Mechanosensitive recruitment of Vinculin maintains junction integrity and barrier function at epithelial tricellular junctions

Abstract

Apical cell-cell junctions, including adherens junctions and tight junctions, adhere epithelial cells to one another and regulate selective permeability at both bicellular junctions and tricellular junctions (TCJs). Although several specialized proteins are known to localize at TCJs, it remains unclear how actomyosin-mediated tension transmission at TCJs contributes to the maintenance of junction integrity and barrier function at these sites. Here, utilizing the embryonic epithelium of gastrula-stage Xenopus laevis embryos, we define a mechanism by which the mechanosensitive protein Vinculin helps anchor the actomyosin network at TCJs, thus maintaining TCJ integrity and barrier function. Using an optogenetic approach to acutely increase junctional tension, we find that Vinculin is mechanosensitively recruited to apical junctions immediately surrounding TCJs. In Vinculin knockdown (KD) embryos, junctional actomyosin intensity is decreased and becomes disorganized at TCJs. Using fluorescence recovery after photobleaching (FRAP), we show that Vinculin KD reduces actin stability at TCJs and destabilizes Angulin-1, a key tricellular tight junction protein involved in regulating barrier function at TCJs. When Vinculin KD embryos are subjected to increased tension, TCJ integrity is not maintained, filamentous actin (F-actin) morphology at TCJs is disrupted, and breaks in the signal of the tight junction protein ZO-1 signal are detected. Finally, using a live imaging barrier assay, we detect increased barrier leaks at TCJs in Vinculin KD embryos. Together, our findings show that Vinculin-mediated actomyosin organization is required to maintain junction integrity and barrier function at TCJs and reveal new information about the interplay between adhesion and barrier function at TCJs.

Keywords: Vinculin; Xenopus; actin; adherens junction; barrier function; epithelium; live microscopy; optogenetics; tension; tricellular junctions.

Figure 1:. Vinculin is mechanosensitively recruited to TCJs.

A) Schematic of the TULIP optogenetic system to activate RhoA. LOVpep is anchored to the plasma membrane (PM). Upon 405 nm light stimulation, LOVpep undergoes a conformational change allowing it to interact with the photo-recruitable GEF (prGEF) and activate RhoA. B) Schematic of extracellular ATP addition. Xenopus embryos are mounted on a custom-made metal slide and are sandwiched between two glass coverslips. The top coverslip only covers 50% of the hole in the slide, allowing an opening to add ATP to the embryo during live confocal imaging. C) Schematic of observed cellular responses to optogenetic activation of RhoA and additional of extracellular ATP. Arrows represent expected forces, with thicker arrows representing more force. D) Live confocal images of epithelial cells expressing Vinculin (Halo-Vinculin with JF646), photo-recruitable GEF (prGEF-YFP), and LOVpep (GFP-silent-LOVpep). Images were captured before and during RhoA activation using optogenetic stimulation. Zoomed in panels (indicated by boxes) highlight changes in Vinculin recruitment at TCJs (right) and BCJs (bottom). E) Quantification of Halo-Vinculin intensity at TCJs before and during RhoA activation. Measured signal was normalized to cytosolic signal, and “before RhoA activation” was set to 1. Statistics, paired t-test; n = 3 experiments, 9 embryos, 55 TCJs; p ≤ 0.0001 (****). Violin plots show the median (dashed line) and the 25^th and 75^th quartiles (dotted lines). F) Live confocal images of cells expressing Vinculin (mNeon-Vinculin) before and after extracellular ATP addition. Zoomed in panels (indicated by boxes) highlight changes in Vinculin recruitment at TCJs (right) and BCJs (bottom). G) Quantification of mNeon-Vinculin intensity at TCJs before and after ATP addition. Measured signal was normalized to cytosolic signal, and “before ATP” was set to 1. Statistics, paired t-test; n = 3 experiments, 5 embryos, 29 TCJs; p ≤ 0.0001 (****). Violin plots show the median (dashed line) and the 25^th and 75^th quartiles (dotted lines). See also Figure S1 and S2 and Video S1 and S2.

Figure 2:. Vinculin’s actin-binding function is needed for proper actomyosin organization and stability at TCJs.

A) Fixed confocal images of epithelial cells from control embryos, Vinculin knockdown embryos (Vinculin KD), Vinculin knockdown embryos injected with wildtype Vinculin mRNA (KD + WT), and Vinculin knockdown embryos injected with mRNA encoding an actin-binding mutant of Vinculin (KD + R1049E) that were stained for F-actin (phalloidin Alexa Fluor 568) and phosphomyosin (α-phosphomyosin light chain 2 antibody). Zoomed in panels (indicated by boxes) highlight changes at TCJs. B) and C) Normalized intensity of phalloidin and α-phosphomyosin at TCJs. Measured signal was normalized, and control was set to 1. Statistics, one-way ANOVA; n = 3 experiments; control = 25 embryos, 174 TCJs; Vinculin KD = 14 embryos, 90 TCJs; KD + WT = 23 embryos, 161 TCJs; KD + R1049E = 20 embryos, 140 TCJs; p-values > 0.05 (ns), ≤ 0.001 (***), ≤ 0.0001 (****). Violin plots show the median (solid line) and the 25^th and 75^th quartiles (dotted lines). D) Raw intensity line scans of phalloidin and α-phosphomyosin adjacent to TCJs in control, Vinculin KD, KD + WT, and KD + R1049E embryos. Locations of line scans are indicated by dashed yellow lines in (A). E) Left: Recovery curve (data points) for mNeon-Actin FRAP at TCJs with a double exponential nonlinear fit (solid line). n = 3 experiments; control = 8 embryos, 21 TCJs; Vinculin KD = 7 embryos, 18 TCJs; errors bars, SEM. Right: Recovery curve (data points) from mNeon-Actin FRAP at BCJs with a double exponential nonlinear fit (solid line). n = 3 experiments; control = 7 embryos, 20 BCJs; Vinculin KD = 8 embryos, 23 BCJs. F) Mobile fractions calculated from (E). Statistics, unpaired t-test; p ≤ 0.0001 (****); error bars, SEM. See also Figure S3 and S4.

Figure 3:. Vinculin stabilizes Angulin-1 at tricellular TJs.

A) Fixed confocal images from control and Vinculin KD embryos that were stained for Angulin-1 (α-Angulin-1). B) Normalized intensity of α-Angulin-1 at TCJs. Statistics, unpaired t-test with Welch’s correction; n = 3 experiments; control = 23 embryos, 170 TCJs; Vinculin KD = 25 embryos, 185 TCJs; p ≤ 0.0001 (****). Violin plots show the median (dashed line) and the 25^th and 75^th quartiles (dotted lines). C) Montage of Angulin-1–3xGFP FRAP in control and Vinculin KD embryos pre-bleach and post-bleach. Dashed circle indicates photobleached region. Images are shown using the FIRE lookup table (LUT). D) Recovery curve (data points) for Angulin-1–3xGFP FRAP at TCJs with a double exponential nonlinear fit (solid line). n = 3 experiments; control = 8 embryos, 15 TCJs; Vinculin KD = 6 embryos, 9 TCJs; errors bars, SEM. E) Mobile fractions calculated from (D). Statistics, unpaired t-test with Welch’s correction; p ≤ 0.0001 (****); error bars, SEM. F) Slow t_1/2 calculated from (D). Statistics, unpaired t-test with Welch’s correction; p ≤ 0.0001 (****); error bars, SEM. G) Fast t_1/2 calculated from (D). Statistics, unpaired t-test with Welch’s correction; p ≤ 0.01 (**); error bars, SEM. See also Figure S3 and S4 and Video S3.

Figure 4:. Vinculin maintains TCJ morphology and integrity under increased tension.

A) Live confocal images of control and Vinculin KD embryos that express an F-actin probe (LifeAct-RFP), photo-recruitable GEF (prGEF-YFP), and LOVpep (GFP-silent-LOVpep). Images of full Z-projections and apical Z-projections are shown at both baseline (before RhoA activation) and increased Rho-mediated tension (during RhoA activation). Yellow arrowheads point to dipped TCJs. Red bars indicate the TCJs that are shown in XZ views in red boxes. B) Live confocal images of control and Vinculin KD embryos that express ZO-1 (Halo-ZO-1 with JF646), α-catenin (α-catenin-mCherry), photo-recruitable GEF (prGEF-YFP), and LOVpep (GFP-silent-LOVpep). Max Z-projections are shown at both baseline tension (before RhoA activation) and increased Rho-mediated tension (during RhoA activation). Yellow arrowheads point to breaks in ZO-1 signal at TCJs. C) Quantification of distance F-actin dip at TCJs before and during RhoA activation in both control and Vinculin KD embryos. Statistics, 2-way ANOVA; n = 3 experiments; control = 4 embryos, 28 TCJs; Vinculin KD = 4 embryos, 28 TCJs; p-values > 0.05 (ns), ≤ 0.0001 (****). D) Quantification of size of ZO-1 breaks at TCJs before and during RhoA activation in both control and Vinculin KD embryos. Statistics, 2-way ANOVA; n = 2 experiments; control = 3 embryos, 30 TCJs; Vinculin KD = 3 embryos, 19 TCJs; p-values > 0.05 (ns), ≤ 0.0001 (****). See also Figure S3 and S5 and Video S4.

Figure 5:. Vinculin is required to maintain barrier function at TCJs.

A) Schematic of Zinc-based Ultrasensitive Microscopic Barrier Assay (ZnUMBA). Disruption of tight junction proteins leads to interaction between Zn²⁺ and FluoZin3, causing increased FluoZin3 fluorescence and indicating a leak in barrier function. B) Live confocal images of FluoZin3 signal in control and Vinculin KD embryos. Images are shown using the FIRE LUT, adjusted in the same way for each image. Gray arrows point to increased FluoZin3 signal at TCJs. C) Quantification of the percent of TCJs that exhibited leaky TCJs during the 25-minute movies. Statistics, unpaired t-test with Welch’s correction; n = 3 experiments; control = 4 embryos, 223 TCJs; Vinculin KD = 5 embryos, 133 TCJs; p-value ≤ 0.05 (*); error bars, SEM. D) Montages of FluoZin3 signal at representative TCJs in control and Vinculin KD embryos. Images are shown using the FIRE LUT; LUT adjustments are indicated for each montage. See also Figure S3 and Video S5.

Figure 6:. Model for Vinculin’s role at TCJs

A) 3D model of a TCJ showing Vinculin anchoring actin bundles at the TCJ to maintain junctional integrity. B) Top-down view of TCJs at baseline and increased tension for both controls and Vinculin KD embryos. Vinculin is mechanosensitively recruited to TCJs. When Vinculin is lost, actomyosin is disorganized, tricellular TJ proteins are less stable, and TCJs have an increased frequency of barrier leaks.