α-Catenin Uses a Novel Mechanism to Activate Vinculin*

Background: In cell-matrix adhesions, vinculin is activated by talin and actin, but the identity of the ligand that activates vinculin in cell-cell adhesions is unknown. Results:α-Catenin activates vinculin and the A50I talin-binding mutant of vinculin. Conclusion: These data suggest that α-catenin employs a novel mechanism to activate vinculin. Significance: These findings explain how vinculin is differentially activated in cell-cell and cell-matrix adhesions. Vinculin, an actin-binding protein, is emerging as an important regulator of adherens junctions. In focal-adhesions, vinculin is activated by simultaneous binding of talin to its head domain and actin filaments to its tail domain. Talin is not present in adherens junctions. Consequently, the identity of the ligand that activates vinculin in cell-cell junctions is not known. Here we show that in the presence of F-actin, α-catenin, a cytoplasmic component of the cadherin adhesion complex, activates vinculin. Direct binding of α-catenin to vinculin is critical for this event because a point mutant (α-catenin L344P) lacking high affinity binding does not activate vinculin. Furthermore, unlike all known vinculin activators, α-catenin binds to and activates vinculin independently of an A50I substitution in the vinculin head, a mutation that inhibits vinculin binding to talin and IpaA. Collectively, these data suggest that α-catenin employs a novel mechanism to activate vinculin and may explain how vinculin is differentially recruited and/or activated in cell-cell and cell-matrix adhesions.

Cadherin-mediated cell-cell adhesion is required for the formation of tissue boundaries and the separation of tissue layers that occurs during embryogenesis (1)(2)(3)(4). In adult tissues, cell-cell adhesions are spatially and temporally regulated to allow for the passage of lymphoid cells across cell layers (4 -6). Developmental and cellular cues alter cell-cell adhesion by facilitating the assembly and disassembly of protein complexes at the cadherins cytoplasmic tail (7), but the molecular events that regulate these processes remain poorly understood.
Vinculin is recruited to the cadherin adhesion complex by binding to ␤-catenin, and it governs cadherin function. For example, our laboratory showed that vinculin is required for stabilizing E-cadherin on the cell surface and that in its absence, cell-cell junctions do not properly assemble (8). Similarly, conditional knock-out of vinculin in the cardiomyocytes of mice resulted in disrupted N-cadherin-mediated cell-cell junctions and depolarized localization of gap junctions (9). Others demonstrated that vinculin acts downstream of myosin VI to regulate the maturation of cell-cell adhesions (10). In addition to regulating the assembly and maintenance of junctions, vinculin is critical for E-cadherin-mediated mechanosensing (11). These studies establish vinculin as a critical component of the cadherin adhesion complex, but how vinculin is regulated at this site is not known.
Vinculin has no enzymatic activity; rather it functions by binding to other proteins. For example, efficient membrane protrusion requires transient recruitment of the Arp2/3 complex to vinculin in adhesions in the leading edge of spreading cells (12), whereas the extent of integrin clustering requires vinculin binding to both talin and actin (13). In order for vinculin to bind ligands, an inactivating intramolecular interaction between its head and tail domains must be relieved (14 -16). The relaxation of this intramolecular interaction has been termed vinculin activation. In cell-matrix or focal adhesions, the mechanism for vinculin activation is somewhat controversial with some groups proposing that a single ligand, talin, activates vinculin and others demonstrating a requirement for talin and actin filaments (17)(18)(19)(20)(21)(22). Talin is not present in adherens junctions. Therefore, another ligand must activate vinculin in this region of the cell. In this study, we tested whether ␣or ␤-catenin, two cadherin adhesion complex components that bind vinculin, can activate it (23,24). We found that ␣-catenin, but not ␤-catenin, activates vinculin, and that this activation requires binding of vinculin to actin filaments. We generated an ␣-catenin point mutant that does not bind vinculin and found that vinculin activation requires direct binding to ␣-catenin. Furthermore, we report that ␣-catenin, unlike all the other known vinculin activators, binds and activates vinculin in the presence of an A50I substitution in the vinculin head. These data suggest that ␣-catenin employs a distinct and novel mechanism to activate vinculin. In addition, the results provide new evidence for stable activation of vinculin by coincident action of two ligands.
FRET Assay of TP3 in Cell Lysates-Cell lysate was made (20) from HEK 293 cells expressing TP3 or TP3A50I. The FRET assays were performed as previously described (20,22) using various concentrations of wild type or mutant ␣-catenins and 5 M actin. The emission spectra of fluorescent proteins in the lysate were acquired at 20°C with a Fluoromax-3 spectrofluorimeter (Jobin Yvon, Edison, NJ). Cerulean emission was traced from 460 to 600 nm with excitation at 440 nm, and Citrine emission was traced from 510 to 600 nm with excitation at 490 nm. The increment was 1 nm, and integration was 0.2 s. The excitation and emission slit widths were 3 mm and 5 mm, respectively. Lysate from an equal number of untransfected HEK293 cells was used to obtain a background emission spectrum.
Determination of the Corrected FRET Emission Ratio, SE/F da -The raw FRET signal is the Citrine emission (peak at 525) stimulated by excitation of Cerulean at 440 nm. It consists of the sensitized emission (SE), 3 the emission from direct excitation of citrine by 440 nm, and the overlap of the cerulean emission spectrum with the citrine emission spectrum. The latter two components of the raw FRET signal are referred to as "spectral cross-talk". The amount of spectral cross-talk was determined as previously described (20) and used to correct the raw FRET data. The corrected FRET emission ratio (ER) is the ratio of the sensitized emission at 525 nm to the emission at 475 nm, after correcting for the citrine and cerulean cross-talk. ER is SE/F da , where SE is the sensitized emission and F da is the fluorescence of the donor in the presence of the acceptor. SE/F da correlates directly with FRET efficiency (20); the higher the SE/F da , the stronger the FRET. FRET results are plotted as the SE/F da ratio.
Pull-down Assay-50 g of purified GST, or GST-FL ␣-catenin, or GST-␣-catenin truncations bound to the glutathione-Sepharose were incubated with MDCK cell lysate at 4°C for 1.5 h. The recovered proteins were washed, fractionated by SDS-PAGE, and analyzed by Western blotting.
Actin Co-sedimentation Assay-HEK 293 cells were detached with 1 mM EDTA in calcium-and magnesium-free PBS at 37°C for 20 min. The pelleted cells were resuspended in ice-cold hypotonic buffer (100 mM KCl, 20 mM Tris, pH 7.5, 2 mM MgCl 2 , 0.2 mM EGTA, 0.5 mM ATP, 0.5 mM DTT, 20 g/ml aprotinin, 2 mM Na 3 VO 4 , and 1 mM PMSF) and at a density of 2 to 4 ϫ 10 6 cells/ml, incubated on ice for 20 min, and homogenized manually. The lysate was cleared by centrifugation at 4°C, 13,000 rpm for 10 min, and then subjected to another centrifugation at 25°C, 80,000 rpm for 30 min. Lysates were mixed with designated purified protein at room temperature for 1 h. Samples were spun at 25°C, 80,000 rpm for 30 min. Equivalent amounts of total sample before spin (T), supernatant (S), and pellet (P) fractions were subjected to SDS-PAGE and recognized with hVIN1 and C4 monoclonal antibodies to vinculin (Sigma) and actin (Millipore), respectively. The amount of GFP-tagged vinculin in the pellet to the total (pellet/ total) was quantified using densitometry and expressed as a ratio. The data presented in the graph are the mean Ϯ S.E. from three independent experiments.
GFP Depletion Assay-HEK293 cells were lysed as described in actin cosedimentation assay. The cell lysate containing GFP-V1-851 or GFP-V1-851 A50I was incubated with increasing concentration of His 6 -tagged ␣-catenin 273-510 for 1 h at room temperature and then incubated with Ni-NTA-agarose. The unbound GFP-V1-851 or GFP-V1-851 A50I was separated from the bound fraction by centrifugation and assayed for concentration by spectrofluorimetry. The fraction of GFP-V1-851 or GFP-V1-851 A50I in complex with ␣-catenin 273-510 was plotted against unbound protein. Data were fitted to the equation B max ϫ X/(K d ϩ X) using Sigmaplot software, where B max ϭ maximum fraction of receptor capable of binding to ligand. The data presented in the graph are the mean Ϯ S.D. from three independent experiments.

␣-Catenin Activates Vinculin
Tris-HCl, pH 7.5, 100 mM KCl, 2 mM MgCl, 0.2 mM EGTA, 0.5 mM ATP, 0.5 mM DTT, 20 g/ml aprotinin, and 1 mM PMSF) at room temperature for 30 min. The mixtures were then incubated with GST Vinexin (amino acids 42-155) bound to glutathione beads that had been blocked with 5 M BSA overnight at 4°C. The samples were centrifuged and the supernatant harvested and saved as the soluble (S) fraction. The beads were then washed and the pellet (P) was recovered. The samples were resolved by SDS-PAGE and analyzed by Western blot.
Immunoprecipitation and Western Blotting-HEK293 cells were lysed as described in the actin cosedimentation assay. Cell lysates were incubated with indicated purified protein(s) for 1 h at room temperature. GFP was then immunoprecipitated with a monoclonal GFP antibody (Roche) at 4°C, and the immunoprecipitates were washed four times in lysis buffer, fractionated by SDS-PAGE, transferred to PVDF, and subjected to Western blot analysis with the appropriate antibody: the p34-Arc subunit of the Arp2/3 complex was blotted using a rabbit polyclonal antibody raised against a peptide that encompassed amino acids 179 -204 of p34-Arc (25). Ponsin was blotted using a rabbit antibody raised against synthetic peptide corresponding to amino acids 192-206 of human CAP (Upstate Cell Signaling Solutions). GFP was blotted with a mouse monoclonal antibody (Roche). Actin was blotted with a mouse monoclonal antibody (clone C4, MP Biomedicals). The blots were developed using ECL Western blot detection reagents (Pierce), and the signal was detected on x-ray film (Kodak).

RESULTS
Vinculin Is Activated by ␣-Catenin and F-Actin-In adherens junctions ␣and ␤-catenin bind to the vinculin head domain suggesting that one or both of these proteins might activate vinculin (24,26). We tested whether vinculin is activated by either catenin. When vinculin is activated, it binds to actin filaments (16,22). Consequently, vinculin activation has reliably been measured by examining the ability of vinculin to co-sediment with actin filaments (16,20,22). We examined whether ␣or ␤-catenin could induce vinculin to co-sediment with actin filaments. For most of these studies, we employed an ␣-catenin fragment 273-510 that binds vinculin because the full-length protein forms intramolecular interactions that preclude access to the vinculin binding site (27,28). Also, fulllength ␣-catenin dimer binds actin and this would prevent knowing whether a ternary complex of vinculin, ␣-catenin, and actin is present in pellets of actin co-sedimentation assays. When cell lysate containing full length EGFP-vinculin was incubated with purified actin filaments and then centrifuged at speeds sufficient to sediment actin filaments, little to no vinculin co-sedimented with actin filaments alone (Fig. 1, A and C and Ref. 22). The addition of ␣-catenin 273-510 to the mixtures of vinculin and actin triggered large amounts of vinculin to pellet with actin filaments (Fig. 1, A and C). This effect was specific for ␣-catenin as little or no vinculin sedimented when ␤-catenin was added (Fig. 1B).
Similar results were found in parallel experiments using FRET probes that report on vinculin activation and actin-binding in cell lysates (16,20,22). For these studies, we employed a vinculin FRET probe, TP3, that contains Citrine inserted before the tail domain and Cerulean on the C terminus, in place of YFP and CFP in the original TP2 construct (20). When vinculin is in the closed conformation, the two fluorophores in this probe are in close contact, which gives a high FRET signal; after vinculin activation and binding to actin, the FRET signal reduces because the binding of actin to vinculin tail domain forces the two fluorophores away from one another (20). We found that the FRET probe has a strong fluorescent signal by itself; the addition of 5 M actin filaments alone, or purified ␣-catenin 273-510 had little effect on the FRET signal (Fig. 1D). In contrast, the addition of both actin filaments and ␣-catenin 273-510 induced a significant reduction in the FRET signal (Fig. 1D). Furthermore, ␣-catenin activation of vinculin was dose dependent with a half-maximal activation observed at 3 M ␣-catenin 273-510 (Fig. 1E). Like wild type vinculin, the vinculin FRET probe also co-sediments with actin in the presence of ␣-catenin or IpaA (supplemental Fig. S1). Thus, ␣-catenin activates vinculin.

␣-Catenin Activates Vinculin
To determine whether ␣-catenin 273-510 activates vinculin for binding to its ligands and to evaluate whether actin is required for this event, we tested whether vinculin bound ponsin, a component of the nectin based adhesions, or the Arp2/3 complex, a potent nucleator of actin polymerization. Both of these ligands bind to sites in the proline-rich region of vinculin and previous studies show that vinculin co-immunoprecipitates with these proteins only when its inactivating intramolecular head-tail interaction has been relieved (12,29). In the absence of ␣-catenin 273-510 or actin filaments, neither ponsin nor the Arp2/3 complex was recovered with vinculin immunoprecipitates (Fig. 1F). In the presence of both ␣-catenin 273-510 and actin, ponsin, and the Arp2/3 complex were recruited to vinculin (Fig. 1F). Taken together, these results show that ␣-catenin activates vinculin for binding to Arp2/3 and ponsin and that this activation requires actin filaments.
To ensure that vinculin activation by ␣-catenin did not require other proteins present in the cell lysates, we tested whether the activation could be recapitulated in vitro using purified proteins. We monitored activation by examining vinculin binding to the SH3 domain of vinexin. We found that only a small fraction of vinculin bound vinexin when ␣-catenin 273-510 alone or F-actin alone were present. When both ␣-catenin and F-actin were present, a large percentage of vinculin bound to GST-vinexin (Fig. 1G). These results indicate that purified ␣-catenin activates vinculin and rules out the possibility that other constituents of the lysates are responsible (Fig. 1G). Hence, ␣-catenin, like talin, employs a combinatorial mechanism to fully activate vinculin.
Generation of an ␣-Catenin Point Mutant That Blocks Vinculin Binding and Activation-To explore whether direct binding of ␣-catenin to vinculin is required for activation, we considered generating a mutant form of ␣-catenin that could not bind vinculin. We mapped the vinculin-binding site on ␣-catenin using a series of ␣-catenin fragments expressed as GST fusion proteins ( Fig. 2A). Previous studies showed vinculin binds to a region of ␣-catenin that contains amino acids 273-510 (27). Using our ␣-catenin GST fusions, we confirmed the presence of a vinculin binding site in this region and further mapped the site to amino acids 326 -377 of ␣-catenin (Fig. 2B). This region of ␣-catenin has been predicted to contain one ␣-helix (Fig. 2D). Alignment of the vinculin binding sites (VBSs) in talin, IpaA, and ␣-actinin revealed that these proteins possess an amphipathic ␣-helix with several conserved hydrophobic residues (Fig. 2C). The ␣-helix of ␣-catenin that binds to vinculin also has these features (Fig. 2, C and D). Substitution of proline for leucine 344 blocked vinculin binding to an ␣-catenin fragment (273-510) and to full-length ␣-catenin (Fig. 3A).
We then tested whether mutation of L344P abolishes ␣-catenin 273-510 activation of vinculin. The mutant ␣-catenin did not induce vinculin co-sedimentation with actin filaments (Fig. 3, B and C, supplemental Fig. S2), and it did not trigger a reduction of FRET signal (Fig. 1D). Furthermore, the mutant ␣-catenin did not induce ponsin or the Arp2/3 complex to co-immunoprecipitate with vinculin (Fig. 1F). Taken together, these findings indicate that direct binding of ␣-catenin to vinculin is required for vinculin activation, and substitution of ␣-catenin L344P blocks this effect.
Vinculin Activation by ␣-Catenin and F-actin Is Independent of Vinculin A50-All of the known vinculin activators contact amino acid A50 lying within the hydrophobic groove between helices ␣1 and ␣2 in the vinculin head (17). We tested whether activation of vinculin by ␣-catenin could be blocked by the A50I substitution which blocks IpaA, talin, and talin VBSs from binding to vinculin (15,16). Surprisingly, ␣-catenin activated A50I vinculin in all of the assays we used (Fig. 4, A-C, E). This observation is in contrast to IpaA and talin, both of which completely fail to activate vinculin in the presence of an A50I substitution (Fig. 4D). These effects were not due to the usage of a fragment of ␣-catenin, because full-length ␣-catenin activated the TP3 vinculin FRET probe (Fig. 4C). In contrast to ␣-catenin 273-510, the addition of full-length ␣-catenin alone partially activated vinculin and the addition of actin further enhanced this activation (Fig. 4C). Similar results were obtained when ␣-catenin was added to TP3 A50I vinculin FRET probe (Fig.  4C). Furthermore, the ability of ␣-catenin to activate TP3 vinculin A50I was not due to an artifact of the probe as this mutation blocked activation stimulated by IpaA and VBS3 (Fig. 4D). Thus, ␣-catenin is the first known protein that is able to activate vinculin containing the A50I point mutation.
To determine whether the A50I substitution affected ␣-catenin binding to vinculin, the dissociation constant for this interaction was measured. For these studies, we used the GFP depletion assay which is commonly used in the field to analyze binding affinities between vinculin and its binding partners. This approach uses GFP-tagged vinculin proteins that are transiently expressed in HEK293 cells to ensure proper protein folding as well as post-translational modification. Increasing amounts of purified His-tagged ␣-catenin fragment were titrated into cell lysate and recovered with Ni-NTA agarose. The decrease of GFP fluorescence in the cell lysate was monitored by fluorimeter. The data were then plotted and fitted to a single binding site equation. We found that ␣-catenin 273-510 binds to vinculin head domain with a K d of 98 Ϯ 19 nM. This result is in good agreement with the previously reported value The ␣-catenin VBS is aligned with two IpaA VBSs, three talin VBSs, and one ␣-actinin VBS using ClustalW. Hydrophobic residues conserved in at least five VBSs are in bold and italicized. The orientation of ␣-actinin VBS is inverted as reported previously (19). D, ␣-catenin VBS is predicted to be amphipathic ␣-helix based on amino acid sequence (27). The ␣-catenin VBS (amino acids 328 -349) was arranged around a helical wheel, which revealed the helix has a hydrophilic face and a hydrophobic face. Hydrophilic residues are shown as circles, hydrophobic residues as diamonds, potentially negatively charged residues as triangles, and potentially positively charged as pentagons. obtained using isothermal titration calorimetry (Fig. 5A and Ref. 15). A50I mutation reduced ␣-catenin binding to a K d of 5.49 Ϯ 0.35 M (Fig. 5B). In further support of ␣-catenin binding to vinculin harboring this mutation, ␣-catenin could be coimmunoprecipitated with vinculin A50I (8). While ␣-catenin binds A50I with a lower binding affinity, the observation that it is still able to activate vinculin is in stark contrast with what has been observed with talin, which does not co-precipitate with vinculin A50I and binds with affinity that is too weak to be measured in pull-down assays (K d Ͼ 20 M) (22). Collectively, these data suggest that ␣-catenin activates vinculin independently of contacting residue A50, and employs a mechanism and vinculin interface that is distinct from those used by talin and IpaA.

DISCUSSION
One challenge to understanding how dynamic changes in cell-cell adhesion are achieved is to identify the cadherin adhesion complex proteins that are essential and to pinpoint mechanisms for regulating their activities. Vinculin is critical for E-cadherin function, but little is known about how it is regulated at this site. Here, we show that ␣-catenin is an adherens junction component that induces activating conformational changes in vinculin (Fig. 1). This effect is specific for ␣-catenin as another adherens junctions component, ␤-catenin, does not activate vinculin (Fig. 1B). Unlike talin which activates vinculin in focal adhesions, ␣-catenin functions independently of an A50 residue in the vinculin head, which forms part of the binding site for talin and IpaA, suggesting the presence of a distinct molecular mechanism for vinculin activation (Fig. 4).
We considered whether ␣-catenin 273-510 is as efficient at activating full length vinculin as other known ligands. The FRET vinculin activation assay showed that ␣-catenin 273-510 achieved half-maximal activation at 3 M ␣-catenin (Fig. 1E). In published titration studies, 2.9 M talin rod domain, 0.48 M VBS3, and 82 nM IpaA were required for half-maximum activation of the HP3 vinculin FRET probe (22). These data indicate that ␣-catenin 273-510 activates vinculin as efficiently as talin rod domain. In support of this notion, the rod domain of talin binds to vinculin 1-851 with a K d ϳ 99 nM (22,30), and ␣-catenin 273-510 binds to vinculin 1-851 with a K d ϭ 98 Ϯ 19 nM (Fig. 5). The difference in the binding affinities of vinculin and ␣-catenin domains for each other versus the concentration of ␣-catenin 273-510 required for half-maximum activation of full-length vinculin FRET probes reflect the difference in accessibility of ligand binding sites on autoinhibited full-length vinculin (15,16,22).
While talin, IpaA, and ␣-catenin activate vinculin, they seem to employ at least two distinct molecular mechanisms for activation. Talin and ␣-catenin 273-510 bind to vinculin head domain and require actin filaments to bind to the tail domain to form stable complexes with vinculin ( Fig. 1 and Ref. 22). In contrast, IpaA (22) and full-length ␣-catenin activate vinculin in the absence of any exogenously added ligand, although addition of actin increases the extent of activation in both cases. In addition to distinct requirements for actin, the known activators bind to partially distinct regions of vinculin. Specifically, substitution of A50I in the vinculin head domain completely abolishes both talin and IpaA binding and vinculin activation (22). However, this mutant version of vinculin still binds ␣-catenin, albeit with a lower affinity. Importantly, ␣-catenin can still activate A50I vinculin (Fig.  4), even though its affinity is reduced ϳ50-fold (Fig. 5). These findings suggest that there are at least two distinct mechanisms for activating vinculin: 1) ligand binding to the A50 groove in the vinculin head and actin filament binding to the tail, and 2) ligand binding independently of A50 with a requirement for actin filaments.  VIN 1-398) was incubated with purified GST or wild-type or mutant ␣-catenin in binding buffer. The protein complexes were recovered using glutathione beads, washed, separated by SDS-PAGE, and immunoblotted with an antibody against His. FL indicates full-length ␣-catenin. B and C, substitution of L344P in ␣-catenin blocks vinculin co-sedimentation with actin. HEK293 cells expressing GFP-tagged fulllength vinculin (WT VIN) were lysed, clarified by centrifugation, and then incubated with the indicated proteins with the same concentration as described in Fig. 1A. The mixtures were analyzed as described in Fig. 1A, and the resulting immunoblot is shown in B. The amount of GFP-tagged vinculin in the pellet and the total sample prior to centrifugation was quantified and expressed a ratio of the pellet/total as described in Fig. 1. The data presented in the graph are the mean Ϯ S.E. from three independent experiments. The co-immunoprecipitation of vinculin A50I with ponsin or the Arp2/3 complex was assessed as described in Fig. 1F. TCL denotes a sample of total cell lysates.
At first glance, the idea that there are two distinct mechanisms for activating vinculin is somewhat surprising given that all three ligands possess and employ an amphipathic ␣-helix with shared structural characteristics for vinculin binding and activation ( Fig. 2C and Ref. 30). However considering that vinculin is comprised almost exclusively of ␣-helical bundles with similar structural features, it is reasonable to believe that these amphipathic ␣-helices could bind to more than one region of the molecule (15,31). Thus, it is likely that there is more than one surface on vinculin that is required for its activation. In support of this notion, both talin and IpaA have multiple vinculin binding sites that could potentially interact with more than one region of vinculin. Also IpaA binds to two distinct sites on vinculin (32). We attempted to identify an additional ␣-catenin binding site on vinculin, but have been unable to detect ␣-catenin binding to vinculin regions other than 1-258 amino acids. It is possible that the second site lies within this region and/or has a low affinity for ligands or is cooperative with the first site.
Irrespective of the mechanism(s) by which ␣-catenin activates vinculin, interaction of these proteins is critical for adherens junction function. It has long been thought that ␣-catenin localizes vinculin to adherens junctions owing to the observation that vinculin is lost from adherens junctions when ␣-catenin is deleted from cells (24,33). However, the requirement for ␣-catenin to localize vinculin to adherens junctions has not been universally supported. For example, work by Hazan et al. (23) showed that in cell lines lacking ␣-catenin, vinculin co-precipitated with ␤-catenin-E-cadherin complexes. Similarly, we found that mutant versions of vinculin that were devoid of ␤-catenin binding, but retained ␣-catenin binding, did not localize to adherens junctions (8). Hence in some contexts, vinculin localizes to adherens junctions independently of ␣-catenin. We believe that these conflicting observations may be explained by the fact that ␣-catenin does not localize vinculin to adherens junctions but does activate it at this site. In this scenario, cells lacking ␣-catenin would not be expected to have vinculin in adherens junctions as a result of vinculin not being activated and stabilized at this site. Alternatively vinculin might be present but the antibodies used to detect it did not recognize the autoinhibited conformation of vinculin.
The new finding that ␣-catenin and actin cooperate to activate vinculin combined with our previous observation that ␤-catenin localizes vinculin to adherens junctions (8) allow for a revised model for vinculin recruitment and activation in adherens junctions to be proposed. According to this model, upon initiation of adherens junction assembly, ␤-catenin localizes vinculin to adherens junctions where it is bound and activated by the combined effects of ␣-catenin and actin. The subsequent unfurling of vinculin allows its numerous interacting proteins to bind and modulate the assembly and maturation of adherens junctions and E-cadherin mechanosensing (8 -11, 34, 35).