Shp2 plays a crucial role in cell structural orientation and force polarity in response to matrix rigidity

Abstract

Cells can sense and respond to physical properties of their surrounding extracellular matrix. We have demonstrated here that tyrosine phosphatase Shp2 plays an essential role in the response of mouse embryonic fibroblasts to matrix rigidity. On rigid surfaces, large focal adhesions (FAs) and anisotropically oriented stress fibers are formed, whereas cells plated on compliant substrates form numerous small FAs and radially oriented stress fibers. As a result, traction force is increased and organized to promote cell spreading and elongation on rigid substrates. Shp2-deficient cells do not exhibit the stiffness-dependent increase in FA size and polarized stress fibers nor the intracellular tension and cell shape change. These results indicate the involvement of Shp2 in regulating the FAs and the cytoskeleton for force maintenance and organization. The defect of FA maturation in Shp2-deficient cells was rescued by expressing Y722F Rho-associated protein kinase II (ROCKII), suggesting that ROCKII is the molecular target of Shp2 in FAs for the FA maturation. Thus, Shp2 serves as a key mediator in FAs for the regulation of structural organization and force orientation of mouse embyonic fibroblasts in determining their mechanical polarity in response to matrix rigidity.

The interactions between adherent cells and the extracellular matrix (ECM) are essential for many cellular processes, including proliferation, differentiation, and migration (1–3). It is known that cells can sense the extracellular mechanical cues and respond to these cues by modulating cellular biochemical activities and intracellular force, a process called mechanotransduction (4–7). The altered intracellular cytoskeletal forces modulate cell shape and control complex cell behaviors that are critical for tissue development and homeostasis (8–10). Focal adhesions (FAs) link the cell to the ECM at sites of integrin binding and play dual roles in force transmission and signal transduction (11, 12). When a nascent adhesion is formed, the activated integrin links to cytoskeleton via a talin-mediated connection and requires actin-related protein-2/3 (ARP2/3) complex-mediated actin polymerization that is independent of myosin II activity (13, 14). Thereafter, some adhesions disassemble, whereas other adhesions mature through myosin-mediated contractility that promotes the recruitment of additional cytoskeletal and signaling proteins to the FAs for adhesion strengthening and signal transmission (15–19). Therefore, the FAs are mechanosensitive, and the regulation of FA dynamics plays an important role in mechanotransduction (11, 12, 20).

Abnormal cell and tissue responses to mechanical stress have been shown to contribute to the etiology and clinical presentation of many diseases or developmental disorders (21). Tyrosine phosphatase Shp2, encoded by the PTPN11 gene, is a ubiquitously expressed, nonreceptor protein tyrosine phosphatase (PTP) (22). Germline mutations in PTPN11 cause Noonan syndrome (gain of function) and Leopard syndrome (catalytically defective), which involve developmental abnormalities and clinical features that are strongly associated with defects in mechanotransduction (23–26). A targeting deletion of PTPN11 exon 3 (encoding 46–110 aa) in homozygous mice led to embryonic death during gastrulation with severe defects in mesodermal patterning (27, 28), indicating the functional deficiency of Shp2 caused by exon 3 deletion. No phenotype was found in heterozygous mice. In addition, mouse embryonic fibroblasts (MEFs) derived from the Shp2^Ex3−/− embryos displayed severe defects in cell migration, which could be rescued by the reintroduction of wild-type Shp2 (29).

The spatiotemporal control of small GTPase RhoA/Rho-associated protein kinase (ROCK)-mediated actomyosin contractility plays a key role in regulating FA dynamics. A reduction in contractility leads to FA disassembly, and enhanced contractility causes FA stabilization and maturation; these changes affect the cellular response to extracellular stimuli such as matrix rigidity. Importantly, RhoA/ROCK signaling is involved in the regulation of stem cell lineage commitment by mechanical forces, emphasizing the importance of RhoA/ROCK signaling in mechanotransduction (2, 30, 31). Our previous studies showed that the RhoA-mediated ROCKII activation is negatively regulated by ROCKII phosphorylation at Y722 residue. Protein kinase Src and Shp2 phosphatase reciprocally modulate this regulation specifically at FAs (32, 33). In this report, we demonstrated that Shp2 plays a key role in the maturation of FAs and the regulation of stress-fiber orientation during the cellular response to matrix rigidity. By traction force microscopy (TFM), we found that Shp2^Ex3−/− cells bear lower intracellular tension than wild-type cells (Shp2^WT), confirming the involvement of Shp2 in regulating the FAs and the cytoskeleton for force maintenance and organization. Moreover, we demonstrated the recruitment of Shp2 in FAs and its correlation with the decrease of ROCKII Y722 phosphorylation in FAs. By rescuing FA maturation via expression of Y722F ROCKII in Shp2^Ex3−/− cells, we also provide evidence that ROCKII is the molecular target of Shp2 in FAs. In summary, our study provides new insight into the role of Shp2 in mechanotransduction by showing that, when cells respond to the increase of matrix rigidity, Shp2 participates in FAs to promote their maturation through ROCKII Y722 dephosphorylation, which in turn contributes to stress-fiber polarization and enhanced intracellular tension.

Results

Shp2 Is Required for Matrix-Stiffness–Mediated FA Maturation and Stress-Fiber Polarization.

FAs anchor cells to the ECM, serving as mechanosensors and providing mechanical support for stress fibers that bear the intracellular tension (11, 12). To test whether Shp2 is involved in the cellular response to matrix rigidity, Shp2^WT and Shp2^Ex3−/− MEFs were plated on fibronectin (FN)-conjugated elastomeric polyacrylamide (PAA) gels with Young’s modulus E = 2.5, 9.0 and 21.5 kPa, as well as on glass for 4 h. We had checked that a steady state of cell spreading of both cells was attained after a 4-h adhesion in each condition. Analysis of immunofluorescence staining of paxillin (Fig. 1A) showed that the size distribution of FAs in Shp2^WT cells varied with the substrate rigidity (Fig. 1B). On the softest gel (2.5 kPa), most FA areas were smaller than 0.5 μm². An increase of rigidity induced the formation of larger FAs (>2 μm²), with a concomitant decrease in small FAs (Fig. 1B). However, as shown in Fig. 1C, Shp2^Ex3−/− cells exhibited numerous small FAs regardless of matrix rigidity (9.0 and 21.5 kPa and glass); no FAs were detectable in the cells seeded on the softest gel (2.5 kPa), as these cells were not spreading. These results suggest that the modulation of FA maturation and clustering in MEFs by matrix stiffness require Shp2 function.

Fig. 1.

Shp2 is required for matrix stiffness-dependent FA maturation in fibroblasts. (A) Shp2^WT and Shp2^Ex3−/− MEFs seeded on PAA gel with a Young’s modulus of 2.5, 9.0, or 21.5 kPa, or on glass coverslips, coated with fibronectin for 4 h, were stained with anti-paxillin antibody for FA determination. All FAs in the field were subjected to ImageJ analysis for FA quantification. (Scale bar, 10 μm.) (B and C) Numbers of FAs per cell in different sizes (areas; in micrometers squared) of Shp2^WT (B) and Shp2^Ex3−/− (C) MEFs seeded on gels with different rigidities or on glass coverslips. Inset shows the mean FA area in each condition. All data are expressed as mean ± SD from 20 representative cells. *P < 0.005; **P < 0.0005.

It has been shown that the alignment of FAs and/or orientation of stress fibers precede cell elongation in response to matrix rigidity (34–36). To calculate the orientation of stress fibers, we analyzed the images of FITC-phalloidin stained cells using a method similar to that described by Karlon et al. (37). In order to combine the data from individual cells on a common coordinate system, each cell image is rotated so that the median stress-fiber orientation coincides with the horizontal axis (0°). Stress-fiber orientation angles between −90° and 90° are then measured counterclockwise with respect to this horizontal axis. A map displaying the local orientation of the stress fibers with respect to the median stress-fiber orientation is shown in Fig. 2A. The polar histograms of stress-fiber orientations are shown in Fig. 2B. The spread of the distribution of stress-fiber orientations, as indicated by the standard deviation, provides a measure of the degree of stress-fiber polarization. Fig. 2B shows that the Shp2^WT cells seeded on the more rigid matrices had higher degrees of alignment of stress-fiber orientations. Shp2^Ex3−/− cells exhibited wider distributions of stress-fiber orientation than Shp2^WT cells and showed little change as the matrix rigidity was increased. These findings suggest that Shp2 plays an important role in the stiffness-dependent FA maturation and anisotropic orientation of stress fibers in MEFs.

Fig. 2.

Shp2 is involved in the organization of stress-fiber orientation in response to matrix rigidity. (A) Images of Shp2^WT (a–c) and Shp2^Ex3−/− (d–f) MEFs seeded on FN-coated glass. Cells were stained with FITC-phalloidin (green) for F-actin, and Hoechst (blue) for nuclei (a and d). After image segmentation, the orientation of each fiber was determined, and the image was rotated so that the median stress-fiber orientation is horizontal (b and e). Different orientations of actin filaments are depicted with different colors (c and f). (Scale bar in d, 10 μm.) (B) Histograms of stress-fiber orientations summarized from 15 Shp2^WT (blue) and 15 Shp2^Ex3−/− (red) cells for each level of matrix rigidity for 4 h as indicated.

Shp2 Deficiency Affects Cell Shape and Intracellular Tension.

The spreading of a cell involves shape changes that produce mechanical stresses inside the cell and in the matrix (35, 36). To further elucidate the role of Shp2 in the cellular response to matrix rigidity, we measured cell spreading shape and traction forces of Shp2^WT and Shp2^Ex3−/− cells seeded on substrates with different rigidities. Fig. 3A shows typical examples of traction stress vector plots (T_xy), maps of traction stress magnitude (T_mag), and DIC cell images. The arrowheads indicate the stress direction, and both the width and length of the arrow stem indicate the stress magnitude (Fig. 3A, Top row). The high NA (60×) used in our DIC microscopy includes the thin ruffles of the lamellipodia, which may not be seen with a lower NA.

Fig. 3.

Cellular traction force analysis of Shp2^WT and Shp2^Ex3−/− MEFs in response to matrix rigidity. Cells were seeded on the FN-coated PAA gel with rigidities of 2.5, 9.0, or 21.5 kPa overnight. Traction fields were computed from the displacement of images of fluorescence beads, before and following cell removal by trypsin treatment. (A) Showing a set of representative images of traction stress vector plots (T_xy), maps of traction stress magnitude (T_mag), and DIC cell images. Arrows in T_xy plots show the direction of the displacement field of the traction. Colors in T_xy and T_mag maps show the magnitude of the traction vectors in the Pascal unit. (Scale bar in DIC, 10 μm.) (B–G) Box and whisker diagrams showing (B) cell spreading area, (C) aspect ratio, (D) shape factor, (E) net contractile moment, (F) intracellular tensions parallel to the long axes (T_L), and (G) intracellular tensions parallel to the short axes (T_S) of the Shp2^WT (blue) and Shp2^Ex3−/− (red) MEFs. The sample numbers are 14, 12, and 20 for Shp2^WT and 15, 13, and 24 for Shp2^Ex3−/− cells on the 2.5, 9.0, and 21.5 kPa gels, respectively. *P < 0.005; **P < 0.0005.

With an increase in matrix rigidity, Shp2^WT cells spread out and became more elongated. Accordingly, the cell spreading area (Fig. 3B), aspect ratio (Fig. 3C), and shape factor (ratio of the perimeter of the cell to that of a circle with an equal area, Fig. 3D) increased with matrix rigidity in Shp2^WT cells. These parameters were significantly lower in Shp2^Ex3−/− cells. While the spreading area of Shp2^Ex3−/− cells also increased with matrix rigidity, the cell aspect ratio and shape factor remained very low and did not vary with the level of matrix rigidity (Fig. 3 C and D), consistent with the decrease of stress-fiber polarization in these Shp2^Ex3−/− cells.

Analysis of the traction stress measurements revealed that the net contractile moment, which is a coordinate-invariant scalar often used to infer cellular contractile strength (38), was significantly lower in Shp2^Ex3−/− cells than in Shp2^WT cells (Fig. 3E). The traction stress measurements were processed as described by Hur et al. (39) to determine maps of intracellular tension. Tension (T) is defined as force per unit length; thus, the total force supported by the cytoskeleton at a given intracellular section is equal to the product of the intracellular tension and the length of that section (Fig. S1 and Eq. S1). At each location, we determined the intracellular tensions parallel to the long and short axes of the cell, T_L and T_S, respectively. While T_L was significantly lower in Shp2^Ex3−/− cells than in Shp2^WT cells, T_S did not show statistically significant differences between the two types of cells (Fig. 3 F and G). This may suggest a minimum fiber length from which Sph2 function becomes activated and/or significantly important.

Whereas the loss of Shp2 function led to lower levels of cell spreading area (Fig. 3B), net contractile moments (Fig. 3E), and intracellular tension T_L (Fig. 3F) at all levels of rigidity, these parameters still showed a dependence on matrix rigidity in Shp2^Ex3−/− cells. Therefore, our data indicate that, although Shp2 plays an important role in the regulation of both cell shape and the directional organization of intracellular tension, the variations of these parameters with matrix rigidity still remain in Shp2^Ex3−/− cells, albeit with significant attenuations. These results indicate that Shp2 is not a major factor for rigidity sensing.

Expression of ROCKII Y722F Rescues FA Maturation in Shp2-Deficient Cells.

We have previously demonstrated that Shp2 is involved in RhoA-dependent ROCKII activation (32). To address whether Shp2 is recruited into FAs to decrease ROCKII Y722 phosphorylation, we isolated the FA fraction from Shp2^WT and Shp2^Ex3−/− cells using a method as described by Kuo et al (40). The cells were hypotonically shocked and then the cell bodies were removed by pulsed flow with PBS. The FA fraction was collected to examine the levels of Shp2 protein and ROCKII Y722 phosphorylation by Western blot analysis. We found that full-length Shp2 protein was clearly detected in the FA fraction isolated from Shp2^WT cells (Fig. S2A). In Shp2^Ex3−/− cells, the total protein level of exon3-deleted Shp2 was significantly lower than full-length Shp2 in Shp2^WT cells (Fig. S2A); this is consistent with other reports (29, 41). Importantly, almost no mutant Shp2 protein was detected in FA fraction isolated form Shp2^Ex3−/− cells. This finding suggests the requirement of the Shp2 domain encoded by exon3 for its FA targeting and may also explain the functional deficiency caused by the deletion of exon3 in Shp2. The FA-targeting of Shp2 was further confirmed by immunofluorescence staining of the FAs from cells after hypotonic treatment (Fig. S2B). The ROCKII protein was detected in the FA fraction of both types of cells. The level of ROCKII Y722 phosphorylation normalized by dividing it by the ROCKII protein level of the same sample was significantly higher in Shp2^Ex3−/− cells than Shp2^WT cells (Fig. S2A). Quantification of the ratio of fluorescence densities of pY722-ROCKII to paxillin in FAs at 4 h showed that the relative level of ROCKII Y722 phosphorylation decreased in Shp2^WT cells in response to the increase of matrix rigidity (Fig. S3). It is recognized that this ratio could vary with the stages of FA maturation. In contrast, the ROCKII Y722 phosphorylation of Shp2^Ex3−/− cells at all matrix rigidities (9.0 kPa, 21.5 kPa, and glass) was at a level higher than those seen in Shp2^WT cells and was insensitive to variations in matrix rigidity. Overall, these data indicate that Shp2 participates in FAs and plays a significant role in ROCKII Y722 dephosphorylation in cellular response to matrix rigidity.

To test whether ROCKII is the molecular target of Shp2 for FA maturation, we expressed the wild type (WT) and Y722F mutant of GFP-ROCKII in Shp2^WT and Shp2^Ex3−/− cells and determined their FA sizes by visualizing their coexpression with RFP-vinculin (Fig. 4A). Y722F GFP-ROCKII expression induced a decrease in the number of small FAs (<0.5 μm²) in both Shp2^WT and Shp2^Ex3−/− MEFs, compared with the corresponding cells expressing WT GFP-ROCKII. Importantly, the number of matured FAs (>2 μm²) in Shp2^Ex3−/− cells was markedly increased by the expression of Y722F GFP-ROCKII (Fig. 4B) to reach the level observed in Shp2^WT cells. The cell shape factor of Shp2^Ex3−/− cells was also partially restored by Y722F GFP-ROCKII (Fig. 4C). These results indicate that expression of Y722F ROCKII was able to rescue the defect of FA maturation in Shp2^Ex3−/− cells, thus further establishing the role of ROCKII (Y722) in FA maturation.

Fig. 4.

Expression of ROCKII Y722F rescues FA maturation in Shp2^Ex3−/− MEFs. (A) Shp2^WT and Shp2^Ex3−/− MEFs were cotransfected with pRFP-vinculin and pGFP-ROCKII WT or Y722F for 24 h, and then replated onto FN-coated coverslips for 2 h. (Scale bar, 10 μm.) (B) Number of FAs per cell in different sizes (area; in micrometers squared) under different conditions are shown. Inset shows the mean FA area in each condition (n = 10). (C) Cellular shape factors are shown (n = 20). *P < 0.005; **P < 0.0005.

Discussion

The rigidity of the ECM influences many cellular processes, including cell adhesion, migration, growth, differentiation, and gene expression (2, 42, 43). In this study, we used MEFs to address the role of Shp2 in their response to ECM rigidity. Comparison of the results on Shp2^WT and Shp2^Ex3−/− MEFs showed that Shp2 is required for the maturation of FAs and the anisotropic orientation of stress fibers in response to matrix rigidity. Concurrently, the loss of Shp2 function in the FAs leads to marked reductions in both the alignment of stress-fiber orientation and net contractile moment inside the cell, thus blunting the cell polarization in response to the increase of matrix rigidity. Furthermore, the responses of MEFs to ECM rigidity involve the participation of Shp2 in FAs, where it contributes to the reduction of ROCKII Y722 phosphorylation. The defect of FA maturation in Shp2^Ex3−/− cells was rescued by expressing unphosphorylable GFP-ROCKII Y722F, suggesting that ROCKII is the molecular target of Shp2 in FAs for the stiffness-dependent FA maturation in MEFs. These results indicate that Shp2 plays a key role in mechanotransduction in response to increased ECM rigidity by decreasing ROCKII Y722 phosphorylation in FAs to enhance FA maturation and facilitate the directional orientation of stress fibers and the spatial organization of intracellular forces.

ECM rigidity can be sensed by the force-sensing molecules that are mechanically linked to ECM-ligand integrins and actin filaments, thus resulting in cell deformation by changes in the mechanical forces transmitted between the ECM and the cytoskeleton (43, 44). These changes may allow the signaling molecules to bind the force-sensing molecules to elicit the biochemical signals involved in the cellular responses to the matrix rigidity (45). Shp2^Ex3−/− cells, as Shp2^WT cells, can increase their spreading area and intracellular tension (T_L) in response to an increase in Young’s modulus of the ECM, indicating that Shp2^Ex3−/− cells are still able to sense ECM rigidity, though in an attenuated manner. When cells sense the rigid matrix, the FAs grow and recruit proteins to enhance contraction force that leads to the rearrangement of FAs and orientation of stress fibers. These changes cause the spatial organization of intracellular forces and hence changes in cell shape and a reinforcement of cellular tension. Therefore, the significantly lower stiffness-dependent changes of FA size, stress-fiber polarization, cell shape, and direction of intracellular tension in Shp2^Ex3−/− cells indicate that the loss of Shp2 function in FAs leads to a defective response to ECM rigidity.

Rigidity response is a case of mechanotransduction in which the variations in ECM compliance are converted into intracellular biochemical signals such as protein activities. It is well known that the differential tyrosine phosphorylations of adhesion molecules and their associated proteins play key roles in regulating FA dynamics (44, 46–48). Prager-Khoutorsky et al. recently showed the roles of different protein tyrosine kinases (PTKs) in the ECM stiffness-dependent fibroblast polarization (34). They delineated the mechanosensitive cell polarization into several stages, beginning with the radial spreading of the cell attached to ECM and followed by the formation of dot-like nascent FAs in its periphery. Owing to the development of mechanical forces at the nascent FAs, these adhesions grow and rearrange their distribution and direction to establish the cell polarization axis. The FAs oriented parallel to the major axis of the cell keep growing and elongating, whereas the FAs oriented in other directions become disassembled. As a result, cells become fully polarized with matured FAs at their leading and trailing edges. Multiple PTKs were found to be involved in the regulation of these stages (34). In our study, we found that Shp2 is involved in the rigidity response by dephosphorylating ROCKII Y722 in FAs to facilitate FA maturation and cause the rearrangement of stress fibers and the directional organization of the intracellular forces. Interestingly, we found that Shp2^Ex3−/− cells have a phenotype similar to one group of PTK knockdowns (protein tyrosine kinases CSK1R, CSK, FER, ERBB4, and FLT3) (34). In particular, CSK serves as an inhibitor of Src kinase, which is responsible for ROCKII Y722 phosphorylation in FAs. CSK-knockdown cells exhibited immature FAs, low tension force, and decreased cell elongation on both rigid and compliant surfaces (34), similar to the Shp2^Ex3−/− cells.

Different studies suggest that Shp2 can either positively (49–51) or negatively (52, 53) regulate RhoA activity. Thus, Shp2 can decrease RhoA activity by inhibiting guanine nucleotide exchange factor Vav2 (53), whereas it can also increase RhoA activity by inhibiting Rho family GTPase-activating protein p190RhoGAP (49, 50). In addition, Shp2 can also promote Src activation directly by dephosphorylating the inhibitory phosphotyrosine on the Src C terminus and indirectly by decreasing Csk recruitment, that activates p190RhoGAP and results in RhoA suppression (54). Under our experimental condition, GST-Rho-binding domain (RBD) pull-down assays showed similar RhoA activities in Shp2^WT and Shp2^Ex3−/− MEFs seeded on soft and rigid gels (Fig. S4). However, the ROCKII activation, as indicated by S1366 phosphorylation (55), increased in response to increasing ECM rigidity in Shp2^WT cells, but not in Shp2^Ex3−/− cells (Fig. S5); this finding is consistent with the stiffness-dependent reduction of ROCKII Y722 phosphorylation in wild-type cells (Fig. S3). Furthermore, expression of catalytic-dead Shp2 (C/S) mutant in NIH3T3 stable line expressing wild-type ROCKII significantly reduced the size of FAs, whereas this manipulation had no effect on NIH3T3 stable line expressing Y722F ROCKII (Fig. S6). Overall, these data suggest that ROCKII is the molecular target of Shp2 in FAs for the stiffness-dependent FA maturation.

Von Wichert et al. demonstrated that Shp2 is required for the formation of force-dependent integrin-cytoskeleton linkages (56). Here, we show that the deficiency of Shp2 leads to decreases in net contractile moments and intracellular tension (T_L) in comparison to WT cells at the same level of ECM rigidity (Fig. 3 E and F). In particular, we demonstrated that the directional organization of the cytoskeleton and the intracellular tension were significantly reduced in the Shp2 mutants. Based on these results, we conclude that, in addition to contributing to FA maturation and the strengthening of mechanical forces at the FAs, Shp2 regulates the spatiotemporal dynamics of FAs and contributes to the directional rearrangement of stress fibers and force organization, thus playing an important role in determining the mechanical polarity in cellular response to matrix rigidity.

Materials and Methods

The experimental materials and methods are described in brief here; detailed information is provided in SI Materials and Methods. Shp2^WT and Shp2^Ex3−/− MEFs were kindly provided by Gen-Sheng Feng (Molecular Pathology Graduate Program, University of California, San Diego) (29). PAA gels were prepared based on the method developed by Wang and colleagues (57). The final concentration of acrylamide/Bis-acrylamide was 5%/0.1%, 10%/0.1%, and 10%/0.2% (vol/vol) for 2.5, 9.0, and 21.5 kPa gels, respectively. The substrate deformation field was obtained from the lateral displacements of fluorescent beads embedded in the gel. The lateral traction stresses exerted by the cells were determined from the measured substrate deformation after solving the equation of static equilibrium for an elastic substrate, as described previously (58). The net contractile moment (M) was calculated from the measured traction stresses as described in SI Materials and Methods (and refs. 34, 38). The intracellular tension along the major (T_L) and minor (T_S) axes were determined as described by Hur et al. (39).