Cellular stiffness sensing through talin 1 in tissue mechanical homeostasis

Abstract

Tissue mechanical properties are determined mainly by the extracellular matrix (ECM) and actively maintained by resident cells. Despite its broad importance to biology and medicine, tissue mechanical homeostasis remains poorly understood. To explore cell-mediated control of tissue stiffness, we developed mutations in the mechanosensitive protein talin 1 to alter cellular sensing of ECM. Mutation of a mechanosensitive site between talin 1 rod-domain helix bundles R1 and R2 increased cell spreading and tension exertion on compliant substrates. These mutations promote binding of the ARP2/3 complex subunit ARPC5L, which mediates the change in substrate stiffness sensing. Ascending aortas from mice bearing these mutations showed less fibrillar collagen, reduced axial stiffness, and lower rupture pressure. Together, these results demonstrate that cellular stiffness sensing contributes to ECM mechanics, directly supporting the mechanical homeostasis hypothesis and identifying a mechanosensitive interaction within talin that contributes to this mechanism.

Fig. 1.. Mutations in rod domains alter their stability.

(A) Domain organization of talin showing destabilizing mutations introduced into the R1R2, R3, and R10 domains. (B and C) CD spectra showing molar ellipticity at a wavelength of 220 nm versus temperature in R1R2 (B) and R3 (C) mutant fragments compared to WT fragments.

Fig. 2.. Effects of rod-domain mutations on focal adhesions.

(A) Representative intensity-inverted images showing focal adhesions in Tln1^−/− MEFs expressing WT and rod-domain mutants plated for 3 hours on a fibronectin-coated glass bottom dish. Scale bars, 20 μm. (B) Quantification of focal adhesion area in (A). N = 25 cells for each sample. Boxes represent 25th to 75th percentile, whiskers represent 10th to 90th percentile, lines denote the median, and dots represent the mean. Statistics were analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (C) Quantified ratio of talin to paxillin within focal adhesions. N = 25 cells in each box plot. Boxes represent 25th to 75th percentile, whiskers represent 10th to 90th percentile, lines denote the median, and dots represent the mean. Statistics were analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (D) FRAP time course for WT and R1R2 mutant talin within individual focal adhesions. Data are means of % signal relative to starting values ± SEM. N = 10 cells per sample. (E) Representative immunofluorescence images of tensin-1 (green) in Tln1^−/− MEFs expressing WT and rod-domain mutants (red) plated on glass for 48 hours. (F) Quantification of mean fibronectin intensity per cell in Tln1^−/− MEFs expressing WT or mutant talin forms. N = 15 fields of view from three experiments. Boxes represent 25th to 75th percentile, whiskers represent 10th to 90th percentile, lines denote the median, and dots represent the mean. Statistics were analyzed by one-way ANOVA with Kruskal-Wallis post hoc test.

Fig. 3.. Effect of talin mutation on tension and stiffness sensing.

(A) FRET index of WT and mutant talin in focal adhesions normalized to a C-terminal Talin sensor (CTS) control. Values are means ± SEM of N = 59 to 77 cells from three independent experiments. Statistics were analyzed by one-way ANOVA Kruskal-Wallis post hoc test. (B) Quantification of spread area of cells expressing WT and talin mutants on polyacrylamide gels of varying stiffness. Values are means ± SEM of N = 52 to 78 cells. Statistics analyzed by two-way ANOVA with Tukey’s post hoc test. (C) Quantification of YAP localization in cells expressing WT and R1R2 mutant talin on compliant substrates. Values are mean ± SEM of N = 25 cells per sample. Statistics analyzed by two-way ANOVA with Tukey’s post hoc test. (D) Average stress (force per unit area) in WT- and R1R2 mutant–expressing cells on 3- and 30-kPa substrates. Values are mean ± SEM of N = 25 cells. Statistics analyzed by two-way ANOVA with Tukey’s post hoc test.

Fig. 4.. Effect of vinculin depletion on R1R2 mutant phenotypes.

(A) Normalized FRET index of WT and R1R2 mutant talin expressed in Tln1^−/− MEFs with vinculin depletion by guide RNA against vinculin (Tln1^−/−/SgVinculin) or Tln1^−/− MEFs with control nontargeting gRNA (NT). Values are means ± SEM, N = 25 cells. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (B) Representative images of WT- and R1R2-expressing Tln1^−/− and Tln1^−/−/SgVinculin MEFS. Scale bar, 5 μm. (C) Focal adhesion area in Tln1^−/− and Tln1^−/−/SgVinculin MEFS. Boxes represent 25th to 75th percentile, whiskers represent 10th to 90th percentile, and lines denote the median. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (D) Quantification of cell spread area in WT- or R1R2-expressing Tln1^−/− and Tln1^−/−/SgVinculin MEFS. Values are means ± SEM of N = 32 to 54 cells. Statistics were analyzed by two-way ANOVA with Tukey’s post hoc test.

Fig. 5.. Effects of R2 core versus R1R2 interface interactions.

(A) Schematic of R2 core versus R1R2 interface mutations. (B) Cell spread area of WT and R1R2 core and interface talin mutants on 3- and 30-kPa substrates. Values are means ± SEM of N = 3 experiments. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (C) Focal adhesion area of WT and R1R2 core and interface mutants in Tln1^−/− MEFs. N = 23 to 28 cells. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (D) Normalized FRET index of WT and R1R2 core and interface mutant talin in focal adhesions. Values are means ± SEM of N = 25 cells. Statistics were analyzed by one-way ANOVA with Kruskal-Wallis post hoc test.

Fig. 6.. ARPC5L mediates talin R1R2-dependent stiffness sensing.

(A) Spread area in cells expressing WT talin versus destabilized mutant R1R2 talin on 3- and 30-kPa stiffness substrates in the presence of dimethyl sulfoxide (DMSO), ARP2/3 inhibitor (CK666), formin inhibitor (SMIFH2), and microtubule polymerization inhibitor (nocodazole). Values are means ± SEM of N = 25 to 50 cells. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (B) Representative images showing focal adhesions with and without CK666 in cells expressing WT and R1R2 mutant. Scale bars, 10 μm. (C) Quantified focal adhesion area from (B). N = 20 to 25 cells. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test. (D) Representative blots showing coimmunoprecipitation of Flag-tagged ARPC5L and Flag-tagged ARPC5 with the indicated GFP-tagged talin constructs. (E) Quantified cell spread area in WT- and R1R2-expressing cells with (SiARPC5L) and without ARPC5L depletion [nontargeting siRNA (NT)] on hydrogels of varying stiffness. Values are means ± SEM of N = 25 to 50 cells. Statistics analyzed by two-way ANOVA with Tukey’s post hoc test. NS, not significant. (F) Quantification and (G) representative images of WT- and R1R2 mutant–expressing cells with and without ARPC5L depletion. Scale bars, 10 μm. N = 37 to 66 cells. Statistics analyzed by one-way ANOVA with Kruskal-Wallis post hoc test..

Fig. 7.. R1R2 mutant knock-in mice.

(A) Schematic showing the insertion of L638D and V722R mutations in exon 17 and exon 19 using a 1.6-kb template. (B) Genotype of progeny from mating heterozygous Tln1^{L638D-V722R/+} mice (Tln1^KI-het). χ² analysis revealed no significant difference. χ² equals 3.658 with 2 degrees of freedom. The two-tailed P value equals 0.1606 and is not statistically significant. (C) Body mass of WT and R1R2 mutant knock-in mice from 3 to 14 weeks of age. Values are means ± SEM. Statistics analyzed by unpaired t test. (D) Compressive radial stiffness of the ascending aorta of P24 WT and R1R2 mutant mice measured by atomic force microscopy applied from the adventitial surface. Values are means ± SEM of N = 6 mice. Statistics analyzed by unpaired t test. (E to L) Ex vivo biomechanical measurements of ascending aorta of P24 WT and R1R2 mutant mice N = 4 for (H) and N = 5 for (E) to (G) and (I) to (L).

Fig. 8.. ECM in R1R2 mutant aorta.

(A) Picrosirius red staining (left) and total Movat’s pentachrome staining (middle). Collagen content in Movat stains extracted by custom MATLAB software (74) (right) for ascending aortas from WT and R1R2 mutant mice at P24. Scale bar, 100 μm. (B) Quantification of total collagen fibers (red, orange, yellow, and green fibrils) identified by picrosirius red staining in the adventitia and media. Zoomed-in images are presented as fig. S9B. Values are means ± SEM of N = 5. Statistics analyzed by unpaired t test. (C) Quantification of total collagen area within adventitia and media of ascending aorta at P24 by Movat staining. Values are means ± SEM of N = 5. Statistics analyzed by unpaired t test.