cAMP-induced Epac-Rap activation inhibits epithelial cell migration by
modulating focal adhesion and leading edge dynamics
Karen S. Lylea,1,2, Judith H. Raaijmakersa,2, Wytse Bruinsmaa,
Johannes L. Bosa,⁎, Johan de Rooijb,⁎
aDepartment of Physiological Chemistry, Centre for Biomedical Genetics and Cancer Genomics Centre, Universiteitsweg 100, 3584 CG Utrecht, the Netherlands
bDepartment of Cell Biology, the Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, the Netherlands
Received 22 November 2007; received in revised form 18 January 2008; accepted 18 January 2008
Available online 31 January 2008
Epithelial cell migration is a complex process crucial for embryonic development, wound healing and tumor metastasis. It depends on
alterations in cell–cell adhesion and integrin–extracellular matrix interactions and on actomyosin-driven, polarized leading edge protrusion. The
small GTPase Rap is a known regulator of integrins and cadherins that has also been implicated in the regulation of actin and myosin, but a direct
role in cell migration has not been investigated. Here, we report that activation of endogenous Rap by cAMP results in an inhibition of HGF- and
TGFβ-induced epithelial cell migration in several model systems, irrespective of the presence of E-cadherin adhesion. We show that Rap
activation slows the dynamics of focal adhesions and inhibits polarized membrane protrusion. Importantly, forced integrin activation by antibodies
does not mimic these effects of Rap on cell motility, even though it does mimic Rap effects in short-term cell adhesion assays. From these results,
we conclude that Rap inhibits epithelial cell migration, by modulating focal adhesion dynamics and leading edge activity. This extends beyond the
effect of integrin affinity modulation and argues for an additional function of Rap in controlling the migration machinery of epithelial cells.
© 2008 Elsevier Inc. All rights reserved.
Keywords: Rap; Epac; cAMP; Migration; Integrin; Cytoskeleton
Epithelial cell migration is a complex process induced by
specific growth factors that takes place during certain stages
of embryonic development, organogenesis and wound healing.
In response to oncogenic signals, epithelial cell migration also
mediates tumor invasion and metastasis . Epithelial cell
migration requires the disruption of cell–cell adhesion , the
modification of the integrin–extracellular matrix (ECM) interac-
tions  and engagement of the actomyosin-based migration
machinery that induces polarized membrane protrusion .
Beneath thisleading edge protrusion ofa migrating cell, integrin-
mediated focal contacts are initiated and subsequently reinforced
by tension generated in the actomyosin cytoskeleton . As a
consequence, they grow larger and alter their composition to
become focal adhesions (FAs) . Contraction of the actomyosin
cytoskeleton attached to leading edge FAs pulls the cell body
forward and is used to disassemble rear-end FAs [6,7]. The effi-
ciency of migration in two-dimensional culture also depends on
the balance between ECM concentration and the extent of
Several growth factors implicated in tumor metastasis can
induce the processes described above in cultured cells, resulting
in the scattering of initially clustered epithelial cells. The most
well-known inducers of epithelial cell scattering are transform-
ing growth factor-β (TGFβ) and hepatocyte growth factor
(HGF) . TGFβ induces scattering in many different cell lines,
Available online at www.sciencedirect.com
Cellular Signalling 20 (2008) 1104–1116
⁎Corresponding authors. J. de Rooij is to be contacted at Hubrecht Institute,
Uppsalalaan 8, 3584 CT, The Netherlands. Tel.: +31 30 2121960; fax: +31 30
E-mail addresses: firstname.lastname@example.org (J.L. Bos), email@example.com
(J. de Rooij).
University of California-San Francisco, 513 Parnassus Avenue, HSW-618,
San Francisco, CA, United States.
2These authors contributed equally to this paper.
0898-6568/$ - see front matter © 2008 Elsevier Inc. All rights reserved.
invariantly accompanied by silencing of the E-cadherin gene
through Smad signaling . The most prominent induction
of scattering by HGF occurs in MDCK cells  and does not
capacity, but correlates with increased integrin-mediated adhe-
sion and depends on actomyosin based tension . Given the
lethal consequences of tumor metastasis, we aim to understand
the cellular machinery that governs epithelial cell migration.
cAMP is a pivotal second messenger that regulates a wide
range of cellular processes. Signaling through cAMP and protein
kinase A (PKA) has been implicated in cytoskeletal regulation
and cell migration . The effects of PKA on cell migration can
matrixused[13–16].cAMPalsoactivates the guaninenucleotide
exchange factor (GEF) Epac that can subsequently activate the
small GTPase Rap . Rap is an important regulator of both
integrin- and cadherin-mediated adhesion (reviewed in [18–20]).
Although it is not yet completely understood how Rap regulates
these two processes, several proteins that interact with its GTP-
bound form have been identified that may serve as effector
proteins . In the case of integrin-mediated adhesion, Rap
regulates both integrin affinity and integrin avidity, or clustering,
depending on the type of integrin and the cell type [21–24]. Two
effectors of Rap1, Riam and RapL, have been shown to be
important in the regulation of integrin affinity [25,26], although
likely recruit junctional proteins to sites of developing cell–cell
and the junctional complex [20,29,30]. Because of its established
function as a regulator of integrin-mediated cell adhesion, a role
for Rap in cell migration has been suggested. Direct evidence
comes from studies in leukocytes, where chemokine-induced
endothelium and subsequent endothelial transmigration [31,32].
Clearly, in addition to PKA, also Epac/Rap signaling may be
involved in regulation of cell migration via cAMP.
Previously, we reported that Rap is involved in cell surface
expression of E-cadherin and the stabilization of cell–cell
junctions and that cAMP-induced activation of Rap through
Epac1 inhibits HGF-induced cell scattering. These observations
suggested that scattering is inhibited by stabilization of adherens
junctions . However, a deficiency in C3G, another RapGEF,
results in increased migration velocity , indicating that Rap
might also have a restraining effect on cell migration itself.
Moreover, Zhang et al. recently described that TGFβ-induced
transformation of cells is inhibited by cAMP, independently of
PKA . As TGFβ, in contrast to HGF, down regulates the
E-cadherin expression levels, cAMP-Epac-Rap signaling may
effect on cell–cell junctions.
Here, we show that activation of Rap through Epac1 inhibits
epithelial cell migration in a number of different model systems
in response to both HGF and TGFβ, irrespective of the presence
of cell–cell junctions. Interestingly, forced integrin activation,
by the integrin-activating antibody TS2/16, does not inhibit
migration, even though it induces adhesion to the ECM to the
same extent as Rap activation does. Apparently, the effects of
Rap on cadherin-mediated adhesion and integrin activation are
not sufficient to inhibit epithelial cell migration, indicating that
inhibition of the basal cell migration machinery is the critical
step downstream of Rap that mediates its effects on scattering.
To further understand the mechanism of Rap-induced inhibition
of cell migration, we studied the migration machinery in more
detail and observed that Rap activation impairs the dynamics of
focal adhesions and blocks protrusive activity at the leading
edge in migrating cells. These effects are also not mimicked by
integrin activating antibodies. Together, these data show that
Rap regulates focal adhesion and leading edge dynamics,
independently of integrin activation, to restrain epithelial cell
2.1. cAMP-induced Rap activation inhibits HGF-stimulated
epithelial cell migration
To investigate the effect of Rap activation on epithelial cell
migration we used MDCK cells, which do not express endoge-
nous Epac, and MDCK cells stably expressing GFP-tagged
well plate coated with collagen (10 µg/mL), simultaneously
filmed by phase-contrast microscopy for 2 h, stimulated with
HGF, and filmed for an additional 18 h. Parental MDCK cells
exhibited a typical response to HGF; the cells initially spread,
disrupted cell–cell contacts and migrated away from their neigh-
bors (Movie S1 and Fig. 1A). MDCK-GFP-Epac cells showed a
similar response to HGF treatment compared to parental MDCK
cells, but scattering was inhibited in HGF-stimulated MDCK-
bromo-cAMP (Movie S1 and Fig. 1A) or the cAMP-elevating
drug forskolin (Movie S2 and Fig. S1). 8-Bromo-cAMP did not
inhibit scattering in parental MDCK cells, suggesting that activa-
tion of Epac mediated this response. To confirm that PKA is not
specific cAMP analog 8-CPT-2OMe-cAMP  and the PKA-
specific cAMP analog N6-Bnz-cAMP . 8-CPT-2OMe-cAMP
inhibited scattering in MDCK-GFP-Epac, but not in parental
MDCK cells (Movie S1 and Fig. 1A). In contrast, activation of
PKA with N6-Bnz-cAMP did not inhibit scattering in MDCK-
GFP-Epac cells (Movie S2 and Fig. S1).
To quantify these observations, the migration velocity of
parental and MDCK-GFP-Epac cells was determined by tracking
cell nuclei using custom-written, automated cell-tracking soft-
ware. For each condition 3 independent time-lapse image series
were analyzed, resulting in the tracking of at least 300 individual
cells per condition. MDCK and MDCK-GFP-Epac cells in-
creased their migration velocity from 0.5 to 1.2 µm/min, within
16 h of HGF stimulation (Fig. 1B, solid and dashed black lines).
MDCK cells treated with 8-bromo-cAMP showed the same re-
sponse to HGF (Fig. 1B, solid grey line). In contrast, MDCK-
GFP-Epac cells exposed to HGF and 8-bromo-cAMP or 8-CPT-
2OMe-cAMP showed no significant increase in cell velocity
(Fig. 1B, grey dashed line and Fig. 1C, black dashed line,
1105K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
1106 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
respectively). To evaluate the statistical significance of the inhi-
bition of migration by cAMP, the velocity values for each time
point after the plateau of maximal velocity was reached (t=750–
1000 min post-HGF) were averaged (Fig. 1E) and subjected to a
Students t-test. Thus we conclude that cAMP-induced activation
of Epac (and not PKA) and subsequent activation of endogenous
Rap in MDCK cells strongly inhibits HGF-induced scattering.
2.2. Rap activation inhibits HGF-induced cell migration in the
absence of cell–cell junctions
As Rap has been previously shown to modulate both cell–cell
and cell-ECM adhesion receptors [21–23,33,38,39], we aimed to
investigate whether the cAMP-induced inhibition of cell migra-
tion is simply caused by the stabilization of cell–cell adhesion
through Rap. To this end, we analyzed the velocity of MDCK-
GFP-Epac cells in the absence of cell–cell junctions using two
approaches. First, we examined the HGF-induced motility in low
Ca2+(20 µM Ca2+)-containing medium that does not support
homotypic E-cadherin adhesion. Second, we used our tracking
software to distinguish clearly single cells from cells in contact
with neighboring cells in high Ca2+conditions to specifically
analyze the HGF-induced motility of single cells.
Fig. 1C (dotted black line and Movie S3) shows that HGF-
stimulated MDCK-GFP-Epac cells in low Ca2+-containing
medium achieved a similar maximum velocity of 1.2 µm/min.
Notably, these cells showed an earlier increase in cell velocity
compared to MDCK-GFP-Epac cells cultured in high Ca2+
(Fig. 1C, solid grey line), indicating that cell–cell junctions do
need to be disrupted before the velocity increase is observed. In
the presence of 8-CPT-2OMe-cAMP, MDCK-GFP-Epac cells in
cAMP-treated cells in the presence of high Ca2+(Fig. 1C, dashed
grey and dashed black lines, respectively and Movie S3). To
corroborate this conclusion, the velocity increase of clearly non-
contacted cells present in high Ca2+conditions (usually less than
10% of all cells) was inhibited as well (Fig. 1D, dashed black
line). Comparison of maximum velocities and statistical analysis
was performed as above and is depicted in Fig. 1E. These data
demonstrate that Rap activation strongly inhibits epithelial cell
migration even in the absence of functional adherens junctions.
2.3. Rap inhibits epithelial cell migration induced by several
different growth factors
To investigate if the inhibitory effect of Rap activation on cell
migration is restricted to HGF-induced motility, we used A549
cells that scatter in response to HGF or TGFβ, another notorious
metastasis-promoting growth-factor  that disrupts cell–cell
junctions by silencing the E-cadherin gene . In analogy to
MDCK cells, we constructed a cell line stably expressing Epac1
8-CPT-2OMe-cAMP.Asshown inFig.2B,bothHGF and TGFβ
induce scattering in A549 and A549-Epac cells. In A549-Epac
cells, scattering is abolished by 8-CPT-2OMe-cAMP, whereas in
parental A549 cells scattering is normal.
We used time-lapse phase-contrast microscopy to further
on it in these cells (Movie S4) and measured cell migration to
solid black and grey lines, respectively), which is completely
abolished by 8-CPT-2OMe-cAMP (Fig. 2C, dotted black and
grey lines, respectively). Quantification and statistical evaluation
is shown in Fig. 2E. Thus, Rap activation efficiently blocks scat-
tering induced by two distinct signaling pathways, one of which
results in a loss of E-cadherin expression (even in the presence of
active Rap, see Fig. 3F).
2.4. cAMP induction of endogenous Epac-Rap signaling
inhibits epithelial cell migration
blocks cell migration, we used renal cell carcinoma RCC10
little to no E-cadherin . RCC10 cells respond to HGF
stimulation by activating ERK1/2 and express the β1 integrin
required for migration on a collagen matrix, as well as Epac1
(Figs. 2A and 3C). These cells showed a 1.5 fold increase in
velocity upon HGF stimulation (Fig. 2D, solid black line and
Movie S5). The motogenic response of these cells was more
rapid compared to MDCK cells, and cells achieved their maxi-
mal velocity after 1 h of HGF stimulation. This probably
reflects the absence of cell–cell junctions in these cells.
Importantly, treating cells with 8-CPT-2OMe-cAMP to speci-
fically activate Epac/Rap, completely inhibited the HGF-
induced increase in cell motility (Fig. 2D, dotted black line and
Movie S5 and quantification in 2E). Together, these data show
that cAMP-mediated activation of endogenous Rap via both
exogenously- and endogenously-expressed Epac1 leads to the
inhibition of epithelial cell scattering induced in different cell
lines and by different growth factors. As this inhibition is
independent of the level of E-cadherin expression and the
presence of cell–cell junctions, this suggests that Rap has an
inhibitory effect on the induction of cell migration itself.
Fig. 1. cAMP-induced Rap activation inhibits HGF-induced cell migration in the absence or presence of cell–cell adhesion. (A) Representative images from phase
contrast timelapse image series showing the inhibitory effect of 8-bromo-cAMP and 8-CPT-2OMe-cAMP on HGF-induced cell motility in MDCK-GFP-Epac, but not
parental MDCK cell lines. Scale bar is 100 µm. (B) Velocity time-course, by automated tracking of approximately 300 cells from three independent time-lapse image
series, to quantify HGF-induced migration and cAMP-induced inhibition in these cells. (C) Velocity time-course by tracking of approximately 600 cells from 3
independent time-lapse image series to quantify the inhibitory effect of 8-CPT-2OMe-cAMP on HGF-induced MDCK and MDCK-GFP-Epac cell velocity in the
presence (1.8 mM Ca2+) and absence (20 µM Ca2+) of cell–cell junctions. (D) Velocity time-course by tracking of approximately 60 cells from 3 independent time-
lapse image series, showing the effect of 8-CPT-2OMe-cAMP on the migration of non-contacted MDCK-GFP-Epac cells in high Ca2+. (E) Average cell velocity
(averaging values from all time-points at the plateau phase of maximum velocity (750–1000 min)) (±SD) from each of the velocity time-courses in B, C, and D. An
asterisk (⁎) indicates pb0.0001.
1107K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
Fig. 2. Rap activation inhibits several types of growth factor-induced epithelial cell migration. (A) Western blot showing Epac1 and β1-integrin expression in A549-
Epac and RCC10cells. (B) 8-CPT-2OMe-cAMP inhibits TGFβ- and HGF-induced scattering of A549-Epac cells. Representative phase contrast images of A549-Epac
cellsshowingtheeffect of 8-CPT-2OMe-cAMP onTGFβ- andHGF-induced cell scattering.Scale baris 100 µm.Cells werestimulated withTGFβ orHGF for 24h, in
the presence or absence of 8-CPT-2OMe-cAMP. (C) Velocity time-course showing the effect of 8-CPT-2OMe-cAMP on TGFβ- and HGF-induced A549-Epac cell
migration. (D) Velocity time-course showing the inhibitory effect of 8-CPT-2OMe-cAMP on HGF-induced RCC10 cell migration. (E) Average cell velocity (±SD) at
the plateau phase (1100–1200 min) of A549 cells and RCC10 cells (400–500 min) from each of the velocity time-courses in C and D, respectively.
1108K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
2.5. 8-CPT-2OMe-cAMP activates Rap, but does not interfere
in growth factor signaling
One of the mechanisms via which Rap may interfere in
growth factor-induced cell migration, is through direct inhibi-
tion of growth factor signaling. Rap pulldowns were used to
confirm that 8-CPT-2OMe-cAMP activates Epac to induce Rap-
GTP in Epac-expressing (endogenous or exogenous) cell lines.
This activation of Rap was present for the complete duration of
our time lapse experiments and was not affected by HGF or
TGFβ stimulation (Fig. 3A–C). Next, we examined if Epac1
expression or Rap activation with 8-CPT-2OMe-cAMP could
suppress HGF- and TGFβ-mediated ERK activation . In
agreement with our previous experiments in MDCK cells ,
activation of Rap in RCC10 cells does not interfere with HGF-
induced ERK1/2 phosphorylation at any of the time-points
investigated (Fig. 3C). Rap activation also did not modulate the
TGFβ-induced phosphorylation of ERK1/2 in A549-Epac cells
(Fig. 3D). Furthermore, neither exogenous Epac1 expression
nor 8-CPT-2OMe-cAMP stimulation suppressed Smad2 phos-
phorylation after TGFβ addition (Fig. 3E), as was reported in
293T cells . Finally, Rap activation did not affect the down
regulation of E-cadherin protein levels that is observed in the
presence of TGFβ (Fig. 3F). These data show that Rap acti-
vation does not interfere in the major growth factor receptor
signaling events important for scattering, but rather acts to
prevent the induction of migration at a more downstream level.
For instance, Rap may directly inhibit the migration machinery.
2.6. Forced integrin activation does not mimic the inhibitory
effects of Rap on cell migration
Besides the disruption of cell–cell adhesion , cells also
that induces polarized membrane protrusion  for efficient cell
downstream of Rap [21,28], could explain the inhibition of cell
Fig. 3. Rap activation does not interfere with growth factor signaling. (A) Rap is activated by 8-CPT-2OMe-cAMP in MDCK-GFP-Epac cells in the presence and
absence of HGF. Cells were stimulated for 16 h with HGF and/or 8-CPT-2OMe-cAMP. Blots shown are representative of at least 3 independent experiments. (B) Rap
activation in parental and Epac-expressing A549 cells. Cells were treated for the indicated time periods with TGFβ and 8-CPT-2OMe-cAMP prior to performing a
Rap-GTP pulldown assay. Blots shown are representative of at least 4 independent experiments. (C) Rap activation does not interfere with HGF-induced ERK
activationand8-CPT-2OMe-cAMP-induced Rapactivityisnot significantlyaffected byshort-or long-termHGFstimulation inRCC10cells.Cellsweretreated for the
indicated time periods with HGF and/or 8-CPT-2OMe-cAMP prior to performing a Rap-GTP pulldown assay and probing of whole cell lysates for Erk activation.
Blots shown are representative of 3 independent experiments. (D) Rap activation does not affect TGFβ-induced ERK activation in A549-Epac cells. Cells were
incubated for 24 h with the indicated stimuli. Whole cell lysates were probed with the appropriate antibodies as indicated to show activation and equal loading. Blots
shown are representative of at least 3 independent experiments. (E) Epac expression and Rap activation do not affect TGFβ-induced phosphorylation of Smad2. Cells
were incubated for 24 h with the indicated stimuli. Whole cell lysates were probed with the appropriate antibodies as indicated to show activation and equal loading.
Blots shown are representative of 3 independent experiments. (F) Rap activation does not restore TGFβ-induced downregulation of E-cadherin in A549-Epac cells.
Cells were incubated for 24 h with the indicated stimuli. Whole cell lysates were probed with the appropriate antibodies as indicated to show activation and equal
loading. Blots shown are representative of at least 3 independent experiments.
1109K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
into their high-affinity conformation and thus mimic inside-out
integrin activation as induced by Rap. Because these antibodies
recognize human, but not dog integrins, we used RCC10 and
A549 cells and not MDCK cells. For RCC10 cells plated on
collagen, which mainly binds to β1 integrins to mediate cell
migration, we used the β1 integrin-activating antibody TS2/16
. For A549 cells plated on fibronectin (Fn), which can bind
to β1 and β3 integrins to facilitate cell migration, we used a com-
binationofTS2/16 andthe β3 integrin-activatingantibodyLIBS6
on collagen, TS2/16 and 8-CPT-2OMe-cAMP induce adhesion to
a similar extent (Fig. 4C), indicating that indeed β1 integrins are
the main collagen receptors in these cells. In A459 cells on Fn,
TS2/16 induces adhesion to a higher extent than 8-CPT-2OMe-
cAMP, whereas LIBS6 results in a very small induction of adhe-
sion (Fig. 4F), indicating that β1 integrins are the main receptors
for Fn in these cells. However, in migration experiments TS2/16
Fig. 4. Forcedintegrin-activation by antibodiesdoesnot mimic the effect of Rap activationon cell migration. (A) Velocityincrease of RCC10 cellsby HGFis inhibited
by 8-CPT-2OMe-cAMP, but not TS2/16. In a live-cell phase-contrast time-lapse, HGF and 100 µM 8-CPT-2OMe-cAMP or 3 µg/ml TS2/16 were added at t=60 min.
the plateau-phase (400–500 min) from the velocity time-course in A. (C) Adhesion time course showing comparable 8-CPT-2OMe-cAMP- and TS2/16-induced
for the indicated time points. Data are means±SD; n=4. (D) Velocity increase of A549-Epac cells by HGF is not inhibited by TS2/16 and LIBS6. HGF and 100 μM
8CPT-2OMe-cAMP or antibodies (3 µg/ml TS2/16 and 3.7 µg/ml LIBS6) were added during imaging at t=120 min and velocity was determined by automated cell-
tracking of approximately 300 cells from 3 independent time-lapse image series. (E) Average cell velocity (±SD) at the plateau-phase (1100–1200 min) of A549 cells
on 1 µg/ml fibronectin in the absence or presence of 100 µM 8-CPT-2OMe-cAMP or 3 µg/ml TS2/16 or 3.7 µg/ml LIBS6, allowed to adhere for 30 min, washed and
quantified. Data are means±SD; n=4.
1110 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
Fig. 5. Rap activation enhances focal adhesion area and stability. (A) Representative images showing the effect of 8-CPT-2OMe-cAMP on FAs in MDCK-Epac cells.
Cells were plated on collagen in the presence or absence of 8-CPT-2OMe-cAMP, fixed and stained for paxillin. Lower panels show images segmented in ImageJ.
(B) Images were processed and the total focal adhesion and cell area for at least 35 cells from 2 independent experiments were quantified in ImageJ as described in the
Materials and methods section. (C) Representative images from timelapse series showing the effect of 8-CPT-2OMe-cAMP and TS2/16 on the stability of GFP-
paxillin-labeled FAs in HGF-treated RCC10cells. A 7.6 µm×7.6 µm area of the cell is highlighted(white dashedbox), and white arrowheads show FAs that persist for
more than 60 min within this area. Scale bar is 5 µm. (D) Focal adhesion persistence analysis was performed in ImageJ as described in the Materials and methods
section for eight cells (~200 FAs tracked per condition) and the data are expressed as the mean±SD.
1111 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
plateau-velocity and statistical evaluation is shown in Fig. 4B
(RCC10 cells, 400–500 min) and 4E (A549 cells, 1100–
1200 min). This shows that simply inducing integrin activation
does not inhibit HGF-induced motility and indicates that a
different effect downstream of Rap is involved.
2.7. Rap activation inhibits focal adhesion turnover and front-
rear polarity in contrast to forced β1-integrin activation
Because just simple integrin activation and ECM adhesion
does not seem to be enough to inhibit migration, we focused on
focal adhesions (FAs), to investigate if Rap might affect integrin
signaling downstream of the integrin-ECM connection. FAs reg-
To determine if Rap activation modulates these structures in
epithelial cells, we examined the effect of 8-CPT-2OMe-cAMP
on FA morphology in MDCK-Epac cells (MDCK cells stably
expressing un-tagged Epac1 ) plated on collagen and stained
for paxillin, a major component of FAs. Three hours after plating,
MDCK-Epac cells (in the absence or presence of 8-CPT-2OMe-
cAMP) were adhered and spread onto the collagen-coated glass
coverslip. In the presence of 8-CPT-2OMe-cAMP, FAs were
larger and more elongated and paxillin staining was more intense
A549-Epac and RCC10 cells (data not shown). Measuring the
area of paxillin intensity that clearly surmounted background
fluorescence, revealed a 1.3-fold increase in the relative focal
adhesion area in 8-CPT-2OMe-cAMP-treated compared to un-
treated cells (Fig. 5B, pb0.0001). These data show that Rap acti-
vation enhances FAs in MDCK-Epac cells.
Next, we examined the dynamic behavior of FAs in migrating
not form tightly clustered cell colonies. RCC10 cells were
transfected with GFP-paxillin, plated and imaged using widefield
fluorescence microscopy. Cells transfected with GFP-paxillin
appeared indistinguishable from untransfected cells in phase-
contrast images demonstrating that exogenous GFP-paxillin ex-
pression does not significantly alter the morphology of RCC10
FAs forming at the leading edge of protrusions, and larger FAs in
the body and trailing edge of the cell (Movie S6 and Fig. 5C
magnified region at the leading edge of the cell, and white
arrowheads identify the FAs that persist for longer than 60 min.
Upon stimulation with 8-CPT-2OMe-cAMP, cells rapidly altered
their focal adhesion dynamics with fewer FAs forming and dis-
appearing within the 90 min time period (Fig. 5C, panels from
edge adhesions and an increase in large central FAs. As a con-
sequence, cells appeared to rapidly lose their distinct front/rear
polarity and migration was inhibited, when stimulated with 8-
CPT-2OMe-cAMP, but not with TS2/16.
The lifetime of individual FAs is highly variable in migrating
cells. To quantify focal adhesion lifetime in a comprehensive
manner, we randomly selected FAs in the first frame of the
image sequence and determined how long these persisted. Cells
treated with HGF showed a nearly equal percentage of FAs that
persisted for a short time (less than 15 min), intermediate times
(16–60 min) and a very long time (N61 min) (Fig. 5D). When
Rap was activated in these cells, the percentage of FAs lasting
for more than 61 min was increased 2-fold (Fig. 5D, grey and
white bars, p=0.002), whereas we observed a concomitant 2-
fold decrease in the percentage of FAs lasting 15 min or shorter
(p=0.04). TS2/16 treatment did not mimic the effect of 8-CPT-
2OMe-cAMP, as it did not result in any increase in the per-
centage of long-lasting FAs, but rather in a 1.5-fold increase of
small, short-lived FAs (Fig. 5C and D, grey and black bars,
p=0.06). Furthermore, the presence of TS2/16 enhanced the
polarized phenotype of these cells (as judged by the asymmetric
distribution of small and large FAs).
in the loss of front rear polarity. This response is very different
from the response to treatment with an integrin-activating anti-
body, which shows increased FA dynamics and no loss of cell-
2.8. Rap activation reduces lamellipodial activity
In RCC10 GFP-paxillin timelapses, we noted a decrease in the
number of leading edge protrusions after Rap activation (Fig. 6A,
arrowheads). To clearly determine whether Rap activation indeed
affects HGF-induced protrusion and retraction, we examined the
membrane dynamics of RCC10 cells in the presence and absence
of 8-CPT-2OMe-cAMP. RCC10 cells were transiently transfected
with the membrane marker GFP-CAAX , plated on collagen-
coated glass-bottom dishes overnight and imaged using widefield
fluorescence microscopy. In the presence of HGF, RCC10 cells
formed approximately equivalent and polarized areas of net
protrusion and retraction (Fig. 6B, solid black areas and Movie
when Rap was activated with 8-CPT-2OMe-cAMP in HGF-
treated cells (Fig. 6B), but remained in cells treated with TS2/16.
Furthermore, protrusion and retraction were not polarized in 8-
CPT-2OMe-cAMP-treated cells as no persistence in any direc-
tion was observed. These differences were not due to toxic or
otherwise disturbing effects of the expression of GFP-CAAX as
expression was similar in HGF-only and HGF+8-CPT-2OMe-
cAMP or HGF+TS2/16-treated cells used for this analysis (as
determined by fluorescence intensity). Furthermore, HGF-only
treated cells that expressed GFP-CAAX migrated indistinguish-
able from untransfected RCC10 cells (data not shown).
To quantify the protrusive activity, we divided the area of
leading-edge extension in consecutive frames by the total cell
area during the 60 min of the experiment (Fig. 6C). The average
membrane activity of HGF-treated cells was 1.4-fold higher than
The membrane activity was reduced 1.4-fold when Rap was
activated with 8-CPT-2OMe-cAMP (Fig. 6C, black dashed line
activation by TS2/16 did not mimic the inhibitory effect of Rap
activation. These data demonstrate that Rap activation blocks the
1112 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
HGF-induced increase in polarized membrane protrusion and
For cells to migrate efficiently, both cell–cell junctions and
actomyosin cytoskeleton needs to be induced [1–3]. Polarized
alsorequiredfor efficient cellmigration . Asthe small GTPase
Rap is a known regulator of cell junctions and integrin-mediated
adhesion  and has been suggested to be involved in cell
migration , we investigated how Rap interferes in epithelial
cell migration. Surprisingly, neither the stabilization of cell–cell
junctions nor the activation of integrins could account for this
effect on migration. Major receptor signaling pathways like ERK
and Smad (unlike previously reported ), are also not affected,
indicating that signaling through Epac/Rap does not interfere in
HGF or TGFβ signaling, but rather acts downstream to block the
induction of cell migration by these transforming growth factors.
During migration on the ECM, integrin-based FAs provide
. Regulated, efficient formation and turnover of FAs is re-
quired for optimal cell migration. Small adhesions at the leading
FAs that provide the tension for the disassembly of adhesions at
the rear of the cell [6,47].
as a marker  and found a rapid alteration of focal adhesion
dynamics. After Rap was activated, fewer FAs formed at the
leading edge and focal adhesion disassembly was inhibited,
Fig. 6. Rap activation inhibits HGF-induced membrane protrusions. (A) 8-CPT-2OMe-cAMP decreases the number of protrusions in RCC10 cells treated with HGF.
The arrowheads indicate newly formed protrusions in GFP-paxillin expressing RCC10 cells in each timelapse frame. (B) Rap activation reduces the HGF-induced
increase in membrane protrusion and retraction area. Representative images of RCC10 cells transiently expressing GFP-CAAX and their corresponding total
protrusion and retraction area (black areas) over 60 min after treatment with HGF alone, HGF+8-CPT-2OMe-cAMP, and HGF+TS2/16. The difference in membrane
area between sequential timelapse images was determined in ImageJ as described in the Materials and methods section. Scale bar is 5 µm. (C) Timecourse showing the
difference in membrane area relative to the total area of the cell, due to protrusion and retraction, in sequential frames of 60 min timelapses. The data are the averages
from 9, 27, 18, and 19 cells for mock-, HGF-, HGF+8-CPT-2OMe-cAMP-, and HGF+TS2/16-treated cells, respectively. (D) Average membrane activity (±SD) of
RCC10 cells during the time-course in C.
1113 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
resulting in an overall increase in FA persistence and an apparent
cells. One explanation for the observed effects on focal adhesion
complexes is an increase in integrin-mediated cell adhesion.
However, the β1-integrin-activating antibody TS2/16 did not
affect FA dynamics in the same way as Rap activation, although
both induced a similar level of adhesion. This suggests that
migration. In accordance with these observations, Huttenlocher
et al. showed that forcing integrins into a high affinity state using
antibodies does not influence FA morphology and does not
strongly inhibit cell migration, but shifts the optimal migration
conditions to a lower concentration of ECM substrate. In contrast,
mutant integrinsthat havelost proper regulationoftheir linkage to
the cytoskeleton and exhibit forced cytoskeletal linkage, show
. We observe similar effects when Rap is activated in epithelial
cells that are treated with growth factors. Our results, therefore,
suggest that Rap has an effect on the integrin-cytoskeletal linkage
and that this effect is more important for the inhibition of cell
achieved by the same molecules that mediate the affinity modu-
lation (Riam and Talin ), remains to be shown. Thus far we
that we used for our experiments.
In addition to stabilizing FAs, Rap inhibits the formation of
polarized membrane protrusions during migration. This process
is driven by actin polymerization (the motor), but also depends
on a regulation of the integrin-cytoskeletal linkage (the clutch)
as has been illustrated by two recent papers [49,50]. Several
reports have indicated that Rap may influence actin polymer-
ization, the driving force behind lamellipodial protrusion. For
instance, Rap has strong stabilizing effects on cortical actin in
endothelial cells [51,52] and the Rap interacting protein RIAM
was shown to increase the amount of filamentous actin,
presumably through its interaction with profilin . Further-
more, Rap interacts with several RacGEFs, known regulators of
actin polymerization, to increase cell-spreading , although it
has also been reported that Rap can function as an antagonist of
Rac signaling .
Because we also observe an effect on FA dynamics, the effect
of endogenous Rap activity on lamellipodial protrusion could be
explained by a stabilization of the connection between the
integrins and the cytoskeleton. In other words, the clutch is
engaged too long, leading to a loss of productive protrusion.
Further studies are required to determine whether Rap regulates
identify the molecular mechanism that mediates the Rap-induced
inhibition of cell migration.
Finally, contractile tension generated within the actomyosin
cytoskeleton is also required for efficient migration. Increased
actomyosin contraction results in increased FA size  and
dynamic regulation of myosin is critical to efficient protrusion
. An alternative explanation for the effects on FA size and
dynamics and for the effects on protrusion could be the recent
finding that Rap1 regulates myosin in Dictyostelium .
However, we could not detect any effects of Rap activation on
myosin light chain phosphorylation downstream of HGF or
TGFβ, arguing against this possibility. Also, the Rap1 effector
that mediates the myosin induction by Rap1 is not conserved
between Dictyostelium and mammals.
We conclude that activation of endogenous Rap leads to an
inhibition of growth factor-induced epithelial cell migration by
targeting the basal migration machinery. This effect is indepen-
affinity modulation of β1-integrins. Rap inhibits epithelial cell
migration through the stabilization of focal adhesions and the
inhibition of membrane protrusion, possibly by stabilizing the
connection between the actin cytoskeleton and integrins.
5. Materials and methods
5.1. Cell lines and culture
Stable MDCK-GFP-Epac cells were created by transfection of MDCK with
pEGFP-C1-Epac1 followed by selection with G418. Polyclonal MDCK cells
stably expressing moderate levels of GFP-Epac were isolated by fluorescence
activated cell sorting (FACS) from this cell line. MDCK-Epac1 cells were
described previously . Stable Epac1-expressing A549 cells were created by
infecting A549 cells with Epac1 ecotrophic virus. The Epac1 gene was linked
via an IRES sequence to a zeocin resistance gene. Forty-eight hours after
infection, cells were placed under selection with zeocin (0.2 mg/ml) to select for
Epac1 expressing cells. Monoclonal A549-Epac1 cell lines expressing moderate
levelsof Epac1 weresingle-cellsorted fromthese polyclonalcell lines by FACS.
MDCK, MDCK-GFP-Epac and MDCK-Epac cells were cultured in DMEM
supplemented with 10% fetal calf serum (FCS), glutamine, and antibiotics.
A549 and RCC10 cell lines were cultured in RPMI supplemented with
glutamine, antibiotics, and 10% or 8% FCS, respectively. RCC10 cells were
transiently transfected using Fugene transfection reagent (Roche) according to
the manufacturer's protocol. Cells were plated in complete medium 24 h after
transfection, and analysis of expressed proteins occurred 24 h thereafter.
The GFP-CAAX construct expresses an N-terminal GFP-tagged tetra-amino
acid motif (CAAX) that localizes to the plasma membrane . The GFP-
paxillin construct expresses an N-terminal EGFP-tagged fusion to human
paxillin and was generously provided by Dr. Marc Ginsberg (University of
California-San Diego). GFP-Epac1 contains amino acids 2-881 of Epac1 fused
at its N-terminus to EGFP in the pEGFP-C1 vector.
5.3. ECM proteins
For analysis of MDCK cell motility, non-tissue culture treated 48-well plates
were coated with collagen type I from calf skin (Sigma) for 2 h at 37 °C, washed
3 times with phosphate buffered saline (PBS), and blocked with 1% heat-
denatured bovine serum albumin (BSA) in PBS for 1 h at 37 °C. For analysis of
A549 cell motility, 48 well plates were coated with 1 µg/mL fibronectin for 2 h
at 37 °C and washed 3 times in PBS. In all other cases, glass-bottomed dishes
and coverslips were coated with collagen type I from rat tendon (Upstate) for
16 h at 4 °C and washed 3 times in PBS.
5.4. Rap activation assays and immunoblotting
detected following Western blotting with polyclonal anti-Rap antibody (Santa
Cruz). Polyclonal phospho-ERK (Thr202/Tyr204) antibody and phospho-Vasp
(Ser157) antibody were obtained fromCell Signaling. Polyclonalphospho-Smad2
1114 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
were from ChemiconInternational. Anti-β1 integrinand anti-β-catenin antibodies
were from BD Transduction Laboratories and anti-tubulin antibody was from
5.5. Live cell microscopy
For phase-contrast imaging, MDCK, MDCK-GFP-Epac, and RCC10 cells
were plated in medium containing 0.5% FCS and 10 mM Hepes, pH 7.4, 24 h
before image acquisition in non-tissue culture–treated polystyrene well plates
coated with 10 µg/mL collagen (or the indicated concentrations). Prior to
imaging, 100 µM 8-CPT-2OMe-cAMP, 10 µM forskolin, 300 µM N6-Bnz-
cAMP or 3 µg/mL TS2/16 were added to the appropriate wells, wells were
completely filled with medium and the plate was sealed using silicon grease and
a glass plate. Images were acquired every 6 min using a 10×0.5 NA Plan
objective lens and a 0.5 NA ELWD condenser with a Zeiss Axiocam camera on
a Zeiss Axiovert 200 M microscope in climate-controlled incubator. A robotic
stage (Zeiss MCU 28) was used to collect images at different stage positions. All
electronic microscope functions were controlled using Axiovision software
(Zeiss). The cells were imaged for 2 h in absence of HGF, and then 5 ng/ml HGF
was added to cells on the microscope stage to prevent loss of the cells of interest
and imaging was continued for 16 h. At least three timelapse series were
acquired for each condition in each separate experiment.
A549 and A549-Epac cells were plated on a fibronectin-coated (1 µg/mL)
48-well plate 24 h before imaging, which was performed as above. At the
indicated timpoint 10 ng/mLTGFβ, 25 ng/ml HGF and 100 µM 007 or 3 µg/ml
integrin-activating antibody TS2/16 and/or LIBS6 were added to the appropriate
wells. GFP-tagged proteins were imaged in an 8-well Lab-tek chambered
coverglass (Nalge Nunc International, Rochester, NY) coated with 10 µg/mL
collagen on a Zeiss Axiovert 200 M microscope using a Lambda DG-4 Ultra
High Speed Wavelength Switcher from Sutter Instruments as a light source.
Fluorescent images were acquired every 2 or 3 min using either a 40X/1.3 oil or
a 63X/1.25 oil Neofluar objective lens.
5.6. Fixation and immunolocalization
Cells were fixed in freshly-prepared 4% paraformaldehyde for 10 min,
permeabilized in 0.2% Triton X-100 for 5 min, and blocked in PBS containing
with monoclonal paxillin antibody (BD Biosciences) for 1 h in PBS containing
0.2% BSA, followed by incubation with the appropriate secondary antibody for
45 min at room temperature. Images were acquired using a Zeiss Axioskop 2
microscope fitted with a Zeiss Axiocam CCD camera and 100X Plan APO
5.7. Image analysis and processing
To determine cell trajectories in phase-contrast timelapse image series, the
centroids of the nuclei were followed. To automate this and allow for unbiased
analysis of many cells in multiple timelapses, a program was written in Matlab
(Mathworks) that segments images based on pixel intensity and determines the
5 consecutive frames are automatically discarded (manuscript in preparation, JdR
and Danuser G.). Detection fidelity in our experiments was over 80%, which was
confirmed by eye for each individual timelapse. To distinguish single cells from
clustered cells in this program, areas occupied by cells were determined by edge-
detection andoverlayedwiththedetected nucleitodetermine ifone (singlecell) or
more (clustered cells) nuclei were present in a detected cell-area. Similar results
were obtained when a smaller number of randomly selected cells from a number
of timelapses were analyzed using the track objects function in MetaMorph
(Universal Imaging Corp.).
Focal adhesion and total cell areas from images of fixed and stained cells
were measured in ImageJ. First, the fluorescence intensities of images from 2
to remove background staining that was consistently observed around the
nucleus.FAsweresegmentedusingthe analyzeparticlesfunction inImageJ,and
our segmentation procedure was inspected visually.
selected in similarly localized regions of the cell and marked in the first frame
FAs were tracked until they were no longer visible and the frame at which the
focal adhesion disappeared was recorded.
To quantify protrusion dynamics, we determined the changes in membrane
area between sequential timelapse images of RCC10 cells expressing GFP-
CAAX. Using Metamorph software, an exclusive threshold was applied to each
normalized image series to define the outer cell membrane. We created a journal
in Metamorph that defined and measured the thresholded areas and produced a
applied to sequential binary timelapse images to determine areas of protrusion
area. Images showing the net protrusion and retraction were made by applying
the sum slices option in the z-project function of ImageJ to sequentially
5.8. Adhesion assays
Cells were trypsinized, washed once in RPMI containing 10% FCS, and
allowed to recover surface proteins for 1.5 h in suspension in RPMI containing
0.5% FCS, glutamine, antibiotics, and 10 mM Hepes, pH 7.4, at 37 °C with
constant, gentle rolling. 8-CPT-2OMe-cAMP (200 µM) and TS2/16 antibody
to the wells of a 48-well polystyrene cell culture dish coated with 3 µg/ml
collagen (for RCC10 cells) or 1 µg/ml fibronection (A549-Epac cells). After
rolling, 50,000 cells in 100 µL were plated per well, making the final
concentrations of 8-CPT-2OMe-cAMP and TS2/16 100 µM and 1 µg/mL,
respectively. Adhesion was allowed to proceed for the indicated times at 37 °C,
and unbound cells were discarded by washing three times with PBS preheated to
37 °C. Adhered cells were lysed in the wells by adding 200 µl of assay buffer
substrate (Sigma-Aldrich). The total amount of cellular protein per well was
was incubated for 20 h at 37 °C and terminated by addition of 100 µl of 1 N
NaOH. Absorbance was measured at 405 nm. Every condition was measured in
5.9. Statistical analysis
Statistical analysis was performed in Kaleidagraph (Synergy Software)
using the unpaired Student's t-test for samples of unequal variance.
We thank members of our labs for comments and discussion,
Livio Kleij for assistance with live cell imaging experiments, Dr
Jeroen Bakkers for providing the GFP-CAAX construct and Dr
Mark Ginsberg for pEGFP-paxillin. We thank Dr Jan Willem
from Dr Mark Ginsberg, whom we thank for it as well. We thank
Arnoud Sonnenberg for support, suggestions and critical reading
of the manuscript. This work was supported by a Long-Term
Fellowship from the Human Frontier Science Program (KSL),
and grantsfromthe Dutch Cancer Society(KWF Kankerbestrijd-
ing) to JHR (2003–2956) and JdR (2006-3714).
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.cellsig.2008.01.018.
1115 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116
References Download full-text
 J.P. Thiery, Nat. Rev., Cancer 2 (6) (2002) 442.
 E.H. Danen, Curr. Pharm. Des. 11 (7) (2005) 881.
 R. Horwitz, D. Webb, Curr. Biol. 13 (19) (2003) R756.
 E. Zamir, B. Geiger, J. Cell Sci. 114 (Pt 20) (2001) 3583.
 D.J. Webb, J.T. Parsons, A.F. Horwitz, Nat. Cell Biol. 4 (4) (2002) E97.
 A.J. Ridley, M.A. Schwartz, K. Burridge, R.A. Firtel, M.H. Ginsberg, G.
Borisy, J.T. Parsons, A.R. Horwitz, Science 302 (5651) (2003) 1704.
 A. Huttenlocher, M.H. Ginsberg, A.F. Horwitz, J. Cell Biol. 134 (6) (1996)
 J.P. Thiery, J.P. Sleeman, Nat. Rev., Mol. Cell Biol. 7 (2) (2006) 131.
 M. Deckers, M. van Dinther, J. Buijs, I. Que, C. Lowik, G. van der Pluijm,
P. ten Dijke, Cancer Res. 66 (4) (2006) 2202.
 J. de Rooij, A. Kerstens, G. Danuser, M.A. Schwartz, C.M. Waterman-
Storer, J. Cell Biol. 171 (1) (2005) 153.
 A.K. Howe, Biochim. Biophys. Acta 1692 (2-3) (2004) 159.
 M.L. Edin, A.K. Howe, R.L. Juliano, Exp. Cell Res. 270 (2) (2001) 214.
 S. Kim, M. Harris, J.A. Varner, J. Biol. Chem. 275 (43) (2000) 33920.
 O. Dormond, C. Ruegg, Thromb. Haemost. 90 (4) (2003) 577.
 J. de Rooij, F.J. Zwartkruis, M.H. Verheijen, R.H. Cool, S.M. Nijman, A.
Wittinghofer, J.L. Bos, Nature 396 (6710) (1998) 474.
 J.L. Bos, Curr. Opin. Cell Biol. 17 (2) (2005) 123.
 E. Caron, J. Cell Sci. 116 (Pt 3) (2003) 435.
 M.R. Kooistra, N. Dube, J.L. Bos, J. Cell Sci. 120 (Pt 1) (2007) 17.
 K. Katagiri, M. Hattori, N. Minato, S. Irie, K. Takatsu, T. Kinashi, Mol.
Cell. Biol. 20 (6) (2000) 1956.
 E. Caron, A.J. Self, A. Hall, Curr. Biol. 10 (16) (2000) 974.
 K.A. Reedquist, E. Ross, E.A. Koop, R.M. Wolthuis, F.J. Zwartkruis, Y. van
 K.M. de Bruyn, S. Rangarajan, K.A. Reedquist, C.G. Figdor, J.L. Bos,
J. Biol. Chem. 277 (33) (2002) 29468.
 E.M. Lafuente, A.A. van Puijenbroek, M. Krause, C.V. Carman, G.J.
Freeman, A.Berezovskaya, E.Constantine, T.A. Springer, F.B. Gertler, V.A.
Boussiotis, Dev. Cell 7 (4) (2004) 585.
 K. Katagiri, A. Maeda, M. Shimonaka, T. Kinashi, Nat. Immunol. 4 (8)
 K. Katagiri, M. Imamura, T. Kinashi, Nat. Immunol. 7 (9) (2006) 919.
 J. Han, C.J. Lim, N. Watanabe, A. Soriani, B. Ratnikov, D.A. Calderwood,
W. Puzon-McLaughlin, E.M. Lafuente, V.A. Boussiotis, S.J. Shattil, M.H.
Ginsberg, Curr. Biol. 16 (18) (2006) 1796.
 T. Hoshino, T. Sakisaka, T. Baba, T. Yamada, T. Kimura, Y. Takai, J. Biol.
Chem. 280 (25) (2005) 24095.
 A. Glading, J. Han, R.A. Stockton, M.H. Ginsberg, J. Cell Biol. 179 (2)
 E.S. Wittchen, J.D. van Buul, K. Burridge, R.A. Worthylake, Curr. Opin.
Hematol. 12 (1) (2005) 14.
 M.J. Lorenowicz, M.Fernandez-Borja, P.L.Hordijk,Arterioscler. Thromb.
Vasc. Biol. 27 (5) (2007) 1014.
 L.S. Price, A. Hajdo-Milasinovic, J. Zhao, F.J. Zwartkruis, J.G. Collard, J.L.
Bos, J. Biol. Chem. 279 (34) (2004) 35127.
 Y. Ohba, K. Ikuta, A. Ogura, J. Matsuda, N. Mochizuki, K. Nagashima, K.
 A. Zhang, Z. Dong, T. Yang, Am. J. Physiol. Renal. Physiol. 291 (6)
Genieser, S.O. Doskeland, J.L. Blank, J.L. Bos, Nat. Cell Biol. 4 (11) (2002)
 A.E. Christensen, F. Selheim, J. de Rooij, S. Dremier, F. Schwede, K.K.
Dao, A. Martinez, C. Maenhaut, J.L. Bos, H.G. Genieser, S.O. Doskeland,
J. Biol. Chem. 278 (37) (2003) 35394.
 C. Hogan, N. Serpente, P. Cogram, C.R. Hosking, C.U. Bialucha, S.M.
Feller, V.M. Braga, W. Birchmeier, Y. Fujita, Mol. Cell. Biol. 24 (15)
 A.L. Knox, N.H. Brown, Science 295 (5558) (2002) 1285.
 H. Kasai, J.T. Allen, R.M. Mason, T. Kamimura, Z. Zhang, Respir. Res. 6
R. Raval, T.S. O'Brien, P.H. Maxwell, Cancer Res. 66 (7) (2006) 3567.
 P. Conrotto, I. Yakymovych, M. Yakymovych, S. Souchelnytskyi,
J. Proteome Res. 6 (1) (2007) 287.
 E. van de Wiel-van Kemenade, Y. van Kooyk, A.J. de Boer, R.J. Huijbens,
P. Weder, W.vande Kasteele,C.J.Melief, C.G.Figdor,J.CellBiol.117(2)
 A.L. Frelinger III, I. Cohen, E.F. Plow, M.A. Smith, J. Roberts, S.C. Lam,
M.H. Ginsberg, J. Biol. Chem. 265 (11) (1990) 6346.
 K. Burridge, M. Chrzanowska-Wodnicka, Annu. Rev. Cell Dev. Biol. 12
 S. von der Hardt, J. Bakkers, A. Inbal, L. Carvalho, L. Solnica-Krezel, C.P.
Heisenberg, M. Hammerschmidt, Curr. Biol. 17 (6) (2007) 475.
 R. Zaidel-Bar, C. Ballestrem, Z. Kam, B. Geiger, J. Cell Sci. 116 (Pt 22)
 D.J. Webb, C.M. Brown, A.F. Horwitz, Curr. Opin. Cell Biol. 15 (5) (2003)
 K. Hu, L. Ji, K.T. Applegate, G. Danuser, C.M. Waterman-Storer, Science
315 (5808) (2007) 111.
 C.M. Brown, B. Hebert, D.L. Kolin, J. Zareno, L. Whitmore, A.R.
Horwitz, P.W. Wiseman, J. Cell Sci. 119 (Pt 24) (2006) 5204.
 X. Cullere, S.K. Shaw, L. Andersson, J. Hirahashi, F.W. Luscinskas, T.N.
Mayadas, Blood 105 (5) (2005) 1950.
 M.R. Kooistra, M. Corada, E. Dejana, J.L. Bos, FEBS Lett. 579 (22)
 W.T. Arthur, L.A. Quilliam, J.A. Cooper, J. Cell Biol. 167 (1) (2004) 111.
 A.M. Valles, M. Beuvin, B. Boyer, J. Biol. Chem. 279 (43) (2004) 44490.
 S.L. Gupton, C.M. Waterman-Storer, Cell 125 (7) (2006) 1361.
 T.J. Jeon, D.J. Lee, S. Merlot, G. Weeks, R.A. Firtel, J. Cell Biol. 176 (7)
 B. Franke, J.W. Akkerman, J.L. Bos, Embo J. 16 (2) (1997) 252.
 M.A. Schwartz, K. Denninghoff, J. Biol. Chem. 269 (15) (1994) 11133.
1116 K.S. Lyle et al. / Cellular Signalling 20 (2008) 1104–1116