Current Genetics

, Volume 48, Issue 1, pp 18–33

The minimum domain of Pho81 is not sufficient to control the Pho85–Rim15 effector branch involved in phosphate starvation-induced stress responses

Authors

  • Erwin Swinnen
    • Functional BiologyKatholieke Universiteit Leuven
  • Joëlle Rosseels
    • Functional BiologyKatholieke Universiteit Leuven
    • Functional BiologyKatholieke Universiteit Leuven
Research Article

DOI: 10.1007/s00294-005-0583-3

Cite this article as:
Swinnen, E., Rosseels, J. & Winderickx, J. Curr Genet (2005) 48: 18. doi:10.1007/s00294-005-0583-3

Abstract

The phosphate regulatory mechanism in yeast, known as the PHO pathway, is regulated by inorganic phosphate to control the expression of genes involved in the acquisition of phosphate from the medium. This pathway is also reported to contribute to other nutritional responses and as such it affects several phenotypic characteristics known also to be regulated by protein kinase A, including the transcription of genes involved in the general stress response and trehalose metabolism. We now demonstrate that transcription of post-diauxic shift (PDS)-controlled stress-responsive genes is solely regulated by the Pho85–Pho80 complex, whereas regulation of trehalose metabolism apparently involves several Pho85 cyclins. Interestingly, both read-outs depend on Pho81 but, while the previously described minimum domain of Pho81 is sufficient to sustain phosphate-regulated transcription of PHO genes, full-length Pho81 is required to control trehalose metabolism and the PDS targets. Consistently, neither the expression control of stress-regulated genes nor the trehalose metabolism relies directly on Pho4. Finally, we present data supporting that the PHO pathway functions in parallel to the fermentable growth medium- or Sch9-controlled pathway and that both pathways may share the protein kinase Rim15, which was previously reported to play a central role in the integration of glucose, nitrogen and amino acid availability.

Keywords

PhosphatePho85Pho80Pho81Sch9Rim15

Introduction

Inorganic phosphate (Pi) is an essential nutrient for all organisms, being required for the biosynthesis of nucleic acids, phospholipids and cellular metabolites as well as for energy metabolism and signal transduction. Microorganisms often encounter limiting concentrations of this essential nutrient and have evolved complex signal transduction networks that enable them to make optimal use of the scarce available phosphate sources. When Pi in the medium becomes limiting, the budding yeast Saccharomyces cerevisiae activates a regulatory mechanism known as the PHO pathway, which results in an increased expression of multiple genes involved in the acquisition, uptake and storage of phosphate (Ogawa et al. 2000; Persson et al. 2003). This transcriptional activation is mediated by the transcription factor Pho4 in cooperation with the coactivator Pho2 (Lenburg and O’Shea 1996). Typical Pho4 target genes include PHO84 and PHO89, encoding the high-affinity phosphate transporters, and PHO5, encoding a secreted acid phosphatase. The activity of Pho4 is regulated through phosphorylation at several serine residues by the cyclin-dependent kinase (CDK) Pho85 when associated with the Pho80 cyclin (Kaffman et al. 1994). Phosphorylation of Pho4 leads to its inactivation via disruption of the Pho4–Pho2 interaction, stimulation of Pho4 nuclear export and inhibition of its nuclear import (O’Neill et al. 1996; Kaffman et al. 1998a, b; Komeili and O’Shea 1999). Regulation of the kinase activity of the Pho85–Pho80 complex in response to phosphate levels occurs through the Pho81 CDK inhibitor (CKI) protein (Schneider et al. 1994; Ogawa et al. 1995). Under low-phosphate conditions, Pho81 inhibits the kinase activity of the Pho85–Pho80 complex, thereby allowing nuclear import and activation of the Pho4 transcription factor. Notably, only one particular domain of Pho81, the so-called Pho81 minimum domain, is essential for inhibition of the Pho85–Pho80 kinase activity towards the Pho4 target (Huang et al. 2001). Since expression of PHO81 is itself controlled by Pho4, it is the target of a positive feedback loop where a low-Pi signal results in enhanced expression of PHO81 (Creasy et al. 1993; Ogawa et al. 1993, 1995). In addition, it was recently shown that the Pho80 and Pho81 proteins are subjects for Pho85-mediated phosphorylation and this phosphorylation appears to be required for the proper in vivo functioning of these components, indicating a multitude of internal feedback loops within the pathway (Knight et al. 2004; Waters et al. 2004).

Aside from its role in the PHO pathway, Pho85 has additional functions in many other pathways (for reviews, see Moffat et al. 2000; Carroll and O’Shea 2002). This is reflected in the multiple phenotypes displayed by a pho85Δ strain. Cells deleted for the PHO85 gene grow slowly on glucose and fail to properly utilize a number of other carbon sources including galactose, sucrose and several non-fermentable carbon sources (Lee et al. 2000). Other phenotypes include hyperaccumulation of glycogen, sporulation defects, morphological abnormalities, hypersensitivity to a large number of stress conditions and aberrant expression profiles during the diauxic shift (Gilliquet and Berben 1993; Timblin et al. 1996; Measday et al. 1997; Huang et al. 2002; Nishizawa et al. 2004). Consistent with the diverse roles of Pho85, ten genes encoding known or putative Pho85 cyclins (Pcl) have been identified (Measday et al. 1997). Hence, the different functions of Pho85 may relate to specific cyclin subunits; and in some cases the molecular targets for particular kinase complexes have indeed been identified (Huang et al. 1996, 1998; Wilson et al. 1999). For most processes involving the Pho85 CDK, however, neither the required cyclin partners nor the molecular targets for the kinase complex are known. Another issue concerns the regulation of the activities of the different Pho85 complexes. Only for the Pho85–Pho80 complex, is regulation via the Pho81 CKI well documented. Other Pho85 complexes have been reported to be regulated by a cell cycle-dependent transcriptional control on cyclin expression (Ogas et al. 1991; Espinoza et al. 1994; Measday et al. 1994, 1997; Aerne et al. 1998; Tennyson et al. 1998; Lee et al. 2000) or by activation of the cyclin in response to nutrient availability (Shemer et al. 2002).

In this paper, we focused on phenotypes known to be dependent on protein kinase A (PKA), including trehalose metabolism and expression of stress-responsive genes (Roosen et al. 2004). We demonstrate that the PHO pathway is required for proper regulation of these read-outs in response to phosphate levels. We provide evidence that different Pho85 complexes may regulate trehalose metabolism, while specifically the Pho85–Pho80 complex seems to be responsible for expression-regulation of several stress-responsive genes, especially those regulated by Gis1, a transcription factor that binds the post-diauxic shift (PDS) DNA element. We also show that Pho81 takes part in this regulation but, in contrast to the regulation of the PHO5 encoded acid phosphatase activity, the Pho81 minimum domain as described by Huang et al. (2001) seems not to be sufficient. Consistently, none of the phenotypic read-outs strictly depends on Pho4. Since expression of stress-responsive genes and trehalose metabolism are known targets of the Sch9-controlled fermentable growth medium (FGM) pathway, which converges downstream of PKA on the protein kinase Rim15 (Reinders et al. 1998; Pedruzzi et al. 2003; Roosen et al. 2005), we analyzed a possible connection between this signalling cascade and the PHO pathway. We demonstrate that, similar to Sch9, Pho85 also appears to regulate Rim15 and the Rim15 effectors. Interestingly, both kinases appear to function in parallel signalling routes, as indicated by the synthetic lethality when the deletions of SCH9 and PHO85 are combined.

Materials and methods

Yeast strains and plasmids

S. cerevisiae strains used in this study are listed in Table 1. Deletions were made using polymerase chain reaction-based disruption cassettes as described by Brachmann et al. (1998). Sporulation was performed by spotting diploid cells onto plates containing 1% K-acetate, 0.1% KHCO3, pH 6.0, for 5–6 days at 24°C. All experiments were carried out with isogenic wild-type and mutant strains. Plasmids EB0730 (pADH1pr-Py2 in pRS316), EB1361 (pADH1pr-PHO81-Py2 in pRS316), EB1362 (pADH1pr-PHO81(aa 400-724)-Py2 in pRS316), EB1363 (pADH1pr-PHO81(aa 400-660)-Py2 in pRS316) and EB1364 (pADH1pr-PHO81(aa 645-724)-Py2 in pRS316) were kindly gifted by E.K. O’Shea and described by Huang et al. (2001).
Table 1

Strains used in this study

Name

Genotype

Reference

BY4741 (wild type)

Matahis3Δ1 leu2Δ0 met15Δ0 ura3Δ0

Y.K.O. collection

pho4Δ

BY4741 with pho4::KANMX4

Y.K.O. collection

pho80Δ

BY4741 with pho80::HIS3

This study

pho81Δ

BY4741 with pho81::KANMX4

Y.K.O. collection

pho85Δ

BY4741 with pho85::KANMX4

Y.K.O. collection

pho81Δpho85Δ

BY4741 with pho81::KANMX4 pho85::KANMX4

This study

pho81Δpho80Δ

BY4741 with pho81::KANMX4 pho80::HIS3

This study

rim15Δ

BY4741 with rim15::KANMX4

Y.K.O. collection

pho85Δrim15Δ

BY4741 with pho85::KANMX4 rim15::KANMX4

This study

pho80Δrim15Δ

BY4741 with pho80::HIS3 rim15::KANMX4

This study

sch9Δ

BY4741 with sch9::HIS3

This study

BY4742

Matα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0

Y.K.O. collection

pho85Δ Matα

BY4742 with pho85::KANMX4

Y.K.O. collection

Culture and starvation conditions

Cells were cultured at 30°C into exponential phase (optical density at 600 nm = 1.5–2.0) in YP medium (1% yeast extract, 2% bacto peptone) with 2% glucose. Mid-exponential cells were harvested and transferred to phosphate-starvation medium (5.7 g l−1 yeast nitrogen base without phosphate) with 4% glucose, uracil (50 mg l−1), leucine (250 mg l−1), histidine (100 mg l−1) and methionine (20 mg l−1). Cells were starved for phosphate for 3 days at 30°C under continuous shaking and starvation medium was refreshed daily. The pho81Δ and pho4Δ strains with constitutive overexpression of specific PHO81 constructs were grown in SD-ura medium (synthetic medium without uracil containing 2% glucose) until the exponential phase and transferred to phosphate-starvation medium (without uracil) containing 4% glucose.

Incubation conditions for phosphate-addition assays

The phosphate-starved, glucose-repressed cells were rapidly cooled on ice and harvested by centrifugation. The pellet was washed twice with ice-cold MES buffer, pH 6.0, resuspended in fresh phosphate-starvation medium with 4% glucose and incubated at 30°C with shaking. After 30 min of incubation, 10 mM KH2PO4 was added to the culture and samples were taken at the indicated time-points.

Determination of trehalose levels

Samples were collected by filtration at the indicated time-points, washed once with ice-cold water, weighed and frozen in liquid nitrogen. The pellets were resuspended in 1 ml of 0.25 M Na2CO3 per 40 mg of cells and boiled at 95°C for 60 min. Samples were neutralized and intracellular levels of trehalose were determined using Humicola trehalase (Neves et al. 1994; Colombo et al. 1998). Glucose concentrations were measured using the GOD-PAP kit (Dialab).

Sampling, extraction and determination of trehalase activity

Samples of 50 mg cells ml−1 were taken at the indicated time-points. Cells were rapidly cooled by the addition of ice-cold water, centrifuged and resuspended in 500 μl of 25 mM MES buffer, 50 μM CaCl2, pH 7, for extraction. Crude cell extracts were prepared as described by Giots et al. (2003) and dialyzed against 25 mM MES buffer, pH 7, with 50 μM CaCl2. Trehalase activity in dialyzed cell extracts was determined as described by Giots et al. (2003). The specific activity is expressed as nmol glucose liberated min−1  mg protein−1 . The total amount of protein in the samples was determined using a standard method (Lowry et al. 1951).

RNA extraction and Northern blot analysis

Samples were taken at the indicated time-points and rapidly cooled by the addition of ice-cold water. The cells were pelleted, washed once with ice-cold water and stored at −80°C. Isolation of total RNA was performed as described by Giots et al. (2003). Probes were labelled with α−32 P-dCTP, using the High Prime kit (Boehringer Mannheim, Monza, Italy). Northern blots were made by separation of total RNA on gels containing 1% (w/v) agarose in 50 mM boric acid, 1 mM sodium citrate, 5 mM NaOH, pH 7.5, 1% formaldehyde. Subsequently, RNA was transferred by capillary blotting to a Hybond-N membrane (Amersham), using 10× SCC buffer (1.5 M NaCl, 0.15 M Na-citrate, pH 7.0). The filters were hybridized with the 32 P-labelled probes consisting of fragments of the desired coding region. The blots were analyzed using PhosphorImager technology (FUJIX, BAS-1000). Quantification of Northern blot images was performed with TINA ver. 2.0.

Reproducibility of results

All experiments were performed at least three times. For trehalose levels, mean values for three independent experiments are shown, with error bars representing the corresponding standard deviation (SD). For trehalase experiments, the absolute values of activities were variable between different experiments, but the relative differences observed between separate experiments and strains were highly reproducible. Therefore, statistical analysis on the basal level of trehalase activity and activation extent was performed on a relative scale, with the wild-type level set to 100% in each experiment, after which mean values and SD were calculated for three independent experiments. These values are represented in Tables 2, 3, 4. For Northern blot experiments, representative blots from three independent experiments are shown. To demonstrate the significance of the difference in expression between the different pho mutants, quantifications of the stress-responsive and PHO genes are included for the first Northern blot experiment. For the indicated time-points, expression levels are quantified using TINA ver. 2.0 and normalized using the corresponding 18S expression levels. Wild-type levels after 3 days of starvation and at the beginning of the repression experiment are set to 100%. Mean values of the relative expression of three independent experiments are shown, along with their SD.
Table 2

Relative basal level and activation of trehalase activity for the different pho mutants described in Fig. 2a. Trehalase activities are expressed as a percentage of wild-type levels (mean values from three independent experiments, with SD in brackets)

 

pho85Δ

pho80Δ

pho81Δ

pho81Δpho85Δ

pho81Δpho80Δ

Basal activity

94.5 (6.4)

106.7 (10.4)

157.1 (5.6)

107.1 (7.6)

155.5 (7.3)

Activation

55.4 (14.7)

62.3 (8.8)

60.2 (10.9)

42.3 (12.9)

41.3 (4.2)

Table 3

Relative basal level and activation of trehalase activity of the pho81Δ strain transformed with different PHO81 constructs (see Fig. 4b). Trehalase activities are expressed as a percentage of wild-type levels (WT + EB0730; mean values from three independent experiments, with SD in brackets). See Materials and methods for a description of the different plasmids

 

Pho81Δ + EB0730

pho81Δ + EB1361

pho81Δ + EB1362

pho81Δ + EB1363

pho81Δ + EB1364

Basal activity

161.0 (2.4)

113.9 (9.5)

174.0 (3.1)

163.2 (12.6)

163.8 (4.7)

Activation

50.0 (7.3)

39.0 (0.3)

54.6 (5.4)

56.8 (5.4)

47.0 (1.1)

Table 4

Relative basal level and activation of trehalase activity of the pho4Δ strain transformed with different PHO81 constructs (see Fig. 5c). Trehalase activities are expressed as percentage of wild-type levels (WT + EB0730; mean values from three independent experiments, with SD in brackets). See Materials and methods for a description of the different plasmids

 

pho4Δ + EB0730

pho4Δ + EB1361

pho4Δ + EB1364

Basal activity

159.1 (12.2)

122.8 (1.4)

174.0 (1.2)

Activation

83.6 (0.9)

52.1 (1.8)

75.5 (8.8)

Results

Pho81-mediated inhibition of the Pho85–Pho80 complex is required for the correct regulation of stress-related phenotypes

When wild-type yeast cells are starved for inorganic phosphate, they arrest growth and enter the G0 phase. During starvation, these cells acquire stress resistance as they accumulate the stress-protectant trehalose (Fig. 1a) and induce the expression of stress-responsive genes. The latter include: (1) genes that contain solely PDS elements in their promoter, such as SSA3 and GRE1, allowing transcriptional control by the transcription factor Gis1 and (2) genes that contain stress response elements (STREs) in their promoter, such as HSP26, which constitute the binding sites for the transcription factors Msn2 and Msn4 (Fig. 1b; Pedruzzi et al. 2000; Cameroni et al. 2004; Roosen et al. 2005). The deletion of PHO80 or PHO85 did not prevent the starvation-induced responses, which was expected since the Pho85–Pho80 complex is inhibited by Pho81 under conditions of phosphate limitation. The deletion of PHO81, in contrast, impaired the accumulation of trehalose (Fig. 1a) and displayed a delayed and reduced induction of the PDS-driven genes SSA3 and GRE1, while effects on the STRE-driven gene HSP26 were minimal (Fig. 1b,c). Subsequent analysis of the pho81Δpho85Δ and pho81Δpho80Δ double-deletion strains showed that the defect in PDS-controlled gene expression of the pho81Δ mutant was due to a constitutive active Pho85–Pho80 complex. However, the additional deletion of PHO80 in the pho81Δ mutant only partially suppressed the defect in trehalose accumulation of the pho81Δ strain, in contrast to the full suppression obtained by the additional deletion of PHO85 (Fig. 1a). This result indicates that other Pho85-associated cyclin(s) are required to maintain increasing trehalose levels during phosphate starvation. Examination of the expression of NTH1, TPS1 and TPS2, encoding respectively the trehalase enzyme, the trehalose-6-phosphate synthase and the phosphatase subunit of the trehalose synthase complex, during phosphate starvation revealed no significant defects in the induction of these genes in any of the pho mutants, compared with the wild-type strain (Fig. 1b). This suggests that the defect in trehalose accumulation of the pho81Δ mutant is likely due to impaired post-transcriptional regulation of the enzymes involved in trehalose metabolism.
Fig. 1

The PHO pathway mediates phosphate starvation-induced effects on known PKA targets. Exponentially growing cells of the wild type (BY4741) and isogenic pho mutants were harvested and transferred to phosphate-depleted medium. The trehalose content in the different strains (a) was determined each day of the starvation period. Filled circles Wild type, open circles pho85Δ, filled triangles pho80Δ, open triangles pho81Δ, filled squares pho81Δ pho85Δ, open squares pho81Δ pho80Δ . Mean values for three independent experiments are shown, with SD (error bars). Samples of the wild-type strain and isogenic pho mutants were also taken during the starvation period for Northern blot and expression analysis (b) of stress-responsive genes (i.e. HSP26, SSA3, GRE1), genes encoding proteins involved in trehalose metabolism (i.e. NTH1, TPS1, TPS2) and known phosphate-responsive genes (i.e. PHO84, PHO89). 18S mRNA levels are shown as a standard. Expression levels for the stress-responsive genes and PHO genes during each day of the starvation period were quantified (c) to demonstrate the significance of the difference between the different pho mutants. Average values for three independent experiments are shown, with SD (error bars). The wild-type level of each gene after 3 days of starvation was set to 100%

The re-addition of phosphate to phosphate-starved wild-type cells reversed the starvation responses and triggered a rapid activation of trehalase (Fig. 2a, Table 2) allowing mobilization of the trehalose reserves (Fig. 2b). The single and combined pho85Δ and pho80Δ mutants, however, did not (efficiently) mobilize their trehalose reserves despite the observation that there was still a considerable activation of trehalase, albeit lower than in the wild type. This uncoupling of trehalose degradation and trehalase activation was even more pronounced in the pho81Δ strain, where trehalose levels appeared to increase upon phosphate re-addition despite clear activation of trehalase. The pho81Δ mutant also displayed a significant higher basal trehalase activity before phosphate re-supplementation, which is consistent with the lower trehalose content found in starved cells. Interestingly, the basal trehalase activity was only restored by the additional deletion of PHO85 but not by the deletion of PHO80. The latter again indicates that, besides Pho80, other Pho85-associating cyclins should be involved to maintain proper trehalose metabolism.
Fig. 2

The effect on trehalose metabolism of phosphate re-supplementation to phosphate-starved cells. Trehalase activation (a) and mobilization of trehalose (b) after the addition of 10 mM KH2PO4 to phosphate-starved cells of the wild-type strain and isogenic pho mutants. Filled circles Wild type, open circles pho85Δ, filled trianglespho80Δ, open triangles pho81Δ, filled squares pho81Δpho85Δ, open squares pho81Δpho80Δ . For trehalose levels, mean values for three independent experiments are shown, with SD (error bars)

Re-supplementation of phosphate to the starved cells also triggered a rapid repression of both PDS-controlled and STRE-controlled genes in the wild-type strain and in the single pho81Δ strain (Fig. 3a,b). In the mutants lacking PHO85 or PHO80, however, this repression was significantly delayed and for some genes even absent, indicating that phosphate-induced repression of the PDS-driven and STRE-driven genes is also dependent on the Pho85–Pho80 complex. Surprisingly, repression of some or all of the genes involved in trehalose metabolism (NTH1, TPS1, TPS2) was defective in all pho mutants tested, even in the pho81Δ strain, and may in part account for the lack of proper trehalose mobilization after the addition of phosphate, despite activation of trehalase under the same conditions. As noted previously, this result indicates that these genes do not behave as typical STRE-controlled genes and that other but yet unknown mechanisms and factors contribute to the regulation of their expression (Winderickx et al. 1996; Zahringer et al. 1998, 2000).
Fig. 3

The effect on gene expression of phosphate re-supplementation to phosphate-starved cells. Northern blot analysis (a) for stress-responsive genes (i.e. HSP26, SSA3, GRE1), genes encoding proteins involved in trehalose metabolism (i.e. NTH1, TPS1, TPS2) and phosphate-responsive genes (i.e. PHO84, PHO89) after the addition of 10 mM KH2PO4 to phosphate-starved cells of the wild-type strain and isogenic pho mutants. Samples were taken before and several minutes after phosphate re-supplementation, as indicated. 18S mRNA levels are shown as a standard. Expression levels for the stress-responsive genes and PHO genes before phosphate re-addition and 60 and 120 minutes after the addition of 10mM KH2PO4 to the phosphate-starved cells were quantified (b) to demonstrate the significance of the difference in repression between the different pho mutants. Average values for three independent experiments are shown, with SD (error bars). The initial wild-type level of each gene was set to 100%

As a control for the experiments described above, the expression of typical Pho4 targets, such as PHO84 and PHO89, was monitored during phosphate starvation (Fig. 1b,c) and after re-supplementation of phosphate to the starved cells (Fig. 3a,b). Consistent with previously reported data (Lenburg and O’Shea 1996), the Pho4 target genes are induced in the wild-type strain upon starvation and repressed when phosphate is again available. Their expression is completely abolished in the pho81Δ mutant, while it is constitutive in the single pho85Δ and pho80Δ mutants and in the double pho81Δpho85Δ and pho81Δpho80Δ mutants.

The Pho81 minimum domain is only sufficient to sustain Pho4-mediated transcription of PHO genes

The Pho81 contains a minimum domain that appears to be necessary and sufficient for the CKI function (Huang et al. 2001). This minimum domain comprises a region of 80 amino acids C-terminal to the ankyrin repeats (amino acids 645–724) and is required for the binding of Pho81 to the Pho85–Pho80 kinase complex. So far, only the regulation of the Pho5 acid phosphatase activity in a pho81Δ strain has been used as a readout to demonstrate the importance of this minimum domain. Therefore, different PHO81 truncations were tested for their ability to complement the defects related to trehalose metabolism and expression of the PDS genes SSA3 and GRE1 in the pho81Δ mutant. The pho81Δ mutant transformed with an empty plasmid (construct EB0730) or with constructs encoding only the ankyrin repeats (construct EB1633), the ankyrin repeats and the minimum domain (construct EB1362) or only the minimum domain (construct EB1364) behaved similarly and failed to accumulate trehalose (Fig. 4a), to restore the enhanced basal trehalase activity during starvation (Fig. 4b, Table 3) or to mobilize carbohydrate upon re-supplementation of phosphate (Fig. 4c). These transformants did also not restore the delayed and reduced induction of the PDS-driven genes upon phosphate starvation, although the overexpression of at least the PHO81 minimum domain (EB1362, EB1364) in the pho81Δ strain enabled the strain to induce again the Pho4 target genes PHO84 and PHO89 under the same conditions (Fig. 4d). Only the re-introduction of full-length PHO81 (construct EB1361) partially restored all defects in the pho81Δ mutant. Thus, while the minimum domain is able to regulate the activity of the Pho85–Pho80 kinase complex towards the Pho4 target, it is ineffective in regulating the activity of the same kinase complex towards trehalose metabolism and expression control of the stress-responsive genes.
Fig. 4

Analysis of different PHO81 constructs for their ability to complement the pho81Δ strain. Exponentially growing cells of the wild-type strain and pho81Δ mutant strains transformed with different Pho81-encoding constructs were harvested and transferred to phosphate-depleted medium. The accumulation of trehalose (a) was followed each day of the starvation period for the different strains. The same strains were analyzed after the addition of 10 mM KH2PO4 to the phosphate-starved cells in order to monitor the trehalase activation (b) and trehalose mobilization (c). Filled circles Wild type (BY4741) + EB0730, open circles pho81Δ + EB0730, filled triangles pho81Δ + EB1361, open triangles pho81Δ + EB1362, filled squares pho81Δ + EB1363, open squares pho81Δ + EB1364. For trehalose levels, mean values for three independent experiments are shown, with SD (error bars). See Materials and methods for a description of the different Pho81-encoding plasmids. Samples were also taken during phosphate starvation of the wild-type strain and the different pho81Δ mutant strains for Northern blot and expression analysis (d) of the stress-responsive genes (HSP26, SSA3, GRE1) and the phosphate-responsive genes (PHO84, PHO89). 18S levels are shown as a standard

The defect of a strain lacking a functional Pho4 can be suppressed by overexpression of PHO81

During phosphate starvation, Pho81 inhibits the kinase activity of the Pho85–Pho80 complex, leading to an activation of PHO gene transcription via the Pho4 transcription factor (Lenburg and O’Shea 1996; Komeili and O’Shea 1999). One of the target genes of the Pho4 transcription factor is PHO81 itself; and it was shown that this positive feedback loop is required for proper signalling through the PHO pathway (Creasy et al. 1993; Ogawa et al. 1993, 1995). Consistently, cells lacking Pho4 failed to accumulate trehalose during phosphate starvation and, like cells deficient for Pho81, this defect was only restored by overexpression of full-length PHO81 and not by overexpression of the minimum domain of PHO81 (Fig. 5a). Also, the abnormal mobilization profile of trehalose after re-supplementation of phosphate to the starved pho4Δ cells could only be corrected by overexpressing full-length PHO81 (Fig. 5b). Note that the trehalose levels initially tended to increase in the pho4Δ upon re-supplementation of phosphate, even though this strain displayed an almost wild-type curve for the activation of trehalase (Fig. 5c, Table 4). Thus, both the pho4Δ and the pho81Δ mutants share the uncoupling of trehalose degradation and trehalase activation. Furthermore, the overexpression of the full-length PHO81 not only reduced the phosphate-induced trehalase activation in the pho4Δ mutant but also partially restored the increased basal trehalase activity before the addition of phosphate to phosphate-starved cells. With respect to the induction of the PDS-driven SSA3 and GRE1 genes during phosphate starvation, the pho4Δ mutant was clearly delayed when compared to the wild-type strain; and again this delay was only restored by overexpression of full-length PHO81 (Fig. 5d).
Fig. 5

Expression of full-length PHO81, but not the minimum domain, can suppress the defects of a pho4Δ strain. Exponentially growing cells of the wild-type strain and pho4Δ mutants transformed with various Pho81-encoding plasmids were harvested and transferred to phosphate-depleted medium. The trehalose content (a) was determined for the different strains each day of the starvation period. The strains were also followed after the addition of 10 mM KH2PO4 to the phosphate-starved cells, to monitor trehalose mobilization (b) and trehalase activation (c). Filled circles Wild type (BY4741) + EB0730, open circles pho4Δ + EB0730, filled triangles pho4Δ + EB1361, open triangles pho4Δ + EB1364. For trehalose levels, mean values for three independent experiments are shown, with SD (error bars). See Materials and methods for a description of the different Pho81-encoding plasmids. Samples were also taken each day of the phosphate-starvation period to follow the expression (d) of the stress-responsive genes (HSP26, SSA3, GRE1) and the phosphate-responsive genes (PHO84, PHO89). 18S levels are shown as a standard

These data led us to conclude that Pho4 was not affecting the targets studied directly, but indirectly via the positive feedback loop for expression of PHO81. In line with this hypothesis, the expression of a constitutively active form of Pho4, containing five serine-to-alanine substitutions at its sites of phosphorylation by Pho85–Pho80 (Kaffman et al. 1998a, b), could not alleviate the defective trehalose accumulation or the retarded induction of stress-responsive genes in the pho81Δ mutant during phosphate starvation, despite the fact that it rescued PHO gene transcription (data not shown).

The PHO pathway functions in parallel to Sch9-controlled signalling to regulate Rim15 and the Rim15 effectors

The observation that aberrant signalling of the PHO pathway affected mainly those stress-responsive genes that contain solely PDS elements and no additional STRE consensus sequences in their promoter incited us to investigate possible cross-talk with other nutrient-controlled pathways. As mentioned above, PDS elements are recognized by the transcription factor Gis1, which is of utmost importance to allow yeast cells to traverse the diauxic shift and adapt metabolism and growth from fermentation to respiration (Pedruzzi et al. 2000). More recent research confirmed that the transcription activity of Gis1 is in turn regulated by the PAS-like protein kinase Rim15 and the protein kinase Sch9 (Pedruzzi et al. 2000; Cameroni et al. 2004; Roosen et al. 2005). In order to illustrate the interplay of these kinases with the PHO pathway and pinpoint a likely candidate to transmit signals derived from the Pho85/Pho80 complex to Gis1, we introduced the additional deletion of SCH9 or RIM15 in the pho85Δ mutant. As illustrated in Fig. 6a, tetrad analysis of a heterozygote diploid indicated that, although spores with a combined deletion of SCH9 and PHO85 could germinate, the strain displays a dramatic synthetic growth defect and therefore could not be analyzed further. Nevertheless, this synthetic phenotype indicates that Sch9 and Pho85 may function in parallel but converging pathways. For RIM15, we found that its additional deletion did not influence the growth properties of pho85Δ cells when these are spotted onto glucose-supplemented media. This could be expected, since Rim15 is immediately downstream and inactivated via phosphorylation by PKA during fermentative growth (Reinders et al. 1998). However, when spotted onto glycerol-containing medium, the additional deletion of RIM15 gave partial suppression of the respiratory growth defect typical for the pho85Δ mutant (Fig. 6b). In line with this, the additional deletion of RIM15 in both the pho85Δ mutant and the pho80Δ mutant also overruled phosphate signalling towards the read-outs studied. Similar to the single rim15Δ mutant, the double-mutants rim15Δpho85Δ and rim15Δpho80Δ failed to accumulate trehalose (Fig. 7a). These strains were also defective for the induction of the stress-responsive genes upon phosphate starvation (Fig. 7b), while they gave a faster repression of these genes upon re-supplementation of phosphate to the starved cells (Fig. 7c). Hence, consistent with a model that Rim15 is negatively regulated by the Pho85–Pho80 kinase, the rim15Δ mutants persistently displayed phenotypes that were better comparable to those of the pho81Δ mutant than to those of the pho85Δ or pho80Δ mutant.
Fig. 6

Pho85 functions in parallel to Sch9 and regulates the protein kinase Rim15. Diploid cells heterozygous for PHO85 and SCH9 were sporulated and subjected to tetrad analysis (a). Haploid cells with a pho85Δsch9Δ genotype displayed a severe growth defect, failing to form healthy colonies. Growth of strains deleted in PHO85 and/or RIM15 was analyzed (b). Cells were spotted onto solid medium for 3–5 days before scanning

Fig. 7

The protein kinase Rim15 is essential for phosphate-dependent signalling. The effect of a rim15Δ on phosphate signalling with or without the additional deletion of a component of the Pho85–Pho80 kinase was determined. Cells were starved for phosphate during 3 days, after which the trehalose content (a) was determined. Mean values for three independent experiments are shown, with SD (error bars). Samples were also taken during the starvation period (b) and after re-supplementation of 10 mM KH2PO4 to the phosphate-starved cells (c), in order to monitor the expression of the stress-responsive genes (HSP26, SSA3, GRE1) and the phosphate-responsive genes (PHO84, PHO89). 18S levels are shown as a standard

Discussion

When yeast cells encounter nutrient limitation, they respond by activating several stress responses, which include the accumulation of the reserve carbohydrates trehalose and glycogen and the induction of a multitude of stress-responsive genes (Lillie and Pringle 1980; Thevelein and de Winde 1999). These properties are known to be tightly regulated by multiple signalling pathways with a major role being played by PKA (Francois and Parrou 2001; Winderickx et al. 2003; Cameroni et al. 2004; Roosen et al. 2004, 2005). The latter was concluded based on several studies showing that cells induce stress responses and become more stress-resistant under conditions that correspond to low PKA activity, i.e. if glucose or one of the essential nutrients is missing, while they display opposed responses and become stress-sensitive under conditions that are considered equivalent to high PKA activity, i.e. when cells are grown on rich glucose-containing medium where all nutrients essential for growth are abundantly present (Hirimburegama et al. 1992; Durnez et al. 1994; Donaton et al. 2003; Giots et al. 2003). The adaptation to the presence of essential nutrients appeared not to be associated with changes in the cAMP level, therefore it was postulated that another pathway could control the activity of the catalytic subunits of PKA. This pathway was called the FGM pathway, with the protein kinase Sch9 being the major controlling component (Thevelein 1994; Crauwels et al. 1997). Most recently, the concept of the FGM pathway was demonstrated to reflect convergence of the Sch9-dependent pathway on Rim15, a protein kinase immediately downstream and under negative control of PKA (Reinders et al. 1998; Pedruzzi et al. 2003; Roosen et al. 2005).

In this paper, we demonstrated the involvement of the yeast PHO pathway for the phosphate-dependent regulation of the abovementioned stress responses. Similar to the control of Pho4-mediated transcription of several PHO genes, Pho81-dependent regulation of Pho85 kinase activity is required for proper trehalose metabolism in response to phosphate availability. Unregulated high activity of Pho85 due to the absence of Pho81 impairs trehalose accumulation during phosphate starvation, whereas the absence of Pho85 activity results in the loss of a quick trehalose mobilization after the re-addition of phosphate to the starved cells. However, in contrast to the known PHO pathway, regulation of trehalose levels apparently requires other Pho85-associating cyclins in addition to Pho80. This could be concluded from the observations that the pho81Δpho80Δ mutant still displays lower levels of trehalose than a wild-type strain, in contrast to the pho81Δpho85Δ mutant and that the increased basal activity of trehalase in the pho81Δ mutant can only be restored by the additional deletion of PHO85 but not by the additional deletion of PHO80. Concerning the rapid phosphate-induced trehalase activation, our results indicate that proper functioning of the PHO pathway is important for maximal signalling. However, other signalling systems are involved, as trehalase activation is not abolished in all the pho mutants tested. For instance, it has been shown that PKA plays an important role in the phosphate-induced activation of trehalase (Giots et al. 2003). Several Pho85 cyclins are already implicated in the regulation of the reserve carbohydrate glycogen and one may argue that they may also be required for correct regulation of trehalose levels. These cyclins include Pcl6, Pcl7, Pcl8 and Pcl10 (Huang et al. 1996, 1998; Wilson et al. 1999, 2005; Lee et al. 2000). Of them, Pcl7 might be a very interesting candidate, since this cyclin has been shown to directly interact with Pho81 in vivo, to regulate the kinase activity of the Pho85–Pcl7 complex in response to phosphate levels (Lee et al. 2000). However, this does not exclude that other Pho85 cyclins are involved in Pho81-dependent processes. For instance, the Pho85–Pcl8 and Pho85–Pcl10 complexes phosphorylate the glycogen synthase enzyme Gsy2 (Huang et al. 1996, 1998; Wilson et al. 1999), but so far it is not known how these kinase activities are regulated. Although Pho81 localizes to the nucleus under phosphate-limiting conditions, where it inhibits the activity of the nuclear Pho85–Pho80 complex, additional localization of Pho81 to the cytosol and plasmamembrane observed under the same conditions may be important for the additional function Pho81 plays in the regulation of trehalose metabolism (Huang et al. 2001). The role of the Pho4 transcription factor in the regulation of trehalose levels is limited to its positive action on Pho81 function via transcriptional induction of the PHO81 gene. This is an interesting result, as it shows that the Pho4-mediated feedback on PHO81 expression (Ogawa et al. 1995) is not only required for continued activation of Pho4 but is also important to allow Pho81 to regulate other Pho85-mediated processes in order to ensure a proper response to phosphate starvation.

A second common response of yeast cells to nutrient-starvation conditions is the transcriptional induction of a large set of stress-responsive genes and our results confirm that this also applies to phosphate starvation (Crauwels et al. 1997; Thevelein and de Winde 1999; Donaton et al. 2003; Giots et al. 2003). Interestingly, only for the expression-regulation of the PDS-driven genes SSA3 and GRE1 did our data clearly demonstrate the importance of the CKI function of Pho81. For the regulation of the STRE-driven gene HSP26, the involvement of the PHO pathway was much less obvious as its induction during phosphate starvation seemed to be largely independent of Pho81 while its repression upon phosphate re-supplementation to the starved cells was delayed in strains lacking Pho85 or Pho80. Note that also the expression of the genes encoding trehalase and the different subunits of the trehalose synthase complex is controlled in part via STRE elements, though they do not seem to act as other typical STRE-controlled genes (Winderickx et al. 1996; Zahringer et al. 1998, 2000). We also observed this phenomenon as repression of NTH1, TPS1 and/or TPS2 after the re-addition of phosphate to the starved cells was delayed for strains lacking either PHO85 or PHO80 as well as PHO81. Hence, it is possible that the uncoupling between trehalase activation and trehalose mobilization that we and others (Giots et al. 2003) observed in mutant strains affected in phosphate signalling may due to an overshoot of trehalose synthesis as compared to trehalose degradation during the first hours after phosphate re-supplementation to starved cells.

The effects on expression of the PDS genes appeared to be largely dependent on the Pho85–Pho80 complex. Although Pho4 is to date the best known physiological effector of the Pho85–Pho80 kinase, our data show that this transcription factor is not directly involved in transcriptional control of the PDS-driven genes but exerts its effect via the positive feedback loop towards Pho81 (Ogawa et al. 1995). In consequence, another effector must link the Pho85–Pho80 kinase activity to phosphate-dependent transcription regulation of the PDS-driven genes. The existence of an additional Pho85–Pho80 effector has already been suggested by Flick and Thorner, who noticed that the temperature-sensitive growth defect of a plc1Δ strain (Flick and Thorner 1993) could be suppressed either by overexpression of PHO81, by disruption of the PHO80 gene or by growing the cells in a low-phosphate environment (Flick and Thorner 1998). Since these suppressive effects were independent of the Pho4 transcription factor, it was concluded that the Pho85–Pho80 kinase inhibits a target which is also downstream of Plc1 and which is required for the proper response to nutrient levels and is involved in growth at elevated temperatures (Flick and Thorner 1998). Further support for the involvement of an additional Pho85–Pho80 effector comes from the data obtained with expression of truncated PHO81 alleles. Regulation of the activity of the Pho4 transcription factor in response to phosphate levels is known to be accomplished through Pho81-mediated regulation of Pho85–Pho80 kinase activity, for which a region of Pho81 containing only 80 amino acids was shown to be both necessary and sufficient (Schneider et al. 1994; Ogawa et al. 1995; Huang et al. 2001). In our studies, however, this Pho81 minimum domain is not sufficient for proper regulation of the PDS-driven genes in response to phosphate availability and thus other domains of Pho81 must be involved. Intriguingly, these domains must then be able to discriminate between the different effectors of the Pho85–Pho80 complex.

The observation that aberrant signalling of the PHO pathway clearly affected the expression of the PDS-driven genes pointed us to the specific involvement of Gis1, the transcription factor that binds the PDS elements. That Gis1 would be specifically required to induce gene expression in response to phosphate starvation may not be so surprising. Indeed, this factor was previously shown to be essential to induce expression when glucose becomes limiting and cells enter the diauxic shift, upon which they acquire stationary-phase characteristics. Therefore, it is feasible to assume that Gis1 is activated whenever cells encounter shortness of an essential nutrient. The activity of Gis1 is positively regulated by the protein kinases Rim15 and Sch9 (Reinders et al. 1998; Pedruzzi et al. 2000; Roosen et al. 2005) Interestingly, apart from being an activator of Gis1, Sch9 also acts as negative regulator of Rim15, indicating that Sch9 may function as a molecular buffer system that regulates the amplitude of Rim15-dependent responses. In addition, Rim15 is also under negative control of PKA and Tor1/2 and thus the kinase integrates signals from at least three nutrient-sensing cascades (Pedruzzi et al. 2003). In this study, we showed that combined deletion of SCH9 and PHO85 results in a severe growth defect, indicating that both protein kinases operate in parallel though converging pathways. Furthermore, we observed that growth of the pho85Δ mutant is hypersensitive to the addition of the immunosuppressant rapamycin (data not shown), as previously demonstrated (Huang et al. 2002), indicating that Pho85 and the yeast Tor kinases may also function in parallel, but converging pathways. These observations are consistent with Rim15 defining the point of convergence. In line with this, the deletion of RIM15 in a pho85Δ background did not lead to an additional growth defect, but in contrast partially alleviated the growth defect of a pho85Δ strain on glycerol medium. This places Rim15 indeed downstream and under the negative control of Pho85; and it suggests that one of the reasons why the pho85Δ cells fail to grow on non-fermentable carbon sources is because of an uncontrolled Rim15 activity. Consistently, a recent high-throughput protein-interaction analysis suggested that Rim15 and Pho85 may interact with each other in vivo (Ho et al. 2002), but this interaction awaits further confirmation. Nevertheless, our data already indicate that the nutritional-integrator function of Rim15 is even broader than previously described (Pedruzzi et al. 2003; Roosen et al. 2005), as we found that Rim15 is indeed required for trehalose accumulation and for correct transcriptional induction of stress-responsive genes during phosphate starvation. As with the pho81Δ mutant, the PDS-controlled genes SSA3 and GRE1 display a much stronger dependency on a functional Rim15 kinase than the STRE-controlled gene HSP26, a finding that supports a model in which Rim15 is under the negative control of the Pho85–Pho80 kinase (Fig. 8).
Fig. 8

Model for the involvement of the PHO regulatory pathway in phosphate-induced signalling and genetic interaction with other signalling components. When yeast cells are starved for phosphate, the PHO pathway activates the Pho4 transcription factor through Pho81-mediated inhibition of the Pho85–Pho80 kinase complex. This leads to induced expression of classic PHO genes. Importantly, the PHO81 gene is also induced, creating a positive feedback loop within the pathway. The Pho85–Pho80 kinase also appears to regulate the protein kinase Rim15, which functions downstream of PKA and Sch9 to orchestrate induction of stress-responsive genes, particularly those genes controlled by Gis1, the transcription factor known to bind the PDS elements. In addition, Pho81 controls the activity of (an) additional Pho85 complex(es), involved in the biosynthesis of the reserve carbohydrate trehalose in response to phosphate levels. See text for further details

Finally, it should be mentioned that induction of the PDS genes in a pho81Δ strain is not totally abolished, but only severely impaired in response to phosphate depletion. Therefore, these genes are subject to additional control allowing the bypass of the defect in Pho81-dependent inhibition of the Pho85–Pho80 kinase. Since this control becomes apparent only after prolonged starvation of the cells, it may be mediated by the lack of metabolic signals when cells are severely starved for phosphate, for instance depletion of ATP. Thus, yeast cells may have a biphasic response to phosphate starvation. In a first phase, the cells rapidly adapt to the lack of phosphate through activation of the PDS-driven genes via the PHO pathway. In a second, slower phase, the lack of metabolic signals generated after prolonged phosphate starvation triggers a stress signal, leading to induction of the PDS genes independent of the PHO pathway. Possibly, the second phase may depend on the stress-activated transcription factors Msn2 and/or Msn4, which are known to mediate part of the Rim15-dependent transcriptional responses redundantly to Gis1 (Cameroni et al. 2004).

Acknowledgements

This work was supported by a fellowship from the Fund of Scientific Research in Flanders (FWO-Vlaanderen) to E.S. and by grants from FWO-Vlaanderen and the research fund of K.U. Leuven

Copyright information

© Springer-Verlag 2005