Portal de Periódicos da CAPES

Phosphorylation of Hsl1 by Hog1 leads to a G2 arrest essential for cell survival at high osmolarity

2006; Springer Nature; Volume: 25; Issue: 11 Linguagem: Inglês

10.1038/sj.emboj.7601095

1460-2075

Josep Clotet, Xavier Escoté, Miquel Àngel Adrover, Gilad Yaakov, Eloi Garí, Martí Aldea, Eulàlia de Nadal, Francesc Posas,

Fungal and yeast genetics research

Article11 May 2006free access Phosphorylation of Hsl1 by Hog1 leads to a G2 arrest essential for cell survival at high osmolarity Josep Clotet Josep Clotet Department of Molecular and Cellular Biology, Universitat Internacional de Catalunya, Sant Cugat del Vallès, Spain Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Xavier Escoté Xavier Escoté Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Miquel Àngel Adrover Miquel Àngel Adrover Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Gilad Yaakov Gilad Yaakov Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Eloi Garí Eloi Garí Departament de Ciències Mèdiques Bàsiques, Universitat de Lleida, Lleida, Spain Search for more papers by this author Martí Aldea Martí Aldea Departament de Ciències Mèdiques Bàsiques, Universitat de Lleida, Lleida, Spain Search for more papers by this author Eulàlia de Nadal Eulàlia de Nadal Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Francesc Posas Corresponding Author Francesc Posas Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Josep Clotet Josep Clotet Department of Molecular and Cellular Biology, Universitat Internacional de Catalunya, Sant Cugat del Vallès, Spain Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Xavier Escoté Xavier Escoté Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Miquel Àngel Adrover Miquel Àngel Adrover Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Gilad Yaakov Gilad Yaakov Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Eloi Garí Eloi Garí Departament de Ciències Mèdiques Bàsiques, Universitat de Lleida, Lleida, Spain Search for more papers by this author Martí Aldea Martí Aldea Departament de Ciències Mèdiques Bàsiques, Universitat de Lleida, Lleida, Spain Search for more papers by this author Eulàlia de Nadal Eulàlia de Nadal Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Francesc Posas Corresponding Author Francesc Posas Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain Search for more papers by this author Author Information Josep Clotet1,2,‡, Xavier Escoté2,‡, Miquel Àngel Adrover2,‡, Gilad Yaakov2, Eloi Garí3, Martí Aldea3, Eulàlia de Nadal2 and Francesc Posas 2 1Department of Molecular and Cellular Biology, Universitat Internacional de Catalunya, Sant Cugat del Vallès, Spain 2Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Barcelona, Spain 3Departament de Ciències Mèdiques Bàsiques, Universitat de Lleida, Lleida, Spain ‡These authors have contributed equally to this work *Corresponding author. Cell Signaling Unit, Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra (UPF), Dr. Aiguader, 80, 08003 Barcelona, Spain. Tel.: +34 93 542 2848; Fax: +34 93 542 2802; E-mail: francesc.p[email protected] The EMBO Journal (2006)25:2338-2346https://doi.org/10.1038/sj.emboj.7601095 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Control of cell cycle progression by stress-activated protein kinases (SAPKs) is essential for cell adaptation to extracellular stimuli. Exposure of yeast to osmostress leads to activation of the Hog1 SAPK, which controls cell cycle at G1 by the targeting of Sic1. Here, we show that survival to osmostress also requires regulation of G2 progression. Activated Hog1 interacts and directly phosphorylates a residue within the Hsl7-docking site of the Hsl1 checkpoint kinase, which results in delocalization of Hsl7 from the septin ring and leads to Swe1 accumulation. Upon Hog1 activation, cells containing a nonphosphorylatable Hsl1 by Hog1 are unable to promote Hsl7 delocalization, fail to arrest at G2 and become sensitive to osmostress. Together, we present a novel mechanism that regulates the Hsl1–Hsl7 complex to integrate stress signals to mediate cell cycle arrest and, demonstrate that a single MAPK coordinately modulates different cell cycle checkpoints to improve cell survival upon stress. Introduction Activation of stress-activated protein kinases (SAPKs) is essential for proper cell adaptation to extracellular stimuli (Kyriakis and Avruch, 2001). In budding yeast (Saccharomyces cerevisiae), the presence of high osmolarity in the extracellular environment results in the activation of the stress-activated Hog1 kinase. Activation of Hog1 is essential for cell survival in response to high osmolarity, because the MAPK elicits an extensive programme required for cell adaptation that includes regulation of gene expression, translation and cell cycle progression (de Nadal et al, 2002; Hohmann, 2002). The HOG pathway comprises a central core of kinases composed of the MAPK Hog1, the MAPKK Pbs2 and a set of three MAPKKKs (Ssk2/Ssk22 and Ste11) activated by two independent upstream mechanisms (de Nadal et al, 2002). The best-understood mechanism of activation involves a ‘two-component’ osmosensor that involves the Sln1 histidine kinase (Posas et al, 1998), which activates the Ssk2 and Ssk22 MAPKKKs. The Sln1 osmosensor utilizes a phospho-relay mechanism inactivated under high osmolarity conditions, thus Sln1 is a negative regulator of the HOG pathway. Mutations that inactivate the Sln1 osmosensor (i.e., sln1ts4) are detrimental for growth and this can be prevented by disruption of the HOG1 gene (Maeda et al, 1994; Posas et al, 1996). The HOG pathway can also be activated by expression of either constitutively active MAP3K mutants, or a constitutively active MAPKK mutant (Pbs2DD)(Maeda et al, 1995; Posas and Saito, 1997). Recent reports have shown that exposure of yeast cells to high osmolarity or mutations that lead to activation of the Hog1 SAPK (i.e. sln1ts4, Pbs2DD) result in arrest of cell cycle progression (Alexander et al, 2001; Belli et al, 2001; Yaakov et al, 2003; Escote et al, 2004). It has been reported that activated Hog1 leads to G1 arrest by a dual mechanism that involves downregulation of cyclin expression and direct phosphorylation of the CDK-inhibitor protein Sic1. This combination results in Sic1 stabilization and inhibition of cell cycle progression to prevent premature entry into S phase without proper cell adaptation (Escote et al, 2004; Zapater et al, 2005). Previous reports suggested that osmostress also induces a delay at G2 (Alexander et al, 2001; de Nadal et al, 2002). However, the molecular mechanism mediating this effect is poorly understood (Alexander et al, 2001). Entry into mitosis is controlled by the activity of the cyclin B–Cdk1 (Clb2–Cdc28) complex, which is held in check by the protein kinase Wee1, Swe1 in S. cerevisiae. Swe1 is thought to delay entry into mitosis until critical cell size has been reached (Rupes, 2002; Kellogg, 2003; Harvey et al, 2005) or defects in bud formation or cytoskeletal function are monitored by the ‘morphogenesis checkpoint’ (Cid et al, 2002; Lew, 2003). Several requirements are critical for regulating Swe1 stability, its phosphorylation by the Clb2–Cdc28 (Asano et al, 2005; Harvey et al, 2005) and also the activity of the Hsl1 checkpoint kinase together with Cdc5 (Hanrahan and Snyder, 2003; Asano et al, 2005). When bound to the septins (Versele and Thorner, 2005), Hsl1 tethers Hsl7 at the bud neck, which is in turn required for recruitment of Swe1 to the bud neck to facilitate spatially controlled Swe1 phosphorylation, prior to ubiquitin-mediated degradation (Lew, 2003). Thus, timely phosphorylation and subsequent degradation of Swe1 are critical for proper activation of the Clb2–Cdc28 complex. Here, we show that Hog1 prevents G2 progression by direct phosphorylation of the Hsl1 kinase protein, which leads to Swe1 accumulation and decrease on Clb-associated Cdc28 activity. Failure to arrest at G2 upon osmostress is deleterious for cell survival. Interestingly, we propose a mechanism to regulate G2 progression that involves components of the morphogenesis checkpoint, such as Hsl1, but with novel regulatory properties to integrate stress signals. Taken together, it is shown that a single MAP kinase is able to modulate different cell cycle checkpoints, at G1 and G2, coordinately to prevent cell cycle progression in the presence of stressful conditions to allow cells to survive upon osmostress. Results Activation of the Hog1 MAPK results in modulation of cell cycle progression in both G1 and G2 Recently, we showed that Hog1 was involved in the control of G1 progression. Briefly, we analysed the effect of Hog1 activation on cell cycle progression in cells allowed to proceed into S phase after release from α-factor arrest. Direct release of α-factor-arrested cells carrying a temperature-sensitive allele of SLN1 (sln1ts4) at nonpermissive temperature resulted in accumulation of a large number of cells in G1 (unbudded cells with 1C DNA content) (Figure 1A) (Escote et al, 2004). This accumulation was not observed in wild-type cells and was prevented by deletion of HOG1 and SIC1 (Escote et al, 2004). Similar results were obtained when cells were treated with osmostress (Figure 1C) (Escote et al, 2004). Figure 1.Activation of the HOG pathway results in G1 and G2 cell cycle arrest. (A) Cells arrest at G1 in response to Sln1 inactivation. The sln1ts4 or wild-type strains were synchronized with α-factor for 2 h, shifted to 37°C for 1 h and then released into YPD medium at 37°C. Total DNA content was assessed by FACS analysis and presented as cell counts (y-axis) versus 1C and 2C DNA content (x-axis). The percentage of budding cells is depicted on the right-hand side of each graph. (B) Cells arrest at G2 in response to Sln1 inactivation. The sln1ts4 or wild-type strains were synchronized with α-factor for 3 h, released into fresh media at 25°C and shifted to 37°C after 50 min in YPD (time 0). Total DNA content was analysed as in (A). (C) Cells arrest at G1 in the presence of osmostress. Wild-type cells were synchronized in G1 phase with α-factor and released into YPD medium containing 0.4 M NaCl. Total DNA content was assessed as in (A). (D) Cells arrest at G2 in the presence of osmostress. Wild-type cells were synchronized with α-factor and released into YPD medium. After 50 min, NaCl was added (0.4 M) (time 0). Total DNA content was assessed as in (A). Download figure Download PowerPoint It was reported that in response to osmostress, cells arrested at G2 (Alexander et al, 2001). To characterize the involvement of Hog1 in this arrest, we took advantage of the system described above in which we analysed the effect of sustained activation of Hog1 in G2 either by mutations in the Sln1 osmosensor or the Pbs2 MAPKK (Maeda et al, 1994; Maeda et al, 1995; Posas et al, 1996; Wurgler-Murphy et al, 1997; Posas and Saito, 1997) and by osmostress. sln1ts4 cells were arrested in G1 phase with α-factor then released from the arrest and monitored until they were in G2 and incubated at nonpermissive temperature. As shown in Figure 1B, a high percentage of cells remained in G2 phase (about 76% after 120 min were budded cells with 2C DNA content, 43% of them contained a single nuclei near the bud neck and showed short spindles as judged by tub1-GFP localization), whereas only a small proportion of wild-type cells remained in G2 (∼39% budded cells with 2C DNA content, and only 3% of them contained a single nuclei near the bud neck after 120 min). Activation of the HOG pathway can also be triggered by expression of an allele of the Ssk2 MAPKKK, which contains a deletion of the N-terminal regulatory domain (Ssk2ΔN), or expression of the active allele of Pbs2 (Pbs2DD). However, the G2 arrest caused by expression of these alleles was not as efficient as observed by the inactivation of Sln1 (Supplementary Figures S1/S4 and data not shown). Osmostress treatment of wild-type cells synchronized at G2 also resulted in delayed G2 exit when compared to nontreated cells (83% of cells in G2 upon 80 min of stress versus 30% without stress) (Figure 1D). Thus, activation of Hog1 results in both G1 and G2 arrest of cell cycle progression. Deletion of SWE1 prevents the G2 arrest caused by activation of the Hog1 MAPK Activation of the HOG pathway at G2 resulted in accumulation of cells with an elongated bud, which is reminiscent of altered Clb2–Cdc28 activity (Figure 2A). The phosphorylation that regulates Cdc28–Clb2 activity is driven by the activities of the Swe1 kinase and the Mih1 phosphatase (Morgan, 1997). Previous results indicated that deletion of SWE1 prevented G2 arrest in response to osmostress (Alexander et al, 2001); however, osmostress causes profound cytoskeleton defects which could be monitored by Swe1 to arrest cell cycle at G2 independently of Hog1. To test whether SWE1 was mediating the G2 arrest caused by Hog1 activation, sln1ts4, sln1ts4 hog1 and sln1ts4swe1 were synchronized at G2 and incubated at nonpermissive temperature. As shown in Figure 2A, upon Hog1 activation, deletion of SWE1 resulted in a clear decrease of cells arrested in G2 when compared to cells carrying wild-type SWE1 (65 versus 93% of cells in G2, respectively, after 80 min at nonpermissive temperature). Correspondingly, deletion of SWE1 prevented bud enlargement. It is worth noting that deletion of SWE1 causes already some morphological defects in cells growing under normal conditions in w303 strain background (Buscemi et al, 2000) (Figure 2B). In contrast, deletion of MIH1 only enhanced the effect of Hog1 activation (not shown). Therefore, Swe1 plays an important role in the G2 arrest caused by activation of the Hog1 MAPK. Figure 2.Swe1 mediates the G2 arrest caused by Hog1 activation. (A) Deletion of HOG1 or SWE1 abolishes cell cycle arrest at G2 caused by Sln1 inactivation. sln1ts4, sln1ts4 hog1Δ, and sln1ts4 swe1Δ strains were synchronized with α-factor for 3 h, released into fresh media at 25°C and shifted to 37°C after 50 min in YPD (time 0). Total DNA content was analysed as in Figure 1A. (B) Wild-type, YPC89 (hog1Δ) and YPC166 (swe1Δ) cells containing an empty vector or PGAL1-Pbs2DD were grown for 7 h in SD-Ura plus galactose and visualized by differential interference contrast. Download figure Download PowerPoint Activity of Cdc28 at G2 can be controlled by its state of phosphorylation and by the levels of Clb2. Clb2 mRNA and protein levels were diminished upon expression of the Pbs2DD (Figure 3A and B). Similar results were observed when Clb2 mRNA levels were monitored in the Sln1ts mutant (Supplementary Figure S1). In addition, the relative levels of Clb2–Cdc28 activity diminished upon expression of the Pbs2DD (Figure 3A and B), as reported to occur in response to osmostress (Alexander et al, 2001). Actually, Cdc28–Clb2 activity was almost not downregulated when the Cdk1 complex was purified from swe1 cells (Supplementary Figure S2). Taken together, Hog1 activation results in a decrease of the activity of the Clb–Cdc28 complex and this modulation is mediated by the downregulation of the Clb2 levels and the activity of the Swe1 kinase. Figure 3.Activation of Hog1 causes a decrease on Clb2 levels and Clb2–Cdc28 activity. (A) The levels of Clb2 protein decreases upon Hog1 activation. Wild-type and YPC89 (hog1Δ) cells containing PGAL1-Pbs2DD were transformed with a plasmid expressing HA-tagged Clb2 under its own promoter. Cells were grown in raffinose and then galactose was added (time 0). Cells were collected at the indicated times and Clb2 was visualized by immunoblotting with monoclonal antibody 12CA5 to HA. Wild-type cells without the Clb2-HA plasmid are shown as negative control (−). (B) Hog1 activation leads to a reduction on the relative levels of Clb2–Cdc28 activity. Cells as in (A); wild type (•) and hog1Δ (▾) were grown in the presence of galactose and collected at the indicated time points. Protein extracts were prepared and Clb2–Cdc28 kinase activity was assessed by an in vitro kinase assay of immunoprecipitated Clb2–Cdc28 using Histone H1 (HH1) as a substrate. Kinase activity was normalized to that of wild type, time 0. Data±s.d. from three independent experiments are shown. Download figure Download PowerPoint Sustained activation of the Hog1 MAPK pathway results in stabilization of the Swe1 kinase We therefore decided to investigate how Swe1 was regulated upon Hog1 activation. It was reported that stability of Swe1 is critical to regulate its activity towards Cdc28–Clb2 (Sia et al, 1996; Sia et al, 1998). Swe1 degradation is stimulated by protein phosphorylation. We therefore analysed the phosphorylation of Swe1 in synchronized cells in the presence or absence of stress (Figure 4A). Osmostress induced a delay in Swe1 phosphorylation, suggesting that Swe1 stability could be affected by Hog1 activation. We then analysed whether Swe1 accumulates in response to Hog1 activation. Expression of Pbs2DD or the addition of NaCl (as reported previously (Sia et al, 1998)) resulted in a strong accumulation of Swe1 protein (Figures 4B and C). Therefore, activation of Hog1 results in accumulation of the CDK regulator, Swe1. Figure 4.Swe1 stabilization is critical for cell survival upon stress. (A) Swe1 phosphorylation is affected by osmostress. Cells expressing epitope-tagged Swe1 from its chromosomal locus were synchronized by pheromone treatment and 50 min after release, cells were subjected (NaCl) or not (control) to 0.4 M NaCl. Swe1 and Cdc28 (loading control) were detected from cell extracts by immunoblotting using specific antibodies. Control strain (without tagged Swe1) was wild type W303 (−). (B) Swe1 accumulates in response to osmostress. Tagged Swe1 was expressed from its chromosomal locus in wild-type cells (YPC130). Cells were synchronized as in (A) and subjected (NaCl) or not (control) to 0.4 M NaCl. Swe1 was detected from yeast extracts using monoclonal anti-myc-specific antibodies. Cdc28 is presented as loading control. Control strain was wild type W303 (−). The percentage of acrylamide on (A) was lower than in (B) to increase the separation between the phosphorylated forms of Swe1. (C) YPC130 strain (which expresses Swe1-myc) containing either control vector (vector) or PGAL1-Pbs2DD plasmid were grown in SD-Ura plus raffinose and after addition of galactose cells were collected at the indicated times. The presence of myc-tagged Swe1 was detected as in (C). The asterisk indicates an unspecific protein detected by the myc antibodies that is kept constant along the culture. It is presented as a loading control. (D) Deletion of SWE1 or SIC1 results in osmosensitive cells. The wild-type (W303) strain and its derivatives hog1Δ, sic1Δ, swe1Δ and sic1Δ swe1Δ mutants were spotted on YPD plates or YPD plates containing 0.4 M, 0.8 M or 1 M NaCl. Growth was scored after 3 days at 30°C. Download figure Download PowerPoint Failure of sic1Δ cells to arrest at G1 upon osmostress causes premature entry into S phase and cells become partially osmosensitive (Escote et al, 2004). Because Swe1 was required for Hog1-mediated G2 arrest and Swe1 was stabilized upon stress, we analysed whether the lack of SWE1 also resulted in cells unable to adapt to osmostress. As shown in Figure 4D, swe1-deficient cells also become partially osmosensitive. It is worth noting that swe1-deficient cells do not arrest at G2 in the presence of high osmolarity and this results in a high percentage of binucleated cells due to a premature entry into mitosis (Figure 2A and data not shown). Synthetic sic1 and swe1 mutations render cells even more osmosensitive than the single mutations, resulting in sensitivity similar to the hog1 mutation (Figure 4D). Therefore, modulation of cell cycle progression at different phases of the cell cycle, in G1 and G2, is required for cell survival upon stress. Hsl7 delocalizes from the bud neck in response to Hog1 activation It is known that localization of the septin cytoskeleton, the activity of the checkpoint protein kinase Hsl1 and the presence of its interacting protein Hsl7 in the bud neck regulate Swe1 stability. Actually, Hsl7 interacts directly with both Hsl1 and Swe1 and mutations that impair either of these interactions, or change the localization of Hsl1 or Hsl7, stabilize Swe1 protein (Barral et al., 1999; McMillan et al, 1999; Shulewitz et al, 1999; Cid et al, 2001; Hanrahan and Snyder, 2003; Lew, 2003; Theesfeld et al, 2003). We then followed the localization of septins (i.e., Cdc11 and Cdc12), of the Hsl1 kinase and the Hsl7 protein fused to GFP. Localization of the septins Cdc11 and Cdc12 (not shown) and Hsl1 kinase in the bud neck were not affected by Hog1 activation, nor by osmostress or expression of Pbs2DD (Figure 5A and C). However, Hsl7 localization dramatically changed upon Hog1 activation. Cells exposed to osmostress showed a dramatic delocalization of Hsl7 from the bud neck. Hsl7 delocalization was transient and correlated to the level of Hog1 activation (Figure 5A and B). In addition, whereas Hsl7 was recruited at the bud neck in more than 70% of control cells, only 15% of these cells displayed Hsl7 in the bud neck when expressing Pbs2DD (Figure 5C). It was reported that when delocalized from the septin ring, Hsl7 is rapidly dephosphorylated (Theesfeld et al, 2003; Sakchaisri et al, 2004). Correspondingly, activation of Hog1 results in rapid Hsl7 dephosphorylation (not shown). Thus, activation of the MAPK results in the delocalization of Hsl7 from the bud neck, which promotes Swe1 stabilization. Figure 5.Localization of Hsl7 is affected by activation of the HOG pathway. (A) Localization of Hsl7 but not Cdc11 or Hsl1 is affected by osmostress. Wild-type, hog1Δ or hsl1Δ cells containing chromosomal GFP-tagged Cdc11, Hsl1 or Hsl7 were grown in YPD and subjected (NaCl) or not (control) to a brief osmostress (10 min at 0.4 M NaCl). GFP proteins were visualized by direct fluorescence. Numbers represent percentage of cells with GFP fluorescence at the bud neck and is the result of the measurement in triplicate of 200 cells each. (B) Time course of Hsl7GFP localization upon osmostress. Wild-type cells containing Hsl7GFP were subjected to 0.4 M NaCl at the indicated times. Hog1 phosphorylation was followed using specific antibodies against phosphorylated p38 in whole-cell extracts. Numbers represent percentage of cells with GFP fluorescence at the bud neck and is the result of the measurement in triplicate of 200 cells each. (C) Localization of Hsl7 is affected by Hog1 activation. Wild-type cells expressing PGAL1-Pbs2DD or a control vector and GFP-tagged Cdc11, Hsl1 or Hsl7 were grown in minimal media containing galactose. Numbers represent percentage of cells with GFP fluorescence at the bud neck and is the result of the measurement in triplicate of 200 cells each. Download figure Download PowerPoint Hog1 interacts and directly phosphorylates the Hsl1 checkpoint kinase To analyse whether the regulation of Hsl7 localization at the bud neck was a direct event regulated by Hog1, we tested whether Hog1 was able to interact with and phosphorylate Hsl1 or Hsl7. In vivo co-precipitation experiments were performed with yeast cells containing a chromosomally HA-tagged Hsl1 or Hsl7 transformed with plasmids that expressed GST-tagged full-length Hog1 or GST control. As shown in Figure 6A, Hog1 is able to interact with Hsl1 but not with Hsl7 (not shown). Therefore, we investigated whether Hsl1 was phosphorylated upon osmostress by Hog1. Yeast cells were exposed to a brief osmotic stress (10 min, 0.4 NaCl) after 50 min of release from pheromone arrest. Osmostress induced a rapid phosphorylation of Hsl1, as seen by a shift on Hsl1 mobility in an SDS–polyacrylamide gel. Interestingly, stress-induced Hsl1 phosphorylation was not observed in a hog1Δ strain (Figure 6B). Figure 6.Hog1 phosphorylates Ser1220 within the Hsl7-binding site in Hsl1. (A) Hog1 interacts with Hsl1 in vivo. Yeast extracts containing GST or GST-Hog1 and untagged or tagged HA-Hsl1 (from its chromosomal locus) were precipitated using glutathione beads and precipitated Hsl1 was probed using anti-HA antibodies. GST proteins were detected using anti-GST antibodies. (B) Hsl1 is phosphorylated upon osmostress in a HOG1-dependent manner. hsl1 cells or hog1 cells were transformed with a plasmid expressing a catalytically inactive Hsl1, contains the K110A mutation (HSL1) or in addition the S1220A mutation (hsl1SA) under the GAL1 promoter. Cells were synchronized with pheromone treatment in the presence of galactose and 50 min after release, cells were subjected (+) to a brief osmotic shock (0.4 M NaCl, 10 min). Extracts were treated (+) with alkaline phosphatase (AP). The effect of AP was prevented by the addition of phosphatase inhibitors (not shown). The presence of Hsl1 and Cdc28 was probed. (C) Hog1 directly phosphorylates the C-terminal region of Hsl1 in vitro. Hog1 and the constitutively activated Pbs2 allele (Pbs2EE) purified from E. coli were incubated in the presence of kinase buffer and ATP. Then, catalytically inactive Hsl1 kinase domain (1–600), an Hsl1 fragment (from aa 600 to 900) or the C-terminal domain of Hsl1 that contains aa 717–1517, purified from E. coli, were added in the presence of radioactive ATP. Phosphorylated proteins were detected by autoradiography (P32) or Coomassie staining. Position of Hsl1 fragments is indicated on the left. (D) Hog1 directly phosphorylates the Ser1220 of Hsl1 in vitro. The wild-type Hsl1 and the Hsl1 S1220A mutant were expressed in E. coli and assayed as in (C). Phosphorylated proteins were detected by autoradiography (P32) or Coomassie staining. Position of Hsl1 proteins is indicated on the left. Download figure Download PowerPoint We then tested whether Hsl1 was directly phosphorylated by the Hog1 MAPK. Several fragments of Hsl1 were expressed and purified from Escherichia coli, and then incubated with activated recombinant Hog1 in presence of radioactive ATP. Those experiments showed that, whereas Hsl7 or Swe1 were not phosphorylated by the MAPK (not shown), Hsl1 was phosphorylated in its C-terminal region (amino acids (aa) 711–1517) but not in the N-terminal domain (aa 1–900) (Figure 6C). It was described that Hsl7 recruitment to the bud neck requires the binding of Hsl7 to the Hsl1 C-terminal region (Shulewitz et al, 1999; Cid et al, 2001). In vitro kinase assays testing several Hsl1 fragments for direct phosphorylation pointed at the Ser1220 as the Hog1 phosphorylation site in the Hsl1 C-terminal region. Mutation of Ser1220 to Ala abolished phosphorylation of Hsl1 by Hog1 (Figure 6D). Correspondingly, in vivo phosphorylation assays showed that the Hsl1S1220A mutant was not phosphorylated upon osmostress (Figure 6B). Interestingly, Ser1220 is situated in the middle of the Hsl7-binding domain in Hsl1 (Shulewitz et al, 1999; Cid et al, 2001). Therefore, the Hog1 MAPK is able to interact and directly phosphorylate a single residue within the Hsl7-binding domain of the Hsl1 kinase. Phosphorylation of Hsl1 S1220 by the Hog1 MAPK is critical for Hsl7 localization, G2 arrest and proper adaptation to osmostress To assess the relevance of the phosphorylation of Hsl1 Ser1220 in the localization of Hsl7 in response to Hog1 activation, we analysed Hsl7 localization in cells expressing wild-type Hsl1 or the unphosphorylatable mutant of Hsl1, Hsl1 S1220A (Hsl1SA). Whereas Hsl7 localization changed dramatically in response to osmostress or in response to Pbs2DD expression in cells containing wild-type Hsl1, it did not change in cells expressing the unphosphorylatable Hsl1SA protein (Figure 7A and B). The amounts of Hsl7-GFP were similar in a wild-type and the Hsl1SA strain (not shown). It is worth noting that in cells carrying the Hsl1 S1220E allele (Hsl1SE, which cannot be phosphorylated by Hog1, but the acid residue mimics the phoshorylated state), the amount of Hsl7 present in the bud neck under normal conditions was already half of that observed in wild-type cells (72% wild type versus 35% HSL1SE). Correspondingly, they showed an abnormal morphology reminiscent to hsl1-deficient cells (data not shown). Figure 7.Phosphorylation of Hsl1 Ser1220 by Hog1 determines Hsl7 localization and is essential for G2 arrest and survival upon stress. (A) Mutation of Hsl1 Ser1220 affects Hsl7 localization. The hsl1Δ HSL7GFP strain transformed with plasmids containing wild-type Hsl1 or the mutated alleles Hsl1 Ser1220A (Hsl1SA) or Hsl1 Ser1220E (Hsl1SE) and PGAL1-Pbs2DD or a control vector were grown as in Figure 5C.