Biochemical Journal

Research article

Ypp1/YGR198w plays an essential role in phosphoinositide signalling at the plasma membrane

Chao Zhai, Kuoyu Li, Valentini Markaki, John P. Phelan, Katherine Bowers, Frank T. Cooke, Barry Panaretou

Abstract

Phosphoinositide signalling through the eukaryotic plasma membrane makes essential contributions to many processes, including remodelling of the actin cytoskeleton, vesicle trafficking and signalling from the cell surface. A proteome-wide screen performed in Saccharomyces cerevisiae revealed that Ypp1 interacts physically with the plasma-membrane-associated phosphoinositide 4-kinase, Stt4. In the present study, we demonstrate that phenotypes of ypp1 and stt4 conditional mutants are identical, namely osmoremedial temperature sensitivity, hypersensitivity to cell wall destabilizers and defective organization of actin. We go on to show that overexpression of STT4 suppresses the temperature-sensitive growth defect of ypp1 mutants. In contrast, overexpression of genes encoding the other two phosphoinositide 4-kinases in yeast, Pik1 and Lsb6, do not suppress this phenotype. This implies a role for Ypp1 in Stt4-dependent events at the plasma membrane, as opposed to a general role in overall metabolism of phosphatidylinositol 4-phosphate. Use of a pleckstrin homology domain sensor reveals that there are substantially fewer plasma-membrane-associated 4-phosphorylated phosphoinositides in ypp1 mutants in comparison with wild-type cells. Furthermore, in vivo labelling with [3H]inositol indicates a dramatic reduction in the level of phosphatidylinositol 4-phosphate in ypp1 mutants. This is the principal cause of lethality under non-permissive conditions in ypp1 mutants, as limiting the activity of the Sac1 phosphoinositide 4-phosphate phosphatase leads to restoration of viability. Additionally, the endocytic defect associated with elevated levels of PtdIns4P in sac1Δ cells is restored in combination with a ypp1 mutant, consistent with the opposing effects that these two mutations have on levels of this phosphoinositide.

  • lipid kinase
  • phosphoinositide
  • Saccharomyces cerevisiae
  • signalling
  • Stt4
  • Ypp1

INTRODUCTION

PPIn (polyphosphoinositide) second messengers regulate numerous cellular processes in eukaryotes, including organization of the actin cytoskeleton, vesicle trafficking and signalling via MAPK (mitogen-activated protein kinase) cascades. There has been significant progress in our understanding of how the rapid turnover of PPIns is controlled both spatially and temporally (reviewed in [1]). Of the seven PPIn species recognized, four have been identified in the budding yeast Saccharomyces cerevisiae. One of these, PtdIns4P, is synthesized by three distinct and highly conserved PI4Ks (phosphoinositide 4-kinases): Pik1 [2,3], Stt4 [4] and Lsb6 [57]. Pik1 is found at the Golgi and in the nucleus, Stt4 localizes to the plasma membrane, and Lsb6 is found at the plasma membrane and the vacuolar membrane. Both Stt4 and Pik1 are essential, synthesizing a discrete location-specific pool of PtdIns4P. Consequently, deletion of one cannot be compensated for by overexpression of the other, even though both proteins catalyse the same biochemical reaction [8]. Lsb6, the third PI4K, is non-essential; its contribution to total cellular PtdIns4P in vegetative cells is uncertain, given that over 90% of PtdIns4P is lost in a stt4tspik1ts temperature-sensitive (ts) double mutant within 1 h of shift to the non-permissive temperature [9]. To date, the only phenotype observed in lsb6Δ cells is a reduction in endosome motility, an investigation prompted by the interaction of this lipid kinase with Las17 (the yeast orthologue of Wiskott–Aldrich syndrome protein), which is essential for endocytosis. In the lsb6Δ background, expression of a catalytically inactive Lsb6 mutant could restore wild-type levels of endosome motility as effectively as expressing wild-type Lsb6. Therefore the role played by Lsb6 in endosome motility is independent of its lipid kinase activity [10].

Pik1 localizes to the Golgi, and the PtdIns4P synthesized by this enzyme is essential for Golgi-to-plasma membrane vesicular trafficking [9,11,12]. In addition, Pik1 shuttles between the cytoplasm and the nucleus. Mutant alleles of PIK1 that restrict location of the protein to membranes of the secretory pathway do not sustain viability of a pik1Δ strain. Therefore the PtdIns4P generated in the nucleus by Pik1 is essential for cell viability, although the nuclear-specific role of PtdIns4P remains unknown [13]. Stt4 localizes to the plasma membrane, in part via an interaction with Sfk1, a membrane-spanning protein [14]. The PtdIns4P pool generated by Stt4 is required for aminophospholipid transport from the endoplasmic reticulum to the Golgi [15], organization of the actin cytoskeleton, maintenance of vacuole morphology [9] and cell wall integrity [14]. Some of these roles are due to PtdIns4P acting as a precursor for PtdIns(4,5)P2 synthesis, which is catalysed by the plasma-membrane-associated phosphoinositide 4-phosphate 5-kinase Mss4 [16]. For instance, plasma membrane PtdIns(4,5)P2 activates the Rho/Pkc1-mediated MAPK cascade that maintains cell wall integrity. More specifically, generation of PtdIns(4,5)P2 recruits the guanine-nucleotide-exchange factor Rom2 to the plasma membrane via its phosphoinositide PH (pleckstrin homology)-binding domain, and is an absolute requirement for activation of the Rho1 GTPase [14].

There has been substantial progress in characterizing mechanisms that terminate PtdIns4P-dependent processes, through identification of lipid phosphatases with enzymatic activity towards phosphoinositides ([17] and reviewed in [1]). In contrast, mechanisms that regulate the generation of PtdIns4P by Stt4 are poorly defined. Insight into mechanisms that regulate lipid kinases has come from identifying accessory factors. For example, Frq1 and Vps15 are required for both membrane localization and optimal activity of the PI4K Pik1 and the PI3K (phosphoinositide 3-kinase) Vps34 respectively [18,19].

The protein encoded by the essential ORF (open reading frame) YGR198W was identified as a binding partner of Stt4 [20]. This ORF was recently named YPP1 (sYnuclein-Protective Protein 1), because it acts as a multicopy suppressor of the toxic effect of expressing a human α-syn (α-synuclein) mutant in yeast [21]. We have tested the possibility that Ypp1 is an accessory factor for Stt4. We characterize the phenotypes of ts mutants of YPP1, and we show that they are similar to phenotypes associated with stt4ts mutants. Moreover, overexpression of STT4 suppresses the ts growth defect of ypp1 mutants. We go on to show that total levels of PtdIns4P, as well as levels of plasma membrane PtdIns4P, are low in ypp1ts mutants. Together, these data show for the first time that Ypp1 makes an essential contribution to Stt4-dependent production of PtdIns4P at the plasma membrane.

MATERIALS AND METHODS

Strains, plasmids and media

Yeast strains used in the present study are listed in Table 1. Plasmids are described in the appropriate sections below. Strains were grown in either rich medium [YPD: 1% (w/v) yeast extract/2% (w/v) peptone/2% (w/v) glucose] or synthetic minimal (SD) medium with appropriate supplements for plasmid maintenance [22]. Gene deletions were performed by replacing the target ORF with hph (conferring resistance to hygromycin B), via one-step gene replacement, as described previously [23]. Integration at the correct locus was confirmed by PCR using appropriate primers [24].

View this table:
Table 1 Yeast strains used in the presnt study

EUROSCARF, European Saccharomyces cerevisiae Archive for Functional Analysis.

Isolation of YPP1 conditional alleles

The heterozygous YPP1/ypp1Δ diploid Y24828 was obtained from EUROSCARF (European Saccharomyces cerevisiae Archive for Functional Analysis) and transformed with a centromeric URA3 S. cerevisiae/Escherichia coli shuttle vector (pYC3P) bearing the YPP1 promoter (a 0.45 kb fragment upstream of the ATG) fused to the YPP1 ORF (a 2.5 kb fragment), followed by the ADH1 transcriptional terminator (a 0.4 kb fragment). The YPP1 promoter and ORF from this vector were then subcloned into a centromeric LEU2 S. cerevisiae/E. coli shuttle vector (again bearing the ADH1 terminator), to give pYC1P. Strain Y24828 was transformed with pYC3P and then sporulated, followed by subsequent dissection of tetrads and selection of a ypp1Δ haploid (strain ZC0) with viability maintained by the presence of the URA3 vector bearing YPP1 (pYC3P). Mutations in YPP1 were generated by error-prone amplification of the YPP1 ORF using pYC3P as template, as described previously [25]. The oligonucleotide primers were designed so that the 5′ primer overlaps the 3′ end of the YPP1 promoter, and the 3′ primer overlaps the 5′ end of the ADH1 transcriptional terminator (15 bp overlaps in both cases). Vectors bearing the mutant ypp1 alleles were generated by transforming strain ZC0 with (i) the mutant ypp1 PCR products plus (ii) a large linear fragment from pYC1P lacking the wild-type YPP1. The two fragments recombine with each other in vivo, owing to the 15 bp overlaps between them. The transformants, selected on SD medium lacking both leucine and uracil, possess the pYC3P URA3 vector bearing wild-type YPP1, and a LEU2 vector bearing mutated ypp1. More than 4000 transformants were picked and incubated overnight at 25 °C in 384-well plates containing SD medium lacking leucine. Those transformants bearing ts mutants of ypp1 were isolated as described previously [26]. To ensure that these ts phenotypes were attributed to mutated ypp1, plasmids from the strains were recovered, transformed into the ZC0 strain and re-screened for the ts phenotype. In this way, a series of isogenic ypp1Δ haploid strains were generated, possessing vector-borne copies of wild-type YPP1 (ZC1, Table 1) or ypp1ts alleles (ZC2–ZC8, Table 1). The coding regions of ypp1ts mutants were sequenced.

Sensitivity to cell-wall-damaging agents

Exponentially growing cultures were adjusted to equal cell density (107 cells/ml) and four successive 6-fold serial dilutions were spotted on YPD, or YPD containing the desired concentration of sorbitol, caffeine or SDS. The plates were incubated at the temperatures specified in the Figure legends.

[3H]Inositol labelling of S. cerevisiae and extraction of labelled phosphoinositides

Phosphoinositides were labelled in vivo as described previously [27]. Briefly, strains were grown at 25 °C in SD medium without inositol plus appropriate supplements, in the presence of 15 μCi/ml [3H]inositol (GE Healthcare). Once cultures had reached exponential phase, they were split into two aliquots: one was incubated for a further 1 h at 25 °C, and the other was shifted to 37 °C for 1 h. Cells were killed with an equal volume of ice-cold methanol and sedimented at 3000 g for 5 min at 4 °C. Lipids were extracted and then deacylated, followed by recovery of the glycerophosphoinositols that were resolved by anion-exchange HPLC on a Partisphere 5-SAX column (Whatman).

Assessing multicopy suppression by selected ORFs

Candidate ORFs were overexpressed from the vector pY195M, which is a modification of the episomal URA3 vector YEplac195 [28]. The vector bears the MET25-inducible promoter and the transcription terminator from PGK1. Candidate ORFs were amplified from S. cerevisiae genomic DNA and ligated into pY195M. The resulting vectors were transformed into the ZC1 strain, and isogenic strains containing vector-borne ypp1ts alleles. Transformants were transferred to SD plates lacking uracil, leucine and methionine, to induce overexpression of the candidate ORF.

Two-hybrid analysis

The YPP1 and STT4 ORFs were amplified from yeast genomic DNA, including appropriate restriction sites in the oligonucleotides. The YPP1 PCR product was subcloned into pGAD-C1 [29] to give pAD-YPP1; STT4 was cloned into pGBDU-C1 [29] to give pBD-STT4. Vectors pAD-YPP1 and pBD-STT4 were transformed into strains PJ69-4a and PJ69-4α respectively [29]. Transformants (a and α) were mated on YPD medium for 8 h at 30 °C, and then diploids selected on SD medium lacking both leucine and uracil. To assess Stt4–Ypp1 interaction, the diploids were inoculated on to SD medium lacking leucine, uracil and histidine, and in the presence of 1 mM 3-AT (3-aminotriazole) (Sigma).

Assessing the activation state of the MAPK cell wall integrity pathway

Exponentially growing cultures were either heat-shocked or treated with 10 mM caffeine (as specified in the Figure legends). Preparation of total yeast protein extracts was as described previously [30]. Western blots were probed with either (i) (Thr202/Tyr204)-p44/p42 MAPK antiserum (New England Biolabs) which specifically recognizes the dually Thr190/Tyr192-phosphorylated (activated) Slt2 in yeast, or (ii) anti-Slt2 (Santa Cruz Biotechnology). Activity of Rlm1 transcription factor activity was assessed using p2-RLM1 a YIL117c promoter-LacZ reporter plasmid [31].

Localization of actin and 4-phosphorylated PtdIns

Cells were grown to mid-exponential phase at 25 °C, followed by further incubation at 25 °C or shift to 37 °C, then fixed in formaldehyde and stained with rhodamine–phalloidin (Molecular Probes) to visualize actin [32]. The cellular distribution of 4-phosphorylated PtdIns was assessed by transforming strains with pTL511, a centromeric URA3 plasmid expressing a tandem dimer of the PH domain of Osh2 fused to GFP (green fluorescent protein), under control of the constitutively active PHO5 promoter [33]. Before observation by fluorescence microscopy, exponential-phase cells were either maintained at 25 °C or shifted to a strongly non-permissive temperature (39 °C) for 15 min, as described previously [33].

Endocytosis of Ste3–GFP and Fui1–GFP

Strains were transformed with pFL38GalFUI1-GFP, a centromeric URA3 vector that enables galactose-inducible synthesis of an Fui1–GFP fusion [34]. For localization of Fui1–GFP, cells were grown to exponential phase in 2% galactose and 0.02% glucose at 25 °C, to induce synthesis of Fui1–GFP. This was followed by addition of 2% glucose, to repress Fui1–GFP synthesis. Cells were then maintained at 25 °C or up-shifted to 37 °C for 30 min, followed by addition of uridine (80 μg/ml) to induce endocytosis. Samples were removed for observation at the intervals described in the Figure legends, and endocytosis was stopped by adding ice-cold energy poisoning buffer (SD medium without carbon source, 10 mM NaF and 10 mM sodium azide) as described previously [35]. GFP-fusions were visualized by fluorescence microscopy.

RESULTS

Isolation of ypp1ts alleles

To determine the function of Ypp1, we generated conditional lethal alleles by PCR-mediated mutagenesis. Seven ypp1ts mutants were isolated, three of which exhibited moderate growth at 37 °C, with the remaining four displaying no growth at all at this temperature (Figure 1A). Three of the latter, exhibiting a robust ts phenotype, were selected for further work and the Ypp1 coding regions were sequenced. Most error-prone PCR techniques are not truly random as they yield more transitions than transversions. The method we used, however, removes the bias [25]. The amino acid substitutions were: ypp1-35ts, G60R, N79I, N102S, E189A, L220P, K329R and S336P; ypp1-39ts, P2A, K197E, F218L, L442S, L453P and C463R; and ypp1-41ts, F343V, V538M, W571R, K577E, I584V, K656E and S678L.

Figure 1 Phenotypes displayed by conditional mutants of YPP1

(A) Osmoremedial temperature-sensitivity. YPP1 wild-type (ZC1) and conditional mutant strains (ZC2–ZC8) were streaked on YPD plates with and without 1 M sorbitol, and incubated for 3 days at 25 or 37 °C. (B) Sensitivity of ypp1ts mutants to caffeine and SDS. YPP1 wild-type and ypp1ts mutants were grown at 25 °C to exponential phase, and diluted to equal cell density. Six-fold serial dilutions were spotted across YPD medium, or YPD medium containing the indicated concentrations of SDS, sorbitol, caffeine and caffeine plus sorbitol, followed by incubation at 30 °C for 3 days.

Phenotypes of ypp1ts mutants are similar to those displayed by stt4ts and mss4ts mutants

The PtdIns(4,5)P2 generated by Stt4 and Mss4 plays an essential role in the cell wall integrity pathway; inactivation of either kinase leads to cell lysis [9,16]. As a result, the lack of growth displayed by stt4ts cells at 37 °C is suppressed in the presence of 1 M sorbitol [14], a property shared by all seven ypp1ts mutants (Figure 1A), albeit to various extents. This osmoremedial phenotype is displayed clearly by six of the mutants, and is poorly displayed (but still detectable) by ypp1-35ts (Figure 1A).

The ypp1ts mutants also display other phenotypes characteristic of defects in the cell wall integrity MAPK cascade, namely caffeine-sensitivity, which can be suppressed in the presence of an osmotic stabilizer such as sorbitol, and sensitivity to low levels of SDS (Figure 1B). Of the three mutants tested, ypp1-35ts was the most sensitive to caffeine and SDS. This was expected, because this mutant displayed a robust temperature-sensitive phenotype, which was poorly suppressed by osmotic stabilization of medium (Figure 1A).

Plasma membrane phosphoinositides generated by the sequential action of Stt4 and Mss4 have also been implicated in the organization of the actin cytoskeleton. In the wild-type, actin distribution is polarized, as indicated by the concentration of cortical actin in the bud (Figure 2). A shift from 25 to 37 °C causes a heat-induced disorganization of the actin cytoskeleton, which is transient in the wild-type, persisting for up to 2 h, followed by recovery in organization of the cytoskeleton. In contrast with wild-type cells, however, this disturbance in actin organization persisted for at least 3 h after temperature up-shift in ypp1ts mutants, with random cortical actin patches evident throughout both mother and daughter cells (Figure 2). Such irregular organization of actin was even noticeable at permissive temperature in the ypp1-35ts mutant (Figure 2). An identical phenotype is displayed by both stt4ts and mss4ts mutants [9,16].

Figure 2 Cells lacking Ypp1 activity fail to reorganize the actin cytoskeleton during heat shock

YPP1 (ZC1) and ypp1ts mutants (ZC6 and ZC8) were grown to exponential phase at 25 °C and then shifted to 37 °C for 3 h. Cells were fixed with formaldehyde. Actin cables and cortical actin patches were stained with rhodamine–phalloidin (A); corresponding phase-contrast images (P) are also presented.

STT4 is a multicopy suppressor of ypp1ts

Overexpression of the PI4K encoded by STT4 is sufficient to restore growth of ypp1-39ts and ypp1-41ts at 37 °C (Figure 3). This confirms the possibility of biological interaction between these two proteins inferred by both their physical interaction [20], and the similarities in phenotypes that arise when the function of either protein is compromised. Growth of ypp1-35ts at 37 °C, however, was not restored. This is probably due to the extent to which Ypp1 function is compromised in this mutant; of the three mutants tested, ypp1-35ts was the most sensitive to SDS and caffeine (Figure 1B), and temperature-sensitivity was barely suppressed by osmotic stabilization (Figure 1A). In S. cerevisiae, there is one other essential PI4K: the Golgi-localized Pik1. Even though Stt4 and Pik1 both produce the same second messenger, deletion of one kinase cannot be rescued by overexpression of the other because the enzymes synthesize essential pools of PtdIns4P on different intracellular membranes, and PtdIns4P is not freely diffusible through the cell [8]. At 37 °C, the ypp1-39ts and ypp1-41ts mutants were rescued by overexpressing STT4, but not by PIK1 (Figure 3) or by LSB6, the third PI4K in S. cerevisiae (results not shown). This implies a role for Ypp1 in Stt4-dependent events at the plasma membrane, as opposed to a general role in the overall metabolism of PtdIns4P. Growth at 37 °C was not restored in any of the mutants when the phosphoinositide 4-phosphate 5-kinase Mss4 was overexpressed (results not shown). Rescue by Stt4 overexpression implies that levels of PtdIns4P are too low to support viability of ypp1ts mutants at 37 °C.

Figure 3 STT4 is a multicopy suppressor of ypp1ts

The YPP1 wild-type and ypp1ts mutants (ZC1 and ZC6–ZC8) were transformed with multicopy vectors bearing the ORFs indicated in the central panel under the control of the MET25 promoter. Gene expression was induced by inoculating on SD medium lacking methionine followed by incubation for 3 days at 25 or 37 °C.

The physical interaction between Stt4 and Ypp1 was first identified by affinity-capture MS [20]. We confirmed this by the yeast-two hybrid interaction trap, using a BD (DNA-binding domain)–Stt4 fusion as bait and an AD (activation domain)–Ypp1 fusion as prey (Figure 4A). Furthermore, the interaction between wild-type Ypp1 and Stt4 is weakened at high temperature (Figures 4A and 4B), as is evident by the lower growth rate at 37 °C compared with 25 °C, on media that select for the Ypp1–Stt4 interaction. In addition, we went on to show that interaction between Stt4 and the ypp1-35ts, ypp1-39ts and ypp1-41ts mutant proteins, was lost at 37 °C. Indeed, the interaction between these mutant proteins and Stt4 was so destabilized that it could not be detected even at 25 °C (Figure 4A). We anticipated, however, that interaction at 25 °C might be displayed between Stt4 and one of the less robust ts mutants, such as ypp1-30ts (Figure 1A). This was found to be the case, as a weak interaction between Stt4 and ypp1-30ts was detectable at 25 °C (Figure 4B).

Figure 4 Interaction with Stt4 is abrogated by ypp1ts mutations

Two-hybrid interaction in diploid his3200/his3200 auxotrophs expressing BD–Stt4 plus AD–Ypp1 or the indicated AD–ypp1ts mutants. Cells were incubated at 25 or 37 °C on SD medium lacking histidine. Interaction between AD- and BD-fusions is scored by growth, as this leads to expression of a HIS3 reporter. All plates contained 1 mM 3-AT, which inhibits basal activity of the HIS3 reporter product.

At the restrictive temperature, ypp1ts cells display diminished levels of plasma-membrane-associated PtdIns4P

Cells lacking genomic YPP1, but expressing wild-type or mutant ypp1, were transformed with a vector expressing a GFP–PHOsh2 (PH domain of Osh2) dimer. PHOsh2 specifically recognizes PtdIns4P and PtdIns(4,5)P2 and lacks an accessory site that would lead to association with a specific cellular membrane, allowing it to be exploited as an unbiased reporter for PtdIns4P and PtdIns(4,5)P2 [33]. In wild-type cells, at both 25 and 39 °C, this reporter localized to both punctate internal structures and the plasma membrane (Figure 5A), in agreement with previously reported localization for this sensor [33]. In the ypp1-41ts mutant, however, 15 min of incubation at 39 °C led to a rapid loss of plasma membrane localization of the reporter. The effect was even more severe in ypp1-35ts cells, as GFP–PHOsh2 dimer localization at the plasma membrane was reduced significantly, even at the permissive temperature (Figure 5A).

Figure 5 At restrictive temperature, ypp1ts mutants display a diminished level of plasma-membrane-associated PtdIns4P

(A) Using a PH domain sensor, YPP1 wild-type and ypp1ts mutants (ZC1, ZC6 and ZC8) were transformed with a vector bearing a tandemly repeated dimer of PHOsh2 bracketed by GFP. Strains were grown at 25 °C to exponential phase. Before imaging, an aliquot of each culture was shifted to a strongly non-permissive temperature (39 °C) for 15 min. (B) [3H]Inositol labelling. YPP1 wild-type (ZC1) and the ypp141ts mutant (ZC8) were incubated with [3H]inositol at 25 °C to label all inositol-based lipids to near steady-state equilibrium. Subsequently, cultures were split into two, with one aliquot maintained at 25 °C and the other shifted to 37 °C for 1 h. [3H]Inositol-containing lipids were extracted and deacylated. The glycerophosphoinositols (GroP) were separated by HPLC and quantified by scintillation counting. All data were corrected to control [3H]PtdIns levels, which were: wild-type (W+), 941361 c.p.m.; ypp141ts, 1011448 c.p.m. Results are means±S.E.M. (n=3).

At the restrictive temperature, ypp1-41ts cells display diminished levels of PtdIns4P

Wild-type and ypp1-41ts cells were incubated with [3H]inositol at 25 °C for up to six generation times, to label all inositol-based lipids to near steady-state equilibrium. These exponential-phase cultures were split into two, with one aliquot shifted to 37 °C for 1 h, and the other maintained at 25 °C. Subsequent analysis of labelled lipids by HPLC allowed us to monitor the levels of glycerophosphoinositols derived from all four PPIns made by S. cerevisiae. As expected, levels of PtdIns(3,5)P2 were almost identical in mutant and wild-type cells (Figure 5B).

At the permissive temperature, the level of PtdIns4P in ypp1-41ts cells was almost 60% lower than that seen in wild-type cells (Figure 5B). A shift to 37 °C for 1 h led to a 50% decline in PtdIns4P in the wild-type. Under these conditions, the level of PtdIns4P in ypp1-41ts cells did not fall appreciably in comparison to levels in ypp1-41ts cells at 25 °C, although levels of PtdIns4P in this mutant were already low before the up-shift in temperature.

The changes in PtdIns4P are reflected by the in vivo protein–protein interaction data because the interaction between the Stt4 lipid kinase and wild-type Ypp1 is weaker at high temperature (Figures 4A and 4B). If Ypp1 is required for optimal function of Stt4, then an up-shift in temperature (diminishing the interaction between the two proteins) would account for the decrease in synthesis of PtdIns4P observed at 37 °C in wild-type cells (Figure 5B). A consequence of reduced interaction between wild-type Ypp1 and Stt4 would mean that the effect of mutant Ypp1 on Stt4 function would be less noticeable at 37 °C than at 25 °C. This explains why, at 37 °C, levels of PtdIns4P are only slightly lower in the ypp1-41ts mutant than they are in the wild-type.

Levels of PtdIns(4,5)P2 were lower in ypp1-41ts cells relative to the wild-type, at both permissive and non-permissive temperatures, although the decline in levels was not as great as that observed for PtdIns4P. This was not surprising, as a similar result was reported when labelled phosphoinositides were assessed in stt4ts cells, namely a decline in levels of PtdIns(4,5)P2 that was not as great as the decline in PtdIns4P levels [14].

At the permissive temperature, levels of PtdIns3P were significantly lower in the ypp1-41ts mutant when compared with the wild-type, although levels of this phosphoinositide declined to similar levels in wild-type and mutant, following heat shock (Figure 5B).

Deletion of SAC1 restores viability of ypp1ts mutants under non-permissive conditions

PtdIns4P is essential for cell viability, but an up-shift in temperature leads to a decrease in its synthesis (Figure 5B). If the lower level of PtdIns4P synthesized at 25 °C in ypp1ts mutants is the principal cause of lethality on temperature up-shift, then limiting the activity of phosphatases that hydrolyse PtdIns4P should raise PtdIns4P levels of this phosphoinositide, thereby restoring viability. The only lipid phosphatases with enzymatic action towards PtdIns4P are those containing the Sac1 homology domain. There are five members of this family in yeast: Sac1, Fig4 and the three synaptojanin-like phosphatases, Sjl1/Inp51, Sjl2/Inp52 and Sjl3/Inp53 [17,36]. Of these, Sac1 is the predominant 4-phosphatase in yeast, as cells lacking the corresponding gene have up to 12-fold more PtdIns4P than wild-type cells [17,37,38]. Moreover, Stt4 (rather than Pik1 or Lsb6) synthesizes the majority of PtdIns4P that accumulates in sac1Δ cells [35,37]. Loss of Sac1 activity restored viability at 37 °C to all three ypp1ts mutants tested (Figure 6), confirming further that the essential function of Ypp1 is to facilitate PtdIns4P synthesis via activation of Stt4.

Figure 6 Viability of ypp1ts mutants under non-permissive conditions is restored in a sac1Δ background

YPP1 wild-type (ZC1), ypp1ts mutants (ZC6–ZC8) and corresponding isogenic sac1Δ strains (ZC21–ZC24) were grown at 25 °C to exponential phase, and diluted to equal cell density. Six-fold serial dilutions were spotted across YPD medium, followed by incubation for 3 days at the temperatures indicated.

The MAPK cell wall integrity pathway operates normally in the ypp1-35ts mutant

Sequential action by Stt4 and Mss4 generates PtdIns(4,5)P2 which tethers the guanine-nucleotide-exchange factor Rom2 to the plasma membrane, via interaction with the Rom2 PH domain. This activates Rho1, which in turn activates Pkc1, initiating a phosphorylation cascade leading to activation of the MAPK Slt2 (reviewed in [39]). Compromising the function of Stt4 causes a decrease in the level of PtdIns4P and, in turn, causes destabilization of the cell wall [14]. The same is true when function of Ypp1 is compromised (Figures 1B and 5B), so one possibility is that these mutants fail to activate the MAPK cell wall integrity pathway. We investigated whether this was the case. As expected, Slt2 in wild-type cells is unphosphorylated at 25 °C, but is phosphorylated in the presence of caffeine, which is known to destabilize the cell wall (Figure 7A). Of the three mutants tested, ypp1-35ts was selected for investigating the activity of components involved in the cell wall integrity pathway, because this mutant was the most sensitive to caffeine (Figure 1B). Slt2 was phosphorylated normally in response to caffeine in ypp1-35ts cells (Figure 7A). Phosphorylated Slt2 activates Rlm1, the major transactivator of genes involved in cell wall maintenance [39]. However, a cell wall integrity defect may still result if there is a defect downstream of Slt2, even though signalling through the MAPK cascade itself appears to be normal in ypp1-35ts shifted to the non-permissive temperature. Given that there are mutants which are defective in Rlm1-mediated transcription, even when Slt2 is activated normally [40], we decided to measure transcription of a LacZ reporter controlled by the Rlm1-dependent promoter of YIL117c. In the presence of caffeine, Rlm1-mediated transcription was stimulated to the same extent in wild-type and ypp1-35ts cells (Figure 7B). Consequently, the cell wall integrity defect in ypp1-35ts is not due to defective signalling downstream of Pkc1. Moreover, overexpression of components of this cascade, including the MAPK kinase kinase Bck1, either of the redundant MAPK kinases (Mpk1 or Mpk2), the MAPK Slt2 or Pkc1 itself, all failed to rescue the no-growth defect at 37 °C of the three ypp1ts mutants (results not shown).

Figure 7 The MAPK cell wall integrity pathway is activated in the ypp135ts mutant, in response to caffeine

(A) YPP1 (ZC1) and the ypp135ts mutant (ZC6) were grown to exponential phase at 25 °C and then incubated for 1 h in the presence or absence of 10 mM caffeine. Cells were harvested, and levels of dually phosphorylated Slt2 (indicated as Slt2-P) and total Slt2, were detected by Western blot using appropriate antisera. (B) The same strains were transformed with a vector bearing a YIL117c-promoter LacZ fusion. LacZ activity was assayed 1 h after the addition of 10 mM caffeine. YIL117cLacZ is a reporter for Rlm1-dependent transcription. The y-axis is LacZ Miller units. Results are means±S.E.M. (n=3).

Endocytosis of plasma membrane proteins is unaffected in the ypp1-35ts mutant

Recently, it has been shown that overexpression of YPP1 suppresses the toxicity of the human A30P α-syn mutant that is associated with early-onset Parkinson's disease. Elevated levels of Ypp1 lead to endocytosis and degradation of A30P α-syn in the yeast vacuole [21]. However, α-syn is found in the cytosol and is not native to yeast, and, as such, it is not a typical substrate for the endocytic process. We assessed the possible role Ypp1 might play in endocytosis by following the localization of yeast protein substrates that use the endocytic route for their delivery to the vacuole.

To examine the role of Ypp1 in endocytosis, we monitored the trafficking of the uridine permease Fui1. The rate of constitutive endocytosis of this transporter is relatively low, which permits visualization of this protein at the plasma membrane (unlike the a factor receptor, Ste3, for example, which undergoes rapid constitutive internalization). However, the rates of endocytosis and subsequent degradation of Fui1 in the vacuole, are increased significantly in the presence of uridine [34]. We monitored the intracellular trafficking of an Fui1–GFP fusion, synthesis of which was under the control of the inducible GAL10 promoter [34]. Cells were grown to exponential phase in galactose to induce synthesis of Fui1–GFP, followed by addition of glucose (to block permease synthesis) and a temperature shift to 37 °C. Cells were maintained at this temperature for 30 min to induce the temperature-sensitive phenotype. At this stage, similar levels of Fui1–GFP were seen at the plasma membrane in both wild-type and ypp1-35ts mutant cells (Figure 8). Also, vacuolar fluorescence was clearly observed because of the relative resistance of GFP to vacuolar hydrolases [34]. Uridine was then added to the cells, and the temperature was maintained at 37 °C. As can be seen in Figure 8, Fui1–GFP staining is lost from the cell surfaces of both wild-type and ypp1-35ts cells after incubation for 15 or 60 min with uridine. Therefore the ypp1-35ts mutation has no dramatic affect on the overall endocytosis rate of Fui1–GFP. To confirm that we were looking at an endocytic process, we assessed the localization of the Fui1–GFP fusion in end3Δ cells treated in the same way. End3 is essential for the internalization step of endocytosis [32]. As expected, the Fui1–GFP in the end3Δ background was not internalized, remaining at the plasma membrane (Figure 8). Additionally, we saw similar results when we monitored another endocytic cargo, Ste3–GFP, in wild-type and ypp1-35ts cells (results not shown), so our data would suggest that low PtdIns4P levels do not inhibit endocytosis.

Figure 8 Endocytosis of Fui1–GFP promoted by uridine at 37 °C

Cells producing Fui1–GFP under control of the GAL10 promoter were grown at 25 °C to exponential phase in standard defined medium with galactose. Glucose (2%) was added to repress synthesis of the fusion, and cells were up-shifted to 37 °C. After 30 min at 37 °C, uridine was added to induce endocytosis. Cells were visualized before the addition of uridine, and at the intervals indicated after addition of uridine. Identical fields are shown under fluorescence (Fui1–GFP) and DIC (differential interference contrast) optics.

In the wild-type and ypp1-35ts cells, we noted that levels of internalized Fui1–GFP after the 30 min heat shock, but before addition of uridine, were particularly high (Figure 8). This was expected, because heat shock decreases the half-life of membrane transporters through stimulation of endocytosis [41,42].

As well as stimulating endocytosis of Fui1, uridine also stimulates exocytic trafficking of this permease by inducing direct traffic from the Golgi to the endosomal system, bypassing the plasma membrane. This type of traffic, however, did not contribute to the changes in localization of Fui1–GFP we observed, because we repressed Fui1–GFP synthesis by addition of glucose, before heat shock and subsequent addition of uridine. Therefore any changes in localization of Fui1–GFP caused by heat shock or uridine are only attributable to endocytic trafficking. This is confirmed by the lack of vacuolar staining in the end3Δ background, which is blocked for the early stages of endocytosis, but not deficient in exocytic traffic [34].

The induction of endocytosis by heat shock meant that subtle changes caused by compromised Ypp1 function may not be observed at 37 °C. For this reason, we assessed localization of Fui1–GFP at 25 °C. As expected, the rate of endocytosis in response to addition of uridine was much lower at 25 °C than at 37 °C. In wild-type cells, there was little change in levels of plasma-membrane-associated Fui1–GFP within 15 min of addition of uridine (Figure 9), in contrast with a complete lack of plasma-membrane-associated transporter at the same stage at 37 °C (Figure 8). At 25 °C, 1 h after the addition of uridine, all of the Fui1–GFP had been internalized. However, there was no difference in Fui1–GFP internalization between the wild-type and ypp1-35ts mutant (Figure 9). Even though 25 °C is a permissive temperature for ypp1ts mutants, the levels of PtdIns4P are significantly lower in these mutants in comparison with levels in wild-type cells (Figure 5B). On the other hand, it has been reported that elevated levels of PtdIns4P, which are found in sac1Δ cells, lead to impaired rates of endocytosis [35]. This was true for endocytosis of Fui1–GFP, with plasma-membrane-associated transporter clearly visible in sac1Δ cells, even after 60 min of exposure to uridine, in contrast with the complete internalization of the transporter in wild-type cells (Figure 9). Combining the sac1Δ with ypp1-35ts restored the wild-type kinetics of Fui1–GFP internalization (Figure 9), which is consistent with our data showing the opposing effects of sac1Δ and ypp1-35ts on PtdIns4P levels. In agreement with this is the compensation of the endocytic defect in sac1Δ cells in a background that is compromised for function of the PI4K encoded by STT4 [35]. Not surprisingly, levels of PtdIns4P are restored to near wild-type levels in an stt4tssac1ts double mutant [37].

Figure 9 Endocytosis of Fui1–GFP promoted by uridine at 25 °C

Cells producing Fui1–GFP under control of the GAL10 promoter were grown at 25 °C to exponential phase in standard defined medium with galactose. Glucose (2%) was added to repress synthesis of the fusion. After 30 min at 25 °C, uridine was added to induce endocytosis. Cells were visualized before the addition of uridine (0 min), and at the intervals indicated after addition of uridine. Identical fields are shown under fluorescence (Fui1–GFP) and DIC (differential interference contrast) optics.

DISCUSSION

Biological interaction between Stt4 and Ypp1

A PSI-BLAST search revealed six putative orthologues of Ypp1 in humans, with one of them, TTC7B (Q86TV6), sharing 15% amino acid sequence identity with Ypp1 [21]. Use of algorithms such as Prosite and Pfam, does not identify any recognizable domains in Ypp1. However, the use of a recently developed search tool [43] identifies two putative TPR (tetratricopeptide repeat) protein–protein interaction domains: one towards the N-terminus and the other towards the C-terminus. We characterized the role played by Ypp1 in cell physiology by generating conditional lethal alleles of the corresponding gene.

A TAP (tandem affinity purification) tag screen revealed a physical interaction between the PI4K Stt4 and Ypp1 [20], which we confirmed using a two-hybrid assay. Stt4 is found at the plasma membrane [14]. Ypp1 localizes nearby, in a punctate pattern around the periphery of the cell [21,44]. Until now, there have been no reports describing a functional interaction between these proteins. Our study is the first to demonstrate a functional interaction between the PI4K Stt4 and Ypp1. We have found that the phenotypes of stt4ts mutants are similar to those of our ypp1ts mutants, namely a reversal of the temperature-sensitivity by osmotic stabilization of media, sensitivity to agents that destabilize the cell wall, such as caffeine (also rescued by osmotic stabilization), and defects in actin organization. In addition, we have shown that overexpression of Stt4 suppressess the growth defect of two ypp1ts mutants. Overexpression of the other PI4Ks in yeast, Pik1 and Lsb6, does not suppress the temperature-sensitive growth defect of ypp1ts mutants at 37 °C. This implies a specific role for Ypp1 in Stt4-dependent events at the plasma membrane, as opposed to a general role connected with the function of all three PI4Ks in S. cerevisiae. In agreement with this, we see reduced binding at the cell periphery by a GFP–PHOsh2 sensor which is specific for PtdIns4P and PtdIns(4,5)P2. This is more evident in the ypp1-35ts mutant, which shows clear reduction of staining even at the permissive, as well as the non-permissive, temperature. Staining at the plasma membrane by this sensor is related directly to Stt4 activity, as plasma membrane localization of GFP–PHOsh2 is lost in an stt4ts mutant shifted to the non-permissive temperature [33]. The clear decrease in PH domain staining at the cell membrane in ypp1ts mutants is accompanied by an increase in the intensity of punctate intracellular staining, especially in the ypp1-35ts mutant at 25 °C and the ypp1-41ts mutant at 37 °C. In addition, this intracellular punctate staining is entirely due to PtdIns4P generated by Pik1, as GFP–PHOsh2 staining is lost completely when Pik1 activity is depleted, but retained when Stt4 is inactivated [33]. The increase in intensity of punctate intracellular staining in the ypp1-35ts mutant at 25 °C and the ypp1-41ts mutant at 37 °C could be due to a compensatory increase of PtdIns4P by Pik1. Alternatively, levels of the PH sensor may not be saturating, so a decrease in plasma-membrane-associated PtdIns4P means that more sensor is free to bind this lipid at intracellular locations.

Audhya et al. [9] reported an increase in the levels of PtdIns4P in wild-type cells, following a 1 h up-shift to 37 °C. This is in contrast with our observation of heat shock leading to a reduction in levels of this phosphoinositide. The difference may be due to the method used for radiolabelling phosphoinositides. Audhya et al. [9] labelled cells by incubation with [3H]inositol for 10 min, so measurable levels of phosphoinositides could be influenced by differences in the rate of incorporation of label. We labelled to near-steady-state equilibrium so any changes seen in phosphoinositide levels relate solely to their mass.

Consequently, our results are in agreement with another study which also used equilibrium labelling [16]. In any case, our data thus far suggest that the primary defect responsible for phenotypes exhibited by ypp1ts mutants is a reduction in levels of PtdIns4P. This was confirmed by loss of the Sac1 phosphoinositide 4-phosphatase restoring growth to all three ypp1ts mutants at 37 °C.

The Pkc1-dependent MAPK cascade is not defective in the ypp1-35ts mutant

Optimal activity of Stt4 is needed for cell wall integrity. The PtdIns(4,5)P2, synthesized at the plasma membrane by sequential action of Stt4 and Mss4, is bound by the PH domain of Rom2, which in turn localizes Rom2 to the plasma membrane, where it activates Rho1. This leads to activation of the Rho1/Pkc1-mediated MAPK cascade. Inactivation of Stt4 leads to mislocalization of Rom2, and the activation of the cascade is severely compromised, leading to limited phosphorylation of the Slt2 MAPK [14]. The ypp1ts mutants were sensitive to agents that destabilize the cell wall, so we anticipated that activation of Slt2 and subsequent activation of the Rlm1 transcription factor would also be abrogated in these strains, but this was not the case. In ypp1-35ts, Slt2 is phosphorylated and Rlm1-mediated transcription is stimulated in response to an external stimulus known to activate the MAPK cell wall integrity cascade, namely incubation with caffeine. Accordingly, although the level of plasma-membrane-associated PtdIns4P in ypp1-35ts may be low (Figure 5A), it is sufficient to allow adequate PtdIns(4,5)P2 synthesis for proper activation of the cell wall integrity cascade.

Compromise of Ypp1 function does not lead to a defect in internalization of endocytic cargo proteins

Phosphoinositides play key roles in endosomal trafficking. For example, PtdIns3P has roles throughout the endocytic pathway (reviewed in [1]). In addition, a number of recent studies have indicated that PtdIns4P is also required for endocytosis. Defects in the endocytic pathway occur in the sac1Δ background, in which levels of PtdIns4P are elevated up to 10-fold [35]. However, we have shown that the fall in PtdIns4P levels observed in the yppts mutants does not affect internalization of Fui1 and Ste3, two well-established transmembrane substrates of the endocytic machinery. This suggests that intracellular levels of PtdIns4P can fall significantly without a major effect on endocytosis. In agreement with this are data from a similar approach using cells compromised for function of the Stt4 PI4K itself. When incubated at non-permissive temperature, Ste6 (a plasma membrane transporter subjected to constitutive endocytosis), is efficiently internalized and targeted to the vacuole in both stt4ts and wild-type cells [9]. Given that endocytosis is particularly sensitive to elevated PtdIns4P, it follows that that the defect in endocytosis we observed in sac1Δ cells was rescued by combination with a ypp1ts mutation that restores cellular levels of this phosphoinositide.

Some components of the clathrin-mediated endocytic machinery are PtdIns(4,5)P2-binding proteins [45,46]. Furthermore, PtdIns(4,5)P2 is also involved in remodelling of the actin cytoskeleton, which in turn facilitates endocytosis [47,48]. PtdIns(4,5)P2 is synthesized by Mss4, the only phosphoinositide 4-phosphate 5-kinase in yeast. As expected, a reduction in levels of PtdIns(4,5)P2 caused by loss of Mss4 function leads to a defect in endocytosis [49]. Levels of PtdIns4P, the precursor of PtdIns(4,5)P2 synthesis, fall in the ypp1ts mutants. However, this is not accompanied by a significant fall in PtdIns(4,5)P2 synthesis. This may explain why endocytosis of Fui1–GFP remained unaffected, given that activity of the PtdIns(4,5)P2 effectors involved in endocytosis do not change significantly.

Expression of the Parkinson's disease A30P α-syn mutant is lethal in yeast. This is suppressed by raising the levels of Ypp1, which elevates endocytosis and degradation of A30P α-syn [21]. However, this may not be due to a stimulation of endocytosis itself. Instead, the enhanced rate of A30P α-syn degradation may be due to recruitment of A30P α-syn to the plasma membrane, where it becomes assimilated into the naturally occurring endocytic processes operating in yeast. Three observations are in agreement with this. First, Ypp1 is localized in a punctate pattern in close proximity to the plasma membrane. Secondly, Ypp1 binds to A30P α-syn, but not to the wild-type or A53T α-syn mutant. This explains why A30P α-syn is degraded, whereas wild-type or A53T α-syn remain intact [21]. Thirdly, loss of Ypp1 function does not affect the endocytosis of yeast membrane proteins that are endocytic substrates (Figures 8 and 9).

We investigated the phenotypes displayed by three ypp1ts alleles. Multiple amino acid substitutions were found in the corresponding mutant proteins, with seven substitutions found in ypp1-35ts and ypp1-41ts and six substitutions in ypp1-39ts. Further work will be required to determine which mutations are responsible for the temperature-sensitive phenotypes. We have shown that Ypp1 contributes to generation of PtdIns4P by the lipid kinase Stt4. Some of the mutations fall within the putative TPR protein–protein interaction domains of Ypp1, which may explain the loss of interaction with the Stt4 lipid kinase. The precise role played by Ypp1 in function of Stt4 remains to be determined, and may involve delivery of PtdIns substrate to Stt4, or a direct effect on the catalytic action of Stt4 mediated by interaction with Ypp1.

Acknowledgments

This work was supported by grant BB/E020550/1 from the Biotechnology and Biological Research Council (BBSRC) to B.P. C.Z. was a recipient of a scholarship from the KC Wong Foundation and the China Scholarship Council. We thank Tim Levine (Institute of Ophthalmology, University College London, London, U.K.) for providing the pTL511 vector, Rob Piper (Department of Physiology and Biophysics, University of Iowa, Iowa City, IA, U.S.A.) for providing the pJLU34 vector, and Rosine Haguenauer-Tsapis (Institut Jacob Monod, Paris, France) for providing the pFL38GalFUI1-GFP vector.

Abbreviations: AD, activation domain; 3-AT, 3-aminotriazole; BD, DNA-binding domain; GFP, green fluorescent protein; MAPK, mitogen-activated protein kinase; ORF, open reading frame; PH, pleckstrin homology; PI4K, phosphoinositide 4-kinase; PPIn, polyphosphoinositide; SD, synthetic minimal; α-syn, α-synuclein; TPR, tetratricopeptide repeat; ts, temperature-sensitive; YPD, yeast extract/peptone/glucose

References

View Abstract