Current Biology 20, 1145–1153, July 13, 2010 ª2010 Elsevier Ltd All rights reservedDOI 10.1016/j.cub.2010.05.049
Transient Frictional Slip
between Integrin and the ECM
in Focal Adhesions under Myosin II Tension
Yvonne Aratyn-Schaus1and Margaret L. Gardel1,2,3,*
1Institute for Biophysical Dynamics
2James Franck Institute
3Department of Physics
University of Chicago, Chicago, IL 60637
mechanosensation. Although adhesion assembly depends on
intracellular and extracellular tension, the biophysical regula-
tion of force transmission between the actin cytoskeleton
and extracellular matrix during this process remains largely
Results: To elucidate the nature of force transmission as
myosin II tension is applied to focal adhesions, we correlated
the dynamics of focal adhesion proteins and the actin cyto-
skeleton to local traction stresses. Under low extracellular
tension, newly formed adhesions near the cell periphery
underwent a transient retrograde displacement preceding
elongation. We found that myosin II-generated tension drives
this mobility, and we determine the interface of differential
motion, or ‘‘slip,’’ to be between integrin and the ECM. The
magnitude and duration of both adhesion slip and associated
F-actin dynamics is strongly modulated by ECM compliance.
Traction forces are generated throughout the slip period, and
adhesion immobilization occurs at a constant tension.
Conclusions: We have identified a tension-dependent, extra-
cellular ‘‘clutch’’ between integrins and the extracellular
matrix; this clutch stabilizes adhesions under myosin II
driven tension. The current work elucidates a mechanism by
which force transmission is modulated during focal adhesion
Focal adhesions (FAs) form hierarchical connections between
the F-actin cytoskeleton and the extracellular matrix (ECM) to
transmit mechanical forces across the plasma membrane.
Force generated within the F-actin cytoskeleton and trans-
mitted at FAs exert traction on the ECM and are important in
cell migration and ECM remodeling [1, 2]. In turn, the mechan-
ical properties of the ECM are sensed by adherent cells and
directly affect the morphology of FAs and the F-actin cytoskel-
eton . Cellular force sensing is thought to be dominated
through the regulation of FA assembly and growth by both
intracellular and extracellular forces [4–7]. Thus, an intricate
feedback exists between the F-actin cytoskeleton, ECM
mechanics, traction force generation, and FA assembly.
FA assembly is thought to modulate force transmission by
regulating the coupling of F-actin motion to the underlying
ECM. Near the cell periphery, F-actin polymerization drives
a rapid retrograde flow of a branched, dendritic network,
termed the lamellipodium. Proximal to the lamellipodium, the
F-actin cytoskeleton transitions into a contractile organelle,
termed the lamella, where retrograde flow is mediated by
myosin II. In the absence of FA formation, lamellar retrograde
flow is quite rapid and uniform throughout the cell body [8, 9].
During FA assembly, F-actin retrograde flow slows , and
traction stress builds on the ECM . This process is consis-
tent with the idea that FAs serve as a molecular ‘‘clutch’’
between dynamic F-actin and the immobile ECM [11–13].
Identification of molecular interactions that regulate the
molecular clutch during FA assembly is critical to further
understanding of cellular force transduction. Generally, clutch
regulation could occur intracellularly, via modulation of FA
proteins that link F-actin to integrins, and/or extracellularly,
via modulation of integrin-ECM binding. During FA initiation,
talin plays an important role in regulating this intracellular
linkage, and talin-deficient cell lines have impaired integrin-
ligand binding, enhanced retrograde F-actin flow, and
reduced traction stress [9, 14]. Likewise, in elongated FAs,
points of disconnect between F-actin and integrins are a-acti-
nin and vinculin [15, 16]. The extent to which these proteins
modulate intracellular clutch engagement as tension is built
during FA assembly and growth is unknown. Furthermore,
although extracellular clutch regulation between the integrin
and the ECM has been suggested , no direct observations
have been made. Thus, the components of the molecular
clutch during tension-dependent adhesion assembly remain
In this study, we identify a tension-dependent clutch at the
integrin-ECM interface on physiologically flexible substrates.
At low levels of tension, FAs undergo rapid, micron-scale
traction force; like others, we term this behavior ‘‘frictional
slip’’ . When the ECM stiffness decreases, the magnitude
and duration of frictional slip is enhanced but adhesions
immobilize at similar levels of tension. Once this extracellular
clutch is engaged, F-actin moves relative to FA components.
Thus, actomyosin retrograde flow and ECM mechanics are
coordinated toregulate an extracellular clutchduring theinitial
stages of tension-dependent stabilization of FAs.
ECM Compliance Regulates the Mobility of Nascent FAs
assembly in human osteosarcoma (U2OS) cells are similar to
copy to image the assembly of GFP-paxillin-rich adhesions
near the periphery of cells plated on fibronectin-coated glass
coverslips. We observed the appearance of newly formed,
or nascent, adhesions near the leading cell edge as small
(w0.5 mm diameter) fluorescent spots (Figure 1, upper panel).
These nascent adhesions appeared to be stationary with
respect to the underlying coverslip and, over the course of
15–20 min, elongated rearward into larger plaques with a
length of 3–5 mm. These morphological events are in accor-
dance with published studies of FA maturation on rigid, glass
substrates in numerous cell types [2, 17, 18].
To explore how ECM stiffness impacts FA assembly, we
visualized FA dynamics in U2OS cells plated on fibronectin-
coated polyacrylamide gels with a shear elastic modulus of
2.8 kPa, which resembles the stiffness of physiological tissue
. On compliant gels, cells spread and formed elongated
FA plaques similar in size to those found in cells plated on
fibronectin-coated glass coverslips; FA maturation consisted
of the appearance of small (w0.5 mm) fluorescent spots near
the cell periphery and subsequent elongation over a period
of 15–20 min. However, on compliant gels, nascent adhesions
underwent a rapid, retrograde displacement from their origin
of appearance prior to immobilization to the ECM (Figure 1,
lower panel). This retrograde movement was only seen in
nascent adhesions within 30–90 s after appearance, and it
occurred in a majority of the nascent adhesions observed
(65%, nFA = 39, ncell = 6). The magnitude of retrograde
displacement was 0.67 6 0.33 mm (nFA= 39, ncell= 6) and,
once immobilized, a majority (77%, nFA= 39, ncell= 6) of adhe-
sions elongated proximally. Thus, on compliant matrices,
a rapid, transient rearward translocation of newly formed
adhesion puncta usually precedes growth to an elongated
Blebbistatin Treatment and Removal as a Methodology
for Studying Myosin II-Driven FA Maturation
The retrograde movement of nascent adhesions observed on
compliant matrices suggests a probable role for myosin II
contractility. To investigate this, we designed an experimental
approach to isolate and synchronize myosin II-driven adhe-
sion maturation by perfusion and removal of the myosin II
ATPase inhibitor blebbistatin . In combination with traction
force and high-resolution confocal microscopy of GFP-actin
and mApple-paxillin (Experimental Procedures), this enabled
a characterization of both cytoskeletal dynamics and force
transmission at adhesions during the first steps of myosin
Under control conditions, condensed actin bundles ranging
from 5 to 70 mm in length terminated in elongated adhesions
with an average length of 3 mm (Figure 2; Figures S1 and
S3C); the average traction stress exerted at peripheral FAs
was 150 Pa (Figure 2; Figure S3B). After treatment with
25–50 mM of blebbistatin for 30 min, only small paxillin- and
vinculin-rich puncta that resembled nascent adhesions
remained (Figure 2; Figures S1–S3). Under blebbistatin
treatment, actomyosin bundles disassembled, and dense
‘‘patches’’ of actin colocalized withadhesions atthe lamellipo-
dial-lamellar border (LLB; Figure 2; Figures S1–S3). Immuno-
stained samples revealed an F-actin meshwork with a dense
ribbon of F-actin at the LLB and myosin II puncta, presumably
minifilaments , across the lamellar region (Figure S1). Trac-
tion stresses were constrained to the lamellipodial base and
were reduced to 20–30 Pa (Figure 2; Figure S3).
0:00 5:300:30 27:450:00
Figure 1. FA Maturation on Rigid and Compliant Substrata
(Upper panel) On glass, a GFP-paxillin punctum appears near the cell edge,
remains immobilized, and elongates within minutes. (Lower panel) On
a compliant substrate (2.8 kPa gel), a GFP-paxillin punctum appears
near the cell edge and undergoes retrograde displacement prior to
immobilization and elongation. Dashed lines highlight initial location of the
distal edge of FA puncta. Arrowheads indicate proximal end of FA. A solid
black line indicates the cell boundary. Color-combined image of FA at
Figure 2. Cytoskeletal Remodeling and Traction-
Stress Recovery upon Myosin Reactivation by
(middle), and traction-stress magnitude at times
prior to blebbistatin treatment (control) and at
times indicated after blebbistatin removal. Time
is indicated in min:s. A gray line indicates the
cell boundary. The scale bar represents 5 mm.
of GFP-actin(top), mApple-paxillin
Current Biology Vol 20 No 13
We synchronized myosin II-driven adhesion maturation
across the entire cell by deactivating blebbistatin by means
of drug removal and imaging with 488 nm light  (Movie S1
and Movie S2). Within 15 s, the traction stress at paxillin-rich
adhesions increased to 57 6 13 Pa, and a marginal increase
in the average adhesion length was resolved (Figure S3).
Two minutes after blebbistatin removal, the traction stress
exerted at individual adhesions had increased by approxi-
mately 4-fold, adhesions had elongated by roughly twice their
myosin II-independent length, and short actin bundles were
resolved (Figure 2; Figure S3). Fifteen minutes after blebbista-
tin removal, adhesions had elongated to 80% of their control
length, condensed myosin-rich actin bundles were observed
to span the cell body, and traction stresses had recovered to
the same order of magnitude as pre-blebbistatin levels (Fig-
ures 2; Figures S1 and S3). Thus, blebbistatin treatment and
removal has a reversible effect on cytoskeletal organization
and can be used for studying the process of myosin II-driven
maturation of FAs [6, 22].
F-Actin and FA Dynamics Reveal Frictional Slip
at the FA-ECM Interface
To examine the mobility of adhesions upon blebbistatin
removal, we compared pairs of mApple-paxillin images in an
aligned time-lapse sequence (Figure 3A; Movie S3). Similar
to newly formed adhesions in drug-free conditions, adhesion
puncta moved retrogradely upon removal of blebbistatin.
0 10203040 5060
Traction stress (Pa)
020406080 100 120
0”, 7”7”,13” 29”,45”
Figure 3. F-Actin and FAs Move Retrogradely as
Myosin II Is Reactivated
(A) (Top) Color-combined images of mApple-
paxillin at successive times after blebbistatin
removal, early (red) and late (green) times. Insets
show magnification of outlined regions. The scale
bar represents 5 mm. (Middle) GFP-actin images
overlaid with actin flow vectors (red arrows).
Time indicates the initial time point used in the
velocity calculation. A red arrow indicates trajec-
tory scale = 4.5 mm/min. A dashed white line
indicates a typical linescan across cell front
used in analysis; the location of adhesion and
actin patches (yellow square with black outline)
and lamellar actin (light blue square) is shown.
(Bottom) Heat-scale map of traction stress with
the scale bar indicated at right (from blue to
red). Times are indicated in seconds.
(B) Traction stress (violet squares), F-actin flow
(green circles), paxillin speed (red), and substrate
speed (black squares) over time after blebbistatin
removal. F-actin flow was tracked at actin
patches associated with FA. Data reflect mean
plus standard error for nFA= 26 and ncell= 5.
(C) Displacement of actin patches colocalized
with adhesions (yellow square, black outline)
and at 1.2 mm proximal to adhesion puncta within
the lamella (blue square). Approximate locations
of these data points are indicated in (A), middle
panel. The horizontal dashed line is a line of zero
slope. Data plotted in figures reflect the mean
plus standard error about the mean.
Comparing images obtained immedi-
ately (0 s) and 7 s after blebbistatin
lin puncta were displaced toward the
cell center, whereas the location of the
cell edge had not changed significantly
(indicated by line traces in Figure 3A).
Between 7 and 13 s after removal, retrograde movement of
adhesions was substantially reduced. At later times, the distal
tip of adhesions did not translocate (Figure 3A).
To quantify this motion, we identified the centroids of the
mApple-paxillin puncta and calculated rates of displacement.
Line scans across individual adhesions during frictional slip
revealed a Gaussian intensity profile with a full-width half
maximum of w300 nm and a significant signal-to-noise ratio
(Figure S4), indicating that these features were nearly diffrac-
tion limited and amenable to computational tracking algo-
rithms that track the centroid with sub-pixel resolution and
provide a resolution of mobility on the order of 0.02 mm/min
[23, 24]. Initially, between 0 and 7 s, the adhesion moved retro-
gradely with a speed of 2.7 mm/min (Figure 3B). Within 15 s of
blebbistatin removal, the apparent movement of paxillin-rich
puncta subsided to values near our resolution limit. Because
cells were adhered to compliant substrates, it is possible
that some retrograde movement could be accommodated by
substrate deformation. However, fiducial markers in the top
surface of the substrate underlying adhesion puncta moved
less than 1 mm/min, and there was a rapid decay thereafter
(Figure 3B); movement of fibronectin that was adhered to the
top surface of the substrate moved at rates similar to that of
the substrate (Figure S4). Thus, by 15 s after blebbistatin
gated with respect to the underlying fibronectin-coated
substrate and defines FA immobilization.
Transient Frictional Slip during Adhesion Assembly
To observe F-actin dynamics during blebbistatin removal,
we monitored fiducial marks of GFP-actin over time. Immedi-
ately after blebbistatin removal, actin patches colocalized
with adhesion puncta at the LLB (yellow square, Figure 3A)
and underwent a rapid, coherent retrograde movement,
bistatin removal (Figures 3A and 3B; Figure S5A). Thus, actin
patches stabilized w0.5 mm proximal from their starting loca-
tion (Figure 3C). Lamellar actin, located approximately 1.2 mm
proximal to these patches (blue square, Figure 3A), exhibited
similar, but slightly enhanced, retrograde flow dynamics and
stabilized at roughly 30 s to a steady-state retrograde flow of
0.25 mm/min, the value observed in compact stress fibers 
(Figure S5A). Within 15 s after blebbistatin removal, lamellar
F-actin moved retrogradely w1 mm and continued to move
toward the cell center at later times (Figure 3C). Thus, the
sions and fiduciary marks in lamellar actin increased under
The traction stress exerted on the extracellular matrix
increased monotonically from about 20 Pa to 50 Pa as
mApple-paxillin underwent retrograde movement (Figures 3A
flow rate, such that traction stress increased while F-actin
retrograde flow speed decayed (Figure S5B); this relationship
is consistent with known mechanochemistry of individual
myosin II motors (Supplemental text; Figure S5). Thus, even
while FAs move relative to the underlying matrix, mechanical
forces are transmitted to the extracellular matrix. We refer to
this transient state as myosin II-driven ‘‘frictional slip’’ of adhe-
sion puncta, a phenomenon previously observed in filopodial
Frictional Slip in Nascent Adhesions Occurs between
Integrin and the ECM
Because FAs are multicomponent ensembles that contain a
diverse group of actin and ECM binding proteins [26, 27], we
occur at different components during frictional slip. We
investigated two classes of integrins, avand a5, known to bind
fibronectin, as well as two FA proteins, talin [9, 14] and vinculin
[15, 16, 28], which establish a molecular link between integrin
and actin. We found that all of these proteins underwent retro-
grade motion comparable to that of paxillin within the first 20 s
after blebbistatin removal and were immobilized thereafter
(Figure 4A, Movie S4 and Movie S5). The retrograde dynamics
of different FA proteins were highly correlated, as measured
by Pearson’s correlation coefficient (0.98), and all proteins
became immobilized to the substrate after approximately
15 s (Figure 4B). Furthermore, the total displacement of each
protein population prior to substrate engagement was similar;
the average displacement was 0.5 mm, approximately 4-fold
greaterthan thatoftheunderlying substrate (Figure4C).These
data indicate that a diverse set of FA-associated proteins,
ranging from associated actin to integrin, displace as a collec-
tive unit prior to their immobilization and that a predominant
slip occurs at the interface between integrin and the ECM.
FRAP Reveals Stable Association of avIntegrin
to Adhesion Puncta during Slip
Micron-sized movements of integrin relative to the underlying
substratum occur simultaneously with increases in traction.
Because these movements are too large to be accommodated
by individual bond deformation or stretching, we hypothesize
this motion could result from either polarized remodeling 
or the nature of binding kinetics between the integrin and
from the distal tip is balanced by incorporation of new integrin
at the proximal end to result in apparent motion . In this
scenario, we would expect nearly complete exchange of
integrins with the diffuse population during frictional slip, as
has been observed in polarized remodeling of mature FAs
We utilized fluorescence recovery after photobleaching
(FRAP) to determine whether integrins are stably associated
within the adhesion puncta during frictional slip. We bleached
GFP-a5and GFP-avupon blebbistatin removal and measured
Displacement ( m)
Speed ( m/min)
Figure 4. FA Displacement Occurs at the Integ-
(A) Color-combined images of FA proteins upon
blebbistatin inactivation: (top) 0 s (red) and 40 s
(green) and (bottom) 20 s (red) and 40 s (green).
The scale bar represents 1 mm. Arrowheads indi-
cate adhesion puncta.
(B) Speeds of FA proteins and beads in substrate
over time after blebbistatin inactivation.
(C) Total displacement of adhesion proteins and
substrate prior to adhesion immobilization. There
is no statistically significant difference between
FA proteins; substrate displacement is statisti-
cally significantly different from FAs (p < 0.0001).
Data reflect the mean and standard error for
all samples. Sample sizes are as follows: a5
(nFA= 8; ncell= 6), av(nFA= 9; ncell= 4), talin
(nFA= 6; ncell= 3), paxillin (nFA= 26; ncell= 16),
and vinculin (nFA= 12; ncell= 3).
Data plotted in the figure reflect the mean plus
standard error about the mean.
Current Biology Vol 20 No 13
changes in fluorescence intensity. To ensure that FRAP
measurements of GFP-tagged proteins were performed on
adhesions undergoing frictional slip, we confirmed adhesion
retrograde movement by visualization with mApple-paxillin
and traction-stress increase (data not shown). Both integrins
underwent marginal fluorescence recovery (w20%–25%)
within the first 20 s after blebbistatin removal (Figures 5A
and 5B). At longer times, a5 underwent partial exchange,
whereas avunderwent negligible turnover.
As a control, we photobleached GFP-talin and GFP-paxillin.
For talin, complete recovery was observed over long time
scales, on the order of 2 min. In stark contrast, paxillin under-
went rapid turnover, which was nearly complete by the end of
the slip period (Figures 5A and 5B). Interestingly, the degree of
dynamic exchange observed for integrin, talin, and paxillin
was similar to that measured in large FAs in U2OS cells under
no drug treatment (data not shown) and in other cell types .
These results show that integrins remain stably associated to
adhesion puncta during frictional slip and do not support
polarized remodeling as the predominate mechanism of adhe-
Force per Available Integrin Increases during Frictional
Our FRAP data indicate stable association of integrin within
the adhesion during retrograde slip, suggesting that motion
is probably accommodated by a population of integrins
that undergo cycles of bond association and dissociation to
the ECM while facilitating force transmission [11, 30]. If the
apparent adhesion mobility reflects a time scale determined
by competing effects of bond association and dissociation,
decreased adhesion mobility would be associated with
enhanced bond association and diminished bond dissociation
for a constant population. Enhanced bond association might
result from factors such as changes in integrin activation, a
conformational change of integrin required for integrin-ECM
binding. However, our data show that enhanced integrin acti-
vationvia Mn2+treatment[33, 34]did notperturb the dynamics
of frictional slip (Figure S6).
Alternatively, if bond dissociation is enhanced under
increased force, then increasing the total number of integrins,
while maintaining a constant binding kinetics would provide
a lower force per integrin and, thus, decrease the rate of integ-
rin unbinding . To probe the changes in the total number of
integrins contained within the adhesion during frictional slip,
we measured the total GFP-avor GFP-a5integrin fluorescence
intensity within FA. During the slip period (t < 15 s), GFP-av
intensity increased by approximately 1.5-fold and attained
a steady state shortly thereafter (Figures 5C and 5D). By
contrast, GFP-a5 intensity did not undergo appreciable
increase during slip or after adhesions engaged the substrate.
Because the traction stress increased 3-fold during this time,
Figure 5. Minimal Exchange of Integrin within FA during Frictional Slip
(A) Vertical montage of images of GFP-tagged av, a5, talin, and paxillin. Prebleached adhesions are shown at t = 0 s (top, white solid arrows), and photo-
bleaching occurred between 0 and 5 s. Post-bleach images are shown for av, a5, talin, and paxillin at the times indicated at left. Times are indicated in
seconds. Reappearance of adhesions after bleach in a5, talin, and paxillin images is indicated by outlined arrows.
(B) Recovery profiles for FA proteins in (A) over time. A dashed line demarcates the termination of the FA slip (t w 15 s). Data are shown for 5 FA per protein.
(C) Intensity profiles of GFP-avfrom line scans subtending an individual adhesion at t = 0 s (red), t = 5 s (green), and t = 15 s (blue). Color-combined image of
GFP-avat times indicated in plot. Dashed lines indicate FA width. The scale bar represents 0.5 mm. LP indicates the lamellipodium, and CB marks the cell
(D) Traction stress and integrin fluorescence were normalizedto data at t = 0 s and plotted over time. Data reflect average values taken from 18 FAs from five
(E) Ratios of the relative increase in traction stress to the relative change in integrin intensity at the time of adhesion immobilization.
Data plotted in the figure reflect the mean plus standard error about the mean.
Transient Frictional Slip during Adhesion Assembly
the average tension per available integrin increased during
frictional slip by approximately 2- to 3-fold for both avand a5
integrins (Figure 5E). These data suggest that changes in
adhesion strength during frictional slip are not mediated
entirely by increases in the total number of integrins available
for ligand binding.
Frictional Slip of FAs Is Modulated by the Elastic Stiffness
of the ECM
substrates during the process of FA maturation suggests that
frictional slip may be intimately coupled to ECM stiffness. We
sought to determine how changing ECM stiffness modifies the
frictional slip of small adhesions near the cell periphery by
plating cells on matrices with varied elastic stiffness and glass
coverslips. Consistent with previous reports, cells plated on
soft (0.6 kPa) gels displayed smaller adhesions and fewer
compact F-actin bundles than cells plated on stiffer (2.8 kPa
and glass) substrates (Figure S7A).
In the presence of blebbistatin, the differences in cytoskel-
were largely eliminated. Cells plated on soft matrices con-
tained an isotropic network of F-actin throughout the lamella
and had a band of small (0.5 mm) adhesions near the LLB
(Figure S7B and Movie S6). For matrices with a stiffness less
than 2.8 kPa, retrograde frictional slip of adhesion puncta
was also observed upon blebbistatin removal (Figure 6A;
Figures S7B and S7C; Movie S6 and Movie S7). By contrast,
retrograde displacement of adhesions was not resolved on
glass (w106kPa) coverslips (Figure 6A; Figure S7D). Perturb-
ing the fibronectin concentration or changing the ECM ligand
to collagen did not alter the magnitude or duration of retro-
grade displacement of adhesions (Figure S6). Thus, ECM stiff-
ness was the predominant regulator of frictional slip. As the
ECM stiffness decreased, the total retrograde displacement
of the adhesion increased from <200 nm to 0.8 mm, and the
time before adhesion immobilization increased from 0 to 25 s
(Figure 6A). Interestingly, during frictional slip, the FAs of cells
in different mechanical environments are indistinguishable in
size; elongation of FA plaques occurs only after engagement
to the ECM.
The initial lamellar F-actin flow rate upon release from bleb-
bistatin also depended on the stiffness of the underlying ECM;
it increased from 1.5 mm/min on glass to nearly 6 mm/min on
0.6 kPa gels (Figure 6B). Furthermore, the extent of lamellar
F-actin retrograde displacement prior to adhesion engage-
ment also varied on matrices of different stiffness; it increased
from 0.8 mm on 2.8 kPa gels to 1.1 mm on 0.6 kPa gels (Fig-
the retrograde displacement of the nascent adhesion, which in
turn was larger than the substrate deformation (Figure 6C).
Collectively, these data indicate that the absolute retrograde
displacement of the F-actin adhesions and ECM during fric-
tional slip depends on the ECM stiffness.
Frictional Slip of Adhesion Is Abrogated at a Constant
Because the abrogation of adhesion frictional slip occurred
concomitantly with increased tension and deformation of the
extracellular matrix, we sought to determine whether the
immobilization of adhesions to the ECM under myosin II driven
tension was a stress (e.g., tension)- or strain (e.g., deforma-
tion)-limited process. To explore this, we examined the stress
and strain incurred in the substrate at the time of FA engage-
ment for cells plated on matrices with varied elastic stiffness.
Modifications to the substrate stiffness change the relation-
ship between the amount of substrate deformation and the
magnitude of tension stored within the substrate. As the
substrate stiffness decreases, the amount of deformation
required to attain a certain amount of tension increases pro-
portionally. We observed that the magnitude of substrate
deformation incurred at the time of FA engagement decreased
monotonically as its stiffness increased from 0.6 to 2.8 kPa
(Figure 6D). By contrast, the traction stress exerted on the
substrate at the time of adhesion engagement was similar for
all three substrate stiffnesses and amounted to an approxi-
mately 30 Pa increase over the traction stress exerted during
blebbistatin treatment (Figure 6D). This indicates that the
myosin II-driven transition between FA frictional slip and
immobilization to ECM occurs at a constant tension, rather
than a constant deformation.
We have identified a key biophysical transition that occurs as
myosin II tension is applied to adhesions. In the absence of
myosin II activity, polymerization-dependent nascent adhe-
sions form at the base of the lamellipodium (Figure 7A).
When myosin II tension is applied to the lamellar F-actin
network, adhesions undergo a transient, rapid, micron-scale
rearward translocation while exerting traction against the
underlying substrate in a process we term ‘‘frictional slip.’’
During frictional slip, traction stress builds and the retrograde
F-actin speed slows. Frictional slip is abrogated once a certain
Actin Flow (μm/min)
normalized to 2.8kPa
Total displacement (μm)
Immobilization Time (s)
Stiffness (kPa)Stiffness (kPa)
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Figure 6. FA Slip Modulated by ECM Stiffness
during adhesion slip (solid circles, right) plotted against substrate stiffness.
(B) Lamellar F-actin retrograde flow at the onset of blebbistatin removal is
plotted against substrate stiffness.
and substrate on 2.8 kPa and 0.6 kPa gels.
(D) Relative (normalized to the 2.8 kPa case) changes in substrate strain
(deformation) and traction stress (force) for different substrate stiffnesses.
Sample size distribution: glass (nFA= 8, ncell= 2), 2.8 kPa (nFA= 36, ncell= 5),
1.5 kPa (nFA= 8, ncell= 3), and 0.6 kPa (nFA= 8, ncell= 3). P values for data
compared to 0.6 kPa case; *p < 0.01 and **p < 0.001.
Data plotted in the figure reflect the mean plus standard error about the
Current Biology Vol 20 No 13
level of force is attained, thereby providing a stable adhesion
to further increase force and adhesion elongation.
Therigidity of the ECM regulates the spatiotemporal proper-
ties of adhesion frictional slip, such that adhesions displace
over longer times and greater distances on softer matrices.
With increasing substrate rigidity, the magnitude of rearward
translocation decreases such that, for matrices stiffer than
w10 kPa, we estimate that the magnitude of adhesion translo-
cation will not be resolved by light microscopy. This is consis-
tent with observations of FA assembly dynamics on rigid glass
coverslips, where nascent FAs seem to be immobilized imme-
diately upon their appearance (Figure 1) .
During the transient frictional slip, the integrins move with
grade motion of the F-actin is efficiently transmitted to the
integrin. After integrins stabilize relative to the ECM, lamellar
F-actin retrograde flow persists. This is consistent with
previous experiments, which have found movement between
actin and FA proteins within mature FAs [15, 16, 36, 37]. This
suggests that the initial steps of myosin II-applied tension to
FAs are characterized by a transition from a state dominated
by slip between integrins and the ECM to one dominated by
slip between intracellular FA proteins and F-actin (Figure 7B).
Wespeculate thatthelocationof slipisreflectiveofthe‘‘weak-
of such ‘‘weak links’’ change during adhesion maturation.
The micron-scale movement of adhesions occurs simulta-
neously with force exertion onto the ECM. This indicates that
bonds between the actin cytoskeleton and ECM exist, but
the scale of motion is much larger than could be accommo-
dated by deformation of individual molecular bonds. The
strong correlation observed between the motion dynamics of
numerous adhesion proteins and the FRAP data demon-
strating minimal integrin turnover both suggest a stable asso-
ciation of integrins to the FAs during frictional slip. This leads
us to speculate that binding kinetics between integrin and
the ECM are regulating frictional slip such that detachment
and reattachment at bonds between integrins and the ECM
would explain the simultaneous force transduction and retro-
grade movement. A population of such integrins, constituting
an adhesion, could allow some to detach, relieve stress,
and reattach while others temporarily transfer the force.
Quantitative models incorporating binding kinetics have
been successfully used for modeling the differential motion
observed between F-actin and FA proteins [11, 30] as well as
for modeling other types of adhesive interactions .
Increasing amounts of force occurs concomitantly with
decreased retrograde movement, indicating that the strength
of the integrin-ECM interface increases during frictional
slip. This strength can be estimated by the frictional drag
coefficient, sTR/[A(vFA 2 vS)], where sTR is the traction
stress exerted to the underlying matrix, vFA2 vSis the differ-
ence in flow speed between FA and the substrate (Figure 7C,
inset), and A is the approximate area of nascent adhesions
(Aw0.1 mm2). A calculation of the frictional-drag coefficient
from interpolation of our experimental data yields a frictional-
drag coefficient of the integrin-ECM interface that rapidly
rises nearly 100-fold as the applied traction increases from
30 to 70 Pa (Figure 7C). Because little to no accumulation of
new integrin is observed within adhesions during this time,
predominately through bond reorganization via integrin clus-
tering [39, 40] or through force-induced changes in individual
bond strength between the integrin and ECM [41, 42]. Future
experiments will be needed to dissect the underlying mecha-
nisms of this strengthening as well as to explore the impact
of different integrin-ECM interactions.
In summary, we have found that bonds between the integrin
and extracellular matrix function as an extracellular ‘‘clutch’’ to
modulate the degree of force transmission from the F-actin
cytoskeleton to the extracellular matrix in early-stage myosin
II-mediated FA maturation. The rapid strengthening of the
at the leading cell edge to be weakly adherent and mechani-
this initial stage of substrate sensing, the retrograde flow
dynamics of the actomyosin cytoskeleton play a prominent
role in rapidly building tension to stabilize integrin binding to
ECMs with varied mechanical compliance. However, under
sufficient tension generated by actomyosin contraction, this
10 20 30 40 50 60 70 80 90
Traction Stress (Pa)
Frictional Drag (Pa-s)
actin of unknown
Figure 7. Force-Dependent ‘‘Clutch’’ between
Integrin and the ECM
(A) Schematic diagram illustrating mobility of
cytoskeletal components at different levels of
force. Adhesion assembly: integrins are weakly
coupled to the ECM. Frictional slip: myosin II
drives fast retrograde velocity of F-actin and
associated integrins (v, open arrow); traction
force magnitude is low (small F, solid arrow).
Adhesion immobilization: at a critical tension,
integrin-ECM immobilization occurs, traction
force increases, myosin-dependent actin flow
slows, and FA elongation commences.
(B) Schematic diagram showing that interfacial
slip between integrin and the ECM (dashed line)
dominates in early adhesions under low force,
whereas slip at the actin-FA interface (solid line)
dominates and persists in large adhesions under
(C) Semi-log plot of frictional-drag coefficient
versus traction stress, determined by calculation
based on interpolated data points shown in
Figure S7C. Inset: A semi-log plot of the integ-
rin-ECM slip rate versus time after blebbistatin
removal was calculated from interpolated data
from Figure S7C.
Transient Frictional Slip during Adhesion Assembly
adhesion would stabilize to impede the retrograde motion of
the F-actin and immobilize FA plaques to mediate further FA
elongation and growth. This underscores the importance of
mechanical feedback between the F-actin cytoskeleton and
the ECM in building force-sensitive adhesions that control
cell motility and morphology.
Cell Culture and Transfection
Human osteosarcoma (U2OS) cells were obtained from American Type
Culture Collection (ATCC) and maintained in McCoy’s medium (HyClone),
supplemented with 10% fetal bovine serum (HyClone), penicillin, and strep-
tomycin (GIBCO). Transient transfections of GFP-actin (C. Waterman, NIH),
mApple-paxillin, GFP-vinculin (M. Davidson, U. of Florida), GFP-a5integrin
(R. Horwitz, U. of Virginia), GFP-avintegrin, and GFP-talin (K. Hu, U. of Indi-
ana) were performed with FuGENE6 Transfection Reagent (Roche) per the
Immunofluorescence of myosin light chain (monoclonal, Cell Signaling),
paxillin (polyclonal, Santa Cruz), vinculin (monoclonal, Sigma), fibronectin
(polyclonal, Sigma), and phalloidin (Molecular Probes) staining of F-actin
was performed as previously described .
Preparation of Polyacrylamide (PAA) Substrates for Traction Force
Fibronectin-coated PAA substrates containing 40 nm fluorescent spheres
were prepared on glass coverslips [10, 43] with varying acrylamide/bis-
acrylamide ratios so that gels with varying elastic moduli could be obtained:
7.5%/0.1% for 2.8 kPa, 7.5%/0.05% for 1.5kPa, and 5%/0.075% for 0.6 kPa,
as characterized previously . Fibronectin or collagen was covalently
attached to the top surface of the PAA gel via the bifunctional cross-linker
Live Cell Microscopy
Coverslips containing transfected cells bound to PAA substrates were
mounted in a perfusion chamber (Warner Instruments) in imaging media
consisting of media supplemented with 30 ml/1 ml Oxyrase (Oxyrase
Enzyme system, EC0050) and 10 mM HEPES (pH 7.5). Cells were imaged
at 37?C24–48 hrafter transfection on amultispectral spinning-diskconfocal
lens (Nikon) and a CSUX scanner (Yokogawa). A CCD camera (Coolsnap
HQ2, Photometrics) controlled with Metamorph acquisition software (MDS
Analytical Technologies) was used. FAs and beads were monitored at the
same confocal section; F-actin was imaged 0.2 mm into the cell interior.
Cells were treated with 25–50 mM blebbistatin (Sigma) for 30 min . For
blebbistatin inactivation, cells were washed with imaging media and visual-
ized with 488 nm light .
Particle Imaging Velocimetry code in MATLAB (mpiv, developed by N. Mori
and K-A Chang) was used for identifying, with sub-pixel accuracy, the
movement of beads embedded in polyacrylamide substrate. Fourier trans-
form traction cytometry methods were used for determining traction stress
from bead displacements . Traction stresses documented in the text
were calculated as an average of the maximum values across individual
Fiducial marks of fluorescent F-actin, centroids of FAs, and substrate-
embedded beads were tracked in separate image channels with computa-
tional tracking software developed by the lab of Gaudenz Danuser .
Manual tracking, via the point tracking function in Metamorph, of F-actin
patches associated with FA puncta was used as an alternative method.
Velocity vector fields were interpolated onto linescans subtending indi-
vidual FA puncta (perpendicular to the cell edge) via a Gaussian-weighted
interpolation, with a full-width half maximum of 10 pixels (1 mm). All vector
fields exhibited a high degree of directional coupling . Lamellar F-actin
flow vectors were measured 1.2 mm proximal to the FA centroids
FA engagement time was defined as the time at which the displacement
rate of FA puncta converged with that of beads within the PAA gel (error
of 0.02 mm/min).
Length changes in FAs and actin bundles were quantified in Metamorph.
Fluorescence intensity of GFP-a5and GFP-avwas background subtracted,
integrated across the FA area, and normalized to the first image plane.
Data plotted in figures reflect the mean plus standard error about the
Photobleaching experiments were performed with a 405 nm laser coupled
through a micromirror array to control the spatial location of illumination
(Mosaic, Photonics Instruments) with a 200 ms exposure time and a rectan-
gular region of approximately 5 mm2. The photobleaching event took place
after acquisition of the first image after blebbistatin removal, and post-
bleach images were recorded thereafter. Analysis was performed as in .
be found with this article online at doi:10.1016/j.cub.2010.05.049.
The authors wish to acknowledge use of computational-analysis algorithms
for analyzing cytoskeletal dynamics and traction forces developed in the
labs of Gaudenz Danuser and Ulrich Schwarz, respectively, as well as
generous gifts of cDNA constructs used in this paper: GFP-actin
(C. Waterman, NIH), mApple-paxillin (M. Davidson, U. of Florida), GFP-a5
integrin (R. Horwitz, U. of Virginia), and GFP-avintegrin and GFP-talin (Ke
Hu, U. of Indiana). Experiments visualizing fibronectin mobility during
adhesion slip were conducted by S.P. Winter. The authors would like to
thank Y. Beckham, P. Oakes, and T. Schaus for helpful comments and care-
ful reading of the manuscript. This work was supported by a Burroughs
Wellcome Career Award, a Keck Foundation Grant, and an NIH Director’s
Pioneer Award (DP10D00354) to M.L. Gardel.
Received: March 9, 2010
Revised: April 20, 2010
Accepted: May 14, 2010
Published online: June 10, 2010
1. Berrier, A.L., and Yamada, K.M. (2007). Cell-matrix adhesion. J. Cell.
Physiol. 213, 565–573.
2. Vicente-Manzanares, M., Choi, C.K., and Horwitz, A.R. (2009). Integrins
in cell migration—The actin connection. J. Cell Sci. 122, 199–206.
3. 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.
4. Schwartz, M.A., and DeSimone, D.W. (2008). Cell adhesion receptors in
mechanotransduction. Curr. Opin. Cell Biol. 20, 551–556.
5. Choquet, D., Felsenfeld, D.P., and Sheetz, M.P. (1997). Extracellular
matrix rigidity causes strengthening of integrin-cytoskeleton linkages.
Cell 88, 39–48.
6. 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.
7. Balaban, N.Q., Schwarz, U.S., Riveline, D., Goichberg, P., Tzur, G.,
Sabanay, I., Mahalu, D., Safran, S., Bershadsky, A., Addadi, L., and
Geiger, B. (2001). Force and focal adhesion assembly: A close relation-
ship studied using elastic micropatterned substrates. Nat. Cell Biol. 3,
8. Alexandrova, A.Y., Arnold, K., Schaub, S., Vasiliev, J.M., Meister, J.J.,
Bershadsky, A.D., and Verkhovsky, A.B. (2008). Comparative dynamics
of retrograde actin flow and focal adhesions: Formation of nascent
adhesions triggers transition from fast to slow flow. PLoS ONE 3,e3234.
9. 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,
10. 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 retrograde flow speed. J. Cell Biol. 183,
Current Biology Vol 20 No 13
11. Chan, C.E., and Odde, D.J. (2008). Traction dynamics of filopodia on Download full-text
compliant substrates. Science 322, 1687–1691.
12. Jurado, C., Haserick, J.R.,and Lee,J. (2005). Slipping orgripping? Fluo-
rescent speckle microscopy in fish keratocytes reveals two different
mechanisms for generating a retrograde flow of actin. Mol. Biol. Cell
13. Mitchison, T., and Kirschner, M. (1988). Cytoskeletal dynamics and
nerve growth. Neuron 1, 761–772.
14. Jiang, G., Giannone, G., Critchley, D.R., Fukumoto, E., and Sheetz, M.P.
(2003). Two-piconewton slip bond between fibronectin and the cyto-
skeleton depends on talin. Nature 424, 334–337.
15. Brown, C.M., Hebert, B., Kolin, D.L., Zareno, J., Whitmore, L., Horwitz,
A.R.,andWiseman, P.W.(2006).Probingthe integrin-actin linkage using
high-resolution protein velocity mapping. J. Cell Sci. 119, 5204–5214.
16. Hu, K., Ji, L., Applegate, K.T., Danuser, G., and Waterman-Storer, C.M.
(2007). Differential transmission of actin motion within focal adhesions.
Science 315, 111–115.
17. Webb, D.J., Donais, K., Whitmore, L.A., Thomas, S.M., Turner, C.E.,
Parsons, J.T., and Horwitz, A.F. (2004). FAK-Src signalling through pax-
illin, ERK and MLCK regulates adhesion disassembly. Nat. Cell Biol. 6,
18. Gupton, S.L., and Waterman-Storer, C.M. (2006). Spatiotemporal feed-
back between actomyosin and focal-adhesion systems optimizes rapid
cell migration. Cell 125, 1361–1374.
19. Straight, A.F., Cheung, A., Limouze, J., Chen, I., Westwood, N.J.,
Sellers, J.R., and Mitchison, T.J. (2003). Dissecting temporal and spatial
20. Verkhovsky, A.B., Svitkina, T.M., and Borisy, G.G. (1995). Myosin II fila-
ment assemblies in the active lamella of fibroblasts: Their morphogen-
esis and role in the formation of actin filament bundles. J. Cell Biol.
21. Sakamoto, T., Limouze, J., Combs, C.A., Straight, A.F., and Sellers, J.R.
(2005). Blebbistatin, a myosin II inhibitor, is photoinactivated by blue
light. Biochemistry 44, 584–588.
22. 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.
23. Danuser, G., and Waterman-Storer, C.M. (2006). Quantitative fluores-
cent speckle microscopy of cytoskeleton dynamics. Annu. Rev. Bio-
phys. Biomol. Struct. 35, 361–387.
24. Ji, L., and Danuser, G. (2005). Tracking quasi-stationary flow of weak
fluorescent signals by adaptive multi-frame correlation. J. Microsc.
25. Hotulainen, P., and Lappalainen, P. (2006). Stress fibers are generated
by two distinct actin assembly mechanisms in motile cells. J. Cell
Biol. 173, 383–394.
26. Geiger, B., Spatz, J.P., and Bershadsky, A.D. (2009). Environmental
sensing through focal adhesions. Nat. Rev. Mol. Cell Biol. 10, 21–33.
27. Le Clainche, C., and Carlier, M.-F. (2008). Regulation of actin assembly
associated with protrusion and adhesion in cell migration. Physiol. Rev.
between force and focal complex development. J. Cell Biol. 159,
29. Ballestrem, C., Hinz, B., Imhof, B.A., and Wehrle-Haller, B. (2001).
Marching at the front and dragging behind: Differential alphaVbeta3-
integrin turnover regulates focal adhesion behavior. J. Cell Biol. 155,
30. Macdonald, A., Horwitz, A.R., and Lauffenburger, D.A. (2008). Kinetic
model for lamellipodal actin-integrin ‘clutch’ dynamics. Cell Adh Migr
31. Wehrle-Haller, B., and Imhof, B. (2002). The inner lives of focal adhe-
sions. Trends Cell Biol. 12, 382–389.
32. Lele, T.P., Thodeti, C.K., Pendse, J., and Ingber, D.E. (2008). Investi-
gating complexity of protein-protein interactions in focal adhesions.
Biochem. Biophys. Res. Commun. 369, 929–934.
33. Mould, A.P., Askari, J.A., Barton, S., Kline, A.D., McEwan, P.A., Craig,
S.E., and Humphries, M.J. (2002). Integrin activation involves a confor-
mational change in the alpha 1 helix of the beta subunit A-domain.
J. Biol. Chem. 277, 19800–19805.
34. Chen, J., Salas, A., and Springer, T.A. (2003). Bistable regulation of
integrin adhesiveness by a bipolar metal ion cluster. Nat. Struct. Biol.
35. Choi, C.K., Vicente-Manzanares, M., Zareno, J., Whitmore, L.A.,
Mogilner, A., and Horwitz, A.R. (2008). Actin and alpha-actinin orches-
trate the assembly and maturation of nascent adhesions in a myosin II
motor-independent manner. Nat. Cell Biol. 10, 1039–1050.
36. Guo, W.H., and Wang, Y.L. (2007). Retrograde fluxes of focal adhesion
proteins in response to cell migration and mechanical signals. Mol.
Biol. Cell 18, 4519–4527.
37. Wang, Y.L. (2007). Flux at focal adhesions: Slippage clutch, mechanical
gauge, or signal depot. Sci. STKE 2007, pe10.
38. Thomas, W. (2008). Catch bonds in adhesion. Annu. Rev. Biomed. Eng.
39. Selhuber-Unkel, C., Lo ´pez-Garcı ´a, M., Kessler, H., and Spatz, J.P.
(2008). Cooperativity in adhesion cluster formation during initial cell
adhesion. Biophys. J. 95, 5424–5431.
40. Roca-Cusachs, P., Gauthier, N.C., del Rio, A., and Sheetz, M.P. (2009).
Clustering of alpha(5)beta(1) integrins determines adhesion strength
whereas alpha(v)beta(3) and talin enable mechanotransduction. Proc.
Natl. Acad. Sci. USA 106, 16245–16250.
41. Kong, F., Garcı ´a, A.J., Mould, A.P., Humphries, M.J., and Zhu, C. (2009).
Demonstration of catch bonds between an integrin and its ligand.J. Cell
Biol. 185, 1275–1284.
42. Friedland, J.C., Lee, M.H., and Boettiger, D. (2009). Mechanically acti-
vated integrin switch controls alpha5beta1 function. Science 323,
43. Sabass, B., Gardel, M.L., Waterman, C.M., and Schwarz, U.S. (2008).
High resolution traction force microscopy based on experimental and
computational advances. Biophys. J. 94, 207–220.
44. Aratyn, Y.S., Schaus, T.E.,Taylor, E.W., andBorisy,G.G. (2007).Intrinsic
dynamic behavior of fascin in filopodia. Mol. Biol. Cell 18, 3928–3940.
Transient Frictional Slip during Adhesion Assembly