Wnt/Fz/Dvl signaling activates Rac and Rho in parallel pathways. (A) Rac activation by specific Fz proteins. Fz1 (rat) and Fz7 (Xenopus) activated Rac, but Fz2 (rat) or Fz5 (human) did not (the slight Rac activation seen in the Fz5 lane was not reproducible). (B) Rac activation by Wnt-1 CM was rapid (5 min) and lasting (3 h), and resistant to cycloheximide treatment (which inhibited cyclin D1 induction by Wnt-1 CM, data not shown). Cdc42 was not activated by Wnt-1 CM. (C) Rac activation by Dvl1 and Dvl2, but not Dvl3; by Xdsh and Dsh; by Dvl2 and Xdsh mutants lacking the DIX domain (ΔN-Dvl2 and ΔN-Xdsh); and by the DEP domain but not the PDZ domain alone. The Dsh¹ mutant had weaker activity than Dsh in Rac activation. (D) Rac-N17, but neither Rho-N19 nor Cdc42-N17, inhibited Rac activation by Wnt-1 CM, Fz1, or Dvl2. Rac-N17, Rho-N19, and Cdc42-N17 each inhibited Ephexin activation of Rac and Cdc42 (and Rho; Habas et al. 2001).

A Wnt-induced Dvl–Rac complex that is independent of the Wnt-induced Dvl–Rho complex. (A) Wnt-1 CM induced formation of a complex between Dvl2 and Rac or Rho, but not Cdc42. T-Daam1 or N-Daam1 disrupted the Dvl2–Rho complex but not the Dvl2–Rac complex. Endogenous Rac, Rho, or Cdc42 was immunoprecipitated from cells treated with/without Wnt-1 CM for 3 h, and coprecipitated endogenous Dvl2 was detected by immunoblotting. (B) Rac activation by Fz1 or Dvl2 was not inhibited by T-Daam1 or N-Daam1. (C) Wnt-1 CM induced a complex formation between endogenous Rac and Dvl1, but not Dvl3. (D) The Dvl2 DEP domain, but not the PDZ domain, formed a complex with Rac with or without Wnt-1 stimulation.

Rac function is required for Xenopus gastrulation. (A) XRac1 MO specifically reduced the protein level of endogenous Rac, but not Rho or Cdc42. A control MO had no effect on the level of any of these proteins. Fifty nanograms of the XRac1 MO or control MO was injected into 2-cell embryos, and whole embryo extracts were harvested at stage 10.5. (B) XRac1 MO, but not the control MO, and Rac-N17, Rho-N19, and Cdc42-N17 each inhibited gastrulation. The figure shows affected embryos in each case where a phenotype was seen; for frequency of effects, see Table 1. Fifty nanograms of either XRac1 MO or the control MO; 1 ng of RNA for Rac-N17, Rho-N19, or Cdc42-N17; or 2 ng of RNA for the wild-type Rac, Rho, or Cdc42 were each injected in the dorsal marginal region of the 4-cell-stage embryo, and resulting phenotypes were examined at stage 35. Injection of RNA (2 ng) for human Rac1, but not Rho or Cdc42, rescued the effect of the XRac1 MO, and had a weak ameliorating influence on the effect of Rac-N17. (C,D) Rho-N19 and Rac-N17 inhibited Rho and Rac activation, respectively. RNA (2 ng) or MO (50 ng) was injected at the 4-cell stage. (C) Dorsal injection, DMZ explants. Endogenous Rac activation was inhibited by RNA for Rac-N17 or Cdc42-N17 but not Rho-N19 or XDaam1 MO, whereas endogenous Rho activation was inhibited by Rho-N19 but not Rac-N17 or Cdc42-N17, and by XDaam1 MO but not the control MO. (D) Ventral injection, VMZ explants. Rac induction by Dvl2 RNA was inhibited by coinjected RNA for Rac-N17 or Cdc42-N17 but not by Rho-N19 or XDaam1 MO, whereas Rho activation by Dvl2 was suppressed by Rho-N19 but not Rac-N17 or Cdc42-N17, and by XDaam1 MO but not the control MO.

Parallel Rac and Rho pathways downstream of Wnt/Fz/Dvl signaling (see Discussion).

Wnt/Fz/Dvl activation of JNK via Rac but not Rho. (A) JNK activation by specific Fz proteins. Fz1 and Fz7 activated JNK, but Fz2 or Fz5 did not or did weakly. (B) Wnt-1 CM, but not control CM, activated JNK rapidly in the presence of cycloheximide. (C) JNK activation by Dvl1 and Dvl2, but not Dvl3; by Xdsh and Dsh; by Dvl2 and Xdsh mutants lacking the DIX domain (ΔN-Dvl2 and ΔN-Xdsh); and by the DEP domain but not the PDZ domain alone. The Dsh¹ mutant had weaker activity than Dsh in JNK activation. (D–F) JNK activation by Wnt-1 CM (D), or by transfected Fz1 (E) or Dvl2 (F) was blocked by Rac-N17 but not Rho-N19 or Cdc42-N17.

Dorsal Rac activation by Wnt-11/Fz signaling during Xenopus gastrulation. (A) XRac1 RNA is expressed maternally and throughout embryogenesis as monitored by RT–PCR. EF-1α was used as a loading control. −RT, without reverse transcriptase. (B) Endogenous Rac activation was detected in DMZ but not in VMZ isolated from stage 10.5 embryos. Cdc42 activation was not detected in DMZ or VMZ. (C) Endogenous Rac activation was inhibited in DMZ by injected RNA for Dn-Xwnt-11 (1 ng) or extra-xFz7 (1 ng). RNA for Xdd1 (1 ng) or N-Daam1 (2 ng) did not inhibit dorsal Rac activation. β-galactosidase RNA (β-gal, 2 ng) injection was used as control. (D) Rac was ectopically activated in VMZ by RNA for Xwnt-11 (200 pg), xFz7 (200 pg), Xdsh (1 ng), or Xdd1 (1 ng), but not by C-Daam1 DNA (200 pg) or by β-gal RNA (2 ng). C-Daam1 DNA was injected instead of RNA, which interfered with early embryo cleavage (Habas et al. 2001).

Rac function is required for convergent extension movements and is independent of the Wnt/β-catenin pathway. (A) In animal explants treated with activin, XRac1 MO but not the control MO (50 ng) inhibited elongation. RNA for human Rac1 (2 ng), but not Rho or Cdc42, rescued the XRac1 MO effect. RNA (1 ng) for Rac-N17, Rho-N19, and Cdc42-N17 also inhibited convergent extension. The Rac-N17 effect was partially rescued by coinjected RNA (2 ng) for wild-type Rac, but not Rho or Cdc42. (B,C) In animal explants, neither XRac1 MO nor dominant negative GTPase mutants affected mesodermal gene expression induced by activin (B), or organizer markers induced by Xwnt8 RNA (20 pg; C), as assayed by RT–PCR at stage 10.5. EF-1α was used as an internal control. WE, whole embryo; −RT, no reverse transcriptase.