Osmotic Activation of the HOG MAPK Pathway via Ste11p MAPKKK: Scaffold Role of Pbs2p MAPKK

See allHide authors and affiliations

Science  13 Jun 1997:
Vol. 276, Issue 5319, pp. 1702-1705
DOI: 10.1126/science.276.5319.1702


Exposure of the yeast Saccharomyces cerevisiae to high extracellular osmolarity induces the Sln1p-Ypd1p-Ssk1p two-component osmosensor to activate a mitogen-activated protein (MAP) kinase cascade composed of the Ssk2p and Ssk22p MAP kinase kinase kinases (MAPKKKs), the Pbs2p MAPKK, and the Hog1p MAPK. A second osmosensor, Sho1p, also activated Pbs2p and Hog1p, but did so through the Ste11p MAPKKK. Although Ste11p also participates in the mating pheromone–responsive MAPK cascade, there was no detectable cross talk between these two pathways. The MAPKK Pbs2p bound to the Sho1p osmosensor, the MAPKKK Ste11p, and the MAPK Hog1p. Thus, Pbs2p may serve as a scaffold protein.

MAP kinase cascades are common eukaryotic signaling modules that consist of a MAP kinase (MAPK), a MAPK kinase (MAPKK), and a MAPKK kinase (MAPKKK) (1). In S. cerevisiae, two independent osmosensors regulate the common HOG (high osmolarity glycerol response) signal transduction pathway, which includes the Pbs2p MAPKK and Hog1p MAPK (2-5). The Sln1p-Ypd1p-Ssk1p two-component osmosensor uses a multistep phosphorelay mechanism to regulate the redundant MAPKKKs Ssk2p and Ssk22p (2, 3, 6,7). Activated Ssk2p or Ssk22p then phosphorylates and activates the Pbs2p MAPKK. The second osmosensor, Sho1p, contains four transmembrane segments and a COOH-terminal cytoplasmic region with an SRC homology 3 (SH3) domain (3). The interaction between an NH2-terminal proline-rich motif in Pbs2p and the Sho1p SH3 domain is essential for the activation of Pbs2p by Sho1p (3). We now show that the activation of Pbs2p by Sho1p is mediated by the Ste11p MAPKKK, which is also an integral component of the mating pheromone response pathway (8-10). We propose a mechanism to explain how potential cross talk between the mating and the HOG signaling pathways is prevented.

Phosphorylation of Pbs2p appears to be required for its activation by the Sho1p osmosensor because Pbs2p containing mutations at the activating phosphorylation sites (Ser514 → Ala and Thr518 → Ala) is not activated by Sho1p (3). To test the possibility that a kinase other than Ssk2p and Ssk22p can phosphorylate and activate Pbs2p, we examined Pbs2p phosphorylation in vivo with the glutathione S-transferase (GST) fusion protein GST-PBS2(K-M) (11). PBS2(K-M) contains a Lys389 → Met mutation that inactivates its kinase activity, and thus it cannot undergo autophosphorylation. In ssk2Δ ssk22Δdouble-mutant cells, GST-PBS2(K-M) was highly phosphorylated after a brief osmotic shock (Fig.1A). This phosphorylation event was apparently dependent on the Sho1p osmosensor, because GST-PBS2(K-M) recovered from the ssk2Δ ssk22Δ sho1Δ triple mutant after osmotic shock was not phosphorylated (Fig. 1A). These results suggest that a protein kinase other than Ssk2p and Ssk22p can phosphorylate Pbs2p in cells exposed to osmotic shock.

Figure 1

Phosphorylation and activation of Pbs2p by the Ste11p MAPKKK. (A) Phosphorylation of Pbs2p in vivo. The indicated mutants expressing GST-PBS2(K-M) were grown in phosphate-depleted medium, incubated with [32P]orthophosphate, and subjected (+) or not (−) to a brief osmotic shock (0.4 M NaCl for 2 min) (11). GST-PBS2(K-M) was purified by association with glutathione-Sepharose beads and subjected to SDS-PAGE. Proteins were transferred to a nylon membrane and detected by autoradiography. The same filter was also probed with a monoclonal antibody to GST (anti-GST). Arrowheads indicate the position of GST-PBS2(K-M). (B) Osmosensitivity and sterility of OS-306 (ssk2Δ ssk22Δ ste11-306). OS-306 was transformed with centromeric plasmids containing the indicated genes. The transformants were spotted on YPD plates with or without 1.5 M sorbitol. Mating competency was assayed by the replica method (18). (C) High osmolarity–induced tyrosine phosphorylation of Hog1p in strain FP50 (MATa ura3 leu2 his3 ssk2::LEU2 ssk22::LEU2 ste11::HIS3) transformed with plasmids containing the indicated genes. Cells were collected before (−) or 5 min after (+) the addition of NaCl to a final concentration of 0.4 M. Tyrosine-phosphorylated Hog1p was detected by immunoblot analysis with monoclonal antibody 4G10 to phosphotyrosine. (D) High osmolarity–induced tyrosine phosphorylation of Hog1p in various mutant strains. Cells were treated as in (C) and tyrosine-phosphorylated Hog1p was detected by immunoblot analysis. (E) Osmosensitivity of various mutants. The indicated genes were disrupted (13) in the wild-type (WT) strain TM141 (MATa ura3 leu2 trp1 his3) and thessk2Δ ssk22Δ strain TM254 (MAT a ura3 leu2 his3 ssk2::LEU2 ssk22::LEU2), and the osmosensitivity of the resulting cells was tested as in (B).

To identify the protein kinase responsible for the Sho1p-mediated phosphorylation of Pbs2p, we screened for mutants on the basis of the assumption that mutational inactivation of the responsible kinase, in conjunction with ssk2Δ and ssk22Δmutations, would render the mutant cell incapable of activating Pbs2p and thus would confer osmosensitivity. One such synthetic mutant, OS-306, was sterile as well as osmosensitive. We isolated 20 genomic clones that complemented the osmosensitivity of OS-306, of which 10 contained SSK2 and another 10 all contained theSTE11 gene (12). The STE11 genomic clones complemented both the sterility and the osmosensitivity of OS-306 (Fig. 1B). To exclude the possibility that STE11 was merely acting as a multicopy suppressor of OS-306, we disrupted theSTE11 gene in an ssk2Δ ssk22Δ strain. Several independently isolated ssk2Δ ssk22Δ ste11Δ triple mutants were all sterile and osmosensitive (Fig. 1E). Disruption ofSTE11 alone had no effect on Pbs2p phosphorylation in vivo, but Pbs2p was not phosphorylated in response to osmotic shock in anssk2Δ ssk22Δ ste11Δ triple mutant (Fig. 1A). Thus, Ste11p contributes to the activation of Pbs2p.

Tyrosine phosphorylation of Hog1p is a sensitive measure of the activation state of Pbs2p (2, 5). However, Hog1p phosphorylation was not observed in response to osmotic shock in ssk2Δ ssk22Δ ste11Δ cells, indicating that this triple mutant has completely lost the capacity to activate Pbs2p (Fig. 1C). Transformation of the ssk2Δ ssk22Δ ste11Δ triple mutant with a plasmid containing eitherSSK2 +, SSK22 +, orSTE11 +, but not SHO1 +, restored tyrosine phosphorylation of Hog1p in response to osmotic shock, indicating the redundant roles of Ssk2p, Ssk22p, and Ste11p in Pbs2p activation. Consistent with the notion thatsho1Δ and ste11Δ mutations inactivate the same upstream signaling branch in the HOG pathway, neitherssk2Δ ssk22Δ sho1Δ nor ssk2Δ ssk22Δ ste11Δ triple-mutant cells showed tyrosine phosphorylation of Hog1p in response to osmotic shock (Fig.1D). In a sho1Δ ste11Δ double mutant, in whichSSK2 and SSK22 are functional, tyrosine phosphorylation of Hog1p in response to osmotic shock was detected (Fig. 1D).

Because Ste11p is a MAPKKK for the mating pheromone response pathway (9, 10), we tested whether other components in the mating pathway also participate in the HOG pathway. Thus, we disrupted either STE20, STE11, STE7, STE5, or the controlSHO1 gene in an ssk2Δ ssk22Δ background (13). Only the ste11Δ and sho1Δmutations were synthetically osmosensitive with ssk2Δ ssk22Δ (Fig. 1E). Furthermore, disruption of two other genes that encode protein kinases similar to Ste20p—CLA4 andYOL113—had no effect on osmosensitivity (14). Thus, STE11 may be the only gene shared between the mating and the HOG pathways.

Kinases in the MAPKKK family can be constitutively activated by eliminating their NH2-terminal noncatalytic domains (9). Indeed, expression of Ssk2p or Ste11p with NH2-terminal truncations (SSK2ΔN and STE11ΔN, respectively) resulted in Pbs2p-mediated tyrosine phosphorylation of Hog1p in the absence of osmotic stress (Fig. 2A). Thus, both SSK2ΔN and STE11ΔN can activate Pbs2p in the absence of upstream stimuli. Both GST-STE11ΔN and GST-SSK2ΔN proteins also efficiently phosphorylated the GST-PBS2(K-M) protein in vitro (Fig. 2B).

Figure 2

Activation of the HOG pathway by a constitutively active Ste11p (STE11ΔN) through Pbs2p phosphorylation. (A) Pbs2p-dependent tyrosine phosphorylation of Hog1p induced by the expression of SSK2ΔN or STE11ΔN (19). The plasmids pGal-SSK2ΔN, pGal-STE11ΔN, or pYES2 (vector) were introduced into wild-type (WT) strain TM141 or its pbs2Δ(pbs2::LEU2) derivative. Cells were grown in synthetic medium with raffinose, and the GAL1 promoter was induced with galactose (2). Samples were taken before (−) or 1 hour after (+) addition of galactose. Wild-type cells were treated (+) or not (−) with 0.4 M NaCl for 5 min. Tyrosine-phosphorylated Hog1p was detected by immunoblot analysis with antibody 4G10. (B) In vitro phosphorylation of Pbs2p by GST-SSK2ΔN or GST-STE11ΔN. Purified GST fusion proteins were incubated with GST-PBS2(K-M) in the presence of [γ-32P]ATP (adenosine triphosphate) and buffer (20). 32P-Labeled GST-PBS2(K-M) was detected by autoradiography after SDS-PAGE. (C) Activation of both the mating and HOG pathways by overexpression of STE11ΔN. Host cells were transformed with the plasmids pGal-SSK2ΔN or pGal-STE11ΔN. The transformants were spotted on selective medium containing either glucose or galactose. Growth was scored after 4 days at 30°C.

Continuous activation of the Hog1p MAPK by SSK2ΔN is lethal to yeast cells, and this lethality is suppressed by disruption of thePBS2 gene (3). Expression of STE11ΔN was also toxic to cells (Fig. 2C). The STE11ΔN lethality was suppressed partially by the pbs2Δ mutation and completely by theste7Δ pbs2Δ double mutation. In contrast,ste7Δ alone had little suppressive effect. Thus, the STE11ΔN lethality is caused by the hyperactivation of both the mating pheromone pathway and the HOG pathway.

We investigated whether the activation of Ste11p by osmotic stress results in the activation of mating responses and whether activation of Ste11p by mating factors results in activation of the HOG pathway. To assess activation of the mating pathway, we measured the expression of the FUS1 gene with a FUS1-lacZ promoter fusion construct (15). Activation of the HOG pathway was assessed by measuring tyrosine phosphorylation of Hog1p. These experiments were performed with ssk2Δ ssk22Δdouble-mutant cells, so that the activation of the Pbs2p MAPKK was dependent solely on the Ste11p MAPKKK. The α mating factor induced expression of FUS1-lacZ, but not tyrosine phosphorylation of Hog1p (Fig. 3). In contrast, osmotic shock induced tyrosine phosphorylation of Hog1p but not FUS1-lacZ expression. In ssk2Δ ssk22Δ ste11Δ triple-mutant cells, no response to either the mating factor or osmotic shock was detected. Thus, although the Ste11p MAPKKK participates in both the mating and HOG pathways, there is little or no cross talk between these pathways.

Figure 3

Lack of cross talk between the mating and HOG pathways. (A) Time course of Hog1p tyrosine phosphorylation. Yeast strains TM254 (MATassk2Δ ssk22Δ) or FP50 (MATa ssk2Δ ssk22Δ ste11Δ) were exposed to 0.4 M NaCl or 5 μM α factor for the indicated times, and tyrosine-phosphorylated Hog1p was detected by immunoblot analysis with antibody 4G10. (B) Expression of FUS1-lacZ. TM254 or FP50 strains were transformed with the pSB231 (FUS1-lacZ) reporter plasmid (15). Transformants were grown to exponential phase in YPD buffered at pH 3.5 (21) and exposed to 0.4 M NaCl or 5 μM α factor for the indicated times. β-Galactosidase activity was measured and expressed in Miller units (22). Data are means + SD of 12 assays (triplicate determinations with four independent transformants).

The scaffold protein Ste5p interacts with the Ste11p MAPKKK, Ste7p MAPKK, Fus3p-Kss1p MAPK (16), and G protein βγ subunits (17). Thus, the complex formed around Ste5p may allow the incoming signal from the mating factor receptor to flow only through this complex. The previous observation that the Sho1p osmosensor interacts with Pbs2p suggests that another signaling complex may be formed by the components of the HOG signaling pathway (3). Indeed, coprecipitation experiments revealed that Sho1p was associated with Pbs2p but not with Ste11p in intact cells (Fig.4A), and that Pbs2p interacts with both Ste11p and Hog1p (Fig. 4B).

Figure 4

Association of Pbs2p with Sho1p, Ste11p, and Hog1p. (A) Coprecipitation of hemagglutinin (HA)–tagged Pbs2p (PBS2HA) with GST-SHO1. The wild-type yeast strain TM141 was cotransformed with either the p426TEG (GST) or p426TEG-SHO1 (GST-SHO1) plasmids, and DNA encoding either HA-tagged Pbs2p (Gal-PBS2HA) or Ste11p (Gal-STE11HA) under the control of theGAL1 promoter in the YCpIF vector (23). Cells were grown in glucose (−) or galactose (+), the latter to induce expression of PBS2HA or STE11HA. Cells were lysed, and proteins were purified by association with glutathione-Sepharose beads (24) and subjected to immunoblot analysis with antibodies to GST (anti-GST) or to HA (anti-HA). (B) Coprecipitation of STE11HA and HOG1HA with GST-PBS2. TM141 was cotransformed with either p426TEG vector (GST) or p426TEG-PBS2 (GST-PBS2) (this plasmid contains the catalytically inactive Lys389 → Met mutation to prevent the toxicity of Pbs2p overexpression), and DNA encoding either HA-tagged Ste11p (Gal-STE11HA) or Hog1p (Gal-HOG1HA) under the control of theGAL1 promoter in the YCpIF vector. Cells and samples were processed as in (A).

Our results are consistent with the formation of a multiprotein complex that includes Sho1p, Ste11p, Pbs2p, and Hog1p, although it remains to be shown that these interactions occur simultaneously. Formation of such a multiprotein complex would restrict the osmotic stress–activated Ste11p MAPKKK to phosphorylating only the Pbs2p MAPKK, like the Ste5p complex ensures that the mating pheromone–activated Ste11p MAPKKK phosphorylates only the Ste7p MAPKK (Fig. 5). In this sense, both Ste5p and Pbs2p appear to serve a similar scaffold function, even though they are not structurally related. Given that several distinct MAP kinase cascades coexist in mammalian cells, formation of similar multiprotein complexes may be a general mechanism to prevent inappropriate cross talk.

Figure 5

Schematic model of the pheromone response pathway and the HOG signal transduction pathway. Pbs2p appears to serve as a scaffold protein that holds Sho1p, Ste11p, and Hog1p in a signaling complex, whereas Ste5p holds Ste11p, Ste7p, and Fus3p-Kss1p in a separate complex.

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


Stay Connected to Science

Navigate This Article