The Hog1 mitogen-activated protein kinase mediates a hypoxic response in Saccharomyces cerevisiae

Abstract

We have studied hypoxic induction of transcription by studying the seripauperin (PAU) genes of Saccharomyces cerevisiae. Previous studies showed that PAU induction requires the depletion of heme and is dependent upon the transcription factor Upc2. We have now identified additional factors required for PAU induction during hypoxia, including Hog1, a mitogen-activated protein kinase (MAPK) whose signaling pathway originates at the membrane. Our results have led to a model in which heme and ergosterol depletion alters membrane fluidity, thereby activating Hog1 for hypoxic induction. Hypoxic activation of Hog1 is distinct from its previously characterized response to osmotic stress, as the two conditions cause different transcriptional consequences. Furthermore, Hog1-dependent hypoxic activation is independent of the S. cerevisiae general stress response. In addition to Hog1, specific components of the SAGA coactivator complex, including Spt20 and Sgf73, are also required for PAU induction. Interestingly, the mammalian ortholog of Spt20, p38IP, has been previously shown to interact with the mammalian ortholog of Hog1, p38. Taken together, our results have uncovered a previously unknown hypoxic-response pathway that may be conserved throughout eukaryotes.

Figure 1.—

Factors and conditions that control PAU induction. PAU mRNA levels were measured by Northern analysis, with SNR190 serving as the loading control. The experiments were done as described in materials and methods. (A) Upc2 is required for PAU induction. Wild-type (FY2609), upc2Δ (FY2869), hap1Δ (FY2611), hap2Δ (FY2867), and hap1Δ hap2Δ (FY2868) strains were grown in the presence (+) or absence (−) of oxygen for 5 hr. (B) Depletion of heme or ergosterol during aerobic growth leads to PAU induction. To test the effect of heme depletion, wild-type (lanes 1 and 2, FY2609) and hem1Δ (lanes 3 and 4, FY2637) strains were grown with or without 280 µg/ml δ-ala for 12 hr and PAU mRNA levels were measured. To test the effect of ergosterol depletion, wild-type (lanes 5 and 6, FY2609) and GAL1_pr-ERG25 (lanes 7 and 8, FY2870) strains were grown for 12 hr with 2% galactose or 2% glucose and PAU mRNA levels were measured. (C) Addition of heme or ergosterol represses the hypoxic induction of PAU genes. A wild-type (FY2609) strain was grown aerobically (lanes 1 and 4), for 4 hr hypoxically (lanes 2 and 3), or for 8 hr hypoxically (lanes 5 and 6). Heme (at 500 µg/ml, lane 3) or ergosterol (at 20 µg/ml, lane 6) was added to the media 40 min before shifting to hypoxia. For the −heme (lane 2) and −erg (lane 5) controls, an equal volume of the solvent was added as described in materials and methods. (D) Upc2 is required for PAU induction during heme or ergosterol depletion. To test whether Upc2 is required during heme depletion, hem1Δ (FY2637, lane 1) and hem1Δ upc2Δ (FY2872, lane 2) strains were grown in the absence of δ-ala for 12 hr and PAU mRNA levels were measured. To test whether Upc2 is required during ergosterol depletion, GAL1_pr-ERG25 (FY2870, lane 3) and GAL1_pr-ERG25 upc2Δ (FY2871, lane 4) were each grown in glucose for 12 hr and PAU mRNA levels were measured. (E) Manipulating membrane fluidity with DMSO or glycerol affects hypoxic PAU induction. Northern analysis is shown of wild-type cells (FY2609) grown in the presence (+O₂) or absence (−O₂) of oxygen for 4 hr. DMSO (5%, 10%, or 15% v/v; lanes 3–5) or glycerol (5%, 10%, or 15% v/v; lanes 7–9) was added to the medium immediately before shifting cells to hypoxic growth.

Figure 2.—

The role of the Hog1 MAP kinase in PAU and UPC2 induction. (A) PAU hypoxic induction is impaired in a hog1Δ mutant. PAU expression was monitored by Northern blot analysis of wild-type (FY2609) or hog1Δ (FY2873) cells grown in the presence (+) or absence (−) of O₂ for 5 hr. (B) Hog1 becomes phosphorylated after a shift to hypoxic growth. The expression and phosphorylation of Hog1 protein were determined by Western blot analysis of wild-type (FY2609) or HOG1-myc (FY2874) cells grown in the presence (+) or absence (−) of O₂ for 5 hr, using the indicated antibodies. (C) UPC2 mRNA levels are not affected by hog1Δ. UPC2 mRNA levels were monitored by real-time PCR in wild-type (FY2609) or hog1Δ (FY2873) cells in the presence (aerobic) or absence (hypoxic) of oxygen for 5 hr.

Figure 3.—

Role of the Hog1 MAPK cascade in PAU induction. (A) PAU hypoxic induction requires Ssk1 but not Ste11. Northern analysis is shown of PAU expression in wild-type (FY2609), hog1Δ (FY2873), ste11Δ (FY2875), ssk1Δ (FY2878), and ssk1Δ ste11Δ (FY2879) strains grown in the presence (+) or absence (−) of O₂ for 5 hr. (B) PAU hypoxic induction requires Pbs2. Northern analysis is shown of PAU expression in wild-type (FY2609), hog1Δ (FY2873), and pbs2Δ (FY2877) strains grown in the presence (+) or absence (−) of O₂ for 5 hr. (C) PAU hypoxic induction does not require Msn2 or Msn4. Northern analysis is shown of wild-type (FY2609), hog1Δ (FY2873), msn2Δ (FY2895), msn4Δ (FY2896), and msn2Δ msn4Δ (FY2897) strains grown in the presence (+) or absence (−) of O₂ for 5 hr. (D) PAU hypoxic induction requires Mga2 but not Spt23. Northern analysis is shown of wild-type (FY2609), hog1Δ (FY2873), mga2Δ (FY2880), and spt23Δ (FY2881) strains grown in the presence (+) or absence (−) of O₂ for 5 hr.

Figure 4.—

Hog1 distinguishes between the hypoxic and osmotic stress signals. (A) Expression of PAU mRNA during hypoxic growth and during osmotic stress. Northern analysis is shown of PAU and GRE2 mRNA levels in wild-type (FY2609) and hog1Δ (FY2873) strains. Cells were grown for the indicated times in hypoxia (lanes 1–6) or 1 m sorbitol (lanes 7–12). (B) Hog1 is activated with different kinetics under osmotic stress and during hypoxic induction. Western analysis is shown of Hog1 phosphorylation in a wild-type (FY2609) strain grown in 1 m sorbitol (lanes 1–5) or hypoxia (lanes 6–12) for the indicated times.

Figure 5.—

Several hypoxic genes are dependent upon Hog1. (A) Expression microarray analysis showing hypoxic induction of genes in wild-type (FY2609, x-axis) and hog1Δ (FY2873, y-axis) strains. Only the genes hypoxically induced more than twofold in two of two experiments are shown on the plot. The solid line represents the same induction in wild-type and hog1Δ strains, while the dotted lines represent a twofold difference in induction. Genes shown in red have mRNA levels decreased by twofold or more in the hog1Δ mutant. The data point labeled “PAUs” represents the average expression of all 24 PAU genes. (B) Genes in the sterol/heme/Hog1/Upc2 pathway are listed with their function and conserved Upc2 binding site(s). The list of genes includes only those that meet three criteria: (1) regulated by O₂ and Hog1 (Figure 5A), (2) induced by depletion of sterols and depletion of heme (Figure S4), and (3) contain a conserved Upc2 binding site in the promoter, on the basis of the previously determined consensus (TCGTATA or TCGTTYAG) (Cohen et al. 2001; Znaidi et al. 2008).

Figure 6.—

The SAGA coactivator complex is required for PAU induction. (A) Northern analysis of PAU mRNA levels in wild-type (FY2609), spt20Δ (FY2883), gcn5Δ (FY2884), spt3Δ (FY2885), and sgf73Δ (FY2886) strains grown in the presence (+) or absence (−) of O₂ for 5 hr. (B) Northern analysis of PAU mRNA levels in wild-type (FY2609), sgf73Δ (FY2886), sus1Δ (FY2887), sgf11Δ (FY2888), ubp8Δ (FY2889), sus1Δ sgf11Δ ubp8Δ (FY2890), and sgf73Δ sus1Δ sgf11Δ ubp8Δ (FY2891) strains grown in the presence (+) or absence (−) of O₂ for 5 hr. (C) Real-time PCR analysis of UPC2 mRNA levels in wild-type (FY2609), spt20Δ (FY2883), gcn5Δ (FY2884), spt3Δ (FY2885), and sgf73Δ (FY2886) strains grown in the presence (aerobic) or absence (hypoxic) of oxygen for 5 hr. The high standard deviation for some samples is due to an outlier experiment that showed the same relative levels for all of the mutants.

Figure 7.—

A model for hypoxic induction of PAU gene expression. During hypoxic growth, when heme and ergosterol levels are low, membrane fluidity is altered. This change is proposed to activate the Hog1 MAPK cascade via the Sln1/Ssk1 pathway. In turn, activated Hog1 may induce PAU transcription directly, by translocating to the nucleus and activating transcription factors, or indirectly, by influencing signaling events in the cytoplasm, such as activation of Upc2. In addition, a SAGA-dependent, Hog1-independent pathway induces UPC2 transcription. The induction of PAU and other genes by this pathway is important in maintaining the cell wall and membrane in the absence of oxygen. Solid arrows indicate parts of the model strongly supported by our results; dashed arrows represent speculative consequences of Hog1 activation. We have left Mga2 out of the current model, pending further information on its relationship to the factors shown.