Vinculin is Essential For Sustaining Normal Levels of Endogenous Force Transmission at Cell-Cell Contacts

Transmission of cell-generated (i.e., endogenous) tension at cell-cell contacts is crucial for tissue shape changes during morphogenesis and adult tissue repair in tissues like epithelia. E-cadherin-based adhesions at cell-cell contacts are the primary means by which endogenous tension is transmitted between cells. The E-cadherin-β-catenin-α-catenin complex mechanically couples to the actin cytoskeleton (and thereby the contractile machinery of the cell) both directly and indirectly. However, the key adhesion constituents required for substantial endogenous force transmission at these adhesions in cell-cell contacts are unclear. Due to the role of α-catenin as a mechanotransducer that recruits vinculin at cell-cell contacts, we expected α-catenin to be essential for the high levels of force transmission normally sustained. Instead, using the traction force imbalance method to determine the inter-cellular force at a single cell-cell contact between cell pairs, we found that it is vinculin that is essential for high endogenous force transmission. Our results constrain the potential mechanical pathways of force transmission at cell-cell contacts and suggest that vinculin can transmit forces at E-cadherin adhesions independent of α-catenin, possibly through β-catenin. Furthermore, we tested the ability of cell-cell contacts to withstand external stretch and found that vinculin is essential to maintain cell-cell contact stability under external forces as well.


Introduction
Cell-cell contacts in tissues like epithelia are interfaces where cell-generated and external forces are transmitted from cell to cell and thereby across the tissue [1,2]. During morphogenesis, cell-generated tension transmitted through cell-cell contacts is essential for cell shape changes as well as supracellular morphological transformations [3]. Even pathological events such as cancer metastasis and cell-to-cell transmission of some pathogens involve changes in the forces transmitted at cell-cell contacts. These cell-cell contacts are bound by many types of adhesions, but E-cadherin adhesions are chiefly important in the integrity and mechanical function of these cell-cell contacts [4]. While biophysical and biomimetic approaches have broadened our understanding of E-cadherin adhesions and their response to forces [5,6], there is much unknown about endogenous force transmission through E-cadherin adhesions in the native context of lateral contacts between cells. It is also unclear as to what specific factors determine the adhesion strength of these lateral cell-cell contacts when subject to external forces. E-cadherin forms a 1:1 complex with β-catenin which in turn binds to α-catenin [7]. This E-cadherin-β-catenin-α-catenin complex can couple to the actin cytoskeleton by directly binding to actin, with enhanced binding of α-catenin to F-actin under force [8][9][10]. The E-cad-catenin complex can also couple (or potentially couple) to actin via the adhesion-associated protein vinculin [11], actin cross-linker α-actinin [12], tight junction protein ZO-1 [13], F-actin binding proteins afadin [14] or EPLIN [15] and the formin Fmn1 [16]. In particular, α-catenin has been shown to function as an elastic link in series with cadherin and actin [17] that transitions to an open conformation under force that then recruits vinculin [18]. Accordingly, such recruitment of vinculin has been shown to depend on non-muscle myosin II activity in cells [19]. Vinculin can also be recruited to E-cadherin adhesions via myosin VI [20] as well as β-catenin dependent ways [11,21,22], suggesting that vinculin may play a significant role in cell-to-cell force transmission.
However, how much vinculin affects the endogenous force transmission at cell-cell contacts between epithelial cells, and how this compares to the contribution of a presumably more constitutive component like α-catenin, has not been directly assessed.
The mechanical function of cell-cell adhesion associated proteins are at least two-fold: Transmission of mechanical forces from cell to cell as well as maintenance of the strength of these adhesions. Many of the approaches used to study the role of cell-cell adhesion proteins in force transmission employ biochemical methods at cell-cell contacts themselves, or quantitative methods with biomimetic interfaces such as cadherin-coated substrates [5,[23][24][25][26][27][28][29][30], cadherin-coated beads [31,32] and suspended cell doublets [33]. Quantitative approaches such as FRET-based sensors can look at force transmission through specific proteins at cell-cell contacts [6,34,35], but the total forces transmitted via cell-cell contacts are not known in this context. Due to the presence of multiple adhesion systems at cell-cell contacts, determining the total endogenous force transmitted at cell-cell contacts as such can help identify the overall effect of perturbations like specific protein knockdowns. In a similar manner, while cadherin-coated substrates and suspended cell doublets have enabled key insights into determinants of E-cadherin adhesion strength, assessment of adhesion strength of lateral cell-cell contacts between epithelial cells is essential to understand how this interface ultimately resists mechanical challenges and the role played by specific cell adhesion associated proteins such as vinculin. Here, we test the importance of putative physical pathways of force transmission by knocking out proteins prominently known to be involved in mechanotransduction at E-cadherin adhesions. We find that, contrary to expectation, vinculin rather than α-catenin is crucial for transmitting high endogenous tension at cell-cell contacts. We also use large external stretching to find that vinculin is essential for maintaining cell-cell contact integrity under external stretch, highlighting the crucial mechanical role of vinculin at epithelial cell-cell contacts. To generate knockout (KO) cells, CRISPR/Cas9 was used with the gRNA sequence CACGAGGAAGGCGAGGTGGA for vinculin (previously shown [36] to knockout vinculin) and the gRNA sequence TCTGGCAGTTGAAAGACTGT for α-catenin (previously shown [36] to knockout α-catenin). The gRNA sequences were used in the Sigma All-in-One U6-gRNA/CMV-Cas9-tGFP Vector). Cells were transiently transfected with this vector (with the appropriate gRNA), followed by clonal expansion. Clones were screened for vinculin or α-catenin loss using Western blotting. For the double (α-catenin and vinculin) KO, vinculin KO cells were used to generate an additional KO of α-catenin. α-catenin with its vinculin binding site (amino acids 316-405 in α-catenin) replaced by a homologous similar sequence from vinculin (amino acids 514-606 in vinculin), called α-catenin DVBS (delta vinculin binding site) [17,37] was used to generate the α-catenin DVBS in α-catenin KO cell line (Addgene plasmid 178649).

Western Blotting
Cells were washed with phosphate-buffered saline (PBS) and lysed using RIPA buffer supplemented with protease as well as phosphatase inhibitors. SDS-PAGE of the proteins was followed by Western blotting using PVDF membranes. The proteins on the PVDF membranes were incubated in 5% bovine serum albumin (BSA) in PBS for 1 h at room temperature followed by primary antibody incubation overnight. Then, the samples were rinsed in 0.2% Tween in PBS and incubated for 45 min with HRP-conjugated secondary antibody in 0.2% Tween in PBS. After rinsing with PBS, blot chemiluminescence was imaged using a Bio-Rad ChemiDoc system. Primary antibodies used for blotting were anti-rabbit α-catenin (catalogue# C2081 from Sigma, St.
Fluorophore-conjugated secondary antibodies from Jackson ImmunoResearch or ThermoFisher were used in all staining experiments.

Preparation of Soft Silicone Substrate
Soft silicone (Qgel 300, CHT USA Inc., Richmond, VA) was prepared by mixing its A and B components at a 1:2.2 ratio. The gel mixture was cured using a heater at 100 0 C for an hour. To use these silicone substrates for traction force microscopy, fluorescent beads and collagen I were coupled as follows. After curing, the silicone was exposed to 305 nm UV light (

Rheology of Soft Silicone
The shear rheology of the soft silicone was characterized using an MCR-302 rheometer (Anton Paar, Ashland, VA). Presence of an air bearing in the rheometer enabled measurement of the moduli of soft samples (kPa and below). The soft silicone was prepared and cured as above and loaded between 25 mm diameter parallel plates. The storage and loss shear moduli were obtained as a function of angular frequency for 1% strain (determined to be in the linear range using a strain sweep). The average of the shear storage modulus (G') in the 0.1 to 1 rad/s range was considered to be the nominal G'.

Traction Force Microscopy and Traction Force Imbalance Method
A phase image of each MDCK cell or cell pair along with the correspondent image of beads beneath were first recorded. After the cells were disintegrated using 1% sodium dodecyl sulfate, an image of the beads on the relaxed substrate were recorded. The stressed substrate bead images (in the presence of cells) and the relaxed bead images (in the absence of cells) were aligned using an ImageJ plugin [38]. The displacement field was then computed using mpiv (https://www.mathworks.com/matlabcentral/fileexchange/2411-mpiv), scripted in MATLAB (MathWorks, Natick, MA). Traction stresses are then reconstructed using regularized Fourier Transform Traction Cytometry that uses the Boussinesq solution, such as in previously published work [38][39][40][41][42][43][44]. The Traction Force Imbalance Method [26,39,41] was then used to compute the inter-cellular force at the cell-cell contact within a cell pair from the vector sum of traction forces under each cell within the cell pair.

Biaxial Stretch of Epithelial Islands
A 0.01" thick silicone sheet (Speciality Manufacturing, Saginaw, MI) was exposed to 305 nm UV light for five minutes and then incubated with collagen I at 37 0 C, under 5% CO2 for 15 minutes.

Statistical Analysis
For statistical analysis, t-test was used to compare wildtype and vinculin KO single cell data ( fig.   1 and 5) and analysis of variance (ANOVA) was used for multiple comparisons of all the cell pair data ( fig. 2, 3 and 4), with * indicating p < 0.05, ** indicating p < 0.01 and *** indicating p < 0.001.

Results and Discussion
Vinculin is known to be recruited to E-cadherin mediated adhesions at cell-cell contacts under the action of endogenous forces [19]. However, it is known that vinculin can also be recruited at cell-cell contacts in the absence of myosin mediated contractility [45]. Thus, we wanted to test whether vinculin is an essential element for sustaining high levels of endogenous force at cell-cell contacts or if its effect on endogenous inter-cellular tension is only marginal. To this end, we first generated a CRISPR knockout (KO) of vinculin in MDCK cells ( fig. S1, fig.   1A,B). Immunofluorescence staining for vinculin marks focal adhesions in wildtype (WT) MDCK cells ( fig. 1A), but this is not the case for MDCK vinculin KO cells ( fig. 1B). Vinculin is an important focal adhesion protein as well [46][47][48], and previous reports [49,50] have shown that the absence of vinculin decreases the traction force exerted by fibroblasts onto the ECM. Vinculin was also recently shown to be essential for high force transmission through focal adhesions in HeLa cells [51]. However, another study [52] reported that vinculin knockdown did not significantly decrease traction forces exerted by mesenchymal stem cells. Therefore, we first We then asked if the absence of vinculin affects the level of endogenous force transmitted via cell-cell contacts. To answer this, we used the Traction Force Imbalance Method (TFIM) [39].
Unlike an isolated cell, where the vector sum of traction forces vanishes (within experimental error), for each cell within a cell pair, the vector sum of traction forces is not balanced as such, but this imbalance in traction force corresponds to the inter-cellular force that is required for physical force balance for each cell in the cell pair. TFIM has been previously used to measure inter-cellular forces within endothelial cell pairs [54] and epithelial cell pairs undergoing dynamic cell rearrangements [55] and epithelial cell sheets [56]. We measured the traction forces for WT ( fig.   2A   Since it is α-catenin that is considered to be the primary recruiter of vinculin to E-cadherin adhesions, we surmised that α-catenin would be at least as important as vinculin in force transmission through cell-cell contacts. The centrality of α-catenin in potential force transmission pathways of E-cadherin to F-actin via many candidate proteins such as afadin, EPLIN and ZO-1, among others, also suggested that it would be a critical component of the effective mechanical pathway at cell-cell contacts. We thus expected the absence of α-catenin to decrease the endogenous forces at cell-cell contacts even more severely than the absence of vinculin. To test this, we generated an MDCK α-catenin KO cell line ( fig. S1). We noticed that cell-cell contacts in α-catenin KO cell monolayers show the localization of both actin and E-cadherin ( fig. S3). We then measured traction forces for cell pairs ( fig. 3A,B) and then used TFIM to determine the endogenous force transmitted in cell-cell contacts within α-catenin KO cell pairs. To our surprise, we found that the inter-cellular tension for α-catenin KO cell-cell contacts was 39 ± 23 nN -not (statistically) significantly lesser than that for WT contacts (fig. 3E). However, our results are consistent with more qualitative laser ablation results -α-catenin knockdown (KD) was previously shown [21] to cause only a minor decrease in cell-cell tension as assessed by the retraction of the ablated ends of cell-cell contacts. α-catenin KD cells were also shown [37] to exert only slightly reduced traction forces on E-cadherin-coated substrates compared to WT cells.  Since vinculin is just one of many potential α-catenin binding partners that can transmit force to E-cadherin adhesion via the actin cytoskeleton, we wanted to test the specific role of the α-catenin-vinculin interaction. We exogenously expressed α-catenin lacking the vinculin binding site (VBS) (α-catenin DVBS) in MDCK α-catenin KO cells. We then measured the traction forces for MDCK α-catenin DVBS cell pairs ( fig. 3C,D) and then used TFIM to determine the intercellular force. The inter-cellular force for MDCK α-catenin DVBS cell-cell contacts was 36 ± 20 nN -not (statistically) significantly lesser than that for WT contacts, similar to α-catenin KO cellcell contacts (fig. 3E). We further tested whether the absence of both α-catenin and vinculin  This result is consistent with the previously reported observation [37] that cells expressing αcatenin DVBS exert similar traction forces on E-cadherin-coated substrates as cells lacking αcatenin. Thus, it is vinculin rather than α-catenin that is essential for transmitting high endogenous forces at cell-cell contacts. While we used α-catenin KO as an experimental tool, it has been suggested that the decoupling of α-catenin from the E-cadherin-β-catenin complex may be physiologically relevant in cadherin junction disassembly in some contexts [16].
Given the essential role that we found for vinculin in the exertion of high endogenous intercellular forces, we wanted to know if vinculin also performs a similar essential role in protecting cell-cell contact integrity under mechanical challenges. In order to test this role for vinculin at cellcell contacts, we wanted to use a method that can directly test the integrity of lateral cell-cell  5I). Thus, vinculin is not only essential for high endogenous force transmission, but also for maintaining cell-cell contact integrity under high external forces. These results are consistent with previous reports of a mechanoprotective role for vinculin at E-cadherin adhesions, suggested by its recruitment to sites of forces exerted via E-cad-beads [32,57]. Our results are also consistent with prior results using suspended doublets [58], detached cell sheets [36] and E-cadherin-coated substrates [59] that indicated a role for vinculin in maintaining cell-cell contact integrity. Our results with cell-cell contacts also complements the known role for vinculin in maintaining the integrity of cell-ECM contacts under cell-generated tension [60].

Conclusion
Force transmission through epithelial cell-cell contacts plays a pivotal role in dynamic events during morphogenesis and adult tissue repair. In this report, we show that vinculin is essential for transmitting high levels of endogenous force through cell-cell contacts. Our results not only suggest that the α-catenin-vinculin complex is not necessary for transmitting high endogenous tension through cell-cell contacts, but also that α-catenin's interaction with other proteins like afadin, EPLIN or ZO-1 is not essential for transmitting high inter-cellular tension.
The α-catenin-vinculin may thus be just one of many active mechanical links for transmitting forces through cell-cell contacts. Through what other links may high levels of cell-generated forces be transmitted from the actomyosin apparatus to E-cadherin adhesions at cell-cell contacts? Our results are consistent with previously proposed interactions such as β-catenin-vinculin playing a mechanical role [21]. In fact, vinculin and α-catenin bind to the same N-terminal region of βcatenin [59,61]. There is also evidence for the β-catenin-vinculin interaction in cancer cells that lack α-catenin [61] and for an α-catenin-independent means by which β-catenin can couple to the actin cytoskeleton [62]. However, the nanoscale positioning of vinculin is a bit displaced from βcatenin in α-catenin KD cells [21], suggesting that other intermediate molecular linkers may play a role. Vinculin can also be recruited to epithelial cell-cell contacts in a myosin VI dependent manner [20]. In any case, our results suggest that vinculin plays a key role at cell-cell contacts in addition to its established role of being recruited to α-catenin under specific force inputs.
Vinculin's ability to support high junctional tension as well as high contact strength, as shown here, is consistent with its essential role not only at cell-cell contacts in epithelia, but also other tissues undergoing dynamic events [63], including those like cardiac tissues [64], where endogenous forces reach even higher values. It is likely that high force transmission through vinculin enables the enhanced adhesion strength of cell-cell contacts, reminiscent of force coupled stabilization reported at focal adhesions [65,66]. Vinculin's key mechanical role at cell-cell contacts may also potentially explain why a bacterial pathogen has evolved to specifically bind to it to reduce cell-cell tension and promote its spread from cell to cell [67].