Cortactin is necessary for E-cadherin–mediated contact formation and actin reorganization

Classical cadherin adhesion molecules are fundamental regulators of tissue organization in metazoan organisms (Takeichi, 1991; Yap et al., 1997). One key function of these proteins is to mediate adhesive recognition between motile cells during tissue patterning. As exemplified by E-cadherin, this characteristically entails a morphological process where limited initial points of cell–cell contact become progressively extended into stable zones of adhesion (contact zone extension; Adams et al., 1998; Vasioukhin et al., 2000; Ehrlich et al., 2002). However, productive contact zone extension is not solely determined by the surface adhesive properties of cadherins. Contact formation also involves intimate functional cooperation between cadherins and the actin cytoskeleton (Adams and Nelson, 1998; Vasioukhin et al., 2000). Of note, it has recently become apparent that cadherin adhesions can mark sites for actin assembly. This cooperation may serve both to initially appose cell surfaces (Vasioukhin et al., 2000; Vaezi et al., 2002) and to productively extend contacts after initial adhesion is made (Ehrlich et al., 2002; Kovacs et al., 2002b).

The molecular mechanisms that link cadherins and actin assembly are now beginning to be identified. One such mechanism is likely to involve interaction between E-cadherin and the Arp2/3 actin nucleator complex. Recently, we found that cadherin adhesive ligation can recruit the Arp2/3 complex to the cell surface (Kovacs et al., 2002b). Strikingly, Arp2/3 preferentially localizes to sites of actin assembly in newly forming cadherin contacts. Such cadherin-directed localization of Arp2/3 provides an attractive potential mechanism to optimally focus the polymerization of actin in newly forming contacts, thereby allowing its force-generating capacity to be concentrated to drive the extension of those contacts.

The molecules that regulate actin assembly at cadherin contacts remain to be elucidated. Cortactin has recently emerged as a potentially critical regulator of actin assembly in a variety of contexts that entail dynamic control of membrane movements (Olazabal and Machesky, 2001; Weed and Parsons, 2001; Higgs, 2002). First discovered as an src-kinase substrate in RSV-transformed fibroblasts (Wu et al., 1991), cortactin is a multidomain actin-binding protein (Wu and Parsons, 1993; Weed and Parsons, 2001). Cortactin can interact directly with both Arp2/3 (via an NH₂-terminal acidic [NTA] domain) and F-actin (notably via the fourth of six tandem repeats located in the NH₂-terminal half of the molecule; Weed et al., 2000). Direct binding of cortactin activates Arp2/3-driven actin nucleation (Uruno et al., 2001; Weaver et al., 2001, 2002), and this is enhanced when WASp-interacting protein associates with the cortactin SH3 domain (Kinley et al., 2003). Cortactin also inhibits disassembly of Arp2/3-generated actin filaments, an effect that can potentially stabilize the cortical actin network (Weaver et al., 2001). These properties of cortactin prompted us to examine its role in cadherin contact formation. We now report that cortactin is necessary for cadherin-directed actin assembly and contact zone extension.

The capacity for cortactin to participate in cadherin function was first suggested by its immunofluorescent localization in cultured cell monolayers that contain endogenous E-cadherin. MDCK (Fig. 1, untreated), as well as NMuMG and MCF-7 cells (unpublished data), displayed extensive cortactin staining in cadherin-based cell–cell contacts. Moreover, cortactin was rapidly recruited to MDCK cell–cell contacts as cadherin adhesion was restored after being disrupted by chelation of extracellular calcium (Fig. 1). However, cortactin did not always precisely colocalize with E-cadherin, typically being distributed somewhat less continuously than E-cadherin in cell–cell contacts. Cortactin also showed vesicular and perinuclear staining, consistent with a reported role in cellular trafficking (Kaksonen et al., 2000; Lynch et al., 2003). These data suggested that cortactin could be rapidly recruited to newly forming cadherin contacts.

To determine if cadherin adhesion was responsible for recruiting cortactin to cell–cell contacts, we first compared the localization of cortactin in CHO cells stably transfected with human E-cadherin (hE-CHO cells). Parental CHO cells, which do not express detectable classical cadherins (Kovacs et al., 2002b), formed loose contacts with one another as detected by phalloidin staining (Fig. 2 A), but these did not accumulate cortactin. In contrast, cortactin frequently concentrated with E-cadherin in contacts between hE-CHO cells (Fig. 2 A). Like MDCK cells, cortactin staining was often less extensive than the cadherin staining.

Then, we used a dimeric recombinant ligand containing the complete ectodomain of human E-cadherin fused to the Fc region of IgG (hE/Fc) to test if cadherin homophilic adhesion was sufficient to recruit cortactin to the plasma membrane (Fig. 2 B). This and similar reagents (Noren et al., 2001; Lambert et al., 2002) support adhesion and cadherin-activated cell signaling (Noren et al., 2001; Kovacs et al., 2002a,b). Latex beads coated with hE/Fc attached efficiently to the dorsal surfaces of hE-CHO cells, thereby providing spatially defined adhesive signals (Goodwin et al., 2003). Prominent cortactin accumulation was detected at the sites of bead binding (Fig. 2 B), where it colocalized with β-catenin (marking the cellular cadherin complex). In contrast, neither cortactin nor β-catenin (unpublished data) accumulated at sites of contact between cells and Con A–coated beads, despite effective binding of beads to cell surfaces (Fig. 2 B). Although 95% of hE/Fc-coated beads (n = 40) recruited cortactin, only 5% of Con A–coated beads (n = 40) showed any cortactin staining. This suggested that homophilic cadherin binding sufficed to recruit cortactin to sites of adhesion.

To better define the precise pattern of cortactin recruitment in cadherin contacts, we used total internal reflection fluorescence (TIRF) microscopy to visualize cortactin in cells adherent to hE/Fc-coated substrata (Fig. 3). In this assay, the cell–substratum interface constitutes the zone of contact and, as cells adhere, they progressively extend their zones of contact by protruding cadherin-based lamellipodia (Fig. 3 and Video 1; Kovacs et al., 2002a,b). Although not designed to reproduce all events likely to arise when native cell surfaces are brought together, these planar spreading assays have the analytic advantage of allowing us to isolate dynamic cellular events that arise as specific responses to cadherin ligation, independent of any potential juxtacrine signals (Noren et al., 2001; Kovacs et al., 2002a,b; Yap and Kovacs, 2003). Moreover, many salient features displayed by cells in these assays, including the spatial confinement of Rac and PI3-kinase signaling to the outer, extending margins of the contact (Kovacs et al., 2002a), are also observed when migrating MDCK cells establish productive contacts with one another (Ehrlich et al., 2002). Combined with the ability of TIRF microscopy to image molecules located within 100 nm of the cell substrate boundary (Steyer and Almers, 2001), these assays thus provided the opportunity to examine dynamic events located at the cadherin adhesive interface itself.

TIRF microscopy of hE-CHO cells adherent to hE/Fc-coated substrata revealed that cortactin stained in prominent bands at the very outer margins of the contact zones (Fig. 3, A and A′). This confirmed that the concentration of staining at outer margins identified by confocal microscopy (Fig. 4) reflects its localization at the contact interface itself. As seen in monolayers, cortactin also stained in the perinuclear region as well as in puncta throughout the bases of lamellipodia. In contrast, relatively little cortactin staining was seen in the bodies of the lamellipodia proximal to the outer margins (Fig. 3 A′). However, by TIRF illumination, β-catenin (marking the cadherin complex) distributed diffusely in clusters throughout the lamellipodia (Fig. 3, B and B′) and contact zones generally (Fig. 3 B). Therefore, the accumulation of cortactin at the outer margins, and its relative paucity immediately proximal to the margins, was not due to changes in the proximity of the cell membrane to the adhesive substratum.

The prominent localization of cortactin at the outer margins was noteworthy because these outer margins represent regions where the zones of contact are being actively extended and remodeled. Indeed, time-lapse TIRF imaging of transiently transfected hE-CHO cells further revealed that GFP-cortactin rapidly accumulated at the outer margins of lamellipodia as cells extended their contacts on hE/Fc-substrata (Fig. 3 C; Video 1). Characteristically, GFP-cortactin appeared to consistently accumulate wherever the outer margin was extended. Again, despite increased perinuclear and cytoplasmic fluorescence due to transient expression of the transgene, accumulation of GFP-cortactin at the outer margins was associated with relative depletion more proximally in the extensions. Thus, cortactin localization at the outer margins appeared to be dynamic, coinciding with outward movement of the outer margin.

This suggested strongly that cortactin might preferentially associate with the outer margins where contact zones were undergoing extension. As a further test of this notion, we used TIRF to assess the patterns of fluorescence recovery in the adhesive interface after photobleaching of GFP-cortactin. As shown in Fig. 3 (D–F) and in Video 2 , fluorescence first returned at the very outer margins of the cadherin-based lamellipodia, being readily detectable within 20 s after photobleaching. Measurement of fluorescence intensity showed clear recovery of fluorescence both at the very outer margins and within the lamellipodium proximal to the outer margin itself (Fig. 3 E). However, fluorescence began to increase earlier, and to a greater degree, at the outer margin than in more proximal regions of lamellipodia (Fig. 3 E), suggesting that this was a site for preferential recovery of GFP-cortactin. As a further test, we compared the distribution of fluorescence recovery in line scans through the photobleached area at various times after photobleaching (Fig. 3 F). This analysis showed that although fluorescence progressively increased throughout the photobleached area, recovery was most pronounced at the very outer margin. Together, these data indicate that cortactin accumulated preferentially at the outer margins of cadherin-based lamellipodia where contact zones were being actively extended.

A major target of cortactin is the Arp2/3 actin nucleator complex (Uruno et al., 2001; Weaver et al., 2001). Consistent with this, we found that cortactin colocalized with Arp2/3 at sites of homophilic cadherin adhesion. In bead adhesion assays, cortactin accumulated with transiently expressed GFP-Arp3 at sites of adhesion to hE/Fc-coated beads (Fig. 2 B), but not after binding of Con A–coated beads (unpublished data). In planar spreading assays, cortactin also colocalized precisely with transiently expressed GFP-Arp3 at the outer margins of cadherin-based lamellipodia (Fig. 4). In both these assays, cortactin coaccumulated with F-actin—at the sites of binding to cadherin-coated beads (Fig. 2 B) and in prominent bands at the outer margins of cadherin-based lamellipodia, which we have previously shown to correspond to sites of active actin assembly (Fig. 4; Kovacs et al., 2002b).

To test whether cortactin could participate in actin organization at cadherin adhesive contacts, we examined the effect of cortactin mutants on F-actin accumulation around hE/Fc-coated beads (Fig. 5). When transiently expressed, these cortactin mutants generated polypeptides of the predicted molecular weight and were readily detected by immunofluorescence (Fig. S1). F-actin accumulated prominently at the sites of adhesion between control cells and hE/Fc-coated beads (Fig. 5 B; hE/Fc, untransfected). Characteristically, intense F-actin staining was observed in a distinct circumferential pattern around the hE/Fc-coated beads (Fig. 2 and Fig. 5 B). In contrast, no circumferential actin staining was seen upon binding of Con A beads, which showed only occasional actin cables that appeared to originate and terminate in different focal planes than the beads (Fig. 5 B, Con A). Importantly, transient expression of the CA domain of N-WASp (Fig. 5 B, GFP-CA) largely abolished circumferential phalloidin staining around cadherin-coated beads. Because this fragment is a potent inhibitor of Arp2/3 activity (Rohatgi et al., 1999), it indicated that Arp2/3 activity was necessary for cadherin-based actin accumulation upon bead binding.

Circumferential actin accumulation around hE/Fc-beads was also significantly inhibited in cells expressing either cortactin Δ4 (lacking the F-actin–binding site) or cortactin ΔNTA (unable to bind Arp2/3; Fig. 5 B). Semi-quantitative assessment of F-actin accumulation around beads showed that expression of either cortactin mutant inhibited actin accumulation around hE/Fc beads to a similar degree as GFP-CA (Fig. 5 C). In contrast, a mutant in which the three tyrosines phosphorylated by src were mutated to phenylalanine (cortactin 3Y, which can inhibit cell motility; Huang et al., 1998) did not affect actin accumulation at cadherin beads. These data strongly suggest that cortactin participates in directing actin assembly at these cadherin contacts, consistent with recent evidence that cortactin participates in Arp2/3-mediated actin assembly (Weaver et al., 2001; Krueger et al., 2003).

Actin assembly by Arp2/3 is commonly implicated in generating the forces necessary for cell surface protrusion (Pollard et al., 2000). This led us to test whether cortactin might participate in cadherin-based contact zone extension. First, we examined whether cortactin mutants affected the ability of cells to form cadherin-based lamellipodia using an assay that measures the dimensions of outer margins as a proportion of the total cell periphery (Fig. 6; lamellipodial index, LI). We found that transient expression of mutants that disrupt either the ability of cortactin to bind F-actin (Δ4) or its ability to interact with the Arp2/3 complex (ΔNTA) reduced the LI of hE-CHO cells by >50% compared with untransfected controls (Fig. 6). In contrast, neither the 3Y mutant nor a mutant that consisted only of the internal repeats region (Fig. 6, repeats) affected cadherin-based contact zone extension compared with untransfected control cells (Fig. 6).

As an independent loss-of-function approach, we used the pSUPER system for RNA interference (RNAi; Brummelkamp et al., 2002) to suppress cortactin expression. Transient expression of pSUPER-Cort significantly reduced cortactin levels detected in hE-CHO cells by immunoblotting, but did not affect cellular E-cadherin or β-catenin levels (Fig. 7 A). Furthermore, in planar spreading assays, cells with reduced cortactin levels could be clearly detected by immunofluorescence (Fig. 7 B). Cadherin-based lamellipodial extension was reduced by ∼70% in these pSUPER-treated cells compared with control hE-CHO cells cotransfected with an empty pSUPER vector and GFP, which efficiently formed lamellipodia on hE/Fc (Fig. 7 B).

Then, we examined whether cortactin depletion affected the ability of cadherin-containing cells to reform cell–cell contacts after depletion of extracellular calcium (Fig. 8). For these experiments, we used NMuMG mouse epithelial cells that contained endogenous E-cadherin. Although pSUPER-Cort did not produce effective knock-down in these cells, synthetic RNA duplexes reduced cortactin levels by ∼50–60% (Fig. 8 A). Compared with untransfected cells, RNAi-treated cells were significantly slower to reestablish cadherin-based contacts after replacement of extracellular Ca²⁺ (Fig. 8, B and C). Together, these findings demonstrate that cortactin plays an important role in the ability of cells to extend nascent cadherin-based adhesive contacts.

To further assess the function of cortactin, we examined the effects of perturbing cortactin function in MDCK cells, a classic polarized epithelium. Expression of the cortactin Δ4 or ΔNTA mutants dramatically altered cell morphology (Fig. 9). Cells expressing these mutants displayed irregular outlines and long cellular extensions, whereas untransfected cells or cells expressing the vector alone showed the regular cobblestone morphology characteristic of this cell line. Moreover, E-cadherin and β-catenin (unpublished data) failed to accumulate at the sites of contact between these cells and surrounding untransfected cells. In contrast, cells transfected with the Repeats or 3Y cortactin mutants showed no changes in cellular morphology and concentrated E-cadherin in cell–cell contacts, exactly as did control cells.

As a first step toward characterizing the molecular basis for cortactin recruitment to E-cadherin, we tested the capacity for E-cadherin and cortactin to interact biochemically. As shown in Fig. 10 A, cortactin and E-cadherin coimmunoprecipitated from monolayers of hE-CHO cells, but not from cadherin-negative parental CHO cells. Cortactin also coimmunoprecipitated E-cadherin from MDCK cell monolayers, both before and during recovery after depletion of extracellular Ca²⁺ (Fig. 10 B).

Further, we found that E-cadherin coimmunoprecipitated from hE-CHO cells plated onto hE/Fc, but not from cells plated onto poly-l-lysine (Fig. 10 C), suggesting that sustained cadherin ligation was sufficient to cause cortactin and E-cadherin to interact biochemically. We estimate that ∼1–2% of total cellular E-cadherin was found in the cortactin immunoprecipitates (Fig. 10 C). However, this is likely to significantly underestimate the true stoichiometry of the cortactin–cadherin interaction because the cortactin immunoprecipitates probably contain a significant amount of cortactin not complexed with cadherin. Finally, the association between cortactin and E-cadherin did not depend on actin filament integrity because the efficiency of coimmunoprecipitation was unaffected by latrunculin A (Fig. 10 D) added to the intact cells or after lysis.

It is increasingly apparent that key aspects of cadherin biology entail close functional cooperation between the surface-adhesive action of cadherin molecules and the force-generating capacity of actin assembly. Our current data identify a necessary role for cortactin in this cooperation, based on four observations: (1) cortactin and Arp2/3 coaccumulated in dynamic cadherin contacts, notably those at the outer margins of extending contact zones; (2) cortactin mutants effectively inhibited F-actin accumulation in cadherin contacts, a process that requires Arp2/3 activity; and (3) perturbing cortactin function profoundly affected the ability of cells to extend cadherin-adhesive contacts and (4) disturbed the biogenesis of adherens junctions in MDCK cells, accompanied by gross distortions of epithelial cell morphology.

Actin assembly has the capacity both to drive cell surfaces together to initiate contacts (Vasioukhin et al., 2000) and to promote extension of cadherin-based cell contacts upon one another after contacts are first made (Kovacs et al., 2002b). In the latter case, E-cadherin ligation itself is likely to be an active participant, by directing where actin assembly concentrates in contact zones. We found that cortactin mutants profoundly inhibit actin accumulation at adhesive contacts with cadherin-coated beads. We chose to use bead assays in these experiments because they provide spatially defined, cadherin-specific adhesive cues that are independent of cell spreading. Similar approaches demonstrated that G-actin incorporates at the adhesive contact with N-cadherin–coated beads (Lambert et al., 2002), and we confirmed that F-actin accumulation at bead adhesions was inhibited by an N-WASp fragment that effectively inhibits Arp2/3-mediated actin assembly in vitro (Rohatgi et al., 1999). Therefore, our results suggest strongly that cortactin activity is necessary for the Arp2/3-dependent actin assembly that occurs in response to E-cadherin homophilic ligation.

In keeping with a role for actin assembly in extending adhesive contacts, we found that the ability of cells to extend cadherin contacts in planar spreading assays was dramatically reduced when cortactin activity was perturbed, either by expression of dominant-negative cortactin mutants or by RNAi-mediated knock-down of endogenous cortactin. Furthermore, depletion of cortactin in NMuMG cells significantly reduced the efficiency with which cells reestablished contacts with one another. Together, these findings indicate that cortactin plays a key role in cadherin-based contact zone extension. In separate analyses, we also recently established that Arp2/3 activity is necessary for extension of cadherin adhesive contacts in these assays (unpublished data). Thus, we interpret these findings to indicate that cortactin is necessary for Arp2/3-mediated actin assembly to effectively drive cell surfaces together during contact zone extension.

In vitro, cortactin can increase net actin assembly by directly or indirectly stimulating the catalytic activity of the Arp2/3 complex (Uruno et al., 2001; Weaver et al., 2001, 2002), as well as stabilizing actin filaments at a post-nucleation step (Weaver et al., 2001). Any of these actions, individually or in concert, may account for the contribution of cortactin to the cadherin-mediated actin assembly that we observed. Notably, our analysis of cortactin mutants suggests that both the Arp2/3-binding and F-actin–binding domains of the molecule are necessary for actin assembly at cadherin contacts. Hence, the ability of cortactin to act as a multidomain scaffold, simultaneously linking Arp2/3 and actin filaments, may be a key to its role in cadherin-directed actin assembly. The inhibitory effect of the cortactin ΔNTA mutant might then be explained by its ability to displace endogenous cortactin from filaments (because this mutant retains the F-actin–binding site), whereas the Δ4 mutant would compete with endogenous cortactin for Arp2/3, but be incapable of binding F-actin.

Ultimately, for cortactin to potentiate Arp2/3 activity, this protein must also be strictly recruited to sites of actin assembly at the cell surface. Indeed, we found that cortactin colocalizes with the Arp2/3 complex exactly where cadherin adhesion induces actin assembly—around nascent contacts with cadherin-coated beads and at the outer margins of contact zones in planar adhesion assays. Furthermore, after photobleaching, GFP-cortactin appeared to preferentially recruit to the outer margins of cadherin-based lamellipodia where contact zones were being actively extended (Kovacs et al., 2002b). Such selective accumulation might reflect either preferential delivery of cortactin to these regions and/or selective retention at these sites. Irrespective of the mechanism, this finding indicates that cortactin is recruited to specific subregions within individual cadherin contacts, suggesting that the adhesive contact zones are not biochemically or functionally uniform structures (Ehrlich et al., 2002; Kovacs et al., 2002b).

Our data further indicate that cortactin recruitment occurs as a consequence of cadherin homophilic ligation and, indeed, is associated with a biochemical interaction between cortactin and the cadherin adhesive complex. Most notably, presentation of a cadherin-specific adhesive signal (on beads or planar substrata) was sufficient to recruit cortactin to adhesive sites and induce cortactin to coimmunoprecipitate with E-cadherin. In contrast, nonspecific adhesive ligands did not recruit cortactin to the membrane, nor did they allow the cadherin–cortactin biochemical interaction to be sustained.

Although we have yet to elucidate the molecular mechanism that links cortactin to the cadherin adhesive complex, our data argue against it being a constitutive interaction. In planar spreading assays, much E-cadherin did not colocalize with cortactin, and in our current experiments only a small proportion of total cellular E-cadherin was found in the cortactin immunoprecipitates. Similarly, integrin-based cell migration also entails a functionally important interaction between small subpopulations of Arp2/3 and vinculin (DeMali et al., 2002). Although cortactin may interact with some members of the p120-catenin family (Martinez et al., 2003), this suggests that the contribution of cortactin to cadherin function involves a specific pool of cadherin molecules, presumably that is responsible for recruiting Arp2/3. Similarly, we observed earlier that only a subpopulation of the cellular pools of cadherin and Arp2/3 interacted biochemically during contact extension (Kovacs et al., 2002b).

Instead, it is likely that recruitment of cortactin to cadherin contacts is signal regulated, consistent with emerging evidence that multiple signals determine both Arp2/3 activity (Machesky and Insall, 1999) and cortactin localization (Head et al., 2003). A good candidate here is Rac, which clearly stimulates cortactin recruitment in motile cells (Weed et al., 1998). Moreover, Rac is directly activated by cadherin ligation (Nakagawa et al., 2001; Kovacs et al., 2002a), and is likely to be spatially confined to the outer margins in extending contact zones (Ehrlich et al., 2002; Kovacs et al., 2002a), exactly corresponding to the sites where cortactin recruits in our analyses. Spatio-temporal localization of signals may then be an important factor that directs cortactin to specific subregions within individual adhesive contact zones.

In summary, we have identified cortactin as a novel determinant of cadherin–actin cooperativity during adhesive contact formation. We envisage that cortactin acts in newly forming contacts to promote Arp2/3-mediated cell surface protrusion, thereby driving the extension of nascent contacts to form stable zones of adhesion. However, the impact of cortactin is not confined to the early stages of contact formation. Cortactin mutants significantly inhibited the accumulation of E-cadherin in contacts between MDCK cells, a hallmark of adherens junctions assembly, accompanied by gross distortions of epithelial cell morphology (residual cortactin likely explains why morphologic changes were less apparent in RNAi-treated NMuMG cells). Interestingly, the cortactin mutants that most effectively perturbed cadherin-directed actin assembly and contact zone extension (Δ4 and ΔNTA) were also the most potent in disrupting junctional assembly and epithelial morphology. These effects of cortactin on later stages of epithelial biogenesis may therefore reflect a requirement for efficient contact zone extension early in the process of contact formation. However, cortactin may also participate in other aspects of junctional biogenesis. For example, cortactin also binds directly to ZO-1, a cytoplasmic scaffolding protein of the tight junction in both invertebrate and vertebrate cells (unpublished data; Katsube et al., 1998), which can also associate with adherens junctions. Given its potential to support multiple intermolecular interactions (Weed and Parsons, 2001), cortactin may then act as a signal integrator in cadherin contacts, coupling adhesion to actin dynamics, cell signaling, and junctional assembly.

hE-CHO cells, hE/Fc purification, and preparation of hE/Fc-coated substrata and latex beads (6-μm-diam) have been described previously (Kovacs et al., 2002a,b). Transient transfections were performed with LipofectAMINE™ (Invitrogen) according to the manufacturer's instructions. MDCK cells were cultured in DME with 10% FCS and 2 mM glutamine; NMuMG cells in DME with 10% FBS and 10 μg/ml insulin.

Primary antibodies were as follows: (1) a pAb directed against the ectodomain of human E-cadherin was raised in rabbits using hE/Fc as an immunogen. In Western blots, this pAb recognized a single 116-kD polypeptide band in lysates from MCF-7 and hE-CHO cells, but not in lysates from parental CHO cells that lack E-cadherin; (2) mouse mAb against the cytoplasmic tail of human E-cadherin (Transduction Laboratories); (3) mAb HECD1 against human E-cadherin (provided by Dr. M. Wheelock [University of Nebraska, Omaha, NE], with the permission of Dr. M. Takeichi [RIKEN CDB, Kobe, Japan]); (4) mouse mAb 3B8 directed against canine E-cadherin (a gift from Dr. Warren Gallin, University of Alberta, Edmonton, Alberta, Canada); (5) rat mAb DECMA-1 against mouse E-cadherin (Sigma-Aldrich); (6) rabbit pAb against the NH₂ terminus of Xenopus β-catenin (a gift of Dr. Barry Gumbiner, University of Virginia, Charlottesville, VA); (7) mouse mAb 4F11 against cortactin (Wu et al., 1991); (8) rabbit pAb C2 raised against chicken cortactin; (9) anti-FLAG mAb M5 (Sigma-Aldrich); (10) rabbit anti-GFP pAb (Molecular Probes, Inc.); and (11) anti-GFP mAb (Roche). F-actin was identified using Alexa^® 488– or Alexa^® 568–labeled phalloidin (Molecular Probes, Inc.). Secondary antibodies were species-specific antibodies conjugated with Alexa^® 350, Alexa^® 488, or Texas red (Molecular Probes, Inc.).

The pEGFP-CA construct bearing the CA fragment of N-WASp was a gift of Drs. H. Miki and T. Takenawa (University of Tokyo, Tokyo, Japan). GFP-Arp3 (constructed by Dr. Matt Welch) was a gift from Dr. D. Schaefer (Washington University, St. Louis, MO). GFP-cortactin and the cortactin mutants Δ4, Repeats, and 3Y were described earlier (Weed et al., 2000; Kinley et al., 2003). The cortactin ΔNTA expression construct was created by PCR amplification of mp85.L7 (Miglarese et al., 1994), with a 5′ primer flanking codon 85 containing a KpnI restriction site and a 3′ primer flanking the stop codon (546) containing an EcoRI restriction site. The amplified PCR product was digested with KpnI and EcoRI and subcloned in frame into KpnI–EcoRI-digested pcDNA3FLAG2AB. The final construct was tested by DNA sequencing.

To generate the pSUPER-Cort construct, the pSUPER vector was digested with BglII and HindIII and then ligated with the annealed oligonucleotides 5′-gatccccCACATCAACATTCACAAGCttcaagagaGCTTGTGAATGTTGATGTGtttttggaaa-3′ and 5′-agcttttccaaaaaCACATCAACATTCACAA-GCtctcttgaaGCTTGTGAATGTTGATGTGggg-3′ (cortactin sequences are in uppercase, corresponding to bases 284–302 in the murine cortactin cDNA; Miglarese et al., 1994). Oligonucleotides were HPLC purified and 5′-phosphorylated before annealing. The pSUPER-Cort sequence was verified by DNA sequencing.

Synthetic RNA duplexes based on the sequences 5′-AAAGCTTCGAGAGAATGTCTT-3′ and 5′-AAGACTGAGAAGCATGCCTCC-3′ were prepared using the Silencer™ kit (Ambion). Cells were transfected with a cocktail of both duplexes using siPORT™ Amine (Ambion) as per the manufacturer's instructions.

Samples were prepared for immunofluorescence analysis as described previously (Kovacs et al., 2002a,b). Fixed material was mounted in 1% N-propyl-gallate in 50% glycerol:PBS for epi-illumination and laser-scanning confocal microscopy; glycerol was omitted from the mounting medium for TIRF. Fixed material was examined by epi-illumination using microscopes (model AX70 or IX81; Olympus) with 60×, 1.4 NA or 100×, 1.4 NA objectives (Olympus), and was imaged with cameras (Orca 1 or Orca-ER; Hamamatsu). TIRF imaging was performed using a microscope (model IX81; Olympus) and an objective (60×, 1.45 NA; Olympus) illuminated with a 10-mW argon laser. For live-cell imaging, cells were incubated in phenol red–free HBSS at 37°C. All acquisition and movies were controlled using MetaMorph^® software (Universal Imaging Corp.). Photobleaching was performed by closing the iris field diaphragm in the epi-illumination pathway and illuminating with a 100-W Hg lamp for 10 s. Laser-scanning confocal microscopy was performed using a scanning system (MRC 600; Bio-Rad Laboratories) mounted on a microscope (Axioskop; Carl Zeiss MicroImaging, Inc.) equipped with a Plan-Apochromat 63× objective (Carl Zeiss MicroImaging, Inc.) running under Bio-Rad Laboratories software. Some epi-illumination images were deconvolved from z-stacks using three-dimensional blind deconvolution (Autoquant). Movie images were compiled and converted to QuickTime stacks using ImageJ. Figures were assembled for presentation in Adobe Photoshop^®.

To quantitate cadherin-based contact formation in planar spreading assays, phalloidin-stained images were analyzed using MetaMorph^® software. For each cell, the lamellipodial index was calculated as follows: lamellipodial outer margins were identified by their broad phalloidin staining, were measured, and the total length of outer margins in each cell was expressed as a percentage of the cell perimeter for that cell. Contact formation in calcium jump experiments was quantitated by measuring the lengths of individual cadherin contacts (defined as lines of continuous cadherin staining between cell vertices). As a semiquantitative estimate of actin accumulation around hE/Fc-coated beads, the degree of phalloidin staining around beads was assessed using a scoring system based on the proportion of the circumference of the bead showing intense F-actin accumulation. 4+, complete circumferential F-actin accumulation; 3+, ∼75% of the perimeter; 2+, ∼50% perimeter; 1+, ∼25%; 0, no actin accumulation.

Cells were lysed in 1 ml lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 50 mM NaF, 2 mM sodium vanadate, 0.1% BSA, and complete protease inhibitors [Roche]). Protein complexes were immunoprecipitated with either a polyclonal E-cadherin antibody or a C2 pAb directed against cortactin bound to protein A–agarose beads and were separated by SDS-PAGE. Immune complexes were blotted for E-cadherin and cortactin. Blots were scanned and compiled in Adobe Photoshop^® (v. 6).

Online supplemental material includes expression characterization of the cortactin mutants used in this work (Fig. S1), graphical illustration of the regions chosen to analyze FRAP (Fig. S2), and the videos of GFP-cortactin localization imaged by TIRF (Videos 1 and 2) that are illustrated in Fig. 3.

We thank all our aforementioned colleagues for the gifts of antibodies and reagents; Drs. Tom Fleming and Bhav Seth for sharing unpublished data; Dr. Joe Loureiro for critical reading of the manuscript; Julie Head for her assistance in generating and validating cortactin plasmids; Roger Daly, Gillian Lehrbach, and the Garvan Institute Cell Culture facility (Sydney, Australia) for NMuMG cells; Teresa Munchow for assiduous tissue culture support; Sandrine Roy for transfection advice; and all our colleagues in the lab for their support, encouragement, and advice.

The work in Australia was supported by grants to A.S. Yap from the National Health and Medical Research Council (Australia), Wellcome Trust, and Human Frontiers Science Program. S.A. Weed is supported by grant K22 DE 14364 from the National Institute of Dental and Craniofacial Research. Support is also provided by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (RO1 DK061397) to A.S. Fanning and James M. Anderson; and A.S. Yap is a Wellcome Trust Senior International Medical Research Fellow.