Interaction with the SH3 domain protein Bem1 regulates signaling by the Saccharomyces cerevisiae p21-activated kinase Ste20

Abstract

The Saccharomyces cerevisiae PAK (p21-activated kinase) family kinase Ste20 functions in several signal transduction pathways, including pheromone response, filamentous growth, and hyperosmotic resistance. The GTPase Cdc42 localizes and activates Ste20 by binding to an autoinhibitory motif within Ste20 called the CRIB domain. Another factor that functions with Ste20 and Cdc42 is the protein Bem1. Bem1 has two SH3 domains, but target ligands for these domains have not been described. Here we identify an evolutionarily conserved binding site for Bem1 between the CRIB and kinase domains of Ste20. Mutation of tandem proline-rich (PxxP) motifs in this region disrupts Bem1 binding, suggesting that it serves as a ligand for a Bem1 SH3 domain. These PxxP motif mutations affect signaling additively with CRIB domain mutations, indicating that Bem1 and Cdc42 make separable contributions to Ste20 function, which cooperate to promote optimal signaling. This PxxP region also binds another SH3 domain protein, Nbp2, but analysis of bem1Delta versus nbp2Delta strains shows that the signaling defects of PxxP mutants result from impaired binding to Bem1 rather than from impaired binding to Nbp2. Finally, the PxxP mutations also reduce signaling by constitutively active Ste20, suggesting that postactivation functions of PAKs can be promoted by SH3 domain proteins, possibly by colocalizing PAKs with their substrates. The overall results also illustrate how the final signaling function of a protein can be governed by combinatorial addition of multiple, independent protein-protein interaction modules.

FIG. 1.

A proline-rich domain between the CRIB and kinase domains of Ste20 contains a Bem1-binding site. (A) Fragments used to map a Bem1-binding motif within Ste20 residues 434 to 499 by two-hybrid analysis. Asterisks indicate PxxP motifs; GA represents mutations in two of these motifs, as shown in panel B. DNA binding domain fusions to Ste20 fragments (from top to bottom, pB20N2, pB20N, pPP1062, pB20BA, pB20BB, pB20BC, and pPP1180) were coexpressed in PPY760 with AD fusions to Bem1^157-551 (pRL51.1), Cdc42^G12V/C188S (pPP1027), or the vector (pGAD424). Interactions were scored as positive (++) or negative (−), in comparison to those for vector controls, by a filter β-galactosidase assay. Bem1 interaction results were similar when full-length Bem1^1-551 (pRL5.2) rather than Bem1^157-551 was used, though in all cases the signal was stronger with the latter. (B) Sequence showing tandem PxxP motifs (boxed) in theminimal Bem1-binding region and mutations of Pro475 and Pro477 that constitute the PP-GA allele. WT, wild type. (C) Quantitative measurements of key two-hybrid interactions from panel A, using the same strain and plasmids (pRL51.1 for Bem1). Measurements are expressed as mean β-galactosidase units (n = 3 or 4; all standard deviations were within 40% of the mean). DBD, DNA binding domain. (D) Coprecipitation of GST-Ste20 and Bem1-myc is disrupted by the PP-GA mutation. Strains expressing myc₁₂-tagged Bem1 (+) (DLY4000) or untagged Bem1 (−) (DLY1) were transformed with a vector (pRD56) or constructs expressing the indicated Ste20 mutants (WT, PP-GA, or ΔN) as galactose-inducible GST fusions. Following galactose induction, glutathione precipitates were analyzed by immunoblotting with anti-myc (top left) or anti-GST (right) antibodies. (Bottom left) Densitometric quantification of Bem1-myc signals from four separate experiments (mean ± standard deviation), with the mean density in the negative control (GST-Ste20^WT in DLY1) assigned a value of zero and that in the positive control (GST-Ste20^WT in DLY4000) assigned a value of 1.

FIG. 2.

Phenotypic consequences of point mutations (PP-GA) in the Bem1-binding site. In all panels, the vector is pRS316, and native STE20 and GFP-STE20 plasmids are as described in Table 2. (A) Mating pathway phenotypes of strains with the Ste20 PP-GA mutation, alone and in combination with mutations in the CRIB domain. Arrest (left) and mating (center) results are shown for both native and GFP-fused versions of each Ste20 mutant. Induction of the transcriptional reporter FUS1-lacZ by α-factor (5 μM; 2 h) is shown for the GFP fusions (right) as means ± standard deviations (n = 4). All assays used strain PPY913. (B) The effect of the PP-GA mutation on mating pathway signaling is independent of pheromone and Gβγ. Ste20-dependent signaling was activated in the absence of pheromone or Gβ (Ste4) by galactose-induced synthesis of Ste5-CTM (pH-GS5-CTM) or Ste5^P44L-GST (pH-SL2); the host strain, PPY866 (ste4Δ ste5Δ ste20Δ), also harbored the indicated GFP-STE20 constructs. After a 3-h galactose treatment, FUS1-lacZ induction was measured (mean ± standard deviation; n = 9). To facilitate comparison between the two Ste5 reagents, results were expressed relative to those for wild-type (wt) Ste20, which yielded 134 (Ste5-CTM) and 88.3 (Ste5^P44L-GST) mean raw units. (C) Hyperosmotic cross talk signaling shows that the effect of the PP-GA mutation is also independent of Ste5. Strain PPY1691 (hog1Δ ste20Δ ste5Δ) harboring the indicated GFP-STE20 constructs was treated with 1 M sorbitol for 3 h, after which FUS1-lacZ induction was measured (mean ± standard deviation; n = 3). Mean expression in cells not induced with sorbitol was <1 U (data not shown). Similar results were observed for hog1Δ ste20Δ and hog1Δ ste4Δ ste5Δ ste20Δ strains (data not shown).

FIG. 3.

Effects of Ste20 mutations in Bem1- versus Cdc42-binding domains on viability, polarization, and localization. (A) Ste20^PP-GA shows a slight defect in performing the Cla4-redundant, essential function of Ste20. (Top) Strain KBY211 (ste20Δ cla4-75^ts) was transformed with the indicated GFP-STE20 plasmids, and then fivefold serial dilutions were spotted onto −Ura plates and incubated for 3 days at 25 or 37°C. (Bottom) KBY211 cultures harboring the indicated GFP-STE20 plasmids were grown in −Ura liquid medium to logarithmic phase at 23°C and then were diluted back and shifted to 37°C. Culture densities were measured over time by optical density at 660 nm (OD₆₆₀) as indicated. Results are means ± standard deviations from three trials; in each trial the same rank order of growth rate was observed. (B) Morphology of cells in panel A after 4 h at 37°C (similar results were observed after 3, 8, or 9 h). Note that the PP-GA mutant generally allows for normal bud neck constriction, compared to that of vector controls, but bud growth is less elongated than with wild-type Ste20, possibly implying a defect in apical polarization. Cells harboring CRIB domain mutants were heterogeneous but were on average more elongated than the PP-GA mutant and showed more examples of wide bud necks (especially the S338A mutant [data not shown]). (C) The Bem1-binding domain can contribute to localization of GFP-Ste20. Localization of the indicated GFP-Ste20 derivatives was examined in cells (PPY1249) overexpressing Bem1 from a high-copy-number plasmid (pPP1202). This method allowed occasional detection of the Ste20 Δ334-369 derivative at polarized growth sites (though rare in comparison to wild-type Ste20), whereas this phenomenon was never observed with the combined Ste20 Δ334-369 PP-GA derivative. The PP-GA mutation alone causes a mild decrease in the intensity of enrichment at bud tips rather than affecting the number of cells showing such enrichment.

FIG. 4.

The Bem1-binding domain of Ste20 also binds another SH3 domain protein, Nbp2. (A) Two-hybrid analysis showing interaction of Nbp2 with the isolated Bem1-binding fragment (Ste20 434-499) but not with a larger Bem1-interacting fragment (Ste20 1-499). DNA binding domain (DBD) fusions (plasmids pB20N2 and pB20BC) were coexpressed in PPY760 with vector (pGAD424) or AD fusions to Bem1^157-551 (pRL51.1) or Nbp2 (pPP1355). Interactions were quantified in parallel by mean β-galactosidase units (n = 3; standard deviations were within 40% of the mean for all means greater than 1). (B) GST coprecipitation assays showing that Bem1 and Nbp2 have similar requirements for binding to full-length Ste20. Strains expressing Bem1-myc₁₂ (DLY4000) (left) or Nbp2-myc₁₃ (PPY1356) (right) were transformed with galactose-inducible constructs expressing GST alone(vector; pRD56) or the indicated GST-Ste20 fusion (pRDSTE20ATG [wild type {WT}], pPP1254 [PP-GA], pRDSTE20RI [ΔN], pPP1327 [L369G], or pRD56-K649M). Following galactose induction, glutathione precipitates were analyzed by immunoblotting with anti-myc (top) or anti-GST (bottom) antibodies. The Bem1-myc and Nbp2-myc experiments for which results are shown here were conducted simultaneously; similar results were observed in multiple independent trials. The relevant portions of the anti-myc blots from the Bem1 and Nbp2 experiments were juxtaposed solely to facilitate comparison and do not imply that Bem1-myc₁₂ (82 kDa) and Nbp2-myc₁₃ (48 kDa) have similar electrophoretic mobilities. Note that for both Bem1 and Nbp2, coprecipitation with GST-Ste20 is disrupted by the PP-GA and ΔN mutations but not by the L369G or K649M mutation.

FIG. 5.

The phenotypic effect of Ste20 PxxP mutation depends on Bem1 rather than on Nbp2. The level of FUS1-lacZ induced by α-factor was assayed for GFP-Ste20 fusions that contained combinations of mutations in two domains—the indicated CRIB domain variant (wild type [WT], S338A, or Δ334-369) combined with a Bem1-binding domain that was either intact (PP) or mutant (PP-GA). Control transformants harbored the pRS316 vector (vect.). Deletion mutant strains were compared to congenic wild-type cells in the same strain background. Because both bem1Δ and nbp2Δ mutations impair growth with severities that depend on the strain background, the experiments for which results are shown here were performed in strain backgrounds where the mutants are the least growth impaired (15Dau for BEM1 versus bem1Δ; W303 for NBP2 versus nbp2Δ), but similar results were obtained for nbp2Δ mutants in the strain background (15Dau; PPY1343 versus PPY1415) where the deletion caused a stronger growth defect (data not shown). (A) Deletion of BEM1 mimics the effect of the PP-GA mutation, and in bem1Δ cells the PP-GA mutation confers no further pheromone response defect, indicating that the effect of the PP-GA mutation can be attributed to a disruption of Bem1-Ste20 interaction. Strains: PPY1343 (BEM1) and PPY1341 (bem1Δ). (B) In nbp2Δ cells, the PP-GA mutation still confers a signaling defect. Strains: PPY913 (NBP2) and PPY1456 (nbp2Δ).

FIG. 6.

Effects of Ste20 PxxP mutations on filamentous growth and hyperosmotic resistance. (A) Agar invasion. Strain PPY1209 (Σ1278b ste20Δ) was transformed with the indicated GFP-Ste20 fusion constructs and assayed for invasive growth as described in Materials and Methods. Representative examples of cells remaining adherent after gentle rinsing with water, as well as before rinsing, are shown. Note that invasive growth conferred by the Δ334-369 allele is eliminated by the PP-GA mutation. (B) Filamentous colony morphology. The transformants used for panel A were grown on low-glucose medium for 24 h. Cells with wild-type Ste20 become elongated and form branched microcolonies under these conditions. Note that the disruption of these features becomes gradually more severe as mutations in the Cdc42-binding and Bem1-binding motifs are combined. (C) Hyperosmotic resistance. Strains PPY1287 (ssk1Δ ste20Δ) and PPY1646 (ssk2Δ ssk22Δ ste20Δ) were transformed with the indicated GFP-Ste20 fusion constructs. Fivefold serial dilutions were spotted onto YPD or YPD plus 1.2 M NaCl and incubated at 30°C for 2 to 7 days (d.) as indicated. Note that only the Δ334-369 PP-GA double mutant has a strong defect in osmoresistance and that even this mutant retains some function, as revealed at longer incubation times.

FIG. 7.

Evolutionary conservation of the Bem1 binding site in Ste20 orthologs and paralogs. Standard single-letter symbols for amino acid residues are used. In addition, the symbols +, @, b, and x stand for basic, aliphatic, hydrophobic, and any residues, respectively. (A) Schematic diagram indicating relative positions of defined domains in Ste20. (B) Alignment of sequences from Ste20 orthologs from fungi of increasing phylogenetic distance. Sequences are from S. cerevisiae (S.cer), Saccharomyces kudriavzevii (S.kud), Saccharomyces bayanus (S.bay), Saccharomyces castelli (S.cas), Candida albicans (C.alb), Ashbya gossypii (A.gos), and Neurospora crassa (N.cra). Each sequence shown lies between the CRIB and kinase domains. Identities to the Ste20 sequence are boxed black. (C) Bem1 binding is disrupted by mutation of conserved residues in the Ste20 motif. The indicated point mutations (from K469A to S481A) were incorporated into a DBD-Ste20^1-499 fusion construct (pB20N2), and interaction with Bem1^157-551 (pRL51.1) or Cdc42 (pPP1027) was tested in the two-hybrid tester strain PPY760 by a quantitative β-galactosidase assay. For ease of comparison, signals were normalized to those of wild-type (wt) pB20N2 (taken as 100%), whose mean values were 15.0 (Bem1) and 271.5 (Cdc42) β-galactosidase units; standard deviations (n = 3)were within 15% of the mean in all cases and are omitted for clarity. Note that the mutations do not disrupt interaction with Cdc42 and thus are specific for Bem1. Also, similar Bem1-binding results were observed by using a full-length Bem1 (residues 1 to 551) construct (pXP-BEM1), and no appreciable interaction signals were observed by using a vector (pGAD424) control (data not shown). (D) Alignment of Ste20 paralogs and orthologs from S. cerevisiae (S.c.) and Schizosaccharomyces pombe (S.p.). A variable number of PxxP motifs (from 1 to 12) are present in each protein. Shown here are the only motifs that are perfect matches to class I (+x@Px@P) or class II (@Px@Px+) consensus sequences (45). Exceptions are as follows: Pak2 has no perfect matches, so its only PxxP motif is shown for comparison; Ste20 has one class I match and one class II match, but only the class I match is shown because it is the one identified in this study as a Bem1-binding site. Pak1 has two class I matches (#1 and #2), each of which is shown. Within the core motif, identities to Ste20 are boxed black, and conservative or common substitutions are boxed grey. The top six sequences are from PAK family kinases and thus represent candidate trans ligands for the SH3-2 domain of Bem1 or its ortholog (with the exception of the Pak2 sequence, which is a poor match and which lies within the CRIB domain). The bottom two sequences lie within the PX domains of Bem1 and its S. pombe ortholog, Scd2, and have been proposed to bind in cis to SH3-2 domains within the same protein (21). These sequences match the class I consensus but lack the additional conserved sequence features common to the PAK motifs.