INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In eukaryotic cells, the endoplasmic reticulum (ER) is the site where folding of the newly synthesized proteins that are destined for cell surface occurs. A number of cellular proteins, such as immunoglobulin binding protein, protein disulfide isomerase, glucose-regulated protein 94, peptidyl-prolyl-cis-trans-isomerase, and Erp72, are localized to the lumen of the ER and are proposed to act as chaperones to promote proper folding and/or to prevent aggregation of the folding intermediates (12). Consistent with the proposed chaperone functions, the accumulation of unfolded or misfolded proteins in the ER activates the transcription of the genes encoding the ER chaperones and thereby upregulates their synthesis. A conserved promoter element, the unfolded-protein response (UPR) element (UPRE), was identified in yeast (23) and mammalian (5) cells as being necessary and sufficient to mediate transcriptional induction in response to unfolded proteins in the ER. In the yeast Saccharomyces cerevisiae, a basic leucine zipper protein, Hac1p (8, 25), that binds to the UPRE and a transcriptional coactivator complex, Gcn5-Ada (42), are required for the transcriptional induction of KAR2 in response to unfolded proteins in the ER.

In yeast, transcriptional induction of the ER chaperone genes also requires a transmembrane serine/threonine kinase, Ire1p (7, 24). Ire1p is structurally similar to class I growth factor receptors and has three distinct domains, an N-terminal lumen domain, a transmembrane domain that spans the ER membrane, and a C-terminal domain that is either in the cytoplasm or in the nucleoplasm. The cytoplasmic-nucleoplasmic domain has intrinsic serine/threonine kinase activity (41), undergoes oligomerization and trans-autophosphorylation in response to unfolded proteins in the ER (30), and contains a region in the extreme carboxy terminus that has homology to RNase L (35). Thus, Ire1p appears to be the proximal sensor of unfolded proteins in the ER that initiates the UPR.

Recently, it was shown that the UPR is regulated by Hac1p levels in the cell (8, 16). When the UPR is inactive, Hac1p is not produced, as the HAC1 precursor mRNA is not translated (16). Upon activation of the UPR, HAC1 mRNA is spliced in an Ire1p-dependent manner to generate Hac1p, which binds the UPRE and activates transcription of the chaperone genes. As HAC1 mRNA splicing is not affected by mutations that affect the splicesome function, a novel splicing pathway appears to be involved in the splicing of HAC1 mRNA. Recently, it was shown that Ire1p has a site-specific endonuclease activity that cleaves HAC1 mRNA (35) and that the cleaved intermediates are subsequently ligated by the tRNA ligase Rlg1p (36).

Reversible protein phosphorylation is a major mechanism that modulates protein function in a variety of signal transduction pathways by the opposing actions of protein kinases and phosphatases. In eukaryotes, dephosphorylation at serine and threonine residues is catalyzed by gene products of two distinct families, PPP and PP2C (serine/threonine phosphatases of type 2C) (2). Members of the PP2C family show significant homology to mitochondrial pyruvate dehydrogenase phosphatase and show no apparent homology to the PPP family, comprised of PP1, PP2A, and PP2B. PP2C enzymes are unique in that they exist as monomers (15, 39), require Mg²⁺ or Mn²⁺ for catalytic activity (15, 39), and have no known inhibitors. The recently described crystal structure of human PP2C demonstrated that the catalytic domain is composed of conserved acidic residues that are proposed to coordinate Mg²⁺ or Mn²⁺ (9). Although the absence of specific inhibitors and genetic approaches has delayed the analysis of these enzymes, recent studies have revealed that they are key components of a variety of cellular signal transduction pathways. In eukaryotes, PP2C has been implicated in the reversal of protein kinase cascades that are activated upon environmental stress. In both fission and budding yeasts, a PP2C-like phosphatase negatively regulates the PBS2/HOG1-MAP kinase pathway that is activated in response to osmotic and heat shock (19, 20, 33, 34). In plants, PP2C positively regulates signal transduction by the plant hormone abscisic acid, leading to stomatal closure (18), seed dormancy (18, 22), and growth inhibition (18). Although these studies have identified pathways that are regulated by the PP2C family of phosphatases, no known specific substrates have been identified for these phosphatases to date.

To gain insight into the regulation of the UPR pathway, we searched for proteins that interact with the cytoplasmic-nucleoplasmic domain of Ire1p kinase. In this paper, we describe the identification of a novel PP2C-like protein serine/threonine phosphatase, Ptc2p, by virtue of its interaction with Ire1p. Ptc2p specifically interacts with phosphorylated Ire1p and dephosphorylates Ire1p in vitro. Cells devoid of Ptc2p have a hyperactive Ire1p receptor that leads to deregulation of the UPR pathway. We propose that Ptc2p downregulates the UPR by dephosphorylating Ire1p, thus functioning as an off switch in the UPR pathway.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Yeast strains, general methods, and plasmid constructions. Escherichia coli DH5 was used for the propagation of plasmids. The genotypes of the S. cerevisiae strains used in this study are shown in Table 1. The genetic methods and standard media used were previously described (31).

ire1-1

Western blot analysis. Yeast cell lysates were made according to Williams et al. (43). Western blotting was performed by standard procedures (14) with anti-influenza HA epitope (Boehringer Mannheim Biochemicals, Indianapolis, Ind.), anti-GST (Santa Cruz Biotechnology, Santa Cruz, Calif.), and anti-T7 epitope (Novagen, Madison, Wis.) primary antibodies and horseradish peroxidase-conjugated goat anti-mouse secondary antibodies (Gibco BRL, Gaithersburg, Md.). Bands were detected with an enhanced chemiluminescence kit (Amersham Corp., Arlington Heights, Ill.) and quantified with National Institutes of Health Image 1.55b program.

Protein phosphatase assays. Ptc2p, Ptc2p^D234A, Ptc2p^E37A/D38A, and Ptc2p^{E37A/D38A/D234A} were expressed as GST fusion proteins in Escherichia coli, purified, and cleaved with thrombin to release the phosphatase fragments. Purified recombinant GST-Ire1p (GST-WC) fusion protein was phosphorylated with [-³²P]ATP as described before (41). Phosphatase assays were performed as described by McGowan and Cohen (21) with buffer containing 50 mM Tris-HCl (pH 7.0), 0.1 mM EGTA, 0.1% (vol/vol) 2-mercaptoethanol, 60 mM magnesium acetate, and 1 mg of bovine serum albumin per ml and with ³²P-labeled GST-WC (1 µg) as the substrate. The phosphorylated proteins were analyzed by sodium dodecyl sulfate (SDS)-10% polyacrylamide gel electrophoresis (PAGE) and autoradiography.

In vitro pull-down assays. Plasmids carrying PTC2, ptc2^D234A, ptc2^E37A/D38A, or ptc2^{E37A/D38A/D234A} under the control of the T7 promoter were used in a coupled in vitro transcription-translation assay to create [³⁵S]methionine-labeled products. Translated products were separated by SDS-PAGE, treated with En³Enhance, and quantified with a PhosphorImager (Molecular Dynamics, Sunnyvale, Calif.). Equal amounts of recombinant proteins were incubated with glutathione-Sepharose beads containing equal amounts of phosphorylated and nonphosphorylated GST-Ire1p cytoplasmic domain fusion proteins at 4°C for 2 h in binding buffer (phosphate-buffered saline [PBS], 10% glycerol, 2 mM EDTA). As the control, beads containing comparable amounts of GST were used. Beads were recovered, washed twice with PBS containing 10% glycerol and 0.05% Triton X-100 and once with PBS, and boiled for 3 min in 1× Laemmli buffer (17). Extracts were electrophoresed on reducing SDS-10% polyacrylamide gels, treated with En³Enhance, and analyzed by autoradiography and PhosphorImager scanning.

Two-hybrid assays. Yeast two-hybrid assays were performed as described previously (13). The yeast reporter strain EGY48 was sequentially transformed with derivatives of pEG202 and pJG4-5 containing different regions of the Ire1p cytoplasmic domain. Interactions were monitored by the ability of the reporter strain to grow on media lacking leucine.

-Galactosidase activity and protein assays. -Galactosidase activity was quantified with a -galactosidase assay kit (Promega Corp., Madison, Wis.) according to the instructions provided by the manufacturer. Proteins were quantified with a protein assay kit (Bio-Rad Laboratories, Hercules, Calif.).

RNase protection analysis. Total RNA was isolated (29) from cells treated or not treated with tunicamycin for 90 min and analyzed with an RNase protection kit (Boehringer). pGEM-4ZHAC1 and pGEM-4ZACT1 were linearized with XbaI, and antisense RNA probes were synthesized with T7 RNA polymerase (Boehringer) and [-³²P]CTP (Amersham). For each sample, 10 µg of RNA was hybridized for 14 h at 45°C with 2 × 10⁵ cpm of HAC1 and ACT1 RNA probes, digested with RNase A and RNase T₁, processed as suggested by the manufacturer, and analyzed on a 6% polyacrylamide gel containing 8 M urea. The band intensities were quantitated by PhosphorImager scanning and normalized to that of actin (ACT1) mRNA. A DNA sequencing ladder of a known template was used as a size marker.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Ptc2p interacts with Ire1p in vivo. To elucidate mechanisms that regulate the UPR, we searched for molecules that interact with Ire1p. As conventional biochemical methods have often failed to identify such serine/threonine kinase receptor molecules, we used a modified version of the yeast two-hybrid system (13). The cytoplasmic-nucleoplasmic domain of Ire1p containing the intact kinase domain was used as the bait (LexA-WC) to screen a yeast genomic library. The kinase that showed the highest homology to Ire1p, Cdc28p (27% identity and 46% similarity within the kinase domains), was used as a negative control bait. PTC2 was isolated three times independently in interactions with Ire1p in this screen. When tested in the yeast two-hybrid system, B42-HA-Ptc2p specifically interacted with LexA-WC (Fig. 1A) but not with LexA-Cdc28p (data not shown). To examine the molecular details of the Ire1p-Ptc2p interaction, a series of N-terminal and/or C-terminal truncations were constructed in both IRE1p and Ptc2. The summary of the interaction data is presented in Fig. 1B. Both LexA-NK (N linker plus the kinase domain) and LexA-KC (kinase domain plus the C tail) interacted with B42-HA-Ptc2p, and the binding properties were quantitatively similar to those observed with LexA-WC, indicating that deletion of either the N-linker region or the C-tail region does not destroy the interaction. In contrast, neither LexA-WK (kinase domain alone) nor LexA-CT (C-tail alone) interacted with B42-HA-Ptc2p. Although we cannot rule out the possibility that these subdomains had altered secondary structures that did not permit detectable interactions, these results suggest that there are multiple Ptc2p interaction sites disseminated within the cytoplasmic-nucleoplasmic domain of Ire1p and that the loss of one or more such sites can significantly perturb the detectable interaction. The N-linker region alone could not be used in the two-hybrid assays because it was transcriptionally active.

View larger version (43K): [in this window] [in a new window]   FIG. 1.   Analysis of Ire1p interaction with Ptc2p and mapping of interaction domains. (A) Two-hybrid analysis. LexA fusion proteins with intact and truncated forms of the cytoplasmic domain of Ire1p (bait) were tested for interaction with the original clone B42-HA-Ptc2p (aa 174 to 464), B42-HA-Ptc2p containing aa 1 to 312, or B42-HA-Ptc2p containing aa 174 to 355. Transformants harboring IRE1 and PTC2 fusions were patched onto His Trp plates and replica plated onto His Trp Leu plates containing either glucose or galactose. The growth of strain EGY48 harboring different regions of the Ire1p cytoplasmic domain as LexA fusions (in pEG202) and as B42-HA fusions (in pJG4-5) on His Trp medium (control) and His Trp Leu medium containing either galactose plus raffinose (Gal) or glucose (Glu) was monitored. (B) Schematic representation of LexA-Ire1p fusion proteins (bait) and their interactions with B42-HA-Ptc2p (prey) as detected by two-hybrid analysis. WC, wild-type cytoplasmic-nucleoplasmic domain; NK, N linker plus kinase domain; WK, wild-type kinase domain; KC, kinase domain plus C-terminal tail; CT, C-terminal tail.

Ptc2p physically associates with phosphorylated Ire1p. To substantiate the genetic evidence for a protein-protein interaction between Ire1p and Ptc2p, affinity adsorption experiments were performed with products from coupled in vitro transcription-translation of PTC2 in the presence of [³⁵S]methionine. The Ire1p cytoplasmic-nucleoplasmic domain was produced as a soluble GST fusion protein (GST-WC) in E. coli, adsorbed to glutathione-Sepharose beads, and used for adsorption of the in vitro-translated Ptc2p from reticulocyte lysates. Glutathione-Sepharose beads impregnated with comparable amounts of GST were used as a control. In these assays, the amount of Ptc2p brought down by GST-WC was not different from the amount obtained with the control beads containing GST alone (Fig. 2, lanes 4 and 2, respectively). However, when GST-WC was autophosphorylated by incubation in kinase buffer with ATP prior to the adsorption, the binding of GST-WC to Ptc2p was increased by 4.5-fold (Fig. 2, lane 5) over that of the GST control. When EDTA was removed from the binding buffer, GST-WC did not bring down Ptc2p (data not shown). It is possible that in the presence of Mg²⁺ the phosphatase activity of Ptc2p dephosphorylates Ire1p and consequently results in the dissociation of Ptc2p from Ire1p (see below). These results suggest that Ire1p physically associates with Ptc2p as a result of a direct interaction between the two proteins. More importantly, these results indicate that Ptc2p discriminates between the phosphorylated and the nonphosphorylated forms of Ire1p and specifically interacts with the phosphorylated receptor kinase.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Direct physical association of Ire1p and Ptc2p in vitro. Aliquots of an in vitro translation reaction mixture containing 35S-labeled Ptc2p were incubated with glutathione-Sepharose beads impregnated with either phosphorylated (lane 5) or nonphosphorylated (lane 4) GST-WC. As a control, beads coated with GST alone and either with (lane 3) or without (lane 2) prior incubation in in vitro kinase buffer (41) were used to adsorb Ptc2p. Beads were collected, washed, and boiled to release the bound proteins. Equal proportions of the samples were analyzed by SDS-PAGE, Western blotting, and fluorography. A portion (8.3%) of the Ptc2p input is shown in lane 1.

Ptc2p is a genuine protein serine/threonine phosphatase that dephosphorylates Ire1p in vitro. Ire1p kinase is autophosphorylated at serine and threonine residues both in vitro (41) and in vivo (30). The specific interaction between the phosphorylated form of Ire1p and Ptc2p, a putative protein serine/threonine phosphatase, implied two possible functional consequences. First, to investigate whether Ire1p is a substrate for Ptc2p, in vitro dephosphorylation assays were performed. Ptc2p was produced by expressing a GST-Ptc2p fusion protein in E. coli; the fusion protein was purified and subsequently cleaved with thrombin to remove GST. Ire1p was also expressed as a GST fusion protein (GST-WC) in E. coli, purified, and autophosphorylated in the presence of [³²-P]ATP. A fixed amount of labeled GST-WC was incubated with increasing amounts of Ptc2p. The degree of phosphorylation or dephosphorylation was analyzed by SDS-PAGE followed by autoradiography. In these assays, Ptc2p efficiently removed covalently attached phosphate groups from GST-WC (Fig. 3). The dephosphorylation was evident, with a GST-WC/Ptc2p ratio as low as 1:0.2 (data not shown), and increased steadily with increasing amounts of Ptc2p (Fig. 3, lanes 2 to 5). Since EDTA completely inactivated the dephosphorylation of GST-WC (see Fig. 5A, lane 3), the phosphatase activity of Ptc2p is Mg²⁺ dependent. These results demonstrate that Ptc2p is indeed a PP2C-like protein serine/threonine phosphatase that can efficiently dephosphorylate Ire1p kinase in vitro.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Dephosphorylation of Ire1p by Ptc2p. A fixed amount of 32P-labeled GST-WC (1 µg) was incubated with increasing amounts of Ptc2p in dephosphorylation buffer (21). Samples were boiled, and equal portions were analyzed by SDS-PAGE and autoradiography. The GST-WC/Ptc2p ratios used were determined by a protein assay and were confirmed by scanning of the Coomassie blue-stained gel with NIH Image. The migration of bovine serum albumin is indicated by an arrow.

Mutations in the metal binding sites affect the substrate binding and catalytic activity of Ptc2p. The crystal structure of human PP2C reveals that its catalytic domain is composed of a binuclear metal center that is coordinated by four invariant aspartate residues and a glutamate residue that are conserved among some members of the PP2C family (9). Since metal ions (Mn²⁺ and Mg²⁺) are required for the catalytic activity of the PP2C-like enzymes, substitution of these residues would predict catalytically inactive enzymes. To establish whether the phosphatase activity of Ptc2p is required for the in vivo regulation of Ire1p, we substituted the Glu³⁷, Asp³⁸, and Asp²³⁴ residues corresponding to the conserved residues that coordinate metal ions in the human counterpart (Fig. 4) with alanine. GST-Ptc2p fusion proteins carrying single (Ptc2p^D234A), double (Ptc2p^E37A/D38A), or triple (Ptc2p^{E37A/D38A/D234A}) mutations were expressed as soluble proteins in E. coli, indicating that the structures of these proteins were not severely misfolded. Dephosphorylation assays revealed that the single (Ptc2p^D234A; Fig. 5A, lane 5), double (Ptc2p^E37A/D38A; Fig. 5A, lane 4), and triple (Ptc2p^{E37A/D38A/D234A}; data not shown) phosphatase mutants could not catalyze the removal of covalently attached phosphate residues from GST-WC, similar to the results obtained in the absence of Mg²⁺ (Fig. 5A, lane 3). These results demonstrate that the above-listed mutations render the phosphatase catalytically inactive and are consistent with its predicted tertiary structure and the deduced catalytic mechanism.

View larger version (72K): [in this window] [in a new window]   FIG. 4.   Sequence alignment of the catalytic domains of human PP2C and Ptc2p. The GenBank accession numbers for the PP2C gene and PTC2 nucleotide sequences are S87759 and U18839, respectively. Invariant residues are shaded in gray. Conserved residues are boxed. Residues that coordinate metal ions are indicated by circles. Residues that were mutated in this study are denoted by diamonds. The alignment was made with the Wisconsin sequence analysis package (Genetics Computer Group Inc., Madison, Wis.).

View larger version (37K): [in this window] [in a new window]   FIG. 5.   Effect of mutations at the predicted Mg2+ binding sites on the Ire1p interaction and phosphatase activity. (A) In vitro dephosphorylation assay. Equal amounts (1 µg) of 32P-labeled GST-WC were incubated with dephosphorylation buffer (control, lane 1), Ptc2p (1 µg, lane 2), Ptc2p (1 µg) with EDTA (lane 3), Ptc2pE37A/D38A (1 µg, lane 4), or Ptc2pD234A (1 µg, lane 5). Samples were boiled, and equal portions were analyzed by SDS-PAGE Coomassie staining, and autoradiography. The migration of the wild-type and mutant phosphatases is indicated by a solid arrow. The open arrow indicates the migration of bovine serum albumin. (B) In vitro binding assay. Beads impregnated with equal amounts of either nonphosphorylated (lanes 4 to 6) or phosphorylated (lanes 7 to 9) GST-WC were tested for interaction with equal amounts of in vitro-translated Ptc2p (lanes 4 and 7), Ptc2pD234A (lanes 5 and 8), or Ptc2pE37A/D38A (lanes 6 and 9). Bound proteins were analyzed by SDS-PAGE, Western blotting, and fluorography. Ten percent of the Ptc2p input is shown in lanes 1 to 3. The phosphorylated form of GST-WC is indicated by an asterisk.

PTC2 is not essential for cell viability. If the interaction between Ire1p and Ptc2p and the dephosphorylation of Ire1p by Ptc2p were physiologically significant for the regulation of the UPR pathway, cells deficient in Ptc2p should have an altered UPR. To investigate this hypothesis, two yeast strains carrying null alleles of PTC2 were constructed. In AWY446, a kan^r gene was inserted into the unique Tth111I site in the 5' end of the PTC2 coding region. In the other null mutant, AWY506, the open reading frame was disrupted by replacing the 935-bp XbaI-NruI fragment with a LEU2 gene. This construction removed amino acid residues 42 to 353 of PTC2. PCR and Northern blot analyses confirmed that these mutants contained the disrupted PTC2 gene (data not shown). Both null mutants were viable, indicating that PTC2 is not an essential gene for cell viability. In rich liquid medium, ptc2 cells exhibited growth rates identical to those of wild-type and ire1 cells (Fig. 6A). Under the same growth conditions, a low concentration of tunicamycin, a drug that inhibits N-linked glycosylation and causes the accumulation of unfolded proteins in the ER, severely impaired the growth of the ire1 cells (Fig. 6A). In contrast, tunicamycin did not significantly affect the growth of either wild-type or ptc2 cells, and the growth curves for these two strains were indistinguishable. These observations suggest that ptc2 cells have no apparent growth defects and are capable of effectively protecting themselves from the lethal consequences of the accumulation of unfolded proteins in the ER.

View larger version (11K): [in this window] [in a new window]   FIG. 6.   Effect of PTC2 on the growth of S. cerevisiae in the presence (closed symbols) or absence (open symbols) of tunicamycin. (A) PTC2 deletion does not affect the growth rate. AWY446 (ptc2-1::kanr) (triangles), AWY500 (PTC2) (squares), AWY506 (ptc2-1::LEU2) (circles), and AWY 516 (ire11) (arrowheads) were grown in 1% yeast extract-2% peptone-2% dextrose (YPD) medium to an A600 of 0.1, and the cultures were divided into flasks. Tunicamycin (final concentration, 0.25 µg/ml) was added to one set of flasks, and incubation was continued at 30°C. Aliquots were removed at the indicated times, and the A600 was measured. (B) Ptc2p overproduction inhibits cell growth. AWY500 harboring PTC2 (arrowheads), ptc2D234A (circles), or ptc2E37A/D38A (squares) in a 2µm vector, pYES2 (diamonds), was grown in Ura medium containing 2% galactose and 1% raffinose for 10 h. Cultures were divided and treated as in panel A. (C) Wild-type Ptc2p and single and double Ptc2p mutants are expressed equally in S. cerevisiae. AWY500 harboring PTC2 (lane 1), ptc2E37A/D38A (lane 2), and ptc2D234A (lane 3) in the pYES2 vector was grown in Ura medium containing 2% galactose and 1% raffinose for 10 h. Cells were harvested and lysed, and the extracts were analyzed by Western blotting with anti-T7 antibody.

Overexpression of PTC2 retards cell growth. To further elucidate the role of PTC2 in yeast biology, T7 epitope-tagged Ptc2p was overexpressed with an inducible GAL1 promoter. Overexpression of wild-type Ptc2p inhibited cell growth, whereas overexpression of both Ptc2p mutants did not affect the growth rate relative to that of wild-type cells (Fig. 6B). Since the levels of expression of the mutant and wild-type Ptc2p proteins were similar (Fig. 6C), this difference was not due to a difference in expression. These results support the conclusion that overproduction of the phosphatase is growth inhibitory due to excessive phosphatase activity. It is possible that Ptc2p is also involved in other signaling pathways that regulate cell growth, and its overproduction may therefore inhibit cell growth. The addition of tunicamycin significantly retarded the growth of the yeast cells overexpressing wild-type Ptc2p but did not significantly affect the growth of either wild-type cells or cells that overexpressed mutant Ptc2p, indicating that the UPR pathway is compromised upon Ptc2p overexpression, a phenotype common to ire1 cells.

The UPR pathway is sensitized in ptc2 cells. In order to investigate whether null mutants of PTC2 have an elevated UPR, yeast strains harboring a single integrated copy of a lacZ reporter gene that has an upstream 22-bp UPRE were constructed. The UPR was monitored by the levels of tunicamycin-inducible -galactosidase expression in the cell extracts. In wild-type cells, the expression of -galactosidase increased steadily with increasing concentrations of tunicamycin (Fig. 7A). Although both the deletion (ptc2-1::LEU2) and the insertion (ptc2-1::kan^r) mutants showed similar dose-dependent levels of -galactosidase expression, the levels of expression were approximately three- to fourfold higher than that of wild-type cells, indicating that cells devoid of Ptc2p were hypersensitive to unfolded proteins. These observations suggest that PTC2 is a negative regulator of the UPR pathway.

View larger version (15K): [in this window] [in a new window]   FIG. 7.   PTC2 downregulates the UPR. (A) Null mutants of PTC2 exhibit an elevated UPR. Cultures of strains AWY446 (ptc2-1::kanr) diamonds, AWY500 (PTC2) (squares), and AWY506 (ptc2-1::LEU2) (circles) were grown in YPD to the early log phase and divided, and tunicamycin (Tm) was added to the final concentrations indicated. After 90 min of incubation at 30°C, cells were harvested and lysed and -galactosidase activity was measured. (B) Overexpression of wild-type Ptc2p and catalytically inactive Ptc2p deregulates the UPR. Strains expressing either wild-type or mutant Ptc2p were grown in Ura medium containing 2% galactose and 1% raffinose for 10 h to the early log phase. Cultures were divided, Tm was added to a final concentration of 2 µg/ml, and incubated was continued for 90 min. Cells were harvested and lysed, and -galactosidase activity was measured. Specific -galactosidase activity represents an average of three independent experiments. Bars indicate standard deviations. (C) Tm dose dependence of the UPR in wild-type cells overexpressing wild-type or mutant Ptc2p. The UPR was monitored as described in panel B, except that different concentrations of Tm were used for induction. Symbols: squares, pYES2 vector; diamonds, pYES2 carrying PTC2; circles, pYES2 carrying the D234A allele; triangles, pYES2 carrying the E37A/D38A allele.

Overexpression of wild-type Ptc2p but not the catalytically inactive mutants of Ptc2p downregulates the UPR. To study the functional effects of the catalytically inactive Ptc2p mutants on the UPR, T7 epitope-tagged versions of the catalytically inactive mutants were overexpressed in a wild-type strain that harbors a UPRE-LacZ reporter construct. In comparison with the vector control, yeast cells that overexpressed Ptc2p displayed a reduction in tunicamycin-dependent -galactosidase expression (Fig. 7B). Although this difference was observed throughout the entire range of concentrations, it was more pronounced at lower doses (Fig. 7C). In contrast, yeast cells that overexpress either Ptc2p^D234A (the single mutant that did not interact with Ire1p) or Ptc2p^E37A/D38A (the double mutant that could interact with Ire1p) displayed increased and tunicamycin-inducible -galactosidase activity (Fig. 7B). These results are consistent with the notion that PTC2 is a negative regulator of the UPR pathway and that the phosphatase activity of Ptc2p is essential for the regulation of the UPR.

PTC2 is required for regulated HAC1 mRNA splicing. Since activation of the UPR pathway is regulated by Ire1p-dependent HAC1 mRNA splicing and subsequent generation of Hac1p (8, 16), we questioned whether PTC2 plays a role in HAC1 mRNA splicing. To investigate this possibility, splicing of HAC1 mRNA was monitored by an RNase protection assay. In this assay, detection of an RNase-resistant 147-nucleotide fragment is diagnostic of cleavage of the 3' splice site of HAC1 mRNA. In agreement with Cox and Walter (8), cleaved HAC1 mRNA was detected in a tunicamycin-dependent fashion in wild-type control cells (Fig. 8, lanes 1 and 2) but not in ire1 cells (lanes 3 and 4). Interestingly, ptc2 cells displayed a 64% increase in tunicamycin-dependent cleavage of HAC1 mRNA (Fig. 8, lane 6), and tunicamycin-dependent cleavage of HAC1 mRNA was reduced to 57% in cells overexpressing wild-type Ptc2p (Fig. 8, lane 8). In addition, cleaved HAC1 mRNA was detected in ptc2 cells as well as in cells overexpressing either Ptc2p^D234A or Ptc2p^E37A/D38A in the absence of tunicamycin (Fig. 8, lanes 5, 9, and 11), indicating that HAC1 mRNA processing is deregulated in these cells. These results confirm the role of PTC2 as a negative regulator of the UPR pathway and, more importantly, demonstrate that PTC2 acts upstream of the HAC1 mRNA splicing event.

View larger version (54K): [in this window] [in a new window]   FIG. 8.   PTC2 regulates HAC1 mRNA splicing. Cultures of strains AWY500, AWY506, and AWY500 overexpressing either wild-type Ptc2p or Ptc2p mutants were grown in synthetic complete medium with or without uracil to the early log phase, divided, and induced with tunicamycin (Tm) (2 µg/ml) for 90 min. Cells were harvested, RNA was isolated, and cleavage of HAC1 mRNA was assayed by an RNase protection assay with a probe that is colinear with the S. cerevisiae HAC1 gene. The 147-nucleotide (nt) fragment represents the 3' portion of (HAC1) mRNA that extends to the 3' cleavage site. The 74-nucleotide fragment derived from the 5' side of the spliced HAC1 mRNA is not shown in this analysis. The abundance of spliced HAC1 mRNA HAC1i relative to ACT1 mRNA is indicated. Products of higher molecular weights may represent intermediates in the splicing reaction that are observed only in ptc2 cells (asterisks). 2µ, 2µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Ptc2p is a novel phosphatase that binds to and dephosphorylates Ire1p. In the budding yeast S. cerevisiae, the Ire1p kinase is activated in response to unfolded protein in the ER to transmit a signal that induces the transcription of genes encoding ER-resident protein chaperones. We have identified a novel serine/threonine phosphatase, Ptc2p, as a specific interactor of Ire1p by interaction-trap (two-hybrid) screening. This is the first demonstration of a protein serine/threonine phosphatase as a component of the signal transduction pathway from the ER to the nucleus. Several genetic and physical methods demonstrated that the association between Ptc2p and Ire1p was specific and selective. First, the yeast two-hybrid assay demonstrated that Ptc2p interacted with Ire1p but not with Cdc28p, a kinase that shows the highest homology to Ire1p. Second, in vitro physical assays demonstrated that Ptc2p interacted with phosphorylated Ire1p but not with nonphosphorylated Ire1p. Third, a single amino acid mutation in the Ire1p interaction domain of Ptc2p disrupted the interaction between the two proteins in vitro. On the basis of the in vitro interaction between the bacterially expressed Ire1p and the in vitro-translated Ptc2p, we suggest that the interaction is direct and specific and depends on the phosphorylation status of Ire1p.

View larger version (10K): [in this window] [in a new window]   View larger version (22K): [in this window] [in a new window]   FIG. 9.   Domain structure of Ptc2p and model for the role of Ptc2p in the UPR pathway. (A) Linear representation of Ptc2p. The myristoylation (Myr) signal and the phosphatase domain are deduced from the sequence of PTC2. The Ire1p interaction domain, as mapped by the two-hybrid analysis, is shown. (B) Model for the activation and inactivation of the UPR pathway in S. cerevisiae. P, covalently attached phosphate.

Ptc2p negatively regulates the UPR pathway upstream of HAC1 mRNA splicing. An increasing body of evidence suggests that the PP2C-like enzymes play multiple roles in regulating a number of signal transduction pathways. In mammals, PP2C is implicated in Ca²⁺-related signaling in the brain (11). In C. elegans, the PP2C enzyme FEM-2 is required to promote male development (6). The ABI1 gene product of A. thaliana, which features an EF hand (two alpha helices) Ca²⁺-binding site at the amino terminus and a PP2C domain at the carboxy terminus, is required for abscisic acid-mediated responses, such as stomatal closure (18), the maintenance of seed dormancy (18, 22), and the inhibition of plant growth (22). The S. cerevisiae genome encodes six PP2C-like enzymes (37) whose physiological functions are unknown, with the exception of Ptc1p and Ptc2p. Ptc1p downregulates the osmosensing signal transduction pathway (20) and is required for tRNA splicing (27). In Schizosaccharomyces pombe, the Ptc1p counterpart is encoded by ptc1⁺ and is important for survival of heat shock as well as for osmoregulation (32). S. cerevisiae Ptc2p is most homologous to Ptc3p of S. pombe, and both proteins are implicated in osmosensing signal transduction pathways (19, 20, 33, 34). In this report, we have described the role of Ptc2p in the UPR, a hitherto-unknown function for this enzyme.