Journ

alof

Cell

Scie

nce

Atomic force microscopy – looking atmechanosensors on the cell surface

Jurgen J. Heinisch1,*, Peter N. Lipke2,*, Audrey Beaussart3, Sofiane El Kirat Chatel3, Vincent Dupres3,David Alsteens3 and Yves F. Dufrene3,*1Universitat Osnabruck, Fachbereich Biologie/Chemie, AG Genetik, Barbarastr. 11, 49076 Osnabruck, Germany2Department of Biology, Brooklyn College of the City University of New York, New York, NY 11210, USA3Universite Catholique de Louvain, Institute of Condensed Matter and Nanosciences, Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium

*Authors for correspondence (Juergen.Heinisch@Biologie.Uni-Osnabrueck.DE; PLipke@brooklyn.cuny.edu; Yves.Dufrene@uclouvain.be)

Journal of Cell Science 125, 1–7� 2012. Published by The Company of Biologists Ltddoi: 10.1242/jcs.106005

SummaryLiving cells use cell surface proteins, such as mechanosensors, to constantly sense and respond to their environment. However, the way inwhich these proteins respond to mechanical stimuli and assemble into large complexes remains poorly understood at the molecular level. In

the past years, atomic force microscopy (AFM) has revolutionized the way in which biologists analyze cell surface proteins to molecularresolution. In this Commentary, we discuss how the powerful set of advanced AFM techniques (e.g. live-cell imaging and single-moleculemanipulation) can be integrated with the modern tools of molecular genetics (i.e. protein design) to study the localization and molecular

elasticity of individual mechanosensors on the surface of living cells. Although we emphasize recent studies on cell surface proteins fromyeasts, the techniques described are applicable to surface proteins from virtually all organisms, from bacteria to human cells.

Key words: Atomic force microscopy, AFM, Cell adhesion, Cell surface, Cell imaging, Mechanosensing, Microscopy, Single molecule

IntroductionProteins located on the cell surface have pivotal roles in adhesion,

sensing of the environment, signaling, communication, transport,energy transformation, embryonic and tissue development, tumourmetastasis and microbial infection (Sheetz, 2001; Discher et al.,

2005; Vogel and Sheetz, 2006; Lecuit and Lenne, 2007). Thesecomplex functions rely on the assembly and interactions of specificproteins, including mechanosensors and cell adhesion molecules.

Whereas much is known about the molecular biology of suchsurface proteins, their biophysical properties, such as their

mechanics and clustering in the cell membrane, remain poorlyunderstood at the molecular level.

Cells can react to external cues by engaging signal transduction

pathways, which detect environmental changes at the cell surfaceand trigger the appropriate intracellular responses. Frequently,these signaling pathways are initiated by sensor proteins that span

the plasma membrane and ultimately trigger a change in geneexpression and, consequently, in the cell’s proteome (McIntoshet al., 2009; Rodicio and Heinisch, 2010). Such sensor proteins

either bind specific ligands (e.g. hormones, adhesion molecules,etc.) or detect mechanical forces and related physical stimuli.

Hence, mechanosensing and mechanotransduction, which convertthese mechanical forces into biochemical signals, have key roles inregulating processes such as cell growth, differentiation, cell shape

and cell death (Schwartz, 2009; Vogel and Sheetz, 2006; Brownand Discher, 2009).

Cell adhesion molecules also have key roles in mediating

cellular processes such as cell–cell communication (Dalva et al.,2007), tissue development (Morgan et al., 2007), inflammation(Weber et al., 2007), cancer (Gray-Owen and Blumberg, 2006)

and microbial infection (Verstrepen et al., 2004; Kolter andGreenberg, 2006; Telford et al., 2006; Sokurenko et al., 2008).

There is growing evidence that the force-induced deformation of

such proteins has an important role in modulating their cellularfunction. Prominent examples of such processes are ‘catchbonds’, i.e. bonds between receptors and ligands that are

strengthened by tensile mechanical force (Sokurenko et al.,2008), as well as the force-induced exposure of cryptic peptidesequences, as observed in, for example, fibronectin and integrin(Vogel and Sheetz, 2006; Brown and Discher, 2009).

Atomic force microscopy (AFM) makes it possible to observe,manipulate and explore the cell surface at a molecular resolution,and therefore has produced a wealth of new opportunities in cell

biology, including understanding the nanoscale organization anddynamics of cell membranes and cell walls, measuring cellmechanics and cell adhesion, unraveling the molecular elasticityof cellular proteins and the mechanisms by which they assemble

into nanodomains in the membrane (Muller and Dufrene, 2011;Muller et al., 2009). In this Commentary, we explain the basicprinciples of AFM, discuss current strategies for imaging live

cells and for probing single functional proteins on their surfaces,provide a critical evaluation of the potential and limitations of thetechnique, and survey recent discoveries in cell surface biology

that were driven by AFM. Although animal cell studies will becovered, we will especially focus on breakthroughs that havebeen made in yeast cell research, because this is wheresubstantial progress in single-cell surface protein analysis has

recently been made.

Atomic force microscopyGeneral principles

Instead of using an incident beam to visualize a sample – aswould be the case in classical microscopy – AFM senses thesmall forces (in the piconewton range, ,10212 N) that act on the

ARTICLE SERIES: Imaging Commentary 1

JCS online publication date 17 October 2012

Journ

alof

Cell

Scie

nce

surface of a sample (Binnig et al., 1986; Muller and Dufrene,

2011; Muller et al., 2009). Images of molecules or cells in buffer

or growth medium are made by scanning a sharp tip over the

sample surface and measuring the interaction force between the

tip and the surface (Fig. 1A). The sample is mounted on a

piezoelectric scanner, which ensures three-dimensional

positioning with high accuracy. While the tip is being scanned

across the sample in the x and y directions, the force between the

tip and specimen can be recorded. The sharp tip is attached to a

soft cantilever, and, as the cantilever bends, its deflection is

detected by movement of a laser beam reflected from the tip.

AFM topographic imaging is widely used in the life sciences, and

has provided high-resolution images of biomolecules, membranes

and cells in buffer at unprecedented resolution (Dufrene, 2008a,

Dufrene, 2008b; Engel and Gaub, 2008; Muller and Dufrene,

2011; Muller et al., 2009).

AFM is also widely used to manipulate and analyze single

biomolecules with a method called single-molecule forcespectroscopy (SMFS) (Hinterdorfer and Dufrene, 2006; Engeland Gaub, 2008; Puchner and Gaub, 2009; Muller et al., 2009;

Dufrene et al., 2011). Here, the tip is brought into proximity ofand retracted from the biological sample, and the cantileverdeflection measures the interaction force (Fig. 1B). The force–distance curves that are obtained with this procedure provide key

insights into the localization, elasticity and binding strength ofsingle molecules. As we will discuss below, the manipulation ofsingle molecules on the surface of living cells often requires

labeling the tip with chemical groups or bioligands using specificprotocols (Fig. 1B).

Imaging living cells with AFM

Soon after its invention, AFM became a valuable tool for imagingcells (Butt et al., 1990; Radmacher et al., 1992). However, AFMimaging of single cells requires their firm attachment to a surface,

which is not always a simple task. A straightforward approach isto exploit the ability of animal cells to spread and adhere to solidsupports (Radmacher et al., 1992). Coating the substrate withadhesion proteins might be used to enhance immobilization, and

this method has made it possible to observe, for example, actinfilament dynamics beneath the plasma membrane of glial cells(Henderson et al., 1992). In some cases, chemical fixation using

cross-linking agents such as glutaraldehyde might be requiredeither to prevent cell damage or detachment from the supportcaused by the scanning tip, or as a means to obtain high-

resolution images (Fig. 2A). Using these different protocols,various cell types have been investigated, includingmacrophages, CV-1 kidney cells, fibroblasts, Madin–Darby

canine kidney (MDCK) cells, platelets and cardiomyocytes(Fig. 2A; Dufrene, 2011; Jena and Horber, 2002).

In recent years, much progress has been made with regards tolive-cell imaging of various microbes (Fig. 2B; Dufrene, 2008b;

Dufrene, 2011). In order to gain reliable high-resolution imagesof microbial cells, sample preparation is of crucial importance.Unlike animal cells, microbes have a well-defined shape and

usually do not spread on surfaces under experimental conditions.As a result, the contact area between a cell and the support is verysmall, which often leads to cell detachment caused by thescanning tip. Therefore, several approaches have been developed

to ensure more stable cell attachment (Dufrene, 2008a). Forinstance, it is possible to attach the cells onto supports thathave been functionalized with either positively charged

macromolecules, like poly-L-lysine or polyethylenimine, orwith molecules containing hydrophobic groups. This methodhas been successfully applied to lactic acid and Gram-negative

bacteria, diatoms and fungi, and has yielded new insights intosurface structure and elasticity of these organisms. Cells can alsobe immobilized mechanically within gelatin-coated supports or

captured in the holes of porous polymer membranes. In the latterapproach, a concentrated cell suspension is driven through aporous membrane with a pore size similar to that of the cells to beinvestigated (Kasas and Ikai, 1995) (Fig. 2B). This procedure is

fairly simple and does not involve a macromolecular support,thus preventing the risk of contamination of the cell surface or theAFM tip.

Although AFM imaging offers key benefits over moreconventional microscopy techniques, newcomers should beaware that the current technology is still limited by a number

A

B

Ni2+-NTA

Fig. 1. Atomic force microscopy. (A) In the imaging mode, a very sharp tip

follows the contours of the cell surface with nanometer resolution. The lipid

bilayer of the plasma membrane is shown, with inserted proteins as yellow

objects. (B) In SMFS, the small interaction force between the AFM tip and

cell surface molecules is measured, while the distance between the tip and cell

is altered. The two examples show a tip labeled with a chemical group

(Ni2+-NTA, left) or a ligand (e.g. an antibody, right) to detect, determine the

location of and manipulate individual cell surface proteins, such as

mechanosensors (shown in magenta and green, respectively).

Journal of Cell Science 125 (0)2

Journ

alof

Cell

Scie

nce

of problems that hamper its widespread use in cell biology (Box

1). Nonetheless, AFM can be employed to address a number of

interesting cell biological questions, as we will discuss below.

Single-protein localization and manipulation

Unlike fluorescence microscopy, AFM topographic imaging lacks

biochemical specificity (Table 1). Thus, specific molecules, such

as a given receptor, cannot be unambiguously identified and

located on cell surfaces. This is an important limitation, because

the organization and interactions of the cell surface machinery is

an important parameter for controlling its functions. However, the

use of spatially resolved SMFS with AFM tips that are labeled with

specific chemical groups or ligands now enables the detection and

localization of single functional proteins on cell surfaces

(Hinterdorfer and Dufrene, 2006). The rationale behind this

approach is to use SMFS to record arrays of force curves across the

cell surface with a specifically labeled tip (Fig. 2C), estimate the

specific molecular recognition force value for each force curve and

then display it in terms of pixels, where the pixel brightness

reflects the magnitude of the recorded force (Fig. 2D). The cell

surface recognition map that can be obtained using this approach

enables researchers to understand the distribution of a given

receptor and to determine whether it is clustered or randomly

distributed across the cell surface (Fig. 2D). However, it should be

A B

C E D

Distance

Forc

e

100 nm 200

pN

1 µm 10 µm

Fig. 2. Atomic force microscopy in cell biology. (A,B) Imaging cells. AFM

images of gently fixed macrophages spread on glass (A) and of a single yeast

cell of Candida albicans trapped in a porous polymer membrane (B). Arrows

highlight a common artifact, namely the alteration of the cell surface by the

scanning tip. (C-E) Imaging and manipulating cell surface proteins: single-

molecule force spectroscopy with AFM tips attached to specific antibodies

makes it possible to detect single proteins on the cell surface (C), to map their

distribution (D) and to measure their mechanical response in relation to

function (E). The recognition map in (D) documents the detection of single

cell-adhesion protein (white pixels), and the notion that they are concentrated

into localized nanodomains. The force–distance curves in (E) document

multiple force peaks that reflect the force-induced unfolding of individual

proteins. Shown here are the results for the Candida albicans Als adhesion

proteins, whose unfolding and clustering are believed to have a role in

cell adhesion.

Box 1. Advantages and limitations of AFM incell biology

AFM shows distinct advantages over optical and electron

microscopy techniques for cell surface biology (Table 1), namely

(1) the ability to observe purified membranes and live cells at

nanometer resolution and under physiological conditions (e.g. in

buffer solution at room temperature and atmospheric pressure),

(2) the capability to track structural dynamics and remodeling of

the cell surface in response to environmental stimulants (e.g.

drugs), and (3) the possibility to determine the location of and

force-probe single constituents of the cell surface. Despite these

key benefits, there are still a number of limitations and

technological issues that must be solved before the full potential

of the technique to address cell biological questions can be

exploited (Table 1).

When observing cells using AFM topographic imaging, an

important source of problems is the resolution and interpretation of

the images, which directly depend on the imaging force, tip

geometry and physical properties of the sample (Dufrene, 2008a).

Large forces acting between tip and sample during imaging can

dramatically reduce the image resolution and cause molecular

damage or displacement (Fig. 2A,B). This is particularly true for

soft corrugated mammalian cells, where the tip generally deforms

the cell surface, pushes cellular material along the scanning

direction and easily becomes blind owing to contamination by

loosely bound macromolecules. Therefore, it is essential to control

the force that is applied during scanning (0.1-0.5 nN) by selecting

appropriate buffer conditions. In addition, unlike microbial cells,

which can be readily imaged in their native state (Fig. 2B), animal

cells generally require chemical fixation in order to be imaged by

AFM (Fig. 2A). The image resolution achieved with AFM on

mammalian cells is generally not better than 50-100 nm. One

approach to solve this problem is to use a nanopipette that does

not contact the surface directly, which gives resolutions from live

cells of ,20 nm (Novak et al., 2009).

Another limitation of AFM imaging compared with fluorescence

microscopy is its rather poor temporal resolution (typically

,1 minute per image), which is much slower than the time scale

at which dynamic processes usually occur in cell biology.

However, remarkable advances are being made in developing

high-speed AFM set-ups that can operate in the millisecond

timescale and thus offer new possibilities to explore cellular

dynamics (Shibata, 2010).

A crucial feature of SMFS is the functionalization of the AFM tips

with biomolecules. This can be achieved using essentially two

types of surface chemistries, which are based either on the strong

chemisorption of thiols onto gold surfaces or on the covalent

attachment of silanes or alcohols onto silicon oxide surfaces

(reviewed by Hinterdorfer and Dufrene, 2006). In addition, well-

defined procedures are available to accurately measure the

localization and forces of single proteins on live cells

(Hinterdorfer and Dufrene, 2006), but accurate data collection

and interpretation often remain technically challenging. In this

context, the main issues are those associated with the quality of

the tip and its possible alteration during the course of an

experiment. As with statistical validation, spatially resolved

SMFS should be performed on different locations of the cell

surface using different tips and the measurements should be

repeated using many cells from independent cultures. Finally, an

important drawback of AFM-based SMFS is that it probes cell

surface biomolecules, but is unable to access the interior of living

cells.

AFM of cell surface mechanosensors 3

Journ

alof

Cell

Scie

nce

noted that, as is the case for cell imaging, single-protein analysis by

SMFS is still a rather new method that suffers from a number of

drawbacks (Box 1).

SMFS cannot only be used for mapping single protein

molecules on a given cell surface, but can also be employed to

subject cell surface proteins to force in order to study their

mechanical properties (Fig. 2E). Early studies have demonstrated

that proteins can be subjected to controlled forces using this

approach, and have yielded details of protein unfolding pathways

(Rief et al., 1997). Since then, single-molecule AFM, combined

with molecular dynamics simulations and protein engineering,

has greatly enhanced our understanding of the mechanical

behavior of a great variety of proteins (Engel and Gaub, 2008;

Puchner and Gaub, 2009; Li and Cao, 2010; Marszalek and

Dufrene, 2012). Most of these single protein manipulation

experiments have been conducted in vitro, where the system

under study can be tightly controlled. However, this also means

that molecules are removed from the native cellular context that

controls their structural assembly and functional state (Dufrene

et al., 2011). In the following section, we will highlight some

recent breakthrough studies in which SMFS has been

successfully used to unravel the nanomechanics and clustering

of single proteins on the surface of living cells.

Measuring mechanics of single proteinsNanomechanics of membrane sensors

The biophysical mechanisms underlying cellular mechanosensing

are poorly understood at the molecular level. In cell wall integrity

(CWI) sensing and signaling in yeast, five plasma-membrane-

spanning proteins (Wsc1, Wsc2, Wsc3, Mid2 and Mtl1)

presumably detect mechanical forces acting on the cell wall

and membrane, and subsequently trigger an intracellular

signaling cascade, which ultimately results in the expression of

genes encoding either cell wall biosynthetic enzymes or cell wall

proteins (Levin, 2005; Rodicio and Heinisch, 2010). However,

the lack of an appropriate probing technique meant that, until

recently, it remained unclear how these membrane sensors

responded to mechanical force.

SMFS studies with genetically manipulated Wsc1 in live

Saccharomyces cerevisiae have recently unraveled the

mechanical behavior of single sensor molecules in relation to

their function (Dupres et al., 2009; Heinisch et al., 2010a). Central

to this discovery was the use of genetic manipulations to address a

conceptual problem: AFM is a surface-based technique, so how

can it probe protein sensors (such as Wsc1) that are embedded

within the cell wall? The yeast cell wall has a thickness of

,100 nm (Backhaus et al., 2010), whereas the maximal length of

the extracellular part of the Wsc1 sensor is 86 nm, meaning

it does not reach the outermost cell surface. Therefore, the

mechanosensors were artificially elongated to a length of 153 nm

by inserting a stiff serine/threonine-rich region from the Mid2

sensor. Furthermore, the addition of a histidine tag to the N-

terminal end of the protein allowed the proteins to be detected by a

AFM tip modified with Ni2+-nitrilotriacetate (Ni2+-NTA). Force–

extension curves obtained by pulling on the modified Wsc1

sensors revealed a fascinating behavior: the sensors display the

characteristics of a Hookean spring (i.e. their force curve contains

a linear region where force is directly proportional to extension)

Table 1. Comparison of high-resolution techniques for imaging single proteins in cells

Technique

AFMSuper-resolution fluorescence

microscopy (TIRF, PALM, STORM) Electron microscopy (TEM)Topographic imaging SMFS

Resolution <50 nma <20 nm <20–50 nm <1-10 nmLive cell Yes Yes Yes NoTime for sample

preparation10 minutes 1 dayb 1-3 hours 2-5 days

Time for imageacquisition

<5 minutes 25-50 minutes <5 minutes 5-10 minutes

Image processing 5 minutes A few hours Up to 24 hours 30 minutesSimultaneous detection

of different proteinsNo No Yes Yesc

Protein tracking No No Yes NoCost of equipment J150,000-250,000 J250,000-500,000 J<500,000Operational costs Low Moderate Moderate HighAdvantages Localization and force response of single proteins; protein

unfolding; dynamic processes; various environments(temperature, buffer, etc.)

Spatial-temporal resolution; onlybasic genetic manipulationsrequired; high sample throughput;study of live protein interactionsand their quantification

Imaging of the cellultrastructures at very highresolution

Disadvantages Only the cell surface is analyzed; only a single cell at atime; slow temporal resolution; various sources ofartifacts like cell or tip alteration

Frequently a bad signal-to-noiseratio; photobleaching; noinformation on physical propertiesof proteins; tendency offluorescent protein tags tomultimerize (i.e. interactionartifacts)

Fixation artifacts; no accessto dynamics; noinformation on physicalproperties of proteins

Only whole cell analyses are considered, i.e. purified proteins or membranes are not included. TIRF, total internal reflection fluorescence; PALM,photoactivated localization microscopy; STORM, stochastic optical reconstruction microscopy; TEM, transmission electron microscopy.

aIn mammalian cells (a 10-nm resolution can be routinely achieved on microbial cells).bIncluding tip preparation.cIndirect; e.g. gold beads of different sizes.

Journal of Cell Science 125 (0)4

Journ

alof

Cell

Scie

nce

(Dupres et al., 2009). These nanospring properties, which contrast

with the properties of most proteins (Rief et al., 1997; Marszalekand Dufrene, 2012), have subsequently been shown to be mediatedby the extracellular serine/threonine-rich region of the sensor

(Dupres et al., 2009). Notably, this technology has also provided atool for determining cell wall thickness in vivo in yeast, becausesensor molecules of different lengths can only be detected oncethey reach the surface (Dupres et al., 2010).

Accordingly, SMFS opens new avenues for investigating howproteins respond to forces in living cells and how mechanosensingevents proceed in vivo, thereby providing new insights into what has

been suggested to be one of the most ancient sensory mechanisms inevolution (Kee and Robinson, 2008). An exciting new developmentin this direction will be the combination of SMFS with fluorescencemicroscopy. Pulling on a sensor from the outside with SMFS should

lead to the recruitment of the corresponding intracellular signalingcomponents to the cytoplasmic tail of the sensor, which cansubsequently be observed by fluorescence microscopy. We also

anticipate that the single protein manipulations described hereshould allow researchers to address the mechanics of a variety ofother membrane sensors that would usually be hidden and not

accessible by AFM per se. For instance, mammalian cell sensors,such as integrins, are usually covered by a polysaccharide andprotein matrix, the glycocalyx, which prevents reliable SMFS

experiments. Elongation and tagging of such proteins shouldprovide a means to solve this issue.

Nanomechanics of cell adhesion proteins

During the past 15 years, SMFS has been widely used to measurethe molecular interactions of cell adhesion molecules, includingselectins, cadherins, integrins, oligosaccharides, proteoglycansand microbial adhesins (see Hinterdorfer and Dufrene, 2006;

Muller and Dufrene, 2011, and references therein). Because theseexperiments were conducted on purified molecules, it is of greatinterest to probe single cell-adhesion proteins directly in living

cells to further understand the interactions and functions ofadhesion molecules in vivo.

Adhesion of the pathogen Candida albicans is mediated by cell-surface glycoproteins referred to as the Als family. Microscopic

assays have revealed that Als-mediated adhesion involves aninitial adhesion step, followed by force-induced proteinconformational changes that result in the formation of strong

amyloid-like bonds that can ‘cement’ the adhering cells together(Dranginis et al., 2007; Lipke et al., 2012). Again, the moleculardetails of this mechanism remained essentially unknown. SMFS

was therefore used to study the adhesive and mechanical propertiesof Als5 in live cells (Alsteens et al., 2009). Pulling on single Als5proteins using a ligand-modified AFM tip has revealedcharacteristic force signatures with multiple well-defined peaks,

with each peak corresponding to the force-induced unfolding of anindividual tandem-repeat domain. AFM measurements alsorevealed that the unfolding probability increases with the number

of repeats and correlates with the level of cell–cell adhesion, whichconfirms that the modular domains have a role in many fungaladhesion molecules. Presumably, the force-induced unfolding of

Als proteins leads to extended conformations in which previouslymasked hydrophobic groups become exposed, thus favoringhydrophobic interactions between opposing cells. These single-

molecule measurements open up new avenues for understandingthe mechanical properties of adhesion molecules in a variety of celltypes. In particular, they should contribute to increasing our

understanding of how force-induced conformational changes in

proteins such as integrins modulate their binding strength (Vogel

and Sheetz, 2006; Brown and Discher, 2009).

Imaging protein clusteringUnderstanding how membrane proteins assemble to form

membrane micro- and nano-domains is another important

question in cell biology. A key example of such clustering is the

formation of adhesion domains (i.e. focal adhesion complexes)

composed of aggregated proteins that mediate the early stage of

adhesion (Vogel and Sheetz, 2006). The small size (,10-500 nm)

and dynamic nature of membrane domains make their direct

visualization in live cells very challenging. However, AFM is

emerging as a valuable tool to map the distribution of single-

proteins on cell surfaces, thereby unraveling the formation and

composition of such protein domains. Prominent examples include

the mapping of the distribution of prostaglandin receptors on live

CHO cells (Kim et al., 2006), the localization of vascular

endothelial cadherin-binding sites (Chtcheglova et al., 2007), and

the measurement of the organization and stiffness of neuron

membrane domains containing glycosylphosphatidylinositol

(GPI)-anchored proteins (Roduit et al., 2008). As discussed

below, functional protein clusters were recently found to form

and increase in size under stress, thereby activating cell signaling

and cell adhesion in yeasts.

Stress-induced clustering of membrane sensors

Fluorescence microscopy has shown that GFP–Wsc1 sensors

form membrane patches, but the structure–function relationships

of such clusters has long remained a mystery (Wilk et al., 2010).

To address this problem, spatially resolved SMFS was used to

map the distribution of single Wsc1 sensors on the surface of live

S. cerevisiae (Heinisch et al., 2010b). Molecular recognition

images were recorded on cells expressing His-tagged elongated

sensors using Ni2+-NTA tips. Many sensor molecules appeared

to form clusters of ,200 nm, which is in the range of the

large patches observed by fluorescence microscopy for

the microdomains containing the marker protein Can1 (MCC)

in S. cerevisiae (Malinsky et al., 2010). Both the surface density

of Wsc1 sensors and their tendency to cluster increases when

cells are stressed by heat or low osmolarity, which suggests that

clustering is a stress response that is intimately connected to

signaling.

An analysis of three different Wsc1 mutants has also indicated

that the Wsc1 cysteine-rich domain has a crucial function in

sensor clustering and signaling. Although mutant proteins display

surface density and spring behaviors that are similar to those of

wild-type Wsc1, they are more evenly distributed over the yeast

cell surface, instead of being clustered. In addition, Wsc1 sensor

clustering mediated by the cysteine-rich domain is dependent on

the formation of disulfide bridges (Dupres et al., 2011). As these

mutant proteins produce non-functional sensors that are unable to

activate the downstream signaling pathway, these findings

demonstrate the importance of the cysteine-rich domain in

sensor clustering and turnover (Heinisch et al., 2010b). The

studies also suggest that in yeast, like in higher eukaryotes, the

function of sensors is coupled to their localized enrichment in

membrane patches that are referred to as ‘sensosomes’. We are

confident that single-protein imaging by SMFS will be useful to

study similar clustering of sensor proteins in mammalian cells,

AFM of cell surface mechanosensors 5

Journ

alof

Cell

Scie

nce

especially following the further development of high-speedSMFS mapping.

Force-induced clustering of cell adhesion proteins

A fascinating phenomenon that is portrayed by cell adhesion

domains in mammalian cells is their ability to grow andstrengthen under force (Bershadsky et al., 2006). In Candida

albicans, macroscopic assays have suggested that adhesion-

triggered changes in the conformation of Als5 propagate aroundthe cell surface, thereby forming ordered adhesion domains(Lipke et al., 2012). This mechanism was recently experimentallyconfirmed by SMFS: mechanical force was shown to trigger the

formation of cell adhesion nanodomains in living yeast cells(Alsteens et al., 2010). Pulling on single adhesins was shown toinduce the formation of adhesion domains of 100-500 nm in size,

and these force-induced nano-domains propagate over the entirecell surface. Als5-mediated remodeling is independent of cellularmetabolic activity, because the protein shows the same behavior

in heat-killed cells, which are no longer metabolically active.Remarkably, a single-site mutation in the conserved amyloid-forming sequence of the protein has revealed that amyloid

interactions represent the driving force that underlies Als5clustering (Garcia et al., 2011). Hence, the strength of yeastcell–cell adhesion results from the force-induced amyloid-likeclustering of hundreds of proteins on the cell surface to form

arrays of ordered multimeric binding sites (Lipke et al., 2012).These results highlight the role that functional amyloids can havein cell adhesion, both in lateral clustering (in cis) of many

adhesion proteins to increase binding avidity, and in the potentialto form stable amyloid-type bonds between cells (in trans), amechanism that we expect to be detected in many more

organisms as more cell adhesion systems are examined(Romero et al., 2010).

Interestingly, the above mechanical restructuring is reminiscentof shear-induced unfolding and refolding of proteins during bloodclotting, where force induces the exposure of interaction sites in

the blood glycoprotein von Willebrand factor (VWF), which hasrecently been demonstrated by single-molecule laser tweezerexperiments (Kim et al., 2010). Shear-induced bond strengthening,

called ‘catch bonding’, is also crucial in arresting leukocyte rollingand extravasation (Alon and Dustin, 2007). Clearly, as the SMFStechnique further evolves and matures, it will contribute to the

discovery of similar remarkable phenomena in many other cellsurface proteins.

ConclusionsUnderstanding protein mechanics and protein clustering in vivo is acrucial challenge in cell biology, which, until now, has been

difficult to tackle owing to the lack of high-resolution probingtechniques. This Commentary has highlighted how the combinationof single-molecule AFM with genetic manipulations is a powerful

tool for probing functional receptors and sensors on cell surfaces, ina way that was not possible previously, and how this approachnicely complements other imaging techniques in cell biology

(Table 1). This integrated platform can be applied to measuring themechanical properties of single proteins, and can be used to imagetheir localization, thus revealing whether they are isolated or form

clusters.

Despite the key benefits of AFM, researchers should realizethat the technology is still under development and that there arestill a number of limitations that must be solved in order for the

technique to become more widespread in the cell biology

community (Box 1). Clearly, a detailed understanding of the

principles and limitations of the different AFM modalities is

essential before users start their first experiments. In SMFS,

protocols have been developed for attaching biomolecules to

AFM tips but they still require specific expertise that is usually

not found in cell biology laboratories. In the future, defining

simple standardized protocols for tip functionalization and data

interpretation, and automation of force spectroscopy analyses,

making them readily available to the biology community, will

contribute to make SMFS accessible for cell biologists.

Undoubtedly, the main limitation of both AFM imaging and

spatially resolved SMFS in cell biology is their rather slow

temporal resolution (Box 1). Hopefully, current efforts in

developing high-speed AFM techniques should soon provide

access to millisecond resolutions using live cells. Another

important direction for future research is to expand the range

of potential applications for single molecule analysis, by

combining AFM with advanced light microscopy techniques,

such as stimulated emission depletion microscopy (Hell, 2009).

Such combined imaging platforms would allow researchers to

simultaneously identify, track, observe and force probe the

individual constituents of the cell surface, thereby helping to

answer many unresolved problems in cell biology.

FundingWork in the Y.F.D. group was supported by the National Foundationfor Scientific Research (FNRS); the Foundation for Training inIndustrial and Agricultural Research (FRIA); the Universitecatholique de Louvain (Fonds Speciaux de Recherche); the FederalOffice for Scientific, Technical and Cultural Affairs (InteruniversityPoles of Attraction Programme); and the Research Department of theCommunaute francaise de Belgique (Concerted Research Action). Y.F. D. and D.A. are Senior Research Associate and PostdoctoralResearcher of the FRS-FNRS, respectively. Work in the P. N. L.laboratory was supported by National Institutes of Health [grantnumber SC1 GM083756]. Work on CWI sensors in the lab of J. J. H.is supported by a grant from the Deutsche Forschungsgemeinschaftin the frame of the SFB944. Deposited in PMC for release after 12months.

ReferencesAlon, R. and Dustin, M. L. (2007). Force as a facilitator of integrin conformational

changes during leukocyte arrest on blood vessels and antigen-presenting cells.

Immunity 26, 17-27.

Alsteens, D., Dupres, V., Klotz, S. A., Gaur, N. K., Lipke, P. N. and Dufrene, Y. F.

(2009). Unfolding individual Als5p adhesion proteins on live cells. ACS Nano 3,

1677-1682.

Alsteens, D., Garcia, M. C., Lipke, P. N. and Dufrene, Y. F. (2010). Force-induced

formation and propagation of adhesion nanodomains in living fungal cells. Proc. Natl.

Acad. Sci. USA 107, 20744-20749.

Backhaus, K., Heilmann, C. J., Sorgo, A. G., Purschke, G., de Koster, C. G., Klis,

F. M. and Heinisch, J. J. (2010). A systematic study of the cell wall composition of

Kluyveromyces lactis. Yeast 27, 647-660.

Bershadsky, A., Kozlov, M. and Geiger, B. (2006). Adhesion-mediated mechan-

osensitivity: a time to experiment, and a time to theorize. Curr. Opin. Cell Biol. 18,

472-481.

Binnig, G., Quate, C. F. and Gerber, C. (1986). Atomic force microscope. Phys. Rev.

Lett. 56, 930-933.

Brown, A. E. X. and Discher, D. E. (2009). Conformational changes and signaling in

cell and matrix physics. Curr. Biol. 19, R781-R789.

Butt, H. J., Wolff, E. K., Gould, S. A. C., Dixon Northern, B., Peterson, C. M. and

Hansma, P. K. (1990). Imaging cells with the atomic force microscope. J. Struct.

Biol. 105, 54-61.

Chtcheglova, L. A., Waschke, J., Wildling, L., Drenckhahn, D. and Hinterdorfer, P.

(2007). Nano-scale dynamic recognition imaging on vascular endothelial cells.

Biophys. J. 93, L11-L13.

Dalva, M. B., McClelland, A. C. and Kayser, M. S. (2007). Cell adhesion molecules:

signaling functions at the synapse. Nat. Rev. Neurosci. 8, 206-220.

Journal of Cell Science 125 (0)6

Journ

alof

Cell

Scie

nce

Discher, D. E., Janmey, P. and Wang, Y. L. (2005). Tissue cells feel and respond to thestiffness of their substrate. Science 310, 1139-1143.

Dranginis, A. M., Rauceo, J. M., Coronado, J. E. and Lipke, P. N. (2007). Abiochemical guide to yeast adhesins: glycoproteins for social and antisocial occasions.Microbiol. Mol. Biol. Rev. 71, 282-294.

Dufrene, Y. F. (2008a). Atomic force microscopy and chemical force microscopy ofmicrobial cells. Nat. Protoc. 3, 1132-1138.

Dufrene, Y. F. (2008b). Towards nanomicrobiology using atomic force microscopy.Nat. Rev. Microbiol. 6, 674-680.

Dufrene, Y. F. (2011). Life at the Nanoscale: Atomic Force Microscopy of Live Cells.Singapore: Pan Stanford Publishing.

Dufrene, Y. F., Evans, E., Engel, A., Helenius, J., Gaub, H. E. and Muller, D. J.

(2011). Five challenges to bringing single-molecule force spectroscopy into livingcells. Nat. Methods 8, 123-127.

Dupres, V., Alsteens, D., Wilk, S., Hansen, B., Heinisch, J. J. and Dufrene, Y. F.(2009). The yeast Wsc1 cell surface sensor behaves like a nanospring in vivo. Nat.

Chem. Biol. 5, 857-862.Dupres, V., Dufrene, Y. F. and Heinisch, J. J. (2010). Measuring cell wall thickness in

living yeast cells using single molecular rulers. ACS Nano 4, 5498-5504.Dupres, V., Heinisch, J. J. and Dufrene, Y. F. (2011). Atomic force microscopy

demonstrates that disulfide bridges are required for clustering of the yeast cell wallintegrity sensor Wsc1. Langmuir 27, 15129-15134.

Engel, A. and Gaub, H. E. (2008). Structure and mechanics of membrane proteins.Annu. Rev. Biochem. 77, 127-148.

Garcia, M. C., Lee, J. T., Ramsook, C. B., Alsteens, D., Dufrene, Y. F. and Lipke,P. N. (2011). A role for amyloid in cell aggregation and biofilm formation. PLoS ONE

6, e17632.Gray-Owen, S. D. and Blumberg, R. S. (2006). CEACAM1: contact-dependent control

of immunity. Nat. Rev. Immunol. 6, 433-446.Heinisch, J. J., Dupres, V., Alsteens, D. and Dufrene, Y. F. (2010a). Measurement of

the mechanical behavior of yeast membrane sensors using single-molecule atomicforce microscopy. Nat. Protoc. 5, 670-677.

Heinisch, J. J., Dupres, V., Wilk, S., Jendretzki, A. and Dufrene, Y. F. (2010b).Single-molecule atomic force microscopy reveals clustering of the yeast plasma-membrane sensor Wsc1. PLoS ONE 5, e11104.

Hell, S. W. (2009). Microscopy and its focal switch. Nat. Methods 6, 24-32.Henderson, E., Haydon, P. G. and Sakaguchi, D. S. (1992). Actin filament dynamics

in living glial cells imaged by atomic force microscopy. Science 257, 1944-1946.Hinterdorfer, P. and Dufrene, Y. F. (2006). Detection and localization of single

molecular recognition events using atomic force microscopy. Nat. Methods 3, 347-355.

Jena, B. P. and Horber, J. K. H. (2002). Atomic Force Microscopy in Cell Biology,

Methods in Cell Biology. San Diego, CA: Academic Press.Kasas, S. and Ikai, A. (1995). A method for anchoring round shaped cells for atomic

force microscope imaging. Biophys. J. 68, 1678-1680.Kee, Y. S. and Robinson, D. N. (2008). Motor proteins: myosin mechanosensors. Curr.

Biol. 18, R860-R862.Kim, H., Arakawa, H., Hatae, N., Sugimoto, Y., Matsumoto, O., Osada, T.,

Ichikawa, A. and Ikai, A. (2006). Quantification of the number of EP3 receptors on aliving CHO cell surface by the AFM. Ultramicroscopy 106, 652-662.

Kim, J., Zhang, C. Z., Zhang, X. and Springer, T. A. (2010). A mechanicallystabilized receptor-ligand flex-bond important in the vasculature. Nature 466, 992-995.

Kolter, R. and Greenberg, E. P. (2006). Microbial sciences: the superficial life ofmicrobes. Nature 441, 300-302.

Lecuit, T. and Lenne, P. F. (2007). Cell surface mechanics and the control of cellshape, tissue patterns and morphogenesis. Nat. Rev. Mol. Cell Biol. 8, 633-644.

Levin, D. E. (2005). Cell wall integrity signaling in Saccharomyces cerevisiae.Microbiol. Mol. Biol. Rev. 69, 262-291.

Li, H. and Cao, Y. (2010). Protein mechanics: from single molecules to functional

biomaterials. Acc. Chem. Res. 43, 1331-1341.

Lipke, P. N., Garcia, M. C., Alsteens, D., Ramsook, C. B., Klotz, S. A. and Dufrene,

Y. F. (2012). Strengthening relationships: amyloids create adhesion nanodomains in

yeasts. Trends Microbiol. 20, 59-65.

Malinsky, J., Opekarova, M. and Tanner, W. (2010). The lateral compartmentation of

the yeast plasma membrane. Yeast 27, 473-478.

Marszalek, P. E. and Dufrene, Y. F. (2012). Stretching single polysaccharides and

proteins using atomic force microscopy. Chem. Soc. Rev. 41, 3523-3534.

McIntosh, B. E., Hogenesch, J. B. and Bradfield, C. A. (2010). Mammalian Per-Arnt-

Sim proteins in environmental adaptation. Annu. Rev. Physiol. 72, 625-645.

Morgan, M. R., Humphries, M. J. and Bass, M. D. (2007). Synergistic control of cell

adhesion by integrins and syndecans. Nat. Rev. Mol. Cell Biol. 8, 957-969.

Muller, D. J. and Dufrene, Y. F. (2011). Atomic force microscopy: a nanoscopic

window on the cell surface. Trends Cell Biol. 21, 461-469.

Muller, D. J., Helenius, J., Alsteens, D. and Dufrene, Y. F. (2009). Force probing

surfaces of living cells to molecular resolution. Nat. Chem. Biol. 5, 383-390.

Novak, P., Li, C., Shevchuk, A. I., Stepanyan, R., Caldwell, M., Hughes, S., Smart,

T. G., Gorelik, J., Ostanin, V. P., Lab, M. J. et al. (2009). Nanoscale live-cell

imaging using hopping probe ion conductance microscopy. Nat. Methods 6, 279-281.

Puchner, E. M. and Gaub, H. E. (2009). Force and function: probing proteins with

AFM-based force spectroscopy. Curr. Opin. Struct. Biol. 19, 605-614.

Radmacher, M., Tillamnn, R. W., Fritz, M. and Gaub, H. E. (1992). From molecules

to cells: imaging soft samples with the atomic force microscope. Science 257, 1900-

1905.

Rief, M., Gautel, M., Oesterhelt, F., Fernandez, J. M. and Gaub, H. E. (1997).

Reversible unfolding of individual titin immunoglobulin domains by AFM. Science

276, 1109-1112.

Rodicio, R. and Heinisch, J. J. (2010). Together we are strong—cell wall integrity

sensors in yeasts. Yeast 27, 531-540.

Roduit, C., van der Goot, F. G., De Los Rios, P., Yersin, A., Steiner, P., Dietler, G.,

Catsicas, S., Lafont, F. and Kasas, S. (2008). Elastic membrane heterogeneity of

living cells revealed by stiff nanoscale membrane domains. Biophys. J. 94, 1521-

1532.

Romero, D., Aguilar, C., Losick, R. and Kolter, R. (2010). Amyloid fibers provide

structural integrity to Bacillus subtilis biofilms. Proc. Natl. Acad. Sci. USA 107, 2230-

2234.

Schwartz, M. A. (2009). Cell biology. The force is with us. Science 323, 588-589.

Sheetz, M. P. (2001). Cell control by membrane-cytoskeleton adhesion. Nat. Rev. Mol.

Cell Biol. 2, 392-396.

Shibata, M., Yamashita, H., Uchihashi, T., Kandori, H. and Ando, T. (2010). High-

speed atomic force microscopy shows dynamic molecular processes in photoactivated

bacteriorhodopsin. Nat. Nanotechnol. 5, 208-212.

Sokurenko, E. V., Vogel, V. and Thomas, W. E. (2008). Catch-bond mechanism of

force-enhanced adhesion: counterintuitive, elusive, but ... widespread? Cell Host

Microbe 4, 314-323.

Telford, J. L., Barocchi, M. A., Margarit, I., Rappuoli, R. and Grandi, G. (2006).

Pili in gram-positive pathogens. Nat. Rev. Microbiol. 4, 509-519.

Verstrepen, K. J., Reynolds, T. B. and Fink, G. R. (2004). Origins of variation in the

fungal cell surface. Nat. Rev. Microbiol. 2, 533-540.

Vogel, V. and Sheetz, M. (2006). Local force and geometry sensing regulate cell

functions. Nat. Rev. Mol. Cell Biol. 7, 265-275.

Weber, C., Fraemohs, L. and Dejana, E. (2007). The role of junctional adhesion

molecules in vascular inflammation. Nat. Rev. Immunol. 7, 467-477.

Wilk, S., Wittland, J., Thywissen, A., Schmitz, H.-P. and Heinisch, J. J. (2010). A

block of endocytosis of the yeast cell wall integrity sensors Wsc1 and Wsc2 results in

reduced fitness in vivo. Mol. Genet. Genomics 284, 217-229.

AFM of cell surface mechanosensors 7