Report

Ste5 RING-H2 Domain: Role in Ste4-Promoted Oligomerization for Yeast Pheromone Signaling

See allHide authors and affiliations

Science  03 Oct 1997:
Vol. 278, Issue 5335, pp. 103-106
DOI: 10.1126/science.278.5335.103

Abstract

Ste5 is a scaffold for the mitogen-activated protein kinase (MAPK) cascade components in a yeast pheromone response pathway. Ste5 also associates with Ste4, the β subunit of a heterotrimeric guanine nucleotide–binding protein, potentially linking receptor activation to stimulation of the MAPK cascade. A RING-H2 motif at the Ste5 amino terminus is apparently essential for function because Ste5(C177S) and Ste5(C177A C180A) mutants did not rescue the mating defect of aste5Δ cell. In vitro Ste5(C177A C180A) bound each component of the MAPK cascade, but not Ste4. Unlike wild-type Ste5, the mutant did not appear to oligomerize; however, when fused to a heterologous dimerization domain (glutathione S-transferase), the chimeric protein restored mating in an ste5Δ cell and anste4Δ ste5Δ double mutant. Thus, the RING-H2 domain mediates Ste4-Ste5 interaction, which is a prerequisite for Ste5-Ste5 self-association and signaling.

Mating in the yeastSaccharomyces cerevisiae is initiated by pheromone binding to heterotrimeric guanine nucleotide–binding protein (G protein)–coupled receptors on cells of opposite mating type, leading to dissociation of the Gβγ complex (Ste4• Ste18) from the inhibitory Gα (Gpa1) subunit (1). Subsequent signal propagation activates an evolutionarily conserved MAPK cascade (2), ultimately causing arrest of the cell cycle in the G1 phase and the production and activation of factors required for cell and nuclear fusion. Ste5 is an essential component of this pathway (3, 4) and is thought to function as a scaffold for the MAPK cascade components (5). Ste5 also associates with Ste4 (6) and thus may link release of Gβγ to activation of the MAPK cascade.

The NH2-terminus of Ste5 (residues 177 to 229) contains a cysteine-rich region that is the prototype for the RING-H2 motif (Fig. 1A). The RING-H2 motif is a variant of the larger class of RING domains (7) but contains a second histidine in place of the cysteine normally found at position 5 (Fig. 1B). Proteins possessing RING and RING-H2 domains participate in diverse cellular processes, but no specific function has yet been ascribed to these domains (7). The crystal structures of two RING domains have been solved (8), and each is a globular pseudosymmetric fold that coordinates two Zn2+ atoms through a cross-bridging element. To disrupt this structure as a means to determine its importance to the function of Ste5, we mutated [to serine (S) or alanine (A), as indicated] either the first, or both the first and second, conserved cysteine (C), yielding Cys177→ Ser177 (C177S) and C177A C180A mutant proteins. The expressed amount and stability of both mutant proteins were comparable to those of wild-type Ste5 (9). Neither theste5(C177S) allele nor the ste5(C177A C180A)allele was able to complement the mating defect of anste5Δ strain, when the mutant genes were expressed from a centromere-based plasmid, driven by either the STE5 promoter or the inducible GAL1 promoter (Table1). Even high-level expression of the mutant proteins from the GAL1 promoter on a multicopy plasmid caused only a small increase in the mating proficiency of theste5Δ cells (Table 1). Thus, an intact RING-H2 domain is essential for Ste5 function. Consistent with this observation,ste5Δ cells containing the ste5(C177A C180A)allele were unable to activate transcription from a pheromone-inducible reporter gene (FUS1-lacZ) (10) in response to the mating pheromone α-factor (9). Likewise,ste5Δ cells expressing the Ste5(C177A C180A) mutant were unable to respond to α-factor (9), as judged by the halo bioassay for pheromone-induced growth arrest (11).

Figure 1

Sequence alignment and derived consensus for the RING-H2 motif. (A) Shown are RING-H2 domains of the following proteins (with the indicated GenBank accession numbers): Ste5 (L23856); Deltex (U09789); Far1 (M60071); Rapsyn (Z33905); Neurodap1 (D32249); Pep3 (M65244); and Pep5 (X54466). Conserved residues are indicated in bold and represent presumptive Zn2+-binding ligands. (B) Consensus sequence for the RING-H2 motif. Positions conserved in all members are given in bold. X represents any amino acid, and the number of such residues is also indicated.

Table 1

Quantitative mating assays. Strain BYB69 (MAT a ste5Δ) (21) was transformed with appropriate vector controls (YCplac33, YCpUGal, or YEp352Gal) or with the derived plasmids indicated (22) expressing either wild-type STE5 + or each of two different RING-H2 domain alleles: the single mutantste5(C177S) or the double mutant ste5(C177A C180A). The mating ability of the resulting transformants was tested, using minor modifications (27) of a quantitative procedure (11), with strain DC17 as theMATα partner. Values given represent the averages (and standard deviations of those means) for at least three independent trials, each performed in triplicate. ND, not determined.

View this table:

When Ste5(C177A C180A) was overexpressed from the GAL1promoter on a multicopy plasmid in a STE5 +strain, the cells formed a halo in response to α-factor, but the halo filled in much more rapidly than did lawns of control cells lacking the plasmid (9), indicating an attenuated G1 arrest response in these cells. Such a dominant-negative effect might result if the mutant protein competed with wild-type Ste5 for the binding of one or more of the factors required for signaling and sequestered them in an inactive complex. To determine directly whether the Ste5(C177A C180A) mutant protein remained competent to interact with any of the known MAPK cascade components, we immunoprecipitated extracts from cells expressing derivatives of Ste5 and Ste5(C177A C180A), tagged at their NH2-terminal end with a c-Myc epitope, with an appropriate monoclonal antibody (9E10) (12). The immune complexes were washed, and the bound proteins were resolved by SDS–polyacrylamide gel electrophoresis (PAGE) and transferred to a membrane filter for immunoblotting. The filters were probed with polyclonal antisera to the MAPK Fus3, the MAPK kinase Ste7, and the MAPK kinase kinase Ste11. Ste5(C177A C180A) associated with Ste11, Ste7, and Fus3 with an affinity virtually indistinguishable from that of wild-type Ste5 (Fig. 2, A through C), indicating that disruption of the RING-H2 domain does not prevent binding of the MAPK cascade components and, therefore, does not compromise the overall structure of the Ste5 protein.

Figure 2

Binding of Ste5(C177A C180A) to Ste11, Ste7, and Fus3, but not Ste4. A protease-deficient strain BYB84 (MAT a ste5Δ) (21) carrying a vector alone (–) or the same plasmid expressing either Myc-tagged Ste5 or Myc-tagged Ste5(C177A C180A) (22) was cotransformed with either another empty vector or the same vector expressing either Ste11 (A), Ste7 (B), Fus3 (C), or Ste4 (D). The cells were grown, induced for protein expression, lysed, and subjected to immunoprecipitation with an anti-Myc monoclonal antibody 9E10 (12). The resulting immune complexes were separated by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and analyzed by immunoblotting (24) with rabbit polyclonal antibodies specific for Ste11 (A, top), Ste7 (B, top), Fus3 (C, top), Ste4 (D, top) or Ste5 [(A) through (D), bottom]. Despite overexpression, all of the untagged proteins displayed only weak nonspecific adsorption to the 9E10 antibody. Results shown are representative of at least three different experiments, each performed with independent transformants. Molecular size markers are in kD.

Because Ste5 can associate with Ste4 (Gβ subunit), and this interaction requires the NH2-terminal region of Ste5 (6), we examined the effect of the RING-H2 domain point mutations on the ability of Ste5 to interact with Ste4. Ste4 was not immunoprecipitated with the Ste5(C177A C180A) mutant, whereas it was with wild-type Ste5 (Fig. 2D). Thus, when the RING-H2 domain is disrupted, interaction of Ste5 with Ste4 is severely diminished. We confirmed that the RING-H2 domain is essential for Ste4 association with Ste5 in the two-hybrid system (6, 13). Both theste5(C177S) and ste5(C177A C180A) alleles (as well as Ste5 mutants with deletions of the sequences encompassing the RING-H2 domain) failed to interact with Ste4, but still interacted with all three of the MAPK components, whereas normal Ste5 interacted with all four partners (9).

Ste5 can also self-associate (14, 15). To test whether the Ste5(C177A C180A) mutant could oligomerize, we performed interallelic complementation tests. We used ste5point mutations that result in a specific defect in the association of Ste5 with one, and only one, component of the MAPK cascade (15). Neither Ste5(I504T), which does not interact with Ste11, nor Ste5(V763A S861P), which fails to interact with Ste7, when expressed alone, complemented a ste5Δ cell, whereas normal Ste5 expressed from the same vector did (Fig.3). However, expression of both mutant proteins in the same cell restored efficient mating (Fig. 3), suggesting that Ste5 action in vivo requires oligomer formation (14, 15). In contrast, the ste5(C177A C180A)allele was unable to rescue the mating debility of ste5Δcells when coexpressed with either the ste5(V763A S861P)allele (Fig. 3) or the ste5(F514L) allele (9), despite the fact that the RING-H2 domain mutant is only defective in its ability to associate with Ste4. These results suggest that, in addition to its role in mediating Ste5 association with Ste4, the RING-H2 domain is also required for Ste5 oligomerization.

Figure 3

Genetic analysis of Ste5-Ste5 interaction by interallelic complementation. Strain BYB69 (MAT a ste5Δ) (21) was transformed with either vector alone (YCp111) or the same vector expressing the ste5(V763A S861P) mutant allele, which is defective for interaction with Ste7 (15). These two derivatives were then cotransformed with either a vector control (YCp33) or with the same vector expressing either wild-type Ste5 or each of two different ste5 mutants: the RING-H2 domain mutant, ste5(C177A C180A); or, theste5(I504T) allele, the latter of which is defective for interaction with Ste11 (15). The resulting plasmid-bearing strains were patched onto selective plates lacking uracil and tryptophan (upper panel), mated to an appropriateMATα tester strain (DC17), and subsequently replica-plated onto a medium selective for diploids (lower panel) (25). Patches from two independent transformants containing each combination of plasmids are shown.

We tested whether fusion of the Ste5(C177A C180A) protein to a heterologous dimerization domain would allow the mutant protein to function. For this purpose, we used Schistosoma japonicumglutathione-S-transferase (GST), which forms a stable dimer, both in solution and in protein crystals (16), and which can functionally substitute for the dimerization domain of a heterologous protein (17). Indeed, a chimeric form of Ste5(C177A C180A), in which GST was fused to the COOH-terminus (at residue 913), did enable ste5Δ cells to mate, and did so as well as a fusion of GST to wild-type Ste5 (Fig. 4A), within the limits of resolution of this assay (15). Also, expression of either Ste5-GST or Ste5(C177A C180A)-GST inste5Δ cells induced FUS1-lacZ transcription (9). Thus, bringing Ste5(C177A C180A) molecules into proximity with each other by GST-mediated dimerization was sufficient to restore function to this mutant.

Figure 4

Dimerization of the RING-H2 domain mutant restored Ste5 function and bypassed the requirement for Ste4. (A) Strain BYB69 (MAT a ste5Δ) (21) transformed with a plasmid expressing GST alone, or either a Ste5-GST or a Ste5(C177A C180A)-GST chimera (26), were patched onto selective plates (left), mated to a MATα tester strain (DC17), and subsequently replica-plated onto a medium selective for diploids (right). Patches from two independent transformants containing each plasmid are shown (25). (B) Strain BYB88 (MAT a ste4Δ ste5Δ) transformed with a plasmid expressing Ste4 alone, Ste5 alone, the Ste5-GST fusion, the Ste5(C177A C180A)-GST fusion, or both Ste4 and Ste5 together, were patched onto selective plates lacking the appropriate amino acid (–AA; left), mated to the MATα tester strain (DC17), and then replica-plated onto a medium selective for diploids (right). Patches from two independent transformants containing each plasmid, or combination of plasmids, are shown (25).

In immunoprecipitation experiments, the Ste5(C177A C180A)-GST chimera did not associate with Ste4 (9), indicating that GST-mediated dimerization of Ste5 did not reconstitute a binding site for Ste4. Because the chimera is functional, this observation implies that the normal order of events in pheromone-induced activation of Ste5 is association with Ste4, followed by oligomerization. If so, expression of the Ste5(C177A C180A)-GST chimera should bypass the need for Ste4, as well as the need for Ste5. Indeed, as predicted on the basis of our findings, a ste4Δ ste5Δ double mutant expressing the Ste5(C177A C180A)-GST fusion was able to mate (Fig. 4B), although not quite as well as cells expressing both wild-typeSTE5 and STE4. Thus, even though Ste5(C177A C180A) cannot interact with Ste4 and cannot self-associate, once dimerized by fusion to GST, the resulting chimera is competent for signaling.

Expression of the wild-type Ste5-GST fusion, which supported mating ofste5Δ cells (Fig. 4A), did not rescue the mating defect of the ste4Δ ste5Δ double mutant (Fig. 4B). We conclude from this finding that the intact RING-H2 domain of wild-type Ste5 may have an inhibitory function that can only be alleviated upon its interaction with Ste4. The C177A C180A mutations, by perturbing the RING-H2 domain, apparently eliminate this negative function, such that artificial dimerization can promote signaling even in a cell that lacks Ste4. Presumably, therefore, in normal cells, interaction of Ste4 with Ste5 relieves the inhibitory function of the RING-H2 domain as well as promotes a conformational change that permits dimerization of Ste5, in that order.

Disruption of the RING-H2 domain in Ste5 by mutation of conserved cysteines abolished the function of Ste5 in mating. Likewise, naturally occurring point mutations in the RING element of the human breast cancer susceptibility–determining protein, BRCA1 (18), ablate its tumor-suppressor function. Our results indicate that association with the components of the MAPK cascade is not sufficient for Ste5 action. The RING-H2 domain is not only required for Ste4 binding to Ste5, but also (directly or indirectly) for Ste5 oligomerization. Thus, the RING-H2 domain of Ste5 serves as a molecular link between G protein activation and stimulation of the MAPK cascade. These observations raise the possibility that RING-H2 and RING domains in other proteins may mediate their multimerization or their association with (and regulation by) Gβγ subunits, or both. Finally, our evidence indicates that the RING-H2 domain in wild-type Ste5 also acts as a negative regulatory element. This inhibitory role may involve the proline-rich regions (residues 1 through 162 and 260 through 337) that immediately flank the RING-H2 domain on either side. In this regard, previous results (4) have shown that a mutation (T52M) in the region NH2-terminal to the RING-H2 motif confers a partial, constitutively hyperactive phenotype, suggesting that this alteration perturbs the inhibitory function of the RING-H2 domain.

  • * These authors contributed equally to this paper.

  • Present address: Laboratory of Molecular Embryology, National Institute of Child Health and Human Development, National Institutes of Health, Building 18T, Room 106, Bethesda, MD 20892–5431, USA.

  • To whom correspondence should be addressed. E-mail: jthorner{at}mendel.berkeley.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article