Report

Formation of Actin Stress Fibers and Focal Adhesions Enhanced by Rho-Kinase

See allHide authors and affiliations

Science  28 Feb 1997:
Vol. 275, Issue 5304, pp. 1308-1311
DOI: 10.1126/science.275.5304.1308

Abstract

The small guanosine triphosphatase (GTPase) Rho is implicated in the formation of stress fibers and focal adhesions in fibroblasts stimulated by extracellular signals such as lysophosphatidic acid (LPA). Rho-kinase is activated by Rho and may mediate some biological effects of Rho. Microinjection of the catalytic domain of Rho-kinase into serum-starved Swiss 3T3 cells induced the formation of stress fibers and focal adhesions, whereas microinjection of the inactive catalytic domain, the Rho-binding domain, or the pleckstrin-homology domain inhibited the LPA-induced formation of stress fibers and focal adhesions. Thus, Rho-kinase appears to mediate signals from Rho and to induce the formation of stress fibers and focal adhesions.

The small GTPase Rho is inactive in its guanosine diphosphate (GDP)-bound form and active in its guanosine triphosphate (GTP)-bound form (1). When cells are stimulated with certain extracellular signals such as LPA, GDP·Rho is believed to be converted to GTP·Rho, which binds to specific targets that mediate its biological functions. Rho participates in signaling pathways that lead to the formation of actin stress fibers and focal adhesions (2). Actin stress fibers are linked to integrins at the inner surface of the plasma membrane through focal adhesions (1, 2). Rho also participates in the regulation of cell morphology (3), cell aggregation (4), cell motility (5), cytokinesis (6), and smooth muscle contraction (7). In budding yeast, RHO1 (a homolog of RhoA) is implicated in the regulation of cell morphology and budding (8). Target proteins, the functions of which are modulated by Rho, include protein kinase N (PKN) (9), Rho-kinase (10) [also called Rho-binding kinase (11)], and the myosin-binding subunit (MBS) of myosin phosphatase (12). Rho-kinase phosphorylates MBS and consequently inactivates myosin phosphatase (12). Rho-kinase phosphorylates myosin light chain and thereby activates myosin adenosine triphosphatase (ATPase) (13). Other targets of Rho with unknown functions include rhophilin, p160 Rho-associated coiled-coil-containing protein kinase (ROCK), rhotekin, citron (9, 14), and phosphatidylinositol-4-phosphate 5-kinase (PIP5-K) (15). To elucidate the functions of Rho-kinase among these targets of Rho, we produced dominant active and negative forms of Rho-kinase.

Serine-threonine kinases such as protein kinase C (PKC) and Raf are usually composed of regulatory and catalytic domains (16). Deletion of the regulatory domains makes PKC and Raf constitutively active, and the regulatory fragments serve as dominant negative forms of the kinases (17, 18). Rho-kinase is composed of catalytic, coiled-coil, Rho-binding, and pleckstrin-homology domains. We produced four fragments containing these domains as glutathione-S-transferase (GST) fusion proteins: GST-CAT (catalytic domain, amino acids 6 to 553), GST-COIL (coiled-coil domain, amino acids 421 to 701), GST-RB (Rho-binding domain, amino acids 941 to 1075), and GST-PH (pleckstrin-homology domain, amino acids 1125 to 1388). Guanosine 5′-O-(3-thiotriphosphate) (GTP-γ-S)·GST-RhoA bound to GST-RB, but GTP-γ-S·GST-RhoAA37 bound to it very weakly (Fig. 1A). RhoAA37 is structurally equivalent to H-RasA35, which has a mutation in the effector-interacting domain (substitution of threonine to alanine) (1, 19). Rho-kinase had kinase activity on myosin light chain that was activated by GTP-γ-S·GST-RhoA (13), whereas GST-CAT showed full kinase activity without addition of GTP-γ-S·GST-RhoA (Fig. 1B). The molecular activities of Rho-kinase in the presence of GTP-γ-S·GST-RhoA and of GST-CAT in the absence of GTP-γ-S·GST-RhoA were 0.32 ± 0.02 s−1 and 0.71 ± 0.02 s−1, respectively, indicating that GST-CAT is constitutively active. GST-CAT mutated at the ATP-binding site (GST-CAT-KD) did not have kinase activity. Although GST-RB inhibited Rho-kinase activity stimulated by GTP-γ-S·GST-RhoA in a dose-dependent manner, it did not affect the kinase activity of GST-CAT (Fig. 1C). The kinase activity of Rho-kinase was not affected by GST-CAT-KD, GST-COIL, or GST-PH (20). The kinase activity of PKN and myosin light chain kinase was not inhibited by GST-CAT-KD, GST-COIL, GST-RB, or GST-PH (20). Rho-induced formation of stress fibers is inhibited by protein kinase inhibitors such as staurosporine (21); staurosporine inhibited the kinase activity of both Rho-kinase and GST-CAT (Fig. 1C).

Fig. 1.

Dominant active and negative forms of Rho-kinase. (A) Nitrocellulose membranes containing the purified native Rho-kinase (lanes 1 and 3) and GST-RB (lanes 2 and 4) separated by SDS-PAGE were probed with [35S]GTP-γ-S·GST-RhoA (lanes 1 and 2) or [35S]GTP-γ-S·GST-RhoAA37 (lanes 3 and 4) (23, 24). The results are representative of three independent experiments. (B) Myosin light chain was phosphorylated by native Rho-kinase or GST-CAT in the presence or absence of GTP-γ-S·GST-RhoA (1.5 μM) (25). Data are means ± SEM of triplicate determinations. (C) Myosin light chain was phosphorylated by native Rho-kinase in the presence of GTP-γ-S·GST-RhoA or by GST-CAT with GST-RB or staurosporine. Data are means ± SEM of triplicate determinations.

Confluent, serum-starved Swiss 3T3 cells had very few stress fibers, which were visualized by phalloidin (Fig. 2A) as described (2). When the cells were stimulated with LPA, new stress fibers appeared and increased in number and diameter (Fig. 2A) (2). Microinjection of GST-CAT also induced stress fiber formation, whereas GST-CAT-KD was inactive in this capacity (20). Injected GST-CAT often caused the formation of a large aggregate of actin filaments connected with stress fibers at the central area. Although it is not clear why hub-like actin filaments were present, they may have resulted from the high contractility of stress fibers induced by injected GST-CAT. When the cells were microinjected with C3 transferase (C3), which causes adenosine diphosphate ribosylation and inhibition of Rho (22), the cells rounded up within 30 min (3, 20). Injected C3 abolished LPA-induced stress fiber formation (Fig. 2A), whereas it did not inhibit GST-CAT-induced stress fiber formation. Coinjection of GST-CAT with C3 prevented the cells from rounding up. Cells stimulated by LPA in the presence of staurosporine showed randomly arranged actin filaments (Fig. 2A) (21), but cells injected with GST-CAT in the presence of staurosporine did not form stress fibers.

Fig. 2.

Actin reorganization and focal adhesion formation induced by Rho-kinase. (A) Actin reorganization caused by Rho-kinase. Actin filaments in confluent, serum-starved Swiss 3T3 cells stimulated by vehicle (a), stimulated by LPA (200 ng/ml) for 15 min without injection (b), microinjected with C3 (80 μg/ml) and stimulated by LPA (c), stimulated by LPA 15 min after treatment with 100 nM staurosporine (d), microinjected with GST-CAT (0.5 mg/ml) alone (e), microinjected with GST-CAT and C3 (f), and microinjected with GST-CAT 15 min after treatment with staurosporine (g) (26). (B) Focal adhesion formation induced by Rho-kinase. Vinculin localization is shown. Images (a) through (g) show the same treatments as in (A), except that actin filaments and vinculin localization are shown in Swiss 3T3 cells microinjected with GST-CAT (h). The arrowheads show the injected cells. Scale bars, 20 μm.

Very few focal adhesions, as visualized by an antibody to vinculin, were observed in confluent, serum-starved Swiss 3T3 cells (Fig. 2B) (2). When the cells were stimulated with LPA, new focal adhesions appeared and increased in number (2). Microinjection of GST-CAT induced focal adhesion formation. Longitudinal stress fibers newly synthesized after injection of GST-CAT were linked to focal adhesions that had an elongated arrowhead shape, as revealed by dual immunofluorescence analysis (Fig. 2B). Thus, it seems clear that GST-CAT-induced focal adhesions exhibit the specific features of adhesions elicited by Rho (21). Microinjection of C3 abolished the LPA-induced formation of focal adhesions, whereas it did not abolish that induced by GST-CAT. Staurosporine inhibited both LPA-induced and GST-CAT-induced formation of focal adhesions. Injection of constitutively active PKN or MBS did not induce formation of stress fibers and focal adhesions, nor did it affect those induced by GST-CAT (20).

Injection of GST-RB or GST-PH inhibited LPA-induced formation of stress fibers and focal adhesions (Fig. 3). About 30% of the cells injected with GST-CAT-KD did not form stress fibers or focal adhesions in the presence of LPA; GST-COIL had no effect. The GST-CAT-induced formation of stress fibers and focal adhesions was not inhibited by injection of GST-CAT-KD, GST-COIL, GST-RB, or GST-PH, which indicated that GST-CAT-KD, GST-RB, and GST-PH inhibited the functions of endogenous Rho-kinase but did not inhibit the functions of the exogenously overexpressed GST-CAT.

Fig. 3.

Effect of various forms of Rho-kinase on LPA-induced actin filament reorganization and focal adhesion formation. Confluent, serum-starved Swiss 3T3 cells were microinjected with GST-CAT-KD (2 mg/ml) (A and E), GST-COIL (5 mg/ml) (B and F), GST-RB (5 mg/ml) (C and G), or GST-PH (5 mg/ml) (D and H), and then stimulated with LPA (200 ng/ml) (26). Actin filaments and vinculin localization are shown. The arrowheads show the injected cells. Scale bar, 20 μm.

Because Swiss 3T3 cells are not suitable for nuclear injection of plasmids, we examined the morphological effects of Rho-kinase in Madin-Darby canine kidney (MDCK) cells microinjected with cDNAs encoding various domains of Rho-kinase. Microinjection of the cDNA encoding RhoV14 into MDCK cells resulted in increased formation of stress fibers (Fig. 4A) and focal adhesions (20, 21). RhoV14 is structurally equivalent to H-RasV12 (substitution of glycine to valine) (1, 19). Stress fibers and focal adhesions formed in cells injected with the cDNA encoding CAT (Fig. 4B) (20). The cDNAs encoding CAT-KD, the constitutively active form of PKN, or MBS had no effect. Coinjection of the cDNAs encoding CAT-KD, RB, or PH inhibited RhoV14-induced stress fiber formation (Fig. 4, C to E) and focal adhesion formation. CAT-KD was less effective than RB or PH. COIL was inactive in this capacity.

Fig. 4.

Actin reorganization in MDCK cells. Actin filaments are shown in MDCK cells microinjected with pEF-BOS-HA-RhoV14 (0.1 mg/ml) + pEF-BOS-myc (1 mg/ml) (A), pEF-BOS-myc-CAT (0.1 mg/ml) + pEF-BOS-myc (1 mg/ml) (B), pEF-BOS-HA-RhoV14 + pEF-BOS-myc-CAT-KD (1 mg/ml) (C), pEF-BOS-HA-RhoV14 + pEF-BOS-myc-RB (1 mg/ml) (D), or pEF-BOS-HA-RhoV14 + pEF-BOS-myc-PH (1 mg/ml) (E) (27). The arrowheads show the injected cells. Scale bar, 20 μm.

We assume that CAT behaves as a dominant active form, and that CAT-KD, RB, and PH behave as dominant negative forms for Rho-kinase but not for CAT. CAT-KD at high concentrations may inhibit interaction of Rho-kinase with substrates such as myosin, as described for other kinases (18). Because the PH domain is supposed to localize molecules at the specified regions, PH may inhibit the proper localization of Rho-kinase in cells. CAT existed in the cytoplasm in MDCK cells, though Rho-kinase is partly localized at cell-to-cell junctions (20). RB may tie up activated Rho and inhibit interaction of Rho with targets such as PKN in addition to Rho-kinase. Consistently, the inhibitory effect of RB could be reversed by overexpression of RhoAV14 in MDCK cells, but the inhibitory effect of CAT-KD and PH could not be rescued by overexpression of RhoAV14 (20). CAT-KD and PH may serve as more specific inhibitors for Rho-kinase. Taken together, these findings strongly suggest that Rho-kinase regulates the formation of stress fibers and focal adhesions in Swiss 3T3 and MDCK cells in response to activation of Rho. Because the organization of the CAT-induced stress fibers is somewhat different from that induced by LPA in Swiss 3T3 cells, additional signals from Rho or the regulatory domain of Rho-kinase may be necessary for the formation of highly organized stress fibers.

Because injection of CAT slightly increased the intensity of phalloidin staining, Rho-kinase appears to induce actin polymerization to a small extent. The cells stimulated by LPA in the presence of staurosporine showed randomly arranged actin filaments, but the cells injected with CAT in the presence of staurosporine did not form stress fibers, indicating that there are additional pathways (such as PIP5-K) that induce actin polymerization downstream of Rho.

REFERENCES AND NOTES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
View Abstract

Navigate This Article