The Skn7 response regulator controls gene expression in the oxidative stress response of the budding yeast Saccharomyces cerevisiae

Brian A. Morgan, Geoffrey R. Banks, W. Mark Toone, Desmond Raitt, Shusuke Kuge, Leland H. Johnston

Author Affiliations

  1. Brian A. Morgan1,2,
  2. Geoffrey R. Banks1,
  3. W. Mark Toone14,
  4. Desmond Raitt1,
  5. Shusuke Kuge3 and
  6. Leland H. Johnston1
  1. 1 Division of Yeast Genetics, National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA, UK
  2. 2 Department of Biochemistry and Genetics, Medical School, University of Newcastle, Newcastle‐upon‐Tyne, NE2 4HH, UK
  3. 3 Institute of Medical Sciences, University of Tokyo, Shirokanedai, Minato‐ku, Tokyo, Japan
  4. 4 Imperial Cancer Research Fund, 44 Lincoln's Inn Fields, London, WC2A 3PX, UK


Deletion of the bacterial two‐component response regulator homologue Skn7 results in sensitivity of yeast to oxidizing agents indicating that Skn7 is involved in the response to this type of stress. Here we demonstrate that following oxidative stress, Skn7 regulates the induction of two genes: TRX2, encoding thioredoxin, and a gene encoding thioredoxin reductase. TRX2 is already known to be induced by oxidative stress dependent on the Yap1 protein, an AP1‐like transcription factor responsible for the induction of gene expression in response to various stresses. The thioredoxin reductase gene has not previously been shown to be activated by oxidative stress and, significantly, we find that it too is regulated by Yap1. The control of at least TRX2 by Skn7 is a direct mechanism as Skn7 binds to the TRX2 gene promoter in vitro. This shows Skn7 to be a transcription factor, at present the only such eukaryotic two‐component signalling protein. Our data further suggest that Skn7 and Yap1 co‐operate on the TRX2 promoter, to induce transcription in response to oxidative stress.


Reactive oxygen species (ROS) are responsible for a wide range of intracellular damage to DNA, proteins and cellular structures (for reviews see Ruis and Schuller, 1995; Moradas‐Ferreira et al., 1996). Free radicals are a normal by‐product of respiring cells and are also produced by a wide range of different environmental chemicals. Hence cells have developed mechanisms to respond to this type of stress, termed the oxidative stress response (OSR). Although a number of the cellular proteins functioning in the OSR have been identified in eukaryotes, little is known regarding the signal transduction pathways that detect and respond to oxidative stress. Much of our knowledge in this field has come from work in the budding yeast Saccharomyces cerevisiae. In yeast, overexpression of two transcription factors, Yap1 and Yap2, was found to confer resistance to a wide range of drugs and oxidizing agents (Moradas‐Ferreira et al., 1996). Yap1 and Yap2 are members of the c‐jun family of proteins which constitute part of the AP1 transcription factor of higher eukaryotes. Both Yap1 and Yap2 contain a basic leucine zipper domain (bZIP) adjacent to a DNA binding domain located at the N‐terminus of the proteins. Null mutants of both YAP1 and YAP2 result in sensitivity of S.cerevisiae to the oxidizing agent hydrogen peroxide (Stephen et al., 1995). Further, Yap1 binds directly to the promoter, and regulates the expression of the TRX2 gene (Kuge and Jones, 1994) encoding thioredoxin which acts in the oxidative stress response to reduce protein disulfides. Yap1 also regulates GSH1, encoding glutamyl cysteine synthetase, in response to hydrogen peroxide. The connection between Yap1 and Yap2, and AP1 of higher cells is further strengthened by the observation that the AP1 is itself regulated by the oxidative stress status of the cell (Moradas‐Ferreira et al., 1996). Other work in S.cerevisiae has also demonstrated that there are distinct signalling pathways to hydrogen peroxide and superoxides, although the transcription factor(s) and signalling pathway for the superoxide response have not been characterized (Collinson and Dawes, 1992; Jamieson, 1992; Flattery‐O'Brien et al., 1993; Stephen et al., 1995). In particular, the GSH1 gene is induced by both hydrogen peroxide and menadione, a superoxide‐generating drug (Stephen et al., 1995). Although YAP1 mutants abolish the hydrogen peroxide‐induced expression of GSH1 they have little effect on menadione‐induced expression (Stephen et al., 1995). Thus yeast mounts a complex defence to oxidative stress involving several different transcription factors and signalling pathways.

Recently it was shown that disruption of the SKN7 gene [also known as POS9 (Krems et al., 1996) and BRY1 (Morgan et al., 1995b)] resulted in yeast cells becoming sensitive to hydrogen peroxide (Krems et al., 1996). SKN7 encodes a potential transcription factor which has been implicated in the regulation of cell wall biosynthesis and the cell cycle (Brown et al., 1993, 1994; Morgan et al., 1995b). Overexpression of SKN7 suppresses the cell wall defect associated with mutation of the KRE9 gene (Brown et al., 1993) and additionally, the growth defect associated with deletion of PKC1 (Brown et al., 1994). Furthermore, deletion of the SKN7 gene was found to be synthetically lethal in a pkc1Δ background (Brown et al., 1994; Morgan et al., 1995b). As the PKC1 MAP kinase pathway is involved in cell wall biosynthesis (reviewed in Herskowitz, 1995; Igual et al., 1996), this supports the notion that SKN7 might play a role in the expression of cell wall genes. In addition, overexpression of Skn7 suppresses the lethality associated with loss of the G1 transcription factors SBF and MBF, suggesting a cell‐cycle role (Morgan et al., 1995b). These transcription factors recognize the sequences CACGAAAA (an SCB element) and ACGCGT (an MCB element) respectively, in the promoters of genes they regulate. These include the G1 cyclin and DNA synthesis genes (for reviews see Johnston, 1992; Koch and Nasmyth, 1994). The lethality associated with loss of SBF and MBF has been shown to be due to the absence of G1 cyclin expression. High copy SKN7 was shown to restore G1 cyclin expression through the MCB and SCB elements present in the cyclin promoters (Morgan et al., 1995b). SKN7 does not appear to bind directly to MCB and SCB elements (Morgan et al., 1995b) and is more likely regulating an MCB/SCB binding factor other than MBF and SBF. The sequence of SKN7 revealed homology to the DNA binding domain of heat shock factor (HSF1) (Brown et al., 1993; Morgan et al., 1995b) hence it is possible that SKN7 binds to a sequence related to the HSE elements that HSF1 recognizes.

The Skn7 protein contains a potential receiver domain found in the two‐component signal transduction family of proteins in prokaryotes (Brown et al., 1993; Morgan et al., 1995b). In bacteria, these signal transduction systems are a common method of detecting and responding to the environment (for reviews see Bourret et al., 1991; Parkinson, 1993). Generally, the first component, a homodimer histidine kinase present in the cell membrane, detects the signal, and phosphorylates a conserved histidine residue on its partner. This phosphate is then transferred to a conserved aspartic acid residue within the 120 amino acid receiver domain of the second component, the response regulator protein. Response regulator proteins are generally transcription factors which are activated by the phosphorylation. In eukaryotes only a few potential two‐component signal transduction proteins have been identified (reviewed in Morgan et al., 1995a). In no case, including Skn7, have genes regulated by these systems been identified. In S.cerevisiae, genetic screens have identified one histidine kinase, Sln1, and two response regulator proteins, Ssk1 and Skn7. Sln1 and Ssk1 were shown to act in the regulation of the response of yeast to osmolarity by regulating the Hog1 MAP kinase pathway (Maeda et al., 1994, 1995). However, deletion of SKN7 does not result in any osmolarity defect (B.A.Morgan, unpublished observations). The only stress defect known for skn7 mutants is in oxidative stress as mentioned above, although the role of Skn7 in the OSR remains unclear. Recently Krems et al. (1996) proposed that the receiver domain was essential for the ability of Skn7 to confer cellular resistance to oxidizing agents suggesting, perhaps, that an upstream histidine kinase could be important for the activity of Skn7 in the OSR.

Here we have identified a direct role for Skn7 in the OSR through the induction of TRX2 and also a gene encoding thioredoxin reductase. The Skn7 protein binds directly to the TRX2 promoter and co‐operates with the Yap1 protein to induce gene expression. This is the first example of a gene known to be directly regulated by a response regulator protein in eukaryotes. However, in contrast to Krems et al. (1996) we find that signalling through the receiver domain of Skn7 appears to have no role in the OSR.


skn7Δ strains are sensitive to a range of oxidizing agents

skn7Δ strains are sensitive to hydrogen peroxide (Krems et al., 1996). We have confirmed this result and find that skn7Δ strains are sensitive to a range of oxidizing agents including tetra‐butyl hydrogen peroxide, cadmium and menadione, but not significantly sensitive to diamide (data not shown). Thus the Skn7 protein is required for the cellular response to a variety of free radicals. Deletion of the YAP1 gene also results in sensitivity of yeast cells to many of these agents (Kuge and Jones, 1994). The sensitivity of yap1Δ strains has been shown to be due to the role of the Yap1 protein in the induction of several genes which function in the OSR. Significantly, deletion of the SKN7 gene does not enhance the sensitivity of a yap1Δ strain to diamide, hydrogen peroxide, cadmium or menadione (Krems et al., 1996; our unpublished observations) suggesting that SKN7 and YAP1 are epistatic, functioning in the same pathway. The basis for the sensitivity of the skn7Δ strain to oxidizing agents is unknown but a likely explanation, based on the epistasis of SKN7 and YAP1, is that Skn7 affects the expression of a set of genes also regulated by Yap1 in response to oxidative stress.

Skn7 is required for the induction of thioredoxin and thioredoxin reductase gene expression by the OSR

Expression of the TRX2 gene is induced in response to several oxidizing agents including hydrogen peroxide; this induction is under the control of the Yap1 protein which binds the TRX2 promoter (Kuge and Jones, 1994). To examine the role of Skn7 in TRX2 gene expression, various yeast strains were treated with hydrogen peroxide and total RNA was isolated and probed with a TRX2‐specific gene probe (Figure 1A). In yap1Δ strains the induction of TRX2 is much reduced, although some residual induction remains as observed previously (Kuge and Jones, 1994). Significantly, deletion of the SKN7 gene mimics the deletion of the YAP1 gene. TRX2 induction is almost abolished but once again residual induction is observed. The skn7Δyap1Δ double mutant is no more defective in TRX2 induction than either single mutant, consistent with the genetic epistasis described above. The residual induction of TRX2 that occurs even in the skn7Δyap1Δ double mutant suggests the existence of a further minor induction mechanism. However, the important result is that skn7Δ and yap1Δ mutants appear to have identical phenotypes with respect to TRX2 induction, consistent with their functioning in the same pathway.

Figure 1.Figure 1.
Figure 1.

Skn7 and Yap1 regulate the expression of TRX2 and TRR1 in response to hydrogen peroxide. (A) Northern blot analysis of RNA isolated from different mid‐log yeast strains prior to (lanes 1, 3, 5, 7 and 9) and following treatment with 1 mM hydrogen peroxide for 1 h at 25°C (lanes 2, 4, 6, 8 and 10). Strains used were W303‐1a (lanes 1 and 2), skn7Δ (lanes 3 and 4), W303‐1b (lanes 5 and 6), yap1Δ (lanes 7 and 8) and skn7Δyap1Δ (lanes 9 and 10). The panels below the data show quantitation of the RNA levels relative to the actin transcript. (B) Comparison of potential Yap1 binding sites (Kuge and Jones, 1994) with a possible Yap1 binding site in the TRR1 promoter.

Next, the expression of a gene encoding thioredoxin reductase (Chae et al., 1994) was tested. No name has yet been assigned to this gene and we will refer to it as TRR1. Thioredoxin reductase recycles the oxidized form of thioredoxin to the reduced form to scavenge more ROS, suggesting that TRR1 expression may also be induced by ROS. In addition, the TRR1 promoter contains a potential Yap1 binding site, identical to the SV40 AP‐1 site (Figure 1B). Northern blot analyses revealed that the TRR1 gene is indeed induced by hydrogen peroxide and that this induction is regulated by the YAP1 gene (Figure 1A). Furthermore, like TRX2 expression, Yap1 appears to act principally on the induced levels of TRR1 and not the basal levels which were reduced by some 30% in each case. The induction of TRR1 gene expression was not as pronounced as TRX2, however, but there was a very slight residual induction in a yap1Δ strain. SKN7 is also required for the induction of the TRR1 gene and, in agreement with the data for TRX2, skn7Δ and yap1Δ strains appear identical (Figure 1A). Again the skn7Δyap1Δ double mutant shows no significant additional loss of TRR1 expression as would be expected if these proteins were acting independently to regulate TRR1 gene expression.

As mentioned above, both the skn7Δ and yap1Δ strains are more sensitive than wild type strains to tetra‐butyl hydrogen peroxide. A TRX2 promoter–LacZ reporter construct has been used previously to confirm that TRX2 expression is induced in response to this treatment (Kuge and Jones, 1994). We therefore used this construct to provide independent evidence for the role of Skn7 in TRX2 expression. The reporter plasmid was introduced into skn7Δ and yap1Δ strains and, as expected for the latter strain, there was no induction of LacZ activity following treatment with tetra‐butyl hydrogen peroxide relative to reporter expression in a wild type background (Figure 2). Consistent with the dual function of Yap1 and Skn7 in control of TRX2 expression, the skn7Δ strain showed a similar absence of induction of β‐galactosidase following oxidative stress.

Figure 2.

Skn7 is required for induction by tetra‐butyl hydrogen peroxide of LacZ fused to the TRX2 promoter. Mid‐log cultures of W303‐1a, skn7Δ and yap1Δ strains, transformed with the TRXLACZ plasmid (Kuge and Jones, 1994), were treated with tetra‐butyl hydrogen peroxide for 1 h. β‐galactosidase assays were performed on the untreated and treated cultures.

Diamide, a thiol oxidant (Plummer et al., 1981), also induces TRX2 gene expression and this expression is Yap1 dependent (Kuge and Jones, 1994). We confirmed these data but found that the effect of the SKN7 gene on TRX2 expression following diamide treatment was less marked, although some loss of induction clearly did occur (Figure 3). A similar result was obtained with the TRR1 gene. The double skn7Δ yap1Δ strain showed a loss of induction of TRX2 and TRR1 similar to that observed with yap1Δ alone. These data might partly explain why skn7Δ strains show no significant sensitivity to diamide (see above) whereas yap1Δ strains are sensitive (Kuge and Jones, 1994) and indicates some separation of Skn7 and Yap1 function.

Figure 3.

Skn7 and Yap1 regulate the expression of TRX2 and TRR1 in response to diamide. Northern blot analysis of RNA isolated from different mid‐log yeast strains prior to (lanes 1, 3, 5, 7 and 9) and following treatment with 1.5 mM diamide for 1 h at 25°C (lanes 2, 4, 6, 8 and 10). Strains used were W303‐1a (lanes 1 and 2), skn7Δ (lanes 3 and 4), W303‐1b (lanes 5 and 6), yap1Δ (lanes 7 and 8) and skn7Δyap1Δ (lanes 9 and 10). The panels below the data show quantitation of the RNA levels relative to the actin transcript.

Skn7 binds the TRX2 promoter

The simplest explanation for the data described above is that the Skn7 protein binds directly the TRX2 promoter in order to regulate transcription. To examine this, crude extracts were prepared from a variety of strains and then incubated with a DNA fragment from the TRX2 promoter consisting of the entire upstream region from the potential TATA to the stop codon of the immediate upstream gene. When these protein–DNA complexes were analysed by electrophoretic mobility shift assays (EMSAs) two bands were found to be sensitive to the presence of the Skn7 protein (Figure 4A). The most highly retarded band, termed band 1, was present only in wild type extracts and not in the skn7Δ, yap1Δ or the skn7Δyap1Δ double mutant extracts, suggesting that band 1 is dependent on both the Skn7 and Yap1 proteins. In contrast, a faster‐migrating band, termed band 2, was present in the wild type and yap1Δ extracts, but not the skn7Δ or skn7Δyap1Δ extracts, suggesting that band 2 is dependent on the presence of Skn7 and not Yap1. The most straightforward interpretation of these data is that band 1 contains both Skn7 and Yap1 whilst band 2 contains only Skn7. In agreement with this suggestion, when the skn7Δyap1Δ strain was transformed with a CEN plasmid carrying only the SKN7 gene and the crude extracts were analysed by EMSA, only band 2 reappeared (Figure 4A).

Figure 4.Figure 4.
Figure 4.

Skn7 and Yap1 bind the TRX2 promoter directly. (A) Crude extracts prepared from W303‐1a (lane 1), skn7Δ (lane 2), yap1Δ (lane 3), skn7Δyap1Δ (lane 4) and skn7Δyap1Δ transformed with pBAM1 (lane 5) were analysed by EMSA using the probe TRXΔ1 (see Figure 5 and Materials and methods). Bands which are sensitive to the presence or absence of the Skn7 and Yap1 proteins are indicated as 1–4. (B) The probe TRXΔ1 was mixed with crude extracts prepared from W303‐1a (lane 1), skn7Δ (lane 2), yap1Δ (lane 3) and skn7Δyap1Δ (lane 4), and analysed by EMSA. In lanes 5–8 the same crude extracts also contained the Yap1 polyclonal antiserum while in lanes 9–12 the crude extracts contained the Skn7 polyclonal antiserum. Tracks 13 and 14 are the W303‐1a and yap1Δ extracts incubated with preimmune serum at the same dilution as the polyclonal antisera. (C) Using the probe TRXΔ1, the yap1Δ crude extract was analysed by EMSA. In lane 1 no antiserum was added, in lane 2 the Skn7 polyclonal antiserum, and in lane 3 the preimmune serum at the same dilution as the Skn7 polyclonal antiserum.

To demonstrate that Skn7 binds directly to the TRX2 promoter, an EMSA was performed as above except that a polyclonal antibody raised against the Skn7 protein was included (see Materials and methods). EMSA revealed that bands 1 and 2 were clearly supershifted by the Skn7 antibody (Figure 4B). To confirm this, the experiment was repeated using only yap1Δ extracts and longer autoradiograph exposures (Figure 4C). The antibody supershifted band 2 completely and the preimmune serum was without effect. Thus Skn7 must bind directly to the TRX2 promoter.

The Yap1 protein also binds directly to the TRX2 promoter at two sites termed site 1 and site 2 (Kuge and Jones, 1994; Figure 5). In our EMSA two bands in addition to band 1, bands 3 and 4, are sensitive to the presence of the YAP1 gene (Figure 4). To determine whether all these Yap1‐dependent bands contain the Yap1 protein, various crude extracts were mixed with a polyclonal antibody against Yap1 and the effect on the bands examined. The wild type extracts clearly show a prominent supershifted band and essentially recreate the pattern obtained from yap1Δ extracts (Figure 4B). Hence all the Yap1‐dependent bands contain the Yap1 protein. In agreement with the suggestion above that bands 1, 3 and 4, but not band 2, contain the Yap1 protein, only band 2 appears not to be supershifted by the Yap1 antibody. This has been confirmed by protracting the EMSA electrophoresis step to ensure complete separation of band 1 from the supershifted Yap1 band (data not shown). Hence band 1 contains at least the Skn7 and Yap1 proteins, band 2 contains Skn7, and bands 3 and 4 contain Yap1. Another important conclusion from these data is that significant binding of Skn7 and Yap1 proteins can be detected in the absence of oxidative stress and, furthermore, that Skn7 can bind in the absence of the YAP1 gene and vice versa.

Figure 5.Figure 5.Figure 5.
Figure 5.

Skn7 binding region in the TRX2 promoter. (A) Schematic diagram of the TRX2 promoter and the probes used to identify the Skn7 binding. Near match MCB and SCB elements are indicated. Probes are described in Materials and methods. (B) To map the Skn7 binding site the probes TRXΔ2, TRXΔ3 and TRXΔ6 were mixed with crude extracts prepared from yap1Δ (lane 1) and skn7Δyap1Δ strains (lane 2), and analysed by EMSA. Band 2, containing the Skn7 protein, could only be detected with the TRXΔ3 probe (lane 1). (C) DNA sequence of the 23 nucleotide region of the TRX2 promoter to which Skn7 binds. The potential SCB element and the double point mutation of this sequence in the TRXΔ5 probe is indicated.

To map the TRX2 promoter site responsible for Skn7‐specific binding, a series of overlapping probes for EMSA was constructed from the upstream region of the gene (Figure 5A). The probes TRXΔ1 (Figure 4A), TRXΔ3 (Figure 5B), TRXΔ4 (Figure 6A), WTMCB (Figure 6B) and MUTMCB (Figure 6B), containing the Skn7 binding site, gave a single band which was not detectable in skn7Δ extracts. In contrast, no Skn7‐sensitive band could be detected with probes TRXΔ2 and TRXΔ6 (Figure 5B). As indicated, Skn7 binds within a 23 nucleotide region between the Yap1 binding site 2 (Kuge and Jones, 1994) and the potential TATA sequence (Figure 5A and C). The potential DNA binding domain of Skn7 has homology to HSF1 and, whilst the 23 nucleotide sequence does not contain a complete HSE element, there is some limited homology to one (data not shown). Within the 23 nucleotides is also some limited homology to an SCB element (Figure 5C). Mutation of the CG to TA within this lowers Skn7 binding to this region 20‐fold (Figure 6A). However, the addition of a large molar excess of a fragment from the HO gene promoter containing SCB elements did not compete with Skn7 binding (Figure 6B). Thus, although the CG nucleotides are important for Skn7 binding, this sequence is unlikely to constitute a functional SCB element.

Figure 6.Figure 6.
Figure 6.

The potential SCB and MCB elements are not important for Skn7 binding to the TRX2 promoter. (A) Crude extracts from strains W303‐1a (lanes 1 and 6), skn7Δ (lanes 2 and 7), yap1Δ (lanes 3 and 8), skn7Δyap1Δ (lanes 4 and 9) and skn7Δyap1Δ transformed with a C‐terminal truncation derivative of Skn7 (lanes 5 and 10, band 5), were analysed by EMSA. Extracts were mixed with the probe TRXΔ4 (wild type sequence; lanes 1–5) and with the probe TRXΔ5 (mutant SCB; lanes 6–10) respectively. (B) The role of the potential SCB and MCB elements was further examined in lanes 1–4 (SCB) and lanes 5–8 (MCB). In lanes 1–4 yap1Δ extracts were mixed with the TRXΔ1 probe and excess unlabelled competitor DNA as follows: no competitor (lane 1), TRXΔ1 unlabelled DNA (lane 2), unlabelled SCB containing fragment from the HO promoter (lane 3) and unlabelled lambda phage DNA (lane 4). In lanes 5–8 yap1Δ (lanes 5 and 7) and skn7Δyap1Δ (lanes 6 and 8) crude extracts were mixed with the WTMCB probe (wild type MCB; lanes 5 and 6) and with the MUTMCB probe (mutant MCB; lanes 7 and 8). Band 2 containing Skn7 is indicated. Note that the MUTMCB probe was labelled with approximately half the efficiency of the WTMCB probe explaining the apparent overall lower binding with this probe.

There is also one potential MCB element in the TRX2 promoter (Figure 5A). To test whether this sequence had any role in Skn7 binding, the essential CG core was mutated to TA and a fragment with the mutation tested in EMSA. There was no effect on Skn7 binding (Figure 6B). Thus neither of the potential MCB or SCB sites appear to be relevant for Skn7 activation of TRX2.

Next, we determined whether any significant changes in binding to the TRX2 promoter occurred following treatment of various strains with hydrogen peroxide. Extracts were prepared from the cells at a time when maximal induction of the TRX2 and TRR1 genes had been observed and mixed with a probe from the TRX2 promoter. No obvious effect on the binding of the Yap1 or Skn7 proteins was observed under these conditions (Figure 7A). This implies that, at least in vitro, DNA binding per se is not the major regulatory step in the induction of gene expression and that another mechanism is involved. In addition, no new bands were evident, suggesting that the induction does not involve the binding of a new protein to the TRX2 promoter.

Figure 7.

The effect of H2O2 and diamide on Skn7 and Yap1 binding to the TRX2 promoter. (A) The TRXΔ1 probe was mixed with crude extracts from W303‐1a (lanes 1 and 3) and a yap1Δ strain (lanes 2 and 4) isolated prior to (lanes 1 and 2) and following treatment with mM H2O2 for 1 h at 25°C (lanes 3 and 4). (B) The TRXΔ1 probe was mixed with crude extracts from W303‐1a (lanes 1 and 2) and a yap1Δ strain (lanes 3 and 4) isolated prior to (lanes 1 and 3) and following treatment with 1.5 mM diamide for 1 h at 25°C (lanes 2 and 4).

Kuge and Jones (1994) demonstrated that Yap 1 binding in vitro is induced significantly by diamide. To test whether Skn7 responds in a similar fashion, crude extracts were prepared from various strains treated with 1.5 mM diamide and analysed for the binding of Yap1 and Skn7. Like Kuge and Jones (1994) we observed induction of Yap1 binding following diamide treatment (Figure 7B, lanes 1 and 2). However, in contrast no significant induction of Skn7 binding occurred (Figure 7B, lanes 3 and 4).

Signalling through the receiver domain of Skn7 plays no role in the OSR

Recently it was suggested that the receiver domain of Skn7 was necessary for the resistance of yeast cells to oxidative stress (Krems et al., 1996). Hence we tested whether the receiver domain of Skn7 was required for the induction of TRX2 following oxidative stress. In Skn7, Asp427 is the phosphorylable amino acid (Brown et al., 1994; Morgan et al., 1995b) and previously we demonstrated that mutation of this residue to an asparagine residue blocked the ability of Skn7 to rescue the temperature sensitivity of a swi4ts swi6Δ strain (Morgan et al., 1995b). Hence CEN plasmids containing either the wild type SKN7 gene or with a D427N mutation were introduced into the skn7Δ strain and the sensitivity to hydrogen peroxide was tested (Figure 8A). Both plasmids clearly behave identically; they fully complemented the sensitivity of the skn7Δ strain to hydrogen peroxide. Next, the same strains were grown in minimal medium and treated with 1 mM hydrogen peroxide for 1 h. Northern hybridization revealed that Skn7D427N behaves identically to wild type and TRX2 expression is induced normally (Figure 8B). Finally, the CEN plasmids were introduced into the skn7Δyap1Δ strain. When crude extracts of these strains were characterized by EMSA the wild type and D427N version of Skn7 demonstrated similar binding abilities to a TRX2 promoter fragment (Figure 8C). Thus signalling by means of the receiver domain has no role in the induction of TRX2 in response to oxidative stress or on the binding of Skn7 to the TRX2 promoter.

Figure 8.Figure 8.
Figure 8.

The Asp427 residue in the response regulator domain of Skn7 is not required for the oxidative stress response. (A) 20 μl of a fresh overnight culture of the strains indicated were streaked radially on a YPD plate and allowed to dry. A filter disc saturated with hydrogen peroxide was placed in the centre of the plate. Following incubation at 30°C the extent of inhibition of growth of each strain from the disc was measured. pBAM1 contains wild type Skn7 and pBAM2 contains Skn7 Δ427N. (B) The skn7Δ strain transformed with pBAM1 or pBAM2 was grown to mid‐log in selective media and then treated with 1 mM H2O2 for 1 h. RNA was extracted preincubation (0 time) and after the 1 h treatment (1 h). A Northern blot of these samples pBAM1 0 time (lane 1), pBAM1 1 h (lane 2), pBAM2 0 time (lane 3) and pBAM2 1 h (lane 4) was probed with a labelled TRX2 fragment. (C) Using the TRXΔ1 probe crude extracts of the yap1Δ strain (lane 1), the untransformed skn7Δyap1Δ strain (lane 2) and the skn7Δyap1Δ strain transformed with pBAM1 (lane 3), pBAM2 (lane 4) and the C‐terminal truncation derivative of Skn7 (lane 5) were analysed by EMSA.

The receiver domain as a whole, however, is essential for the response to oxidative stress. A C‐terminal truncation of Skn7, missing amino acids 353–623 containing the receiver domain and the glutamine‐rich region, was isolated by transposon mutagenesis (Morgan et al., 1996). This construct was unable to rescue the sensitivity of skn7Δ to hydrogen peroxide (data not shown). Nevertheless this truncated protein, which contains the domain with homology to the HSF1 DNA binding domain, was still able to bind the TRX2 promoter (band 5 in Figures 6A and 8C). In summary, the receiver domain may play a structural role that is necessary for the oxidative stress response but phosphorylation of Asp427 seems unimportant.


The Yap1 transcription factor is required for the OSR through the induction of gene expression (reviewed in Moradas‐Ferreira et al., 1996). Deletion of the SKN7 gene also results in the increased sensitivity of yeast to oxidative stress. Since response regulator proteins like Skn7 in bacteria are transcription factors, it seemed likely that Skn7, in addition to Yap1, was also required for the induction of gene expression. This is indeed the case: two genes normally induced by oxidative stress, TRX2 and TRR1, require the SKN7 gene for induction. This is a direct mechanism as Skn7 was shown to bind at least the TRX2 promoter. This is the first gene that Skn7 has been shown to regulate directly and is the first example of a direct gene target for any eukaryotic two‐component signal transduction protein.

The binding of Skn7 to a specific region within the TRX2 promoter also, of course, confirms that it is a transcription factor. Despite response regulators being almost entirely transcription factors in prokaryotes, the two‐component systems so far characterized in eukaryotes lie upstream of MAP kinase pathways (Morgan et al., 1995a) and appear to function in regulating MAPKKK activity. This is certainly true of the only other eukaryotic response regulator protein, Ssk1, which controls the Hog1 MAP kinase pathway in budding yeast (see Introduction). Thus our demonstration that Skn7 is a transcription factor makes it the only example known at present that functions as a classical prokaryotic response regulator.

The precise role of Skn7 in the regulation of TRX2 and TRR1 expression is unclear at present. The fact that binding of Yap1 is not significantly affected by the absence of Skn7 and vice versa suggests that no interaction is required for binding to occur. However, EMSA analysis revealed one band that clearly contained both the Yap1 and Skn7 proteins. Attempts to co‐immunoprecipitate Skn7 with Yap1 failed to identify an interaction (unpublished observations). Moreover, Krems et al. (1996) could detect no association using the two‐hybrid technique. Possibly the interaction of Yap1 and Skn7 may require prior binding to DNA, hence a direct contact between Yap1 and Skn7 cannot be excluded. There is indirect evidence that Skn7 may interact with at least one other transcription factor. Skn7 induces G1 cyclin expression through SCB and MCB promoter elements and activates these elements in a reporter construct (Morgan et al., 1995b). The basis of this activation is unclear but since Skn7 does not bind directly to MCBs and SCBs, it may recruit or activate an MCB/SCB binding factor. In spite of this, the near matches to these elements in the TRX2 promoter seem irrelevant to induction of the gene.

Apart from Skn7 and Yap1 function in TRX2 expression, other transcriptional events are also taking place as residual TRX2 induction occurs even in the absence of both proteins. Nevertheless they are responsible for bulk induced levels of the gene. The dual role of Skn7 and Yap1 in control of TRX2 is consistent with their genetic epistasis and also the Northern hybridization data on TRX2 induction. In response to hydrogen peroxide, skn7Δ strains behave identically to yap1Δ strains, the skn7Δyap1Δ double mutant strain behaving like the single deletion of either SKN7 or YAP1. However, when the oxidizing agent used was diamide, deletion of the SKN7 gene had a less deleterious effect than deletion of the YAP1 gene for sensitivity to the agent and for both TRX2 and TRR1 induction. The double delete combination in this case behaves identically to the deletion of YAP1 alone. Moreover, Yap1 binding increases detectably in response to diamide in contrast to Skn7. This increase in Yap1 binding occurs even in the absence of Skn7 (data not shown). Thus there must be some separation in function of the two proteins and some activation of Yap1 which is Skn7 independent. This is not observed with hydrogen peroxide treatment and indicates that Yap1 is regulated differently by these treatments.

The existence of the receiver domain in Skn7 suggests that it responds to phosphorylation from a histidine kinase. Asp427 is the residue within the receiver domain that would normally be phosphorylated by the histidine kinase (Bourret et al., 1991; Parkinson, 1993). In general, a mutation of this aspartate to asparagine is used to test functionality of the receiver; the similarity of the two amino acids ensuring the structural integrity of the domain. We and others have used a D427N mutation to show that signalling to the receiver domain is essential for the function of Skn7 both in cell wall biosynthesis (Brown et al., 1994) and in G1 cyclin expression (Morgan et al., 1995b). Thus the normal behaviour of Skn7 D427N in the OSR is surprising. Krems et al. (1996) used D427A and D427R mutations in examining the Skn7 response to oxidative stress and concluded that the receiver domain was essential. These particular mutations are rarely used in studying the phosphorylable aspartate residue and conceivably they might have disrupted the tertiary structure of the domain, producing the conflicting results. Certainly when we deleted the receiver domain, this destroyed the function of Skn7 in the OSR. In any event, our result with the D427N mutation indicates that the dephosphorylated form of Skn7 is active in the OSR, which does not, of course, necessarily preclude a role for a histidine kinase.

The only known histidine kinase in budding yeast is Sln1, with a role in the response to osmotic stress (Maeda et al., 1994, 1995). The recent completion of the S.cerevisiae genomic sequence has revealed no other likely histidine kinases (B.A.Morgan, unpublished observations), implicating Sln1 in control of Skn7. Deletion of SKN7 does not lead to osmotic sensitivity (B.A.Morgan, unpublished observations), although whether mutation of SLN1 results in increased sensitivity to oxidative stress is not yet clear. Interestingly, Sln1 appears to be constitutively active, the kinase activity becoming inactivated in high osmolarity. Thus, its cognate response regulator, Ssk1, is activated when dephosphorylated. If Sln1 controlled Skn7 activity and responded to oxidative stress as it does to osmotic stress, this might lead to dephosphorylation of the Skn7 receiver domain by analogy with Ssk1. Since the Asp427 is not essential for regulation of OSR genes, this could direct Skn7 to these genes rather than cell wall or cyclin genes. Note, however, that Skn7 is also phosphorylated on serine and/or threonine residues (Brown et al., 1994) so that control of the protein could involve alternative pathways without any histidine kinase involvement. Some form of modification of Skn7 seems to be required for its DNA binding activity, as bacterially produced Skn7 is unable to bind the TRX2 promoter (data not shown).

It is clear that Skn7 regulates at least two groups of genes, those requiring phosphorylation of Asp427, which include cell wall and G1 cyclin genes, and the OSR genes which do not require the phosphorylation (Figure 9). The role of Skn7 in regulation of this disparate group of genes remains obscure. The functional receiver domain argues that the protein transduces signals rather than acting as a non‐specific transcription factor. The feature common to the genes mentioned above might be an involvement in stress. Apart from the OSR, skn7Δ strains are lethal in combination with pkc1Δ, the protein kinase C gene (Brown et al., 1994; Morgan et al., 1995). PKC1 regulates a MAP kinase pathway that responds to hypotonic stress (reviewed in Herskowitz, 1995) and that regulates the synthesis of cell wall genes (Igual et al., 1996). There is also the Skn7 DNA binding domain which is related to that of the heat shock factor. Indeed we have recently found that skn7Δ strains are sensitive to specific heat stresses (D.Raitt, unpublished observations). Concerning the G1 cyclin genes, certain stresses lead to a cell cycle delay in G1 (Rowley et al., 1993; unpublished observations) and we are currently exploring the possibility that Skn7‐induced cyclin expression is necessary for recovery from this arrest.

Figure 9.

Regulation of gene expression by Skn7. Two pathways are envisaged for the regulation of Skn7. In one pathway, a histidine kinase, potentially Sln1, regulates the activity of Skn7 on G1 cyclin and cell wall gene expression in response to unidentified stress. In the other pathway, in response to oxidative stress, Skn7 is regulated independently of phosphorylation of Asp427 in the response regulator domain to activate TRX2 and TRR1 gene expression.

Materials and methods

Strains and growth conditions

The strains of S.cerevisiae used in this study were as follows: W303‐1a (a ade2‐1 trp1‐1 can1‐100 leu2‐3,112 his3‐11 ura3), skn7Δ (a ade2‐1 trp1‐1 can1‐100 leu2‐3,112 his3‐11 ura3 skn7::HIS3), W303‐1b (α ade2‐1 trp1‐1 can1‐100 leu2‐3,112 his3‐11 ura3), yap1Δ (α ade2‐1 trp1‐1 can1‐100 leu2‐3,112 his3‐11 ura3 yap1::TRP1), skn7Δyap1Δ (α ade2‐1 trp1‐1 can1‐11 leu2‐3,112 his3‐11 ura3 skn7::HIS3 yap1::TRP1).

Minimal and rich media used for yeast growth have been described previously (Sherman et al., 1986). Chemicals used in the sensitivity assays were obtained from the following manufacturers: 1.76 M hydrogen peroxide from BDH, 7.7 M tetra‐butyl hydrogen peroxide, diamide, menadione and cadmium sulfate from Sigma. The chemicals were used at the concentrations indicated in the relevant experiments.

Yeast techniques

The skn7Δyap1Δ strain was constructed by disrupting the SKN7 gene in the yap1Δ strain. This was performed by transforming a restriction fragment carrying a HIS3 insertion in the wild type SKN7 gene into the his3 yap1Δ strain and selecting for his+ (Morgan et al., 1995b). Possible double mutants were then tested by PCR to confirm the disruption. The yeast transformations were performed by a derivation of the lithium acetate technique described previously (Schiestl and Gietz, 1989).

β‐galactosidase assays were performed on mid‐log phase cells as described previously (Guarente, 1983). Activities are given in OD units at 420 nm per min per mg of protein. Values represent the averages for duplicate samples in three independent experiments.

Plasmid constructs

An SstI–XbaI fragment was isolated from plasmids pB‐BRY1 and pB‐BRY1 Asn427 respectively (Morgan et al., 1995b) and ligated with the URA3 CEN vector YCplac33 (Gietz and Sugino, 1988) digested with SstI–XbaI to create plasmids pBAM1 (wild type SKN7) and pBAM2 (skn7 D427N) respectively. The C‐terminal truncation derivative of Skn7 was constructed by transposition of a derivative of the bacterial transposon Tn1000 (Morgan et al., 1996) into the plasmid YEp24/BRY1 (Morgan et al., 1995b).

RNA analysis

RNA was extracted from yeast strains grown under described conditions, Northern blotted and probed with various gene‐specific probes as described previously (Morgan et al., 1995b). Equal amounts of RNA were loaded in all lanes. This was confirmed by probing all Northerns with the ACT1 gene encoding actin. The probes for the TRX2 and TRR1 genes were obtained by PCR from the yeast genome using gene‐specific oligonucleotides.

Oligonucleotides used for gel mobility shift assays


Probes were obtained by PCR using the combinations of these oligonucleotides described below. PCR was performed on the TRXLACZ plasmid described in Figure 2. The following oligonucleotides were used to obtain probes 1/2, oligonucleotides Shift 1 and Shift 2: TRXΔ1, oligonucleotides TRXΔ1 and Shift 2; TRXΔ2, oligonucleotides TRXΔ2 and Shift 2; TRXΔ3, oligonucleotides TRXΔ3 and Shift 1; TRXΔ4, oligonucleotides TRXΔ4 and Shift 2; TRXΔ5, oligonucleotides TRXΔ5 and Shift 2; TRXΔ6, oligonucleotides TRXΔ6 and Shift 2; WTMCB, oligonucleotides WTMCB and Shift 1; MUTMCB, oligonucleotides MUTMCB and Shift 1.

Protein extracts and gel mobility shift assays

Yeast cell extracts were prepared from mid‐log cultures. Cell pellets were vortexed for 8×30 s with glass beads in 100–200 μl of 200 mM Tris–HCl pH 8.0, 10 mM MgCl2, 10% glycerol and protease inhibitor mix (PI mix‐100 μg/ml phenylmethyl sulfonyl fluoride, 2 μg/ml aprotinin, 2 μg/ml leupeptin, 2 μg/ml pepstatin A, 50 μg/ml TLCK and 100 μg/ml TPCK). After centrifugation, protein concentrations were determined and extracts were stored at −70°C. EMSAs of DNA–protein complexes were conducted as follows: 7 μg protein extract were incubated with 0.5 ng 32P 5′‐end‐labelled DNA fragment in 10 μl of 25 mM Tris–HCl pH 7.5, 50 mM NaCl, 2 mM EDTA, 7 mM MgCl2, 10% glycerol, PI mix, 1 mg/ml 3‐[(3‐cholamidopropyl)dimethylammonio]‐1‐propanesulfonate (CHAPS) and 50 ng/ml poly(dI:dC) at room temperature for 15 min and on ice for a further 15 min. Competition experiments included a 50‐fold molar excess of unlabelled competitor DNA over labelled probe. In EMSA supershift experiments, anti‐Skn7 or anti‐Yap1 polyclonal serum was added to binding reactions at a serum dilution of 10−2. Reaction mixes were loaded directly onto a 4% (37.5:1) non‐denaturing polyacrylamide gel and electrophoresed at 200 V in 0.6× TBE buffer at 7°C until free DNA reached the bottom of the gel. The gel was dried onto Whatman 3MM paper and autoradiographed.


We would like to thank all our colleagues for helpful advice and discussions, particularly Nicolas Bouquin. We would also like to thank Nic Jones and Derek Jamieson for the gift of strains and plasmids. Desmond Raitt was supported by an EC Network grant, contract number ERBCHRXCT930248.