Depletion of yeast PDK1 orthologs triggers a stress-like transcriptional response

Pkh proteins are the PDK1 orthologs in S. cerevisiae. They have redundant and essential activity and are responsible for the phosphorylation of several members of the AGC family of protein kinases. Pkh proteins have been involved in several cellular functions, including cell wall integrity and endocytosis. However the global expression changes caused by their depletion are still unknown. A doxycycline-repressible tetO7 promoter driving the expression of PKH2 in cells carrying deletions of the PKH1 and PKH3 genes allowed us to progressively deplete cells from Pkh proteins when treated with doxycycline. Global gene expression analysis indicate that depletion of Pkh results in the up-regulation of genes involved in the accumulation of glycogen and also of those related to stress responses. Moreover, genes involved in the ion transport were quickly down-regulated when the levels of Pkh decreased. The reduction in the mRNA levels required for protein translation, however, was only observed after longer doxycycline treatment (24 h). We uncovered that Pkh is important for the proper transcriptional response to heat shock, and is mostly required for the effects driven by the transcription factors Hsf1 and Msn2/Msn4, but is not required for down-regulation of the mRNA coding for ribosomal proteins. By using the tetO7 promoter we elucidated for the first time the transcriptomic changes directly or indirectly caused by progressive depletion of Pkh. Furthermore, this system enabled the characterization of the transcriptional response triggered by heat shock in wild-type and Pkh-depleted cells, showing that about 40 % of the observed expression changes were, to some degree, dependent on Pkh.


Background
The 3-phosphoinositide-dependent kinase (PDK1), a master kinase ubiquitously present in eukaryotic life forms, has critical and diverse roles in cells and organisms. In multicellular organisms, PDK1 phosphorylates its substrate PKB (also termed Akt) in response to increased levels of phosphatidylinositol (3,4,5) triphosphate, which triggers co-localization of the kinases mediated by their pleckstrin homology (PH) domains. Deregulation of PKB in mammals leads to important disorders such as cancer, diabetes, cardiovascular and neurological diseases. PDK1 also phosphorylates the activation loop of at least 20 other members of the socalled AGC family of protein kinases, including SGK, p70 S6 kinase, and PKC. However, this activity does not require the direct interaction of PDK1 with phosphoinositides [1][2][3]. This set of AGC kinases plays important cellular roles in all studied organisms. In mammals they are involved in cell growth, proliferation, differentiation and survival, tumor growth, aldosterone and insulin release, glucose metabolism, gastric acid secretion, regulation of ion transporters and channels, blood pressure, ribosome biogenesis, protein synthesis, cell cycle progression and metabolism, among other important functions [4,5]. It has been demonstrated that in plants PDK1 is required for the activation of not less of 16 AGC kinases and, similar to what happens in animals, PDK1 regulates signaling pathways necessary for proper growth in normal and stress conditions [6,7].
The yeast orthologs Pkh1 and Pkh2 proteins have a kinase domain that is 47.7 % identical to that of human PDK1 (73.5 % and 74.6 % of similarity, respectively). Pkh1 and Pkh2 have redundant and essential functions since cells carrying the double pkh1 and pkh2 deletions are not viable but any single mutant is viable [8]. Pkh3 is a third and more distantly related protein whose catalytic domain displays 42.9 % identity with that of PDK1 (65.6 % similar). The single pkh3 deletion does not display an obvious phenotype [9]. Accordingly, and in contrast to what occurs with the pkh1 pkh2 double-mutant cells, the pkh1 pkh3 and pkh2 pkh3 double-mutant strains grow normally. Lack of PKH3, however, exacerbates the growth defect of the pkh1 D398G pkh2 mutant at the restrictive temperature [10].
As in other organisms, yeast Pkh proteins exert pleiotropic effects by phosphorylating the activation loop of diverse AGC protein kinases and by the direct phosphorylation of other regulatory proteins. Among the identified substrates of the yeast Pkh kinases are the protein kinases Ypk1, Pkc1, Sch9 [11] and Tpk1, one of the catalytic subunits of PKA [10,12]. Sch9 phosphorylation by Pkh regulates lifespan and oxidative stress sensitivity in a process that depends on sphingolipids [13,14]. Phosphorylation of Pkc1 activates the Slt2 MAPK cascade and is crucial for maintaining cell wall integrity (CWI) [9,15].
One of the more commonly used methods to identify the functions of a protein is to analyze the phenotypes caused by the lack of this protein. When redundant proteins are present it will be necessary to simultaneously remove all those redundant proteins, usually by deleting the corresponding genes. However, alternative strategies should be used when the elimination of the genes coding for the redundant proteins leads to non-viable cells. In the case of the essential family of Pkh protein kinases most of the information obtained has been acquired using a Pkh1 D398G temperature-sensitive allele in combination with deletion of the PKH2 gene and incubation at the restrictive temperature of 37°C. We have recently reported an alternative genetic strategy to deplete cells of Pkh activity circumventing the incubation at stressful temperatures [15]. In the newly generated strains the expression of PKH2, controlled by the tetO 7 promoter, is decreased by addition of doxycycline. Therefore, elimination of the essential Pkh activity in this system can be achieved by deleting PKH1 and by incubation of cells in the presence of doxycycline. We decided, however, to delete also the PKH3 gene, since it was identified as a multicopy suppressor of the lethality caused when pkh1 D398G pkh2 cells were incubated at the restrictive temperature [9]. Our strategy allows the phenotypic analysis caused by depletion of Pkh with no need of incubation of cells at 37°C, a temperature that triggers the activation of the CWI pathway. This approach has been successfully used to demonstrate the importance of Pkh in the activation of the CWI pathway in the absence of a heat shock [15].
The identification of the global changes in the expression profile caused by the lack of the studied protein is a more comprehensive approach to identify potential new functional roles for that protein. As expected for the case of redundant proteins, the single deletion of the PKH1 or PKH3 genes does not significantly modify the expression profiles of cells cultivated in optimal conditions (our unpublished results and [16,17]). In fact, pkh1 and pkh3 single mutant cells have been included in the set of 784 non-responsive mutants because three or less significant mRNA expression changes were detected as a result of the single deletion of PKH1 or PKH3 [16]. Our aim is to discern the global transcriptional changes triggered by the lack of all of the three yeast Pkh proteins. For this purpose we used the tetO 7 -based system to produce a progressive depletion of Pkh by incubation of pkh1 pkh3 double mutant cells in the presence of doxycycline for 8 and 24 h. We show that the levels of mRNA involved in the glycogen accumulation and in the responses to heat and oxidative stresses, as well as the unfolded protein response, are increased, whereas the levels of those related with ion transport are decreased. In addition, we validated the transcriptional data with experimental information using gene reporters and cell sensitivity to several environmental stresses.
Finally, we determined the need of Pkh for the proper Hsf1 and Msn2/4-driven transcriptional response to heat stress, but not for the down-regulation of genes coding for ribosomal proteins under these conditions.

Depletion of Pkh proteins causes specific transcriptional changes
Previous large-scale studies have identified minor changes in the transcriptional patterns by the single deletion of PKH1 or PKH3 [16,17]. This is not unexpected, reinforcing the notion that Pkh proteins have redundant functions and that the lack of one Pkh can be probably compensated by the presence of any of the other two Pkh proteins. Now, we have constructed the SDP8 yeast strain (Table 1) in which PKH1 and PKH3 genes are deleted and the quantity of Pkh2 can be progressively decreased by incubation in the presence of doxycycline [15]. According to quantitative RT-PCR experiments, incubation of the cells containing the tetO 7 -PKH2 construct for 24 h in the presence of 100 μg/ml doxycycline reduced more than 200-fold the expression of PKH2 (not shown). At this point most of the cells are still viable [15], probably because enough amounts of Pkh2 protein could still be present in the cell (the estimated half-life of Pkh2 is longer than 20 h). In order to gain insight into the cellular roles of the yeast PDK1 kinases, we decided to use this cellular system to identify the transcriptomic changes triggered by progressive depletion of Pkh proteins. To this end, we compared the expression pattern of the strain SDP8 with that of parental wild-type CML476 cells, grown both in the presence of 100 μg/ml doxycycline for 8 or 24 h. The expression patterns of antibiotic-treated versus untreated SDP8 cells were not compared because, in contrast to the previously published data [18], we have detected a small but significant number of transcriptional changes induced by doxycycline (our unpublished results).
Our data indicated that incubation of SDP8 cells with doxycycline for 8 h changed the expression of 113 genes (3.1 % of the genes with valid data). Eighty-one were found up-regulated and 32 down-regulated ( Fig. 1a and Additional file 1: Table S1). A longer incubation time in the presence of the antibiotic (24 h) increased the number of expression changes, affecting 9.7 % of the analyzed genes (228 and 140 up-and down-regulated, respectively). The expression patterns for the total number of genes whose expression was considered changed at any time-point are shown in the Fig. 1b. As hypothesized, deletion of PKH1 and PKH3 and progressive depletion of Pkh2 had more important transcriptional effects that deletion of any single PKH gene.
Depletion of Pkh alters mRNA levels of genes required for glucose storage, ion transport and ribosomal proteins Functional analysis of the set of 81 genes found upregulated after 8 h of incubation of SDP8 cells with doxycycline revealed an excess of genes involved in the energetic metabolism (19 genes; p-value: 1.09e-05) and in the metabolism of energy reserves (7 genes; p-value: 7.77e-05) (Fig. 2a). The mRNAs corresponding to genes coding for the high affinity glucose and maltose transporters (HXT2, HXT4, HXT7 and MAL31) typically induced by glucose limitation, were also found among the most abundant (Additional file 1: Table S1). Incubation of SDP8 cells in the presence of doxycycline for 24 h resulted in the up-regulation of genes involved in carbohydrate metabolism (72 genes; p-value: 5.09e-16), energy (54 genes; p-value: 4.52e-14), fermentation (15 genes; p-value: 9.41e-10) and stress response (59 genes; p-value: 1.09e-08) among others (Fig. 2b).
Most of the genes involved in glycogen biosynthesis have been found up-regulated when SDP8 cells were incubated in the presence of doxycycline for 8 h (Fig. 3). GAC1, which encodes for the regulatory subunit of Glc7, that tethers Glc7 to the glycogen synthase [19], is among these genes (Fig. 3). Expression of genes involved in the glycogen metabolism normally increases when cells approach the stationary phase in a PKA-dependent manner, involving the transcription factors Msn2 and Msn4 [20]. Glycogen accumulation also plays an important role in response to several stresses [21]. The mRNA levels of TPS1 and TPS2, required for the synthesis of trehalose, were also found increased (Fig. 3). Trehalose is a reserve carbohydrate which is proposed to function as stress protectant that has been involved in stress responses by stabilizing proteins and membranes. Although the cellular levels of trehalose and glycogen are not directly related to the expression levels of the trehalose and glycogen synthase genes, high expression of those genes have been reported in cells under several stressful conditions including, but no limited to, heat shock, oxidative stress, alkaline pH and high osmolarity, [22][23][24].
Analysis of the 32 down-regulated genes in SDP8 cells treated with doxycycline for 8 h (Additional file 2: Table  S2) revealed that the category of iron transport (such as SIT1, ARN1, ARN2, FRE1, FIT2, FIT3 and PHO84 among others; p-value: 7.94e-05) were over-represented (Fig. 2b). This category had similar relevance when SDP8 cells were treated for 24 h with the antibiotic (p-value: 3.77e-05). These expression patterns recall the transcriptional changes described under iron-surplus conditions, controlled by the Aft1and Aft2 transcription factors [25]. In fact, there is a good correlation (R 2 = 0.68) between the expression changes of the genes down-regulated by incubation of SDP8 cells with doxycycline for 24 h and the set of Aft1-dependent iron transport genes [26] (Fig. 4a).
The expression of genes coding for the high affinity inorganic phosphate transport system (PHO84), and the secreted acid phosphatases (PHO5, PHO11, PHO12 and PHO3) that mediate the response to phosphate acquisition, were also found down-regulated at both, 8 and 24 h of SDP8 cells treatment with doxycycline (Additional file 2: Table S2). We confirmed the SDP8 MATa CML476 KanMX4-(tetO 7 ):PKH2 pkh1::HIS3 pkh3::nat1 - [ 15] decreased expression levels of the high-affinity inorganic phosphate transporter gene PHO84 by using the LacZ reporter gene fused to its promoter when SDP8 cells were incubated with doxycycline for 24 h (Fig. 4b).
When the set of 140 genes down-regulated by incubation of SDP8 cells with doxycycline for 24 h were considered, it is noticeable that an excess of genes involved in the translation process was manifested (36 genes; p-value: 1.26e-19; Fig. 2b and Additional file 2: Table S2). Among them, the sub-category of genes coding for ribosomal proteins was the more relevant (31 genes; p-value: 2.69e-15). Genes coding for ribosomal proteins were only moderately down-regulated as a consequence of the longterm depletion of Pkh, as shown in Fig. 4c, and their RNA levels were not affected when SDP8 cells were treated with doxycycline for 8 h.
Our results provide evidence for the existence of a set of genes whose mRNA levels changed just 8 h after starting the depletion of Pkh. Expression of other genes was not altered at this time-point and only changed after prolonged incubation with doxycycline, which could be an indirect consequence of early cellular events caused by depletion of Pkh.

Progressive depletion of Pkh increases the transcription of genes required for fermentation and stress responses
Depletion of Pkh led to progressive accumulation of mRNA of those genes involved in carbohydrate metabolism. Among the 81 and 228 genes found up-regulated (See figure on previous page.) Fig. 1 Gene expression changes triggered by gradual depletion of Pkh. a Number of genes found up-(empty bars) and down-regulated (filled bars) in SDP8 versus WT cells after 8 and 24 h of treatment with 100 μg/ml doxycycline. The percentage respect the total number of genes with valid expression value data are denoted in brackets. b Cluster analysis of the expression profiles caused by depletion of Pkh. Genes whose expression was found changed after treatment of SDP8 cells with doxycycline (either 8 or 24 h) when compared to wild-type cells under the same conditions were hierarchically clustered (complete linkage clustering, uncentered correlation) by means of the Gene Cluster (v. 2.11) software [72] and visualized with Java TreeView (v 3.0) [73]. Relevant functional categories of genes in each cluster are denoted  Fig. 2a), genes involved in the fermentation sub-category were preferentially up-regulated after longterm treatment of SDP8 cells. In fact, only two genes of this sub-category were found up-regulated after 8 h of doxycycline treatment (ALD4 and BDH2; p-value: 1.68e-01) in contrast to 15 genes found after treatment for 24 h, mainly involved in redox processes (AAD3, AAD6, AAD14, AAD15, AAD16, ACS1, ALD2, ALD3, ALD4, ALD6, ARO10, BDH2, DLD1, NDE2 and NGL3; p-value: 9.41e-10).
In SDP8 cells incubated for 8 h with doxycycline we have found increased the mRNA levels of a set of 13 genes involved in the stress response but not specifically implicated in any particular sub-category of stress (AGP2, CYC7, FMP43, GAC1, GAD1, GCY1, HSP26, MGA1, MSN4, STF2, TMA10, PRB1 and YFL054C; p-value: 1.96e-01). Much more relevant, however, was the set of genes up-regulated when the same cells were incubated for 24 h with the antibiotic (59 genes; p-value: 1.09e-08) (Fig. 2a). Among these genes the more significant sub-categories represented, shown in Table 1, are: i) the unfolded protein response (14 genes; p-value: 4.02e-05) containing genes coding for proteins with chaperone/co-chaperon activities (APJ1, CPR6, HSP26, HSP33, HSP42, HSP78, HSP82, MDJ1 among other; ii) oxidative stress response (11 genes; pvalue: 4.58e-04) including CTA1, coding for catalase A, the cytoplasmic peroxiredoxin TSA1, the mitochondrial thioredoxin peroxidases PRX1, the mitochondrial  [74]. Numbers indicate the -fold induction of the specified genes in SDP8 when incubated in the presence of 100 μg/ml doxycycline for 8 and 24 h when compared to CML476 wild-type cells under the same treatment. Genes considered up-regulated are denoted in red. ND: no valid data were obtained superoxide dismutase SOD2, the genes involved in the 4-aminobutyrate metabolism GAD1 and UGA2 and iii) heat-shock response (8 genes; p-value: 9.80e-04). Changes in expression for several of the genes detected in our microarrays analysis were also verified by semiquantitative RT-PCR (Additional file 3: Figure S1).
To validate the significance of the described transcriptional changes, a publication enrichment analysis with the set of 228 genes found up-regulated after treating SDP8 cells with doxycycline for 24 h was performed. The results indicated that 45 of these genes have been previously found up-regulated in response to mild heat shock (p-value: 6.13e-23) [27] among other stress conditions that also induced comparable stress response [28]. However, the most similar expression pattern found was that triggered by treatment of wildtype cells with 0.6 M NaCl for 45 min [29]. From the set of 136 genes induced after the salt stress, 54 genes (mostly involved in oxidative stress response and metabolism of carbohydrates such as the pentosephosphate pathway), were also up-regulated in SDP8 cells treated with doxycycline for 24 h (Additional file 1: Table S1).

Depletion of Pkh mimics a transcriptional response triggered by environmental stressors
Our data suggest that depletion of Pkh triggers, directly or indirectly, a common yeast response to different stresses. Examination of the transcriptional response to many types of environmental stresses allowed the identification of a gene expression program known as environmental stress response (ESR) [22]. A total number of 867 ESR genes were reported, being the expression of 585 and 282 genes, respectively, up-or down-regulated by environmental stresses [22]. From these subsets, 86 and 56 have been found to be also up-and downregulated by depletion of Pkh (Additional file 1: Table S1 and Additional file 2: Table S2), which represent an enrichment of 5-and 2.5-fold over the expected number in case of independent events. Thus, the transcriptional response triggered by depletion of Pkh correlated to that of the ESR program.
Several of the ESR genes have paralogs in the yeast genome whose expression does not respond to ESR [30]. In order to know if the up-regulated genes correspond or not to the ESR genes we calculated expression changes for the set of ESR and non-ESR paralogs. Our analysis indicated that the averages of the expression changes (in log 2 ) observed for the ESR genes were 0.72 and 1.01 for the cells treated with doxycycline for 8 and 24 h, respectively. By contrast, the averages for the non-ESR paralogs were −0.05 and 0.03 respectively. This indicates that Pkh depletion causes the specific up-regulation of ESR genes but, in most cases, it does not increase the expression of their Fig. 4 Depletion of Pkh causes down-regulation of genes involved in ion transport. a Aft1 dependence of S. cerevisiae iron transport genes and their down-regulation by depletion of Pkh. Comparison of the transcriptional changes, in log 2 , triggered by treatment of SDP8 cells with 100 μg/ml doxycycline for 24 h to CML476 when compared to wild-type cells under the same treatment (x axis) with data obtained when the expression pattern of aft1 mutant cells was compared to that of a wild-type isogenic strain carrying the constitutively activated AFT1-1 up allele [26]. b β-Galactosidase activity of cells containing the PHO84-LacZ reporter gene. SDP8 cells transformed with the plasmid PHO84-LacZ and liquid cultures of the resulting strain were treated with doxycycline (100 μg/ml) or the vehicle for 8 or 24 h, as specified in the Methods section. β-Galactosidase activity was measured in non-treated (empty bars) or treated (filled bars) cells. Data are mean ± SEM from three independent clones. Treated cells were compared with untreated cells at each time point using un-paired t-test (n = 3; one-tailed P value; *P < 0.05). c The 140 genes downregulated by incubation of SDP8 cells with doxycycline for 24 h have been ordered according to their expression values when compared to the expression in wild-type cells under the same conditions. Genes coding for ribosomal proteins are denoted by empty squares non-ESR paralogs (Additional file 4: Table S3). For example, typical ESR genes involved in the metabolisms of carbohydrates such as the hexokinase HXK1, the glucokinase GLK1 or the phosphoglucomutase PGM2, were upregulated when Pkh was depleted but the expression of their non-ESR paralogs HXK2, EMI2 and PGM1 was not increased. These results suggest that, directly or indirectly, progressive depletion of Pkh preferently triggers the expression of ESR genes.
Several transcription factors are involved in the transcriptional changes caused by specific stress conditions, being Msn2 and Msn4 major players in this response. These transcription factors recognize and bind the Stress Response Element (STRE), sequences found in most of the promoters of the stress-responsive genes including glycogen and trehalose synthesis [30,31]. To test if the transcription factors Msn2/Msn4 are activated during the stress-related transcriptional response found in Pkhdepleted cells we used the STRE sequence fused to the lacZ reporter gene. As shown in Fig. 5a, activity of the β-galactosidase gene product driven by the STRE was slightly higher after depletion of Pkh. From these results we can conclude that depletion of Pkh triggers a stresslike transcriptional response that is, at least in part, mediated by the Msn2/Msn4 transcription factors.
Lethality caused by depletion of Pkh could be partially suppressed by overexpression of YAP1 Pkh-depleted cells transcriptionally respond to stress. It is reasonable to think that the observed response to oxidative stress could be the consequence of a situation of cellular oxidative stress directly or indirectly caused by depletion of Pkh. We next tested the tolerance to oxidative stress of SDP8 cells incubated in the presence of low (non-lethal) doses of doxycycline. Our results indicate that partial depletion of Pkh made cells hypersensitive to the oxidative agent hydrogen peroxide and to the thiol oxidant diamide, but cells were less affected by the superoxide generator menadione (Fig. 5b). These results could be explained assuming that Pkh-deficient cells suffer of oxidative stress, which can explain the increased levels of reactive oxygen species (ROS) previously observed in these cells [15]. To test if the high levels of ROS could be the cause of Pkh-deficient cells lethality, we overexpressed the transcription factor Yap1, which is specifically required for oxidative stress tolerance [32] in MB005 cells ( Table 1). As observed in Fig. 5c, expression of YAP1 from a centromeric plasmid partially rescued the lethality provoked by the combined deletion of PKH1 and depletion of Pkh2 induced by incubation of MB005 cells with 50 μg/ml of doxycycline.
This result shows that the lethality driven by depletion of Pkh is, at least in part, due to the cellular oxidative stress since it can be partially alleviated by overexpression of YAP1. The transcriptional changes observed in Pkh-depleted cells dedicated to respond to the oxidative stress are probably insufficient to cope with stress and cells become hypersensitive to environmental oxidative stresses.

Pkh-depleted cells are hypersensitive to UPR-inducers and toxic ions
The accumulation of unfolded or misfolded proteins in the endoplasmic reticulum triggers the conserved unfolded protein response (UPR). The objective of the UPR is to reduce the translation and, simultaneously, to initiate a signaling pathway that leads to activation of Hac1. Hac1 is a key transcription factor that increases the transcription of genes involved in proper protein folding (see [33] for a recent review). Depletion of Pkh1 caused the up-regulation of a significant number of genes involved in the UPR ( Table 2). Comparison of the full expression pattern of cells treated with the tunicamycin [34], an inhibitor of the protein glycosylation that triggers the UPR, with that of Pkh-depleted cells evidences that the similarities also extended to the down-regulated genes. In fact, the number of genes up-or down-regulated by treatment with tunicamycin and by depletion of Pkh, was higher than the expected in case of independent events (2.1-and 7.2-fold, respectively; Additional file 1: Table S1 and Additional file 2: Table S2).
To further study the possible connection between depletion of Pkh and the UPR we tested the sensitivity to tunicamycin of cells containing low levels of Pkh. Incubation of SDP8 or MB005 cells in the presence of low doses of doxycycline greatly increased the hypersensitivity to tunicamycin when compared to wild-type cells (Fig. 6a). Pkh-deficient cells were also hypersensitive to β-mercaptoethanol, another agent that activates the UPR, although in a less-specific manner (Fig. 6a). These results indicate that partial depletion of Pkh somehow mimics an ER-stress-induced transcriptional response that could be caused by the progressive accumulation of unfolded proteins or by interfering with the UPR signaling pathway, among other options. Partial depletion of Pkh, however, did not change the basal transcriptional activity driven by the UPRE, although it did increase the cell duplication time (Fig. 6b). Furthermore, decreased levels of Pkh, did not interfere with the normal response triggered by tunicamycin (Fig. 6b).
The transcriptional response of Pkh-depleted cells was also to some extent similar to that induced by saline stresses [29]. We then investigated the effect of low levels of Pkh on salt tolerance. As shown in the Fig. 6c, SDP8 and MB005 cells treated with low doses of doxycycline were hypersensitive to sodium and lithium toxic ions, suggesting that cells expressing low doses of Pkh could be already challenged by an ionic stress, and cannot handle additional environmental ionic stress.

Progressive depletion of Pkh affects the expression of specific transcription factors
From the present data we concluded that depletion of Pkh triggers a transcriptional response addressed to cope with several types of stress. In order to know which transcription factors are mainly affected by the lack of Pkh, the sets of 82 and 228 genes whose expression was increased by incubation of SDP8 cells with doxycycline for 8 and 24 h, respectively were first analyzed. Our results showed that the transcription factors involved in (See figure on previous page.) Fig. 5 Cells lacking Pkh are hypersensitive to oxidative stress. a β-Galactosidase activity of cells containing the STRE-LacZ reporter gene. SDP8 were transformed with the pGM18/17 plasmid, which was previously linearized at the URA3 gene marker by digestion with NotI, and treated as described for the Fig. 4a. β-Galactosidase activity was measured in non-treated (empty bars) or treated (filled bars) cells. Data are mean ± SEM from three independent clones. Treated cells were compared with untreated cells at each time point using un-paired t-test (n = 3; one-tailed P value; *P < 0.05). b Effect of partial depletion of Pkh on the sensitivity to oxidative agents. Wild type (WT) CML476 cells and the indicated mutants were grown, and four different (1:5) dilutions of the cultures containing the same number of cells were spotted on YPD plates containing 10 μg/ ml doxycycline and the indicated concentrations of H 2 O 2 , menadione and diamide. Growth was monitored after 3 days. c Overexpression of YAP1 partially rescues the lethal effect of depletion of Pkh. MB002 and MB005 cells were transformed with the empty centromeric pRS316 or the same plasmid containing the promoter and coding region of YAP1. Serial dilutions of liquid cultures were spotted on YPD plates containing 50 μg/ml doxycycline (DOX) or the vehicle. Cell growth was recorded after 3 days stress responses, such as Msn2, Msn4, Adr1, Hsf1, etc., were the most affected (Additional file 5: Figure S2).

Pkh-depleted cells displayed an attenuated transcriptional response to heat shock
We have shown that depletion of Pkh triggers a heat stress-like transcriptional response. It has been described that Pkh activity is regulated by sphingolipids and these lipids play an important role in the cellular response to a series of stresses, including heat stress [35]. For this reason we next characterized the transcriptional responses of wild-type and Pkh-depleted cells to heat stress. Incubation of doxycycline-treated CML476 wild-type cells at 40°C for 40 min caused the up-regulation of 11.6 % of genes with valid data, mainly those involved in the stress response and protein folding and stabilization. We also observed that 6.0 % of the genes, mostly coding for ribosomal proteins, were down-regulated. When doxycyclinetreated SDP8 cells were heat stressed in the same way, only 9.5 % and 4.1 % were found up-and down-regulated, respectively, suggesting the attenuation of the transcriptional response (Additional file 6: Figure S3), although the functional categories were the same than those of stressed wild-type cells. Similar results were obtained when only genes with valid data for the two strains were considered ( Fig. 7a and Additional file 7: Figure S4), indicating a general attenuation of the transcriptional response to heat stress in Pkh-depleted cells. We next identified the genes whose expression changes under heat stress were dependent on the presence of Pkh. We have found that the expression changes of 271 genes (41.2 %) were somehow dependent on the presence of Pkh: 193 upregulated (44.6 %, Additional file 8: Table S4) and 78 were down-regulated (34.7 %, Additional file 9: Table S5). It is worth mentioning that genes involved in the response to stress and in protein folding and stabilization were up-regulated preferably in a Pkh-dependent manner (Additional file 8: Table S4). Among the genes down-regulated in a Pkh-dependent manner, those involved in the metabolism of the aspartate family of amino acids were found overrepresented. The regulation of most of the genes encoding ribosomal proteins was essentially independent on Pkh ( Fig. 7b and Additional file 10: Table S6). We have also found a set of genes whose expression levels under heat shock changed only in Pkh-depleted but not in wild-type cells. For example, an excess of the seripauperin multigene family (PAU genes) encoded mainly by subtelomeric regions of the genome were found up-regulated in response to heat shock only in doxycycline-treated SDP8 cells (Fig. 7b  and c). It has been suggested that this family of proteins might possess specific roles in the cell adaptation to certain environmental stresses [36]. Similarly, the mRNA levels of a set of genes involved in the maturation of rRNA were down-regulated after heat shock in SDP8 but not in wild-type cells (Fig. 7b). These results indicate that there is a relevant set of genes whose expression changes under heat-shock conditions depends on the presence of Pkh.

The full transcriptional response to heat shock requires Pkh proteins
The above data indicate that depletion of Pkh influences, directly or indirectly, the transcriptional response elicited by heat stress, affecting genes involved in the stress response and the metabolism of energy reserves and amino acids. It is plausible to think that the already elevated levels of mRNA involved in stress responses found in Pkh-depleted cells could prevent further up-regulation of this set of genes in response to heat shock. In fact, we have found a high degree of correlation when the set of genes up-regulated by the depletion of Pkh were compared to those up-regulated by heat shock in wild-type cells (slope: 0.735). This argument, however, is not valid for the set of genes down-regulated by depletion of Pkh1 since there is no correlation at all with those downregulated by heat shock (slope: −0.163). To test the abovementioned hypothesis we analyzed the set of 193 genes whose up-regulated expression in wild-type cells by heat shock was classified as dependent on Pkh. The mRNA quantity of 52 of these genes was found already increased in Pkh-depleted cells, when SDP8 cells were incubated with doxycycline for 24 h (Additional file 6: Figure S3A). The low induction levels of this set of genes under heat shock conditions could be then explained by their high expression levels in Pkh-depleted cells under basal conditions. However, the expression levels of 91 genes, also up-regulated by heat shock in a Pkh-dependent manner, were not found increased under basal conditions in Pkh-depleted cells (Additional file 6: Figure S3B). We can consider that the lack of induction of this set of genes by heat shock in Pkh-depleted cells could be a consequence of Pkh depletion. Although genes involved in protein folding and stabilization and stress response are present in both subsets of genes, its relevance is higher in the subset of already up-regulated genes in Pkh-depleted cells. By contrast, we found higher proportion of genes involved in detoxification in the subset that were not found up-regulated by Pkh depletion (7 genes: CUP1-1, BSD2, GRX6, GRX4, TPO1, TPO3 and SGE1; p-value: 8.09e-03) than in the subset of genes found up-regulated by the depletion of Pkh (2 genes: RTA1 and TPO4; p-value: 4.06e-01). From these analyses we can conclude that there is a set of genes whose up-regulation under heat shock stress is, directly or indirectly, dependent on the presence of normal levels of Pkh.
Therefore, the lower up-regulation detected for this set of genes in SDP8 cells subjected to heat stress is not due to the already elevated levels resulting from Pkh depletion in non-stressed conditions. On the contrary, the attenuated transcriptional response after heat shock of this set of genes could be considered the consequence of Pkh deficiency.

Discussion
In this work we have contributed to the identification of new cellular roles of Pkh by describing the transcriptional patterns triggered by the absence of the Pkh proteins under various conditions. Pkh proteins provide an essential function in diverse yeast species by regulating an increasing number of cellular functions. Most of the studies on Pkh have been performed using a temperaturesensitive allele [9]. However, this strain does not allow the identification of the transcriptional response of Pkhdeficient cells in optimal growth conditions since a shift to a stressful temperature is required to reduce the cellular Pkh activity. We have previously described that Pkhdeficient cells are hypersensitive to heat stress [15]. For this reason we have developed a different approach based on a doxycycline-regulated promoter that allowed us to identify for the first time the transcriptional changes The set of genes up-regulated by incubation of SDP8 cells with 100 μg/ml doxycycline for 8 or 24 h are ranked by transcription factor, according to the Rank by TF application of the Yeastract platform and using the DNA binding plus expression evidence and TF acting as activator. The number of genes whose expression is controlled by each transcription factor is shown. A score for each transcription factor is given by a p-value that denotes the overrepresentation of the given transcription factor targeted genes in the list of up-regulated genes relative to the regulations of that transcription factor targeted genes in the whole Yeastract database (See figure on previous page.) Fig. 6 Cells lacking Pkh are sensitive to UPR inducers and ionic stress. a Effect of partial depletion of Pkh on the sensitivity to the UPR inducers tunicamycin and β-mercaptoethanol. Cultures of CML476 cells (WT) and its specified derivatives were serially diluted and spotted onto YPD plates containing 10 μg/ml doxycycline (DOX), or the vehicle, and the indicated concentrations of tunicamycin (Tn) or β-mercaptoethanol (β-ME). Control plate in the absence of doxycycline of cells incubated with β-ME is shown in the Fig. 5b. Cell growth was recorded after 3 days. b Left panel. β-Galactosidase activity of SDP8 cells containing the UPRE-LacZ reporter gene. SDP8 were transformed with the pMCZ-Y plasmid and incubated in medium without uracil in the presence of 10 μg/ml of doxycycline ("SDP8-Dox") or the vehicle ("SDP8 + Dox") for a total time of 24 h. Six h before the treatment with tunicamycin, new cells cultures at OD 660~0 .05 were prepared containing fresh doxycycline. Finally, cells were resuspended in medium containing fresh doxycycline (or the vehicle) and 2.5 μg/ml tunicamycin (or DMSO). β-Galactosidase activity was measured for cells non-treated (empty bars) or treated with tunicamycin (filled bars). Data are mean ± SEM from six independent clones. Right panel: Representation of the cell density of the cultures used in the left panel obtained by measuring their absorbance at 650 nm in a microplate reader, after the treatment of SDP8 cells with tunicamycin (filled bars) or with the vehicle (empty bars). Data are mean ± SEM from six independent clones. c Effect of partial depletion of Pkh on the salts sensitivity. Four dilutions (1:5) of cells from the strains specified in a were spotted onto YPD plates containing 10 μg/ml doxycycline (DOX) and the indicated concentrations of NaCl or LiCl. Control plate for the absence of doxycycline is shown in the Fig. 5b. Cell growth was recorded after 3 days triggered by the progressive depletion of Pkh proteins in the yeast S. cerevisiae. It is worth mentioning that depletion of Pkh proteins by this strategy does not imply a stressful growth condition for the cells and that, in any case, we have compared the expression profiles of two strains cultivated in the presence of doxycycline. Furthermore, the use of the doxycycline-regulated promoter allowed us to identify the transcriptional response of Pkhdeficient cells to heat stress, an experiment that cannot be properly done by using a strain carrying a temperaturesensitive allele. Until now, transcriptional changes caused only by the single deletion of PKH1 or PKH3 have been identified in large-scale studies [16,17]. Using the published data, and according to our criteria, lack of PKH1 triggers the upregulation of two contiguous genes located in the chromosome IV, BSC1 (YDL037C) and PRM7 (YDL039C). PHO84 was the only down-regulated gene in this mutant strain. In the same study lack of PKH3 up-regulates the same genes described for the pkh1 mutant cells plus YDL196W, and down-regulates PHO84 and also SPL2. It is worth to note that the mRNA levels of genes like FIT3, SPL2 and FIT2 were found reduced in pkh1 mutant cells (0.50, 0.57 and 0.60-fold, respectively) [16]. In another study lack of Pkh1 or Pkh3 had not relevant consequences on the expression profiles [17]. Our study demonstrates that expression changes caused by the double deletion of PKH1 and PKH3 and depletion of Pkh2 are reminiscent of those triggered in response to environmental stresses [37]. We show that the expression of genes involved in the metabolism of glycogen and trehalose as well as in stress responses were already elevated in the first time-point assessed, and increased moderately or did not increase along the rest of the experiment. This early transcriptional response, is probably a direct consequence of the decreased levels of Pkh and is comparable to that described for several environmental stresses, where the expression levels for genes involved in glycogen metabolism increases soon immediately after the stress and then the mRNA levels for these genes progressively decrease [38,39]. Upregulation of the genes required for high-affinity glucose transport and phosphorylation as well as for glycogen accumulation have also been found in heat-stressed cells [30]. Genes involved in the defense against oxidative stress and accumulation of unfolded protein increased their expression along the experiment, suggesting that depletion of Pkh could lead to cellular stress or, at least, trigger a response inducible by stress. In this agreement, we also observed a decrease in the quantity of mRNA coding for ribosomal proteins and other mRNA related to protein synthesis. Oxidative stress and stress driven by the accumulation of improperly folded or unfolded proteins are coupled since the former disrupts the proper disulfide formation thus inducing the UPR [40]. Then, the observed transcriptional response could be primarily driven by the oxidative stress. Oxidative stress can be triggered by several factors, including a malfunction of the pentose phosphate pathway, which function to regenerate NADPH for recycling the oxidized glutathione [41]. According to this, deletion of the gene coding for the glucose 6-phosphate dehydrogenase (ZWF1), enzyme of the first reaction of the pentose phosphate pathway, renders respiring cells with high levels of ROS [42]. We observed that lower than normal levels of Pkh does not affect neither the basal levels of transcription driven by the UPRE nor the activation of this pathway by UPR-inducers. Hypersensitivity to UPR-stressors of cells with low levels of Pkh could be explained by other reasons. Low Pkh levels might, for example, affect the recovery of the UPR after removal of the stress. A delay in the recovery of the UPR is the cause of the hypersensitivity to UPR stressors displayed by reg1 mutant cells [43].
We also detected the short-term down-regulation of genes required for phosphate and iron uptake. It is known that members of the regulon PHO are downregulated as soon as 15 min after the beginning of a heat stress [44]. Expression of the genes coding for the highaffinity phosphate transporters PHO84 and PHO89, and the secreted acid phosphatases PHO11, PHO12 and PHO3 have been found down-regulated when cells were shifted from 17°C or 21°C to 37°C (GEO Samples GSM883 and GSM930) [22]. This fact is compatible with the notion that Pkh-depleted cells suffer a response similar to that triggered by heat shock. Low levels of mRNA required for the iron uptake have also been found by exposure of yeast cells to arsenite, an agent that affects the transcription of genes involved in oxidative stress defense and redox maintenance [45]. However, we cannot discard that the down-(See figure on previous page.) Fig. 7 Transcriptional response of Pkh-depleted cells to heat shock. a Venn diagrams of the number of genes with considered up-regulated (upper panel) and down-regulated (lower panel) mRNA expression under heat sock stress in the wild-type CML476 (WT) and SDP8 strains. b The expression patterns of the wild-type CML476 (WT) and SDP8 cells treated with doxycycline for 24 h and incubated at 40°C for 40 min were compared with those of the same cells in the absence of the heat stress. A set of 803 genes with data in both strains that are differentially expressed at least in one of the strains were hierarchically clustered (complete linkage clustering, uncentered correlation) using the Gene Cluster software (v. 2.11) and visualized with Java TreeView (v 3.0). The relevant functional categories of genes in some clusters are denoted. c The mRNA levels (in -fold change) determined by DNA microarrays for the PAU family of genes after heat shock are represented for the CML476 (WT, empty circles) and SDP8 (filled circles) strains regulation of genes involved in iron uptake could be related to the recently identified connection between high levels of iron and cell death by ferroptosis, in a process that requires the Pkh1-Ypk1 signaling cascade [46].
Cells containing reduced quantities of Pkh were additionally demonstrated hypersensitive to several environmental stresses such as oxidative agents. Overexpression of the transcription factor Yap1 partially rescued the lethality displayed by Pkh-deficient cells, suggesting that these cells suffer oxidative stress, and the observed transcriptional response cannot cope with it. These results are in agreement with i) the high levels of ROS found in Pkh-deficient cells [15] and ii) the hypersensitivity of Pkh-depleted cells to the oxidizing agents diamide and hydrogen peroxide. The fact that depletion of Pkh did not affect the growth in the presence of menadione is not completely surprising since cells have diverse mechanisms to protect against different oxidizing agents and none of the oxidants is representative of general oxidative stress. For example, it has been described that the single mutation of 78 genes confers sensitivity to diamide and hydrogen peroxide but not to menadione [47]. Cells containing low levels of Pkh are also hypersensitive to other stresses tested such as lithium and sodium ions and to UPR-inducers.
Interestingly, expression of INO1 increased about 95fold in SDP8 cells incubated with doxycycline for 24 h. Semiquantitative RT-PCR analysis confirmed a strong expression induction (Additional file 3: Figure S1). Ino1 is required for the synthesis of phosphoinositides and inositol phosphates, playing central roles in membrane traffic and signal transduction pathways. Although mammalian GST-PDK1 binds in vitro to several phosphoinositides, we have found that neither yeast GST-Pkh1 or GST-Pkh2 bind to any phosphoinositide under the same conditions (unpublished results). The impaired endocytosis observed in Pkh-deficient cells [48] and the role of phosphoinositides in membrane trafficking [49] could explain the high expression level of INO1 found in Pkh-depleted cells.
Transcriptional changes observed by incubation of SDP8 cells with doxycycline for 24 h could also be due to indirect effects caused by the decreased levels of Pkh. A plausible explanation for the transcription pattern observed in Pkh-depleted cells is based in the fact that yeast PKA inhibits Msn2/Msn4 and Hsf1 transcription factors [32,50]. Since it has been recently shown that Pkh is necessary for the activation of PKA [10,12], cells depleted of Pkh would have lower-than-normal levels of PKA activity. Low PKA activity would lead to activation of the Msn2/Msn4 and Hsf1 transcription factors, increasing thus, the expression of the genes required for stress defense.
Pkh proteins were also found necessary for the complete transcriptional response triggered by heat stress. We have identified a set of genes whose up-regulation after stress depends on the presence of normal levels of Pkh. Expression levels of a sub-set of these genes were already increased in Pkh-depleted cells, and this could avoid further elevation of their expression in Pkh-depleted cells under heat shock conditions. However, we have found a relevant set of genes with impaired up-regulation in Pkh-deficient cells that could be considered as directly or indirectly affected by the lack of Pkh after heat stress (their expression is not increased in Pkh-depleted cells). We conclude thus that Pkh proteins influence, directly or indirectly, the transcriptional response elicited by heat stress. This is not completely unexpected, since Pkh can be associated with heat shock adaptation in different ways. First, the kinase activity of Pkh was reported to increase under heat shock stress [51]. Second, Pil1 and Lsp1, two components of eisosomes important for heat stress tolerance [52], are phosphorylated by Pkh under heat stress [51]. Third, the PH domain-containing proteins Slm1 and Slm2, regulators of actin cytoskeleton organization in response to stress, are also substrates of Pkh activity under heat shock conditions [53]. Fourth, the phosphorylation status of the Orm1 and 2 proteins, substrates for the Pkh-Ypk cascade, regulate the sphingolipid biosynthesis in response to heat stress [54]. In addition, Pkh-depleted cells are hypersensitive to incubation at 37°C [15]. All these data indicate that the use of the temperature sensitive allele of Pkh might not be appropriate for some studies since Pkh proteins seems to be required for the proper heat stress response. Our work shows that the depletion of Pkh affects the transcription of numerous genes involved in a broad spectrum of cellular activities. Although the methodology of depletion of Pkh employed here cannot clearly distinguish direct from indirect effects, the study reveals the overall effects of Pkh depletion which is relevant for understanding the broad functions of these kinases. Since Candida, Aspergilus and other fungal infectious organisms have conserved the Pkh signaling pathway, inhibition of Pkh will also produce pleiotropic effects. In Candida sp there is only one gene, CaPKH2, representing S. cerevisiae PKH1/2. We have determined, in our previous work, that there are differences at the ATPbinding site and at the PIF-pocket regulatory sites that can enable the development of drugs that inhibit CaPkh2 without affecting the human ortholog PDK1 [15]. Together, our results point to fungal Pkh orthologs for selective anti-fungal drug discovery.

Conclusion
Our studies show that depletion of Pkh affects the transcription of more than four hundred genes, increasing the mRNA levels of genes involved in the response to heat shock and oxidative stress. A relevant fraction of these up-regulated genes are known to be transcribed by the Msn2, Msn4 and Hsf1 transcription factors, suggesting that the absence of Pkh mimics a cellular response to a diversity of stresses. Our system allows, for the first time, the characterization of the transcriptomic changes caused by heat stress in Pkh-deficient cells, since the use of a thermosensitive allele did not allow such study. Thus, we demonstrate that Pkh is required for the proper transcriptional response after heat stress, although the down-regulation of genes coding for ribosomal proteins is independent of Pkh.

Yeast strains and culture conditions
Yeast strains used in this study ( Table 1) were derivatives of the wild type CML476 [55] and have been already characterized in a previous study [15]. Yeast cells were grown at 28°C in solid or liquid YPD medium (10 g/l yeast extract, 20 g/l peptone and 20 g/l glucose). Sensitivity of yeast cells to several agents (oxidizing, salts and UPR-inducers) in solid media was evaluated by drop test growth on YPD plates as previously described [56]. Saturated liquid cultures grown in YPD medium were initially diluted until OD 660 0.05, and serial dilution were prepared as specified in each case. Three μl of each dilution were spotted onto the plates containing the indicated concentrations of the agents plus doxycycline and growth was registered after incubation for the indicated time at 28°C.

Plasmids used
The PHO84-lacZ reporter is the YEp357 plasmid containing the region comprising from −603 to +19 from the PHO84 gene fused to LacZ [57]. The STRE(7x)-lacZ reporter system was prepared as previously described [58]. The pMCZ-Y plasmid contains the UPRE fused to lacZ gene [59]. To express Yap1 the promoter and coding regions of the YAP1 gene and GFP were cloned into the pRS316 plasmid as follows: yeast cells were co-transformed with the next three overlapping DNA fragments: A) a 2.6 kbp PCR-amplified genomic fragment, using the primers YAP1-GFP_fw2 and YAP1-GFP_rv1 (Table 4) as a forward and reverse respectively, containing the promoter and coding region of YAP1. B) A 0.9 kbp PCR-amplified product using the plasmid pFA6a-GFP(S65T)-HIS3MX6 [60] as a template and the primers GFP_UP and GFP_rv1 (Table 4). C) And the EcoRI/XhoI digested pRS316 plasmid [61]. Transformed cells were selected in medium without uracil and their plasmid purified and verified by restriction analysis. The functionality of the final construct was verified by its ability to complement a yap1 deletion mutant and by visualization of the translocation to the nucleus of the Yap1-GFP fusion protein under oxidative stress.

Cultures treatments for RNA preparation
Cultures of the wild-type CML476 and its derivative SDP8 strains grown for 8 h in the presence of doxycycline (100 μg/ml) were collected at OD 660~0 .6. For the 24-h time point the appropriate quantity of cells were collected after 8 h of incubation with doxycycline, resuspended in 50 ml of YPD at an OD 660 of 0.01 and grown for an additional 16 h in the presence of the same quantity of fresh antibiotic. For the heat-stressed cells, wild-type and SDP8 cells treated with doxycycline for 24 h as above were incubated either at 40°C or 28°C for 40 min. In all cases, cell samples were collected by sedimentation (5 min at 1500 g), washed with cold water, and the dried cell pellet was kept at −80°C until RNA purification.

Purification of RNA, RT-PCR and microarray analysis
Extraction of total RNA and transcription analysis by DNA microarrays were performed essentially as previously described [62] using 8 μg of total RNA and an indirect labeling kit (CyScribe Post-Labeling kit, GE-Amersham Biosciences) for the cDNA synthesis, in conjunction with Cy3-dUTP and Cy5-dUTP fluorescent nucleotides. The yeast genomic microchips were constructed in our laboratory and contain at least two copies for each of the 6014 different PCR-amplified ORFs from S. cerevisiae [GEO Platform GPL10039] [63,64]. Pre-hybridization, hybridization, and washes were carried out as recommended by The Institute for Genomic Research [65] with the minor modifications described in [62]. The slides were scanned with a ScanArray 4000 apparatus (Packard BioChips Technologies), and the output was analyzed using GenePix Pro 6.0 software. Spots with either a diameter smaller than 120 μm or fluorescence intensities for Cy3 and Cy5 lower than 150 units were not considered for further analysis. A technical replica was performed for each experiment, where dyes were swapped to avoid dye-specific bias.
In a first set of experiments, we compared the expression profile of SDP8 cells with that of wild-type CML476 cells; both cell types were exposed to doxycycline for a total time of 8 or 24 h in YPD media. In the second series of experiments, we compared the transcriptomic profiles SDP8 and wild-type CML476 cells treated with doxycycline for 24 h and heat-shocked for 40 minutes at 40°C with those of the equally-treated cells but in the absence of the stress. Data from technical replicas were combined and the mean was calculated. A given gene was considered to be induced or repressed when the expression ratio was higher than 2.0 or lower than 0.50, respectively. According to the expression of the genes after heat shock in the wild-type CML476 and SDP8 strains, different levels of dependence on Pkh were defined, as previously described [23]. Thus, genes showing an SDP8/wild-type ratio of 0.67 > X > 0.50 were considered "weakly dependent" (WD), those with a ratio of 0.50 > X > 0.25 were ranked as "strongly dependent" (SD) and those with a ratio ≤0.25 were defined as "totally dependent" (TD). Similarly, genes induced more than 2.5-fold in wild-type cells and considered not induced (i.e., the ratio of stress/no stress <1.3) in SDP8 cells were also considered as TD.
Expression changes for several genes were confirmed by RT-PCR analyses using the Ready-To-Go RT-PCR Beads kit (GE Healthcare) with 50 ng of total RNA in 30 cycles. Gene-specific pairs of oligonucleotides (Table 4) were used to determine the levels of the corresponding mRNAs.

Beta-galactosidase assays
To evaluate the promoter activity of the PHO84, as well as the STRE-and the UPRE-driven transcriptional response in the presence and absence of doxycycline, wild-type and SDP8 cells incubated for 8 or 24 h in the presence of 100 μg/ml were harvested, β-Galactosidase activity measured as described previously [57] and results expressed as Miller Units [66]. Activation of the unfolded protein response pathway was assessed in the same way by using the pMCZ-Y plasmid contains the UPRE fused to lacZ gene.

Informatics tools
The ProtParam algorithm at the Expasy server was used to estimate the protein half-life [67]. The pre-processing tool of the Babelomics platform (http://v4.babelomics.org/ ), was used to pre-process the microarray data, obtaining only one expression value for each gene [68]. The MIPS Functional Catalogue Database (MIPS FunCatDB) [69], available http://mips.helmholtz-muenchen.de/funcatDB/, was used for the functional distribution of gene lists. The Publication Enrichment tool available at the YeastMine website [70] was used to identified significant overlaps between customized set of genes and genes from published expression patterns. The YEASTRACT (Yeast Search for Transcriptional Regulators And Consensus Tracking; http://www.yeastract.com/) tools were used for grouping genes based on their regulatory associations with documented transcription factors [71]. We searched for transcription factors as activators and filtrated for DNA binding plus expression evidence.

Availability of supporting data
The data set supporting the results of this article is available in the Gene Expression Omnibus repository, [GSE32623, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE32 623].

Additional files
Additional file 1: Table S1.