14-3-3 proteins form a family of highly conserved eukaryotic proteins involved in a wide variety of cellular processes, including signalling, apoptosis, cell-cycle control and transcriptional regulation. More than 150 binding partners have been found for these proteins. The yeast Saccharomyces cerevisiae has two genes encoding 14-3-3 proteins, BMH1 and BMH2. A bmh1 bmh2 double mutant is unviable in most laboratory strains. Previously, we constructed a temperature-sensitive bmh2 mutant and showed that mutations in RTG3 and SIN4, both encoding transcriptional regulators, can suppress the temperature-sensitive phenotype of this mutant, suggesting an inhibitory role of the 14-3-3 proteins in Rtg3-dependent transcription [van Heusden and Steensma (2001) Yeast 18, 1479–1491]. In the present paper, we report a genome-wide transcription analysis of a temperature-sensitive bmh2 mutant. Steady-state mRNA levels of 60 open reading frames were increased more than 2.0-fold in the bmh2 mutant, whereas those of 78 open reading frames were decreased more than 2.0-fold. In agreement with our genetic experiments, six genes known to be regulated by Rtg3 showed elevated mRNA levels in the mutant. In addition, several genes with other cellular functions, including those involved in gluconeogenesis, ergosterol biosynthesis and stress response, had altered mRNA levels in the mutant. Our data show that the yeast 14-3-3 proteins negatively regulate Rtg3-dependent transcription, stimulate the transcription of genes involved in ergosterol metabolism and in stress response and are involved in transcription regulation of multiple other genes.
- 14-3-3 proteins
- Saccharomyces cerevisiae
The 14-3-3 proteins form a family of highly conserved acidic dimeric proteins that are present, often in multiple isoforms, in all eukaryotic organisms investigated (reviewed in [1–5]). They bind to more than 150 different proteins and play a role in the regulation of many cellular processes, including signalling, cell-cycle control, apoptosis, exocytosis, cytoskeletal rearrangements, regulation of enzymes and transcription. Although the exact function of the 14-3-3 proteins is still not completely understood, three main mechanisms appear to be important. First, 14-3-3 proteins positively or negatively regulate the activity of enzymes; secondly, 14-3-3 proteins may act as localization anchors, controlling the subcellular localization of proteins; and thirdly, 14-3-3 proteins can function as adaptor molecules or scaffolds, thus stimulating protein–protein interactions. Binding motifs have been identified in a number of proteins that bind to the 14-3-3 proteins. Many, but not all, of these binding motifs contain a phosphorylated serine residue [6–10].
The yeast Saccharomyces cerevisiae has two genes, BMH1 and BMH2, encoding 14-3-3 proteins [11–14]. A bmh1 bmh2 disruption is lethal in most, but not all, laboratory strains, and the lethal bmh1 bmh2 disruption can be complemented by at least four of the Arabidopsis isoforms and by a human and a Dictyostelium isoform [15,16]. As in higher eukaryotes, the S. cerevisiae 14-3-3 proteins are involved in many cellular processes, and many different binding partners have been identified . These include the protein kinases Ste20p  and Yak1p , the protein phosphatase regulator Reg1p , the filament-forming protein Fin1p [20–22], the Mks1 protein , and the transcription factors Rtg3p , Msn2p and Msn4p . Recently, it has been shown that the yeast 14-3-3 proteins bind to cruciform DNA .
In a previous study, we constructed a temperature-sensitive bmh2 mutant by the disruption of both BMH genes and the introduction of a mutated bmh2 allele . We used this mutant to identify extragenic suppressor mutations bypassing the requirement of active 14-3-3 proteins. Recessive mutations in RTG3 and SIN4 resulted in growth at the restrictive temperature. RTG3 encodes a basic helix–loop–helix transcription factor involved in the expression of CIT2 and other genes in respiratory-deficient yeast cells (retrograde signalling) . The Rtg3 protein forms a heterodimer with another basic helix–loop–helix transcription factor (Rtg1p) and binds to the core binding site 5′-GTCAC-3′ (R box) . The expression of Rtg1- and Rtg3-regulated genes is negatively influenced by the target of rapamycin (TOR) signalling pathway . SIN4 encodes a global transcriptional regulator, which can stimulate or repress the expression of several genes and which is a component of the RNA polymerase II complex [29–31]. We showed that the yeast 14-3-3 proteins bind to the Rtg3 protein. Our genetic and biochemical studies suggested that the Rtg3 protein is inactivated by the 14-3-3 proteins. Recently, it was shown that the yeast 14-3-3 proteins also bind to the Mks1 protein , another regulator of retrograde signalling . These studies indicate that the 14-3-3 proteins may have a major role in Rtg3-regulated gene expression. It has also been shown that the activity of the Msn2 and Msn4 transcription factors is influenced by 14-3-3 proteins. The 14-3-3 proteins sequester the phosphorylated forms of the Msn proteins into the cytoplasm .
In the present study, we investigated further the role of 14-3-3 proteins in the regulation of transcription. To this end, we investigated the effect of mutation of the BMH genes on the steady-state mRNA levels in S. cerevisiae at a genome-wide scale. As deletion of both BMH genes is lethal in most laboratory strains, we used a strain with the temperature-sensitive bmh2 allele, which partly complements the bmh1 bmh2 double disruption. We showed that in this bmh mutant, Rtg3-regulated genes have elevated mRNA levels, indicating that the 14-3-3 proteins inhibit the transcription of these genes. In addition, genes involved in gluconeogenesis were activated, and many genes involved in ergosterol synthesis and stress response were down-regulated in the mutant, showing a regulatory role of the 14-3-3 proteins in the transcription of these genes.
MATERIALS AND METHODS
Strains and culture media
Construction of bmh2(Ts) strain GG3096
BMH2 was replaced by URA3 in the MATa strain CEN-PK113-13D as described previously . Subsequently, the bmh2::URA3 allele was replaced by the bmh2(Ts) allele using a DNA fragment obtained by PCR on plasmid YCplac22[bmh2(Ts)]  and selection for 5-fluoro-orotic acid resistance, yielding strain GG3093. BMH1 was deleted in the MATa strain CEN-PK113-7D by replacing the coding region by the kanMX cassette as described by Güldener et al. , yielding strain GG3094. Strains GG3093 and GG3094 were crossed and the resulting diploid was sporulated. After dissection of the asci, MATa haploids were selected having the bmh1::kanMX, bmh2(Ts) and URA3 alleles. One of these haploids, strain GG3096, was analysed further, e.g. the correct integration of the bmh2(Ts) allele was confirmed by sequencing, and used in this study.
Steady-state chemostat cultures were grown in laboratory fermentors (Applicon) of 1 litre working volume, essentially as described in . The cultures were fed with a defined mineral medium containing glucose as the growth-limiting nutrient at a dilution rate of 0.1 h−1 at a temperature of 30 °C. The pH was kept at 5.0±0.2 by addition of 2 M KOH, the airflow was 0.6–0.8 l·min−1. Culture purity was checked by phase-contrast microscopy. For each strain, two independent cultures were run. Steady-state cells were harvested after 9–12 volume changes by pouring samples of approx. 100 ml of culture into a beaker containing approx. 500 ml of liquid nitrogen. The mixture was stirred vigorously, allowing instant freezing of the sample. Frozen samples were broken into pieces and stored at −80 °C.
Pieces of frozen culture containing approx. 2×109 cells were thawed on ice and cells were harvested by centrifugation. Total RNA was isolated using the RNeasy midi kit (Qiagen) after disruption of the cells in the FastPrep instrument (Bio101). Cy3 (cyanine 3)- and Cy5 (cyanine 5)-labelled cDNAs were made using the Fluorescent direct label kit from Agilent Technologies (Stockport, Cheshire, U.K.). The labelled cDNAs were hybridized to Agilent yeast oligonucleotide microarrays containing 10807 60-mer oligonucleotide probes representing 6256 known ORFs (open reading frames) from the S288C strain of S. cerevisiae, according to the instructions from Agilent Technologies. The microarrays were scanned using an Agilent Technologies dual-laser microarray scanner, and the data were extracted using Agilent Feature extraction software. Four microarrays were used: (i) hybridized to Cy5-labelled cDNA from culture 1 of CEN-PK113-7D and Cy3-labelled cDNA from culture 2 of CEN-PK113-7D; (ii) hybridized to Cy5-labelled cDNA from culture 1 of CEN-PK113-7D and Cy3-labelled cDNA from culture 1 of GG3096; (iii) hybridized to Cy5-labelled cDNA from culture 1 of CEN-PK113-7D and Cy3-labelled cDNA from culture 2 of GG3096; and (iv) hybridized to Cy5-labelled cDNA from culture 1 of GG3096 and Cy3-labelled cDNA from culture 1 of CEN-PK113-7D. cDNA labelling, microarray hybridization, scanning and data extraction were performed by ServiceXS, Leiden, The Netherlands. The data are presented as the mean of the data obtained from microarrays 2, 3 and 4. The data are excluded if logCy3/Cy5 obtained from microarray 1 is >0.2 or <−0.2.
RESULTS AND DISCUSSION
Construction of a bmh2 mutant
In order to study the role of 14-3-3 proteins in transcription regulation, we investigated the effect of mutation of the BMH genes on the genome-wide transcription profile. Such studies are complicated by the fact that deletion of both BMH genes is lethal in most laboratory strains [12,13]. In our previous study, we constructed a temperature-sensitive bmh2 mutant by deleting both BMH genes and introduction of a plasmid containing a temperature-sensitive bmh2 allele . In this allele, a single point mutation resulted in the replacement of the serine residue at position 189 by a proline residue. This mutant allele partly complements the lethal bmh1 bmh2 double disruption, allowing growth at 22 °C and 30 °C, but not at 37 °C. To analyse the effect of this bmh mutation on the genome-wide transcription, we preferred to use a mutant lacking auxotrophic markers and plasmids. Therefore we constructed a new mutant (GG3096) in the CEN-PK background in which the BMH1 gene has been deleted and the temperature-sensitive bmh2 allele is integrated at the BMH2 locus. GG3096 has been constructed in the CEN-PK113-7D background, a strain used by several laboratories for transcriptome analyses . At 22 °C and 30 °C, GG3096 grows slower than the wild-type strain CEN-PK113-7D (growth rates at 30 °C: GG3096, 0.13 h−1; CEN-PK113-7D, 0.24 h−1). At 37 °C, GG3096 grows very poorly, although slightly better than our original temperature-sensitive bmh2 strain. Similar to our original mutant, GG3096 is sensitive to 0.02 μg/ml rapamycin, and forms chains of cells and cells with irregular buds at 22 °C (results not shown).
To allow optimal comparison of the expression profile of the mutant GG3096 with that of the wild-type CEN-PK113-7D, both strains were grown in duplicate in glucose-limited chemostat cultures at 30 °C, as it is known that important cultivation conditions, such as dissolved oxygen, metabolite concentrations and pH, change over time in shake-flask cultures. To exclude effects of different growth rates , we grew both strains at the same dilution rate of 0.1 h−1. RNA was extracted from each steady-state culture, labelled with Cy3 or Cy5 and hybridized to commercial oligonucleotide microarrays (Agilent Technologies) representing 6256 S. cerevisiae ORFs. The resulting data set is shown in Table S1 (available at http://www.BiochemJ.org/bj/382/bj3820867add.htm), and is deposited at the GEO-NCBI database under accession numbers GSM13009 to GSM13012. As a control, a hybridization was performed using Cy3- and Cy5-labelled cDNA from two independently grown cultures of CEN-PK113-7D. Only 2% of the ORFs, mainly with a low expression, had a Cy3/Cy5 ratio lower than 0.62 or higher than 1.6. These ORFs were excluded from further analysis. As expected, BMH1 was not expressed in the mutant. The steady-state mRNA levels of 60 ORFs were increased at least 2.0-fold in the bmh2(Ts) mutant. The largest increase (6.8-fold) was found for PCK1, encoding phosphoenolpyruvate carboxykinase, involved in gluconeogenesis (Table 2). The steady-state mRNA levels of 78 ORFs were decreased at least 2.0-fold in the mutant. The largest decrease (8.9-fold) was found for SPS100, involved in spore wall assembly (see Table 4). Similar results could be obtained by other methods. For at least two genes, PCK1 and CIT1, identical results were obtained by Northern blot analysis (Figure 1). The ratio of PCK1 mRNA levels in the mutant relative to those of wild-type was 7.3 for the Northern blot compared with 6.8 for the microarrays (Table 2). For CIT1, these values were 2.1 compared with 1.7 (see Table S1 at http://www.BiochemJ.org/bj/382/bj3820867add.htm). Previously, using a β-galactosidase assay, we showed that the expression of CIT2 was 3.3-fold higher in the bmh2(Ts) mutant (37 °C) than in the wild-type (referred to in ), while the microarrays gave a ratio of 3.8 (Table 2).
Classification of affected genes
The ORFs with a more than 2.0-fold increase in steady-state mRNA levels in the mutant were catalogued according to the functional category defined at the MIPS yeast genome database (http://mips.gsf.de/genre/proj/yeast/index.jsp) and are shown in Table 2. The largest groups of ORFs with increased mRNA levels belong to the ‘metabolism’ (20 ORFs), ‘unclassified proteins’ (15 ORFs), ‘energy’ (eight ORFs) and ‘transcription’ (eight ORFs) functional categories. On the other hand, many of the other functional categories did not contain ORFs with more than 2.0-fold increased mRNA levels (Table 3). After correction for the number of ORFs in each category, the most obvious effects were found for the ‘energy’ (3.1% of the ORFs in this category), ‘metabolism’ (1.9%) and ‘cell rescue, defence and virulence’ (1.8%) functional categories (Table 3).
The ORFs with a more than 2.0-fold decrease in mRNA level in the mutant were catalogued and are shown in Table 4. The largest groups of ORFs with a more than 2.0-fold reduced mRNA level in the bmh2 mutant GG3096 belong to the ‘unclassified proteins’ (27 ORFs), the ‘cell rescue, defence and virulence’ (18 ORFs) and the ‘metabolism’ (16 ORFs) functional categories (Tables 3 and 4). After correction for the number of ORFs in each category, the most obvious effects were found for the ‘cell rescue, defence and virulence’ (6.4% of the ORFs in this category), ‘energy’ (2.7%) and ‘transport facilitation’ (2.2%) functional categories (Table 3). A number of other functional categories did not contain ORFs with more than 2.0-fold reduced mRNA levels.
Effect on Rtg3-regulated genes
Genetic evidence from our previous study suggested an inhibitory role of the 14-3-3 proteins on the Rtg3-dependent transcription . Rtg3 is a basic helix–loop–helix transcription factor involved in the expression of CIT2 and other genes in yeast cells with mitochondrial dysfunction (retrograde signalling) . We showed a physical interaction between the 14-3-3 proteins and the Rtg3 transcription factor. Recently, it was shown that the 14-3-3 proteins also bind to the hyperphosphorylated form of the Mks1 protein . This protein is a negative regulator of retrograde signalling by inactivating Rtg2, a positive regulator of retrograde signalling. The inactive form of Rtg3 is sequestered in the cytoplasm . The expression of six genes involved in glyoxylate, the first steps of gluconeogenesis and glutamate metabolism is known to be regulated by Rtg3 . As shown in Table 5, the steady-state mRNA levels of these six genes are increased between 1.6- and 4.3-fold in GG3096. These data indicate that in agreement with our genetic observations and the observations by Liu et al. , the yeast 14-3-3 proteins have an inhibitory effect on the expression of Rtg3-regulated genes. Probably, many more genes are regulated by Rtg3, as the R box sequence is present in the upstream region of many ORFs. The expression of DIP5, encoding a glutamate–aspartate transporter, is also increased in the bmh2(Ts) mutant (4.9-fold), as well as in respiratory deficient cells . As this gene is involved in glutamate metabolism and it contains two R boxes (in the reverse orientation) in its upstream region, DIP5 is a possible candidate. An RTG3 mutation can suppress the temperature-sensitive phenotype of the bmh2(Ts) mutant. Therefore the increased expression of the Rtg3-regulated genes is most likely, at least partly, to be responsible for the temperature-sensitivity of this mutant. On the other hand, the abnormal morphology of the bmh2(Ts) mutant is not influenced by the RTG3 mutation .
Effect on genes involved in gluconeogenesis
The most prominent increase (6.8-fold) in mRNA levels in the bmh2(Ts) mutant was observed for PCK1, encoding phosphoenolpyruvate carboxykinase involved in gluconeogenesis. Two other genes involved in gluconeogenesis also showed an at least 2.0-fold increased mRNA level, i.e. MDH2 (2.1-fold) and TPI (2.0-fold). These data suggest an inhibitory effect of the yeast 14-3-3 protein on gluconeogenesis.
Effect on genes involved in sterol metabolism
Steady-state mRNA levels of many other genes involved in metabolism are reduced after mutation of the BMH genes (Table 4). This is especially the case for a number of genes involved in ergosterol metabolism: HES1 (4.8-fold), ERG11 (2.7-fold), ERG1 (2.3-fold) and ERG28 (2.1-fold). In addition, mRNA levels of ERG25 and HMG1 were reduced 1.8-fold. These data suggest a stimulatory effect of the yeast 14-3-3 proteins on ergosterol synthesis. Recently, cluster analysis of many data sets of yeast genome-wide expression analyses revealed a set of overlapping transcriptional modules . One of these modules with 27 ORFs (module 67 in ) contains many genes involved in ergosterol synthesis. As shown in Figure 2(B), almost all ORFs in this module had decreased mRNA levels. The mRNA levels of five ORFs in this module (out of 27 ORFs) were decreased more than 2.0-fold, including YPL272c (5.4-fold), HES1 (4.8-fold), ERG11 (2.7-fold), ERG1 (2.3-fold) and ERG28 (2.1-fold) (Figure 2B). In contrast, in the total data set, ORFs with decreased and increased mRNA levels were present in almost equal amounts (Figure 2A). These data support a stimulatory role of the 14-3-3 proteins in the transcription of the ORFs in this module.
Effect on genes encoding stress-related proteins
Mutation of the BMH genes has a very prominent effect on genes in the ‘cell rescue, defence and virulence’ category as the mRNA levels of 18 genes, which is 6.4% of the genes in this category, are reduced more than 2.0-fold. Many of the genes encode proteins belonging to the Pau-protein family (PAU4, 4.0-fold; PAU7, 4.0-fold; YOL161c, 3.7-fold; PAU6, 3.6-fold; PAU3, 3.5-fold; PAU1, 3.5-fold; PAU5, 3.4-fold; YHL046c, 3.1-fold and PAU2, 2.4-fold). Also many ‘unclassified proteins’ with a more than 2.0-fold reduction in mRNA levels have similarity to Pau proteins (YGL261c, 4.7-fold; DAN2, 4.4-fold; YGR294w, 4.2-fold; YMR325w, 4.1-fold; YOR394w, 3.8-fold; YIR041w, 3.7-fold; YKL224c, 3.6-fold; YIL176c, 3.5-fold; YLL064c, 3.5-fold and YDR542w, 3.3-fold). These observations suggest a stimulatory role of the 14-3-3 proteins on the expression of these genes. In addition to genes encoding Pau proteins, many other genes encoding stress-related proteins are affected, both positively and negatively, by mutation of the BMH genes. Reduced expression of stress-related genes may contribute to the temperature-sensitive phenotype of the bmh2(Ts) mutant.
Effect on Msn2- and Msn4-regulated genes
14-3-3 proteins are known to regulate the Msn2 and Msn4 transcription factors by sequestering the phosphorylated forms of these proteins to the cytoplasm . Phosphorylation of the Msn transcription factors is regulated by the RAS-protein kinase A as well as the TOR signalling pathways [25,40–43]. The Msn2 and Msn4 transcription factors bind to a stress-response element in the promoters of many stress-related genes. Thus it is conceivable that in the bmh2(Ts) mutant, the expression of genes having a stress-response element in their promoters is altered. A computer search revealed 81 genes having a promoter containing at least two stress-response elements . Out of these 81 genes, only three showed a more than 2.0-fold increase in mRNA levels, i.e. YNR014w (2.7-fold), CYC7 (2.1-fold) and MDH2 (2.1-fold) and three showed a more than 2.0-fold reduction in expression, i.e. SPS100 (8.9-fold), PAU6 (3.6-fold) and HXK1 (2.0-fold) (Figure 2C). These results indicate that decreased 14-3-3 protein activity did not have a clear effect on the expression of Msn2- and Msn4-regulated genes. On the other hand, it is certainly possible that under the growth conditions used for our experiments, the stress-response signal transduction pathway is not activated, and that the activity of the Msn2 and Msn4 transcription factors is independent from the 14-3-3 proteins.
Our data are consistent with a role of the yeast 14-3-3 proteins in the regulation of the transcription of genes involved in different processes. First, 14-3-3 proteins have an inhibitory effect on the transcription of genes involved in the retrograde response by inhibition of the Rtg3 transcription factor. The bmh2(Ts) mutation has a major positive effect (6.8-fold increase) on the expression of PCK1, involved in gluconeogenesis, and a major negative effect on the expression of genes involved in ergosterol synthesis (up to 4.8-fold decreased expression of HES1). In addition, the expression of many stress-related genes is affected, both positively and negatively. Despite the 14-3-3 proteins binding to the Msn2 and Msn4 transcription factors, the bmh2(Ts) mutation does not have a clear effect on the steady-state mRNA levels of genes regulated by these transcription factors. The mRNA level of a number of genes encoding transporter proteins is decreased. These observations indicate that 14-3-3 proteins regulate transcription at multiple levels. However, the molecular mechanisms of the 14-3-3-protein-dependent regulation of transcription, other than that of the Rtg3-dependent transcription, remain to be established.
We thank M. Hummel for her excellent technical assistance. This study was supported in part by grant BMBF-LPD/8-55 from the Deutsche Akademie der Naturforscher Leopoldina/BMBF.
Abbreviations: Cy3, cyanine 3; Cy5, cyanine 5; ORF, open reading frame; TOR, target of rapamycin
- The Biochemical Society, London