SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Transcriptional status of the genes needed for galactose utilization in Saccharomyces cerevisiae is controlled by the interplay between the prototypical transcriptional activator Gal4p and the inhibitor protein Gal80p. Relief of the inhibition from Gal4p requires the interaction between Gal80p and the galactokinase paralog, Gal3p. Here, we present evidence that decrease in the intracellular levels of ATP or NADP(H) impairs the GAL gene expression. All these induction defects are rescued by overproducing Gal3p or producing Gal4p mutants with reduced interaction with Gal80p. We further demonstrate that removal of Gal80p from the GAL gene promoter is impaired in these mutants, and that NADP(H) cooperates with Gal3p in causing the dissociation of Gal80p from the in vitro preformed DNA-bound Gal80p–Gal4p complex. We also show that Gal80p is only partially removed from the GAL gene promoter in a mitochondria fusion-deficient mutant where the cotranscriptional mRNA processing is crippled. The efficient dissociation is restored by Gal4p mutants with altered interaction with Gal80p and is correlated with the recovered GAL gene expression. These results indicate that multiple metabolic signals exist to facilitate the efficient and appropriate dissociation of Gal80p from Gal4p by Gal3p to achieve the fully active state of Gal4p in the nucleus.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The ability to recognize and to adapt to the changes in environmental cues is essential for the viability of cells. Intrinsic genetic switches operate in eukaryotic cells to enable appropriate activated and repressed expression of specific groups of genes in response to environmental signals. Macromolecular interactions are often the key to these systems that provide multiple layers of flexibility to co-ordinate gene expression. The transcription of the GAL genes encoding the enzymes of the Leloir pathway in the yeast Saccharomyces cerevisiae is controlled by such a genetic switch and is often considered as a paradigm for eukaryotic transcriptional regulation (Johnston, 1987; Lohr et al., 1995; Sellick and Reece, 2005).

Galactose signalling is believed to operate through a three component switch consisting of a transcription activator Gal4p, a transcription repressor Gal80p and a signal transducer Gal3p. When yeast cells are grown in the absence of galactose, the prototypical transactivator Gal4p is prevented from activating transcription of GAL genes by binding of the inhibitor Gal80p to the 34-aa activation domain at its C terminus (Johnston et al., 1987; Ma and Ptashne, 1987; Lohr et al., 1995). Upon switching the cells to galactose, Gal80p's inhibition of Gal4p is relieved and a rapid induction of the GAL genes occurs. The process requires Gal3p that shows a large degree of homology with galactokinase, Gal1p and has thus been shown to directly interact with Gal80p in a galactose- and ATP-dependent manner (Zenke et al., 1996; Yano and Fukasawa, 1997; Platt and Reece, 1998). The resultant dislocation of Gal80p from its binding sites in Gal4p is thought to expose the Gal4p activating domain allowing its interaction with the transcription initiation machinery (Platt and Reece, 1998; Sil et al., 1999). However, the precise molecular mechanism by which the interaction between Gal3p and Gal80p transduces the galactose signal to the promoter-bound Gal80p–Gal4p complex to activate Gal4p has been unclear. Two, somewhat conflicting, models have thus been proposed concerning the molecular mode-of-action for the relief of Gal80p inhibition of Gal4p. The non-dissociation model supported by one body of evidence has been that a tripartite complex forms between Gal3p and the promoter-bound Gal4p–Gal80p complex in the nucleus and subtle changes occur in the conformation of the Gal4p–Gal80p complex enabling it to be competent to activate transcription (Leuther and Johnston, 1992; Platt and Reece, 1998; Bhaumik et al., 2004). Other more recent evidence, however, points to a dissociation model in which Gal80p dissociates from Gal4p and is sequestrated by cytoplasmically localized Gal3p upon galactose induction. In favour of this model are the observations that Gal3p is predominantly localized in the cytoplasm and that the expression of a plasma membrane-tethered Gal3p does not unduly impair the induction of GAL genes. In addition, chromatin immunoprecipitation experiments suggest that reduced binding of Gal80p to Gal4p occurs after galactose induction (Peng and Hopper, 2000; 2002). The most recent live cell imaging assays further revealed that Gal80p rapidly dissociates from Gal4p in response to galactose and such a dissociation depends on the interaction between Gal3p and Gal80p (Jiang et al., 2009). Although the ability of Gal3p to interact with Gal80p in a galactose-dependent manner is essential for the transcriptional induction of the GAL genes in either of these two models, the events that initiate the dissociation and the following redistribution of Gal80p from the nucleus has been unresolved. Gal3p was recently reported to be detectable in the nucleus and such a nuclear pool of Gal3p is expected to facilitate the rapid dissociation of Gal80p from Gal4p in response to galactose (Jiang et al., 2009). However, the seemingly less Gal3p compared with Gal80p within the nucleus raised the possibility that other signals may exist to synergize with Gal3p to enable the efficient dissociation process. Indeed, evidence has been presented suggesting that NADPH as well as mitochondria-derived signals are involved in regulating the interaction between Gal4p and Gal80p (Bhat and Murthy, 2001; Kumar et al., 2008; Li et al., 2010).

In this report, we present evidence that decrease in the intracellular levels of ATP or NADP(H) caused by specific mutations impairs the GAL gene induction. Our results further show that all these induction defects are rescued by overproducing Gal3p or producing Gal4p mutants with weakened interaction with Gal80p. By using an in vitro pull-down analysis, we demonstrate that NADP(H) cooperates with Gal3p to mediate the efficient dissociation of Gal80p from the preformed DNA-bound Gal80p–Gal4p complex. Finally, we demonstrate that removal of Gal80p from the Gal4p at the GAL gene promoter is abolished in these mutants, which is correlated with the impaired GAL gene transcription. Moreover, we show that Gal80p was only partially removed from Gal4p in dsg1-null in which transcription does occur, but cotranscriptional mRNA processing is affected. These results indicate that multiple metabolic signals exist to facilitate the efficient and appropriate dissociation of Gal80p from Gal4p by Gal3p to achieve the fully active state of Gal4p in the nucleus.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Intracellular ATP shortage causes defective GAL gene induction

In order to assess the importance of cellular energetic status on the induction of the GAL regulon in response to galactose, a rip1Δ strain, which lacks the functional respiratory chain (van den Brink et al., 2009), and an atp1Δ strain, which carries a defective ATP synthase, were constructed (Mabuchi et al., 2000). The intracellular concentrations of ATP in both strains dramatically decreased after the glucose-to-galactose switch (Fig. 1A). Productive activation of transcription by Gal4p as measured by the endogenous GAL1 mRNA as well as the GAL1-LacZ reporter gene expression was almost abolished in both strains in response to galactose (Fig. 1B and C). The lack of transcription upon induction was further found to correlate with the drastically reduced distribution of RNA polymerase II (pol II) across the GAL1 gene (Fig. 1D). To ask whether other gene targets might be affected in these mutants, we examined both the transcription of and polymerase density across two relatively unregulated genes, ADH1 and PMA1, as well as a highly regulated gene, MAL12 (Fig. 2A–C and data not shown). This analysis showed that neither the transcription nor the appearance of pol II across the genes was affected in these deletion mutants under both non-inducing and inducing conditions. To further test the ability of these mutant strains to use galactose as a carbon source, rip1Δ and atp1Δ cells were spotted onto plates containing glucose or galactose. While there was no growth defect for both strains on glucose, disruption of rip1 or atp1 abolished the yeast growth on galactose (Fig. 2D). Taken together, these results indicate that maintenance of an appropriate level of intracellular energy charge is essential for the yeast cells to rapidly turn on the GAL genes probably through enabling the efficient Gal3p–Gal80p interaction, and adapt to a new carbon substrate.

image

Figure 1. Intracellular ATP shortage causes defective GAL gene induction. A. Levels of ATP in parental, Δrip1 and Δatp1 strains. Yeast strains bearing a plasmid pRJR197 expressing WT GAL4 were either cultured in glucose (2%), raffinose(2%) or induced with 2% galactose for 1 h. Intracellular ATP was measured using the ATP Bioluminescence Assay kit CLS II. B. WT (JN1), Δrip1 or Δatp1 yeast, expressing WT GAL4, were induced with 2% galactose for 1 h. RNAs corresponding to endogenous GAL1 were quantified by RT-PCR. Error bars represent SEM. C. GAL1-LacZ reporter activity in galactose media. WT (JN1), Δrip1 or Δatp1 yeast, expressing WT GAL4 were induced with 2% galactose for 1 h. β-galactosidase activity is expressed as Miller units. Error bars represent SEM. D. Deletion of rip1 or atp1 prevents the occupancy of RNA polymerase II on the GAL1 gene. Chip was used to determine the distribution of total CTD.

Download figure to PowerPoint

image

Figure 2. Rip1 or Atp1 is not required for the transcription of ADH1 (A) and MAL12 (B) genes. WT (JN1), Δrip1 or Δatp1 yeast, expressing WT GAL4 was cultured in glucose (2%) or induced with 2% galactose or 2% maltose for 1 h. RNAs corresponding to ADH1 or MAL12 were quantified by RT-PCR. (C) Chip reactions analysed in (Fig. 1D) were used to determine the distribution of total CTD across ADH1. (D) rip1- and atp1-null yeast cannot utilize galactose as a carbon source. Ten-fold seral dilutions of WT or Δrip1 or Δatp1 yeast were spotted onto synthetic media plates containing either glucose or galactose as the carbon source.

Download figure to PowerPoint

Decrease in cellular NADP(H) results in defective GAL gene induction

It has been recently revealed that Gal80p binds the dinucleotide NAD and that the presence of NADP causes a reduced binding of the transcription activation domain (TAD) peptide of Gal4p to Gal80p (Kumar et al., 2008). However, the physiological role that the dinucleotide plays with respect to the regulation of GAL gene expression is unclear. To address this question, we constructed yeasts deleted for the two major NAD kinase genes, pos5 and utr1, whose products play a critical role in determining the levels of NADP in mitochondria and cytoplasm respectively (Pollak et al., 2007). The coding sequence of Nma2, which participates in a salvage pathway of NAD synthesis, was also deleted as a control (Bieganowski and Brenner, 2004). In comparison with the relatively constant cellular levels of NAD(H), the intracellular levels of NADP(H) dropped dramatically in the pos5Δ strain, and were only barely reduced in the utr1Δ strain (Fig. 3A). Analysis of the GAL1 transcription as well as the GAL1-LacZ reporter gene expression revealed that, in contrast to deletion of nma2, transcriptional activation in response to galactose did not occur in the pos5 deletion mutant and was dramatically reduced in the utr1Δ strain (Fig. 3B and C). Similar to ATP synthesis mutants, distribution of pol II across the GAL1 gene was drastically reduced in both mutants (Fig. 3D). Analysis of the transcription of and polymerase density across ADH1 and PMA1 genes revealed no difference between these deletion mutants and wild-type yeasts under both non-inducing and inducing conditions (data not shown). To further test the effect of changes in levels of NADP(H) on the ability of yeasts to use galactose as a carbon source, yeasts cells were spotted onto plates containing various concentrations of hydrogen peroxide. In contrast to deletion of pos5, which virtually eliminated the yeast growth on galactose, yeasts deleted for utr1 are still able to use galactose as a carbon source. This moderate growth was, however, almost abolished by the presence of 1 mM of H2O2 (Fig. 3E). Similarly, growth of WT and nma2-null strains on galactose was also severely affected by the presence of 2 mM of H2O2. This effect of hydrogen peroxide on galactose utilization was also observed for yeast deleted for dsg1 or fzo1, whose gene products have been shown to be required for maintenance of fusion-competent mitochondria, and also to participate in the regulation of GAL gene expression (data not shown) (Fritz et al., 2003; Muratani et al., 2005). Notably, the potential non-specific, toxic effect of H2O2 is not responsible for the inability of yeast to grow on galactose media because all strains grew well on glucose media at concentrations up to 2 mM (Fig. 3E). These results suggest that an appropriate level of intracellular NADP(H) is critical for the yeast cells to efficiently turn on the GAL genes.

image

Figure 3. Decrease in cellular NADP(H) causes defective GAL gene induction. A. Levels of NADP(H) in parental, Δpos5, Δutr1 and Δnma2 strains. WT (JN1), JN1-pos5, JN1-utr1 or JN1-nma2 expressing WT GAL4 was cultured in glucose (2%), raffinose (2%) or induced with 2% galactose for 1 h. Total cellular NADP(H) and NAD(H) was measured by the NAD(H)/NADP(H) assay kit. B. WT (JN1), Δpos5, Δutr1 or Δnma2 yeast, expressing WT GAL4, were induced with 2% galactose for 1 h. RNAs corresponding to endogenous GAL1 were quantified by RT-PCR. Error bars represent SEM. C. GAL1-LacZ reporter activity in galactose media. WT (JN1), Δrip1 or Δatp1 yeast, expressing WT GAL4 were induced with 2% galactose for 1 h. β-galactosidase activity is expressed as Miller units. Error bars represent SEM. D. Deletion of pos5 or utr1 impairs the occupancy of RNA polymerase II on the GAL1 gene. Chip was used to determine the distribution of total CTD. E. Yeast growth on galactose media is impaired by H2O2. Ten-fold seral dilutions of WT, Δpos5, Δutr1 or Δnma2 yeast were spotted onto synthetic media plates containing different concentrations of H2O2 with either glucose or galactose as the carbon source.

Download figure to PowerPoint

Overproduction of Gal3p or Gal4p mutants with altered Gal80p-binding characteristics restores GAL gene expression in mutants deficient in ATP or NADP(H) synthesis

Because the ability of Gal3p to interact with Gal80p constitutes a fundamental event driving the Gal80p–Gal4p dissociation, and thereby transcriptional induction of the GAL genes, the finding that GAL gene transcription is impaired in yeasts deficient in ATP or NADP(H) synthesis prompted us to ask whether expression of GAL3 itself in these yeasts is affected. Analysis of RNA species for GAL3 gene or GAL80 gene revealed that, although the modest induction of both genes did not occur in Δrip1 and Δpos5 cells in response to galactose as did in wild type, the expression levels of GAL3 and GAL80 in these yeasts were not affected under non-inducing conditions (Fig. 4A). Together with previous reports, the results indicated that the perturbed GAL switch due to the disabled dual feedback loops is not responsible for the inability of yeast to respond to galactose induction (Ramsey et al., 2006). To probe further into the possibility that the normal dissociation of Gal80p from Gal4p might be affected upon induction in mutant yeasts, we asked whether destabilizing Gal80p–Gal4p interaction rescues the defective GAL gene expression. In contrast to the induction defect displayed by pos5 or rip1 deletion in GAL80+ cells, induction of a GAL1-LacZ reporter in pos5 or rip1-deleted strains simultaneously lacking Gal80p (gal80-) was as efficient as in wild-type yeast (Fig. 4B). Moreover, activation of a GAL1-LacZ reporter gene was largely recovered after galactose induction in mutant cells overexpressing Gal3p. Similar results were also observed for mutant cells expressing Gal4p mutants, Δ683 or K23R, both of which have been shown to have a weakened interaction with Gal80p (data not shown) (Li et al., 2010). This rescued activation of expression was consistent with the increased distribution of the pol II across the GAL1 gene (Fig. 4C). These results demonstrate that cellular ATP and NADP(H) are involved in regulating the interaction between Gal4p and Ga80p.

image

Figure 4. Defective GAL gene expression is rescued by altering Gal80p–Gal4p interaction. A. The induced but not the non-induced expression GAL3 and GAL80 is abolished in rip1- or pos5-null yeast. WT (JN1), Δrip1 or Δpos5 yeast, expressing WT GAL4 were cultured in raffinose (2%) or induced with 2% galactose for 1 h. RNAs corresponding to endogenous GAL3 or GAL80 were quantified by RT-PCR. Error bars represent SEM. B. GAL1-LacZ reporter activity in galactose media. WT (JN1), Δrip1 or Δpos5 yeast, expressing WT GAL4 simultaneously with Gal3p overproduced or in the absence of Gal80p were induced with 2% galactose for 1 h. β-galactosidase activity is expressed as Miller units. Error bars represent SEM. C. Overexpression of GAL3 or deletion of GAL80 restores the occupancy of RNA polymerase II on the GAL1 gene in rip1-null yeast. Chip was used to determine the distribution of total CTD.

Download figure to PowerPoint

Gal3p cooperates with NADP(H) to dissociate Gal80p from DNA-bound Gal80p–Gal4p complex

Either Gal3p or NADP(H) has been shown to destabilize the Gal80p–Gal4p interaction. To address whether NADP(H) cooperates with Gal3p in the destabilization of DNA-bound Gal4p–Gal80p complex, we performed the following experiment. A preformed DNA–Gal4p–Gal80p complex with recombinant miniGal4p and Gal80p bound to oligonucleotides containing three consensus Gal4p binding sites (UASGAL) was incubated either with NADP alone or together with Gal3p in the presence of ATP and galactose. The amount of Gal80p retained by biotin-conjugated DNA was determined by Western blot. As shown in Fig. 4A, in accordance with previous results, NADP alone at a concentration of up to 10 mM caused a marked decrease in the retention of Gal80p by the DNA-bound Gal4p. In contrast, such a destabilizing effect did not occur with the addition of NAD (Fig. 5A). A similar destabilizing effect was achieved at an even lower concentration of NADP (1 mM) when Gal4p K23R mutant was applied instead of wild-type Gal4p. When the preformed complex was further incubated with varying concentrations of NADP in the presence of a constant concentration of Gal3p plus galactose and ATP, the dissociation of Gal80p occurred at a lower concentration of NADP (Fig. 5B). Importantly, an efficient dissociation of Gal80p from the complex was achieved by Gal3p in the presence of NADP at concentrations neither of which caused detectable dissociation of Gal80p (Fig. 5C). In accordance with these in vitro observations, overexpression of pos5 in Δrip1 strain partly restores the GAL1-LacZ reporter gene expression as evidenced also by the almost normal transcription of GAL1 gene after galactose addition (Fig. 5D). On the basis of these results, we suggest that NADP(H) cooperates with ATP-Gal3p in assuring the efficient dissociation of Gal80p from Gal4p in response to galactose induction.

image

Figure 5. Gal3p cooperates with NADP(H) to dissociate Gal80p from DNA-bound Gal80p–Gal4p complex. A. NADP causes a decrease in the retention of Gal80p by the DNA-bound Gal4p. Biotin-conjugated DNA probes were incubated with recombinant miniGal4p or its derivatives and purified His-tagged Gal80p. The preformed complex was isolated with Streptavidin agarose resin, and further incubated with different concentrations of NAD or NADP. The amount of Gal80p retained by biotin-conjugated DNA was determined by Western blot. B. Gal3p synergize with NADP to reduce the retention of Gal80p in the presence of galactose and ATP. The preformed DNA–Gal4p–Gal80p complex was incubated with different amounts of NADP in the presence of 1 nM Gal3p. Both galactose and ATP were included in the binding buffer. The amount of Gal80p retained by biotin-conjugated DNA was determined by Western blot. C. Retention of Gal80p assayed in (B) was determined with different amounts of Gal3p without or in the presence of 5 mM NADP. D. Activated transcription is rescued by overexpressing POS5 in rip1-null yeast. WT (JN1) or Δrip1 yeast, expressing WT GAL4 or simultaneously expressing WT GAL4 and overexpressing POS5 were induced with 2% galactose for 1 h. RNAs corresponding to endogenous GAL1 were quantified by RT-PCR and β-galactosidase activity is expressed as Miller units. Error bars represent SEM.

Download figure to PowerPoint

Gal80p dissociation from the GAL gene promoter on galactose induction is compromised with cellular ATP or NADP(H) shortage

The results above implicate that the defect in GAL gene activation in ATP or NADP(H) synthesis mutant is to a larger extent caused by the failure of Gal80p to dissociate from Gal4p. Because the association of Gal80p with Gal4p on the promoter of GAL gene under non-inducing conditions and rapid dissociation from Gal4p in response to galactose is critical for the GAL gene transcription regulation, we asked whether the association of Gal80p with Gal4p is changed in vivo with cellular ATP or NADP(H) shortage after the addition of galactose. Chromatin immunoprecipitation (Chip) was used to monitor the extent of the association Gal4p and Gal80p with the UASGAL of the GAL1 promoter. In wild-type cells, the association of Gal80p with the promoter decreased dramatically upon galactose induction whereas there was almost no change for the occupancy by Gal4p before and after galactose addition (Fig. 6A). In rip1Δ or pos5Δ cells, however, the Gal80p association with the UASGAL was only slightly decreased in response to galactose (Fig. 6B). A comparable dissociation of Gal80p to that in wild type was observed when Gal3p was overproduced or Gal4p mutants (Δ683 or K23R) was produced in these mutants (Fig. 6B and data not shown). The failure of rip1Δ or pos5Δ yeasts to achieve the dissociation of Gal80p from Gal4p demonstrates that ATP and NADP(H) are physiological relevant in playing a role in regulating transcriptional activation of GAL genes in response to the change of carbon sources.

image

Figure 6. Gal80p dissociation from the GAL gene promoter on galactose induction is compromised with cellular ATP or NADP(H) shortage. A. Gal4p and Gal80p occupancy at UASGAL region in non-induced and induced sites. Chip assays were performed in Sc319 (GAL80) and Sc320 (GAL80-VP16) expressing an HA-tagged Gal4p under its own promoter under both non-inducing and inducing conditions. B. Dissociation of Gal80p from the UASGAL region is compromised in rip1- or pos5-null yeast and is restored by overproducing Gal3p or producing Gal4p K23R. Chip reactions analysed in (A) was probed for the presence of UASGAL sequences in Δrip1 and Δpos5 cells either expressing WT GAL4 or GAL4 K23R alone or simultaneously expressing WT GAL4 and overexpressing GAL3.

Download figure to PowerPoint

Mitochondria fusion-incompetence impairs Gal80p dissociation from the GAL gene promoter on galactose induction

Earlier studies have shown that deletion of dsg1 coding for the mitochondria-associated F box protein Dsg1/Mdm30 allows the transcription of GAL genes to occur, but the cotranscriptional processing of Gal4p target RNAs is impaired (Muratani et al., 2005). The defect of dsg1Δ yeast in producing functional GAL mRNAs is attributed to the disturbance of Ser5 phosphorylation of pol II. However, the underlying mechanism remains unclear. Analysis of the intracellular levels of ATP and NADP(H) revealed no significant decrease in dsg1Δ yeast (data not shown). To probe further into the status of the UASGAL-bound Gal80p–Gal4p in Δdsg1 cells in response to galactose, we examined the Gal80p association with the UASGAL region. This analysis showed that, unlike Δrip1 or Δpos5 yeast, a significant portion of Gal80p remained associated with the UASGAL region although a remarkable galactose-induced Gal80p dissociation occurred as early as half an hour after galactose addition (Fig. 7A). Analysis of the GAL1 mRNA revealed that the efficient transcription of Gal4p target genes in the absence of dsg1 was severely impaired and was correlated with increased association of Gal80p with the promoter when Δdsg1 cells were induced with galactose under microaerobic conditions (Fig. 7B and C). These results suggest that the incomplete dissociation of Gal80p as compared with that in wild-type strain is sufficient for the transcriptional activation in the absence of dsg1. Considering that the induction defect in dsg1-null was rescued by Gal4 mutants with decreased interaction with Gal80p (Li et al., 2010), we asked whether the dissociation of Gal80p as well as the distribution of Ser5-phosphorylated pol II are restored in Δdsg1 cells by Gal4p K23R. In accordance with previous results (Muratani et al., 2005), while Ser5-phosphorylated CTD of pol II was dramatically reduced within the GAL1 ORF in the absence of Dsg1, robust Ser5 phosphorylation was restored by Gal4p K23R (Fig. 7D). Importantly, the dissociation of Gal80p from the promoter occurred as efficiently as in wild-type strain upon galactose addition (Fig. 7A). These results indicate that Dsg1 participates in the efficient dissociation of Gal80p from the GAL promoter, and also implicate that the precise spatial distribution of Gal80p relative to Gal4p may also be involved in assuring the successful production of mature mRNAs.

image

Figure 7. Dsg1 is involved in regulating the association of Gal80p with the GAL gene promoter. A. Gal80p and Gal4p occupancy at UASGAL region in dsg1-null yeast or in Δdsg1 cells expressing Gal4p K23R. Chip reactions analysed in (Fig. 6B) was probed for the presence of UASGAL sequences at different time points after galactose addition. Error bars represent SEM. B. Activated transcription analysis under microaerobic conditions in the absence of dsg1. WT (JN1) or JN1-dsg1 yeast, expressing WT GAL4 were cultured in raffinose (2%) or induced with 2% galactose under aerobic or microanaerobic for 1 h. RNAs corresponding to endogenous GAL1 were quantified by RT-PCR. Error bars represent SEM. C. Gal80p occupancy at UASGAL region in dsg1-null yeast under microanaerobic induction. Chip reactions analysed in (A) was probed for the presence of UASGAL sequences in Δdsg1 cells. D. Ser5 phosphorylation is restored in dsg1-null yeast by expressing Gal4p K23R. Chip was used to determine the distribution of total and Ser5 phosphorylated CTD at the activated GAL1 gene in WT or Δdsg1 yeast expressing HA-tagged GAL4 or Gal4 K23R. Error bars represent SEM.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Gal4p–Gal80p–Gal3p transcription switch plays a pivotal role in controlling Gal4p activity upon galactose induction (Sil et al., 1999). In this study, we demonstrate that decrease in the intracellular levels of ATP or NADP(H) impairs the GAL gene expression. As well, in the present work we show that all these induction defects are rescued by overproducing Gal3p or producing Gal4p mutants with weakened interaction with Gal80p. We further demonstrate that NADP(H) cooperates with Gal3p in the presence of ATP and galactose to cause the efficient dissociation of Gal80p from the preformed DNA-bound Gal4p. Importantly, as in the case of in vitro Gal80p–Gal4p dissociation, the significant decrease in Gal80p's association with the GAL gene promoter does not occur with cellular ATP or NADP(H) shortage, which is correlated with the impaired GAL gene transcription. Moreover, we show here that Gal80p was only partially removed from the GAL promoter in the dsg1 deletion mutant, suggesting that Dsg1-mediated signals other than ATP and NADP(H) may also exist in regulating the Gal80p–Gal4p interaction.

Relieving the Gal80p's inhibition of Gal4p has been the core of assuring the rapid and efficient response of yeast cells to change of carbon sources in the environment. Among others, the ability of Gal3p to interact with Gal4p-bound Gal80p in a galactose-dependent manner is considered to be essential for the transcriptional induction of GAL genes (Sil et al., 1999; Peng and Hopper, 2002). Although biochemical studies have shown that Gal3p directly interacts with Gal80p in a galactose- and ATP-dependent manner (Platt and Reece, 1998), the expected direct binding of ATP and/or galactose is mainly based on the extraordinary levels of similarity between Gal3p and Gal1p. Therefore, direct proof for the physiological role of cellular energy charge in regulating the GAL regulon has not been established yet. Because activation of GAL genes by wild-type Gal4p is impaired in ATP synthesis mutants and the induction defect is largely due to the compromised dissociation of Gal80p from the Gal4p upon induction, it seems reasonable to conclude that the intracellular energy status plays a key role in regulating the interaction between Gal80p and Gal4p, thus assuring the metabolic flexibility of S. cerevisiae to rapidly adapt to utilization of alternative carbon sources upon the depletion of favoured substrates.

On the other hand, the structure of Gal80p in complex with a 9-mer representing the transcription activation domain of Gal4p clearly shows the presence of NAD. Moreover, the presence of NADP(H) instead of NAD(H) has been reported to destabilize the otherwise in vitro formed complex (Kumar et al., 2008). Nevertheless, it remains unclear as to the physiological role that dinucleotide plays in the regulation of GAL gene expression. Here, for the first time, we have shown that decreasing the cellular levels of NADP(H) impairs galactose-triggered Gal4p-mdeiated GAL gene expression. These results would be difficult to explain by the non-specific effect of the deletion of genes encoding NAD(H) kinases in light of our results showing that the growth of the deletion strains on glucose remained essentially the same as wild-type strain. Moreover, the observation that the decreased association of Gal80p with the GAL gene promoter upon galactose induction is compromised in the presence of NADP(H) shortage indicates that an appropriate mitochondria pool of NADP(H) is necessary to initiate the active state of Gal4p. In this respect, although the decrease in total cellular NADP(H) is barely detectable with deletion of utr1, the relatively mild growth defect of Δutr1 on galactose is probably due to a NADP(H) drain leading to a perturbation of the equilibrium between the mitochondria and the cytoplasmic pool of NADP(H) (Pollak et al., 2007). Given that the nucleus-localized Gal3p is apparently less than those localized in cytoplasm (Jiang et al., 2009), our observation that NADP cooperates with Gal3 in destabilizing the Gal80p–Gal4p complex leads us to propose that maintaining an appropriate cellular, especially a mitochondria pool of NADP(H) facilitates the Gal3p-mediated dissociation of Gal80p from promoter-bound Gal4p and the following set-up of a linked equilibrium between cytoplasm and nucleus that favours the maintenance of the induced state. Note that, at this point, we cannot exclude the possibility that cellular ATP or NADP(H) shortage exerts effect on steps other than activated transcription in GAL gene expression given the observation that cellular ATP shortage also cripples translation of GAL mRNAs (van den Brink et al., 2009). Taken together, these data provide in vivo evidence for maintaining an appropriate intracellular energy charge and a mitochondria NADP(H) pool in S. cerevisiae as the adaptive mechanisms involved in its transition between glucose and galactose metabolism.

GAL induction takes place inefficiently in respiratory deficient yeast (Douglas and Pelroy, 1963). In this respect, the mitochondria-associated F box protein, Dsg1, has been identified to be responsible for the ubiquitylation and destruction of Gal4p engaged in activation. More specifically, the step in Gal4p-mediated activation that is dependent on Dsg1 has been proposed to be related to events governing mRNA processing although the exact molecular mechanism remains unclear (Muratani et al., 2005). In this work, we further extend our previous results showing that deletion of dsg1 indeed compromises the efficient dissociation of Gal80p from Gal4p in vivo. Although we can not at present completely exclude the possibility that Dsg1 effects Gal4–Gal80 interaction through specifically modifying Gal4p by ubiquitylation, a simpler explanation based on our data is that an as yet unidentified signal derived from Dsg1-involved mitochondria metabolism may exist to participate in this regulation. Moreover, our present results together with the observation that Gal80p specifies the interaction between Gal4p and coactivators (Carrozza et al., 2002), provide a possible explanation for the observed transcription defect ascribed to the absence of Dsg1 in that the inefficiently dislocated Gal80p may interfere with steps following transcription initiation. We therefore propose that the efficient dissociation of Gal80p from Gal4p not only determines the on–off of transcription of GAL genes, but the appropriate spatial distribution of Gal80p relative to Gal4p may also be involved in the fine tuning of the whole transcription process including the formation of mature mRNAs.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Strains and plasmids

Yeast strains and plasmids used in this work are listed in Tables 1 and 2 respectively. Yeasts were transformed by the high-efficiency method of Gietz et al. (Gietz et al., 1992). Strains JN1 and dsg1-deletion mutant JN3, containing an integrated GAL1-LacZ reporter, were generously provided by Dr WP Tansey. Strains JN1-80, JN1-rip1, JN1-atp1, JN1-pos5, JN1-utr1, JN1-nma2, JN1-rip-80, JN1-pos5-80 were constructed by replacing respective ORFs with TRP1 or HIS3 cassette amplified by nested PCR. Sc320-dsg1, Sc320-rip1 and Sc320-pos5 were from Sc320 by replacing the chromosomal DSG1, RIP1 or POS5 gene with TRP1 cassette amplified by nested PCR.

Table 1.  Yeast strains used in this study.
StrainGenotypeReference or source
JN1MATα GAL1-LacZ::URA3 trplΔ63 leu2ΔIMuratani et al. (2005)
his3Δ200 lys2Δ385 gal4Δ11
JN3MATα GAL1-LacZ::URA3 trplΔ63 dsg1::LEU2Muratani et al. (2005)
his3Δ200 lys2Δ385 gal4Δ11
JN1-80MATα GAL1-LacZ::URA3 gal80::TRP1 leu2ΔIThis study
his3Δ200 lys2Δ385 gal4Δ11
JN1-rip1MATα GAL1-LacZ::URA3 rip1::TRP1 leu2ΔIThis study
his3Δ200 lys2Δ385 gal4Δ11
JN1-atp1MATα GAL1-LacZ::URA3 atp1::TRP1 leu2ΔIThis study
his3Δ200 lys2Δ385 gal4Δ11
JN1-pos5MATα GAL1-LacZ::URA3 pos5::TRP1 leu2ΔIThis study
his3Δ200 lys2Δ385 gal4Δ11
JN1-utr1MATα GAL1-LacZ::URA3 utr1::TRP1 leu2ΔIThis study
his3Δ200 lys2Δ385 gal4Δ11
JN1-nma2MATα GAL1-LacZ::URA3 nma2::TRP1 leu2ΔIThis study
his3Δ200 lys2Δ385 gal4Δ11
JN1-rip1-80MATα GAL1-LacZ::URA3 rip1::TRP1 leu2ΔIThis study
gal80::HIS3 lys2Δ385 gal4Δ11
JN1-pos5-80MATα GAL1-LacZ::URA3 pos5::TRP1 leu2ΔIThis study
gal80::HIS3 lys2Δ385 gal4Δ11
W303-1AMATa ade2-1 can1-100 ura3-1 leu2-3112 his3-11,15Nehlin et al. (1989)
trp1-1
Sc319MATa ura3 leu2 trp1 his3 ade1 gal4Δ gal80ΔLeuther and Johnston (1992)
gal80::GAL80 GAL1-lacZ
Sc320MATa ura3 leu2 trp1 his3 ade1 gal4Δ gal80ΔLeuther and Johnston (1992)
gal80::GAL80-VP16 GAL1-lacZ
Sc320-rip1MATa ura3 leu2 his3 ade1 gal4Δ gal80ΔThis study
gal80::GAL80-VP16 GAL1-lacZ rip1::TRP1
Sc320-pos5MATa ura3 leu2 his3 ade1 gal4Δ gal80ΔThis study
gal80::GAL80-VP16 GAL1-lacZ pos5::TRP1
Sc320-dsg1MATa ura3 leu2 his3 ade1 gal4Δ gal80ΔThis study
gal80::GAL80-VP16 GAL1-lacZ dsg1::TRP1
DH5aF-,ϕ80dlacZΔM15,Δ(lacZYA-argF)U169,deoR,recA1,endA1 
hsdR17(rk−,mk+),phoA,supE44,λ-,thi-1,gyrA96,relA1
BL21(DE3) pLysSF-, ompT hsdS (rBB-mB-), gal, dcm (DE3,pLysS,Camr)Novagen
Table 2.  Plasmids used in this study.
PlasmidDescriptionReference or source
pRJR197CEN ARS1 HIS3 GAL4pro-2HA-GAL4 (1–881aa)Nehlin et al. (1989)
pRS415CEN6 ARS LEU2Sikorski and Hieter (1989)
pYES22u URA3 GAL1pro-CYC1terminatorInvitrogen
pRS3G4CEN ARS1HIS3 GAL4pro-2HA-GAL4 (1–881aa)Li et al. (2010)
pRS5G4CEN ARS1LEU2 GAL4pro-2HA-GAL4 (1–881aa)This study
pRS3G4K23RCEN ARS1 HIS3 GAL4pro-2HA-GAL4 (1–881aa)Lys23-ArgLi et al. (2010)
pRS425ADH12u LEU2 ADH1pro-ADH1terYu et al. (2009)
pRS5G32u LEU2 ADH1pro-GAL3 (1–521aa)This study
pRS5P52u LEU2 ADH1pro-POS5 (1–415aa)This study
pYG32u URA3 GAL1pro-GAL3 (1–521aa)This study
pET22b Novagen
pE22-MG42HA-Gal4(1-93aa + 768-881aa)Li et al. (2010)
pE22-MG423R2HA-Gal4(1-93aa + 768-881aa) Lys23-ArgLi et al. (2010)
pE22-MG425R2HA-Gal4(1-93aa + 768-881aa)Lys25-ArgLi et al. (2010)
pET15b Novagen
pE15-G86His-Gal80This study

Gal4 and its mutant derivatives were expressed under the control of its own promoter with two tandem copies of HA-epitope being fused at the amino terminus of Gal4 (Wu et al., 1996). The DNA cassette for expressing HA-GAL4 was released from pRS3G4 with SacI and ApaI, and was ligated into pRS415 treated with the same enzymes to obtain pRS5G4 (Li et al., 2010). For overproduction of Gal3p and Pos5p, DNA fragments containing the coding sequence for GAL3 or POS5 were amplified from the yeast genome and the amplified fragment was then inserted between the Sal I and BamH I sites of pRS425 under the control of ADH1 promoter to obtain pRS5G3 and pRS5P5. For expression and purification of Gal3p in W303-1a, the DNA fragment containing the coding sequence for GAL3 with a His-tag being fused at its amino terminus was inserted between the Kpn I and Not I sites of pYES2 to obtain pYG3. Plasmids for expression of miniGal4p and Gal80p in Escherichia coli were constructed as described except that the coding sequence of GAL80 was fused with a His-tag in pET-15b (Li et al., 2010). Oligonucleotides including gene-specific primers used in this study for plasmid constuctions, gene deletion or quantitative PCRs are available on request.

Yeast methods

Yeasts were shake-cultured at 30°C either in YEP (0.5% yeast extract, 1.0% peptone) or synthetic complete (SC) medium lacking specific amino acids as required for plasmid selection (Rose et al., 1990). The carbon sources used were glucose (2.0%), galactose (2.0%), 5% glycerol and 2% lactate, and raffinose (2.0%). For plate growth assay, yeast cells with serial dilutions were spotted onto SC plates containing either 2% glucose or 2% galactose as the carbon source, and growth was allowed for the indicated times at 30°C.

For β-galactosidase assay, yeast cells were grown in SC medium containing 2% glucose to an OD600 of ∼1.0 before being collected and washed with sterile H2O twice. Equal amount of cells was transferred into fresh media containing 2% galactose. After shaking for the indicated times at 30°C, was performed as described (Muratani et al., 2005). All results were an average of at least three independent determinations.

For analysis of steady-state levels of RNA by RT-PCR, total RNA was extracted with a RNeasy Mini Column Kit (Qiagen) according to the manufacturer's instructions. Random hexamer-primed and oligo (dT)-primed reverse-transcriptions were performed with the Reverse-Transcription Kit (Takara) followed by real-time PCR using a Bio-Rad IQ5 (Bio-Rad) thermocycler and the Cyber Green PCR Kit (Takara). All samples were also analysed with primers specific for the constitutively expressed PMA1 or NUP170 gene and this result was used as a ‘loading control’ to normalize results from GAL or MAL12 genes. For analysis of ADH1 and PMA1 mRNAs, equal amount of RNAs were used for RT-PCR and 25s rRNA amplified with primers specific for 25S rDNA gene was used as a control. All experiments were performed at least in triplicate.

Protein purification

Production and purification of Gal80p in E. coli were essentially as described previously (Li et al., 2010). For purification of Gal3p, W303-1a was transformed with pYG3 and transformants were grown in 4 litres of selective medium containing 2% glucose at 30°C to an OD600 of 1.0. The cells were collected and washed with sterile H2O twice, and then continued the culture in 2% galactose for 12 h. The yeasts were collected and resuspended in buffer A [Tris pH7.9, 10 mM MgCl2, 1 mM EDTA, 5% Glycerol (V/V), 1 mM DTT, 300 mM (NH4)2SO4], and lysed with glass beads in Fastprep 120 (Bio101). The lysate was clarified by centrifugation at 14 000 r.p.m. for 20 min and the expressed Gal3p was purified on the Ni2+-NT Aagarose column.

Chromatin immunoprecipitations (Chip)

Chip assays were performed according to the protocol described previously (Gonzalez et al., 2002). Briefly, yeast cells were grown in selective SC medium containing 2% glucose to an OD600 of 1.0. The cells were then collected and washed with H2O twice before being transferred to fresh SC medium containing 2% galactose and allowed to be shake-cultured for different time intervals. Formaldehyde was added to a final concentration of 1% and cross-linking was performed for 15 min before the samples were collected and broken in FA-lysis buffer (50 mM HEPES/KOH pH7.5, 150 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 1 mM PMSF, 1 µg ml-1 leupeptin, 1 µg ml-1 pepstatin) with glass bead (0.45 mm) lysis. Broken cells were further sonicated to obtain an average DNA fragment size of ∼500 bp. Immunoprecipitation was carried out by incubating appropriate antibodies with an aliquot of the clarified yeast lysates containing equal amount of protein (1 mg) at 4°C for 4 h. Fifteen microlitres of protein G/A or proein L beads pre-coated with 1 mg ml-1 BSA and 1 mg ml-1 fish sperm DNA were used per IP. Following immunoprecipitation and extensive sequential washes with FA-lysis, FA500 (50 mM Hepes-KOH, pH 7.5; 500 mM NaCl; 1 mM EDTA; 1% Triton X-100; 0.1% Na Deoxycholate), LiCl (10 mM Tris-HCl, pH8; 250 mM LiCl; 0.5% NP-40;0.5% Na Deoxycholate; 1 mM EDTA), TES (10 mM Tris-HCl, pH 7.5; 1 mM EDTA; 100 mM NaCl) buffers, DNA was eluted with the elution buffer (100 mM Tris-HCl pH 7.8; 10 mM EDTA; 1% SDS; 400 mM NaCl) and recovered by proteinase K treatment of the pelleted samples at 65°C, phenol-chloroform extraction and ethanol precipitation. Antibodies recognizing the largest subunit of pol II CTD (8WG16, Santa Cruz Biotechnology), CTD-Ser5P (H14, Covance), DBD (1-147aa) of Gal4p (Santa Cruz Biotechnology), VP16 (Santa Cruz Biotechnology) and His-tag (Novagen) were used in immunoprecipitations. Quantitative PCR of precipitated chromatin DNAs were performed using a Bio-Rad IQ5 thermocycler (Bio-Rad) and the SYBR Green Supermix (Takara). Relative enrichment of DNAs specific for the promoter (prom), 5′ ends (5P) or 3′ ends (3P) of target coding sequences was calculated by comparing products amplified with negative primer sets for non-transcribed intragenic regions (Ahn et al., 2004).

Gal4p–Gal80p interaction assays

The DNA probe used for DNA-mediated Gal4p–Gal80p complex assembly was a double-stranded (5′TACGGATTAGAAGCCGCCGAGCGGGTGACAGCCCTCCG AAGGAAGACTCTCCTCCGTGC 3′) with the template strand being conjugated by a biotin moiety at the 3′ end, representing three consensus Gal4p binding sites (Wu et al., 1996). The assembly of DNA-bound Gal4p–Gal80p complex was carried out essentially as described (Li et al., 2010). NAD, NADP (Sigma) or Gal3p were then added separately or in combination to equal aliquots of the assembled complex to different final concentrations. The volume was adjusted to 1 ml with buffer A supplemented with or without 4 mM ATP and 4% (W/V) galactose. The sample was then rotated for 1 h at 4°C before being pelleted and washed three times with buffer A. The proteins retained on the beads were eluted by heating at 100°C with 4 × SDS sample buffer. Quantification of the bands detected by the chemiluminescence immunoblots was performed by scanning the immunoblots with an ImageQuant 400 scanning densitometer (GE Healthcare).

Measurement of Adenosine phosphates and NAD(H)/NADP(H)

Intracellular ATP was determined using the ATP Bioluminescence Assay kit CLS II (Roche Diagnostics) in black Costar 96-well microtitre plates according to the manufacturer's instructions. Luminescence was read on a Victor 1420 multi-label counter (Perkin-Elmer Instruments).

Intracellular NAD(H)/NADP(H) was determined using the NAD/NADH Assay Kit and NADP/NADPH Assay Kit (Abcam) in Costar 96-well microtitre plates according to the manufacturer's instructions. Yeast cells (107) were collected and washed twice with cold PBS, and then were broken with glass beads in a Bead beater. The extracts were further homogenized with NAD/NADH extraction buffer or NADP/NADPH extraction buffer followed by centrifugation at 14 000 r.p.m. for 5 min. The extracted solutions were then used to determine the protein concentrations by Bradford assay (Amersco) and concentrations of NAD(H) or NADP(H) by measuring the absorbance at OD450nm on a Microplate Spectrophotometer (Molecular Devices instruments).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We are extremely grateful to W. Tansey, S. Johnston, T. Kodadek for yeast strains and plasmids. We thank J. Svejstrup for help with the chromatin IP protocol and helpful discussions and comments. This work was supported by grants from National Science Fund (No. 30670027), Independent Innovation Foundation of Shandong University (IIFSDU) and a fund from Shandong Science and Technology Program (2007JY02).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References