Role of the Unfolded Protein Response Pathway in Regulation of INO1 and in the sec14 Bypass Mechanism in Saccharomyces cerevisiae

Corresponding author: Susan A. Henry, Cornell University, 245 Biotechnology Bldg., Ithaca, NY 14853., sah42{at}cornell.edu (E-mail)

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

INO1, encoding inositol 1-phosphate synthase, is the most highly regulated of a class of genes containing the repeated element, UAS_INO, in their promoters. Transcription of UAS_INO-containing genes is modulated by the availability of exogenous inositol and by signals generated by alteration of phospholipid metabolism. The unfolded protein response (UPR) pathway also is involved in INO1 expression and the ire1 and hac1 mutants are inositol auxotrophs. We examined the role of the UPR in transmitting a signal generated in response to inositol deprivation and to alteration of phospholipid biosynthesis created in the sec14^ts cki1 genetic background. We report that the UPR is required for sustained high-level INO1 expression in wild-type strains, but not for transient derepression in response to inositol deprivation. Moreover, the UPR is not required for expression or regulation of INO1 in response to the change in lipid metabolism that occurs in the sec14^ts cki1 genetic background. Thus, the UPR signal transduction pathway is not involved directly in transcriptional regulation of INO1 and other UAS_INO-containing genes. However, we discovered that inactivation of Sec14p leads to activation of the UPR, and that sec14 cki1 strains exhibit defective vacuolar morphology, suggesting that the mechanism by which the cki1 mutation suppresses the growth and secretory defect of sec14 does not fully restore wild-type morphology. Finally, synthetic lethality involving sec14 and UPR mutations suggests that the UPR plays an essential role in survival of sec14 cki1 strains.

THE promoters of many yeast phospholipid structural genes contain variants of a 10-bp repeated cis-acting promoter element, the inositol-sensitive upstream activating sequence (UAS_INO), which controls transcription in response to the soluble precursors inositol and choline (CARMAN and HENRY 1989 ; GREENBERG and LOPES 1996 ). Among the UAS_INO-containing genes, the INO1 gene, encoding inositol 1-phosphate (I1-P) synthase (DONAHUE and HENRY 1981 ), exhibits the most dramatic regulation and is, therefore, frequently used as a reporter for the entire regulon (HIRSCH and HENRY 1986 ; CARMAN and HENRY 1989 ; LOPES et al. 1991 ).

Transcription of INO1 and other UAS_INO-containing genes also responds to a signal generated from alteration of phospholipid metabolism (HENRY and PATTON-VOGT 1998 ; see Fig 1 for phospholipid metabolic pathways). The overproduction of inositol (Opi^-) phenotype, indicative of overexpression of the INO1 gene, is associated with mutants defective in biosynthesis of phosphatidylcholine (PC) via the methylation of phosphatidylethanolamine (PE; GREENBERG et al. 1983 ; SUMMERS et al. 1988 ; GRIAC et al. 1996 ). The fact that the Opi^- regulatory phenotype is conferred by mutations in structural genes involved in phospholipid biosynthesis suggests that the regulatory mechanism controlling expression of INO1 and other UAS_INO-containing genes might involve a signal generated by ongoing phospholipid metabolism (HENRY and PATTON-VOGT 1998 ). This idea was strengthened when a strong Opi^- phenotype was observed in sec14 cki1 strains (PATTON-VOGT et al. 1997 ). The SEC14 gene encodes a phosphatidylinositol (PI)/PC transporter essential for viability and secretion (BANKAITIS et al. 1989 , BANKAITIS et al. 1990 ). At the restrictive temperature, sec14^ts mutants arrest at the late Golgi stage of the secretory pathway (NOVICK et al. 1980 , NOVICK et al. 1981 ). Mutations in the CDP-choline pathway for PC biosynthesis (cki1, cct1, and cpt1) suppress the growth and secretory defects of sec14 mutants via a bypass mechanism (CLEVES et al. 1991 ). Thus, sec14^ts strains carrying these suppressors are viable at the sec14^ts restrictive temperature, but they exhibit both Opi^- and overproduction of choline (Opc^-) phenotypes (PATTON-VOGT et al. 1997 ). The Opc^- phenotype is directly related to elevated phospholipase D1 activity, which results in increased production of choline and phosphatidic acid (PA; SREENIVAS et al. 1998 ). The Opc^- phenotype of sec14 cki1 strains is eliminated if the SPO14 (PLD1) gene, which encodes phospholipase D1 (ROSE et al. 1995 ), is deleted (SREENIVAS et al. 1998 ). However, the Opi^- phenotype is also eliminated when SPO14/PLD1 is deleted (SREENIVAS et al. 1998 ). The fact that the Opi^- phenotype and elevated INO1 expression are dependent upon Pld1p supports the hypothesis that a signal related to PA production is responsible for INO1 induction in sec14^ts cki1 strains elevated to the restrictive temperature (PATTON-VOGT et al. 1997 ; HENRY and PATTON-VOGT 1998 ; SREENIVAS et al. 1998 ). Furthermore, active Pld1p is essential for the sec14 bypass mechanism and sec14^ts cki1 spo14 strains fail to grow at the sec14^ts restrictive temperature (SREENIVAS et al. 1998 ; XIE et al. 1998 ).

View larger version (20K): In this window In a new window Download PPT slide   Figure 1. Phospholipid metabolism in Saccharomyces cerevisiae. Water-soluble molecules that can be taken up from media are circled (I, inositol; C, choline). Dashed arrows indicate potential flux across the plasma membrane. Intercellular water-soluble metabolites are indicated by ovals (Glu-6-P, glucose 6-phosphate; I-1-P, inositol 1-phosphate; CP, choline phosphate; CDP-C, cytidinediphosphate choline). Lipids are indicated by rectangles (PA, phosphatidic acid; DAG, diacylglycerol; CDP-DG, cytidine-diphosphate diacylglycerol; PS, phosphatidylserine; PE, phosphatidylethanolamine; PC, phosphatidylcholine; PI, phosphatidylinositol; PIP, phosphatidylinositol phosphate; PIP2, phosphatidylinositol bisphosphate; SL, sphingolipids). Solid arrows represent routes of metabolic conversion. The names of genes encoding products catalyzing specific metabolic conversions are shown adjacent to the solid arrows.

The unfolded protein response (UPR) signal transduction pathway also influences INO1 expression. Under endoplasmic reticulum (ER) stress, Ire1p, a transmembrane kinase spanning the ER membrane, is activated and carries out site-specific endoribonucleolytic cleavage of HAC1 mRNA (COX et al. 1993 ; MORI et al. 1993 ; SHAMU and WALTER 1996 ; SIDRAUSKI and WALTER 1997 ; RUEGSEGGER et al. 2001 ). Only the spliced form of the HAC1 transcript is known to be effectively translated (CHAPMAN and WALTER 1997 ; KAWAHARA et al. 1997 ), and consequently Hac1p is detectable only in UPR-activated cells (COX and WALTER 1996 ). Hac1p functions as a basic leucine zipper (bZIP) transcription factor by binding to the upstream activating sequence (UAS) of target genes known as UPREs (unfolded protein response elements; MORI et al. 1992 ; COX and WALTER 1996 ). UPREs were originally found in promoters of chaperone family genes such as KAR2, PDI1, and EUG1 (MORI et al. 1992 ; KOHNO et al. 1993 ). A single UPRE element is sufficient to activate transcription from a heterologous promoter in response to the accumulation of unfolded proteins in the ER lumen (COX et al. 1993 ). In addition to being defective in UPR activation, ire1 and hac1 mutants require exogenous inositol for growth and express low levels of INO1 transcript (NIKAWA and YAMASHITA 1992 ; COX et al. 1993 , COX et al. 1997 ; MORI et al. 1993 ). COX et al. 1997 reported that under inositol-depleting conditions, the UPR is activated and suggested that the activation of the UPR leads to expression of INO1 in response to inositol deprivation. However, MORI et al. 2000 reported that Hac1p produced from an unspliced form of HAC1 mRNA is able to suppress inositol auxotrophy of hac1, suggesting that activation of the UPR per se may not be required for INO1 expression.

In this report, we examine the relationship between the UPR and signals generated from phospholipid metabolism in the sec14 genetic background. We report that a functional UPR is not necessary for INO1 activation or regulation in the sec14^ts cki1 genetic background. However, a functional UPR is required for the bypass mechanism by which the cki1 mutation suppresses the secretory defect of sec14 mutants.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains, media, and growth conditions:
The genotypes and sources of strains used in this study are listed in Table 1. All of the yeast strains listed in Table 1 are of the W303 genetic background. Strains were constructed by standard tetrad analysis (SHERMAN et al. 1978 ; ROSE et al. 1990 ). The HCY399 and HCY400 triple mutants containing sec14^ts cki1 in conjunction with ire1 or hac1 were generated by crosses of SHY653 containing sec14^ts cki1 with JCY147 or JCY408 containing ire1 or hac1. To ensure a uniform genetic background, HCY399 and HCY400 were then backcrossed to the wild-type strains, SHY629 and SHY652, respectively. HCY006, HCY029, HCY030, HCY031, and HCY032 were obtained as spores from the cross between HCY400 and SHY652. HCY401, HCY402, HCY403, and HCY404 were obtained as spores from the cross between HCY399 and SHY629. HCY136 was obtained as a spore colony from a cross between JPV110 and HCY006. Two sets of four strains from each cross (HCY401–HCY404 and HCY029–HCY032, each set derived from a single tetratype ascus) were used to assess the growth phenotypes. Rich YEPD (yeast extract, peptone, dextrose), synthetic complete, synthetic minimal, and sporulation media were prepared as previously described (CULBERTSON and HENRY 1975 ; GREENBERG et al. 1982 ). Synthetic complete medium was prepared without inositol (I^-) or supplemented with 75 µM inositol (I⁺) as previously described (CULBERTSON and HENRY 1975 ; GREENBERG et al. 1982 ). Drop-out medium (the same as complete synthetic medium with inositol but lacking a specific amino acid, adenine, or uracil) was used to test auxotrophic requirements. To select for deletion mutations carrying the kanMX6 marker, 200 mg/liter geneticin (YEPD + G418; Calbiochem, La Jolla, CA) was added to the media described above. To test tunicamycin (tm) sensitivity, 1 mM tunicamycin (Boehringer Mannheim, Indianapolis) was added to I⁺, I^-, and YEPD media (COX and WALTER 1996 ). To analyze growth phenotypes of yeast strains, cells were cultured in YEPD or I⁺ liquid medium to the mid-logarithmic phase of growth. Cells were collected by centrifugation and washed twice with sterile distilled water (dH₂O). The concentration of the cells was adjusted to OD₆₀₀ = 0.7 with sterile dH₂O. The cells were initially diluted 1:100 using dH₂O followed by 1:10 serial dilutions. From each dilution 10 µl of cells were spotted on an appropriate plate and allowed to grow at the designated temperature.

Synthetic lethal analysis:
To test the genetic interaction between the sec14 and hac1 mutations, JPV110 was crossed to HCY006. Diploids were sporulated and dissected. All resulting spores contained the deletion mutation cki1, which has no effect on viability, but suppresses the growth defect conferred by sec14. The genotypes of viable spores were determined by their ability to grow on amino-acid drop-out media and YEPD + G418. Once the genotypes of viable spores were determined, the tetrads were categorized into parental ditype, nonparental ditype, and tetratype with respect to the sec14 and hac1 mutations, and the genotypes of spores that failed to germinate were deduced. In the initial cross, 75% of spores containing the sec14 cki1 hac1 genotype failed to germinate, whereas the sec14 cki1 segregants exhibited >90% viability. The relatively low viability of the sec14 cki1 hac1 segregants suggested that this genotype might be inviable except in the presence of a suppressor segregating in the cross. To test this hypothesis, a sec14 cki1 spore colony (HCY136) was selected from a tetrad that contained a viable sec14 cki1 hac1 segregant, and the cross with HCY006 was repeated. Analysis of the synthetic lethality of sec14 and hac1 was assessed as described above in a cross between HCY136 and HCY006.

ß-Galactosidase assays:
For INO1-CYC-lacZ expression assays, yeast strains harboring a leu2 or ura3 mutation were transformed to leucine or uracil prototrophy with the autonomously replicating plasmids, pMR1036 (RUIS-NORIEGA 2000 ) or pJH359 (LOPES et al. 1991 ), using the standard lithium acetate method (HILL et al. 1991 ). The transformants were pregrown in I⁺ medium to mid-logarithmic phase. Cells were collected by centrifugation, washed with sterile dH₂O, diluted to OD₆₀₀ = 0.1 in repressing (I⁺) or derepressing medium (I^-), and allowed to grow at the designated temperature. Samples of 1 ml were taken at the designated time points and immediately frozen at -80°. For analysis, samples were thawed on ice, suspended in 1 ml of sterile dH₂O, and analyzed for ß-galactosidase activity using yeast ß-galactosidase assay kit (Pierce, Rockford, IL). For UPRE-CYC-lacZ expression assays, the strains containing the integrated UPRE-CYC-lacZ (Leu) or 2 µM UPRE-CYC-lacZ (Ura) reporter construct (COX and WALTER 1996 ) were precultured overnight at 25° and diluted to OD₆₀₀ = 0.1 in YEPD medium. The cells were then shifted to 30° and 37°, and samples were collected at the designated time points. The samples were processed as described above.

Tests for Opi^- and Opc^- phenotypes:
The method for detection of the Opi^- phenotype has been described previously (GREENBERG et al. 1983 ; SWEDE et al. 1992 ). Strains were patched onto I^- medium and allowed to grow at 30° for 2 days. The plates were then sprayed with a suspension of a diploid tester strain (AID), which is homozygous for ino1 and ade1. The cells were incubated at 30° for another 2 days. The Opc^- test was performed in a similar fashion, except that the media used in this assay were either I^- or I⁺, lacking choline, and the tester strain was a choline auxotroph, cho2 opi3 (PATTON-VOGT et al. 1997 ).

Lipid analysis:
Strains were grown in I⁺ medium at 30°, harvested at mid-logarithmic phase of growth by centrifugation, washed twice with sterile dH₂O, resuspended in 5 ml of I^- medium to OD₆₀₀ = 0.25, and allowed to grow for 1 or 4 hr at 30°. A total of 100 µCi of [³²P]orthophosphate/ml was then added to I^- medium, and the cells were incubated for 20 min. The cells were harvested and treated with trichloroacetic acid. Labeled lipids were extracted as previously described (ATKINSON et al. 1980 ). The individual phospholipid species were resolved by two-dimensional paper chromatography (STEINER and LESTER 1972 ) and quantified by PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

Electron microscopy:
Cells subjected to electron microscopy were processed by the procedure of WEBB et al. 1997 , with modifications. This method is specifically optimized for visualizing vacuolar structures. Other structures such as the ER may not be conspicuous. Briefly, the cells were grown in a 5-ml culture of YEPD to an OD₆₀₀ of 0.5. The cells were diluted to OD₆₀₀ = 0.1 in YEPD and shifted to 37°. Samples of 5 ml were taken at 0, 1, and 2 hr after the temperature shift. Each sample collected was fixed for 2 hr at 30° by the addition of 3% glutaraldehyde and 5 mM CaCl₂ buffered with 100 mM sodium cacodylate, pH 6.8. The cells were collected by centrifugation, washed once in 100 mM Tris-HCl (pH 8.0), 25 ml dithiothreitol, 5 mM EDTA, and 1.2 M sorbitol, and incubated in the same solution for 10 min at 30°. The cells were then washed once in 0.1 M K₂HPO₄ adjusted to pH 5.8 with citric acid, and 1.2 M sorbitol. The cells were resuspended in 0.5 ml of the same buffer. A total of 50 µl of ß-glucuronidase-type H-2 (114,000 units/ml, Sigma, St. Louis) and 2.5 mg of Zymolyase 20T (20,000 units/g, Seikagaku, Tokyo) was added. The cells were incubated for 2 hr at 30° to allow the cell wall to be removed. After washing three times with 100 mM sodium cacodylate buffer containing 5 mM CaCl₂, the cells were postfixed for 30 min at room temperature in 1% OsO₄, 1% K-ferrocyanide, and 5 mM CaCl₂ buffered with 100 mM sodium cacodylate, pH 6.8. The samples were washed four times with dH₂O, resuspended for 5 min in 1% thiocarbohydrazide in dH₂O, washed four times with dH₂O, and fixed for 5 min in 1% aqueous OsO₄. The samples were washed four times with dH₂O and dehydrated through a series of ethanol dilutions (50, 70, 80, 90, and 100%), followed by two washes with 100% propyleneoxide. The samples were infiltrated in a (1:1) propylene oxide:LR White resin mixture for several hours, transferred to 100% LR White resin, and infiltrated overnight at room temperature. The next day, a fresh change of 100% LR White resin was added, and infiltration was extended for an additional 8 hr. The resin was polymerized in gelatin capsules at 60° for 24 hr. Thin (80-nm) sections were cut using a DDK diamond knife on a Reichert-Jung Ultracut E. The sections were placed on copper grids, stained with Reynolds lead citrate, viewed, and photographed in a Hitachi 7100 transmission electron microscope operated at an acceleration voltage of 50 keV.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The hac1 and ire1 mutations do not confer inositol auxotrophy in the sec14 cki1 genetic background:
Triply mutant strains, sec14^ts cki1 ire1 and sec14^ts cki1 hac1, were created by standard genetic crosses, as described in MATERIALS AND METHODS. Tetratype spore colonies from single tetrads derived from crosses of HCY399 to SHY629 and HCY400 to SHY652 were tested for growth on I⁺ and I^- medium and for Opi^- and Opc^- phenotypes. The genotypes of the analyzed spore colonies were the following: wild type (HCY403), ire1 (HCY401), sec14^ts cki1 (HCY402), and sec14^ts cki1 ire1 (HCY404), or wild type (HCY032), hac1 (HCY030), sec14^ts cki1 (HCY031), and sec14^ts cki1 hac1 (HCY029). Detailed genotypes of these strains are listed in Table 1. In addition, we tested related sec14^ts, sec14^ts ire1, and sec14^ts hac1 strains (HCY363, HCY364, and HCY365; Table 1). All of the strains grew normally in I⁺ medium at 25° and 30°. The ire1 and hac1 single mutants exhibited slow or defective growth on I^- medium (i.e., have an Ino^- phenotype; Fig 2A), as previously reported (NIKAWA and YAMASHITA 1992 ; COX et al. 1993 , COX et al. 1997 ; COX and WALTER 1996 ). The sec14^ts mutant grew normally on I⁺ and I^- media at 25° and 30° and failed to grow on any medium at its restrictive temperature of 35° or higher, as previously reported (BANKAITIS et al. 1989 , BANKAITIS et al. 1990 ; data not shown). At 33°, the sec14^ts strain exhibited a slightly leaky Ino^- phenotype, but grew normally on I⁺ medium (data not shown). In contrast to sec14^ts strains, sec14^ts ire1 and sec14^ts hac1 strains grew poorly on I⁺ media at 33° and exhibited an Ino^- phenotype more stringent than that of the sec14^ts strain grown at this temperature (data not shown).

View larger version (91K): In this window In a new window Download PPT slide   Figure 2. The sec14ts cki1 ire1 and sec14ts cki1 hac1 strains are able to grow on I- medium at 25°and 30°, but fail to grow at 37°. (A) Two sets of strains, each derived from a single tetrad (HCY029–032 and HCY401–404), were spotted as serial dilutions on I+ and I- medium and incubated at 25°, 30°, and 37° for 4 days. (B) Serial dilutions of the same set of strains were spotted on YEPD medium and incubated at 25°, 30°, and 37° for 4 days.

Similar to sec14^ts cki1 strains, but in contrast to their ire1 and hac1 parents, the sec14^ts cki1 ire1 and sec14^ts cki1 hac1 triple mutants grew normally on I^- media at 25° and 30° (Fig 2A). This Ino⁺ phenotype suggests that the INO1 gene is expressed in the sec14^ts cki1 ire1 and sec14^ts cki1 hac1 genetic background, a topic that will be discussed subsequently. Unexpectedly, the sec14^ts cki1 ire1 and sec14^ts cki1 hac1 strains failed to grow on any medium, including YEPD, I⁺, and I^- media at 37°, the restrictive temperature for the sec14^ts allele (Fig 2A and Fig B). This temperature-sensitive phenotype of the triple mutants resembles that of sec14^ts strains, as opposed to that of sec14^ts cki1 strains (CLEVES et al. 1991 ), and suggests that a negative genetic interaction is occurring between sec14 and the UPR mutations hac1 and ire1.

Moreover, both sec14^ts cki1 hac1 and sec14^ts cki1 ire1 strains exhibited Opi^- phenotypes at 30°, indicative of INO1 overexpression, similar to the phenotype previously observed in the sec14^ts cki1 parental strains (PATTON-VOGT et al. 1997 ; Fig 3). The triply mutant strains also exhibited Opc^- phenotypes (data not shown), similar to the sec14^ts cki1 parent (PATTON-VOGT et al. 1997 ), suggesting that the mutation in the UPR does not affect the elevated level of Pld1p (Spo14p) activity in the sec14^ts cki1 genetic background.

View larger version (63K): In this window In a new window Download PPT slide   Figure 3. Opi- phenotypes of the wild-type (SH629), cki1 (SH627), sec14ts cki1 (SH630), sec14 cki1 (JPV110), the wild-type (HCY403), ire1 (HCY401), sec14ts cki1 ire1 (HCY404), and sec14ts cki1 hac1 (HCY029) strains. Overexpression of INO1 and excretion of inositol by strains being tested results in growth of the inositol auxotrophic tester strain at 30°. Growth of the tester strain is indicated by a halo around the strain being tested.

INO1 is expressed and regulated by inositol in the sec14^ts cki1 genetic background in the absence of a functional UPR pathway:
To assess INO1 expression in the triple mutants, wild-type, hac1, sec14^ts cki1, and sec14^ts cki1 hac1 strains were transformed with the INO1-CYC-lacZ reporter gene, as described in MATERIALS AND METHODS. Cells were first grown under repressing conditions (I⁺) at the sec14^ts permissive temperature of 25° and then shifted to derepressing (I^-) conditions at 30° or 37° (semipermissive and restrictive temperatures, respectively, for sec14^ts). Following a shift from I⁺ at 25° to I^- medium at 30°, hac1 and ire1 cells exhibited significant initial derepression of INO1, reaching a maximum level after 4 hr that approached 60–70% of the level achieved by wild-type cells under identical conditions. Within 5 hr following the shift to I^- medium, however, ß-galactosidase expression from the INO1 reporter construct in hac1 and ire1 cells plateaued at a level 40% of that observed in wild-type cells (Fig 4; ire1 data not shown). This pattern of INO1 expression is similar to that reported by COX et al. 1997 for INO1 expression in ire1 cells. Similar to wild-type cells, hac1 and ire1 cells did not express INO1 at all in I⁺ medium (Fig 4). Thus, despite the overall decrease in INO1 expression in the hac1 and ire1 mutants compared to wild-type cells in I^- medium, the mechanism of regulation of INO1 in response to inositol appears to be intact in the absence of a functional UPR.

View larger version (23K): In this window In a new window Download PPT slide   Figure 4. Induction of the INO1-CYC-lacZ reporter gene in wild-type (SH652), sec14ts cki1 (HCY402), hac1 (HCY030), and sec14ts cki1 hac1 (HCY400) cells. A schematic diagram depicting the construction of pMR1036 is shown above the graph. Plasmid pMR1036 contains a fusion of two UASINO elements derived from INO1, a minimal CYC1 promoter element, and the coding sequence for the Escherichia coli ß-galactosidase enzyme. Expression of ß-galactosidase from the INO1-CYC-lacZ in ire and sec14ts cki1 hac1 cells was similar to that observed in hac1 and sec14ts cki1 ire1 cells, respectively, and those data are not shown. Cells were transformed with the INO1-CYC-lacZ reporter gene (pMR1036), and transformants were shifted from I+ to I+ () or I- () medium at the indicated temperatures. Samples were taken 5 hr following the shift. ß-Galactosidase activity was measured as described in MATERIALS AND METHODS. The ß-galactosidase activity unit was defined as OD420/min/ml.

INO1 expression in sec14^ts cki1 cells was found to be similar to patterns reported by PATTON-VOGT et al. 1997 . In general, INO1 expression levels were found to be higher in sec14^ts cki1 cells grown in I^- medium at 25°, 30°, and 37° than in wild-type cells grown under identical conditions and the degree of INO1 overexpression tended to increase with temperature, as previously reported (PATTON-VOGT et al. 1997 ; Fig 4). At the sec14^ts restrictive temperature of 37°, sec14^ts cki1 cells expressed INO1 even in the presence of inositol (PATTON-VOGT et al. 1997 ; Fig 4), whereas wild-type cells did not express INO1 at any appreciable level, at any temperature, in the presence of inositol.

The pattern of INO1 expression in the sec14^ts cki1 hac1 strain at 25° and 30° was similar to the pattern observed in the sec14^ts cki1 parent. When shifted from I⁺ medium at 25° to I^- medium at 30°, the sec14^ts cki1 hac1 strain achieved a level of INO1 expression 2.5-fold higher than that of the wild-type control grown under identical conditions, while the sec14^ts cki1 strain expressed a level 2.9-fold higher than that of wild type (Fig 4). This high level of INO1 expression is in contrast to the low level of expression observed in the hac1 single mutant (Fig 4). INO1 expression patterns were similar in sec14^ts cki1 ire1 (data not shown). These results are consistent with the growth (Fig 2A) and Opi^- phenotypes of the sec14^ts cki1 hac1 and sec14^ts cki1 ire1 strains (Fig 3) and confirm that an intact UPR is not required for INO1 derepression and overexpression in the sec14^ts cki1 genetic background. Furthermore, since INO1 is not expressed in sec14^ts cki1 hac1 and sec14^ts cki1 ire1 strains grown in I⁺ medium at 30° (Fig 4; sec14^ts cki1 ire1 data not shown), the mechanism of regulation in response to inositol also appears to be intact in the absence of a functional UPR in the sec14^ts cki1 genetic background at the sec14^ts semipermissive temperature of 30°.

Phosphatidylinositol synthesis is compromised in hac1 and ire1 mutants, but not in sec14^ts cki1 ire1 strains:
Since inositol serves as a precursor to the synthesis of PI, the effect of UPR mutations on phospholipid synthesis was assessed. Wild-type, hac1, ire1, and sec14^ts cki1 ire1 strains were pulse labeled with [³²P]orthophosphate ([³²P]H₃PO₄), as described in MATERIALS AND METHODS and Table 2. Cells were grown to logarithmic phase in I⁺ medium at 30° and shifted to I^- medium at 30°. They were then incubated for an additional 1 or 4 hr and then labeled for 20 min with [³²P]H₃PO₄. As a control, cells grown in I⁺ medium were shifted to I⁺ medium for 1 hr and labeled for 20 min. When cells were pulse labeled in I⁺ medium at 30°, the phospholipid labeling pattern was similar in all strains (Table 2) and similar to labeling patterns previously described (ATKINSON et al. 1980 ; PATTON-VOGT et al. 1997 ). During the 20-min pulse-labeling period in I⁺ medium, the wild-type strain incorporated >30% of lipid-associated ³²P into PI (Table 2). Also comparable to previously published reports (KELLEY et al. 1988 ), in wild-type cells shifted to I^- medium, the relative incorporation into PI declined dramatically to 3% of total incorporation, while label accumulated in the precursors, PA and cytidine-diphosphate diacylglycerol (CDP-DG). The labeling pattern of other phospholipids was largely unaffected.

The ire1 and hac1 strains exhibited a labeling pattern similar to wild type when pulse labeled 1 hr following the shift to I^- medium. Within 4 hr after the shift to I^- medium, however, the proportion of label incorporated into PI in wild-type cells recovered to 9% of the total label incorporated into phospholipids, while the proportion of label accumulated in PI in the ire1 and hac1 strains remained low (Table 2). The failure of PI synthesis to recover in ire1 and hac1 cells correlates with the failure to achieve and sustain a high level of INO1 expression in the absence of a functional UPR (Fig 4).

The proportion of label associated with CDP-DG also remained significantly higher in hac1 cells shifted to I^- medium, suggesting that CDP-DG, the immediate precursor of PI, was accumulating to a greater extent in hac1 cells than in wild-type cells grown in I^- medium. However, in hac1 cells as compared to wild type, PA labeling declined somewhat and PC labeling increased, indicating that there may be subtle differences in many aspects of lipid metabolism in UPR mutants as compared to wild type. The labeling of PI in the sec14^ts cki1 ire1 strain shifted to I^- medium recovered in a fashion similar to that seen in the wild-type strain. However, as expected, the proportion of label incorporated in PC in the triple mutant was lower than that in wild type due to the cki1 mutation, which blocks PC synthesis via the CDP-choline pathway. PLD-mediated turnover of PC is also elevated in sec14^ts cki1 strains, a factor that also influences PC labeling (PATTON-VOGT et al. 1997 ; SREENIVAS et al. 1998 ).

The UPR and sec14 mutations exhibit synthetic lethality:
The failure of the sec14^ts cki1 ire1 and sec14^ts cki1 hac1 strains to grow at 37° (Fig 2A and Fig B) suggested the possibility of a negative genetic interaction involving sec14 in combination with hac1 or ire1 mutations. To test this hypothesis, a cross of a sec14 cki1 strain to a cki1 hac1 strain was conducted, as described in MATERIALS AND METHODS.

As a control, sec14 cki1 and cki1 spo14 strains were crossed to each other, since functional phospholipase D1 encoded by SPO14 (PLD1) is known to be required for the viability in strains carrying sec14^ts in combination with cki1 or other bypass suppressors (PATTON-VOGT et al. 1997 ; SREENIVAS et al. 1998 ; XIE et al. 1998 ). Thirty tetrads were dissected from a cross of sec14 cki1 to cki1 spo14, and, as expected, no spores of the sec14 cki1 spo14 genotype survived (data not shown), whereas the survival rates of all other genotypes exceeded 90%. We also tested for synthetic lethality between spo14 and hac1 or ire1. The double mutants, hac1 spo14 and ire1 spo14, were found to be viable and exhibited normal growth (data not shown).

From the diploid generated by the cross of sec14 cki1 to cki1 hac1, 67 tetrads were dissected. While the survival rate of the cki1 and cki1 hac1 spores was 100% and the survival rate of the sec14 cki1 spores was >90%, only five sec14 cki1 hac1 spores survived, a survival rate of 7%. A representative set of spore colonies derived from 10 tetrads from this cross is shown in Fig 5. Among the five surviving sec14 cki1 hac1 spore colonies, four gave rise to small colonies, one of which is shown in Fig 4. One of the sec14 cki1 hac1 colonies failed to propagate after it was restreaked on YEPD medium. Out of the total of 67 tetrads, only one of the five surviving sec14 cki1 hac1 segregants exhibited apparently normal growth.

View larger version (72K): In this window In a new window Download PPT slide   Figure 5. Genetic interaction between sec14 and hac1 mutations. Tetrads from the cross between the sec14 cki1 (HCY136) and cki1 hac1 (HCY006) strains were dissected and incubated for 4 days at 30° and the genotypes of the individual spores were subsequently determined. A photograph of spore colonies derived from 10 representative tetrads is depicted. The genotypes corresponding to the individual spore columns are shown in the box below a photograph of the colonies, using the following abbreviations: c, cki1; ch, cki1 hac1; sc, sec14 cki1; and sch, sec14 cki1 hac1. The symbol * indicates a rare surviving sec14 cki1 hac1 spore from the cross. The genotypes of dead colonies () were deduced.

sec14 mutations confer tunicamycin sensitivity:
Mutations in the UPR pathway are known to confer sensitivity to tm, which compromises the N-linked glycosylation process and leads to the accumulation of unfolded proteins in the ER (COX et al. 1993 ). The apparent synthetic lethality involving UPR and sec14 mutations led us to explore the tunicamycin sensitivity of strains carrying sec14 and sec14^ts mutations. Wild-type, hac1, cki1, and sec14 cki1 strains were tested at 30° for growth on YEPD medium containing 1 mM tunicamycin (YEPD + tm). As expected, the wild-type strain grew normally on YEPD + tm, while the hac1 strain was tm sensitive (Fig 6).

View larger version (74K): In this window In a new window Download PPT slide   Figure 6. The sec14 cki1 and sec14ts cki1 mutants display sensitivity to tunicamycin (tm). Serial dilution of strains were spotted on YEPD and YEPD containing 1 mM tunicamycin (YEPD + tm) plates and incubated at the indicated temperatures for 5 days. Wild-type (SHY629), hac1 (HCY030), sec14 cki1 (JPV110), and cki1 (SHY627) strains were incubated at the sec14ts permissive (25°), semipermissive (30°), and restrictive (37°) temperatures. Growth of wild type (HCY403), sec14ts (SHY625), and sec14ts cki1 (SHY630) at 25° (data not shown) was similar to growth at 30°.

No growth defect was observed in the cki1 strain grown on tm-containing medium (Fig 6). However, the sec14 cki1 strain exhibited a previously unreported sensitivity to tm (Fig 6), a phenotype that was further analyzed in the sec14^ts conditional mutant by examining growth on YEPD + tm medium at 25°, 30°, and 37° (Fig 6). The sec14^ts strain exhibited a temperature-sensitive phenotype and failed to grow at 37°, as previously reported (NOVICK et al. 1980 ; BANKAITIS et al. 1989 ), and grew normally at 25° and 30°, regardless of the presence or absence of tunicamycin (Fig 6). The sec14^ts cki1 strain grew normally in the presence of tunicamycin at the sec14^ts permissive and semipermissive temperatures of 25° and 30°, but was very sensitive to tm at the sec14^ts restrictive temperature of 37° (Fig 6), suggesting that inactivation of Sec14p leads to tunicamycin sensitivity.

sec14 cki1 strains exhibit elevated UPRE expression:
The unexpected finding of synthetic lethality between sec14 and UPR mutants and tunicamycin sensitivity in sec14 cki1 strains led us to examine UPR induction in sec14 cki1 strains. The fusion reporter construct UPRE-CYC-lacZ has been used previously to measure expression levels driven by the UPR responsive element, UPRE (COX and WALTER 1996 ). As expected, wild-type and cki1 strains exhibited low levels of UPRE induction when grown in YEPD medium (Fig 7). The UPR mutants, ire1 and hac1, exhibited low levels of UPRE expression, similar to uninduced levels observed in the wild-type strain grown at 30°, as described previously by COX and WALTER 1996 . However, the sec14^ts cki1 strain, even at the sec14^ts semipermissive temperature of 30° in YEPD medium, exhibited an induction of the UPRE reporter construct that is 3-fold higher than that of the wild-type or the cki1 strain grown under the same conditions (Fig 7). When the sec14^ts cki1 strain was grown at 37°, the restrictive temperature for the sec14^ts allele, UPRE expression was 6-fold higher than in the wild-type strain grown under the same conditions (Fig 7). In the sec14 cki1 strain, UPRE expression was 13-fold higher than in wild type at 30° and 15-fold higher at 37° (Fig 7). These results suggest that sec14 cki1 cells experience stress, which correlates with the increased sensitivity of sec14^ts cki1 and sec14 cki1 cells to tunicamycin and with the synthetic lethality of sec14 and UPR mutations.

View larger version (20K): In this window In a new window Download PPT slide   Figure 7. The UPR is induced in sec14 cki1 (JPV110) and sec14ts cki1 (SHY630) strains. A schematic diagram of pJC104 is shown above the graph. Plasmid pJC104 contains a fusion of four UPRE elements derived from KAR2, a minimal CYC1 promoter element, and the coding sequence for the E. coli ß-galactosidase enzyme. Yeast cells of the indicated strains were transformed with the UPRE-CYC-lacZ reporter construct carried on pJC104, and fresh transformants were cultured for 7 hr at the indicated temperatures. ß-Galactosidase activity was measured as described in MATERIALS AND METHODS. The ß-galactosidase activity unit was defined as OD420/min/ml.

sec14 cki1, sec14 cki1 ire1, and sec14 cki1 hac1 cells exhibit abnormal vacuolar morphology:
Induction of UPRE in sec14^ts cki1 and sec14 cki1 cells suggested that they experienced abnormal stress, leading us to examine their subcellular morphology. Wild-type, ire1, hac1, sec14^ts, sec14^ts cki1, sec14 cki1, sec14^ts cki1 ire1, and sec14^ts cki1 hac1 cells were examined using electron microscopy. Wild-type and hac1 cells were found to have similar normal morphology at 25° and 37° (Fig 8, A–F) and ire1 cells were similar, except that the vacuole seemed to be slightly enlarged compared to the wild-type strain (Fig 8E and Fig F). At 25° and 37°, the cki1 strain exhibited normal subcellular morphology similar to that observed in the wild-type strain (data not shown).

View larger version (324K): In this window In a new window Download PPT slide   Figure 8. sec14 cki1 and sec14ts cki1 cells exhibit vacuolar abnormalities. Depicted are thin-section electron micrographs of wild-type (HCY403) cells grown at 25° and 37° (A and B); hac1 (HCY030) at 25° and 37° (C and D); ire1 (HCY401) at 25° and 37° (E and F); sec14 cki1 (JPV110) at 25° and 37° (G and H); sec14ts cki1 (SHY630) at 25° (I) and 37° (J–L); sec14ts cki1 hac1 (HCY029) at 25°and 37° (M and N); sec14ts cki1 ire1 (HCY404) at 25° and 37° (O and P); and sec14ts (SHY625) at 25° and 37° (Q and R). Cells were initially grown in YEPD medium at 25° and shifted to 37° during the mid-logarithmic phase of growth (OD600 = 0.5), and samples were collected before (25°) and 2 hr following the shift to 37°. Samples were fixed, strained, and subjected to electron microscopy as described in MATERIALS AND METHODS. V, vacuole; N, nucleus. Bars, 1 µm.

However, regardless of the presence of the cki1 suppressor, severe defects in subcellular morphology (Fig 8, G, H, J, K, L, N, P, and R) appeared upon sec14 inactivation. Subcellular morphology of the sec14^ts, sec14^tscki1, sec14^ts cki1 ire1, and sec14^ts cki1 hac1 strains appeared to be normal when the cells were grown at the sec14^ts permissive temperature of 25° (Fig 8I, Fig M, Fig O, and Fig Q). However, defective subcellular morphology became evident in sec14^ts, sec14^ts cki1, sec14^ts cki1 ire1, and sec14^ts cki1 hac1 strains within 1 hr following a shift to the sec14^ts restrictive temperature of 37°. An elevated population of large (250–500 nm) membrane-bound structures, possibly enlarged Golgi, was detected in the sec14^ts cki1 cells within 2 hr following the temperature shift (Fig 8L). Similar defects were observed in all strains carrying sec14 mutations and were not correlated with ire1, hac1, or cki1 mutations. Interestingly, vacuolar structure defects were prominent in the sec14^tscki1 strain. The sec14 cki1 cells also exhibited severe fragmentation of the vacuole at 25° (Fig 8G). However, the fragmented vacuolar morphology of the sec14 cki1 strain appeared different from the defect observed in the sec14^ts cki1 strain at 37°. The vacuoles of the sec14^ts cki1 strain grown at 37° looked invaginated (Fig 8, J–L). We have no independent evidence on what processes are occurring with the vacuoles of these cells.

Activation of the UPR is not sufficient to fully derepress the INO1 gene in the presence of exogenous inositol:
COX et al. 1997 suggested that activation of the UPR in inositol-deprived cells might be related to, or required for, the mechanism for induction of INO1 transcription in the absence of inositol. The fact that UPRE expression is elevated in sec14 cki1 cells (Fig 7), which also express high levels of INO1 (Fig 4), is consistent with this hypothesis. However, the fact that INO1 is overexpressed in sec14^ts cki1 hac1 cells appears to negate this hypothesis. To further explore the relationship between UPRE induction and INO1 expression, wild-type cells transformed with the UPRE-CYC-lacZ or the INO1-CYC-lacZ reporter gene were exposed to various types of stress, including heat, hypo-osmotic, hyperosmotic, ethanol, and oxidative stress. Cells were also shifted to I^- or I⁺ containing 1 mM tunicamycin, conditions under which induction of UPRE-CYC-lacZ has previously been reported (COX et al. 1997 ).

As expected, cells grown in I⁺ medium exhibited very low levels of UPRE and INO1 expression (Fig 9). Consistent with the report by COX et al. 1997 , wild-type cells exhibited a 4.8-fold higher induction of UPRE when grown in I^- medium than when grown in I⁺ medium. However, this effect was not observed until 3 hr following the shift to I^- medium and derepression of INO1 clearly preceded derepression of UPRE (Fig 9). Within 1 hr following a shift from I⁺ to I^- medium, the wild-type strain exhibited a 10-fold induction of the INO1-CYC-lacZ reporter construct gene, whereas the UPRE reporter gene showed no change in expression during the first hour in I^- medium and only modest induction after 3 hr, by which time INO1-driven expression was elevated 20-fold (Fig 9).

View larger version (45K): In this window In a new window Download PPT slide   Figure 9. Comparison of induction kinetics of the INO1 and UPRE reporter genes in wild-type cells grown under various stress conditions. Wild-type strain SH629 was transformed with pJH359 (INO1-CYC-lacZ) or pJC104 (UPRE-CYC-lacZ). Transformants were precultured in I+ medium at 30°. Wild-type transformants were then shifted to stress conditions including inositol deprivation (I- medium at 30°), ER stress (I+ containing 1 mM tm at 30°), ethanol treatment (I+ containing 7.5% ethanol at 30° in the presence of 2% glucose), and heat shock (I+ at 37°), and I+ medium at 30° was used as a control. Samples were taken at 1 and 3 hr following the shift and analyzed for ß-galactosidase activity as described in MATERIALS AND METHODS. The ß-galactosidase activity unit was defined as OD420/min/ml.

In contrast to the relatively slow and modest induction of UPRE in wild-type cells shifted to I^- medium, cells shifted to I⁺ medium containing 1 mM tunicamycin exhibited a 7.5-fold induction of UPRE within 1 hr and an 11.4-fold induction after 3 hr (Fig 9). When wild-type cells were shifted to the dehydration condition imposed by the presence of 7.5% ethanol in I⁺ medium, UPRE expression had increased 5.3-fold by 1 hr and 6.9-fold by 3 hr (Fig 9). In contrast, a sudden change of temperature to 37° produced only a transient effect on UPRE expression in wild-type cells. A 2.8-fold induction of UPRE was seen within the first hour following a shift of wild-type cells from 25° to 37°, but this effect had almost disappeared by 3 hr (Fig 9). UPRE-CYC-lacZ was not significantly induced under hyper- (0.3 M NaCl) or hypo-osmotic (0.4 M sorbitol) stress or the oxidative stress imposed by the presence of H₂O₂ (0.4 mM; data not shown).

Despite the significant induction of UPRE under a number of the stress conditions described above, insignificant, or very minor, induction of the INO1 reporter gene was observed in I⁺ medium under all of these conditions, including the presence of 7.5% ethanol and 1 mM tunicamycin (Fig 9). Wild-type cells grown in I⁺ medium containing 1 mM tunicamycin exhibited minor induction of INO1 in comparison to the dramatic induction of UPRE. The level of INO1 induction in wild-type cells under these conditions is also insignificant compared to INO1 induction in cells shifted to I^- medium (Fig 9). These findings are consistent with the conclusion that the activation of the UPR is not sufficient to derepress INO1 transcription in the presence of inositol. Furthermore, activation of UPR does not appear to underlie the mechanism of INO1 activation during inositol deprivation, nor is an active UPR necessary for this regulation.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Transcription of INO1 and other UAS_INO-containing genes is regulated by the availability of exogenous inositol and is also affected by a signal generated from alteration of phospholipid metabolism (PATTON-VOGT et al. 1997 ; HENRY and PATTON-VOGT 1998 ). In this report, we have examined the role of the UPR in transmitting the signal generated by altered phospholipid metabolism in the sec14 cki1 genetic background, as well as the signal generated by the presence or absence of inositol. COX et al. 1997 reported that a functional UPR is necessary for sustained expression of INO1 in the absence of inositol, but not for its initial derepression. Our examination of INO1 expression and PI synthesis in hac1 and ire cells shifted to I^- medium is consistent with this interpretation. However, the residual INO1 expression in hac1 and ire cells is still regulated in response to inositol, and we conclude, therefore, that a functional UPR is not necessary for the transmission of the signal controlling INO1 transcription in response to inositol.

Moreover, unlike the ire1 and hac1 single mutants, both sec14^ts cki1 ire1 and sec14^ts cki1 hac1 strains were able to grow without exogenous inositol at 25° and 30° (Fig 2A) and exhibited an Opi^- phenotype similar to sec14^ts cki1 strains at the semipermissive temperature of 30° (Fig 3 and Fig 4). The expression of the INO1 reporter gene was elevated compared to wild type in sec14^ts cki1 hac1 and sec14^ts cki1 ire1 cells shifted to I^- medium at 30°, confirming that derepression and overexpression of the INO1 gene in the sec14^ts cki1 genetic background does not require a functional UPR (Fig 4). The Opi^- phenotype and INO1 overexpression in the sec14^ts cki1 strain has been shown to be due to elevated PC turnover caused by activation of Pld1p (PATTON-VOGT et al. 1997 ; SREENIVAS et al. 1998 ). Thus, the UPR pathway is not essential for INO1 transcription in response to elevated turnover of PC in the sec14^ts cki1 genetic background. Clearly, the signal generated by the altered phospholipid metabolism in the sec14^ts cki1 genetic background is not transmitted via the UPR pathway. Moreover, consistent with the results obtained using the hac1 single mutant, INO1 is not expressed in the presence of inositol at 25° or 30° in the sec14^ts cki1 hac1 triple mutant, confirming that the signal controlling INO1 expression in response to the presence or absence of inositol does not require a functional UPR (Fig 4).

Role of the UPR in INO1 expression:
The analysis of INO1 expression and regulation in hac1, ire1, sec14^ts cki1 ire1, and sec14^ts cki1 hac1 cells described above indicates that a functional UPR pathway is not necessary for INO1 expression or regulation. Despite previous reports to the contrary (COX et al. 1997 ; HYDE et al. 2002 ), our results do not support the conclusion that UPRE induction is sufficient to fully activate INO1 transcription, at least when inositol is present in the medium. In wild-type cells grown in I⁺ medium containing tunicamycin, we observed that UPRE was dramatically induced, while INO1 expression was minimally affected (Fig 9). Other conditions such as 7.5% ethanol treatment that resulted in dramatic induction of transcription of UPRE transcription had little or no effect on INO1 expression when inositol was present (Fig 9). Moreover, the INO1 reporter gene was induced earlier and to a greater extent than the UPRE reporter gene in wild-type cells shifted from I⁺ to I^- medium, suggesting that UPRE induction is not causally related to INO1 activation under these conditions (Fig 9).

Nevertheless, it is clear that the level of INO1 expression is influenced by the HAC1 and IRE1 gene products. The failure to sustain INO1 expression (Fig 4) and the consequent effect that this has on phospholipid metabolism (Table 2) explains the inositol auxotrophy of hac1 and ire1 strains. Yet, the mechanism by which the HAC1 and IRE1 gene products influence INO1 expression remains elusive. Our results suggest that the effect of the UPR on INO1 expression must be somewhat indirect, since INO1 transcriptional derepression in I^- medium precedes UPRE induction in cells shifted to I^- medium and UPRE induction does not elicit high levels of INO1 transcription in cells supplied with inositol (Fig 9). MORI et al. 2000 reported that Hac1p produced from an unspliced form of HAC1mRNA is able to suppress inositol auxotrophy of hac1 mutants. However, they also demonstrated that Hac1p produced from the spliced form of HAC1mRNA is a more efficient transcription factor for UPRE than Hac1p produced from the unspliced form of HAC1mRNA (MORI et al. 2000 ). The results of MORI et al. 2000 suggest that once Hac1p is produced, regardless of which form, it not only functions as a transcription factor for the UPRE-containing genes, but it also plays a separate, as yet undefined, role in sustaining a high-enough level of INO1 expression to alleviate the inositol auxotrophy of hac1 cells. Thus, it may be that it is the production of Hac1p, per se, rather than the induction of the UPRE, that is critical for sustained INO1 expression.

A functional UPR is required for the survival of sec14 cki1 strains:
An unanticipated finding of this study was the involvement of the UPR in survival of sec14 cki1 strains following Sec14p inactivation. Both sec14^ts cki1 and sec14 cki1 strains exhibited high levels of UPRE-CYC-lacZ transcription relative to the wild-type or cki1 strain (Fig 7), indicating that the UPR is induced when Sec14p is inactivated. The tunicamycin sensitivity of the sec1