Referent: Dr. P. Hinterdorfer

(University of Linz, Austria) Overige leden: Dr. Th. J. Aartsma

Prof. dr. P.H. Kes

Dr. R. McKendry

(University College London, UK) Prof. dr F.W. Saris

Prof. dr. H.P. Spaink

Contents

Chapter

Page

1 Introduction 1

1.1 Introduction and motivation 2 1.2 Receptor-ligand interactions 4

Historical background of AFM in receptor-ligand research 5 Sugar-lectin interactions in this thesis 5 1.3 Atomic force microscopy 6

Molecular recognition force microscopy 7

1.4 Carbon nanotubes 8

1.5 Outline of this thesis 10

2 An AFM study of FIN filaments: 13 filaments in between the nuclei of dividing

Saccharomyces cerevisiae (yeast) cells

2.1 Introduction 14

Yeast cell division 14

Microtubuli 15

Fin1 protein 16

Confocal fluorescence microscopy 16

Electron microscopy 17

Goals 17

2.2 Methods 17

Fin 1 preparation 17

Sample preparation 18

Atomic force microscopy 18

Control experiments 19

2.3 Discussion and conclusion 20

3 Imaging biological molecules and membranes under 23 physiological conditions

3.1 Introduction 24

3.2 Contact mode in liquid 24

Native purple membrane 27

Experimental details 28

Light harvesting 2 complex (LH2) 30

Acoustic driving 32

Pea lectin dimers 33

MAC mode: magnetic driving 34

Chinese hamster ovary (CHO) cells 34

3.4 Conclusions 36

4 Covalent immobilization of single proteins on mica for 39 molecular recognition force microscopy

4.1 Introduction 40

4.2 Experimental procedures 42

Surface modification 43

Protein immobilization 43

AFM imaging 44

4.3 Results and discussion 45

4.4 Conclusions 49

5 Carbon nanotubes as nanometer-sized probes for AFM 51

5.1 Introduction 52

5.2 Single-walled carbon nanotube AFM tips 52 Production: chemical vapor deposition 52

Mounting procedure 54

Intermezzo: carbon nanotube mechanics 56

Nanotube shortening 59

Making the tip water-proof 60

5.3 Imaging antibody molecules with SWNT AFM tips 61 5.4 Multi-walled carbon nanotube AFM tips 64

Production: arc discharge 64

Mounting procedure 65

Making the tip water-proof 65

Nanotube shortening 66

5.5 High-resolution TEM results 67 Open versus closed nanotubes 67

5.6 Conclusions 69

6 Carbon nanotube functionalization for high-resolution 71 molecular recognition force microscopy

6.1 Lateral resolution in molecular recognition force 72 microscopy

6.2 Chemical functionalization of AFM tips for 72 high-resolution MRFM

Experimental procedures 80

6.6 Conclusions 81

7 Molecular recognition force microscopy on pea lectin 83 and mannan binding lectin 7.1 Introduction 84

7.2 Interpretation of phase images; qualitative and 85

quantitative aspects MRFM with short spacers: how to detect unbinding 88 events 7.3 Pisum sativum lectin 88

Experimental details 89

In air 89

In liquid 90 MRFM with short spacers: how to detect unbinding 92 events, part II 7.4 Mannan binding lectin 95

Experimental details 96

In air 96

Preliminary experiments in liquid 97

7.5 Conclusions 102

8 Summarizing discussion and outlook 105

8.1 Summarizing discussion 106

8.2 Technical outlook 106

8.3 Scientific challenges 107

Appendices

Carbon nanotube mechanics Gold crystal preparation

Acknowledgements Summary

Glossary

Samenvatting voor de leek Curriculum vitae

Chapter 1

Introduction

1.1 Introduction and motivation

The development of atomic force microscopy (AFM) for application to biological systems has initiated a new form of molecular biology or biomolecular physics. With the direct observation of single molecules at work, such as the activity of RNA polymerase1_{, and real-time observation of the complex formation of}

single GroEL and GroES molecules2_{, a new approach to study}

biological processes at the level of individual molecules has been introduced. Combining its ability to study single proteins, membranes and whole cells in physiological conditions with its inherent high resolution, AFM is an excellent tool to study biological processes on a (sub-) molecular level.3,4

Currently, many techniques are being developed or refined to investigate a large multitude of biological systems on the molecular and submolecular level, that have so far only been studied by

traditional, indirect biochemical methods. It is now becoming possible to “see” the molecular mechanisms of biology. Both microscopy and cell biology have now arrived at the single molecule level. This new field, which is often referred to with the buzz word “bio-nanotechnology”, is developing very quickly, because of investments by many research groups in the world.

Our contribution to this new and exciting field is the combination of the high lateral resolution of atomic force microscopy with chemical resolution. In the future, this tool can be applied to complex biological systems and, in addition to “simple” topography, to obtain direct quantitative information such as binding constants of proteins and their dependence on the direct surroundings of a receptor. A key element is the high spatial resolution, which will allow us to extract such information of individual molecules.

In order to compare AFM to other techniques, it is useful to distinguish two different criteria: on the one hand the spatial resolution a technique can achieve and on the other hand the degree in which the conditions, while measuring, approximate real life situations.

Electron microscopy (EM) is a technique that can provide images of biological structures, such as organelles in cells with a remarkably high resolution in all three dimensions.

Single molecule fluorescence microscopy is a powerful technique that allows to extract information about single molecules in their native environment. A very nice example is the assessment of the degree of dimerization of a G-protein-coupled receptor prior to stimulation and after stimulation, which confirmed the connection between dimerization and functionality that had been suggested for a long time.5 _{The resolution that can be obtained is in the order of 200}

nm. In fluorescence resonance energy transfer (FRET), distances of only a few nm between fluorescent labels can be measured.

In terms of resolution, AFM fills a niche between single molecule fluorescence and EM on the one hand and X-ray diffraction and NMR on the other hand. In addition, measuring under physiological conditions, can only be performed by NMR, single molecule fluorescence microscopy and AFM.

The first part of this thesis has an exploratory character. We first acquaint ourselves with the possibilities of bio-AFM by studying various proteins and membrane fragments. In the second part we select one topic from the large multitude of relevant biological questions that can be answered with the AFM techniques described here. We focus on the specific interaction between sugars and sugar-binding proteins, which serves as an example of a more general approach that can be applied to many other biological systems.

The ultimate goal of the research that is described in this thesis is to recognize binding sites on proteins, with (sub-) molecular resolution. In order to reach this goal, very detailed topographical information is necessary, combined with specific chemical binding information. In order to make this possible, standard AFM had to be adapted. Shortly: a small ligand was bound to the sharp needle of the AFM, and the interaction between this ligand and receptors on a surface was probed.

single experiment with all ingredients working simultaneously, we investigated two different kinds of sugar-binding proteins.

1.2 Receptor-ligand interactions

Our present knowledge about receptor-ligand interactions is mainly the result of decades of research on the scale of many cells. This research provided information about the average behavior of many cells. The behavior of single cells or, on an even smaller scale, single receptors, remained unexplored for a long time. New techniques make it possible to study not only single cells, but even single molecules (1.3). This allows us to gain fundamental understanding of the binding process of a ligand to a receptor on a (sub-) molecular level, which provides us with new insight on how single ligands and receptors interact and how this affects the activity of the receptor. For example, coupling between receptors, as in dimerization5 _{or even between different types of receptors}6_{, seems to}

have a large influence on the activity of the receptor. At present, the molecular mechanisms at play are not known. These new developments hold a significant promise for future biology and medicine. Being able to study receptor behavior on a single receptor level may, for instance, ultimately provide drug development researchers with new strategies to design more effective treatment with less side-effects, by targeting only those receptors that are involved in the disease.

Figure 1 Schematic overview of the adaptations necessary to perform high-resolution molecular recognition force microscopy with an AFM on a receptor-ligand system.

Sugar binding pockets 5 nm

Ligand on AFM tip: mannose Short spacer: carbon nanotube OHO HOHO O HO O H O H O HO O H O

Sugar binding pockets 5 nm

Historical background of AFM in receptor-ligand research

Discrete interactions between single ligand and receptor molecules in liquid environment have been measured using AFM with chemically functionalized tips. The first papers in this field demonstrated discrete (strept-) avidin-biotin interactions7,8,9_{. In this}

early work, AFM tips were completely coated with ligands. Later, the interaction between e.g. antibodies and intercellular adhesion molecules10_{, and lectin and red blood cells}11 _{has been studied. Also,}

single molecular interactions in cell-cell adhesion12 _{and the}

cooperativity of molecular adhesion13 _{have been studied. Several}

reviews have been published on biomolecular interactions measured with AFM; see for instance Willemsen14 _{and Allison}15_{. Hinterdorfer et} al. have pioneered the use of a single flexible linker between the AFM

tip and a coupled ligand16_.

Although single-molecule interactions were detected, in most of the above studies the lateral resolution was not optimal. For example, in reference 11, the unbinding events were correlated with the position on the surface with a lateral resolution of 400 nm. A breakthrough came when Raab et al. combined molecular recognition with dynamic AFM17_{, a technique called molecular recognition force}

microscopy (MRFM)18_{. The lateral resolution was improved from 400}

nm to 20 nm and the imaging speed was increased from several tens of minutes to a few minutes for a scan of 500 nm x 500 nm. Recently, single-molecule recognition imaging of histones in nucleosomal arrays was realized.19

Sugar-lectin interactions in this thesis

1.3 Atomic force microscopy

The atomic force microscope was invented by Binnig, Quate and Gerber in 1986.20 _{A sketch of a commonly used AFM design is shown}

in Figure 2.

Shortly, in an atomic force microscope, a very sharp needle at the end of a cantilever is scanned over a surface. The deflection of the cantilever is recorded by a laser beam that is reflected off the cantilever and detected by a 4-segment photodiode. In constant force contact mode, the deflection is kept constant by a feedback system that adjusts the height of the sample surface. The vertical movement of the piezo element provides topographical information about the sample surface. Under optimal conditions, a lateral resolution of 1 nm and a vertical resolution of 0.1 nm can be obtained, when scanning in liquid.

Initially, the AFM was mostly used to image surfaces in air, but soon it was modified in order to make it possible to use the AFM in liquid.21 _{This makes the AFM the ideal tool to study the relationship}

between microstructure and function, because it allows imaging biological samples with (sub-) molecular resolution, while imaging can be performed in liquid, which prevents the biological samples from denaturation.

Piezo scan tube

Cantilever Metal disk Mica Molecules Plastic ring Laser source Mirror 4 Segment photo- detector 10 µm Tip

Molecular recognition force microscopy

In addition to protein dynamics and (sub-) molecular topographical information, the AFM can also provide chemical contrast. Chemical force microscopy (CFM)22,23,24,25 _{is realized by}

binding a molecule to the AFM tip, scanning it over a surface that contains other molecules, and detecting the interaction between the molecule on the tip and molecules on the surface. CFM can be used to study the specific interaction between ligands and receptors, a technique which is called molecular recognition force microscopy (MRFM).26,27 _{The first MRFM experiment was performed in the group}

of Hinterdorfer in Linz. Lysozyme molecules were probed with an AFM tip containg an antibody against lysozyme. The lateral resolution was approximately 25 nm as is shown in Figure 3. Important here is the

role of the spacer, which couples the ligand to the AFM tip and provides it with rotational freedom. Several polymers have been used as a spacer. A high density of spacers on the tip resulting in a tip containing many ligands was used to probe multiple interactions.28 _A

low density of spacers on the tip, resulting in a tip containing only a few ligands, was used to probe single molecular interactions.29

We aimed for a higher lateral resolution (< 5 nm), which implied using a short spacer, although it should remain flexible enough to let the ligand find and bind the binding site. Decreasing the length of the spacer makes the design for a functionalized tip more

difficult, because the ligand should be attached to the very apex of the tip, as is illustrated in Figure 4.

In order to make sure that the ligand can reach the surface, a tip with a very sharp end that has chemical properties different from the rest of the tip should be used, making it possible to bind the ligand to this sharp tip end. We chose to use a carbon nanotube as a spacer, because it has a diameter in the order of 2-10 nm. Furthermore, covalent chemistry can be used to bind a ligand to the nanotube tip end. Using a carbon nanotube as a spacer makes it possible to position a ligand at the very apex of the tip.

1.4 Carbon nanotubes

Carbon nanotubes are wrapped up sheets of graphene with special mechanical and electronic properties.30 _{Carbon nanotubes}

have a Young modulus of the order of 1 TPa, and they buckle elastically under applied load rather than to fracture, which makes carbon nanotubes a uniquely tough and energy-absorbing material.31

Carbon nanotubes are members of the fullerene family, molecules that were discovered by Kroto, Curl and Smalley, who were awarded the Nobel Prize in chemistry in 1996 for this discovery.32,33 _Carbon

nanotubes were discovered by Iijima in 1991, and the first TEM images of nanotubes are shown in Figure 5.34

Figure 4 Schematic picture of chemically modified AFM tips, using a long spacer (a) and a short spacer (b). As can be seen in (b), the short spacer has to be positioned at the very tip end, otherwise the ligand will not be able to reach the receptor on the surface.

Why are carbon nanotubes interesting for us? The high stiffness of the nanotubes makes these molecules a good candidate to use them as a linker between an AFM tip and a molecule that one wants to bind to the tip. Secondly, an open ended carbon nanotube that reacted with oxygen, has a carboxylic acid end group, and can be chemically functionalized in a covalent way, using amide chemistry.35

Carbon nanotubes can be chemically functionalized by binding molecules to the walls or to the end of the nanotube. In the first case, antibodies against fullerenes, that also bind to nanotubes, can be used to functionalize nanotube walls.36 _{Or, nanotubes can be}

coated with polymers that are receptive to certain proteins.37 _For

scanning probe applications, the end of a nanotube can be functionalized covalently, as was developed by Lieber.35,38 _{For this}

covalent functionalization, the nanotube should have an open end.

1.5 Outline of this thesis

This thesis consists of eight chapters. In Chapters 2 and 3, the application of AFM to biological systems is explored, both in air and in liquid. Chapter 2 deals with atomic force microscopy in air and, as an example, contains a study of filaments found in dividing yeast cells. Chapter 3 describes atomic force microscopy in liquid, both in contact mode and in tapping mode.

In Chapters 4, 5, and 6 we discuss all steps that have to be taken in order to make the high-resolution MRFM experiment described in Chapter 7 possible. As a first step, Chapter 4 provides a method for immobilizing well-separated single molecules for AFM studies in liquid. Next, in Chapter 5, the use of carbon nanotubes as AFM tips for molecular recognition force microscopy is discussed, and a description is given of the fabrication process of single-walled and multi-walled carbon nanotube AFM tips. In Chapter 6, the last step is discussed: chemical characterization and functionalization of carbon nanotubes. In Chapter 7, all steps come together, and first results of molecular recognition force microscopy experiments on pea lectin and on mannan binding lectin are shown and discussed.

To conclude, Chapter 8 provides a summarizing discussion and an outlook.

1 _{Kasas, S., Thomson, N.H., Smith, B.L., Hansma, H.G., Zhu, X., Guthold, M.,}

Bustamante, C., Kool, E.T., Kashlev, M., and Hansma, P.K., Biochem 36 (1997) 461-468

2 _{Viani, M.B., Pietrasanta, L.I., Thompson, J.B., Chand, A., Gebeshuber, I.C., Kindt,}

J.H., Richter, M., Hansma, H.G., and Hansma, P.K., Probing protein-protein interactions in real time, Nature Structural Biology 7 (8) (2000), 644-647

3 _{Engel, A., and Müller, D.J., Observing single biomolecules at work with the atomic}

force microscope, Nature structural biology 7 (9) (2000), 715-718

4 _{Dufrêne, Y.F., Using nanotechniques to explore microbial surfaces, Nature reviews 2}

(2004), 451-460

5 _{Blab, G.A., Single-molecule techniques in biological and biophysical research, PhD}

thesis Leiden University (2004)

6 _{Mello, B.A., and Tu, Y., Quantitative modeling of sensitivity in bacterial chemotaxis:}

The role of coupling among different chemoreceptor species, PNAS 100 (2003), 8223-8228

7 _{Lee, G.U., Kidwell, D.A., and Colton, R.J., Sensing discrete streptavidin-biotin}

interactions with atomic force microscopy, Langmuir 10 (1994), 354-357

8 _{Florin, E.L., Moy, V.T., and Gaub, H.E., Adhesion forces between individual}

ligand-receptor pairs, Science 264 (1994), 415-417

9 _{Moy,V.T., Florin, E.L., and Gaub, H.E., Intermolecular forces and energies between}

ligands and receptors, Science 266 (1994), 257-259

10 _{Willemsen, O.H., Snel, M.M.E., Werf, K.O. van der, Grooth, B.G. de, Greve, J.,}

Hinterdorfer, P., Gruber, H.J., Schindler, H., Kooyk, Y. van, and Figdor, C.G.,

Simultaneous height and adhesion imaging of antibody-antigen interactions by atomic force microscopy, Biophysical Journal 75 (1999), 2220-2228

11 _{Grandbois, M., Dettmann, W., Benoit, M., and Gaub, H.E., Affinity imaging of red}

adhesion measured by single-molecule force spectroscopy, Nature Cell Biology 2 (2000), 313-317

13 _{Chen, A., and Moy, V.T., Cross-linking of cell surface receptors enhances}

cooperativity of molecular adhesion, Biophysical Journal 78 (2000), 2814-2820

14 _{Willemsen, O.H., Snel, M.M.E., Cambi, A., Greve, J., Grooth, B.G. de, and Figdor,}

C.G., Biomolecular interactions measured by atomic force microscopy, Biophysical Journal 79 (2000), 3267-3281

15 _{Allison, D.P., Hinterdorfer, P., and Han, W., Biomolecular force measurements and}

the atomic force microscope, Current Opinion in Biotechnology 13 (2002), 47-51

16 _{Hinterdorfer, P., Baumgartner, W., Gruber, H.J., Schilcher, K., and Schindler, H.}

Detection and localization of individual antibody-antigen recognition events by atomic force microscopy, PNAS 93 (1996), 3477-3481

17 _{Raab, A., Han, W., Badt, D., Smith-Gill, S.J., Lindsay, S.M., Schindler, H., and}

Hinterdorfer, P. Antibody recognition imaging by force microscopy, Nature Biotechnology 17 (1999), 902-905

18 _{Hinterdorfer, P. in Methods in cell biology 68, Eds.: Jena, B.P., Hörber, H., Elsevier}

Science New York (2002), 115

19 _{Stroh, C., Wang, H., Bash, R., Ashcroft, B., Nelson, J., Gruber, H., Lohr, D., Lindsay,}

S.M., and Hinterdorfer, P., Single-molecule recognition imaging microscopy, PNAS 101 (2004), 12503-12507

20 _{Binnig G, Quate CF, Gerber C: Atomic force microscope, Phys Rev}

Lett 56 (1986), 930-933.

21 _{Drake, B., Prater, C.B., Weisenhorn, A.L., Gould, S.A.C., Albrecht, T.R., Quate, C.F.,}

Cannel, D.S., Hansma, H.G., and Hansma, P.K., Imaging crystals, polymers, and processes in water with the atomic force microscope, Science 243 (1989), 1586-1588

22 _{Lee, G.U., Kidwell, D.A., and Colton, R.J., Sensing discrete streptavidin-biotin}

interactions with atomic force microscopy, Langmuir 10 (1994), 354

23 _{Florin, E.L., Moy, V.T., and Gaub, H.E., Science 264 (1994), 415}

24 _{Wong, S.S., Joselevich, E., Woolley, A.T., Cheung, C.L., and Lieber, C.M., Covalently}

functionalized nanotubes as nanometre-sized probes in chemistry and biology, Nature

394 (1998), 52-55

25 _Mc_{Kendry, R., Theoclitou, M.-E., Rayment, T., and Abell, Ch., Chiral discrimination by}

chemical force microscopy, Nature 391 (1998), 566-568

26 _{Raab, A., Han, W., Badt, D., Smith-Gill, S.J., Lindsay, S.M., Schindler, H., and}

Hinterdorfer, P. Antibody recognition imaging by force microscopy, Nature Biotechnology 17 (1999), 902

27 _{Hinterdorfer, P. in Methods in cell biology 68, Eds.: Jena, B.P., Hörber, H., Elsevier}

Science New York (2002), 115

28 _{Grandbois, M., Dettmann, W., Benoit, M., and Gaub, H.E., Affinity imaging of red}

blood cells using an atomic force microscope, The Journal of Histochemistry & Cytochemistry 48(5) (2000), 719

29 _{Hinterdorfer, P., Baumgartner, W., Gruber, H.J., Schilcher, K., and Schindler, H.}

Detection and localization of individual antibody-antigen recognition events by atomic force microscopy, PNAS 93 (1996), 3477

30 _{Dekker, C., Carbon nanotubes as molecular quantum wires, Physics Today May}

(1999), pp22-28

31 _{Wong, E.W., Sheehan, P.E., and Lieber, C.M., Nanobeam mechanics: Elasticity,}

strength, and toughness of nanorods and nanotubes, Science 277 (1997), 1971-1975

32 _{Kroto, H.W., Heath, J.R., O’Brien, S.C., Curl, R.F., and Smalley, R.E., C} 60:

Buckminsterfullerene, Nature 318 (1985), 162-163

33 _{Thess, A, Lee, R., Nikolaev, P., Dai, H., Petit, P., Robert, J., Xu, Ch., Lee, Y.H., Kim,}

S.G., Rinzler, A.G., Colbert, D.T., Scuseria, G.E., Tománek, D., Fischer, J.E., and Smalley R.E., Science 273 (1996), 483-487

34 _{Iijima, S. Helical microtubules of graphitic carbon, Nature 354, 1991, 56-58} 35 _{Wong, S.S., Joselevich, E., Woolley, A.T., Cheung, Ch.L., and Lieber, C.M.,}

monoclonal antibody to single wall carbon nanotubes, Nano Letters 1(2001), 465

37 _{Shim, M., Kam, N.W.S., Chen, R.J., Li, Y., and Dai, H., Functionalization of carbon}

nanotubes for biocompatibility and biomolecular recognition, Nano Lett. 2, 285-288 (2002)

38 _{Hafner, J.H., Cheung, C.-L. Woolley, A.T., and Lieber, C.M., Structural and functional}

Chapter 2

An AFM study of FIN filaments:

filaments in between the nuclei of

dividing Saccharomyces cerevisiae

(yeast) cells

Fin is a protein that is present in yeast cells. During cell division, Fin forms filaments that co-localize with microtubuli. Two questions were addressed in this chapter: Can Fin polymerize on its own, or does it use microtubuli, or possibly DNA or RNA as a template? Does Fin need ATP or GTP for polymerization?

By AFM imaging in air, we found that Fin can polymerize on its own, in the absence of tubulin, DNA and RNA. No GTP or ATP was needed for polymerization. We show that two different types of Fin filaments exist: a rigid filament, in which a periodic substructure was resolved, and a flexible filament. The rigid filament type has a subunit length of around 13 nm and a measured width of 9 nm. This measured width which corresponds closely to the measured width of 10 nm that was found with the electron microscope. Rigid filaments could be bundles or super-coils of single FIN filaments.

This Chapter is partly based on:

Martijn J. van Hemert, Gerda E.M. Lamers, Dionne C.G. Klein, Tjerk. H. Oosterkamp, H. Yde Steensma, and G. Paul van Heusden

The Saccharomyces cerevisiae Fin1 protein forms cell cycle-specific filaments between spindle pole bodies

2.1 Introduction

Filaments in between the nuclei in dividing cells of

Saccharomyces cerevisiae (baker’s yeast) were discovered in the

yeast group at the Institute of Biology in Leiden. Several aspects such as the interaction between these Fin-filaments and well-studied microtubuli were examined.

From confocal fluorescence images, it is known that Fin protein is present in a non-filamentous form in small-budded yeast cells, and in a filamentous form in large budded yeast cells. Confocal fluorescence images now show that the Fin filaments co-localize with microtubuli, but no evidence for a direct interaction between the Fin proteins and Tub proteins was found. Electron microscopy studies demonstrated that Fin protein polymerizes in the absence of tubulin, DNA and RNA, and ATP nor GTP is required for polymerization. With the AFM we wanted to get more information on the structure of the Fin filaments, and several control experiments were done in order to check if the filaments were indeed composed of Fin, which had not been checked for the electron microscopy results.

In the next section, a short introduction to cell division of yeast is given, especially focusing on the role of microtubuli.

Yeast cell division

We studied cell division of Saccharomyces cerevisiae (baker’s yeast), which is a small, single-celled fungus. Although yeast is a eukaryotic cell, cell division of yeast is different from normal eukaryotic cell division. Instead of cell enlargement during interphase

(S-phase) when the chromosomes replicate, a bud emerges from the mother cell that will eventually become the daughter cell, as is shown in Figure 1.

When the bud is nearly as big as the mother cell, the nucleus divides (this happens later than in other eukaryotic cells), and two cells form after separation of the bud. At this point, the cell cycle starts over again. The role of microtubuli in cell division is described shortly in the next section.

Microtubuli

Like in other eukaryotic cells in the prophase of cell division, microtubuli form the mitotic spindle to enable the separation of the sister chromatids, such that the mother and daughter cell both get one copy of the chromosomes, as is illustrated in Figure 2. In addition

to the well-known microtubuli, P. van Heusden et al. found another type of filament, which was called Fin1: filaments in between nuclei1_.

Fin1 filaments form directly after formation of the mitotic spindle, align with it, and disappear when the cell cycle is completed. In dividing cells, Fin1 proteins are present in the nucleus in a non-filamentous form. The next section gives a short overview of everything that is known up to now about the Fin1 protein.

a b

Figure 2 (a) Phase-contrast light microscopy image of a dividing cell during metaphase, clearly showing the chromosomes that are aligned between the two mitotic spindles (b) Schematic drawing of a mitotic spindle, showing the two spindle poles and the chromosomes in between. Copyright © 1998 from Essential Cell Biology by Bruce Alberts et al. Reproduced by permission of Garland Science/ Taylor & Francis Books, Inc.

Sister chromatids

Fin1 protein

Fin1 is a 30 kDa protein that mainly consists of α-helices, and putative coiled-coil regions. It has a remarkably high iso-electric point (pI) of 10, which means that it is a basic protein. The C-termini of Fin monomers show hydrophobic interaction. How the Fin monomers form a filament and what the structure (linear, helical,…) of a Fin filament is, is not known.

Confocal fluorescence microscopy

In order to find the localization of Fin relative to microtubules, localization of both Fin1 proteins and tubulin (Tub1) proteins in yeast cells was visualized using confocal fluorescent microscopy. Fin1 protein was tagged with cyano fluorescent protein (CFP) and Tub1 protein was tagged with green fluorescent protein (GFP). CFP was excited at 457 nm and emission was detected at 464-500 nm, and GFP was excited at 488 nm and emission was detected at 464-500 nm. In Figure 3, yeast cells that are in different stages of the cell cycle can be seen. Small-budded cells are at the start of cell division.

Figure 3 a shows the fluorescence of Fin1, and Figure 3 c shows the fluorescence of Tub1. In small-budded cells, CFP-Fin1 protein is non-filamentous, while GFP-Tub1 protein is present in small filaments. In Figure 3 b, Figure 3 a and c are superimposed, in order to show that Fin1 protein and Tub1 protein co-localize: In large-budded cells, both filaments of CFP-Fin1 protein (Figure 3 a, upper cells) and GFP-Tub1 protein Figure 3 c are present with apparent co-localization.

Electron microscopy

A drop of 10 µl containing 200 ng of Fin was put on a carbon grid, dried and stained with 2% phosphotungstic acid. With the electron microscope, images were taken and filaments with a width of around 10 nm were found, as is shown in Figure 4.

Goals

In order to learn more about the Fin1 protein and its polymerization process, we would like to answer two fundamental questions:

1. Can Fin1 polymerize on its own, or does it use microtubuli, or possibly DNA or RNA as a template?

2. Does Fin1 need ATP or GTP for polymerization?

With AFM in air, we were able not only to answer these questions, but we also found two different types of structure of the Fin1 filaments: a rigid and a flexible type.

2.2 Methods

Fin 1 preparation

Fin 1 monomers were isolated from yeast cells by vortexing the cells together with glass beads in an 8M Urea and 0.5 M NaCl solution, in order to open the cells and denature the proteins, which means that Fin filaments are not present anymore, only monomers. Fin1 proteins were separated using Ni-NTA-agarose affinity

chromatography, where Ni binds to a sequence of 6 histidines that were inserted in Fin1. After this, Fin1 proteins were dialyzed against low salt, in order to remove Urea. This renatures the protein, and filaments can possibly be formed. For single molecule studies with the AFM, it is crucial that the protein solutions that are used are extremely pure, especially when the impurities tend to absorb very well to mica. Otherwise it will be very hard to determine what kind of, and how many different proteins are present on the test surface.

Sample preparation

Fin is a basic protein (pI=10), which means that it is positively charged at neutral pH. This implies that the proteins can be immobilized electrostatically on freshly cleaved mica, which is negatively charged. A droplet of 10 µl containing 200 ng of Fin1 protein in 25 mM NaCl/5mM sodium phosphate buffer at pH 7.5 was applied on a freshly cleaved mica surface, rinsed with deionized water after 30 s and dried in a nitrogen flow.

Atomic force microscopy

To investigate if the Fin1 protein can polymerize on its own, without the help of tubulin, DNA, or RNA present, Fin filaments from yeast cells were broken down by denaturing the proteins using Urea as was described above, and subsequently the Fin1 protein was renatured before AFM investigation.

Two types of Fin1 filaments were found in AFM studies performed in air. A rigid filament was found, an example of which can be seen in Figure 5a, with a measured width of on average 9 nm, a height of 1.5 nm, and a subunit length of 13 nm. The measured width is a convolution between the tip diameter and the filament diameter, and this is one of the reasons that the measured width exceeds the measured height. Another reason is that the filament is attached to the surface, which influences the height-width ratio. Finally, the presence of a thin water layer reduces the measured height in AFM imaging in air. Also a flexible type of filament was found, of which an example is depicted in Figure 5b.

For the flexible filament type, we found a measured width of on average 9 nm, and a height of 0.6 nm. We found both types of filaments together on one sample surface (Figure 5 c), which implies that different filament formation is not concentration dependent2_.

Control experiments

In order to make sure that the filaments that were imaged by AFM were not DNA or RNA, we added the enzymes DNAse (DNAse: Fin, 1:20), that breaks up DNA, and RNAse (RNAse: Fin, 1:20), which breaks up RNA, to the Fin solution. After incubation, we applied the mixture to a mica surface. In these positive control experiments, we still observed both types of filaments as is shown in Figure 6.

Another positive control was performed by filtering the Fin solution, using a 300 kDa filter. This filter removes DNA and RNA that are possibly present in the Fin solution. After incubation of the filtrate on mica we still observed both types of filaments, which implies that the filaments that we are imaging are not DNA or RNA.

As a negative control, we incubated Fin monomers in the presence of Urea, which means that the Fin proteins are denatured and no filaments should be able to form. Indeed, no filaments were found by AFM.

a b

c

2.3 Discussion and conclusion

Care has to be taken when interpreting the AFM images, because we look at Fin that is immobilized on a surface, and immobilization changes the shape that the filaments have compared to their shape in solution. When a long and flexible polymer is immobilized from solution onto a surface, the transition from three to two dimensions implies loss of one degree of freedom. This results in a large reduction in the number of possible configurations that the polymer can access. Transport from solution onto a surface can happen relatively slowly, which is called equilibration, or relatively fast, called kinetic trapping.3 _{Equilibration implies that the transport is}

determined only by diffusion. In this case, the shape of the immobilized filaments (2D) resembles the shape of the filaments in liquid (3D), and it is possible to extract valuable information about the polymer such as the persistence length.

But, as Fin is a basic protein and therefore positively charged at low pH, we expect that the protein immediately binds to the negatively charged mica through kinetic trapping. This makes it more difficult to interpret the images, because in addition to intrinsic conformations, also surface-induced conformations are present. For Figure 5b for instance, one could imagine that the filament landed on the surface at a few points first, and then the rinsing moved the filament into one direction (see the arrow in Figure 5 b), creating the form of a chain that is pinned at several points. In the case of kinetic

Figure 6 AFM images taken in tapping mode in air of Fin filaments after treatment with (a) DNAse and (b) RNAse. The enzymes are clearly resolved against the background. Scales are (a) 580nmx 580nm x 3.8nm and (b) 800nmx 800nmx 3.5nm.

trapping, it is still possible to extract some information. The length of the subunits of a filament in 3D is the same as that in 2D.

To conclude: From the AFM images we learned that Fin can polymerize in vitro on its own, without the help of tubulin, without a DNA or RNA template, without ATP or GTP. It forms two types of filaments, a rigid and a flexible filament. The rigid filament type has a subunit length of around 13 nm and a measured width of 9 nm. This measured width which corresponds closely to the measured width of 10 nm that was found with the electron microscope.

Rigid filaments could be bundles or super coils of single FIN filaments. An important remark here is that charges on the FIN filaments could prevent them from aggregating. But, ions from the buffer could screen the charges on the filaments and in this way make aggregation or super coiling possible.

Although the presence of Fin filaments in yeast cells has been demonstrated, the role of Fin in the yeast cell is unclear up to now. Knock-out experiments show that yeast cell-division is unaltered in the absence of Fin. Maybe Fin can take over the function of the microtubuli in extreme conditions, such as abnormal pH values or temperature. Now that we have verified that FIN forms filaments on its own, FIN can indeed have a back-up function and take over the role of microtubuli. It would be interesting to see if for extreme temperatures or pH-values, yeast cells can still divide normally.

This chapter shows the value of AFM use with unmodified tips in air for the study of protein structure. However, study of structure-function relationships of proteins requires use of AFM in liquid. This will be discussed in the next chapter.

1 _{Hemert, M.J. van, Lamers, G.E.M., Klein, D.C.G., Oosterkamp, T.H., Steensma, H.Y.,}

and Heusden, P.H. van. The Saccharomyces cerevisiae Fin1 protein forms cell cycle-specific filaments between spindle pole bodies. PNAS 99 8 (2002), 5390-5393

2 _{Van Noort, J., Verbrugge, S, Goosen, N., Dekker, C., Thei Dame, R., Dual}

architectural roles of HU: Formation of flexible hinges and rigid filaments, PNAS 101 (2004), 6969-6974

3 _{Rivetti, C., Guthold, M., and Bustamante, C., Scanning force microscopy of DNA}

Chapter 3

Imaging biological molecules and

membranes under physiological

conditions

In this chapter, the use of atomic force microscopy in liquid is discussed. Obtaining high-resolution images is facilitated by tuning the concentration and valency of ions in the buffer in order to reduce the electrostatic forces between tip and sample, and in this way making the tip susceptible only to short range forces, which contain information about the atomic/molecular scale structure of the sample. Several examples are shown in order to illustrate the use of AFM in liquid.

AFM contact mode images of fragments of native purple membrane are shown. In these images the bacteriorhodopsin trimers, which arrange in a hexagonal lattice, are resolved. High-resolution images were also obtained on light-harvesting complex 2, an important unit in the photosynthetic apparatus.

In tapping mode, single pea lectin dimers were imaged. Pea lectin is a sugar binding protein, which binds symbiotic Rhizobium bacteria to root hairs of the host plant. On a larger scale, whole Chinese hamster ovary cells were imaged, using magnetic driving of the cantilever.

This chapter is partly based on:

Stamouli, A., Kafi, S., Klein, D.C.G., Oosterkamp, T.H., Frenken, J.W.M., Cogdell, R.J., and Aartsma, Th.J.,

The ring structure and organization of light harvesting 2 complexes in a reconstituted lipid bilayer, resolved by atomic force microscopy

3.1 Introduction

Atomic force microscopy (AFM) can be carried out in liquid and at room temperature or body temperature, which provides the opportunity of studying biomolecules under physiological conditions.1,2 _{Two different modes of operation can be used. In}

contact mode, the cantilever continually touches the surface while scanning. This implies that a lateral force is exerted on the sample, which can damage soft samples or remove loosely bound molecules from the surface. Membrane fragments that are absorbed tightly to the surface can be imaged without damage in contact mode, as long as the force applied by the tip is kept as low as possible. As was discussed in § 1.3, in constant force contact mode, the deflection is kept constant by a feedback system that adjusts the height of the sample surface.

In order to reduce lateral forces, tapping mode AFM was developed.3,4 _{In tapping mode, the cantilever is oscillated during}

scanning, so the contact between tip and sample is intermittent, and in-plane forces are very much reduced. Tapping mode is used e.g. to image single proteins in liquid, because this type of sample is very sensitive to lateral forces, which can remove the proteins from the surface. In tapping mode, feedback is done on the oscillation amplitude of the cantilever.

The resolution that can be obtained both in contact mode and in tapping mode is approximately 0.1 nm vertically and 1 nm laterally.5,6,7,8,9

We show AFM images of several different biological systems. Native purple membrane fragments were imaged in contact mode. Light-harvesting complex 2 (LH2) transmembrane proteins, reconstituted in a lipid bilayer, were also imaged in contact mode. Pea lectin dimers were imaged in tapping mode, and Chinese hamster ovary cells were imaged in magnetic AC mode.

3.2 Contact mode in liquid

High resolution AFM images are obtained only if the tip can approach the sample close enough to be able to “feel” the short-range interactions that provide submolecular resolution10_{. In order to}

accomplish this situation, electrostatic repulsion between tip and sample has to be minimized, as is depicted in Figure 1 a. This can be done by adjusting the valency and the concentration of ions in the buffer, as illustrated in Figure 1 b. The repulsive electrostatic interaction decreases with increasing ion concentration. Adding a multivalent ion, in this case Mg2+_{, helps to inhibit long-range}

The thickness of the diffusive electrical double layer formed by counterions is defined by the Debye length, as described in Equation 3.1,

∑

=

q

c

e

T

k

ε

ε

λ

where ε0 is the permittivity of the vacuum, εe is the dielectric constant

of the medium (in this case water), kB is Boltzmann’s constant, T is

the absolute temperature, e is the unit charge, and ci and qi are the

concentration and charge of the i-th electrolyte component. Remarkable is the strong influence of the valency of the ions on the thickness of the double-layer, which explains why a buffer solution containing e.g. Mg2+ _{can screen surface charges very efficiently.}

The distance between tip and sample at zero force is the sum of the Debye length of tip and surface, as is shown in Figure 1 b. As soon as the double layers of tip and sample start to overlap, an electrostatic interaction occurs. The distance between tip and sample for an applied force can be estimated from force-distance curves. For the upper curve in Figure 1 b, the distance between tip and sample at 100 pN is approximately 18 nm, while for the 300 mM curve this distance is 2 nm. For the lowest curve, the tip touches the sample at a force that is even lower than 100 pN. A force of only 25 pN is required to bring the tip into contact with the membrane, which means that the repulsive double layer force is only 25 pN.

At smaller distances, in addition to the electrostatic force, the Van der Waals force should be taken into account. This is described by the DLVO (Derjaguin-Landau-Verwey-Overbeek) forces, but effects of ionic radius, hydration forces, steric forces and specific interactions are neglected. The DLVO force between a local protrusion on the AFM tip and a spherical protein of comparable radius, as well as the DLVO force between a macroscopic sphere interacting with a planar surface are given in Reference 10.

Distance z between AFM tip and sample (i.e. purple membrane) at an applied force of 100 pN:

(20 mM) z= 18 nm

(300 mM) z= 2nm

(50 mM) z= 0 nm

Figure 1 (a) Sketch showing the short–range Van der Waals interaction and long-range electrostatic interaction between AFM tip and sample (b) Force-distance curves taken on purple membrane, for different electrolyte concentrations. At arrow (1), the electrostatic repulsion between tip and sample starts, and at arrow (2) tip and sample are in contact. The dotted lines show force-distance curves on purple membrane in the absence of repulsive interaction. Reproduced by permission of Daniel Müller, Reference 10.

b a

λDtip+λDsample

Force

Native purple membrane

The cytoplasmic membrane of purple bacteria, called purple membrane, contains bacteriorhodopsin, a light-driven proton pump, which helps to regulate the concentration of protons in the bacterial cell by pumping protons out of the cell. Bacteriorhodopsin is a transmembrane protein that is composed of seven α-helices, connected by loops, as shown in Figure 2.11 _{Three bacteriorhodopsin}

molecules together form a trimer, and these trimers arrange into a hexagonal lattice.

We imaged purple membrane with the AFM in contact mode, under physiological conditions, resolving the hexagonal arrangement of the bacteriorhodopsin trimers, as shown in Figure 3. We noticed that two parameters were crucial for high-resolution imaging. First, the membrane fragments had to be firmly absorbed to the surface, as will be described in the next section. Secondly, the force applied to the membrane with the scanning tip had to be kept as low as possible, in order to prevent disruption of the membrane. For this reason, we chose a cantilever with a spring constant of 0.06 N/m. Furthermore, the setpoint was manually adjusted during scanning to compensate for thermal drift that changes the distance between tip and sample. As can be seen in parts of Figure 3 a, sometimes the force applied on the sample was too low, resulting in a loss of contact between tip and sample, which caused streaky scan lines or even complete loss of the image. We measured an average lattice constant of 6.4 nm, which is close to the value of 6.2 nm published by Mϋller

et al. 5

Figure 2 Structure of bacteriorhodopsin. (a) Schematic picture of the seven transmembrane helices that are connected by loops. (b) Ribbon representation of bacteriorhodopsin, a top view and a bottom view showing the position of the helices and the loops in the membrane. Reproduced by permission of Filipp Oesterhelt, Reference 11.

Experimental details

Purple membrane patches that were isolated from

Halobacterium halobium were kindly provided to us by Dr. D.J. Müller

from Basel (now in Dresden).12 _{We absorbed the native}

membrane patches on mica as described below. Both for absorption of the membrane on the mica surface as well as for high resolution imaging, we used a buffer solution with monovalent ions at a concentration of 300 mM. Sample preparation was as follows. Purple membrane with a concentration of 6 µg/µl was kept in a -80 °C freezer. Immediately before sample preparation, it was defrosted and diluted 25 times in 300 mM KCl, 10 mM Tris (Tris (hydroxymethyl) aminomethane), pH=7.6. A 75 µl droplet of diluted purple membrane was put on a freshly cleaved mica surface. This sample was stored at 4 °C during 20 min. The sample was rinsed with buffer (6 times), while preventing drying of the sample.

Figure 3 Purple membrane imaged in contact mode in liquid (k=0.06 N/m). Scales: (a) 300 nm x 300 nm x 10 nm, (b) 65 nm x 65 nm x 3 nm. Image (b) is a software zoom of the squared area in (a). In (b), the bacteriorhodopsin monomers in a trimer are circled.

As was discussed above, high resolution imaging requires a good choice of buffer solution. When using a buffer of 300 mM KCl, 10 mM Tris, pH 7.6, the forces between tip and sample were balanced quite well, as is shown in Figure 4. The hysteresis in Figure 4 between the approach and retraction curves is caused by slight drift of the position of the tip relative to the sample.

Another critical point for high resolution imaging is the AFM tip shape. We were not able to achieve high resolution for all the tips that we used. Unfortunately, the quality of the tip can only be assessed via the imaging itself, and in addition to that, it can even change during imaging because the tip can pick up some contamination or a protein from the sample. Also, the tip can become blunt when too high forces are applied. AFM tips with a better defined tip end are e.g. oxide sharpened tips, but these tips also have the problems mentioned above. Another example of better defined tips are carbon nanotube tips (see Chapters 5). Carbon nanotube tips will not wear, because they buckle elastically under applied load and thus will stay sharp13_{. The contamination of the nanotubes will be different}

from that of silicon oxide or silicon nitride tips, because of the hydrophobic nature of these tips.

Figure 4 Force-distance curve taken on a fragment of purple membrane in 300 mM KCl, 10 mM Tris buffer solution at pH 7.6. The distance between the AFM tip and the membrane at an applied force of ≥ 100 pN is 0 nm, which is required for high-resolution imaging. Approach curve is grey, retraction curve is black. The dashed line shows force-distance curves on purple membrane in the absence of repulsive interaction.

Force 250

pN/div

Z-piezo displacement 23 nm/div 300 mM KCl, 10 mM Tris, pH 7.6

Light harvesting 2 complex (LH2)

As a second sample, we studied the transmembrane protein LH2 of the purple photosynthetic bacterium Rhodopseudomonas

acidophila. LH2 is a light-harvesting complex which is part of the

photosynthetic system. In purple bacteria, the photosynthetic unit is localized in a system of intracytoplasmic membranes. The photosynthetic complex transfers solar energy into chemical energy, by producing ATP, which is the molecule that stores the energy used for many forms of biological work, from protein synthesis to muscle contraction. The LH2 complex absorbs light, and transfers the energy to light-harvesting complex 1, which is associated with a reaction centre. In the reaction centre, a charge separation occurs, which is followed by electron transport across the membrane. This electron transfer establishes a proton gradient over the membrane, which in the end results in production of ATP. The precise molecular mechanism of this electron transfer is not known. Using the ability of conductive AFM to combine high-resolution imaging with electrical sensitivity, the process of electron transport can be followed on a molecular scale. Using this type of AFM, a clear difference between the electron conduction on LH2 and on reaction centers was measured by Stamouli et al.14

In order to be able to see how the electron conduction varies within one LH2 complex, and identify where in the complex most of the conduction takes place, high-resolution imaging under physiological conditions needs to be made possible. LH2 has to be inserted in a bilayer, and this bilayer needs to be stably absorbed to a surface. In the next section it will be discussed how high-resolution imaging of LH2 was realized, and AFM images will be shown and discussed.

LH2 complexes were reconstituted in a lipid bilayer (egg phosphatidylcholine), and imaged in a buffer containing 10 mM Tris-HCl, pH 7.2, 150mM KCl, 25 mM MgCl2. Details are described in

Reference 6. A buffer containing only monovalent ions turned out not to be sufficient to absorb the lipid bilayer containing LH2 to mica. From Formula 3.1, we know that the Debye length is inversely proportional to the charge squared. This implies that reducing the thickness of the double layer, which results in better absorption of the bilayer to the surface, can be done more efficiently when using bivalent ions. For this reason, we used a buffer solution containing Mg2+ _{to absorb the membrane fragments to mica.}

Si3N4 cantilevers with a spring constant of 0.06 N/m were

used, and the applied force was kept lower than 100 pN, by manually adjusting the setpoint to compensate for thermal drift, as was described above.

images, we found that the inner diameter of a ring of LH2 is 3.6 ± 0.2 nm, and the outer diameter is 7.3 ± 0.5 nm. One ring of LH2 consists of 9 subunits, which clearly can be seen in Figure 5 b.

3.3 Tapping mode in liquid

So far, we have shown AFM images of membrane fragments, recorded in liquid. Imaging single proteins in contact mode in liquid is more difficult. This is because membranes that are absorbed on a surface are more stable than single proteins that are absorbed on a surface, and the in-plane forces applied by the scanning tip can remove loosely bound proteins from the surface. In tapping mode, in-plane forces are very much reduced, by vibrating the cantilever while scanning over the surface. In tapping mode AFM, the cantilever is vibrated close to its resonance frequency, because in this regime the oscillation amplitude decreases linearly with the average tip-sample distance.15 _{When a cantilever is immersed in fluid, the hydrodynamic}

damping results in a lower resonance frequency (3-5 times)16, 17_{, and}

it decreases the Q-factor with several orders of magnitude.16

Figure 5 Light harvesting complex 2, reconstituted in a lipid bilayer, imaged in liquid in contact mode. The inner diameter of an LH2 ring was on average 3.6 nm, outer diameter 7.3 nm. Grey scale (a) 0-1.5 nm, (b) 0-2 nm. The applied force was kept lower than 100 pN. The image was made by Amalia Stamouli.

Acoustic driving

In tapping mode AFM, a cantilever is driven by a resonating piezo in the tip holder. The acoustic waves induced in the fluid medium cause the cantilever to be oscillated. A “forest of peaks” shows up in the frequency sweep, as is shown in Figure 6.

Several groups investigated the origin of the peaks in these resonance spectra. Putman et al. changed both the size of the liquid cell, as well as the level of liquid, and the type of liquid (water, alcohol) used in these experiments. They noticed that new resonances appear in the frequency characteristics. 18 _{From this, it}

was concluded that acoustic modes are present in the medium and that the cantilever is excited acoustically. Schäffer et al. found that the material of the fluid cell also influences the position and height of peaks in the spectrum. Furthermore, they found, very interestingly, that the cantilever response spectrum is a product of the fluid drive spectrum and the cantilever’s thermal noise spectrum. 19

In order to find the resonance frequency of the cantilever, which is quite hard to do because, as was discussed above, also many other frequencies show up in the frequency sweep, we can take the thermal spectrum of the cantilever. An example is shown in Figure 7. Knowing the resonance frequency from the spectrum shown in Figure 7, we can choose a driving frequency close to resonance.

Figure 6 Frequency sweep for a silicon nitride cantilever (k = 0.58 N/m) that is driven acoustically, in liquid (solid line), and in air (dashed line). Horizontal scale is a log scale from 10-100 kHz, vertical scale, also a log scale, are arbitrary units. Reprinted from Reference 18.

cantilever deflection

Pea lectin dimers

Pea (Pisum sativum) lectin is a sugar-binding protein that is present on the root hairs of the pea plant. It binds to sugar moieties on the surface of Rhizobium bacteria present in the soil, and in this way the bacteria stick to the root hairs. After attachment, rhizobia infect the roots and multiply in root nodules. This symbiotic interaction provides the plant with fixed nitrogen, and the bacteria with nutrients from the plant.

In order to study the interaction between sugar and pea lectin, we first immobilized single pea lectin dimers and imaged them with the AFM in liquid. In order to prevent the tip from scraping the proteins away, we used tapping mode. An example of an AFM image of pea lectin is shown in Figure 8. More experimental details are described in Chapter 4.

Figure 7 Thermal spectrum of a biolever (gold-coated silicon nitride, k = 0.03 N/m) in liquid. Frequency scale is from 2-16 kHz. Vertical scale are arbitrary units.

frequency cantilever

MAC mode: magnetic driving

A way to avoid the forest of peaks (see Figure 6) that appear in acoustic driving of a cantilever in liquid is driving the cantilever directly without shaking the whole fluid cell. This can be done by mounting a coil underneath the sample, and using AFM cantilevers with a magnetic coating. An oscillating field in the coil makes the cantilever vibrate, which results in just one peak in the frequency spectrum.20,21

Chinese hamster ovary (CHO) cells

A central goal in liquid AFM is to identify single molecules on a living cell. Before single molecular resolution on a living cell can be obtained with the AFM, stable images on a larger scale have to be made first. Important questions such as how to immobilize cells without disrupting them, and which scanning mode to use need to be answered first.

Chinese hamster ovary (CHO) cells were imaged on the same surface that they were grown on, a petri dish, which provides cells

Figure 8 Pea lectin dimers imaged in tapping mode (f = 9 kHz, amplitude reduction 5-10 %), in liquid with a Si3N4 cantilever (k = 0.03

that are stably attached to a surface. As cells are two orders of magnitude higher than the tapping amplitude, scanning should be performed slowly enough, otherwise the lateral forces will be too high and the cell can be disrupted.

AFM images of living cells as shown in Figure 9 can be obtained, when scanning occurs slowly enough for the feedback to follow the cell contours. In order to keep the force applied to the cell as low as possible during scanning, a relatively soft cantilever was chosen (k= 0.5 N/m). In Figure 9 (a) a height image is shown. In Figure 9 (b) an amplitude image is shown, which is the “error” image as feedback was done on the amplitude. Figure 9 (c) is a phase image, which gives the phase lag between driving signal and response of the cantilever. Phase imaging will be discussed in detail in chapter 7.

From the shape of the cell it can be concluded that the cell is still alive. The highest part of the cell shows the location of the nucleus. The amplitude and phase images (Figure 9 b and c) show more contrast than the height image (Figure 9 a). Notice that the cells touch each other; this is marked by black arrows in Figure 9 b.

Figure 9 CHO cells on plastic, imaged in MAC mode, in isotonic buffer at 25 °C. Shown are height (a), amplitude (b), and phase(c). Scales are 35 µm x 53 µm x (a) 2.25 µm (b) 15 nm (c) 50.4 ˚. Imaging was done in MAC mode (k=0.5 N/m). The image was recorded by Maarten van Es.

Also, parts of the cytoskeleton of the cell can be seen at the white arrows in Figure 9 b.

3.4 Conclusions

In this Chapter, I discussed the advantages and the difficulties of AFM in liquid. Two frequently used modes of AFM were discussed: contact mode and tapping mode. We imaged biological samples ranging from single proteins, to membrane fragments and whole cells.

We imaged single proteins in tapping mode. We chose tapping mode because of the low forces that are applied laterally, in order to prevent removing the proteins from the surface.

Membrane fragments were imaged in contact mode. We kept the applied force as low as possible to prevent damage to the sample. This was achieved by choosing cantilevers with a low spring constant and adjusting the setpoint manually to be “on the verge of losing contact”, and to compensate for thermal drift. In Figure 3 it can be seen that the cantilever lost contact with the surface during some scan lines.

The resolution obtained on the cells (hundreds of nanometers) was much lower than that on the membrane fragments or single molecules (a few nanometers). This is because the cells are large and soft objects, while the membrane fragments and proteins are much stiffer.

A key requirement for obtaining high resolution images in liquid is appropriate sample preparation. Immobilization of membranes can be done by absorption of the membranes in the presence of a buffer solution containing mono-, bi-, or tri- valent ions, depending on the charges on the membrane. Single proteins though, are more easily picked up by the scanning tip. For this reason, electrostatic immobilization is not sufficient, and covalent immobilization is preferred. Covalent immobilization of proteins on a surface will be discussed extensively in the next chapter.

1 _{Drake, B., Prater, C.B., Weisenhorn, A.L., Gould, S.A.C., Albrecht, T.R., Quate, C.F.,}

Cannel, D.S., Hansma, H.G., and Hansma, P.K., Imaging crystals, polymers, and processes in water with the atomic force microscope, Science 243 (1989), 1586-1588

2 _{Engel, A., and Müller, D.J., Observing single biomolecules at work with the atomic}

force microscope, Nature Structural Biology 7 (2000), 715-718

3 _{Martin, Y., Williams, C.C., Wickramasinghe, H.K., Atomic force microscope-force}

mapping and profiling on a sub 100-Å scale, J.Appl. Phys. 61 (10) (1987), 4723-4729

4 _{Zhong, Q., Inniss, D., Kjoller, K., and Elings, V.B., Fractured polymer/silica fiber}

surface studied by tapping mode atomic force microscopy, Surface Science 290 (1993), L688-L692

5 _{Müller, D.J., Sass, H.-J., Müller, S.A., Büldt, G., and Engel, A., Surface structures of}

and Aartsma, T.J., The ring structure and organization of light harvesting 2 complexes in a reconstituted lipid bilayer, resolved by atomic force microscopy, Biophysical Journal 84 (2003), 2483

7 _{Engel, A., Schoenenberger, C.-A., and Müller, D.J., High-resolution imaging of native}

biological sample surfaces using scanning probe microscopy, Current opinion in structural biology 7 (1997), 279-284

8 _{Seelert, H., Poetsch, A., Dencher, N.A., Engel, A., Stahlberg, H., and Müller, D.J.,}

Proton-powered turbine of a plant motor, Nature 405 (2000), 418-419

9 _{Möller, C., Allen, M., Elings, V., Engel, A., and Müller, D.J., Tapping-mode atomic}

force microscopy produces faithful high-resolution images of protein surfaces, Biophysical Journal 77 (2) (1999), 1150-1158

10 _{Müller, D.J., Fotiadis, D., Scheuring, S., Müller, S.A., and Engel, A., Electrostatically}

balanced subnanometer imaging of biological specimens by atomic force microscope, Biophysical Journal 76 (1999), 1101-1111

11 _{Oesterhelt, F., Oesterhelt, D, Pfeiffer, M., Engel, A., Gaub, H.E., and Müller,}

Unfolding pathways of individual bacteriorhodopsins, Science 288 (2000), 143-146

12 _{Oesterhelt, D., and Stoeckenius, W., Isolation of the cell membrane of}

Halobacterium halobium and its fraction into red and purple membrane, Methods

Enzymol. 31 (1974), 667-678

13 _{Hafner, J.H., Cheung, C.-L., Woolley, A.T., and Lieber, C.M., Structural and}

functional imaging with carbon nanotube AFM probes, Progress in Biophysics & Molecular Biology 77 (2001), 73

14 _{Stamouli, A., Frenken, J.W.M., Oosterkamp, T.H., Cogdell, R.J., and Aartsma, T.J.,}

The electron conduction of photosynthetic protein complexes embedded in a membrane, FEBS letters 560 (1-3) (2004), 109-114

15 _{Anczykowski, B., Krüger, D., Babcock, K.L., Fuchs, H., Basic properties of dynamic}

force spectroscopy with the scanning force microscope in experiment and simulation, Ultramicroscopy 66 (1996), 251-259

16 _{Butt, H.J., Siedle, P., Seifert, K., Fendler, K., Seeger, T., Bamberg, E., Weisenhorn,}

A.L., Goldie, K., and Engel, A., J. Microsc. 169 (1993), 75

17 _{Walters, D.A., Cleveland, J.P., Thomson, N.H., Hansma, P.K., Wendman, M.A.,}

Gurley, G., and Elings, V., Short cantilevers for atomic force microscopy, Rev. Sci. Instrum. 67 (1996), 3583

18 _{Putman, C.A.J., Van der Werf, K.O., De Grooth, B.G., Van hulst, N.F., and Greve, J.,}

Tapping mode atomic force microscopy in liquid, Applied Physics Letters 64 (1994), 2454-2456

19 _{Schäffer, T.E., Cleveland, J.P., Ohnesorge, F., Walters, D.A., and Hansma, P.K.,}

Studies of vibrating atomic force microscopy cantilevers in liquid, J. Appl. Phys. 80 (1996), 3622

20 _{Jarvis, S.P., Oral, A., Weihs, T.P., Pethica, J.B., A novel force microscope and point}

contact probe, Rev. Sci. Instrrum 64 (1993), 3515-3520

21 _{Han, W., S.M. Lindsay and T. Jing, A magnetically driven oscillating probe}