Search tips
Search criteria 


Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. 2010 April; 30(8): 1971–1983.
Published online 2010 February 12. doi:  10.1128/MCB.01247-09
PMCID: PMC2849475

Cadherins and Pak1 Control Contact Inhibition of Proliferation by Pak1-βPIX-GIT Complex-Dependent Regulation of Cell-Matrix Signaling [down-pointing small open triangle]


It is crucial for organ homeostasis that epithelia have effective mechanisms to restrict motility and cell proliferation in order to maintain tissue architecture. On the other hand, epithelial cells need to rapidly and transiently acquire a more mesenchymal phenotype, with high levels of cell motility and proliferation, in order to repair epithelia upon injury. Cross talk between cell-cell and cell-matrix signaling is crucial for regulating these transitions. The Pak1-βPIX-GIT complex is an effector complex downstream of the small GTPase Rac1. We previously showed that translocation of this complex from cell-matrix to cell-cell adhesion sites was required for the establishment of contact inhibition of proliferation. In this study, we provide evidence that this translocation depends on cadherin function. Cadherins do not recruit the complex by direct interaction. Rather, we found that inhibition of the normal function of cadherin or Pak1 leads to defects in focal adhesion turnover and to increased signaling by phosphatidylinositol 3-kinase. We propose that cadherins are involved in regulation of contact inhibition by controlling the function of the Pak1-βPIX-GIT complex at focal contacts.

The maintenance of the highly organized architecture of epithelial cells within a tissue is crucial for epithelial function and organ homeostasis. Epithelial organization is controlled by cell-matrix and cell-cell contacts. Cells can restrict cell migration and proliferation in a cell-cell contact-dependent manner by a process called contact inhibition, which is thought to be a critical process for maintaining epithelial organization. On the other hand, epithelial cells need to retain the ability to transiently acquire a more mesenchymal phenotype, with high levels of cell motility and proliferation, to repair epithelial damage. Since epithelial repair generally occurs by epithelial sheet movements in which cell-cell contacts are still present (16), precise control mechanisms are required to regulate the temporal release of contact inhibition in this process. The molecular mechanisms that control contact inhibition, however, are still not completely understood.

Calcium-dependent interactions between the extracellular domains of the adhesion molecule E-cadherin and interactions of β-, α-, and p120-catenins with its cytoplasmic domain mediate adhesion between adjacent epithelial cells (48). E-cadherin has been implicated in contact inhibition (25, 30, 38) and likely functions as a tumor suppressor. While contact inhibition may ultimately be regulated by cadherin-mediated upregulation of the cyclin-dependent kinase 2 inhibitor p27kip1 (25, 38), the intermediate steps between cadherin-mediated cell-cell contact and growth arrest are unclear. The elusive nature of E-cadherin in the regulation of contact inhibition is likely due to the fact that E-cadherin can control cell proliferation by different mechanisms. First, E-cadherin can act via β-catenin, which not only functions in cell-cell adhesion, as a component of adherens junctions, but also can translocate to the nucleus to act as a transcriptional activator of proliferation-promoting genes under the control of the TCF/LEF (T-cell factor/leukocyte enhancing factor) family of transcription factors (48). However, even though numerous studies have reported that the growth-inhibitory effects of E-cadherin depend on the interaction of E-cadherin with β-catenin, many of these studies excluded a role for TCF/LEF involvement, indicating that β-catenin controls proliferation by both TCF/LEF-dependent and -independent pathways (15, 17, 30, 34). Second, E-cadherin can directly interact with growth factor receptor tyrosine kinases (29, 32, 42). These interactions inhibit the mitogenic response upon stimulation of these receptors by soluble ligands in a cell-cell adhesion-dependent manner (30, 32, 42), although stimulation of growth factor signaling has also been reported (29). Finally, E-cadherins engage in functional cross talk with signaling pathways induced by cell-matrix adhesion via integrins. Integrins can cluster in different types of adhesion plaques. For the purposes of this paper, we collectively call these structures focal contacts. In general, integrin-mediated signaling promotes cell proliferation by activating different mitogenic pathways. These include pathways that involve phosphatidylinositol 3-kinase (PI3K), a focal adhesion kinase (FAK)-Src module, and small GTPases of the Rho family, such as Rac. As is the case for cadherins, there is also a high degree of cross talk between integrin and growth factor receptor signaling (7).

It is unclear how cell-cell and cell-matrix signaling coordinates the signaling pathways that control contact inhibition. A detailed understanding of this cross talk is hampered by the fact that cell-cell and cell-matrix adhesion processes activate similar pathways but that the resulting cellular response depends on the site of activation. A case in point is the small GTPase Rac, which is required for epithelial wound healing and promotes cell migration when activated at cell-matrix contact sites but promotes cell-cell adhesion and formation of adherens junctions when activated at cell-cell junctions (45).

We previously showed that translocation of a protein complex containing the Rac effector protein Pak1 and the Rac/cdc42 guanine exchange factor (GEF) βPIX from focal contacts to areas of cell-cell contact is involved in the establishment of contact inhibition (53). This complex, which also contains the multifunctional linker protein GIT1, has been implicated in regulation of cell motility and turnover of focal adhesions in many cell types (52). However, it also regulates adherens junctions and vascular permeability (11, 41) in endothelial cells. Based on our previous work, we hypothesized that βPIX coordinates cell signaling pathways at cell-matrix and cell-cell junctions. In this study, we further investigated the role of βPIX in E-cadherin-dependent establishment of contact inhibition. We found that cadherin-mediated cell-cell adhesion is required for lateral recruitment of βPIX via a mechanism that does not depend on direct interactions with the Pak-PIX-GIT complex. Furthermore, we provide evidence for a role of the Pak-PIX-GIT complex in cadherin-dependent maturation of focal contacts, which in turn may be involved in the downregulation of PI3K-dependent signaling pathways.


Antibodies and other reagents.

Polyclonal rabbit anti-β-catenin, antihemagglutinin (anti-HA), anti-Pak1 (Pak1-N20 and Pak1-C19), and anti-GIT1 and goat anti-cadherin-6 (K-cadherin) antisera were from Santa Cruz Biotechnology. Mouse monoclonal antibodies (MAbs) against E-cadherin (C-terminal), paxillin, βPIX, and p27kip1 were from Becton Dickinson. Polyclonal anti-βPIX antiserum was from Chemicon. Mouse anti-β-tubulin, anti-desmosomal protein, and rat anti-E-cadherin (extracellular domain) antibodies were from Sigma-Aldrich. Mouse antibromodeoxyuridine (anti-BrdU) MAb was from Calbiochem. Rat anti-ZO-1 was obtained via the Developmental Studies Hybridoma Bank at the University of Iowa. Rabbit anti-MEK, anti-pS298-MEK, anti-pS218/222-MEK, anti-Akt, anti-p-S473-Akt, and anti-Pak1 and -2 isoform-specific antibodies were from Cell Signaling. Secondary antibodies comprised Alexa Fluor 488/555-conjugated donkey anti-mouse IgG (heavy plus light chains [H+L]) and anti-rabbit IgG and Alexa Fluor 488/633-conjugated goat anti-rat IgG (H+L) (Invitrogen).

Cell culture and cell lines.

MDCK cell lines that inducibly express Pak1-KD-K299R, Pak1-K299R,R193A,P194A (called Pak1-K299RΔPIX in this paper), Pak1-AID, Pak1-AID-L107F, Pak1-T423E, Pak1-R193A,P194A,T423E, and DN-E-cadherin (DN-Cad) by use of the Tet-off system were described previously (35, 43, 53) and were grown in modified Eagle's medium (MEM) with l-glutamine, 10% fetal calf serum, and penicillin-streptomycin. Cells inducibly expressing DN-Cad were a gift from James Marrs, Indiana University. Cells were maintained in growth medium with 20 ng/ml doxycycline (DOX). To induce gene expression, doxycycline was removed by washing the cells three times, followed by addition of growth medium. Cells were induced 2 days before being plated.

To generate MDCK cells in which expression of E-cadherin was ablated, a short hairpin RNA (shRNA) against canine E-cadherin in pSuper-puro (5′-GGACGTGGAAGATGTGAAT-3′) (6), under the control of the H1 promoter, was subcloned into the pcDNA6/V5 vector. As a negative control, shRNA against green fluorescent protein (GFP) was used. MDCK cells were transfected using the calcium phosphate coprecipitation method, and shRNA-E-cadherin-expressing clones were selected using 6 μg/ml blasticidin S hydrochloride (Invitrogen). The extent of knockdown compared to that in the shRNA-GFP controls was determined by quantitative Western blotting, using β-tubulin as a loading control.

Western blotting.

Unless indicated otherwise, cell lysates were from confluent cells grown for 6 days on 12-mm by 0.4-μm-pore-size polycarbonate Transwell filters (Corning-Costar) and prepared in lysis buffer containing 1% SDS. After protein determination by bicinchoninic acid assay (Pierce), equal amounts of protein were diluted in 2× Laemmli buffer, loaded onto SDS-polyacrylamide gels, and transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore). Quantitative Western blotting was performed using an Odyssey detector (Li-Cor, Lincoln, NE). Secondary antibodies for Odyssey detection comprised Alexa Fluor IRDye 800-conjugated or Alexa Fluor 680-conjugated donkey anti-mouse and donkey anti-rabbit IgGs (Invitrogen). Secondary antibodies for Western blots stained for phosphorylated Akt and MEK1/2 were horseradish peroxidase (HRP)-conjugated donkey anti-rabbit IgGs (Jackson Immunoresearch Laboratories), and blots were visualized by enhanced chemiluminescence (ECL; GE Healthcare).

Rac1 activation assay.

Levels of Rac-GTP were determined using Pak3 CRIB RBD assays as previously described (13).


To knock down βPIX expression with small interfering RNA (siRNA) oligonucleotides, we used Silencer custom siRNA ID214584 from Ambion (siRNA βPIX KD1; target sequence, CGACAGGAAUGACAAUCAC) or a Stealth siRNA (Invitrogen) against canine βPIX (βPIX KD2; target sequence, TAAACTTTGCTCTCACTACCAGTTG). An oligonucleotide against GFP (Con-1) or a scrambled Stealth oligonucleotide (Con-2) was used as a control. All Pak siRNAs were Stealth siRNAs (Invitrogen) and comprised the following target sequences: Pak1 KD-1, GGAUGCGGCUACAUCUCCUAUUUCA; Pak1 KD-2, CAGCCGAAGAAAGAGCUGAUUAUUA; Pak2 KD-1, GCCAGAACAGUGGGCUCGAUUAUUA; and Pak2 KD-2, CAGAGGUGGUUACACGGAAAGCUUA. Scrambled Stealth oligonucleotides were used as controls for Pak knockdown experiments.

For transient transfections, 4 × 106 cells were electroporated with 100 pmol of siRNA, using Amaxa Nucleofector reagent (program T23, buffer T).

Gradient ultracentrifugation.

Confluent 6-day-old cultures of MDCK cells in three 10-cm dishes (about 1 × 107 cells) were rinsed with phosphate-buffered saline (PBS). Next, cells were harvested by cell scraping, using ice-cold TNE buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA) containing protease inhibitors [5 μg/ml pepstatin, 10 μg/ml chymostatin, 5 μg/ml leupeptin, 10 μg/ml antipain, 5 mM benzamidine, 100 U/ml aprotinin, 2 mM 4-(2-aminoethyl) benzenesulfonyl fluoride-hydrochlorine (AEBSF)] and phosphatase inhibitors (1 mM sodium vanadate, 1 mM sodium fluoride). Cells were spun down (5 min, 800 × g, 4°C) and resuspended in 800 μl TNE buffer. Cells were broken by repeated passage (25 to 30 times) through a 27.5-gauge syringe, and postnuclear supernatants (PNS) were prepared by centrifugation at 800 × g for 5 min at 4°C. Nonequilibrium gradient centrifugation was performed essentially as described previously (49). Briefly, PNS was diluted with TNE and mixed at a 1:1 ratio with 60% iodixanol in water (OptiPrep; Axis-Shield PoC As, Oslo, Norway). When indicated, n-octyl-β-d-glucoside was added to a final concentration of 1% (wt/vol). Samples were centrifuged in a near-vertical NVT65.2 rotor (Beckman Coulter Co.) at 350,000 × g at 4°C for 4 h. Fractions of 500 μl were collected from the top to the bottom of the gradient for further analysis, which included protein determination and density analysis using a refractometer. Equal volumes of gradient fractions were loaded onto SDS-polyacrylamide gels and further analyzed by Western blotting.

Calcium switch assay.

To disrupt cell-cell adhesion, cells were grown in low-calcium medium. Briefly, confluent cells on a Transwell filter were washed with PBS, followed by three 10-min washes in 2 mM PBS-EGTA. Next, cells were incubated in low-calcium medium (MEM with 5 μM Ca2+) for 4 h, which results in a loss of cell-cell contacts. Cell-cell contacts were restored by replacing low-calcium medium with normal growth medium, which contains 1.8 mM Ca2+.

TCF/LEF reporter gene assay.

β-Catenin/TCF-activated transcription was determined essentially as described previously (19). MDCK cells conditionally expressing DN-Cad or Pak1-K299R were induced for 3 days, whereas controls were kept in the presence of doxycycline. For subconfluent cells, 5 × 105 cells were plated in a 12-well plate and assayed 16 h after plating. For confluent cells, 1 × 106 cells were plated and assayed 6 days after plating. Cells were transfected with 1 μg of the reporter construct pTOPFLASH, using Lipofectamine 2000 (Invitrogen). As a negative control, we used the FOPFLASH reporter plasmid, containing a mutated TCF/LEF binding site. All cells were cotransfected with 0.5 μg pTK-Renilla (Promega). Luciferase activity was determined by a dual-luciferase reporter assay system (Promega) according to the manufacturer's instructions. MDCK cells transiently transfected with 1 μg of a stabilized form of β-catenin (β-catenin-S37A) were used as a positive control. Samples were normalized for transfection efficiency, using the Renilla luciferase signal, and values were plotted as ratios of the luminescence in pTOPFLASH- and FOPFLASH-transfected cells.

Confocal fluorescence microscopy.

Unless indicated otherwise, confocal images were taken through the middle section of the cell, which precludes imaging of the focal contacts at the basal side of the cell. Focal contacts were visualized after a 30-s extraction with CSK buffer {50 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 10 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)], pH 6.8, 0.5% (vol/vol) Triton X-100} prior to fixation. For staining of paxillin, Pak1, and βPIX, samples were washed in PBS and fixed in 2% paraformaldehyde in PBS for 2 min at room temperature, followed by a 5-min fixation in methanol at −20°C. Samples were blocked and permeabilized with 10% normal goat serum and 0.1% Triton X-100 in PBS for 30 min, and incubations with primary and secondary antibodies were done in the same buffer. Samples were mounted in FluorSave (Calbiochem) supplemented with 10 μg/ml DAPI (4′,6-diamidino-2-phenylindole) to stain nuclei. Cells were imaged on a Zeiss 510 LSM confocal microscope with an Axiovert 200 M microscope and a C-Apochromat 63×/1.2W Corr lens. Images were cropped and adjusted for brightness with Adobe Photoshop CS, version 9.0, and composite images with scale bars were made with Adobe Illustrator CS2, version 12.0.0.

BrdU incorporation assay.

Determination of cell proliferation by incorporation of BrdU was performed as previously described (53). Since confluent Pak1-K299R-expressing cells form a multilayer, we used a relatively short incubation time (3 h) in order to reliably determine the percentage of BrdU-positive cells.


βPIX is required for contact inhibition of proliferation.

When plated at confluent densities, MDCK cells continued to proliferate after establishing initial cell-cell contacts (~12 h after plating). Over a time course of 1 to 6 days, they formed a monolayer in which cells gradually became taller and tightly packed and eventually established contact inhibition of proliferation, as assessed by decreased BrdU incorporation (Fig. (Fig.1A)1A) and highly increased protein expression of p27kip1 (Fig. (Fig.1B).1B). We previously showed that dominant-negative (DN) mutants of Pak1 abrogate contact inhibition of proliferation (53) (Fig. (Fig.1).1). When we inhibited Pak1 function by expressing kinase-dead Pak1-K299R (without DOX [−DOX]) or the autoinhibitory domain of Pak1 (Pak1-AID), cells continued to proliferate and failed to upregulate p27kip1 (53) (Fig. 1C and D). As a result, cells expressing these DN-Pak1 mutants formed a multilayer (53). In contrast, control cells grown in the presence of DOX (+DOX) or cells expressing wild-type Pak1 or a Pak1-AID construct with a point mutation that abolishes autoinhibition (Pak1-AID-L107F) were still able to establish contact inhibition (Fig. (Fig.1C)1C) and did not form a multilayer (53). We previously also showed that cells expressing the constitutively active phospho-mimetic Pak1 mutant Pak1-T423E had phenotypes similar to those of the DN-Pak1 mutants (53). While we originally hypothesized that this was due to kinase-dependent deregulation of Pak1 dynamics at focal adhesions, we later found that these cells also express an ~50-kDa product (arrow in Fig. Fig.1E),1E), which is detected by polyclonal antibodies against the Pak1 N terminus but not by antibodies against the C terminus, which we initially used (Fig. (Fig.1E).1E). Based on its size, we expect this N-terminal fragment to contain Pak1-AID but not an intact Pak1 kinase domain. Thus, this fragment will likely act as an inhibitor of endogenous Pak1, which may explain the similar phenotypes of the Pak1-K299R-, Pak1-AID-, and Pak1-T423E-expressing cells. Consistent with this, Pak1-R193A,P194A,T423E, the corresponding active Pak1 construct, which cannot bind βPIX, does not generate the N-terminal fragment and lacks an obvious phenotype. The same was true for cells expressing Pak1-K299RΔPIX (53).

FIG. 1.
DN-Pak1 perturbs contact inhibition and inhibits lateral βPIX recruitment. (A and B) Parental MDCK cells were plated on Transwell filters at confluent densities. (A) Cell proliferation was determined by BrdU incorporation assays 1, 3, and 6 days ...

Pak1 binds and regulates the Rac/cdc42 GEF βPIX (or p85-cool1), which in turn forms a tight complex with proteins of the GIT family (5, 52). We demonstrated that the inability of DN-Pak1 mutant-expressing cells to establish contact inhibition in scrape-wound healing assays depends on the interaction of Pak1 with βPIX and correlates with an inhibition of lateral recruitment of Pak and βPIX (53) (Fig. (Fig.1F).1F). We thus hypothesize that the lateral recruitment of βPIX-containing protein complexes as cells become confluent is required for contact inhibition of cell proliferation. To further analyze the relationship between lateral βPIX recruitment and the establishment of contact inhibition, we compared the kinetics of lateral recruitment of βPIX and the formation of adherens junctions, using lateral accumulation of β-catenin as a marker for the latter. For this purpose, we plated cells at near-confluent densities on Transwell filters and fixed and immunostained cells on different days after plating (Fig. (Fig.2A).2A). Using confocal microscopy, we acquired optical sections at the midplane of cells, thus precluding visualization of βPIX in focal contacts. In contrast to β-catenin, which localized almost exclusively at cell-cell contacts within 24 h after plating, the kinetics of lateral recruitment of βPIX were much slower. Only after 2 to 3 days did lateral βPIX staining appear in a subset of cells, and 5 to 6 days after establishment of cell-cell contacts, lateral βPIX staining was widespread within the culture (>99% of cells). Strikingly, the kinetics of lateral recruitment of βPIX correlated precisely with the establishment of contact inhibition of proliferation, as determined by BrdU incorporation and the gradual increase of p27kip1 levels during this time course (Fig. 1A and B).

FIG. 2.
Kinetics of lateral βPIX recruitment and establishment of contact inhibition. (A) Parental MDCK cells were plated on Transwell filters at confluent densities. Cells were fixed after 1 to 6 days and costained for βPIX and β-catenin. ...

To investigate if βPIX was required for contact inhibition, we knocked down βPIX expression by using specific siRNA oligonucleotides. Knockdown of βPIX with two different oligonucleotides increased proliferation in 3-day-old confluent cells, which was rescued by inducible expression of human βPIX that was resistant to siRNA-mediated knockdown (Fig. 2B and C). Knockdown of βPIX did not affect cell growth of subconfluent cells (not shown). Together, these data suggest a specific role of βPIX in cell-cell contact-mediated growth control.

Lateral βPIX is not required for junctional integrity.

Since junctional recruitment of Pak, PIX, and GIT has been implicated in the stability of adherens junctions (11, 22), we tested if inhibition of lateral βPIX recruitment destabilized cell-cell junctions in MDCK cells. Knockdown of βPIX did not result in changes in protein levels of E-cadherin and ZO-1 (Fig. (Fig.3A)3A) or affect the integrity of adherens or tight junctions, as determined by staining for E-cadherin (Fig. (Fig.3B),3B), β-catenin, ZO-1, or a desmosomal marker (not shown). Similarly, expression of Pak1-K299R, which also inhibits lateral recruitment of βPIX (Fig. (Fig.1E),1E), did not affect junction protein expression or junctional integrity, as judged by analysis of the same markers (Fig. 3C and D and data not shown). Furthermore, we found no effect on transepithelial resistance in monolayers expressing Pak1-K299R (not shown).

FIG. 3.
Knockdown of βPIX or expression of DN-Pak1 does not affect cell-cell junctions. (A and B) Cells were transfected with siRNA against GFP (control) or βPIX and plated on Transwell filters. (A) Western blots of cell lysates prepared 3 days ...

Lateral recruitment of βPIX and establishment of contact inhibition require functional adherens junctions.

Since E-cadherin-mediated cell-cell adhesion has been implied in contact inhibition of proliferation in epithelial cells, we tested if lateral recruitment of βPIX required E-cadherin-based adherens junctions. For this purpose, we made MDCK cell lines stably expressing an shRNA against E-cadherin (E-cad-KD cells). Surprisingly, even though E-cadherin protein levels were knocked down >95% (Fig. (Fig.4A,4A, compare third and fourth lanes), E-cad-KD cells had no obvious morphological phenotype. Thus, confocal immunofluorescence analysis showed that E-cad-KD cells still formed adherens junctions, tight junctions, and desmosomes, as judged by lateral localization of ZO-1, a desmosomal marker protein, and the adherens junctional markers β-catenin, α-catenin, and p120 (not shown). Furthermore, E-cad-KD cells still recruited βPIX laterally and established contact inhibition (not shown). The lack of a clear phenotype of the E-cad-KD cells was recently also reported by others (10) and suggests that other cadherins compensate for the loss of E-cadherin. Indeed, cadherin-6 (also known as K-cadherin), another major cadherin in MDCK cells (39), was increased >2-fold in E-cad-KD cells (Fig. (Fig.4A,4A, compare third and fourth lanes). Since we were unable to create stable E-cadherin/cadherin-6 double-knockdown cells, we used a well-established dominant-negative approach to downregulate cadherins in MDCK cells (43, 44). Inducible overexpression of a dominant-negative E-cadherin mutant with a truncated extracellular domain (DN-Cad) resulted in effective downregulation of both E-cadherin and cadherin-6 in MDCK cells (Fig. 4A and B), with a small (~20 to 30%) increase of β-catenin and α-catenin (Fig. (Fig.4B).4B). DN-Cad effectively inhibited the formation of adherens junctions, as judged by a strong decrease of β-catenin at cell-cell contacts and its accumulation in the cytoplasm (43) (Fig. 4C to F). The formation of tight junctions was not inhibited or was even slightly increased (44; data not shown). When confluent, DN-Cad cells exhibited a rounded phenotype (Fig. (Fig.4D"4D" and F"), and many cells were expelled into the medium.

FIG. 4.
Lateral recruitment of βPIX requires functional adherens junctions. (A) Western blot showing levels of E-cadherin and cadherin-6 in cells expressing DN-Cad and E-Cad-KD. Controls for DN-Cad expression were cells grown in the presence of DOX (+dox), ...

DN-Cad greatly impaired the lateral recruitment of βPIX (Fig. (Fig.4D′4D′ and F′). The lack of lateral βPIX recruitment was not due to downregulation of the Pak-PIX-GIT complex (Fig. (Fig.4G).4G). Rather, our results suggest that lateral βPIX recruitment depends on the formation of functional cadherin-based adherens junctions. This was also apparent when we disrupted adherens junctions by growing cells in low concentrations of calcium, which caused a rapid translocation of βPIX to the cytosol (Fig. (Fig.4H4H).

We next tested if lateral recruitment of βPIX was mediated by direct interaction with adherens junction components. Reciprocal immunoprecipitation experiments between components of the Pak-PIX-GIT complex (Pak1, βPX, and GIT1) and the adherens junction core complex (E-cadherin, β-catenin, α-catenin, and p120-catenin) failed to show any interaction between the two protein complexes (not shown). Furthermore, OptiPrep nonequilibrium gradient centrifugation of PNS from confluent MDCK cells showed that both complexes did not cofractionate and, in fact, almost entirely excluded each other (Fig. (Fig.5A).5A). Thus, whereas E-cadherin and β-catenin fractionated into a large low-density pool at the top of the gradient, representing the plasma membrane (49), and a much smaller high-density pool at the bottom of the gradient, GIT1 and βPIX were found exclusively in an intermediate-density fraction that did not overlap with either E-cadherin-containing pool. Also, solubilization of the plasma membrane with the detergent octylglucoside resulted in the redistribution of the low-density, membrane-bound fraction of E-cadherin and its interacting partner β-catenin but did not affect the behavior of GIT-PIX-containing complexes (Fig. (Fig.5B).5B). Interestingly, GIT-PIX-containing complexes substantially cofractionated with the focal adhesion proteins paxillin and FAK in both control and detergent-treated gradients (Fig. 5A and B). Collectively, these results indicate that βPIX does not localize to the plasma membrane via direct interaction with the adherens junction.

FIG. 5.
E-cadherin and βPIX-GIT1 complex do not cofractionate during gradient centrifugation. (A and B) Western blots of fractions collected after nonequilibrium gradient centrifugation of postnuclear supernatants run in the absence (A) or presence (B) ...

Inhibition of cadherin function impairs contact inhibition of proliferation.

Since we previously linked the lateral recruitment of βPIX to the establishment of contact inhibition of proliferation (53), we next analyzed this process in DN-Cad-expressing cells. DN-Cad strongly affected cell proliferation in confluent cells. Thus, in contrast to noninduced control cells (+DOX), which gradually established contact inhibition over time, proliferation in cells expressing DN-Cad (−DOX) remained constant over the same time course (Fig. (Fig.6A).6A). Consistent with its inability to establish contact inhibition, DN-Cad-expressing cells also failed to downregulate p27kip1 (Fig. (Fig.6B).6B). When we determined the growth rate in subconfluent cells, either by quantification of viable cells over a time course of 7 days or by BrdU incorporation assays of cells grown for 24 h at intermediate densities (~50% confluence), we found no differences between DN-Cad and control cells (not shown). Thus, the increased proliferation in DN-Cad cells was not due to intrinsic differences in growth rate of these cells but occurred only after cells reached confluence.

FIG. 6.
Perturbation of contact inhibition in DN-Cad-expressing cells by a β-catenin-TCF/LEF-independent mechanism. (A) Control cells grown in the presence of DOX (+dox) and cells expressing DN-Cad (−dox) were plated on Transwell filters. ...

The mechanisms that control contact inhibition downstream of E-cadherin are still poorly understood (30). Loss of E-cadherin has been proposed to increase β-catenin-TCF/Lef signaling by increasing the cytosolic levels of β-catenin (17, 40). However, even though β-catenin accumulated in the cytoplasm of DN-Cad-expressing cells (Fig. 4B and F), we did not observe any nuclear β-catenin (Fig. (Fig.4F)4F) or β-catenin-dependent TCF/Lef-mediated transcriptional activation (Fig. (Fig.6C).6C). TCF/Lef-mediated transcription was also not affected in cells expressing DN-Pak1 (not shown).

Cadherins can also engage in positive or negative cross talk with growth-promoting signaling pathways emanating from focal contacts (7). We therefore tested which signaling pathways were downregulated in contact-inhibited cells. Using phospho-specific antibodies against the activated forms of MEK, extracellular signal-regulated kinase (ERK), and Akt, we found that both the MEK-ERK (phospho-Ser218/Ser222-MEK1,2, phospho-Thr202,Tyr204-ERK1,2) and PI3K (phospho-Ser473-Akt) signaling pathways were strongly downregulated in confluent cells 6 days after plating. Pak1-mediated phosphorylation of MEK1 on Ser298, which primes MEK1 for activation (8), was also downregulated in these cells, albeit to a lesser extent (Fig. (Fig.7A7A).

FIG. 7.
Cell density-dependent downregulation of PI3K signaling is abrogated in cells expressing DN-Cad and Pak1-K299R. (A) Downregulation of MEK-ERK and PI3K signaling pathways in confluent cells. Western blots prepared from lysates of subconfluent (SC) (50% ...

We next analyzed if downregulation of these signaling pathways was affected in DN-Pak1- and DN-Cad-expressing cells. As expected, Pak1-dependent phosphorylation of MEK1-S298 was diminished in subconfluent DN-Pak1-expressing cells (not shown), but apart from that, we did not observe any effect on the downregulation of MEK-ERK signaling in confluent cells. In contrast, the downregulation of Akt activity in confluent cells was impaired in cells expressing DN-Cad or the DN-Pak1 forms Pak1-K299R (Fig. (Fig.7B)7B) and Pak1-AID (not shown). Both confluence-dependent downregulation of the PI3K signaling pathway and the impairment thereof by DN-Cad or DN-Pak1 were independent of the presence of serum (Fig. (Fig.7B),7B), but Akt activation was inhibited by the PI3K inhibitor LY294002 under all conditions (Fig. (Fig.7C7C and data not shown). Furthermore, the sustained Akt activity in confluent Pak1-K299R-expressing cells was dependent on its interaction with βPIX, as cells expressing Pak1-K299RΔPIX still downregulated Akt activity when confluent (Fig. (Fig.7D).7D). Confluent cells expressing this mutant were also able to recruit βPIX laterally and to establish contact inhibition of proliferation (53).

Cadherins and Pak1 control morphology and signaling from focal contacts.

In epithelial cells, PI3K can be activated in response to cell-cell adhesion, growth factor stimulation, and cell-matrix adhesion (18). Since our results indicated that stable cell-cell adhesion in MDCK cells actually downregulated PI3K signaling and that this occurred in a serum-dependent manner, we hypothesized that the impairment of PI3K downregulation in confluent DN-Cad and DN-Pak1 cells was mediated by alterations in cell-matrix signaling. Consistent with this, we found that the structure of focal contacts in DN-Cad and DN-Pak1 cells was grossly altered compared to that in controls. Thus, whereas noninduced confluent control cells formed fairly large structures resembling mature focal adhesions (Fig. 8A and C), focal contacts in induced confluent DN-Cad- and Pak1-K299R-expressing cells were much smaller and/or disorganized compared to those in controls (Fig. 8B and D). The colocalization of βPIX and focal contact markers was also changed: in confluent control cells, where the largest pool of membrane-associated βPIX was found laterally, the smaller, remaining pool of βPIX at the basal surface partially colocalized with focal contact markers such as paxillin and FAK (data for FAK not shown). Interestingly, this pool of βPIX was found in dot-like structures that overlapped with the more rod-shaped focal adhesions, and areas of colocalization were mostly found oriented toward the cell cortex, close to the lateral membrane (Fig. 8A and C; see high-resolution image in inset of Fig. Fig.8A′).8A′). In contrast, this highly structured organization of focal adhesions was disrupted in cells expressing DN-Cad or Pak1-K299R. Specifically, we observed large disorganized clusters in which paxillin and βPIX colocalized to a much greater extent, as well as small βPIX-containing spots that did not colocalize with paxillin (Fig. 8B" and D"). We repeated these experiments with cells costained for GIT1 and paxillin and found that βPIX and GIT1 behaved essentially the same (not shown). The aberrant phenotypes of βPIX-GIT1-containing focal adhesions were not observed in cells expressing Pak1-K299RΔPIX (Fig. (Fig.8E),8E), indicating that improper localization of the Pak1-βPIX-GIT1 complex, rather than inhibition of Pak1 kinase activity per se, determined the regulation of Pak1-βPIX-GIT1-specific focal adhesion turnover.

FIG. 8.
DN-Cad or Pak1-K299R expression induces similar changes in focal contacts. (A to E") Confluent control cells (A and C) and cells expressing DN-Cad (B), Pak1-K299R (D), or Pak1-K299RΔPIX (E) were plated on glass coverslips and stained for βPIX ...

Pak1 is part of a family of highly conserved group I Pak kinases, consisting of Pak1, Pak2, and Pak3. We found that MDCK cells express Pak1 and Pak2 (Fig. (Fig.8F),8F), but we could not detect Pak3, consistent with the mainly neuron-specific expression and functions of this isoform (20). The substrate specificities of the Pak1 and Pak2 isoforms are virtually identical (33), and recent evidence indicates that specific inclusion of Paks in signaling complexes at focal contacts is a determining factor in substrate specificity and functional differences of Pak1 and Pak2 (9). We therefore tested if knockdown of either or both Pak isoforms elicited similar phenotypes to that obtained with DN-Pak1. Nucleofection of two different siRNA oligonucleotides for either Pak isoform resulted in over 95% knockdown of Pak1 and/or Pak2. There was no compensatory upregulation of Pak2 in Pak1-depleted cells or vice versa (Fig. (Fig.8F).8F). Examination of focal contact morphology upon depletion of either or both Pak isoforms revealed strikingly different phenotypes. Loss of Pak1 resulted in a very sparse distribution of large and highly elongated paxillin-positive focal contacts that often no longer localized at the cell cortex but maintained the typical paxillin-βPIX colocalization seen in controls (compare Fig. 8G to G" to Fig. 8H to I"). In contrast, depletion of Pak2 resulted in a large number of paxillin-positive small focal contacts, which localized mostly at the cell cortex but did not substantially costain for βPIX (Fig. (Fig.8J).8J). A second siRNA for Pak2 yielded identical results (not shown). Finally, knockdown of both Paks resulted in very small focal contacts that localized both distally and in the middle of the basal surface. These structures for the most part stained positive for both paxillin and βPIX, although the staining intensity for βPIX was very low (Fig. (Fig.8K).8K). Importantly, knockdown of Pak1, Pak2, or both did not result in the extensive colocalization of paxillin and βPIX in large clusters seen in DN-Cad- or Pak1-K299R-expressing cells (Fig. (Fig.8B"8B" and D"). This suggests that sequestration of Pak in focal adhesions, rather than its loss, gives rise to the phenotypes observed in these cells, which is also consistent with a lack of phenotype in Pak1-K299RΔPIX-expressing cells (Fig. (Fig.8E8E).

The large focal contacts in DN-Pak1-expressing cells accumulate both Pak1 and βPIX (53). Since the GEF activity of βPIX depends on its interaction with Pak (23), we hypothesized that the increased colocalization may result in increased Rho GTPase signaling at these sites. Indeed, we found increased levels of active Rac1 (Rac1-GTP) in cells expressing DN-Cad, Pak1-K299R, and Pak1-AID but not in cells expressing an inactive Pak1-AID (Pak1-AID-L107F) or wild-type Pak (Fig. (Fig.99).

FIG. 9.
Elevated Rac1-GTP levels in cells expressing DN-Pak1 and DN-Cad. Cells expressing DN-Cad, Pak1-K299R, Pak1-AID, Pak1-AID, and wild-type Pak (Pak1-wt) were cultured for 5 days with or without DOX, and Rac1-GTP levels were determined as described in Materials ...

Since the alterations in focal adhesions induced by DN-Cad and Pak1-K299R were remarkably similar, they may be regulated by a common pathway. Since the absence of mature and highly organized focal adhesions coincided with the inability to downregulate PI3K signaling in these cells, we tested if PI3K was required for the regulation of βPIX localization in focal adhesions and for Pak1-βPIX-GIT1-specific focal contact turnover. Indeed, when we inhibited PI3K activity in cells 1 day (i.e., before lateral recruitment) after plating, using the PI3K inhibitor LY294002, βPIX accumulated in large focal adhesions (compare Fig. 10A and D) and failed to be recruited to areas of cell-cell contacts (compare Fig. 10B and B′ with Fig. 10E and E"). In addition, this treatment increased cell spreading (Fig. 10C and F). In contrast, LY294002 did not inhibit the lateral localization of β-catenin (Fig. 10B and B′) or E-cadherin (not shown). The accumulation of βPIX in focal adhesions was reversible, as washout of LY294002 restored lateral βPIX localization within 24 to 48 h (Fig. 10G to J′). These results indicate that PI3K activity is required for turnover of focal adhesions and the disassembly of βPIX from focal complexes. Finally, disruption of adherens junctions in confluent cells by lowering extracellular calcium in a “calcium switch assay” resulted in a quick removal of βPIX from the lateral membrane (compare Fig. 10K and L). Repletion of calcium rapidly recovered lateral βPIX, which was not inhibited by LY294002 (Fig. 10M and N). Together, the data in Fig. Fig.1010 suggest that the lack of lateral recruitment of βPIX in LY294002-treated cells is due to its inability to disassemble from focal adhesions.

FIG. 10.
Inhibition of PI3K inhibits lateral recruitment of βPIX by inhibiting turnover of focal adhesions. (A to B") Parental MDCK cells were plated at confluent densities on glass coverslips and grown for 4 days before being fixed. (D to J") Cells were ...


Our results provide evidence that establishment of contact inhibition of proliferation involves cadherin-dependent translocation of βPIX from focal contacts to areas of cell-cell contact. βPIX or other members of the Pak-PIX-GIT complex localize to focal contacts in subconfluent cells but are found at areas of cell-cell contact in confluent monolayers in epithelial and endothelial cells (1, 24, 41, 53). The functions of the complex at these different sites are likely distinct. At cell-matrix contact sites, it has well-established but incompletely understood roles in regulating the turnover of focal contacts and in cell migration (reviewed in reference 52). In addition, it has been implicated in growth-promoting signaling to ERK pathways, such as the mitogen-activated protein kinase (MAPK) and PI3K pathways (26, 36, 37, 50). Our data show that both pathways are downregulated after cells have formed confluent mature monolayers. Expression of DN-Cad or DN-Pak1 resulted in a lack of downregulation of PI3K but not ERK signaling and appeared to induce abnormal turnover of Pak-PIX-GIT-dependent focal contacts. Together, our data suggest that the lack of contact inhibition in these cells is due to deregulation of PI3K signaling, resulting from aberrant turnover of Pak-PIX-GIT-containing focal contacts.

Cadherin-based adherens junctions are required to recruit βPIX to areas of cell-cell contact, but recruitment is not mediated by direct interaction of the Pak-PIX-GIT complex with these junctions. These findings are consistent with the largely different kinetics of lateral recruitment of β-catenin and βPIX. Since adherens junctions and barrier function are not affected in DN-Pak1-expressing MDCK cells or in cells in which βPIX is knocked down, we concluded that lateral recruitment of the Pak-PIX-GIT complex not only requires the presence of adherens junctions but also depends on a normal turnover of Pak1-βPIX-GIT1-specific focal contacts. Depending on cell type and experimental context, Rac may both promote and inhibit the formation and maintenance of adherens junctions (31). Even though we did not find that ablation of βPIX or the presence of DN-Pak1 mutants affected adherens junctions, we cannot exclude a potential role for the Pak1-PIX-GIT complex in regulating cell-cell adhesion. Pak1 is required but not sufficient for Rac-dependent disruption of adherens junctions in keratinocytes (22). Furthermore, the Pak1-PIX-GIT complex decreases endothelial barrier function by mechanisms that rely on Pak1- or ERK-mediated activation of myosin light chain kinase and cell contractility (4, 41). Pak1 also mediates degradation of adherens junctions in endothelial cells by phosphorylating a conserved sequence in VE-cadherin that induces its internalization (11). Since this sequence is lacking in E-cadherin, a similar role for Pak1 in degradation of E-cadherin is unlikely.

Cadherins not only mediate lateral recruitment but also control the localization of the Pak1-βPIX-GIT1 complex at focal contacts. Signaling from both cell-cell and cell-matrix adhesion sites must be coordinated tightly to maintain epithelial function during cellular rearrangements, and there is considerable overlap in the functions of both adhesion complexes. Not only do they share several structural proteins, such as vinculin, and are both ultimately linked to the actin cytoskeleton, but they are also under positive and negative feedback control of highly similar signaling pathways. In particular, signaling pathways that involve Rho GTPases, PI3K, and Src family kinases are well known to be activated downstream of integrin- or cadherin-mediated adhesion and have been implicated in both positive and negative cross talk between the different adhesion complexes (2). Interestingly, since Rho GTPases, PI3K, and Src are all upstream regulators of the Pak-PIX-GIT complex (5, 52), the complex may be an important player in regulating cross talk between cell-cell and cell-matrix signaling. Our data are consistent with a role for PI3K and the Pak-PIX-GIT complex in cadherin-dependent maturation of focal contacts. While PI3K signaling remains high in cells that fail to recruit βPIX to the lateral membrane, PI3K is also required for lateral recruitment. Inhibition of PI3K results in accumulation of βPIX at focal contacts and blocks lateral recruitment of βPIX when added to cells 1 day after cell plating. Both Pak1 (26) and the PIX family member αPIX can bind PI3K or its lipid products (27, 51), and it is possible that lateral recruitment is mediated by a cadherin-dependent activation of PI3K (28). However, we consider this unlikely, as the rapid removal and re-recruitment of βPIX from the lateral membranes of mature monolayers in response to disruption and reformation of adherens junctions in a “calcium switch” experiment are not sensitive to PI3K inhibition. Furthermore, we were unable to demonstrate an interaction between PI3K and either Pak1 or βPIX in control MDCK or DN-Pak1- or DN-Cad-expressing cells.

We hypothesize that cadherins regulate the translocation of the Pak-PIX-GIT complex indirectly, most likely by controlling Pak1-βPIX-GIT1-specific focal complex turnover and maturation via a mechanism that also involves PI3K. Indeed, others have shown that platelet-derived growth factor (PDGF)-mediated activation of PI3K disrupts the organization of mature focal adhesion and that PI 3,4,5-triphosphatase (PI-3,4,5-P3) lipids are sufficient for this process (12). We propose the following model. When cell-cell adhesion is absent or of low abundance, cell-matrix signaling is dominant, and Pak-PIX-GIT complexes will localize at focal contacts. There, the complex is involved in feed-forward stimulation of PI3K signaling, likely by a mechanism that also involves Rac (47). In response to PI3K activation, turnover of Pak-PIX-GIT complexes at focal complexes is promoted, freeing the complex to become re-recruited either to focal contacts or to other sites, such as cell-cell contacts. At the same time, PI3K signaling stimulates cell proliferation. We thus propose that PI3K signaling in subconfluent cells serves a dual function. It promotes Pak1-βPIX-GIT1-specific focal complex turnover, and it stimulates proliferation, perhaps via activation of Akt. Indeed, we found that inhibition of PI3K leads to large, aberrant focal complexes in which βPIX accumulates (Fig. 10D) but that it also inhibits the increased Akt activation in confluent DN-Cad and Pak1-K299R cells (not shown). Next, as cells become confluent, signaling from adherens junctions will dominate, which will inhibit the feed-forward signaling at focal complexes. As a result, the turnover of Pak1-βPIX-GIT1-specific focal complexes is inhibited, thus abrogating PI3K signaling at these sites, which will lead to maturation of these complexes and a translocation of Pak-PIX-GIT complexes to cell-cell contacts. Deregulation of this mechanism, either by inhibiting focal contact dynamics with DN-Pak1 or by cadherin function, will promote signaling from focal contacts. Thus, even though adherens junctions are able to form, as in the case of DN-Pak1, PI3K signaling from cell-matrix adhesion sites will remain dominant, which will lead to continuing proliferation. Our model is consistent with several studies of different cancer cells showing that E-cadherin induced contact inhibition only in cells grown in three-dimensional aggregates in which cell-matrix interactions were prevented (38, 46), suggesting that, in some cancers, proliferation-promoting signaling from cell-matrix contacts has become dominant over the inhibitory signals from E-cadherin.

Knockdown of Pak1, Pak2, or both had markedly different effects on focal contact morphology but did lead to a sequestration of βPIX in these structures. Thus, it is likely that the effect of Pak1-K299R on cell matrix-mediated signaling is due mainly to its scaffolding effects, rather than to a loss of Pak kinase-dependent phosphorylation of specific substrates at focal contacts. It was previously shown that Pak1-AID inhibits both Pak1 and Pak2 (3). Since Pak1-AID- and Pak1-K299R (which contains an intact AID)-expressing cells exhibit similar phenotypes, it is likely that these DN-Pak1 mutants effectively sequester both Pak1 and Pak2 in focal contacts. The main difference between both Pak isoforms lies in the N-terminal region, which contains several protein interaction sites. Specifically, Pak1 contains five canonical proline-rich SH3 binding sites, whereas Pak2 contains only two (5). To elucidate the function of Pak kinases in contact inhibition, it would be interesting to know which Pak-interacting proteins mediate these effects, whether such interaction proteins have a preference for either Pak isoform, and/or whether binding to such interacting proteins depends on the localization of Pak kinases.

Numerous studies have implicated Pak kinases in epithelial rearrangements during development, and Pak's deregulation or overexpression is thought to underlie epithelial transformation during carcinogenesis (14, 21). Our work emphasizes the fact that understanding its spatial roles is crucial in understanding its function in the regulation of cellular behaviors such as proliferation, apoptosis, and cell migration.


We thank James Marrs for MDCK cells expressing DN-Cad and Kathleen Goss for help with the β-catenin promoter assay.

M.M.P.Z. was supported by the National Institutes of Health (GM076363) and the American Heart Association (0565309B).


[down-pointing small open triangle]Published ahead of print on 12 February 2010.


1. Audebert, S., C. Navarro, C. Nourry, S. Chasserot-Golaz, P. Lecine, Y. Bellaiche, J. L. Dupont, R. T. Premont, C. Sempere, J. M. Strub, A. Van Dorsselaer, N. Vitale, and J. P. Borg. 2004. Mammalian Scribble forms a tight complex with the betaPIX exchange factor. Curr. Biol. 14:987-995. [PubMed]
2. Avizienyte, E., and M. C. Frame. 2005. Src and FAK signalling controls adhesion fate and the epithelial-to-mesenchymal transition. Curr. Opin. Cell Biol. 17:542-547. [PubMed]
3. Beeser, A., Z. M. Jaffer, C. Hofmann, and J. Chernoff. 2005. Role of group A p21-activated kinases in activation of extracellular-regulated kinase by growth factors. J. Biol. Chem. 280:36609-36615. [PubMed]
4. Birukova, A. A., I. Malyukova, A. Mikaelyan, P. Fu, and K. G. Birukov. 2007. Tiam1 and betaPIX mediate Rac-dependent endothelial barrier protective response to oxidized phospholipids. J. Cell. Physiol. 211:608-617. [PubMed]
5. Bokoch, G. M. 2003. Biology of the p21-activated kinases. Annu. Rev. Biochem. 72:743-781. [PubMed]
6. Capaldo, C. T., and I. G. Macara. 2007. Depletion of E-cadherin disrupts establishment but not maintenance of cell junctions in Madin-Darby canine kidney epithelial cells. Mol. Biol. Cell 18:189-200. [PMC free article] [PubMed]
7. Chen, X., and B. M. Gumbiner. 2006. Crosstalk between different adhesion molecules. Curr. Opin. Cell Biol. 18:572-578. [PubMed]
8. Coles, L. C., and P. E. Shaw. 2002. PAK1 primes MEK1 for phosphorylation by Raf-1 kinase during cross-cascade activation of the ERK pathway. Oncogene 21:2236-2244. [PubMed]
9. Coniglio, S. J., S. Zavarella, and M. H. Symons. 2008. Pak1 and Pak2 mediate tumor cell invasion through distinct signaling mechanisms. Mol. Cell. Biol. 28:4162-4172. [PMC free article] [PubMed]
10. den Elzen, N., C. V. Buttery, M. P. Maddugoda, G. Ren, and A. S. Yap. 2009. Cadherin adhesion receptors orient the mitotic spindle during symmetric cell division in mammalian epithelia. Mol. Biol. Cell 20:3740-3750. [PMC free article] [PubMed]
11. Gavard, J., and J. S. Gutkind. 2006. VEGF controls endothelial-cell permeability by promoting the beta-arrestin-dependent endocytosis of VE-cadherin. Nat. Cell Biol. 8:1223-1234. [PubMed]
12. Greenwood, J. A., A. B. Theibert, G. D. Prestwich, and J. E. Murphy-Ullrich. 2000. Restructuring of focal adhesion plaques by PI 3-kinase. Regulation by PtdIns (3,4,5)-p(3) binding to alpha-actinin. J. Cell Biol. 150:627-642. [PMC free article] [PubMed]
13. Hansen, S., M. Zegers, M. Woodrow, P. Rodriguez-Viciana, P. Chardin, K. Mostov, and M. McMahon. 2001. Induced expression of Rnd3 is associated with transformation of polarized epithelial cells by the Raf-MEK-extracellular signal-regulated kinase pathway. Mol. Cell. Biol. 20:9364-9375. [PMC free article] [PubMed]
14. Harden, N. 2002. Signaling pathways directing the movement and fusion of epithelial sheets: lessons from dorsal closure in Drosophila. Differentiation 70:181-203. [PubMed]
15. Herzig, M., F. Savarese, M. Novatchkova, H. Semb, and G. Christofori. 2007. Tumor progression induced by the loss of E-cadherin independent of beta-catenin/Tcf-mediated Wnt signaling. Oncogene 26:2290-2298. [PubMed]
16. Jacinto, A., A. Martinez-Arias, and P. Martin. 2001. Mechanisms of epithelial fusion and repair. Nat. Cell Biol. 3:E117-E123. [PubMed]
17. Jeanes, A., C. J. Gottardi, and A. S. Yap. 2008. Cadherins and cancer: how does cadherin dysfunction promote tumor progression? Oncogene 27:6920-6929. [PMC free article] [PubMed]
18. Katso, R., K. Okkenhaug, K. Ahmadi, S. White, J. Timms, and M. D. Waterfield. 2001. Cellular function of phosphoinositide 3-kinases: implications for development, homeostasis, and cancer. Annu. Rev. Cell Dev. Biol. 17:615-675. [PubMed]
19. Korinek, V., N. Barker, P. J. Morin, D. van Wichen, R. de Weger, K. W. Kinzler, B. Vogelstein, and H. Clevers. 1997. Constitutive transcriptional activation by a beta-catenin-Tcf complex in APC−/− colon carcinoma. Science 275:1784-1787. [PubMed]
20. Kreis, P., and J. V. Barnier. 2009. PAK signalling in neuronal physiology. Cell Signal. 21:384-393. [PubMed]
21. Kumar, R., A. E. Gururaj, and C. J. Barnes. 2006. p21-activated kinases in cancer. Nat. Rev. Cancer 6:459-471. [PubMed]
22. Lozano, E., M. A. Frasa, K. Smolarczyk, U. G. Knaus, and V. M. Braga. 2008. PAK is required for the disruption of E-cadherin adhesion by the small GTPase Rac. J. Cell Sci. 121:933-938. [PubMed]
23. Manser, E., T. H. Loo, C. G. Koh, Z. S. Zhao, X. Q. Chen, L. Tan, I. Tan, T. Leung, and L. Lim. 1998. PAK kinases are directly coupled to the PIX family of nucleotide exchange factors. Mol. Cell 1:183-192. [PubMed]
24. Miura, K., J. M. Nam, C. Kojima, N. Mochizuki, and H. Sabe. 2009. EphA2 engages Git1 to suppress Arf6 activity modulating epithelial cell-cell contacts. Mol. Biol. Cell 20:1949-1959. [PMC free article] [PubMed]
25. Motti, M. L., D. Califano, G. Baldassarre, A. Celetti, F. Merolla, F. Forzati, M. Napolitano, B. Tavernise, A. Fusco, and G. Viglietto. 2005. Reduced E-cadherin expression contributes to the loss of p27kip1-mediated mechanism of contact inhibition in thyroid anaplastic carcinomas. Carcinogenesis 26:1021-1034. [PubMed]
26. Papakonstanti, E. A., and C. Stournaras. 2002. Association of PI-3 kinase with PAK1 leads to actin phosphorylation and cytoskeletal reorganization. Mol. Biol. Cell 13:2946-2962. [PMC free article] [PubMed]
27. Park, H. S., S. H. Lee, D. Park, J. S. Lee, S. H. Ryu, W. J. Lee, S. G. Rhee, and Y. S. Bae. 2004. Sequential activation of phosphatidylinositol 3-kinase, beta Pix, Rac1, and Nox1 in growth factor-induced production of H2O2. Mol. Cell. Biol. 24:4384-4394. [PMC free article] [PubMed]
28. Pece, S., M. Chiariello, C. Murga, and J. S. Gutkind. 1999. Activation of the protein kinase Akt/PKB by the formation of E-cadherin-mediated cell-cell junctions. Evidence for the association of phosphatidylinositol 3-kinase with the E-cadherin adhesion complex. J. Biol. Chem. 274:19347-19351. [PubMed]
29. Pece, S., and J. S. Gutkind. 2000. Signaling from E-cadherins to the MAPK pathway by the recruitment and activation of epidermal growth factor receptors upon cell-cell contact formation. J. Biol. Chem. 275:41227-41233. [PubMed]
30. Perrais, M., X. Chen, M. Perez-Moreno, and B. M. Gumbiner. 2007. E-cadherin homophilic ligation inhibits cell growth and epidermal growth factor receptor signaling independently of other cell interactions. Mol. Biol. Cell 18:2013-2025. [PMC free article] [PubMed]
31. Popoff, M. R., and B. Geny. 2009. Multifaceted role of Rho, Rac, Cdc42 and Ras in intercellular junctions, lessons from toxins. Biochim. Biophys. Acta 1788:797-812. [PubMed]
32. Qian, X., T. Karpova, A. M. Sheppard, J. McNally, and D. R. Lowy. 2004. E-cadherin-mediated adhesion inhibits ligand-dependent activation of diverse receptor tyrosine kinases. EMBO J. 23:1739-1748. [PubMed]
33. Rennefahrt, U. E., S. W. Deacon, S. A. Parker, K. Devarajan, A. Beeser, J. Chernoff, S. Knapp, B. E. Turk, and J. R. Peterson. 2007. Specificity profiling of Pak kinases allows identification of novel phosphorylation sites. J. Biol. Chem. 282:15667-15678. [PubMed]
34. Sasaki, C. Y., H. Lin, P. J. Morin, and D. L. Longo. 2000. Truncation of the extracellular region abrogates cell contact but retains the growth-suppressive activity of E-cadherin. Cancer Res. 60:7057-7065. [PubMed]
35. Sells, M. A., J. T. Boyd, and J. Chernoff. 1999. p21-activated kinase 1 (Pak1) regulates cell motility in mammalian fibroblasts. J. Cell Biol. 145:837-849. [PMC free article] [PubMed]
36. Shin, E. Y., K. S. Shin, C. S. Lee, K. N. Woo, S. H. Quan, N. K. Soung, Y. G. Kim, C. I. Cha, S. R. Kim, D. Park, G. M. Bokoch, and E. G. Kim. 2002. Phosphorylation of p85 beta PIX, a Rac/Cdc42-specific guanine nucleotide exchange factor, via the Ras/ERK/PAK2 pathway is required for basic fibroblast growth factor-induced neurite outgrowth. J. Biol. Chem. 277:44417-44430. [PubMed]
37. Slack-Davis, J. K., S. T. Eblen, M. Zecevic, S. A. Boerner, A. Tarcsafalvi, H. B. Diaz, M. S. Marshall, M. J. Weber, J. T. Parsons, and A. D. Catling. 2003. PAK1 phosphorylation of MEK1 regulates fibronectin-stimulated MAPK activation. J. Cell Biol. 162:281-291. [PMC free article] [PubMed]
38. St. Croix, B., C. Sheehan, J. W. Rak, V. A. Flørenes, J. M. Slingerland, and R. S. Kerbel. 1998. E-cadherin-dependent growth suppression is mediated by the cyclin-dependent kinase inhibitor p27(KIP1). J. Cell Biol. 142:557-571. [PMC free article] [PubMed]
39. Stewart, D. B., A. I. Barth, and W. J. Nelson. 2000. Differential regulation of endogenous cadherin expression in Madin-Darby canine kidney cells by cell-cell adhesion and activation of beta-catenin signaling. J. Biol. Chem. 275:20707-20716. [PubMed]
40. Stockinger, A., A. Eger, J. Wolf, H. Beug, and R. Foisner. 2001. E-cadherin regulates cell growth by modulating proliferation-dependent beta-catenin transcriptional activity. J. Cell Biol. 154:1185-1196. [PMC free article] [PubMed]
41. Stockton, R., J. Reutershan, D. Scott, J. Sanders, K. Ley, and M. A. Schwartz. 2007. Induction of vascular permeability: betaPIX and GIT1 scaffold the activation of extracellular signal-regulated kinase by PAK. Mol. Biol. Cell 18:2346-2355. [PMC free article] [PubMed]
42. Takahashi, K., and K. Suzuki. 1996. Density-dependent inhibition of growth involves prevention of EGF receptor activation by E-cadherin-mediated cell-cell adhesion. Exp. Cell Res. 226:214-222. [PubMed]
43. Troxell, M. L., Y. T. Chen, N. Cobb, W. J. Nelson, and J. A. Marrs. 1999. Cadherin function in junctional complex rearrangement and posttranslational control of cadherin expression. Am. J. Physiol. 276:C404-C418. [PubMed]
44. Troxell, M. L., S. Gopalakrishnan, J. McCormack, B. A. Poteat, J. Pennington, S. M. Garringer, E. E. Schneeberger, W. J. Nelson, and J. A. Marrs. 2000. Inhibiting cadherin function by dominant mutant E-cadherin expression increases the extent of tight junction assembly. J. Cell Sci. 113:985-996. [PubMed]
45. Van Aelst, L., and M. Symons. 2002. Role of Rho family GTPases in epithelial morphogenesis. Genes Dev. 16:1032-1054. [PubMed]
46. Vizirianakis, I. S., Y. Q. Chen, S. S. Kantak, A. S. Tsiftsoglou, and R. H. Kramer. 2002. Dominant-negative E-cadherin alters adhesion and reverses contact inhibition of growth in breast carcinoma cells. Int. J. Oncol. 21:135-144. [PubMed]
47. Weiner, O. D., P. O. Neilsen, G. D. Prestwich, M. W. Kirschner, L. C. Cantley, and H. R. Bourne. 2002. A PtdInsP(3)- and Rho GTPase-mediated positive feedback loop regulates neutrophil polarity. Nat. Cell Biol. 4:509-513. [PMC free article] [PubMed]
48. Wheelock, M. J., and K. R. Johnson. 2003. Cadherins as modulators of cellular phenotype. Annu. Rev. Cell Dev. Biol. 19:207-235. [PubMed]
49. Yeaman, C. 2003. Ultracentrifugation-based approaches to study regulation of Sec6/8 (exocyst) complex function during development of epithelial cell polarity. Methods 30:198-206. [PubMed]
50. Yin, G., Q. Zheng, C. Yan, and B. C. Berk. 2005. GIT1 is a scaffold for ERK1/2 activation in focal adhesions. J. Biol. Chem. 280:27705-27712. [PubMed]
51. Yoshii, S., M. Tanaka, Y. Otsuki, D. Y. Wang, R. J. Guo, Y. Zhu, R. Takeda, H. Hanai, E. Kaneko, and H. Sugimura. 1999. alphaPIX nucleotide exchange factor is activated by interaction with phosphatidylinositol 3-kinase. Oncogene 18:5680-5690. [PubMed]
52. Zegers, M. 2008. Roles of p21-activated kinases and associated proteins in epithelial wound healing. Int. Rev. Cell Mol. Biol. 267:253-298. [PMC free article] [PubMed]
53. Zegers, M. M., M. A. Forget, J. Chernoff, K. E. Mostov, M. B. ter Beest, and S. H. Hansen. 2003. Pak1 and PIX regulate contact inhibition during epithelial wound healing. EMBO J. 22:4155-4165. [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)