Coupling of Ras and Rac Guanosine Triphosphatases Through the Ras Exchanger Sos

See allHide authors and affiliations

Science  23 Jan 1998:
Vol. 279, Issue 5350, pp. 560-563
DOI: 10.1126/science.279.5350.560


The Son of Sevenless (Sos) proteins control receptor-mediated activation of Ras by catalyzing the exchange of guanosine diphosphate for guanosine triphosphate on Ras. The NH2-terminal region of Sos contains a Dbl homology (DH) domain in tandem with a pleckstrin homology (PH) domain. In COS-1 cells, the DH domain of Sos stimulated guanine nucleotide exchange on Rac but not Cdc42 in vitro and in vivo. The tandem DH-PH domain of Sos (DH-PH-Sos) was defective in Rac activation but regained Rac stimulating activity when it was coexpressed with activated Ras. Ras-mediated activation of DH-PH-Sos did not require activation of mitogen-activated protein kinase but it was dependent on activation of phosphoinositide 3-kinase. These results reveal a potential mechanism for coupling of Ras and Rac signaling pathways.

Ras guanine nucleotide binding proteins regulate cell growth through the activation of signaling pathways that control gene expression and actin polymerization (1, 2). The effects of Ras on the actin cytoskeleton are mediated by Rac, another small guanine nucleotide binding protein (3), and Rac proteins function downstream of Ras in the pathways leading to cellular proliferation and oncogenic transformation (4-7). However, the mechanisms linking Ras activation to Rac activation are unknown.

In mammalian cells, growth factor–induced activation of Ras is mediated by the guanine nucleotide exchange factor Sos (8,9). Sos catalyzes the exchange of guanosine diphosphate (GDP) for guanosine triphosphate (GTP) on Ras through a central domain of about 400 amino acids that is highly conserved among guanine nucleotide exchange factors (GEFs) for Ras. The NH2-terminal domain of Sos is about 600 amino acids long and contains regions of homology to Dbl (DH) and pleckstrin (PH) domains. Because Dbl family proteins function as GEFs for specific members of the Rho family of guanosine triphosphatase (GTPases) (10, 11), we tested whether the DH domain of Sos (DH-Sos) can function as an activator of Rac.

Stimulation of Rac results in activation of the c-Jun NH2-terminal kinase (JNK) (12, 13). To test whether DH-Sos can activate JNK, we transfected COS-1 cells with expression plasmids encoding T7 epitope-tagged DH-Sos (14) and a FLAG epitope-tagged version of JNK1. JNK activity was assayed in an immunocomplex kinase assay with c-Jun coupled to glutathione S-transferase (GST–c-Jun) as a substrate (15). Expression of constitutively activated Rac in which amino acid 12 was substituted to Val (RacV12) induced a 25-fold activation of JNK (Fig.1A), and expression of DH-Sos led to a 15-fold stimulation of JNK activity. Activation of JNK induced by DH-Sos was inhibited (80%) by coexpression of dominant negative Rac in which residue 17 was changed to Asn (RacN17), indicating that the stimulation of JNK by DH-Sos depends on the activation of Rac. Expression of DH-Sos had no effect on the activity of another mitogen-activated protein kinase family member, ERK (16).

Figure 1

Activation of Rac by DH-Sos. (A and C) Stimulation of JNK activity by DH-Sos is dependent on Rac but not Cdc42. COS-1 cells were transfected with FLAG epitope-tagged JNK1 in combination with expression plasmids encoding T7 epitope tagged RacV12, DH-Sos, RacN17, or Myc epitope-tagged Cdc42N17. JNK activity was measured by immunocomplex kinase assay with GST–c-Jun as the substrate and visualized by autoradiography (15). Protein expression was determined by protein immunoblotting with polyclonal antibodies to JNK and monoclonal antibodies to T7 or Myc (35). (B andD) Stimulation of guanine nucleotide dissociation from Rac but not Cdc42 by cell lysates containing DH-Sos. Lysates prepared from human kidney 293 cells expressing DH-Sos (closed circle) or vector-only control (closed square) were incubated with [α-32P]GTP-bound GST-Rac (B) or GST-Cdc42 (D). At the indicated times, portions of the incubation mixture were removed and the amount of [α32P]GTP remaining bound to each protein was determined (17). Results are expressed as percentages of the values obtained at 0 min. Results of two independent experiments are shown.

To determine whether DH-Sos might activate Rac by functioning as a GEF, we tested the ability of DH-Sos to stimulate the dissociation of guanine nucleotide from Rac. Human kidney 293 cells were transfected with an expression plasmid encoding T7 epitope-tagged DH-Sos or control vector. Lysates prepared from the transfected cells were incubated with a GST-Rac1 fusion protein bound to [α-32P]GTP, and guanine nucleotide exchange activity was determined by measuring the amount of [α-32P]GTP that remained bound to GST-Rac1 (17). In the presence of lysates from cells expressing DH-Sos, the rate of release of guanine nucleotide from Rac was about twice as fast as that in the presence of lysates from control cells (Fig. 1B), making it probable that DH-Sos functions as a GEF for Rac. Attempts to express and purify recombinant DH-Sos were not successful, thus preventing analysis of GEF activity with purified proteins.

To investigate the specificity of DH-Sos GEF activity, we examined the effects of DH-Sos expression on the activity of Cdc42. Activation of JNK by DH-Sos was not affected by coexpression of the dominant negative Cdc42, Cdc42N17 (Fig. 1C). Moreover, lysates from cells expressing DH-Sos displayed no GEF activity toward GST-Cdc42 (Fig. 1D). Similarly, DH-Sos failed to stimulate guanine nucleotide exchange on RhoA (18). Together, these data suggest that DH-Sos has a preferential activity toward Rac.

Rac proteins regulate the organization of the actin cytoskeleton, and activation of Rac induces polymerization of cortical actin and formation of membrane ruffles in fibroblasts (3, 19). We microinjected COS-1 cells with expression plasmids encoding T7 epitope-tagged wild-type Rac (RacWT) and DH-Sos (20). When microinjected alone, neither DH-Sos nor RacWT induced membrane ruffling (Fig. 2). However, membrane ruffling was observed when cells were injected with both RacWT and DH-Sos. DH-Sos differs from the DH domain of Tiam1, an activator of Rac, which requires an NH2-terminal PH domain for induction of Rac-dependent membrane ruffling (21).

Figure 2

Membrane ruffling induced by DH-Sos. Serum-deprived COS-1 cells were microinjected with expression plasmids encoding T7 epitope-tagged RacWT or DH-Sos or both, as indicated. Three hours after injection cells were fixed and stained with rhodamine-phalloidin to visualize membrane ruffles (20).

Most Dbl family members contain a DH domain in tandem with a PH domain (10, 11). In some cases, these PH domains have been shown to regulate the targeting of DH domains to the appropriate subcellular location (21, 22). The PH domain of Sos (PH-Sos) is located immediately to the COOH-terminal side of DH-Sos, binds phosphoinositides, and functions in membrane localization and activation of Sos (23, 24). To investigate the functional relationship between DH-Sos and PH-Sos, we analyzed the activity of a hemagglutinin (HA) epitope-tagged Sos construct containing the DH and PH domains (DH-PH-Sos) (14). When expressed in COS-1 cells, DH-PH-Sos failed to stimulate JNK activation (Fig.3A). DH-PH-Sos did not induce membrane ruffling when microinjected either alone or with RacWT into quiescent COS-1 cells (16). Thus, the presence of the PH domain of Sos appears to inhibit the activity of the DH domain. The solution structure of the PH domain of Sos suggests that the PH and DH domains make specific structural contacts (25, 26), indicating that the activity of DH-Sos may be regulated by intramolecular interactions with PH-Sos.

Figure 3

Dependence on Ras for activation of JNK by DH-PH-Sos. (A) Failure of DH-PH-Sos to activate JNK. COS-1 cells were transfected with expression plasmids encoding FLAG epitope-tagged JNK1 and either T7 epitope-tagged RacV12 or DH-Sos or HA epitope-tagged DH-PH-Sos or PH-Sos. JNK activity was measured as described (Fig. 1A). (B) Requirement of RasV12 for DH-PH-Sos–induced JNK activation. COS-1 cells were transfected with FLAG epitope-tagged JNK1 in combination with expression plasmids encoding T7 epitope-tagged RacV12 or DH-Sos or HA epitope-tagged DH-PH-Sos or RasV12. (C) Effects of Ras effector binding loop mutants on JNK activation by DH-PH-Sos. COS-1 cells were cotransfected with FLAG epitope-tagged JNK1 and HA epitope-tagged DH-PH-Sos in combination with expression plasmids encoding HA epitope-tagged RasV12, RasV12C40, or RasV12S35. JNK activity was measured as described (Fig. 1A). Results are expressed as fold JNK activation relative to the activation measured in cells transfected with vector alone. Results of two independent experiments are shown.

Because Ras activation is linked to Rac activation (3, 27,28), we examined the role of Ras in DH-Sos–mediated activation of Rac. Coexpression of activated Ras, RasV12, with DH-Sos had no effect on the extent of JNK activation (29). However, DH-PH-Sos, which by itself failed to induce JNK activation (Fig. 3A), did activate JNK when it was coexpressed with RasV12 (Fig. 3B). Thus, Ras-dependent signals appear to enhance the activity of DH-Sos through a mechanism involving PH-Sos. We also tested the effects of Ras effector mutants that interact differentially with downstream effectors because of specific amino acid substitutions in the effector binding loop (30). The Ras mutant in which amino acid 35 was changed to Ser (RasV12S35) is able to bind to Raf-1 but not to RalGDS or the p110α subunit of phosphoinositide 3-kinase (PI 3-kinase), whereas the Ras mutant in which amino acid 40 was changed to Cys (RasV12C40) binds to p110α but not to Raf-1 and RalGDS (28, 31). When expressed with DH-PH-Sos, only the RasV12C40 mutant enhanced activation of JNK (Fig. 3C). Thus, PI 3-kinase might be the effector by which Ras induces activation of DH-PH-Sos. The PH domain of Sos binds to the lipid product of PI 3- kinase phosphatidylinositol-3,4,5-trisphosphate [PtdIns(3,4,5)P3] (32). Thus, activation of PI 3-kinase and the subsequent binding of PtdIns(3,4,5)P3 to the PH domain could provide a mechanism for the activation of DH-Sos. Consistent with this suggestion are the findings that the activity of the DH domain of Vav is directly controlled by substrates and products of PI 3-kinase (33). A similar mode of regulation has been demonstrated for the PH domain of protein kinase B (34).

Our results suggest that Sos activates Ras through the Ras GEF domain and Rac through the DH domain. Ras-mediated PI 3-kinase signaling may provide a coupling mechanism by which Ras activation can control Rac activation. In support of such a mechanism is the observation that a Ras-induced actin rearrangement, which is mediated by Rac, requires functional PI 3-kinase acting upstream of Rac (28). Thus, the multidomain structure of Sos may allow the coordinated activation of signaling pathways involved in growth control and cytoskeletal organization.

  • * To whom correspondence should be addressed: E-mail: barsagi{at}


Stay Connected to Science

Navigate This Article