Journ

alof

Cell

Scie

nce

Nanopatterning reveals an ECM area threshold forfocal adhesion assembly and force transmission thatis regulated by integrin activation andcytoskeleton tension

Sean R. Coyer1,*, Ankur Singh1,*, David W. Dumbauld1, David A. Calderwood2, Susan W. Craig3,Emmanuel Delamarche4 and Andres J. Garcıa1,`

1Woodruff School of Mechanical Engineering and Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, Atlanta,GA 30330, USA2Department of Pharmacology, Yale University School of Medicine, New Haven, CT 06520-8066, USA3Department of Biological Chemistry, Johns Hopkins Medical School, Baltimore, MD 21205, USA4IBM Research GmbH, Zurich Research Laboratory, Zurich 8803, Switzerland

*These authors contributed equally to this work`Author for correspondence (andres.garcia@me.gatech.edu)

Accepted 15 July 2012Journal of Cell Science 125, 5110–5123� 2012. Published by The Company of Biologists Ltddoi: 10.1242/jcs.108035

SummaryIntegrin-based focal adhesions (FA) transmit anchorage and traction forces between the cell and the extracellular matrix (ECM). To gain

further insight into the physical parameters of the ECM that control FA assembly and force transduction in non-migrating cells, we usedfibronectin (FN) nanopatterning within a cell adhesion-resistant background to establish the threshold area of ECM ligand required forstable FA assembly and force transduction. Integrin–FN clustering and adhesive force were strongly modulated by the geometry of the

nanoscale adhesive area. Individual nanoisland area, not the number of nanoislands or total adhesive area, controlled integrin–FNclustering and adhesion strength. Importantly, below an area threshold (0.11 mm2), very few integrin–FN clusters and negligibleadhesive forces were generated. We then asked whether this adhesive area threshold could be modulated by intracellular pathways

known to influence either adhesive force, cytoskeletal tension, or the structural link between the two. Expression of talin- or vinculin-head domains that increase integrin activation or clustering overcame this nanolimit for stable integrin–FN clustering and increasedadhesive force. Inhibition of myosin contractility in cells expressing a vinculin mutant that enhances cytoskeleton–integrin coupling alsorestored integrin–FN clustering below the nanolimit. We conclude that the minimum area of integrin–FN clusters required for stable

assembly of nanoscale FA and adhesive force transduction is not a constant; rather it has a dynamic threshold that results from anequilibrium between pathways controlling adhesive force, cytoskeletal tension, and the structural linkage that transmits these forces,allowing the balance to be tipped by factors that regulate these mechanical parameters.

Key words: Cell adhesion, Fibronectin, Vinculin, Talin, Focal adhesion

IntroductionIntegrin-mediated adhesion to extracellular matrix (ECM)

components such as fibronectin (FN) transmits mechanical

forces and ‘outside-in’ biochemical signals regulating tissue

formation, maintenance, and repair (Hynes, 2002; Danen and

Sonnenberg, 2003; Wozniak and Chen, 2009). Integrin receptors

regulate their affinity for ligands by undergoing conformational

activation through ‘inside-out’ signaling via binding of the

cytoskeletal protein talin to the b-integrin tail (Hynes, 2002;

Calderwood and Ginsberg, 2003; Shattil et al., 2010). Following

activation and ligand binding, integrins cluster together into

nanoscale adhesive structures that function as foci for the

generation of strong anchorage and traction forces in stationary

and migrating cells (Balaban et al., 2001; Beningo et al., 2001;

Galbraith et al., 2002; Tan et al., 2003; Gallant et al., 2005).

These focal adhesion (FA) complexes consist of integrins and

actins vertically separated by a ,40 nm core that includes

cytoskeletal elements, such as vinculin, talin and a-actinin, and

signaling molecules, including FAK, src, and paxillin (Geiger

et al., 2001; Zamir and Geiger, 2001; Kanchanawong et al.,

2010). Integrin clustering is a crucial step in the adhesive process

promoting recruitment of cytoskeletal components, activation of

signaling molecules, and enhancing adhesive force (Miyamoto

et al., 1995a; Miyamoto et al., 1995b; Maheshwari et al., 2000;

Roca-Cusachs et al., 2009; Bunch, 2010; Petrie et al., 2010).

These integrin-based FAs serve as mechanosensors converting

environmental mechanical cues into biological signals (Geiger

et al., 2009).

Two central questions in mechanobiology are: (i) what

information is encoded by the physical properties of ECM

molecules?, and (ii) how are these physical properties sensed and

interpreted by cells as instructions to assemble a force-

transmitting adhesion complex (Bershadsky et al., 2006; Vogel

and Sheetz, 2006; Gardel et al., 2010; Parsons et al., 2010)?

5110 Research Article

Journ

alof

Cell

Scie

nce

Previous work has shown that both ECM rigidity and ligand

spacing influence FA and stress fiber assembly, cell spreading,

migration speed, and adhesive forces (Massia and Hubbell, 1991;

Garcıa et al., 1998a; Maheshwari et al., 2000; Wang et al., 2001;

Coussen et al., 2002; Koo et al., 2002; Jiang et al., 2003; Arnold

et al., 2004; Griffin et al., 2004; Engler et al., 2006; Cavalcanti-

Adam et al., 2007; Petrie et al., 2010; Selhuber-Unkel et al.,

2010). For instance, different average minimal RGD ligand

spacings are required for spreading (440 nm) and FA assembly

(140 nm) (Massia and Hubbell, 1991). More recent studies with

precisely spaced RGD ligands confirmed that FA assembly,

spreading, and migration require close integrin ligand spacing

(,90 nm) and showed that cells cannot integrate signals from

integrin–ligand complexes spaced more than 58 nm from each

other (Arnold et al., 2004; Cavalcanti-Adam et al., 2006).

However, the minimal area of integrin–ligand complexes

required to assemble an adhesion complex and transmit force

remains unknown.

The influence of the size of adhesive complexes on force

transmission has also been investigated, but a generalized

relationship between FA area and force has not been

established. Nascent punctate nanoscale (,0.2 mm2) FAs

assemble at the tips of filopodia and the leading edges of

lamellipodia (Fig. 1A) in an actin polymerization-dependent but

myosin-II-independent process to transmit traction forces

(Beningo et al., 2001; Cai et al., 2006; Choi et al., 2008;

Stricker et al., 2011). These nascent adhesions either undergo

rapid turnover within the lamellipodium or mature into larger,

elongated FAs under the influence of myosin-dependent

cytoskeletal tension (Chrzanowska-Wodnicka and Burridge,

1996; Riveline et al., 2001; Cai et al., 2006; Choi et al., 2008;

Gardel et al., 2008). Mature, large (1.0–10 mm2) FAs (Fig. 1A)

Fig. 1. Cells adhere to nanopatterned FN and form adhesive structures. (A) Nanoscale nascent adhesions (vinculin) formed at the lamellipodium and leading

edge of a fibroblast (mature adhesions: red arrowhead; nascent adhesions: white arrowhead). (B) FN patterns on PDMS stamps are transferred onto the substrates

with NHS esters resulting in the transfer and tethering of FN molecules onto the substrate. (C) Adhesive zone of nanoislands (500 nm64 shown); each zone

consists of a center square (262 mm) surrounded by eight radially distributed adhesion pads. Each adhesion pad consists of 1, 2, 4 or 9 square nanoislands.

Adhesive pads are presented in two orientations (45 , 90 ) relative to the center pad. (D) FN nanopatterns retain high spatial fidelity in culture as visualized by

direct imaging of Alexa-Fluor-555-labeled FN (FN 555) or indirect immunostaining using a polyclonal antibody against FN (poly FN). (E) Cells expressing GFP–

vinculin assemble vinculin-containing FAs on nanopatterns.

ECM area limit for adhesion assembly and force 5111

Journ

alof

Cell

Scie

nce

anchor cells to the substrate to maintain cell morphology and

tensional homeostasis (Balaban et al., 2001; Beningo et al., 2001;Tan et al., 2003). Using deformable substrates and traction forcemicroscopy, a linear relationship between FA area and traction

force was reported for large (.1 mm) FAs, indicating a constanttraction stress at these adhesive structures (Balaban et al., 2001;Tan et al., 2003). However, other studies have measured widelyvariable adhesive forces versus FA area for FAs (Beningo et al.,

2001; Goffin et al., 2006; Stricker et al., 2011). Recently, Gardeland colleagues elegantly demonstrated a strong correlationbetween FA size and traction force only during the initial

stages of FA maturation and growth, whereas mature adhesionsdid not exhibit this relationship (Stricker et al., 2011). Moreover,this group found a spatial dependence for traction forces across

an entire cell with higher traction forces transmitted by matureFAs near the cell periphery. Together, these studies show that FAfunction (i.e. force transmission) is not reliably predicted by FA

size.

To define additional ECM physical parameters that determineadhesive complex function, we examined whether there is aminimal area of integrin–ligand complexes required to assemble

an adhesive complex and transmit adhesive force. In the presentstudy, we used novel nanopatterned adhesive arrays of FNnanoislands within a non-adhesive background to address this

question. We found that stable assembly of integrin–FN clustersand generation of adhesive force depend on the area of individualadhesive nanoislands and not the number of adhesive contacts(number of nanoislands) or total adhesive area. Importantly, the

minimal area of integrin–FN clusters required for FA assemblyand force transmission exhibits a threshold that is not constant,but is instead regulated by recruitment of talin and vinculin and

the cytoskeleton tension applied to these adhesive clusters. Wepropose that this dynamic area threshold for stable FA assemblyand force transmission results from an equilibrium between

pathways controlling adhesive force, cytoskeletal tension, and thestructural linkage that transmits these forces. We further suggestthat perturbation of this force equilibrium is a local regulatory

mechanism for the assembly/disassembly of adhesive structuresand the transmission of adhesive forces.

ResultsFN nanoislands direct assembly of nanoscale adhesiveclusters

To study mechanobiology responses at the scale of individualadhesive structures, we engineered cell adhesion arrays with 250–

1000 nm square FN islands using modified subtractive contactprinting to covalently immobilize FN into defined nanopatterns ona cell adhesion-resistant background (Coyer et al., 2011) (Fig. 1B;supplementary material Fig. S1A). Several FN nanopattern

configurations were prepared to vary the nanoscale geometry interms of nanoisland size, number, and spacing (supplementarymaterial Fig. S2). We generated 10 mm diameter adhesive zones

with a center square (262 mm) surrounded by 8 radially distributedadhesion pads containing nanoislands (Fig. 1C). This configurationallowed for manipulation of the adhesive nanoisland geometry

independently from cell shape/spreading which was maintainedconstant for all nanopattern configurations. This is a criticalconsideration because both anchorage and traction forces are cell

shape and position dependent (Gallant et al., 2005; Stricker et al.,2011). Two adhesion pad orientations (90 , 45 ) relative to thecenter square are present in the adhesive zone. Each adhesion pad

consists of 1, 2, 4 or 9 square nanoislands. These nanoislands

present adhesive areas (0.06–1.0 mm2) corresponding to the size ofsmall, nascent FAs (Beningo et al., 2001; Choi et al., 2008) and arebelow the size of mature FAs (1–10 mm2) (Goffin et al., 2006;

Sniadecki et al., 2006). The edge-to-edge distance betweennanoislands is equal to the nanoisland side dimension. Adhesivepattern geometry is designated by the length of the side of thenanoislands and number of nanoislands; for example, 500 nm64

refers to adhesive pads with 4 nanoislands with sides of 500 nm(Fig. 1C). Printed nanoisland dimensions were confirmed byatomic force microscopy imaging (Coyer et al., 2007; Coyer

et al., 2011). Adhesive zones were spaced 100 mm apart from eachother in order to support adhesion of a single cell (supplementarymaterial Fig. S1B).

NIH3T3 murine fibroblasts cultured overnight (.16 h) adhered

to FN adhesive zones as single cells and remained nearly spherical(supplementary material Fig. S1B). To examine the stability andfidelity of FN nanopatterns in the presence of cells, fibroblasts were

cultured overnight in serum-containing medium on patterns printedwith human FN labeled with Alexa Fluor 555. Examination byfluorescence microscopy demonstrated that the printed pattern of

Alexa-Fluor-555-labeled FN was retained with high-fidelity anduniform intensity (Fig. 1D). Furthermore, immunostaining with apolyclonal antibody that reacts with human, bovine and murine FN

showed no changes in nanopattern integrity or intensity (Fig. 1D).No staining for FN outside the nanopatterned islands, includingbetween nanoislands, was detected. Vinculin-null mouse embryonicfibroblasts expressing GFP–vinculin showed patterns of vinculin

recruitment to FN nanoislands that corresponded to the size andlocation of the FN nanopatterns (Fig. 1E). These results demonstratethat the FN nanopatterns retain high fidelity in terms of spatial

distribution and density in overnight culture, indicating that cellscannot lay down secreted or serum-derived FN or reorganize theprinted FN on the surface.

Assembly of stable integrin–FN clusters exhibits an ECMarea threshold

We first examined whether assembly of stable, steady-state

integrin–FN clusters on the nanoislands is modulated by thenanoscale geometry of the adhesion interface. Clustering ofligand-bound integrins is an early step in the formation of dot-like nascent adhesions (Wiseman et al., 2004; Gardel et al.,

2010). We performed integrin immunostaining using a cross-linking and detergent extraction method that selectively retainsintegrins bound to FN and removes non-ligated integrins (Garcıa

et al., 1999; Keselowsky and Garcıa, 2005). Cells were culturedon FN nanopatterns overnight (.16 h) to allow the formation ofstable integrin–FN clusters. We have previously shown that both

integrin–FN complexes and adhesion strength reach stable,steady-state values after 8 hours (Gallant et al., 2005; Michaelet al., 2009). Adherent cells were incubated in the cell-

impermeable reagent sulfo-DTSSP to cross-link integrins toFN. Following SDS detergent extraction of cellular components,including uncross-linked integrins, immunostaining for a5integrin was performed. Previous analyses with function-

perturbing antibodies demonstrated that adhesion in theNIH3T3 cell model is primarily mediated by a5b1-integrin–FNwithout significant contributions from other receptors or

extracellular ligands (Gallant et al., 2005).

For the present analysis, Alexa-Fluor-555-labeled FN andAlexa-Fluor-488-labeled secondary antibodies were used to

Journal of Cell Science 125 (21)5112

Journ

alof

Cell

Scie

nce

simultaneously visualize FN nanoislands (red) and bound

integrins (green). After screening different nanopattern

configurations, we selected four adhesive zone configurations

for detailed analyses of integrin recruitment and localization.

Three patterns presented the same adhesive pad area (1.0 mm2)

but the pad area was distributed over 1 (1000 nm61), 4

(500 nm64) or 9 nanoislands (333 nm69). Because less is

known about the structure/size of nascent adhesions, a pattern

(250 nm64) with smaller adhesive pad area (0.25 mm2)

distributed over 4 smaller nanoislands was also examined.

Immunostaining revealed that integrin localization was

restricted to the nanoislands within the adhesive pads and the

center square only, and areas between the adhesive regions were

mostly devoid of integrin staining (Fig. 2A). In some instances,

Fig. 2. Nanoscale adhesive geometry modulates integrin–FN clustering. (A) Integrin recruitment to Alexa-Fluor-555-labeled FN nanopatterns was analyzed

using a cross-linking/extraction method that selectively retains FN-bound integrins. Fluorescence microscopy images for integrin binding (green) to FN (red)

adhesive clusters with different geometrical configurations. Scale bars: 1 mm. (B) Frequency maps for integrin recruitment to adhesive zones generated by

stacking individual images. The pseudo-color range represents the frequency of integrin localization to a spatial location with ‘warmer’ colors reflecting a higher

frequency of localized integrin staining. (C) High magnification frequency maps for integrin recruitment to adhesive pads for the 45˚and 90˚orientations. (D) FN

nanoislands in adhesive pads. Scale bars: 1 mm. (E) Binary images of the area corresponding to colocalization of integrin recruitment and FN nanoislands for the

45˚ and 90˚ pad orientation; images are oriented such that the edge closer to the center pad is at the top. (F) Frequency histograms of pad occupancy. Statistical

analyses of frequency distributions revealed the following differences (P,0.05): 1000 nm61.500 nm645333 nm69.250 nm64.

ECM area limit for adhesion assembly and force 5113

Journ

alof

Cell

Scie

nce

integrin staining was present over non-adhesive areas spanningadhesive nanoislands corresponding to ‘cell bridging’ (Lehnert

et al., 2004; Zimmermann et al., 2004; Rossier et al., 2010). Wedo not expect that integrins over the non-adhesive areascontribute significantly to adhesive force since there is no FNunderneath them (Fig. 1D) to support anchorage to the substrate.

Notably, the distribution and intensity of integrin staining onnanoislands differed significantly among the differentgeometrical patterns, with higher signal intensity to the

peripheral adhesive pads for the larger nanoislands (500 and1000 nm) and higher integrin localization to the center square forthe smaller nanoislands (250 and 333 nm).

In order to perform quantitative analyses of integrinrecruitment and clustering to FN nanopatterns, individualimages were stacked and color coded to generate frequencymaps of integrin–FN clustering (Fig. 2B,C). This analysis

exploits the controlled spatial arrangement of the FN islandsand allows us to extract the dominant spatial localization ofintegrins across multiple cells. As such, these frequency map

images represent a better descriptor of integrin localization forthe cell population compared to images of individual cells. Thistype of analysis has been used to identify spatial patterns of

proliferation and cell density in multicellular assemblies andfilters out low frequency occurrences, such as noise (Nelson et al.,2005; Nelson et al., 2006). Fig. 2B presents frequency maps forcells adhering to the four adhesive zone configurations, and

higher magnification images for the adhesive pads in the 45˚ and90˚ orientations are shown in Fig. 2C. In addition, the frequencymaps were overlaid onto images of the FN nanoislands (Fig. 2D)

to identify colocalization of integrin and FN within the adhesivepads (Fig. 2E). On 1000 nm61 patterns, integrins were recruitedto the adhesive pads and center square, and integrins uniformly

localized to the available single FN nanoisland in the adhesivepads. On 500 nm64 patterns, integrins were recruited to thecenter square as well as the nanoislands in the adhesive pads.

Although integrins localized to the 4 nanoislands in the adhesivepad, there was higher frequency of localization to the nanoislandsthat were closer to the center square. This enrichment in integrinrecruitment is most evident on the adhesive pad with the 45˚orientation. For 333 nm69 patterns, integrin recruitment wasenriched to the center square compared to the adhesive pads,where the frequency maps revealed more diffuse integrin

localization. In addition, integrins recruited to adhesive padscolocalized with the FN nanoislands, but only to those that werecloser to the center square. On the 250 nm64 patterns, integrins

clustered exclusively on the center square and very little integrinstaining was evident on the nanoislands. These resultsdemonstrate that stable integrin–FN clusters exhibit a nanoscale

area threshold (333 nm islands corresponding to 0.11 mm2)below which no stable complexes are formed.

It was clear from examining multiple images that the numberof adhesive pads occupied by integrins for a given adhesive zone

(each zone presents 8 adhesive pads to one cell) was dependenton the adhesive pad geometric configuration. We thereforescored the number of adhesive pads (out of 8) staining positive

for integrin–FN clustering and generated histograms for adhesivepad occupancy by integrins (Fig. 2F). For 1000 nm61 patterns,integrin–FN clusters predominantly occupied three or more

adhesive pads with over 50% of cells showing occupancy on all 8adhesive pads. In contrast, for 250 nm64 patterns integrinsshowed low pad occupancy with over 90% of adhesive zones

having two or less adhesive pads occupied by integrins. The

500 nm64 and 333 nm69 patterns resulted in integrin padoccupancies that were equally distributed from partial to fulloccupancy. These findings also support our conclusion that stable

integrin–FN clusters exhibit a nanoscale area threshold(0.11 mm2), below which no stable complexes are formed.

To further characterize these integrin–FN clusters, weexamined the contributions of Rho-kinase activity to steady-

state integrin–FN cluster formation on nanopatterned substrates.Contractility inhibitors have divergent effects on nascentadhesions compared to mature FAs. Inhibitors of Rho-kinase

and actomyosin contractility dissolve mature FAs and reducecorresponding adhesive forces (Burridge and Chrzanowska-Wodnicka, 1996; Amano et al., 1997; Balaban et al., 2001; Tanet al., 2003; Dumbauld et al., 2010). In contrast, inhibition of

Rho-kinase upregulates the number of nascent adhesions in thelamellipodium (Alexandrova et al., 2008) whereas the formationrate of nascent adhesions is myosin-II-independent (Choi et al.,

2008). We examined whether inhibition of contractility using Y-27632, a specific inhibitor of Rho-kinase that reduces myosinlight chain phosphorylation, cell contractility, FA assembly and

FA-dependent forces (Dumbauld et al., 2010; Kuo et al., 2011),influences assembly of integrin–FN clusters on nanoislands. Cellswere cultured on FN nanopatterns overnight and treated with Y-

27632 (10 mM) 30 min prior to analysis. For 500 nm64 patterns,treatment with Y-27632 reduced the frequency of integrinlocalization within the nanoislands (Fig. 3A,B). However,treatment with Y-27632 did not eliminate integrin localization

to FN nanoislands (Fig. 3D) or alter pad occupancy (Fig. 3E),especially when compared to the unoccupied adhesive pads forthe 250 nm64 islands (Fig. 2E). These results demonstrate that

Rho kinase activity reduces, but does not eliminate, steady-stateintegrin–FN clustering, and does not affect pad occupancy. Thepartial myosin dependence suggests that these engineered

nanoscale, steady state adhesions reflect the properties of initialadhesions that are constrained from subsequent increase in areaby the geometric arrangement of the ECM.

Nanoscale adhesive geometry modulates adhesive force

To examine the effects of nanoscale adhesive geometry onadhesive forces, we quantified the force required to detach cellsfrom the adhesive zones after a 16-hour culture time point using a

spinning disk device (Garcıa et al., 1998b; Gallant et al., 2005). Adetachment profile (adhesion fraction f versus shear stress t) wasfit to a sigmoid curve to obtain the shear stress for 50%

detachment (t50), defined here as the cell adhesion strength.Fig. 4A presents typical detachment profiles showing sigmoidaldecreases in the fraction of adherent cells as a function of shear

stress for two nanopattern configurations. The right-ward shift inthe detachment profile for the 1000 nm61 pattern compared tothe center square-only pattern (no adhesive pads) reflects a 2.2-

fold increase in adhesive force.

Cell adhesion strength was quantified for adhesive zoneconfigurations with different adhesive pad areas, nanoislandssizes, and number of nanoislands. Fig. 4B summarizes results for

adhesive force as a function of adhesive pad area and number ofnanoislands per adhesive pad. The upper bound (top dashed line)represents the adhesion strength for a 10 mm diameter

micropatterned area (adhesive area 78.5 mm2), whereas thelower bound (bottom dashed line) corresponds to the adhesionstrength for a pattern with 262 mm center square but no adhesive

Journal of Cell Science 125 (21)5114

Journ

alof

Cell

Scie

nce

pads or nanoislands (adhesive area 4.0 mm2). For most

nanopattern configurations, adhesion strength values were

higher than the lower bound, indicating that FN nanoislands

significantly contribute to adhesive force. A 650% reduction in

total available adhesive area (10 mm diameter circle versus

1000 nm61 pattern) resulted in only a 25% reduction in adhesiveforce. This result is consistent with our previous work

demonstrating that adhesive strength is controlled by small

adhesive areas at the periphery of the cell (corresponding to FAs)

and that the majority of the available adhesive interface does not

contribute significantly to adhesive force (Gallant et al., 2005).

The adhesion strength value for all patterns with nanoisland

dimensions below 333 nm was equivalent to the lower bound (no

adhesive pads), indicating no appreciable contributions to

adhesive force for these nanoislands (Fig. 4B). For example,there are no differences in adhesion strength for 250 nm islands

regardless of whether each pad contained 2, 4 or 9 islands, andthe adhesion strength for these nanoislands is equivalent tocenter-only patterns that have no nanoislands. This result isconsistent with the integrin recruitment results and shows the

functional consequences of the area threshold of integrin–FNclustering to adhesive force. Furthermore, we noticed that the500 nm61 and 500 nm64 patterns, which have same nanoisland

dimensions but different number of nanoislands (1 versus 4), andtherefore, different adhesive pad areas (0.25 versus 1.0 mm2),produced equivalent adhesion strength values (245615 versus

252611 dynes/cm2). This result suggests that individualnanoisland area, independently from the number of nanoislandsand pad adhesive area, controls adhesion strength. Indeed, asshown in Fig. 4C, when adhesion strength is plotted as a function

of the individual nanoisland area, the data points from differentnanopattern configurations collapse into a single curve. Non-linear regression with a logarithmic curve indicated that this

functional dependence accounts for over 92% of the variance inthe data (P,0.0007). Additionally, no differences in adhesionstrength were observed between configurations with the same

adhesive pad area (0.25 mm2) and number of islands (4) but withdifferent inter-island spacings (0.75 versus 1.25 mm; Fig. 4D),indicating that the spacing between nanoislands does not

contribute appreciably to adhesive force. This analysisdemonstrates that, above an area threshold of 0.11 mm2,individual nanoisland area, and not the geometric arrangementof the adhesive area (number of islands, island spacings, total pad

area), regulates the adhesive force generated by stable integrin–FN clusters. Below this area threshold, no appreciable adhesiveforces are generated, correlating with the lack of stable integrin–

FN clusters at steady state.

Talin controls the stable assembly of integrin–FN clustersat nanoscale dimensions

What determines the area threshold for assembly of stableintegrin–FN clusters? Extrinsic, cell-independent factors relatedto ligand presentation/density could dictate integrin clustering.

However, the observed area threshold corresponds to ananoisland area that is considerably (at least 50-fold) largerthan the minimal ligand spacing (,100 nm) required for focal

adhesion assembly and spreading (Massia and Hubbell, 1991;Arnold et al., 2004; Cavalcanti-Adam et al., 2007). Alternatively,cell-dependent, intrinsic mechanisms could regulate this ECM

area ‘switch’ in the assembly of integrin–FN clusters. We firstpostulated that the nanoscale area threshold in integrin–FNclustering results from an insufficient number of activated and

bound integrins required to establish a stable nascent integrincluster. Therefore, we hypothesized that increased integrinactivation would overcome this ‘nanolimit’ for stable integrinrecruitment.

Talin is an elongated (,60 nm) flexible anti-parallel dimerthat interacts with several proteins in FAs (Critchley, 2009).Notably, the FERM domain in the globular N-terminus of talin

[talin1(1–405)] binds to an NP(I/L)Y motif in the cytoplasmictail of integrin-b subunits to activate the integrin (Calderwoodet al., 1999; Tadokoro et al., 2003; Bouaouina et al., 2008; Goult

et al., 2010). To determine whether integrin activation canovercome the ‘nanolimit’ of integrin–FN clustering on 250 nmislands, cells were transfected with plasmids encoding GFP–

Fig. 3. Effects of Rho-kinase inhibition on integrin–FN clustering on

nanoscale patterns. (A) Frequency maps for integrin clustering to 500 nm64

patterns for control and Y-27632 (10 mM)-treated cells. (B) High

magnification frequency maps for integrin recruitment to adhesive pads for

the 45˚and 90˚orientations. Scale bars: 1 mm. (C) FN nanoislands in adhesive

pads. (D) Binary images of area corresponding to colocalization of integrin

recruitment and FN nanoislands for the 45˚ and 90˚ pad orientation; images

are oriented such that the edge closer to the center pad is at the top.

(E) Histograms of pad occupancy.

ECM area limit for adhesion assembly and force 5115

Journ

alof

Cell

Scie

nce

talin1(1–405) or GFP–talin1(1–405)A/E with an integrin

binding-defective mutation (Bouaouina et al., 2008). Expression

of talin1(1–405) significantly enhanced integrin clustering to the

250 nm islands compared to untransfected control cells, whereas

the integrin binding-defective talin mutant had no effects on

integrin recruitment (Fig. 5A,B). There were no significant

differences in integrin recruitment on the larger 500 nm islands

between talin head domain and controls (Fig. 5A,B). Expression

of talin head domain also significantly altered the frequency of

pad occupancy for the 250 nm64 patterns. Talin1(1–405)-

expressing cells on 250 nm64 patterns mostly occupied three

to six adhesive pads (Fig. 5C). Cells expressing the A/E mutant

or control cells showed low pad occupancy on 250 nm64 islands

with over 95% of cells having two or less adhesive pads occupied

by integrin clusters. The 500 nm64 patterns exhibited integrin

pad occupancies that were equally distributed from partial to full

occupancy, but talin1(1–405) expression increased the full pad

occupancy by 45% compared to A/E mutant and control cells

(Fig. 5D). Importantly, adhesion strength values for talin1(1–

405)-expressing cells on 250 nm64 patterns were 1.8-fold higher

than control talin1(1–405)A/E-expressing cells (Fig. 5E;

P,0.05). These results demonstrate that talin-head domain

triggers assembly of integrin–FN nanoclusters below the ECM

area threshold and significantly increases adhesive force.

Vinculin regulates nanoscale assembly of integrin–FN

clusters by balancing adhesive force and cytoskeletal

tension

We next examined the role of the FA protein vinculin in the

nanoscale organization of integrin–FN clusters. Vinculin contains

a talin binding site in its globular head and actin binding sites in

the tail domain, and interactions with these partners are regulated

by an auto-inhibited conformation arising from strong head–tail

binding (Cohen et al., 2006). Importantly, vinculin is a force-

carrying component between adhesive sites and the cytoskeleton

(Grashoff et al., 2010). The vinculin head domain promotes

integrin clustering and increases residence times in mature FAs in

spread cells (Cohen et al., 2005; Cohen et al., 2006; Humphries

et al., 2007). We examined the effects of the vinculin head

domain (VH, 1–851) on the area threshold of integrin–FN

clustering using vinculin-null cells expressing VH. In contrast to

cells expressing wild-type vinculin, cells expressing VH

assembled integrin–FN clusters on the 250 nm nanoislands

(Fig. 6A,B). These results show that the vinculin head domain

Fig. 4. Nanoscale adhesive geometry regulates cell adhesion strength. Cell adhesive force to FN nanopatterns was measured using a spinning disk assay.

(A) Detachment profiles (adhesive fraction versus shear stress) for cells adhering to 1000 nm61 and center-only patterns. Experimental points were fitted to a

sigmoid curve to calculate the shear stress for 50% detachment, which represents the mean adhesion strength. Vertical dashed lines show the shear stress for 50%

detachment for each profile. (B) Adhesion strength as a function of adhesive pad area for different nanoisland configurations. Values are means 6 s.e.m. The top

and bottom dashed lines correspond to the adhesion strengths for a 10 mm diameter circular area and the center-only pattern, respectively. (C) Adhesion strength

as a function of individual nanoisland area (log scale). Values are means 6 s.e.m., and logarithmic (natural base) fit is shown (solid line). The dashed line

corresponds to the adhesion strength for the center-only pattern. Results for 0.0625 mm2 and 0.250 mm2 comprise 3 (250 nm62, 250 nm64, 250 nm69) and 2

(500 nm61, 500 nm64) nanoisland patterns, respectively. (D) Adhesion strength values for 250 nm64 patterns with different inter-island spacings (0.75 versus

1.25 mm), showing no differences in adhesive force.

Journal of Cell Science 125 (21)5116

Journ

alof

Cell

Scie

nce

can overcome the nanoscale limit of integrin–FN clustering,

presumably via interactions with talin. The finding that cells

expressing wild-type vinculin did not assemble stable integrin–

FN clusters on the 250 nm nanoislands could be explained by the

auto-inhibitory interaction between vinculin head and tail

domains. To examine this possibility, we cultured vinculin-null

cells expressing the vinculin T12 mutant (VinT12) on the

nanopatterned substrates. VinT12 has a mutated head-tail

interface that reduces the head-tail affinity 100-fold and thus

exposes binding sites for talin and actin (Cohen et al., 2005;

Cohen et al., 2006). This mutant also drives integrin clustering

and FA assembly (Cohen et al., 2005; Cohen et al., 2006;

Humphries et al., 2007). Surprisingly, expression of VinT12

failed to promote assembly of integrin–FN clusters on the

250 nm nanoislands even though robust clusters were assembled

on the 500 nm patterns (Fig. 6A,B). Expression of VH also

increased pad occupancy (Fig. 6C) on 250 nm islands, whereas

VinT12 expression did not enhance pad occupancy. Pad

occupancy was similar for VH- and VinT12-expressing cells on

500 nm islands (Fig. 6D). This result supports an alternative

explanation in which the nanoscale area threshold for integrin-FN

clustering is regulated by cytoskeletal tension.

Whereas the head domain of VinT12 interacts with talin in a

similar fashion as VH to drive integrin clustering, the tail domain

of VinT12 can also interact with actin filaments to transmit

cytoskeletal forces to adhesive structures. We therefore

hypothesized that the area threshold for assembly of stable

integrin–FN clusters at nanoscale adhesions is also regulated by

cytoskeletal tension. Below the area threshold, the adhesive force

generated by the integrin–FN clusters cannot support the

Fig. 5. Talin head expression drives integrin–FN clustering at the nanoscale. Integrin binding analysis for talin-expressing cells. (A) Fluorescence microscopy

images for integrin binding (green) to FN (red) adhesive zones. Scale bars: 1 mm. (B) Frequency maps for integrin recruitment on 250 nm64 and 500 nm64

patterns generated by stacking individual images. For 250 nm patterns, expression of talin1(1–405) induced recruitment and clustering of integrins

compared to control cells. (C,D) Frequency histograms for pad occupancy on (C) 250 nm64 and (D) 500 nm64 patterns. (E) Cell adhesive force response to FN

nanopatterns with talin head expression. Bar graphs represent fold change in adhesion strength over talin1(1-405)A/E-transfected cells adhering to 250 nm64

patterns (means 6 s.d.; *P,0.05).

ECM area limit for adhesion assembly and force 5117

Journ

alof

Cell

Scie

nce

cytoskeletal tensile force and the integrin clusters are

disassembled or detached from the adhesive interface because

of limited adhesive size of the nanoislands. To test this

hypothesis, vinculin-null cells expressing VinT12 cultured

overnight on the 250 nm nanoislands were treated with

blebbistatin (20 mM) 60 min prior to analysis. Blebbistatin is

an inhibitor of non-muscle myosin IIA that blocks actin–myosin

contractility (Allingham et al., 2005). Cells expressing VinT12

and treated with blebbistatin displayed integrin–FN clusters on

250 nm islands similar to those in VH-expressing cells (Fig. 7A),

whereas untreated VinT12-expressing cells did not exhibit

integrin–FN clusters on 250 nm islands. Blebbistatin treatment

had no effect on integrin–FN cluster formation on 250 nm

nanoislands for cells expressing wild-type vinculin

(supplementary material Fig. S3); this result is not surprising

given the inability of cells expressing wild-type vinculin to form

integrin–FN clusters on the 250 nm islands. Blebbistatin

treatment also increased adhesive pad occupancy compared to

control VinT12-expressing cells (Fig. 7B). These results

demonstrate that a reduction in the cytoskeletal tension applied

to integrin–FN nanoclusters in the presence of a vinculin mutant

that drives integrin clustering allows for stable assembly of

adhesive structures below the ECM area nanolimit. Taken

together, these findings support a model where stable FA

assembly and ECM–cell adhesive forces are regulated by the

force equilibrium between cytoskeletal tension and adhesive

forces.

DiscussionA standing question in mechanobiology is how cells sense

geometrical ECM cues and transmit local forces at FAs

(Bershadsky et al., 2006; Vogel and Sheetz, 2006; Gardel et al.,

2010; Parsons et al., 2010). Several studies have shown that ECM

ligand spacing regulates FA and stress fiber assembly, cell

spreading and migration, and adhesive forces (Massia and

Hubbell, 1991; Garcıa et al., 1998a; Maheshwari et al., 2000;

Fig. 6. Vinculin domains regulate stable assembly of integrin–FN nanoclusters. (A) Fluorescence microscopy images for integrin binding (green) to FN (red)

adhesive zones for wild-type (WT), VH and VinT12-expressing cells on 250 nm64 (top) and 500 nm64 patterns (bottom). Scale bars: 1 mm. (B) Frequency maps

for integrin recruitment on 250 nm64 and 500 nm64 patterns generated by stacking individual images. For 250 nm islands, expression of VH induced clustering

of integrins to FN nanoislands, whereas VinT12 did not induce any recruitment. (C,D) Frequency histograms for pad occupancy for VH and VinT12 cells on

(C) 250 nm64 and (D) 500 nm64 patterns.

Journal of Cell Science 125 (21)5118

Journ

alof

Cell

Scie

nce

Coussen et al., 2002; Koo et al., 2002; Jiang et al., 2003; Arnold

et al., 2004; Cavalcanti-Adam et al., 2007; Petrie et al., 2010;

Selhuber-Unkel et al., 2010). However, whether the geometric

organization of the ECM ligand regulates FA assembly and force

transmission has not been addressed. Here, we used

nanopatterned substrates with defined geometrical arrangements

of FN to restrict the formation of adhesive structures to the

nanoscale dimensions characteristic of FAs to study how

nanogeometry regulates assembly of integrin–FN clusters and

the generation of ECM–cell anchorage forces. The smallest

nascent adhesion size reported is 0.19 mm2 (Choi et al., 2008),

although these authors speculated that the actual size is smaller.

Our nanopatterning approach offers the ability to identify even

smaller adhesive structures and we showed formation of tiny

0.06 mm2 adhesions (Figs 6, 7). We demonstrated that integrin

clustering on FN islands and adhesive force were modulated by

nanoscale ECM area; below a threshold of 0.11 mm2, no stable

integrin–FN clusters were assembled or adhesive forces

generated. Expression of talin head or vinculin head domains

that increase integrin activation or clustering overcame this nano-limit. Inhibition of myosin contractility in cells expressing a

vinculin molecule that enhances the coupling between the force-generating actin cytoskeleton and integrins also restoredintegrin–FN clustering below the area threshold. We concludethat the size of the integrin–FN clusters required for stable FA

assembly and force generation has an ECM area threshold that isnot constant, but is instead regulated by intracellular proteins(talin, vinculin) that influence the force equilibrium between the

adhesive force generated by the integrin–FN clusters, which isrelated to the nanoscale area and number of FN-bound integrins,and the cytoskeletal tension applied to the clusters.

A critical advantage of our patterning strategy is that theoverall cell shape is constrained within the 10 mm diameteradhesive zone for all nanopattern configurations. This approachallows decoupling of integrin–FN cluster formation from cell

spreading and cells cannot spread (or retract). Importantly, thedistance between the adhesive pads containing the nanoislandsand the center square is fixed across all nanopattern

configurations, ensuring equivalent force loading in allexperiments. Although we observed a higher frequency ofintegrin–FN clusters on nanoislands that were closer to the

center square, we do not expect that this reduced distance altersforce loading because this distance is significantly smaller thanthe overall distance between the adhesive pad and the centersquare. Indeed, our calculations estimate that enrichment of

integrin–FN clusters towards the center square altered theresultant adhesive force by less than 1.3%. Finally, while thefocus of the present study was to analyze the assembly of stable,

steady-state integrin–FN clusters and adhesion strength, in thefuture, it will be of interest to perform time course analyses ofintegrin–FN complex assembly and adhesive forces.

Unfortunately, this experiment is not presently possible due totechnical limitations associated with live cell imaging ofnanometer-scale adhesive clusters on the gold-coated substrates.

We propose a force equilibrium model for the ECM-area-regulated assembly of stable integrin–FN clusters (Fig. 8).Activated integrins bind to ECM ligands in the adhesiveinterface and assemble into small adhesive clusters containing

talin and vinculin. The actin cytoskeleton applies tensile force tothese nascent adhesive clusters via actin-myosin contractility.This cytoskeletal force is balanced by the adhesive force

generated by the integrin–FN clusters. Below an area threshold(0.11 mm2), the adhesive force cannot support the cytoskeletaltension and the integrin cluster is unstable and is disassembled or

detached from FN. Above this area threshold, the integrin–FNcluster is large enough to generate sufficient force to support theapplied tension (Fig. 8A). The ECM area threshold for assemblyof stable integrin–FN clusters is therefore regulated by (i) the

adhesive force generated by the integrin–FN clusters, which isrelated to the nanoscale ECM area and number of boundintegrins, and (ii) the cytoskeletal tension applied to the clusters.

Increases in the adhesive force generated by the integrin–FNclusters via integrin activation or clustering through binding totalin head or vinculin head result in a stable adhesive cluster that

can support cytoskeletal tension with smaller adhesive areas(Fig. 8A). Conversely, decreases in the applied cytoskeletaltension via inhibition of myosin contractility result in a reduction

in the force driving disassembly of the integrin clusters (Fig. 8A).Fig. 8B provides quantitative relationships that fully support thismodel. Because the size and position of the adhesive clusters and

Fig. 7. Actomyosin contractility controls stabilization of integrin–FN

clusters at nanoscale dimensions. (A) Fluorescence microscopy images for

integrin binding (green) to FN (red) adhesive zones of VinT12 cells on

250 nm64 patterns in the presence or absence of blebbistatin (20 mM). Scale

bars: 1 mm. (B) Frequency maps for integrin recruitment on 250 nm64 for

VinT12 cells in the presence or absence of blebbistatin (20 mM), generated by

stacking individual images. (C) Frequency histograms for pad occupancy of

VinT12 cells on 250 nm64 patterns in the presence or absence of blebbistatin

(20 mM).

ECM area limit for adhesion assembly and force 5119

Journ

alof

Cell

Scie

nce

cell morphology are defined in our experimental system, we can

convert the measured values for adhesion strength (Fig. 4) to the

cell–ECM adhesive force generated on the nanoislands (FAdh) as

a function of nanoisland dimension (solid black line). For the

cytoskeletal tension (FCSK), we considered two estimates based

on traction forces measured on deformable substrates. Gardel and

colleagues recently reported the traction stresses during the initial

stages of FA assembly (Stricker et al., 2011); the range for

maximum traction forces for these nascent complexes is

indicated by the gray box in Fig. 8B. Plotted also is the

traction force for larger, mature FAs derived by Geiger and

colleagues (Balaban et al., 2001) that increases with FA size

(dashed gray line) (Balaban et al., 2001). Several noteworthy

observations are evident from Fig. 8B. First, FCSK is larger than

FAdh for small nanoislands but there is a cross-over point for

which FAdh becomes larger than FCSK. This cross-over point

corresponds to the area threshold. For either FCSK model, the area

threshold occurs below areas with dimensions below 333 nm

(,280–300 nm) in excellent agreement with our experimental

data for integrin–FN cluster assembly (Fig. 2). Remarkably,

when the value for adhesion strength for the talin head domain on

250 nm nanoislands is included (open symbol), the resulting FAdh

exceeds the FCSK, also in agreement with the result that

expression of talin head promotes assembly of stable integrin–

Fig. 8. Force equilibrium model for ECM-area-regulated assembly of integrin–FN clusters. (A) On nanoislands (side dimension L), activated integrins bind to FN

and assemble into small adhesive clusters containing talin and vinculin. The stability of these nascent adhesive clusters is dependent on the balance between the force of

integrin–FN adhesion (FAdh) and the cytoskeleton-mediated tensile force (FCSK). For nanoislands with dimensions larger than the threshold area (L§Lcritical), the

integrin–FN adhesive force can support the applied cytoskeletal tension and stable integrin–FN clusters assemble on nanoislands. For nanoislands with dimensions

smaller than the threshold area (L#Lcritical), the applied cytoskeletal tension exceeds the integrin–FN adhesive force and no stable integrin–FN complexes are assembled.

Recruitment of talin/vinculin and modulation of the applied cytoskeletal tension to the integrin–FN clusters alters the force equilibrium to regulate integrin–FN complex

assembly. (B) Plot for FAdh and FCSK as a function of nanoscale area showing area threshold (cross-over point). FAdh (solid black line) was derived from experimental

adhesion strength values (filled circles); open circle shows the FAdh for talin1(1–405)-expressing cells. Estimates for FCSK based on the work of Gardel and colleagues

(Stricker et al., 2011) are shown as a gray box and those based on the work of Geiger and colleagues (Balaban et al., 2001) are shown as a dashed line (see text).

Journal of Cell Science 125 (21)5120

Journ

alof

Cell

Scie

nce

FN clusters on the 250 nm nanoislands (Fig. 5). This findingshows that talin-based integrin activation stabilizes integrin–FN

clusters and enables these nanoscale structures to transmit largeadhesive forces.

This force equilibrium model for ECM-area-controlledassembly of integrin–FN clusters provides a simple, local

regulatory mechanism for the assembly/disassembly of adhesivestructures. During leading edge protrusion in cell migration, small(,0.19 mm2) nascent adhesions (Choi et al., 2008) assemble and

either disassemble or become stable and grow into matureadhesions (Parsons et al., 2010). The mechanism(s) regulatingthese adhesion dynamics is not clear but two models have been

proposed (Parsons et al., 2010). In the first model, nucleation ofadhesion clusters is initiated by integrin binding, clustering andrecruitment of cytoskeletal proteins (e.g. vinculin, talin), and thesenascent structures grow and mature in response to contractile

forces. In the second model, FA assembly is coupled to actinpolymerization wherein vinculin and FAK bind to Arp2/3 andthese complexes then bind ECM-bound integrins to stabilize the

nascent adhesion in a myosin-independent fashion. The forceequilibrium model is consistent with elements of first model andprovides a simple biophysical nanoscale switch to control the

nucleation and growth of adhesive structures. The finding thatintegrin activation via binding of talin head or driving clusteringvia vinculin head domain overcomes the nanoscale limit for stableintegrin–FN cluster assembly supports the explanation that

integrin activation and binding drive assembly of nascentadhesive clusters (Cohen et al., 2005; Cohen et al., 2006;Humphries et al., 2007; Zhang et al., 2008). The assembly of

integrin clusters on 250 nm nanoislands in the presence ofblebbistatin is also consistent with previous reports showing thatmyosin activity is not required for formation of nascent adhesions

(Alexandrova et al., 2008; Choi et al., 2008). Our results with thevinculin head domain driving assembly of integrin–FNnanoclusters suggest that Arp2/3 is not directly involved in this

process because this vinculin domain lacks that the proline-richstrap (amino acids 851–880) that contains the Arp2/3 binding site(DeMali et al., 2002). Finally, the force-balance model alsoprovides an attractive explanation for mechanosensitive changes in

FA assembly. For instance, application of local external forces toadhesive clusters would perturb the local force balance and driveFA growth to increase the net adhesive force. This prediction is

consistent with the observation that application of external forcesto adhesive plaques results in Rho-dependent directional focaladhesion growth (Riveline et al., 2001). Similarly, changes in

substrate stiffness could alter the local force balance to regulate thesize of FA structures without significant changes in localcytoskeletal forces, in good agreement with published results

(Yeung et al., 2005; Aratyn-Schaus et al., 2011).

In conclusion, we demonstrate that integrin–FN clustering andadhesive force are strongly modulated by the geometry of thenanoscale adhesive area. Stable assembly of integrin–FN clusters

and adhesive force depend on the area of individual adhesivenanoislands and not the number of adhesive contacts or totaladhesive area. Importantly, the minimal size of integrin–FN clusters

required for FA assembly and force transmission exhibits an areathreshold that is not constant, but is instead regulated by recruitmentof talin and vinculin and the cytoskeleton tension applied to these

adhesive clusters. We propose that this dynamic area thresholdresults from an equilibrium between pathways controlling adhesiveforce, cytoskeletal tension, and the structural linkage that connects

these forces. This force equilibrium acts as a simple, local regulatory

mechanism for the assembly/disassembly of nascent adhesive

structures and the transmission of adhesive forces.

Materials and MethodsCells and reagents

NIH3T3 fibroblasts (American Type Culture Collection, Manassas, VA, USA) werecultured in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 10%newborn calf serum (HyClone, Logan, UT, USA) and 1% penicillin-streptomycin(Invitrogen). Cell culture reagents, including human plasma FN and Dulbecco’sphosphate-buffered saline (PBS), were purchased from Invitrogen. BSA waspurchased from Sigma-Aldrich (St Louis, MO, USA). Antibodies against a5 integrin(ab1921, Millipore, Billerica, MA, USA), and human FN (anti-hFN polyclonalantibody, Sigma-Aldrich) were used for immunostaining. Alexa-Fluor-488-conjugated secondary antibodies and Alexa Fluor 555 succinimidyl ester werepurchased from Invitrogen. Cross-linker 3,3-dithiobis(sulfosuccinimidylpropionate)(DTSSP) was purchased from Pierce Chemical (Rockford, IL, USA).Poly(dimethylsiloxane) (PDMS) elastomer and curing agent (Sylgard 184) wereproduced by Dow Corning (Midland, MI, USA). ZEP520A was purchased fromZeon Chemicals (Tokyo, Japan). Amyl acetate was produced by Mallinckrodt Baker(Phillipsburg, NJ, USA), and n-methyl pyrrolidinone (NMP, 1165 Remover) wasobtained from MicroChem (Newton, MA, USA). Tri(ethylene glycol)-terminatedalkanethiol [HS-(CH2)11-(OCH2CH2)3-OH; EG3] and carboxylic acid-terminatedalkanethiol [HS-(CH2)11-(OCH2CH2)6-OCH2-COOH; EG6-COOH] were purchasedfrom ProChimia Surfaces (Sopot, Poland). N-hydroxysuccinimide (NHS), N-(3-dimethylaminopropyl)-N9-ethylcarbodiimide hydrochloride (EDC), and 2-(N-morpho)-ethanesulfonic acid were purchased from Sigma-Aldrich.

Talin plasmids and expression in cells

NIH3T3 fibroblasts were transfected using an Amaxa Nucleofector II (Lonza, Basel,Switzerland). Constructs for GFP–talin1(1–405) and GFP–talin1(1–405)A/E withintegrin binding-defective mutation have been described (Bouaouina et al., 2008).Talin1(1–405) included the entire FERM and has been previously shown to activatea5b1 integrin (Bouaouina et al., 2008). For each sample, 2 million cells wereresuspended in 100 ml of Nucleofector II Solution R (Lonza) with 2.5 mg of plasmidand transfected using program U-30. Immediately after transfection, cells weretransferred to 1.5 ml centrifuge tube containing 500 ml of pre-warmed RPMI 1640 andincubated for 15 min. Cells were then transferred into 100 mm plates containingnormal growth medium. 24 h after transfection, cells were enriched using flowcytometry for GFP expression and studies were performed 48 hours after transfection.

Vinculin cell lines

Retroviral plasmids pTJ66-tTA and pXF40 were previously described (Gersbachet al., 2006). eGFP-C1 WT vinculin and eGFP-T12 vinculin plasmids have beendescribed (Chen et al., 2005). One AgeI restriction site was inserted into the multiplecloning site of pXF40, the retroviral expression vector. The oligonucleotides59-AGCTTGTCAGCTACCGGTGCTACTGCA-39 and 59-AGCTTGCAGTAGC-ACCGGTAGCTGACA-39 (AgeI sequences underlined) were annealed together,creating HindIII-compatible overhangs at each end. This product was then ligatedinto a linearized pXF40 vector which had been digested with HindIII. Finally, theeGFP–vinculin constructs were digested from the pEGFP-C1 with AgeI and SalI andligated into the SalI and AgeI-digested pXF40 vector. The pXF40-eGFP-vinculinWT, VH and T12 vectors transcribe the eGFP-vinculin gene from the tetracycline-inducible promoter. All vectors were verified by sequencing the ligation points.

Retroviral stocks were produced by transient transfection of helper virus-free øNXamphotropic producer cells with plasmid DNA as previously described (Byers et al.,2002). Vinculin-null mouse embryonic fibroblasts (Xu et al., 1998), a kind gift fromEileen Adamson (Sanford-Burnham Medical Research Institute), were plated ontissue culture polystyrene at 26104 cells/cm2 24 h prior to retroviral transduction.Cells were transduced with 0.2 ml/cm2 of equal parts pTJ66-tTA and pXF40-eGFP-vinculin retroviral supernatant supplemented with 4 mg/ml hexadimethrine bromideand 10% fetal bovine serum, and centrifuged at 1200 g for 30 min in a Beckman GS-6R centrifuge with a swinging bucket rotor. Retroviral supernatant was replaced withgrowth medium (DMEM, 10% fetal bovine serum, 100 U/ml penicillin G sodium,100 mg/ml streptomycin sulfate, 1% non-essential amino acids, 1% sodiumpyruvate). Five days after transduction, eGFP-expressing cells were FACS sorted,expanded, and either used for experimentation or cryopreserved in liquid nitrogenfor later use. Expression of vinculin constructs was verified by western blot andimmunofluorescence microscopy (data not shown).

Nanopatterned surfaces

Mixed self-assembled monolayers (SAMs) of tri(ethylene glycol)-terminated(EG3) and carboxylic acid-terminated (EG6-COOH) alkanethiols on gold wereused to present anchoring groups for covalent immobilization of FN within a non-fouling background as reported earlier and showed in supplementary material Fig.S1 (Coyer et al., 2007; Coyer et al., 2011).

ECM area limit for adhesion assembly and force 5121

Journ

alof

Cell

Scie

nce

Cell seeding and integrin cross-linking

Alexa-Fluor-555-conjugated FN to was used to visualize patterns of printed

protein. In order to leave free primary amines on FN for tethering to mixed self-

assembled monolayers, a ratio of 25:1 w/w of FN to amine-reactive Alexa Fluor

555 succinimidyl ester was used in the reaction. Cells were seeded on FN-

patterned substrates at 235 cells/mm2 in DMEM supplemented with 10% newborn

calf serum and cultured overnight (.16 h) for all experiments.

Bound integrins were visualized using a cross-linking and extraction method

(Garcıa et al., 1999; Keselowsky and Garcıa, 2005). Cells patterned on substrates

were incubated in cold DTSSP cross-linker (1.0 mM in PBS) for 30 min to cross-

link integrins to their bound ligand. DTSSP was then quenched using 50 mM

Tris-buffered saline, followed by extraction of uncross-linked components of

the cell with 0.1% SDS supplemented with protease inhibitors (16.7 mg/ml

phenylmethylsulfonyl fluoride, 10 mg/ml leupeptin, 10 mg/ml aprotinin). Afterextraction, samples were fixed in cold paraformaldehyde (3.7% in PBS) for 5 min,

blocked in blocking buffer (5% goat serum in PBS) for 30 min, and incubated with

primary antibody (anti-a5-integrin, 5 mg/ml) diluted in blocking buffer for 1 h at

37 C. Primary antibodies were visualized using Alexa-Fluor-488-conjugated

secondary antibodies (anti-rabbit IgG, 10 mg/ml) diluted in blocking buffer for 1 h

at 37 C. Images were captured using a Nikon Eclipse E400 fluorescence microscope,

Nikon CFI Apo TIRF 606oil/1.49 NA objective and Image-Pro image acquisition

software (Media Cybernetics, Bethesda, MD, USA) at fixed exposures andillumination conditions. Frequency map images were produced by stacking and

averaging individual images for a given condition using Image Pro software. In order

to evaluate pad occupancy, individual images with positive integrin staining for the

center area (indicating a cell) were scored for the number of adhesive pads within an

adhesive cluster with positive integrin localization and normalized by the total

number of patterns counted to generate a cumulative distribution.

For talin studies, 48 h after transfection, cells were seeded onto nanopatterns

and cross-linking/extraction was performed 72 h after initial transfection. For

vinculin studies, eGFP–vinT12-, eGFP–WT-, or eGFP–VH-expressing cells were

seeded overnight on the nanopatterns followed by cross-linking extraction as

described earlier. Inhibition experiments were performed on cells cultured

overnight on nanopatterns and treated with Y-27632 (10 mM) and

(2)-blebbistatin (20 mM) for 30 min and 60 min prior to analysis, respectively.

Adhesive force measurements

Cell adhesion strength was measured using a spinning disk system (Garcıa et al.,

1998b; Gallant et al., 2005). Patterned coverslips with adherent cells cultured

overnight were spun in PBS + 2 mM dextrose for 5 min at a constant speed in acustom-built device. The applied shear stress (t) is given by the formula

t50.8r(rmv3)1/2, where r is the radial position, r and m are the fluid density and

viscosity, respectively, and v is the spinning speed. After spinning, cells were fixed

in 3.7% formaldehyde, permeabilized in 0.1% Triton X-100, stained with ethidium

homodimer-1 (Invitrogen), and counted at specific radial positions using a 106objective lens in a Nikon TE300 microscope equipped with a Ludl motorized stage,

Spot-RT camera, and Image Pro analysis system. Sixty-one fields (80–100 cells/field

before spinning) were analyzed and cell counts were normalized to the number ofcell counts at the center of the disk, where the applied force is zero. The fraction of

adherent cells (f) was then fit to a sigmoid curve f51/(1+exp[b(t2t50)]), where t50

is the shear stress for 50% detachment and b is the inflection slope. t50 represents the

mean adhesion strength for a population of cells.

Calculation of adhesive forces for force equilibrium model

Adhesion strength data (Fig. 4) was converted to adhesive forces using a simple

mechanical equilibrium model for a nearly spherical adherent cell (Gallant et al.,

2005; Gallant and Garcıa, 2007). Traction forces based on those derived by Geiger

were obtained by multiplying the nanopattern area by the stress constant (5.5 nN/

mm2) reported by this group (Balaban et al., 2001). Estimates for traction forces

based on Gardel were obtained from the range of forces corresponding to the FA

growth phase in figure 3 in Stricker et al. (Stricker et al., 2011).

Statistics

Non-linear regression analysis was performed using SigmaPlot (Systat Software,

San Jose, CA, USA). Results were analyzed using one-way ANOVA and Tukey’s

post-hoc test for pairwise comparisons unless otherwise stated. Statisticalcomparisons for pad occupancy were completed using Kruskal–Wallis

nonparametric tests followed by pair-wise comparisons adjusted by multiplying

the P-value by the number of comparisons. All statistical analysis was completed

using SYSTAT 11.0 (Systat Software).

AcknowledgementsWe thank D. Brown, G. Spinner, and N. D. Gallant for their supportwith the fabrication of templates. We thank K. L. Templeman forpreparation of plasmids.

FundingThis work was supported by the National Institutes of Health [grantnumber R01-GM065918 to A.J.G. and R01-GM41605 to S.W.C.].Deposited in PMC for release after 12 months.

Supplementary material available online at

http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.108035/-/DC1

ReferencesAlexandrova, A. Y., Arnold, K., Schaub, S., Vasiliev, J. M., Meister, J. J.,

Bershadsky, A. D. and Verkhovsky, A. B. (2008). Comparative dynamics ofretrograde actin flow and focal adhesions: formation of nascent adhesions triggerstransition from fast to slow flow. PLoS ONE 3, e3234.

Allingham, J. S., Smith, R. and Rayment, I. (2005). The structural basis of blebbistatininhibition and specificity for myosin II. Nat. Struct. Mol. Biol. 12, 378-379.

Amano, M., Chihara, K., Kimura, K., Fukata, Y., Nakamura, N., Matsuura, Y. andKaibuchi, K. (1997). Formation of actin stress fibers and focal adhesions enhancedby Rho-kinase. Science 275, 1308-1311.

Aratyn-Schaus, Y., Oakes, P. W. and Gardel, M. L. (2011). Dynamic and structuralsignatures of lamellar actomyosin force generation. Mol. Biol. Cell 22, 1330-1339.

Arnold, M., Cavalcanti-Adam, E. A., Glass, R., Blummel, J., Eck, W., Kantlehner,

M., Kessler, H. and Spatz, J. P. (2004). Activation of integrin function bynanopatterned adhesive interfaces. ChemPhysChem 5, 383-388.

Balaban, N. Q., Schwarz, U. S., Riveline, D., Goichberg, P., Tzur, G., Sabanay, I.,Mahalu, D., Safran, S., Bershadsky, A., Addadi, L. et al. (2001). Force and focaladhesion assembly: a close relationship studied using elastic micropatternedsubstrates. Nat. Cell Biol. 3, 466-472.

Beningo, K. A., Dembo, M., Kaverina, I., Small, J. V. and Wang, Y. L. (2001).Nascent focal adhesions are responsible for the generation of strong propulsive forcesin migrating fibroblasts. J. Cell Biol. 153, 881-888.

Bershadsky, A., Kozlov, M. and Geiger, B. (2006). Adhesion-mediated mechanosensitivity:a time to experiment, and a time to theorize. Curr. Opin. Cell Biol. 18, 472-481.

Bouaouina, M., Lad, Y. and Calderwood, D. A. (2008). The N-terminal domains oftalin cooperate with the phosphotyrosine binding-like domain to activate b1 and b3integrins. J. Biol. Chem. 283, 6118-6125.

Bunch, T. A. (2010). Integrin aIIbb3 activation in Chinese hamster ovary cells andplatelets increases clustering rather than affinity. J. Biol. Chem. 285, 1841-1849.

Burridge, K. and Chrzanowska-Wodnicka, M. (1996). Focal adhesions, contractility,and signaling. Annu. Rev. Cell Dev. Biol. 12, 463-519.

Byers, B. A., Pavlath, G. K., Murphy, T. J., Karsenty, G. and Garcıa, A. J. (2002).Cell-type-dependent up-regulation of in vitro mineralization after overexpression ofthe osteoblast-specific transcription factor Runx2/Cbfal. J. Bone Miner. Res. 17,1931-1944.

Cai, Y., Biais, N., Giannone, G., Tanase, M., Jiang, G., Hofman, J. M., Wiggins,

C. H., Silberzan, P., Buguin, A., Ladoux, B. et al. (2006). Nonmuscle myosin IIA-dependent force inhibits cell spreading and drives F-actin flow. Biophys. J. 91, 3907-3920.

Calderwood, D. A. and Ginsberg, M. H. (2003). Talin forges the links betweenintegrins and actin. Nat. Cell Biol. 5, 694-697.

Calderwood, D. A., Zent, R., Grant, R., Rees, D. J., Hynes, R. O. and Ginsberg,

M. H. (1999). The Talin head domain binds to integrin b subunit cytoplasmic tailsand regulates integrin activation. J. Biol. Chem. 274, 28071-28074.

Cavalcanti-Adam, E. A., Micoulet, A., Blummel, J., Auernheimer, J., Kessler, H.and Spatz, J. P. (2006). Lateral spacing of integrin ligands influences cell spreadingand focal adhesion assembly. Eur. J. Cell Biol. 85, 219-224.

Cavalcanti-Adam, E. A., Volberg, T., Micoulet, A., Kessler, H., Geiger, B. and

Spatz, J. P. (2007). Cell spreading and focal adhesion dynamics are regulated byspacing of integrin ligands. Biophys. J. 92, 2964-2974.

Chen, H., Cohen, D. M., Choudhury, D. M., Kioka, N. and Craig, S. W. (2005).Spatial distribution and functional significance of activated vinculin in living cells.J. Cell Biol. 169, 459-470.

Choi, C. K., Vicente-Manzanares, M., Zareno, J., Whitmore, L. A., Mogilner, A.

and Horwitz, A. R. (2008). Actin and a-actinin orchestrate the assembly andmaturation of nascent adhesions in a myosin II motor-independent manner. Nat. Cell

Biol. 10, 1039-1050.Chrzanowska-Wodnicka, M. and Burridge, K. (1996). Rho-stimulated contractility

drives the formation of stress fibers and focal adhesions. J. Cell Biol. 133, 1403-1415.Cohen, D. M., Chen, H., Johnson, R. P., Choudhury, B. and Craig, S. W. (2005).

Two distinct head-tail interfaces cooperate to suppress activation of vinculin by talin.J. Biol. Chem. 280, 17109-17117.

Cohen, D. M., Kutscher, B., Chen, H., Murphy, D. B. and Craig, S. W. (2006). Aconformational switch in vinculin drives formation and dynamics of a talin-vinculincomplex at focal adhesions. J. Biol. Chem. 281, 16006-16015.

Coussen, F., Choquet, D., Sheetz, M. P. and Erickson, H. P. (2002). Trimers of thefibronectin cell adhesion domain localize to actin filament bundles and undergorearward translocation. J. Cell Sci. 115, 2581-2590.

Coyer, S. R., Garcıa, A. J. and Delamarche, E. (2007). Facile preparation of complexprotein architectures with sub-100 nm resolution on surfaces. Angew. Chem. Int. Ed.

Engl. 46, 6837-6840.Coyer, S. R., Delamarche, E. and Garcıa, A. J. (2011). Protein tethering into

multiscale geometries by covalent subtractive printing. Adv. Mater. 23, 1550-1553.

Journal of Cell Science 125 (21)5122

Journ

alof

Cell

Scie

nce

Critchley, D. R. (2009). Biochemical and structural properties of the integrin-associatedcytoskeletal protein talin. Annu. Rev. Biophys. 38, 235-254.

Danen, E. H. and Sonnenberg, A. (2003). Integrins in regulation of tissue developmentand function. J. Pathol. 201, 632-641.

DeMali, K. A., Barlow, C. A. and Burridge, K. (2002). Recruitment of the Arp2/3complex to vinculin: coupling membrane protrusion to matrix adhesion. J. Cell Biol.

159, 881-891.Dumbauld, D. W., Shin, H., Gallant, N. D., Michael, K. E., Radhakrishna, H. and

Garcıa, A. J. (2010). Contractility modulates cell adhesion strengthening throughfocal adhesion kinase and assembly of vinculin-containing focal adhesions. J. Cell.

Physiol. 223, 746-756.Engler, A. J., Sen, S., Sweeney, H. L. and Discher, D. E. (2006). Matrix elasticity

directs stem cell lineage specification. Cell 126, 677-689.Galbraith, C. G., Yamada, K. M. and Sheetz, M. P. (2002). The relationship between

force and focal complex development. J. Cell Biol. 159, 695-705.Gallant, N. D. and Garcıa, A. J. (2007). Model of integrin-mediated cell adhesion

strengthening. J. Biomech. 40, 1301-1309.Gallant, N. D., Michael, K. E. and Garcıa, A. J. (2005). Cell adhesion strengthening:

contributions of adhesive area, integrin binding, and focal adhesion assembly. Mol.

Biol. Cell 16, 4329-4340.Garcıa, A. J., Takagi, J. and Boettiger, D. (1998a). Two-stage activation for a5b1

integrin binding to surface-adsorbed fibronectin. J. Biol. Chem. 273, 34710-34715.Garcıa, A. J., Huber, F. and Boettiger, D. (1998b). Force required to break a5b1

integrin-fibronectin bonds in intact adherent cells is sensitive to integrin activationstate. J. Biol. Chem. 273, 10988-10993.

Garcıa, A. J., Vega, M. D. and Boettiger, D. (1999). Modulation of cell proliferationand differentiation through substrate-dependent changes in fibronectin conformation.Mol. Biol. Cell 10, 785-798.

Gardel, M. L., Sabass, B., Ji, L., Danuser, G., Schwarz, U. S. and Waterman, C. M.

(2008). Traction stress in focal adhesions correlates biphasically with actin retrogradeflow speed. J. Cell Biol. 183, 999-1005.

Gardel, M. L., Schneider, I. C., Aratyn-Schaus, Y. and Waterman, C. M. (2010).Mechanical integration of actin and adhesion dynamics in cell migration. Annu. Rev.

Cell Dev. Biol. 26, 315-333.Geiger, B., Bershadsky, A., Pankov, R. and Yamada, K. M. (2001). Transmembrane

crosstalk between the extracellular matrix and the cytoskeleton. Nat. Rev. Mol. Cell

Biol. 2, 793-805.Geiger, B., Spatz, J. P. and Bershadsky, A. D. (2009). Environmental sensing through

focal adhesions. Nat. Rev. Mol. Cell Biol. 10, 21-33.Gersbach, C. A., Le Doux, J. M., Guldberg, R. E. and Garcıa, A. J. (2006). Inducible

regulation of Runx2-stimulated osteogenesis. Gene Ther. 13, 873-882.Goffin, J. M., Pittet, P., Csucs, G., Lussi, J. W., Meister, J. J. and Hinz, B. (2006).

Focal adhesion size controls tension-dependent recruitment of a-smooth muscle actinto stress fibers. J. Cell Biol. 172, 259-268.

Goult, B. T., Bouaouina, M., Elliott, P. R., Bate, N., Patel, B., Gingras, A. R.,

Grossmann, J. G., Roberts, G. C., Calderwood, D. A., Critchley, D. R. et al.

(2010). Structure of a double ubiquitin-like domain in the talin head: a role in integrinactivation. EMBO J. 29, 1069-1080.

Grashoff, C., Hoffman, B. D., Brenner, M. D., Zhou, R., Parsons, M., Yang, M. T.,McLean, M. A., Sligar, S. G., Chen, C. S., Ha, T. et al. (2010). Measuringmechanical tension across vinculin reveals regulation of focal adhesion dynamics.Nature 466, 263-266.

Griffin, M. A., Sen, S., Sweeney, H. L. and Discher, D. E. (2004). Adhesion-contractile balance in myocyte differentiation. J. Cell Sci. 117, 5855-5863.

Humphries, J. D., Wang, P., Streuli, C., Geiger, B., Humphries, M. J. and

Ballestrem, C. (2007). Vinculin controls focal adhesion formation by directinteractions with talin and actin. J. Cell Biol. 179, 1043-1057.

Hynes, R. O. (2002). Integrins: bidirectional, allosteric signaling machines. Cell 110,673-687.

Jiang, G., Giannone, G., Critchley, D. R., Fukumoto, E. and Sheetz, M. P. (2003).Two-piconewton slip bond between fibronectin and the cytoskeleton depends on talin.Nature 424, 334-337.

Kanchanawong, P., Shtengel, G., Pasapera, A. M., Ramko, E. B., Davidson, M. W.,

Hess, H. F. and Waterman, C. M. (2010). Nanoscale architecture of integrin-basedcell adhesions. Nature 468, 580-584.

Keselowsky, B. G. and Garcıa, A. J. (2005). Quantitative methods for analysis ofintegrin binding and focal adhesion formation on biomaterial surfaces. Biomaterials

26, 413-418.Koo, L. Y., Irvine, D. J., Mayes, A. M., Lauffenburger, D. A. and Griffith, L. G.

(2002). Co-regulation of cell adhesion by nanoscale RGD organization andmechanical stimulus. J. Cell Sci. 115, 1423-1433.

Kuo, J. C., Han, X., Hsiao, C. T., Yates, J. R., 3rd and Waterman, C. M. (2011).Analysis of the myosin-II-responsive focal adhesion proteome reveals a role for b-Pixin negative regulation of focal adhesion maturation. Nat. Cell Biol. 13, 383-393.

Lehnert, D., Wehrle-Haller, B., David, C., Weiland, U., Ballestrem, C., Imhof, B. A.

and Bastmeyer, M. (2004). Cell behaviour on micropatterned substrata: limits ofextracellular matrix geometry for spreading and adhesion. J. Cell Sci. 117, 41-52.

Maheshwari, G., Brown, G., Lauffenburger, D. A., Wells, A. and Griffith, L. G.(2000). Cell adhesion and motility depend on nanoscale RGD clustering. J. Cell Sci.

113, 1677-1686.

Massia, S. P. and Hubbell, J. A. (1991). An RGD spacing of 440 nm is sufficient for

integrin avb3-mediated fibroblast spreading and 140 nm for focal contact and stress

fiber formation. J. Cell Biol. 114, 1089-1100.

Michael, K. E., Dumbauld, D. W., Burns, K. L., Hanks, S. K. and Garcıa, A. J.

(2009). Focal adhesion kinase modulates cell adhesion strengthening via integrin

activation. Mol. Biol. Cell 20, 2508-2519.

Miyamoto, S., Akiyama, S. K. and Yamada, K. M. (1995a). Synergistic roles for

receptor occupancy and aggregation in integrin transmembrane function. Science 267,

883-885.

Miyamoto, S., Teramoto, H., Coso, O. A., Gutkind, J. S., Burbelo, P. D., Akiyama,

S. K. and Yamada, K. M. (1995b). Integrin function: molecular hierarchies of

cytoskeletal and signaling molecules. J. Cell Biol. 131, 791-805.

Nelson, C. M., Jean, R. P., Tan, J. L., Liu, W. F., Sniadecki, N. J., Spector, A. A. and

Chen, C. S. (2005). Emergent patterns of growth controlled by multicellular form and

mechanics. Proc. Natl. Acad. Sci. USA 102, 11594-11599.

Nelson, C. M., Vanduijn, M. M., Inman, J. L., Fletcher, D. A. and Bissell, M. J.

(2006). Tissue geometry determines sites of mammary branching morphogenesis in

organotypic cultures. Science 314, 298-300.

Parsons, J. T., Horwitz, A. R. and Schwartz, M. A. (2010). Cell adhesion: integrating

cytoskeletal dynamics and cellular tension. Nat. Rev. Mol. Cell Biol. 11, 633-643.

Petrie, T. A., Raynor, J. E., Dumbauld, D. W., Lee, T. T., Jagtap, S., Templeman,

K. L., Collard, D. M. and Garcıa, A. J. (2010). Multivalent integrin-specific ligands

enhance tissue healing and biomaterial integration. Sci. Transl. Med. 2, 45ra60.

Riveline, D., Zamir, E., Balaban, N. Q., Schwarz, U. S., Ishizaki, T., Narumiya, S.,

Kam, Z., Geiger, B. and Bershadsky, A. D. (2001). Focal contacts as

mechanosensors: externally applied local mechanical force induces growth of focal

contacts by an mDia1-dependent and ROCK-independent mechanism. J. Cell Biol.

153, 1175-1186.

Roca-Cusachs, P., Gauthier, N. C., Del Rio, A. and Sheetz, M. P. (2009). Clustering

of a5b1 integrins determines adhesion strength whereas avb3 and talin enable

mechanotransduction. Proc. Natl. Acad. Sci. USA 106, 16245-16250.

Rossier, O. M., Gauthier, N., Biais, N., Vonnegut, W., Fardin, M. A., Avigan, P.,

Heller, E. R., Mathur, A., Ghassemi, S., Koeckert, M. S. et al. (2010). Force

generated by actomyosin contraction builds bridges between adhesive contacts.

EMBO J. 29, 1055-1068.

Selhuber-Unkel, C., Erdmann, T., Lopez-Garcıa, M., Kessler, H., Schwarz, U. S.

and Spatz, J. P. (2010). Cell adhesion strength is controlled by intermolecular

spacing of adhesion receptors. Biophys. J. 98, 543-551.

Shattil, S. J., Kim, C. and Ginsberg, M. H. (2010). The final steps of integrin

activation: the end game. Nat. Rev. Mol. Cell Biol. 11, 288-300.

Sniadecki, N. J., Desai, R. A., Ruiz, S. A. and Chen, C. S. (2006). Nanotechnology for

cell-substrate interactions. Ann. Biomed. Eng. 34, 59-74.

Stricker, J., Aratyn-Schaus, Y., Oakes, P. W. and Gardel, M. L. (2011).

Spatiotemporal constraints on the force-dependent growth of focal adhesions.

Biophys. J. 100, 2883-2893.

Tadokoro, S., Shattil, S. J., Eto, K., Tai, V., Liddington, R. C., de Pereda, J. M.,

Ginsberg, M. H. and Calderwood, D. A. (2003). Talin binding to integrin b tails: a

final common step in integrin activation. Science 302, 103-106.

Tan, J. L., Tien, J., Pirone, D. M., Gray, D. S., Bhadriraju, K. and Chen, C. S.

(2003). Cells lying on a bed of microneedles: an approach to isolate mechanical force.

Proc. Natl. Acad. Sci. USA 100, 1484-1489.

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

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

Wang, H. B., Dembo, M., Hanks, S. K. and Wang, Y. Y. (2001). Focal adhesion

kinase is involved in mechanosensing during fibroblast migration. Proc. Natl. Acad.

Sci. USA 98, 11295-11300.

Wiseman, P. W., Brown, C. M., Webb, D. J., Hebert, B., Johnson, N. L., Squier,

J. A., Ellisman, M. H. and Horwitz, A. F. (2004). Spatial mapping of integrin

interactions and dynamics during cell migration by image correlation microscopy. J.

Cell Sci. 117, 5521-5534.

Wozniak, M. A. and Chen, C. S. (2009). Mechanotransduction in development: a

growing role for contractility. Nat. Rev. Mol. Cell Biol. 10, 34-43.

Xu, W., Baribault, H. and Adamson, E. D. (1998). Vinculin knockout results in heart

and brain defects during embryonic development. Development 125, 327-337.

Yeung, T., Georges, P. C., Flanagan, L. A., Marg, B., Ortiz, M., Funaki, M., Zahir,

N., Ming, W., Weaver, V. and Janmey, P. A. (2005). Effects of substrate stiffness

on cell morphology, cytoskeletal structure, and adhesion. Cell Motil. Cytoskeleton 60,

24-34.

Zamir, E. and Geiger, B. (2001). Molecular complexity and dynamics of cell-matrix

adhesions. J. Cell Sci. 114, 3583-3590.

Zhang, X., Jiang, G., Cai, Y., Monkley, S. J., Critchley, D. R. and Sheetz, M. P.

(2008). Talin depletion reveals independence of initial cell spreading from integrin

activation and traction. Nat. Cell Biol. 10, 1062-1068.

Zimmermann, H., Katsen, A. D., Ihmig, F. R., Durst, C. H., Shirley, S. G. and Fuhr,

G. R. (2004). First steps of an interdisciplinary approach towards miniaturised

cryopreservation for cellular nanobiotechnology. IEE Proc. Nanobiotechnol. 151,

134-138.

ECM area limit for adhesion assembly and force 5123