When glucose-repressed, Saccharomyces cerevisiae cannot use acetic acid as a carbon source and is inhibited in growth by high levels of this compound, especially at low pH. Cultures exposed to a 100 mM acetate stress activate both the Hog1p and Slt2p stress-activated MAP kinases. Nevertheless, only active Hog1p, not Slt2p, is needed for the acquisition of acetate resistance. Hog1p undergoes more rapid activation by acetate in pH 4.5, than in pH 6.8 cultures, an indication that the acid may have to enter the cells in order to generate the Hog1p activatory signal. Acetate activation of Hog1p is absent in the ssk1Δ and pbs2Δ mutants, but is present in sho1Δ and ste11Δ, showing that it involves the Sln1p branch of the high-osmolarity glycerol (HOG) pathway signaling to Pbs2p. In low-pH (pH 4.5) cultures, the acetate-activated Hog1p, although conferring acetate resistance, does not generate the GPD1 gene or intracellular glycerol inductions that are hallmarks of activation of the HOG pathway by hyperosmotic stress.
Weak organic acids can act either as inhibitory agents or as carbon sources for microbial growth, depending on their concentrations, their ability to enter the cell, and the capacity of the microorganism to degrade the acid. Interest in the weak-acid resistance mechanisms of certain Zygosaccharomyces yeasts and of Saccharomyces cerevisiae stems mainly from the problems that this resistance causes in food preservation. These are yeasts that pose a significant spoilage threat for many materials preserved by low pH, low water activity, and/or the presence of the highest weak organic acid preservative (acetate, propionate, sorbate or benzoate) levels allowed in food and beverage preservation (Fleet, 1992; Steels, 1999, 2000; Piper, 2001).
Certain Zygosaccharomyces yeasts possess the capacity to degrade acetic acid in the presence of glucose (Sousa, 1998). Saccharomyces cerevisiae, in contrast, cannot use acetate as a carbon source in the presence of fermentable sugars. Its plasma membrane monocarboxylate-proton symporters for acetate uptake, its tricarboxylic acid cycle enzymes for the catabolism of acetate in the mitochondrion, and its glyoxylate cycle enzymes for acetate assimilation in the peroxisome, are all subject to glucose repression (Casal, 1996; Paiva, 2004). Instead, glucose-repressed S. cerevisiae is inhibited by high levels of acetic acid, especially at low pH, when this acid exists substantially in the undissociated form. It is possible that S. cerevisiae is frequently inhibited in this way in nature. Acetate is often present in high concentrations in the rotting plant materials where this yeast is in competition with bacteria and other fungi, being a product both of the acetic acid bacteria (notably Acetobacter spp.) and of certain acetic acid-secreting yeasts (the latter organisms are also often encountered as spoilage agents of wine and tea) (Pretorius, 2000; Steels, 2002).
Recent work has shown that Z. bailii and S. cerevisiae undergo apoptotic cell death when exposed to lethal amounts of acetic acid (Ludovico, 2001, 2003). Whether glucose-repressed S. cerevisiae mounts any adaptive response to help it withstand high acetic acid levels is still unclear, although it has been shown that preconditioning at low pH is protective against acetic acid-induced apoptosis (Giannattasio, 2005). Pdr12p, the S. cerevisiae plasma membrane ATP-binding cassette transporter that is strongly induced by, and that confers resistance to, the more lipophilic weak organic acid preservatives with reasonable water solubility (propionate, sorbate, benzoate) does not confer any resistance to acetic acid (Bauer, 2003; Hatzixanthis, 2003).
By screening of protein kinase deletion strains of S. cerevisiae, this study identified a number of the kinases of the high-osmolarity glycerol (HOG) mitogen-activated protein (MAP) kinase cascade as being important in acetic acid resistance. It revealed that an active Hog1p, the terminal MAP kinase of this cascade, is required for cells to become resistant to acetic acid stress.
Materials and methods
Strains and plasmids
The deletion mutant yeast strains used in this study were from Euroscarf (http://www.uni-frankfurt.de/fb15/mikro/euroscarf/), in the BY4741 haploid (his3-Δ1; leu2-Δ0; met15-Δ; ura3-Δ0) genetic background. pES86-HA-HOG1 (harboring an HA-tagged HOG1 coding sequence under ADH1 promoter control) and various mutant derivatives of this same plasmid (T174A, Y176F, and the double T174A Y176F mutant) were kind gifts of Dr David Engelberg.
Yeast growth was on either YPD [2% (w/v) Bacto peptone, 1% yeast extract, 2% glucose, 20 mg L−1 adenine], or dropout 2% glucose (DO) medium (Adams, 1997), the pH of these media being adjusted to 4.5 or 6.8 with HCl or NaOH, respectively, prior to autoclaving. Acetate was added from 8.7 M acetic acid solutions, and titrated to either pH 4.5 or pH 6.8 with NaOH. For growth inhibition in liquid culture, overnight cultures grown at pH 4.5 or pH 6.8 were diluted to OD600 nm=0.05, and then grown for 24 h and 30°C at the same pH in the presence of different acetate concentrations. For agar growth acetate sensitivity assays, c. 5 μL aliquots of serial dilutions of overnight YPD, pH 4.5, cultures were spotted onto YPD pH 4.5, 1.5% agar plates supplemented with the indicated level of acetate. Growth was monitored over 3–5 days at 30°C.
Protein analysis and immunoblots
Total protein extracts were prepared and analyzed by Western blotting, as previously described (Martin, 2000). Western blot analysis of yeast Hog1p, Slt2p and Sba1p used, as primary antibody, rabbit polyclonal antisera raised against these proteins. Slt2p and Hog1p were analyzed with polyclonal anti-Mpk1p (Y-244) and anti-Hog1 (Y-215) (Santa Cruz Biotechnology), respectively. The secondary antibody was horseradish peroxidase antirabbit or antimouse IgG (Amersham) diluted 2000-fold. Analysis of the active forms of Hog1p and Slt2p used antibodies against the dually phosphorylated (Thr180/Tyr182)-p38 MAP kinase or anti-(Thr202/Tyr204)-p44/42 MAP kinase (New England Biolabs). These recognize, respectively, the dually Thr174/Tyr176-phosphorylated Hog1p and the Thr190/Tyr192-phosphorylated Slt2p in yeast extracts (Siderius, 1997; Martin, 2000). Enhanced chemiluminescence reagents (Amersham) were used for detection.
Glycerol accumulation measurements
Cultures in exponential 30°C growth (5 × 107 cells per mL) on pH 4.5 or pH 6.8 DO medium were harvested and transferred to SD medium of the same pH plus 100 mM acetate. After 20 and 90 min, intracellular glycerol levels were measured using the Glycerol assay kit (Megazyme International, Ireland).
GPD1 mRNA analysis
Total yeast RNA was isolated, 10 μg samples were Northern blotted, and the blots were hybridized using standard methods (Piper, 1994) to a GPD1 gene probe, the latter as a PCR product amplified from S. cerevisiae BY4741 total genomic DNA using primers ATGTCTGCTGCTGCTGAT and CTAATCTTCATGTAGATCTA, and then radiolabeled using the Megaprime labeling Kit (Amersham, UK).
Culture pH exerts a strong influence on the response of S. cerevisiae to inhibitory acetic acid stress
Acetic acid (pKa 4.75) is substantially undissociated at pH 4.5, but almost entirely dissociated to the acetate anion at pH 6.8. Its capacity to inhibit S. cerevisiae growth is considerably more marked at the former pH (Fig. 1a). It is generally accepted that this is because as the undissociated, uncharged form it enters cells much more readily than the charged acetate anion (Piper, 2001) (see Discussion). In pH 6.8 cultures, the acetate concentration needed to appreciably inhibit growth (500 mM; Fig. 1a) was found to generate inductions of GPD1 mRNA and intracellular glycerol (Fig. 1). Such a high acetate level (effectively, in these pH 6.8 experiments, a sodium acetate salt stress) is therefore generating events that are characteristic of the yeast response to hyperosmotic stress (Albertyn, 1994; Norbeck, 1996; Ansell, 1997). At the lower pH of 4.5, however, a much lower acetate level is needed to cause comparable growth inhibition. This lower acetate level (100 mM; Fig. 1a) had very different effects on GPD1 mRNA and glycerol, GPD1 mRNA displaying only a very slight, transient induction and the intracellular glycerol declining in these pH 4.5 cultures exposed to inhibitory acetic acid stress (Figs 1b and c). Despite this, pH 4.5 cultures could still induce GPD1 mRNA and intracellular glycerol in response to salt (NaCl) stress (Figs 1b and c). Figure 1 therefore shows that inhibitory levels of acetate stress generate a very different response in pH 4.5, as compared to pH 6.8, cultures.
Acetic acid induces the dually phosphorylated state of both stress-activated MAP kinases of yeast, Hog1p and Slt2p(Mpk1p), but acetate sensitivity arises only with the loss of Hog1p, not of Slt2p
To identify the signaling pathways required for acetic acid resistance, we screened the 101 viable protein kinase deletion mutants of S. cerevisiae (http://mips.gsf.de/genre/proj/yeast/index.jsp) for defective pH 4.5 growth in the presence of 100 mM acetic acid (the maximal concentration that would allow growth of the BY4741 wild type at this pH; Fig. 1a). hog1Δ was one of four acetic acid-sensitive deletion strains identified in this screen, the others being pbs2Δ, ssk1Δ and ctk2Δ (Figs 2a and 3a, and data not shown). Of the five S. cerevisiae MAP kinases, only the loss of Hog1p generated acetate sensitivity, while loss of the cell integrity MAP kinase (Slt2p/Mpk1p) slightly increased acetate resistance (Fig. 2a).
Pbs2p, the activity lost in the pbs2Δ mutant, is the only kinase capable of activating Hog1p (Bell & Engelberg, 2003; O'Rourke & Herskowitz, 2004). This Pbs2p is, in turn, controlled by two different signaling cascades emanating from the Sho1p and Sln1p osmosensors of the plasma membrane (Maeda, 1994). Sho1p controls Pbs2p via Ste20p, Ste50p and Ste11p, whereas Sln1p signals to Pbs2p via Ypd1p, Ssk1p and Ssk2p/Ssk22p (Hohmann, 2002). We found that, while the ssk1Δ mutant is almost as sensitive to acetate as pbs2Δ, the sho1Δ and ste11Δ mutants are relatively insensitive (Fig. 3a). Adaptation to acetate resistance appears, therefore, to require the Sln1p-directed cascade. This is a cascade that is also activated by moderate osmostress, the Sho1p-directed branch of signaling to Pbs2p requiring more severe conditions of osmostress for its activation (O'Rourke & Herskowitz, 2004).
We next investigated whether acetate stress would lead to activation of the Hog1p and Slt2p MAP kinases, initially analyzing pH 4.5 cultures exposed to different levels of acetic acid for 20 min (see Materials and methods). MAP kinases are activated through the dual threonine/tyrosine phosphorylation of a TXY motif (Canagarajah, 1997; Tanoue & Nishida, 2003), in reactions catalyzed by the relevant MAP kinase kinase, either Pbs2p for Hog1p or Mkk1/2p for Slt2p. Levels of the phosphorylated Hog1p were low in the absence of stress, or with 40 or 60 mM acetate addition (Fig. 2b). In contrast, the acetate levels inhibitory for growth of the hog1Δ mutant (80 or 100 mM; Fig. 2a) caused appreciable activation of Hog1p and Slt2p (Fig. 2b). This induction of the active states of both of these MAP kinases was unexpected, since most studies indicate that these two MAP kinases are activated by different conditions of stress. Simultaneous phosphorylation of both Hog1p and Slt2p has, however, been previously observed with heat shock and following hydrogen peroxide addition (Staleva, 2004). Hog1p is generally activated under hyperosmotic conditions (Kapteyn, 2000; Hohmann, 2002). Slt2p is, however, activated in a number of diverse situations, such as under hyperosmotic (Garcia-Rodriguez, 2005) and hypo-osmotic (Davenport, 1995; Alonso-Monge, 2001) conditions, exposure to high temperature (Davenport, 1995; Kamada, 1995; Martin, 2000), exposure to caffeine and vanadate (Martin, 2000), agents that cause weakening of the cell wall (Ketela, 1999; Reinoso-Martin, 2003; Garcia, 2004), endoplasmic reticulum stress (Bonilla & Cunningham, 2003), the polarized growth of budding and mating (Zarzov, 1996), and the operation of a morphogenesis checkpoint, induced by cytoskeletal disorder (Harrison, 2001).
Despite inhibitory acetate stress at pH 4.5 generating the active states of both Hog1p and Slt2p (Figs 2b and 4a), only the former MAP kinase is needed for acetic acid resistance (slt2Δis, unlike hog1Δ, not sensitive to acetic acid; Fig. 2a). Acetate activation of Hog1p was found to be absent in ssk1Δ and pbs2Δ, but not in sho1Δ or ste11Δ mutant cells (Fig. 3b), providing further evidence that the acetate activation of HOG pathway signaling involves the Sln1p/Ssk1p branch of the upstream signaling to the Pbs2p activator of Hog1p. In pH 4.5 cultures, this acetate-activated Hog1p does not generate the GPD1 gene and glycerol inductions that are hallmarks of HOG pathway activation by hyperosmotic stress (Figs 1b and c). The reason why these inductions are not apparent may be a decline in intracellular pH in these cells (see Discussion).
Hog1p activation has been most extensively characterized under conditions of hyperosmotic shock. With the addition of 0.4 M NaCl it is extremely rapid, yet transient (apparent by 1 min, yet gone after 30 min) (Maeda, 1994; Siderius, 1997; Ferrigno, 1998). With more severe osmostress, Hog1p phosphorylation is slower, yet more sustained [with 0.8 M NaCl it is detectable over 60 min (Tamas, 2000), while with 1.4 M NaCl it remains high for several hours before declining (Hohmann, 2002)]. To determine whether acetic acid induces a slow or a rapid response, phosphorylations of Hog1p and Slt2p were measured at intervals following the addition of 100 mM acetic acid to pH 4.5 and pH 6.8 cultures (Fig. 4a). At pH 4.5, increases in activated Hog1p were detectable within 5 min, this phosphorylated Hog1p becoming maximal at 20–40 min but decreasing by 60 min (Fig. 4a). At pH 6.8, acetate induced a comparable degree of Hog1p activation, but the induction was much slower, only becoming apparent some 20–30 min after the acetate was added to the cells (Fig. 4a). At the latter pH the acetic acid is also less inhibitory to growth (Fig. 1a), probably due to slower entry of the acid into the cells (see above). The more rapid Hog1p activation in 100 mM acetate-treated pH 4.5, as compared to pH 6.8, cultures is therefore an indication that the acid may have to enter the cell in order to activate Hog1p. In pH 6.8 cultures it required a much higher, inhibitory acetate concentration of 500 mM (Fig. 1b) for Hog1p to be activated within 5 min (data not shown).
As shown in Fig. 4a, the kinetics of Slt2p phosphorylation with acetic acid stress were essentially similar in cultures maintained at pH 4.5 and pH 6.8. It appears, therefore, that culture pH exerts a much smaller effect on Slt2p activation by acetate than Hog1p activation by acetate.
Expressed in the hog1Δ mutant, the nonphosphorylatable T174A, Y176F single-mutant or T174A/Y176F double-mutant forms of Hog1p could not confer the Hog1p requirement in acetate resistance (Fig. 4b). It is therefore the active Hog1p MAP kinase induced by the stress (Figs 2b, 3b and 4a), not merely the presence of a Hog1p protein, that confers this resistance.
Culture pH strongly affects the ability of acetic acid to inhibit glucose-repressed S. cerevisiae (Fig. 1a), as well as the response of the cells to this acetate stress. Only at pH 6.8, not at pH 4.5, did inhibitory levels of acetate generate appreciable inductions of GPD1 mRNA and intracellular glycerol (Figs 1b and c). In pH 4.5 cultures, undissociated acetic acid that enters the cell will encounter the higher-pH environment of the cytosol (normally close to neutral), leading to its dissociation to protons and the acetate anion. The antimicrobial effects of weak organic acids at low pH have generally been attributed to this intracellular acidification and anion accumulation (Fleet, 1992; Steels, 1999, 2000; Piper, 2001). The proton release can acidify the cytosol, reductions in S. cerevisiae intracellular pH having been demonstrated following acetic acid addition (Arneborg, 2000). This intracellular acidification is thought to lead, in turn, to inhibition of cell metabolic activity (Cassio, 1987; Pampulha & Loureiro-Dias, 2000). This is possibly the reason why inhibitory acetate stress at pH 4.5, while it activates Hog1p (Figs 2b and 3a), does not generate the GPD1 gene and glycerol inductions that are symptomatic of Hog1p becoming activated by osmostress (Fig. 1c). Nevertheless, the Hog1p activation at this pH is essential for acquisition of acetate resistance (Figs 2a and 4b). Hog1p is also essential for S. cerevisiae resistance to another organic acid, citric acid (Lawrence, 2004). Nevertheless it generally requires a considerably higher concentration of citric acid, as compared to acetic acid, to inhibit yeast growth (unpublished observations).
The inducer of the Hog1p-mediated protective response to acetic acid stress may be moderate intracellular levels of the acetate anion. Still higher intracellular acetate levels may trigger apoptotic cell death, such as was seen in earlier studies that employed lethal levels of acetate stress (Ludovico, 2001; Giannattasio, 2005). Hydration of this acetate anion also has the potential to create a high intracellular turgor pressure, which is more likely to cause water entry into the cell (as with hypo-osmotic stress), rather than the water loss of hyperosmotic stress. Such a hypo-osmotic stimulus may be the signal for Slt2p activation with 100 mM acetate stress (Figs 2b and 4a). It might also trigger autophosphorylation of Sln1p, such as is thought to occur with the hypo-osmotic stress that is caused when the fps1Δ mutant accumulates elevated intracellular glycerol (Tao, 1999). Our observation that the ssk1Δ mutant is sensitive to acetate, while the sho1Δ and ste11Δ mutants are relatively insensitive (Fig. 3a), is consistent with the acetate activation of Hog1p being signaled through the Sln1p branch of HOG pathway signaling.
We are currently investigating whether the acquisition of acetate resistance is an indirect effect of Hog1p activation (e.g. requiring Hog1p-dependent transcription), or is mediated through this MAP kinase catalyzing the direct, regulatory phosphorylation of key regulators of homeostasis, such as occurs when Hog1p regulates the Nha1p and Tok1p plasma membrane proteins during NaCl stress (Proft & Struhl, 2004). Activated Hog1p can potentially alter the expression level of a large number of yeast genes, since roughly half of the gene induction or repression events in response to osmostress (involving some 579 genes, roughly 10% of the genome) exhibit a strong dependence on Hog1p (O'Rourke & Herskowitz, 2004).
We are extremely grateful to David Engelberg for plasmids and strains. This work was supported by BBSRC grant BB/A500329/1.