Distinct intramolecular interactions regulate autoinhibition of vinculin binding in αT-catenin and αE-catenin

α-Catenin binds directly to β-catenin and connects the cadherin–catenin complex to the actin cytoskeleton. Tension regulates α-catenin conformation. Actomyosin-generated force stretches the middle (M)-region to relieve autoinhibition and reveal a binding site for the actin-binding protein vinculin. It is not known whether the intramolecular interactions that regulate epithelial (αE)-catenin binding are conserved across the α-catenin family. Here, we describe the biochemical properties of testes (αT)-catenin, an α-catenin isoform critical for cardiac function and how intramolecular interactions regulate vinculin-binding autoinhibition. Isothermal titration calorimetry showed that αT-catenin binds the β-catenin–N-cadherin complex with a similar low nanomolar affinity to that of αE-catenin. Limited proteolysis revealed that the αT-catenin M-region adopts a more open conformation than αE-catenin. The αT-catenin M-region binds the vinculin N-terminus with low nanomolar affinity, indicating that the isolated αT-catenin M-region is not autoinhibited and thereby distinct from αE-catenin. However, the αT-catenin head (N- and M-regions) binds vinculin 1000-fold more weakly (low micromolar affinity), indicating that the N-terminus regulates the M-region binding to vinculin. In cells, αT-catenin recruitment of vinculin to cell–cell contacts requires the actin-binding domain and actomyosin-generated tension, indicating that force regulates vinculin binding. Together, our results show that the αT-catenin N-terminus is required to maintain M-region autoinhibition and modulate vinculin binding. We postulate that the unique molecular properties of αT-catenin allow it to function as a scaffold for building specific adhesion complexes.

The cadherin-catenin complex that forms the core of the adherens junction (AJ) is required for intercellular adhesion and tissue integrity (1-3). Classical cadherins are single-pass transmembrane proteins with an extracellular domain that forms trans-interactions with cadherins on adjacent cells (4)(5)(6). The adhesive properties of classical cadherins are driven by the recruitment of cytosolic catenin proteins to the cadherin tail: p120-catenin binds to the juxtamembrane domain, and β-catenin binds to the distal part of the tail. β-Catenin recruits α-catenin, a mechanoresponsive actin-binding protein (7)(8)(9)(10)(11)(12)(13). The AJ mechanically couples and integrates the actin cytoskeletons between cells to allow dynamic adhesion and tissue morphogenesis (3).
Three α-catenin family proteins are expressed in mammals: the ubiquitous αE-catenin, neuronal (αN)-catenin, and testes (αT)-catenin (36,37). αE-catenin and αN-catenin are 81% identical and 91% similar in amino acid sequence. αT-catenin is 58% identical and 77% similar to αE-catenin and αN-catenin, making it the most divergent of the family (36)(37)(38). αT-catenin is predominantly expressed in the heart, testes, brain, and spinal cord (39,40). In the heart, it localizes to the intercalated disc (ICD), a specialized adhesive structure that maintains mechanical coupling and chemical communication between cardiomyocytes (41,42). In mice, loss of αT-catenin from the heart causes dilated cardiomyopathy, and mutations in αTcatenin are linked to arrhythmogenic ventricular cardiomyopathy in humans (43,44). In addition to cardiomyopathy, αTcatenin is linked to multiple human diseases, including asthma, neurological disease, and cancer (40,45,46). Despite a growing awareness of its importance in human disease, the molecular properties and ligand interactions of αT-catenin remain poorly understood. Our previous work revealed that αT-catenin, unlike mammalian αE-catenin, is a monomer in the solution that can bind to F-actin with low micromolar affinity in the absence of tension. F-actin binding is also not allosterically regulated, as the β-catenin-αT-catenin complex binds to F-actin with the same affinity as the αTcatenin monomer (38). Single-molecule pulling experiments have shown the αT-catenin M-region to be mechanoresponsive as it unfurls in a force range similar to αE-catenin (47).
Here we show that αT-catenin associates with the components of the cadherin-catenin complex with an affinity similar to αE-catenin in vitro, revealing that they may compete with one another for binding β-catenin at the plasma membrane. We also show that the M-region of αT-catenin is not autoinhibited and can bind the vinculin N-terminus in the absence of tension with strong affinity. Unlike αE-catenin, however, when the N-terminus of αT-catenin is attached to the M-region, the affinity for vinculin drops significantly. This indicates that interdomain interactions between the N-terminus and the M-region of αTcatenin regulate its interaction with vinculin. We measured the recruitment of vinculin to cell-cell contacts and found that, despite the distinct mechanism of regulation, recruitment of vinculin is still tension dependent. Our findings indicate that the way in which tension regulates vinculin binding differs between αE-catenin and αT-catenin. We postulate that this mechanism is important for the ability of αT-catenin to build specific and distinct molecular complexes at AJs.

Results
αT-catenin binds tightly to the β-catenin-N-cadherin core complex We characterized the interaction between αT-catenin and β-catenin by isothermal titration calorimetry (ITC) using purified recombinant proteins. We used the head region (comprising the N-and M-domains) of αT-catenin (aa 1-659, Fig. 1A) for these experiments because it is more stable than full-length αT-catenin and yields sufficiently high protein concentrations for ITC. Past studies revealed that the αE-catenin head region (aa 1-651) binds β-catenin and the β-catenin-E-cadherin tail complex with a similar affinity to full-length αE-catenin (12). We observed that the αTcatenin head binds β-catenin with a dissociation constant 250 nM ( Fig. 1B; Table 1). The affinity of αT-catenin for βcatenin is an order of magnitude weaker than the association of αE-catenin or αN-catenin for β-catenin (15-20 nM; (12)).
αT-catenin autoinhibition αT-catenin-β-catenin interaction by titrating the β-catenin-Ncad cyto complex into αT-catenin head (Fig. 1C). The affinity of this interaction was 5 to 6 nM (Table 1), indicating that αTcatenin binds to the cadherin-β-catenin complex an order of magnitude more strongly than to β-catenin alone. This affinity is similar to the 1-to 2-nM affinity observed between the cadherin tail-β-catenin complex and αE-catenin or αN-catenin (12) and suggests that αT-catenin can effectively compete with αE-catenin for binding to the membrane-associated cadherin-β-catenin complex.

αT-catenin N-terminus is monomeric
Full-length αT-catenin is primarily a monomer in the solution, although it does have homodimerization potential in vitro (38). The best evidence for dimerization potential   Increasing concentrations of αT-catenin V94D N1-N2 were incubated with or without 1-mM BS3 at 4 C for 30 min, separated by SDS-PAGE, and stained with Coomassie dye. D, αT-catenin N1-N2 binding to β-catenin-N-cadherin cytoplasmic tail (β-cat-Ncad cyto ) was measured by ITC. Thermodynamic properties derived from this trace are shown in Table 1. ITC, isothermal titration calorimetry; Ncad cyto , N-cadherin tail; SEC, size-exclusion chromatography.
We then analyzed the oligomeric state of the αT-catenin N-terminus by cross-linking. Increasing concentrations of αT-catenin N1-N2 were incubated with or without the crosslinker bis(sulfosuccinimidyl)suberate (BS3) at 4 C or 37 C and the resulting products analyzed by SDS-PAGE. As expected, the αT-catenin N1-N2 migrated as a 25-kDa protein in the absence of a cross-linker (Fig. 2B). Incubation with BS3 did not affect migration at low concentrations, although at higher concentrations (4 and 8 μM), larger species were detected at both temperatures. In contrast, αTcatenin V94D N1-N2 ran as 50-kDa protein in the presence of BS3 at all concentrations tested (Fig. 2C), indicating a cross-linked dimer. We conclude that the αT-catenin N terminus, similar to full-length protein, is primarily a monomer in the solution.
We then tested the ability of the αT-catenin N1-N2 to bind to the β-catenin-Ncad cyto complex by ITC. The affinity of this interaction was 7 nM (Fig. 2D, Table 1), similar to the αTcatenin head and confirming that this fragment contains the complete β-catenin binding site.  Table 2. ITC, isothermal titration calorimetry; SEC, sizeexclusion chromatography.

αT-catenin autoinhibition αT-catenin M-region binds vinculin
The αE-catenin M region is autoinhibited with respect to vinculin binding: mechanical force is required to break an internal salt bridge network within M1-M3 to reveal the vinculin-binding site in M1 (25,30). The αT-catenin M region also recruits vinculin and force is required to unfurl M1 to promote high-affinity binding (47). However, our previous proteolysis experiments of full-length αT-catenin revealed that the M2-M3 region existed in a more open, protease-sensitive state (38). Notably, the amino acids that form the salt bridges required for autoinhibition in the αE-catenin M-region are conserved in αT-catenin, with the exception of E277 in M1 that pairs with R451 in M2. In αT-catenin, the arginine is conserved at the corresponding residue (aa 446), but the glutamic acid is replaced by a threonine at aa 272 (25), preventing these residues from interacting. We questioned if αT-catenin M1-M3 adopted an autoinhibited conformation.
We examined the organization and vinculin-binding properties of the complete αT-catenin M-region (M1-M3, aa 260-626). αT-catenin M1-M3 eluted as a single, discrete peak by size-exclusion chromatography, identical to αE-catenin M1-M3 (aa 273-651; Fig. 3A). We then probed the M-region flexibility by limited trypsin proteolysis. Tryptic digestion of αE-catenin M1-M3 revealed that it was largely protease resistant: nearly 50% of the fragment was still intact after 120 min of digestion, consistent with it forming a closed, autoinhibited domain (Fig. 3B). In contrast, αT-catenin M1-M3 was completely digested after 60 min into three stable fragments at 23, 16, and 12 kDa (Fig. 3B). Note that both M regions contain a similar number of lysine and arginine residues (47 in αE-catenin and 39 in αT-catenin), the majority of which are conserved. N-terminal sequencing revealed that the 23-kDa and 16-kDa fragments both started at aa 379 and represented the M2-M3 and M2 bundles, respectively. The 12-kDa fragment started at aa 485 and corresponded to the M3 bundle (Fig. 3C). Similar protease-resistant fragments were observed from digest of full-length protein (38) and are consistent with the αT-catenin M-region adopting a more open, protease-sensitive state relative to αE-catenin, despite five of the six salt bridge residue pairs being conserved. Likewise, the isolated αT-catenin M1-M3 fragment does not appear to adopt a stable, autoinhibited conformation.

αT-catenin N-terminus regulates vinculin binding
Recent in vitro single-molecule stretching experiments revealed that force is required to expose the vinculin-binding site in αT-catenin M1-M3 (47). However, our ITC results with the αT-catenin M1-M3 fragment indicated that tension was not required to release M1. In the stretching experiments, the αT-catenin M1-M3 fragment was flanked by a pair of titin I27 immunoglobulin-like domains, and the N-terminus of the fusion protein was tethered to a substrate (47). We questioned if N-terminal associations with M1-M3 stabilize M1 and regulate vinculin binding.
We first characterized the interaction between αT-catenin head domain and vinculin D1 by ITC. The binding shifted from exothermic to endothermic, and the affinity was 5 μM, two orders of magnitude weaker than those observed with M1-M3 (Fig. 4A, Table 2). This suggested that the addition of N1-N2 stabilized M1 and inhibited vinculin binding.
Recent work revealed allosteric coupling between the Nterminal domains and M-region of αE-catenin (16). Specifically, the presence of β-catenin caused changes in the accessibility of cysteine residues in the N2-M2 interface and in M3. Because β-catenin binding alters the relative positions of the αE-catenin N1 and N2 domains (12), we tested if β-catenin binding to the αT-catenin N-region affected vinculin binding to the M-region. Vinculin D1 was titrated into a solution of the purified αT-catenin head-β-catenin complex. The presence of β-catenin had little impact on the affinity (3 μM, Fig. 4B, Table 2), indicating that N-terminal-mediated autoinhibition was maintained. This result is consistent with past work

αT-catenin autoinhibition
showing that the β-catenin-cadherin complex had no effect on αE-catenin binding to vinculin (23). We then asked if the entire N-terminus is required to regulate M1 binding. The addition of N2 to αT-catenin M1-M3 (aa 147-626) did not weaken vinculin D1 binding relative to M1-M3 (K d = 50 nM, Fig. 4C, Table 2). The reaction switched from exothermic to endothermic, suggesting partial compensation of M1 stability in this fragment. Thus, all or part of N1 is required to regulate M1 interactions with vinculin. This is consistent with the observation that removing the first 56 residues of N1 from full-length αE-catenin reduces the inhibition of vinculin binding by about 50-fold (23).
We next questioned if the titin repeats attached to M1-M3 in the construct used by Pang et al. (47) stabilized M1. We titrated vinculin D1 into a solution of the 2I27-αT-catenin M1-M3 (aa 259-667)-2I27 construct. Notably, the binding was endothermic and the affinity was 250 nM, markedly weaker than that observed with αT-catenin M1-M3 alone (Fig. 4D, Table 2). We speculate that the well-structured titin repeats promote M1 stability and may sterically occlude vinculin D1 binding, thus reducing the affinity. The M1 stability and steric hindrance provided by the I27 repeats partially mimic the complete N-terminus and likely explain why tension is needed to promote vinculin binding in the single-molecule stretching experiments, whereas αT-catenin M1-M3 in the solution binds vinculin readily.
Given that the entire N-terminus is required to regulate vinculin binding in αT-catenin, we asked if N1-N2 could regulate M1-M3 in trans. We titrated vinculin D1 into an equimolar mixture of αT-catenin N1-N2 and M1-M3. The binding reaction was similar to M1-M3 alone: exothermic with a K d of 50 nM (Fig. 4E, Table 2). This result indicates that the M-region must be physically coupled to the N-terminus to stabilize M1 and regulate ligand accessibility.
We then examined the correlation between EGFP fusion expression and vinculin recruitment. As expected, EGFP-αE-catenin expression levels did not correlate with vinculin levels because vinculin was only recruited when αE-catenin was activated by sufficient tension (24,28). However, we observed a positive correlation between EGFP-αT-catenin expression and vinculin recruitment. We observed a similar positive correlation with constitutively open αE-catenin constructs in cardiomyocytes (34). Together, these results suggest that while vinculin recruitment to αT-catenin is tension dependent, the force threshold for binding may be low, permitting all/most AJ-incorporated αT-catenin molecules to recruit vinculin.
Finally, we tested if actin binding was required to relieve αTcatenin autoinhibition and if removal of M3 opened the Mregion to permit vinculin recruitment. We expressed EGFPtagged truncations of αE-catenin and αT-catenin: the head domain lacking the ABD (N1-M3) or the head domain minus M3 and ABD (N1-M2). All EGFP-tagged αE-catenin and αTcatenin fragments expressed as soluble proteins of the predicted size (Fig. 5N). αE-catenin and αT-catenin N1-M3 (aa 1-670 and aa 1-659, respectively) were both cytosolic and, as expected, unable to form cell-cell contacts and recruit vinculin (Fig. 5, J and K). Deletion of M3 in αE-catenin relieves autoinhibition, and αE-catenin N1-M2 (aa 1-510) was able to recruit vinculin and restore cell-cell adhesion, as shown previously (49). Here vinculin provides the necessary actinbinding activity to link the cadherin-catenin complex to Factin through αE-catenin. In contrast, αT-catenin N1-M2 (aa 1-502) was cytosolic and failed to recruit vinculin to cell-cell contacts. This suggests that in αT-catenin, loss of M3 does not release M1 for ligand binding and underscores the importance of the N-terminus in regulating M1 binding to vinculin.

αT-catenin forms a strong AJ core with cadherin-β-catenin
Our ITC results show that the β-catenin-Ncad cyto complex binds with high, 5-nM affinity to αT-catenin similar to 1-to 2-nM affinity previously reported between the β-catenincadherin tail complex and αE-catenin (12). Thus, αT-catenin forms a strong cadherin-catenin core complex like αE-catenin to link actin to the AJ. αT-catenin is coexpressed with αEcatenin in multiple mammalian tissues and, in the heart, it is enriched along the ICD with αE-catenin (39,45,46,50). It is unknown if αE-catenin and αT-catenin bind stochastically to the cadherin-β-catenin complex in cardiomyocytes or if binding is regulated to favor one α-catenin or the other. Notably, loss of either αE-catenin or αT-catenin from the mouse heart causes dilated cardiomyopathy (43,51), suggesting that each α-catenin has a unique, critical role in ICD function and heart physiology. In addition, αE-catenin and αTcatenin (Heier and Kwiatkowski, unpublished data) also bind plakoglobin, the desmosome-associated β-catenin homolog that is enriched at the ICD (52,53). Plakoglobin binds the cadherin tail with an affinity similar to β-catenin, and both plakoglobin and β-catenin are enriched at cardiomyocyte AJs (50). The downstream consequences of α-catenin binding to plakoglobin or β-catenin are not clear, but recent evidence that the N-terminus and M-region of αE-catenin are allosterically αT-catenin autoinhibition coupled (16) raises the possibility that differential binding could affect downstream ligand recruitment. Future work is expected to reveal how specific cadherin-β-catenin/plakoglobin-α-catenin complexes are formed and function to regulate adhesion and signaling.
In the absence of the cadherin tail, β-catenin alone binds αT-catenin an order of magnitude weaker than αE-catenin (250 nM versus 20 nM). This suggest that αT-catenin does not compete with αE-catenin for binding cytosolic β-catenin. Although there is evidence of a cytosolic αE-catenin-β-catenin complex in epithelial cells (54), the physiological relevance of this interaction is not clear and a similar α-catenin-β-catenin complex in cardiomyocytes has not been reported.
αE-catenin can homodimerize and the cadherin-free, cytosolic homodimer pool regulates actin dynamics and cell motility (9,10,12,54,55). In contrast, full-length αT-catenin is a monomer in the solution and has limited dimerization potential with no evidence of homodimerization in vivo (38). Our results here show that the isolated αT-catenin N-terminus is also monomeric. We speculate that limited dimerization potential and weaker affinity for β-catenin favor a cadherin-free, cytosolic pool of an αT-catenin monomer. The role of such a monomer pool is unclear.
αT-catenin autoinhibition M3 maintain the autoinhibited form. Tension breaks these interactions to release M1, allowing it to unfold and bind strongly to vinculin. More recent biophysical data suggest that the αT-catenin M-region adopts a similar conformation with force being required to free M1 for high-affinity binding to vinculin (47).
Our limited proteolysis experiments with the isolated αTcatenin M-region indicate that it exists in a more open, protease-sensitive conformation. Consistent with this, ITC results revealed that the αT-catenin M-region binds vinculin with low nanomolar affinity and is not autoinhibited. αT-catenin possesses five of the six residue pairs that form the salt bridge network that mediates αE-catenin autoinhibition (25). The E277-R451 bridge between M1 and M2 in αE-catenin is not conserved in αT-catenin, with the glutamic acid replaced by a threonine at aa 272. The threonine at aa 272 is conserved across the αT-catenin family. Assuming no major structural differences between the αE-catenin and αT-catenin M-regions, the T272 and R446 residues would not be able to interact. In simulations, the αE-catenin E277-R451 salt bridge is predicted Figure 5. αT-catenin recruitment of vinculin to cell-cell contacts is tension dependent. A-D, R2/7 cells were transfected with EGFP-αE-catenin full length (αEcat FL; A and C) or EGFP-αT-catenin full length (αTcat FL; B and D) and treated for 30 min with dimethyl sulfoxide (A and B) or 50-μM blebbistatin (C and D) before fixation. Cells were stained for F-actin and vinculin. Individual and merged vinculin (magenta) and GFP (green) channels are shown. E, quantification of vinculin intensity at cell-cell contacts. Signal intensity at contacts was divided by the average cytoplasmic intensity, and a scatter plot of all data points is shown. The black horizontal line is the median, and the error bars define the interquartile range. One-way ANOVA with Tukey's comparisons; n ≥ 24 images from two independent experiments. F-I, vinculin contact/cytoplasmic ratio (C/C) was plotted against the average GFP intensity of each fusion construct from the masked cell contact region. Linear regression analysis was performed to calculate the slope (red line), 95% confidence intervals (black lines surrounding the slope), and R 2 value. The slope deviation from zero was analyzed for significance (p value). n ≥ 24 images from two independent experiments. J-M, R2/7 cells were transfected with EGFP-tagged αE-catenin head (aa 1-670, J), αT-catenin head (aa 1-659, K), αE-catenin N1-M2 (aa 1-510, L), and αT-catenin N1-M2 (aa 1-502, M), fixed and stained for vinculin. Individual GFP (top panel) and vinculin (bottom panel) channels are shown. Scale bar is 10 μm in all images. N, R2/7 cells transfected with EGFP-tagged αE-catenin or αT-catenin full-length (FL), head, or N1-M2 constructs. Cell lysates separated by SDS-PAGE and blotted for GFP (top) or GAPDH (bottom). n.s., not significant.
αT-catenin autoinhibition to rupture quickly under external force (16,25). Unfortunately, we lack a structure of the αT-catenin M-region to determine if a similar network regulates domain organization. However, our data demonstrate that such interbundle interactions within the αT-catenin M-region are insufficient for autoinhibition.
The N-terminus is required for αT-catenin M-region autoinhibition The αT-catenin head fragment (N1-M3) showed weak, micromolar binding to vinculin, indicating that the N-terminus regulates M-region autoinhibition. The entire N-terminus (N1-N2) is required for autoinhibition because deleting N1 restored strong vinculin binding. Notably, replacing the Nterminus with another well-structured protein moiety, the titin I27 repeats, reduced vinculin binding and caused the binding reaction to switch from exothermic to endothermic, suggesting the M1 was stabilized. This observation explains the forcedependent vinculin binding to the αT-catenin M-region observed in recent biophysical experiments (47). The ability of the titin I27 repeats to partially restore autoinhibition also suggests that steric, nonspecific interactions rather than specific interdomain associations (e.g., salt bridges) between the αT-catenin N-terminus and M-region promote M1 folding and M-region autoinhibition. Interestingly, the first 56 residues of αE-catenin contribute to the inhibition of vinculin binding in addition to the M3 domain (23).

αT-catenin recruits vinculin
Expression of EGFP-αT-catenin was sufficient to restore cell-cell contacts and recruit vinculin in α-catenin-deficient R2/7 cells. Vinculin binding to αT-catenin requires actin binding by the latter and is tension dependent, similar to αEcatenin. This is consistent with the vinculin-binding site in the αT-catenin M1 domain being occluded in the absence of force. Notably, deletion of M3 in αT-catenin did not relieve autoinhibition, offering additional evidence that the intramolecular interactions required for autoinhibition in αT-catenin differ from αE-catenin.
Vinculin is recruited to αE-catenin to bolster the AJ connection to actin under mechanical load (24,33,34,56). Neither the physiological context nor the functional role of vinculin recruitment to αT-catenin have been established in vivo. Loss of αE-catenin from the mouse heart disrupts AJ formation and causes a marked decrease in vinculin expression and recruitment, despite the presence of αT-catenin, resulting in cardiomyopathy (51). The inability of αT-catenin to compensate for αE-catenin in the heart underscores how the two α-catenins, despite sharing core properties, are likely regulated by distinct mechanisms and play unique roles in junction organization and signaling.
Together, our data support a model in which the αT-catenin N-terminus functions to regulate M-region stability and autoinhibition. Our in situ data indicate that force is required to reveal the vinculin-binding site. We speculate that it may do so by separating the M-region from the N-terminus to remove steric hinderance rather than breaking a network of internal M-region salt bridges. This could provide cadherin-β-catenin-αT-catenin complexes in cardiomyocytes with distinct mechanical properties, allowing ligand binding and allosteric signaling over a unique force scale relative to αE-catenincontaining complexes. Although we have focused on vinculin binding to M1, intrinsic differences in αT-catenin M region organization could also affect other M1 as well as M2 and M3 ligand interactions, possibly independent of force. Intramolecular and allosteric interactions are emerging as an important factor in the regulation of α-catenin conformation and molecular complex formation at the AJ. Further work is needed to define how molecular differences between the αcatenin protein family regulate mechanical adhesion and signaling.

Recombinant protein expression and purification
GST-tagged and His-tagged fusion proteins were expressed in BL21-Gold E. coli cells and purified as described (38,57). GST-tagged proteins were bound to glutathione-agaroseconjugated beads, whereas His-tagged proteins were bound to Ni-NTA beads. Bound beads were then equilibrated in cleavage buffer (20-mM Tris, pH 8.0, 150-mM NaCl, 2-mM EDTA, 10% glycerol, and 1-mM DTT or BME) and incubated with tobacco etch virus protease overnight at 4 C to cleave proteins from the respective tag. Proteins were then purified by Mono Q or Mono S ion-exchange chromatography at 4 C, followed by S200 gel-filtration chromatography at 4 C. Purified proteins were eluted in 20-mM Tris, pH 8.0, 150-mM NaCl, 10% glycerol, and 1-mM DTT, concentrated to working concentrations using a Millipore column concentrator and flashfrozen in liquid nitrogen.

Limited proteolysis and Edman degradation sequencing
Proteins were diluted to 12 μM in 20-mM Tris, pH 8.0, 150-mM NaCl, and 1-mM DTT and incubated at room temperature (RT) in 0.05 mg/ml sequencing grade trypsin (Roche Applied Science). Digestions were stopped with 2X Laemmli αT-catenin autoinhibition sample buffer and placed on ice until analysis. Samples were boiled and run by SDS-PAGE and then stained with 0.1% Coomassie Brilliant Blue R-250, 40% ethanol, and 10% acetic acid. Gels were scanned on a LI-COR scanner. For N-terminal sequencing, digested peptides were blotted onto polyvinylidene difluoride membrane, stained (0.1% Coomassie Brilliant Blue R-250, 40% methanol, and 1% acetic acid), destained, and dried. Individual bands were excised from the membrane and sequenced by Edman degradation (Iowa State University Protein Facility).

Crosslinking experiments
αT-catenin protein fragments were incubated with or without 1-mM BS3 (Thermo Scientific) in 20-mM Hepes, pH 7.4, 150-mM NaCl, and 1-mM DTT for 30 min at 4 C or 37 C, separated by SDS-PAGE, stained with the Coomassie dye, and imaged on a LI-COR scanner.

ITC titration calorimetry
Proteins used for ITC were purified as described except the S200 buffer was replaced with the ITC buffer (20-mM Hepes, pH 8.0, 150-mM NaCl, 1-mM TCEP). An identical buffer was used to purify both cell and titrant samples to ensure buffer match. Only fresh, unfrozen proteins were used for ITC. Measurements were performed on a Malvern MicroCal PEAQ-ITC or MicroCal VP-ITC calorimeter (Malvern Panalytical). For experiments on the MicroCal PEAQ-ITC, titration occurred by an initial 0.5-μl injection followed by 18 × 2μl injections of 110-to 150-μM β-catenin, 103-to 158-μM βcatenin-Ncad cyto complex, or 396-to 600-μM vinculin aa 1 to 259 (D1) into the cell containing 9-to 55-μM of αT-catenin or αE-catenin. For experiments on the MicroCal VP-ITC, the ligand was added with an initial 2-μl injection followed by 32 × 9-μl injections. The concentration of αT-catenin head or αTcatenin head-β-catenin complex in the cell varied between 22 and 56 μM. Vinculin D1 concentrations in the syringe varied between 240 and 546 μM. All calorimetry experiments were performed at 25 C. All data analyses were performed using Malvern MicroCal ITC analysis software. For baseline correction, a mean baseline value, calculated by averaging the data points at saturation, was subtracted from each data point.

Immunostaining and confocal microscopy
Cells were fixed in 4% paraformaldehyde in PHEM buffer (60-mM Pipes, pH 7.0, 25-mM Hepes, pH 7.0, 10-mM EGTA, pH 8.0, 2-mM MgCl, and 0.12 M sucrose), washed with PBS, permeabilized with 0.2% Triton X-100 in PBS for 2 min, and then blocked for 1 h at RT in PBS +10% BSA. Samples were washed three times in PBS, incubated with the primary antibody in PBS +1% BSA for 1 h at RT, washed three times in PBS, incubated with the secondary antibody in PBS +1% BSA for 1 h at RT, washed three times in PBS, and mounted on the ProLong Diamond mounting medium. Cells were imaged on a Nikon Eclipse Ti inverted microscope outfitted with a Prairie swept-field confocal scanner, Agilent monolithic laser launch, and Andor iXon3 camera using NIS-Elements imaging software.

Image analysis
To quantify vinculin recruitment to EGFP-tagged αEcatenin and αT-catenin, a maximum projection was created from four planes of the z-stack (600-nm total distance) where cell-cell contacts were best in focus. IsoJ Dark thresholding was used to create a mask of the GFP channel to define the region of analysis in ImageJ. Vinculin signal intensity was then measured within the masked region. Next, three random intensity measurements of vinculin staining were taken in the cell cytoplasm and these values averaged. Finally, the vinculin intensity within the mask was divided by the cytoplasmic signal to normalize between samples and calculate the contact/cytoplasmic ratio. Colocalization data were plotted with Prism software (GraphPad). A one-way ANOVA with Tukey's comparisons was performed to determine significance; p < 0.05.
To examine the relationship between EGPF-αE-catenin or EGFP-αT-catenin levels and vinculin recruitment, the vinculin contact/cytoplasmic ratio was plotted against the average GFP intensity of each fusion construct within the masked cell contact region (described above). Linear regression analysis was performed to calculate the slope, 95% confidence intervals, R 2 value, and p value using Prism software (GraphPad).

Data availability
All data are contained within the article.
Acknowledgments-We thank Jie Yan for sharing the titin I27-αTcatenin construct and Cara Gottardi for the R2/7 cells.