Role of the Unfolded Protein Response Pathway in Secretory Stress and Regulation of INO1 Expression in Saccharomyces cerevisiae

Yeast strains:

The genotypes and sources of strains are listed in Table 1. Yeast strains listed in Table 1 are of the S288C and W303 genetic backgrounds. Strains were constructed by standard tetrad analysis (Sherman et al. 1978; Rose et al. 1990). Briefly, HCY104, HCY105, HCY106, HCY107, HCY126, and HCY362 were obtained as meiotic spore colonies from diploid deletion strains obtained from the Research Genetics (Huntsville, AL) strain collection. To obtain an ire1Δ strain isogenic with the genetic background of Sec⁻ strains used in this study, the HCY405 strain was constructed using PCR-based gene disruption of IRE1 in the CKY8 genetic background. A 1.5-kb deletion cassette containing the bacterial KanMX gene was amplified by PCR using pFA6-KanMX4 as a template and IRE1-S1 and IRE1-S2 as the primers (Wach et al. 1994). The KanMX4-containing PCR fragment was included at each end, 45 bp (85–129) and 43 bp (3039–3081), of the IRE1 ORF, resulting in replacement of the IRE1 gene with KanMX. The 1.5-kb deletion cassette was transformed into CKY8 using the lithium acetate (LiAc) method as described previously (Hill et al. 1991). Transformants were selected on YEPD containing 200 mg/liter geneticin (YEPD + G418; Calbiochem, La Jolla, CA). To verify the correct ORF replacement, the size of the insert (1.5 kb) was checked by PCR using a set of primers flanking the disrupted gene (IRE1-S3 and IRE1-S4) and genomic DNA isolated from each of the putative deletion mutants as a template. HCY423 was obtained by mating HCY405 and CKY46. To obtain a hac1Δ mutant in the same genetic background, a hac1Δ meiotic spore generated from the heterozygous diploid strain collection (Research Genetics) was backcrossed to CKY49 to generate HCY126.

Double mutants, Sec⁻ ire1Δ or Sec⁻ hac1Δ, were constructed by crossing the various ire1Δ or hac1Δ strains to the various Sec⁻ strains. HCY280 and HCY283 were obtained by mating HCY423 and RSY263. HCY427 and HCY428 were obtained by mating HCY423 and RSY267. HCY425 and HCY422 were obtained by mating HCY405 and CKY46. HCY437 and HCY435 were obtained by mating HCY423 and RSY269. HCY442 and HCY444 were obtained from the mating between HCY423 and RSY271. HCY455 and HCY457 were obtained from the cross of HCY423 and RCY279. HCY448 and HCY446 were obtained by mating HCY423 and RCY281. HCY470 and HCY473 were obtained by mating HCY405 and JBY318. HCY178, HCY179, HCY180, and HCY181 were obtained from the cross of HCY126 and CKY49. HCY462, HCY463, HCY465, and HCY464 were the products of a cross between HCY362 and HCY401. HCY495 and HCY496 were obtained from a mating between HCY107 and RCY243. Two sets of four spores containing the tetratype genotypes derived from a single tetrad (HCY462–HCY465 and HCY466–HCY469) were generated by crossing HCY362 (erg6Δ) to HCY401 (ire1Δ) and HCY030 (hac1Δ), respectively.

Transformation with pHAC1 and pHAC1ⁱ:

Various Sec⁻ strains were transformed with the HAC1- and HAC1ⁱ-containing plasmids (pHAC1 and pHAC1ⁱ, CEN/LEU2 marker; gifts from Dr. Kazutoshi Mori, Kyoto University, Kyoto, Japan). The sec13Δ strain, CKY480 (see Table 1 for the full genotype), carrying the wild-type SEC13 gene on a plasmid (pRS316; CEN; URA3 marker; provided by Dr. Chris Kaiser) was transformed with pHAC1ⁱ or an empty vector (YCp50; CEN; LEU2 marker). Since the SEC13 gene is essential for growth, the SEC13 null mutant is obligated to carry and express the wild-type SEC13 gene on a plasmid for its survival. The double transformants simultaneously carrying pSEC13 and an empty vector or pSEC13 and pHAC1ⁱ were grown for 1 day on plates containing synthetic medium lacking uracil and leucine and then replica plated onto plates lacking leucine and containing 5-fluoroorotic-acid (5-FOA) and incubated for an additional 3 days at room temperature. The expression of the URA3 gene is toxic to cells exposed to 5-FOA, forcing the transformants to lose the pSEC13 plasmid, thus obligating them to depend on expression of pHAC1ⁱ for survival.

β-Galactosidase assays:

To assay UPRE-CYC-lacZ expression, strains were transformed to uracil prototrophy with pJC104, containing UPRE-CYC-lacZ, provided by Dr. Peter Walter (Cox and Walter 1996). Transformants were precultured to midlogarithmic phase of growth at 25° in synthetic medium lacking uracil. Cells were collected by centrifugation, washed with sterile dH₂O, and diluted to OD₆₀₀ = 0.1 in medium lacking uracil and shifted to the designated temperature to grow for 3 additional hours. Samples (1 ml) were taken and analyzed for β-galactosidase activity using the Yeast β-galactosidase assay kit (Pierce, Rockford, IL). For UPRE-CYC-lacZ expression assay in BFA-treated erg6Δ cells, HCY104 and HCY362 were transformed with pJC104 (UPRE-CYC-lacZ). Cells were prepared as described above except that their initial OD₆₀₀ was 0.5. Indicated amounts of BFA were added to medium lacking uracil and cells were grown at 30°, and β-galactosidase activity was measured as described above.

To assess INO1 expression, strains were transformed to uracil prototrophy with pJH359 (INO1-CYC-lacZ) and transformants were precultured at 25° to midlog phase of growth in synthetic medium with (I⁺) or without (I⁻) inositol. Cells were then shifted to the indicated temperature for 3 hr and β-galactosidase was measured.

Northern blot analysis:

Cells were precultured overnight to midlogarithmic phase of growth at 25° in YEPD medium. Wild-type cells were shifted to YEPD medium at 30° in the presence or absence of 1 mm tunicamycin. Sec⁻ strains were shifted to YEPD medium at their semipermissive temperatures. Isolation of RNA was performed by hot phenol extraction. Northern analysis was performed by running 10-μg samples of RNA loaded onto 1% agarose, 6% formaldehyde, 1× MOPS gels and capillary transferred to Nylon Plus (QIAGEN, Valencia, CA) membrane. Prehybridization and hybridization conditions were as described in Hirsch and Henry (1986). Hybridization probes for ACT1 (Marykwas and Fox 1989) and HAC1 were prepared using a riboprobe in vitro transcription system. ACT1 was linearized with BamHI and riboprobe was synthesized with SP6 RNA polymerase according to manufacturer's instructions (Promega, Madison, WI). The 449-bp SalI fragment from YCp-HAC1 was subcloned into pGEM1 (Promega) through the SalI site (pGEM-HAC1). pGEM-HAC1 was linearized with HindIII. Riboprobe was synthesized with T7 RNA polymerase according to manufacturer's instructions.

INO1 expression is reduced in Sec⁻ mutants experiencing secretory stress:

INO1 expression was assayed in a selection of the Sec⁻ mutants as described in materials and methods using an INO1-CYC-lacZ reporter gene. All of the mutants tested exhibited reduced INO1 expression compared to wild type when cultured at semipermissive growth temperatures in inositol-free (I⁻) medium (Figure 2A). This effect was most pronounced in the sec13^ts-1 and sec16^ts-2 strains, where a three- to fivefold reduction in INO1-CYC-lacZ expression was observed in cells grown in I⁻ medium shifted to 30° compared to the same cells grown continuously at 25° (Figure 2A). The sec18^ts-1, sec22^ts-3, and sec21^ts-1 mutants also exhibited levels of INO1 expression that were 50% of wild type or lower, even at the permissive temperature of 25° (Figure 2A). A further reduction in INO1-CYC-lacZ expression was observed in these strains when they were shifted to the semipermissive temperatures of 30° or 33°. Of the strains tested (Figure 2), the sec22^ts-3 strain showed the lowest level of INO1 expression at 30° (∼10% of wild-type levels; Figure 2A). Somewhat reduced levels of INO1 expression were also observed at all temperatures in the sec1^ts-1, sec6^ts-4, and sec15^ts-1 strains. In these cases, β-galactosidase activity from the INO1-CYC-lacZ reporter gene was reduced compared to wild type even at 25°, but was not significantly further reduced at 30° or 33°. In the sec14^ts-3 strain, expression from the INO1 reporter gene was reduced compared to wild type at 25° and 30°. Expression of INO1-CYC-lacZ increased somewhat after a shift to 33°, but did not reach the level of expression observed in the wild-type strain grown under comparable circumstances (Figure 2A).

Open in a separate window

Figure 2.—

INO1 expression is reduced in Sec⁻ cells under conditions of secretory stress. (A) Sec⁻ strains were transformed with pJH359 (INO1-CYC-lacZ) as described in materials and methods and transformants were precultured at 25° to midlogarithmic phase of growth in I⁻ medium lacking uracil and shifted to the indicated temperatures. Samples were taken after 3 hr following the temperature shift and analyzed for β-galactosidase activity. β-Galactosidase activity is expressed as OD₄₂₀/min/ml. (B) sec13-1 cells carrying pJC104 (UPRE-CYC-lacZ) were precultured at 30° to midlogarithmic phase of growth in I⁺ medium lacking uracil. Cells were filtered and resuspended in I⁺ or I⁻ medium, which also lacked uracil, at an OD_600nm = 0.5. The cultures were incubated for an additional 3 hr at 30° and then analyzed for β-galactosidase activity. (C) sec13^ts-1 cells carrying pJH359 (INO1-CYC-lacZ) were precultured at 30° to midlogarithmic phase of growth in I⁺ medium lacking uracil. Cells were filtered and resuspended in I⁺ or I⁻ medium lacking uracil at an OD_600nm = 0.5. Cultures were incubated at 30° for an additional 3 hr and analyzed for β-galactosidase activity.

Since the UPR is known to be activated under inositol-limiting conditions (Cox et al. 1997), we examined both UPRE-CYC-lacZ and INO1-CYC-lacZ expression in sec13^ts-1 and wild-type cells shifted from inositol-containing (I⁺) to inositol-free (I⁻) medium. Wild-type and sec13^ts-1 cells, transformed with INO1-CYC-lacZ or UPRE-CYC-lacZ reporter constructs, were first grown to logarithmic phase at 30° in I⁺ medium and then shifted for 3 additional hours to I⁻ medium (Figure 2, B and C). Wild-type transformants treated in this fashion showed modest expression from the UPRE reporter gene following the shift to I⁻ medium at 30°, as well as derepression of the INO1 reporter construct (Figure 2, B and C), consistent with the report of Cox et al. (1997).

In contrast to wild type, at 30°, sec13^ts-1 cells exhibited high levels of β-galactosidase expression from the UPRE reporter gene whether inositol was present or not (Figure 2B). In fact, the level of UPRE expression in sec13^ts-1 cells grown at 30° in I⁺ medium, exceeded the level of UPRE expression in wild-type cells shifted to I⁻ medium by >8-fold. Following the shift to I⁻ medium, a further increase of ∼30% in expression from the UPRE-CYC-lacZ reporter gene was observed in sec13^ts-1 cells (Figure 2B). Overall, the level of β-galactosidase activity driven by the UPRE construct in sec13^ts-1 cells grown at 30° in I⁻ medium exceeded that observed under the same conditions in wild-type cells by ∼12-fold (Figure 2B). In contrast to the high level of UPRE expression observed in sec13^ts-1 cells grown at 30° in I⁺ medium (Figure 2B), INO1-CYC-lacZ expression in the sec13^ts-1 cells at 30° was fully repressed when inositol was present (Figure 2C). When sec13^ts-1 cells were shifted to I⁻ medium at 30°, INO1-CYC-lacZ derepression occurred, but the level of β-galactosidase activity achieved was only about one-third of the level seen in the wild-type strain (Figure 2C).

The hac1Δ and ire1Δ mutations exhibit partial synthetic lethality with mutations conferring defects in membrane trafficking:

Sec⁻ ire1Δ and Sec⁻ hac1Δ strains were generated for each of the Sec⁻ mutants used in this study and, in most cases, double-mutant progeny were found to have growth defects that were more extreme than those of the corresponding parental Sec⁻ single mutant. Examples of Sec⁻ ire1Δ double mutants, from each of the categories of Sec⁻ mutants used for the temperature-shift experiments described above, are shown in Figure 3A. Sec⁻ hac1Δ double-mutant strains, which are not shown, had phenotypes very similar to the corresponding Sec⁻ ire1Δ strains. No change was detected in the growth pattern on YEPD plates in response to temperature of double mutants involving sec1^ts-1 (Figure 3A) or pep12Δ with ire1Δ or hac1Δ, and slight, if any, difference was observed in the case of the double mutants involving sec4^ts-8, sec14^ts-3, and sec15^ts-1, as compared to the corresponding single mutants (data not shown). Double mutants involving sec2^ts-59 and sec21^ts-1 were not generated. In all other cases, the double mutants, Sec⁻ ire1Δ or Sec⁻ hac1Δ, failed to grow at a temperature ∼2°–3° lower than that of the corresponding Sec⁻ parent (Figure 3A; Sec⁻ hac1Δ data not shown).

Open in a separate window

Open in a separate window

Figure 3.—

(A) Genetic interactions between ire1Δ and Sec⁻ mutations and Ino⁻ phenotypes of Sec⁻ Upr⁻ mutants. The ire1Δ (either HCY405 or HCY423) strain was crossed to Sec⁻ mutants (CKY46, RSY267, RSY269, and JBY318) to generate Sec⁻ ire1Δ double-mutant strains. From each cross, four spore colonies from a single tetratype ascus were used to assess the growth phenotype. For simplicity, only growth of wild-type and ire1Δ spores generated from a cross of CKY46 and HCY405 is shown here, since all other wild-type and ire1Δ spores generated from other crosses showed similar growth phenotypes. HCY424 (wt), HCY423 (ire1Δ), HCY425 (sec13^ts-1), HCY422 (sec13^ts-1 ire1Δ), HCY280 (sec12^ts-4), HCY283 (sec12^ts-4 ire1Δ), HCY427 (sec16^ts-2), HCY428 (sec16^ts-2 ire1Δ), HCY437 (sec17^ts-1), HCY435 (sec17^ts-1 ire1Δ), HCY470 (sec6^ts-4), HCY473 (sec6^ts-4 ire1Δ), HCY495 (sec1^ts-1), and HCY496 (sec1^ts-1 ire1Δ) were precultured to midlogarithmic phase of growth at 20° in YEPD medium. The concentration of the cells was adjusted to OD₆₀₀ = 0.7 with sterile dH₂O. Cells were initially diluted 1:100 using dH₂O followed by 1:10 serial dilutions. Four microliters of cells from each dilution were spotted on YEPD plates and allowed to grow at the designated temperatures for 3 days. (B) Strains generated as described in A, above, were precultured to midlogarithmic phase of growth at 25° in I⁺ medium, washed twice, and spotted as a series of dilutions (1:10) on I⁺ or I⁻ medium and incubated at the temperatures shown. In the top, the phenotypes of four spore colonies from the cross of ire1Δ to sec17-1 are shown. Below, the wild-type and ire1Δ colonies are omitted.

Mutations affecting COPII vesicle formation exhibited some of the strongest negative interactions with ire1Δ and hac1Δ. For example, the sec13^ts-1 ire1Δ and sec13^ts-1 hac1Δ strains failed to grow at a temperature of 30° or higher, while the restrictive temperature for the sec13^ts-1 strain is 33° (Figure 3A). The sec16^ts-2 ire1Δ and sec16^ts-2 hac1Δ strains grew poorly even at 25°, while the sec16^ts-2 parental strain was able to grow, although poorly, at temperatures up to 30° (Figure 3A). Double mutants simultaneously defective in the UPR and sec12^ts-4 (Figure 3A) or sec23^ts-1 (data not shown) also failed to grow at temperatures ∼2° lower than the restrictive temperature for the corresponding parental Sec⁻ strain. The sec23^ts-1 ire1Δ and sec23^ts-1 hac1Δ double mutants exhibited particularly poor growth and could grow only at 20° or lower on YEPD plates. They could not be cultured in liquid YEPD medium even at 20° and they failed to grow entirely on YEPD plates at 25° or higher. The sec23^ts-1 parental strain, by contrast, was capable of normal growth on YEPD plates at temperatures up to 28° (data not shown). Double mutants carrying ire1Δ or hac1Δ in combination with mutations affecting vesicle docking and fusion, sec17^ts-1 (Figure 3A), sec18^ts-1, or sec22^ts-3 (data not shown), also failed to grow at temperatures 2°–3° lower than the minimum restrictive temperature of the corresponding parental Sec⁻ mutant. Among the mutations that affect distal secretion from the Golgi to the plasma membrane, only sec6^ts-4 exhibited a clear negative interaction with ire1Δ or hac1Δ. The sec6^ts-4 ire1Δ and sec6^ts-4 hac1Δ strains failed to grow at 33°, whereas the sec6^ts-4 parental strain has a restrictive temperature of 35° (Figure 3A).

The ire1Δ and hac1Δ mutations were also crossed into the erg6Δ genetic background to assess their effect on BFA sensitivity. While growth of the erg6Δ parental strain was completely inhibited by the presence of 100 mm BFA, as previously reported (Graham et al. 1993; Shah and Klausner 1993; Vogel et al. 1993), and was unaffected by 25 or 35 mm BFA in YEPD medium (Figure 4), the erg6Δ ire1Δ and erg6Δ hac1Δ double-mutant strains were inhibited by 35 mm BFA (Figure 4). Thus, the presence of a UPR mutation results in an increased sensitivity of erg6Δ to BFA.

Open in a separate window

Figure 4.—

Mutations in the UPR increase the sensitivity of erg6Δ cells to BFA. (A) As controls, HCY104 (wt), HCY105 (ire1Δ), HCY126 (hac1Δ), and HCY362 (erg6Δ) were tested for their ability to grow on YEPD medium containing 100 mm BFA. (B) Two sets of four spores containing the tetratype genotypes derived from a single tetrad (HCY462–HCY465 and HCY466–HCY469) were generated by crossing HCY362 (erg6Δ) to HCY401 (ire1Δ) or HCY030 (hac1Δ). Cells were precultured in YEPD medium at 30° to midlogarithmic phase (OD₆₀₀ = 0.5) of growth, spotted as a series of dilutions (1:10) on YEPD plates containing 25 mm and 35 mm BFA, and allowed to grow at 30° for 4 days.

Growth of a subset of the Sec⁻ strains was also examined at the semipermissive temperature of 30° on media containing or lacking inositol. As reported by Nikawa and Yamashita (1992), ire1Δ (Figure 3B) and hac1Δ strains (data not shown) exhibit somewhat “leaky” inositol auxotrophy (Ino⁻ phenotype; Figure 3B). The sec13-1 strain also has a leaky Ino⁻ phenotype at its semipermissive temperature of 30° (Figure 3B) as reported by Gilstring et al. (1999). The sec14^ts-3 strain also grows more poorly in I⁻ medium at the semipermissive temperature of 33° (Figure 3B), as reported by Kearns et al. (1997) and Chang et al. (2002). In a number of cases tested, the double mutants (i.e., Sec⁻ ire1Δ or Sec⁻ hac1Δ) exhibited somewhat tighter Ino⁻ phenotypes than either single-mutant parent did (Figure 3B). Growth of five such double mutants, sec14 ^ts-3, sec13^ts-1, sec6^ts-4, sec16^ts-2, and sec17^ts-1, in combination with ire1Δ is shown in Figure 3B.

Constitutive activation of the UPR by expression of the HAC1ⁱ gene rescues growth defects of some Sec⁻ mutants:

The sec1^ts-1, sec4^ts-8, sec13^ts-1, sec13^ts-4, sec14^ts-3, sec15^ts-1, and sec22^ts-3 strains were transformed with the HAC1ⁱ construct, lacking the intron sequence. The lack of the intron in HAC1ⁱ results in constitutive activation of the UPR, even in ire1Δ strains (Cox and Walter 1996; Kawahara et al. 1997, 1998; Mori et al. 2000).

The sec15^ts-1 and sec14^ts-3 mutants transformed with pHAC1ⁱ appeared to grow more poorly than the same strains carrying vector alone (data not shown), a result consistent with the reports of Kawahara et al. (1997) and Mori et al. (2000), who observed that wild-type strains transformed with pHAC1ⁱ exhibited slower growth than controls. No change was observed in the growth of the sec1^ts-1, sec22^ts-3, or sec4^ts-8 strains when transformed with pHAC1ⁱ (data not shown). While the growth of the sec1^ts-1 strain was not affected by transformation with pHAC1ⁱ (data not shown), the sec1^ts-1 ire1Δ strain transformed with pHAC1ⁱ was able to grow at 33° (Figure 5B), a temperature at which the sec1^ts-1 single mutant is unable to grow (see Figure 3A). This result was not expected, since the sec1^ts-1 ire1Δ and sec1^ts-1 strains had similar phenotypes at 33° (Figure 3).

The most dramatic effects of expression of pHAC1ⁱ were obtained in the sec13^ts-1 and sec13^ts-4 strains. When transformed with pHAC1ⁱ, the sec13^ts-4 and sec13^ts-4 hac1Δ strains were able to grow at 37° (Figure 5A), a temperature that is restrictive for sec13^ts-4. The sec13^ts-1 strain was also able to grow at 32° when transformed with pHAC1ⁱ (data not shown), whereas sec13^ts-1 will not grow above its semipermissive temperature of 30° when transformed with vector alone. HAC1ⁱ was also able to rescue growth of sec13^ts-1 in the absence of inositol at 30° and 32° (Figure 5C).

However, HAC1ⁱ was not able to restore sec13^ts-1 growth under any condition (with or without inositol) at temperatures of 33° or greater (Figure 5C), suggesting that HAC1ⁱ expression cannot suppress complete loss of Sec13p function. To test this idea, a sec13Δ strain carrying the wild-type SEC13 gene on plasmid pRS316 was transformed with pHAC1ⁱ or an empty vector (YCp50) and a plasmid “shuffling” experiment was performed, as described in materials and methods. When the transformants were forced to lose pSEC13, they failed to grow even when pHAC1ⁱ was present (data not shown), indicating that expression of the HAC1ⁱ gene cannot compensate for the total loss of Sec13p.

The role of the UPR and the secretory pathway in expression and regulation of the INO1 gene:

INO1 is one of a large number of coregulated genes, which contain the promoter element UAS_INO and respond to the availability of inositol in the growth medium (Carman and Henry 1999). While the mechanism by which the cell regulates INO1 in response to inositol and other precursors of lipid metabolism has not been fully elucidated, it is known that the Ino2p and Ino4p transcription factors bind to UAS_INO to activate transcription (Carman and Henry 1989; Lopes et al. 1991; Ambroziak and Henry 1994; Schwank et al. 1995; Greenberg and Lopes 1996). Furthermore, the negative regulator, Opi1p, is required for repression of UAS_INO-containing genes in response to inositol (White et al. 1991). Recent evidence suggests that changes in the pattern of membrane phospholipid synthesis, produced in response to the incorporation of exogenous inositol, result in translocation of the negative regulator, Opi1p, to the nucleus from the endoplasmic reticulum and that Opi1p translocation coincides with repression of INO1 (Loewen et al. 2004).

Cox et al. (1997) suggested that activation of INO1 transcription in the absence of inositol might be linked to activation of the UPR signal transduction pathway. In support of this hypothesis, Cox et al. (1997) showed that the UPR is activated in wild-type cells growing in inositol-free medium and that deletion of OPI1 suppressed the Ino⁻ phenotype of ire1Δ and hac1Δ mutants. Previously, consistent with the hypothesis of Cox et al., we reported that expression of both UPR and INO1 is elevated in sec14-3^ts strains carrying the cki1Δ suppressor (Chang et al. 2002). However, deletion of HAC1 or IRE1 did not eliminate overexpression of INO1 in sec14^ts cki1Δ cells (Chang et al. 2002), suggesting that activation of the UPR, while correlated with INO1 expression, is not obligatory for INO1 activation.

It has been assumed that the Ino⁻ phenotype of UPR mutants, such as hac1Δ and ire1Δ, is due to the inability to maintain wild-type INO1 expression levels in the absence of UPR activation (Cox et al. 1997; Chang et al. 2002). Another possibility is that inositol deprivation results in a stress condition that elicits UPR activation and that UPR activation under these circumstances is essential in much the same way that it is in wild-type cells experiencing stress due to the buildup of unfolded proteins following exposure to tunicamycin. We propose that providing inositol to UPR mutants under such circumstances alleviates the underlying stress condition that necessitates UPR activation for survival, thus explaining the Ino⁻ phenotype. Since inositol limitation (Cox et al. 1997; Chang et al. 2002) and secretory stress (Figure 1) both result in UPR activation, we questioned whether these two stress conditions might have additive effects upon the level of UPRE and/or INO1 expression. Indeed, shifting the sec13^ts-1 strain to inositol-free medium at the semipermissive temperature of 30° resulted in an increase of ∼30% in UPRE expression over the already elevated levels seen in this strain at 30° in the presence of inositol (Figure 2).

Inositol limitation and secretory stress, thus, affect UPR activation in an additive or synergistic fashion. In Sec⁻ cells growing near their restrictive temperatures, the additive stress caused by lack of inositol may, in some cases, exceed the stress tolerance of even those cells with an intact UPR. Provision of inositol under such circumstances would reduce the stress that results in UPR induction, potentially explaining the conditional Ino⁻ phenotype of sec13^ts-1 cells near their restrictive temperature. Consistent with this idea, transformation of sec13^ts-1 with HAC1ⁱ permits growth at temperatures up to 32° with or without inositol (Figure 5), whereas the parent sec13^ts-1 strain transformed with the vector grows well only up to ∼30° and only if inositol is present (Figure 5).

While inositol limitation and secretory stress had an additive effect upon UPRE expression levels, no correlated additive effect on INO1-CYC-lacZ expression was observed in sec13^ts-1 cells grown in the absence of inositol at semipermissive temperatures (Figure 2A). To the contrary, in sec13^ts-1 cells shifted to 30° in the absence of inositol, INO1-CYC-lacZ expression was reduced to less than one-half of the level observed in the wild-type control grown under identical conditions (Figure 2C). Similar reductions in INO1-CYC-lacZ expression levels, relative to the wild-type control were observed in all of the Sec⁻ mutants assayed after growth in inositol-free media at semipermissive conditions (Figure 2A). The absence of a correlation between UPRE and INO1 expression was even more apparent in sec13^ts-1 cells grown at 30° in the presence of inositol (compare Figure 2B and 2C). Under these growth conditions, UPRE expression was greatly elevated in sec13^ts-1 cells (Figure 2B). Yet, INO1-lacZ expression in the sec13^ts-1 strain at 30° was fully repressed when inositol was present (Figure 2C). Clearly, activation of the UPR during secretory stress does not result in activation of INO1 when inositol is present.

The sec14-3^ts mutation, unlike the other Sec⁻ mutations analyzed here, does not affect an immediate component of the secretory apparatus itself. Rather, Sec14p is a lipid transfer protein that binds both phosphatidylinositol and phosphatidylcholine (Aitken et al. 1990). sec14 mutations result in a wide range of changes in both lipid metabolism and membrane trafficking (Novick et al. 1980, 1981; Bankaitis et al. 1989; Aitken et al. 1990; Cleves et al. 1991; Kearns et al. 1997; Patton-Vogt et al. 1997; Henry and Patton-Vogt 1998; Sreenivas et al. 1998; Xie et al. 1998). Surprisingly, despite the wide range of Sec⁻ mutants that exhibited UPRE activation at their semipermissive temperatures, we did not observe this effect in the sec14-3^ts single mutant. The failure to observe UPRE activation in the sec14-3^ts strain at the semipermissive temperatures of 30° and 33° was very surprising given that high levels of UPRE activation were previously observed at both 30° and 37° (Chang et al. 2002) in sec14-3^ts strains carrying the bypass suppressor, cki1Δ. Inada and Guthrie (2004), however, recently reported active splicing of HAC1 mRNA, indicative of UPR activation, in the sec14-3^ts mutant shifted to its restrictive temperature. Thus, it is possible that we failed to observe UPR activation in the sec14-3^tsmutant because we used a less direct assay (i.e., a UPRE reporter construct vs. HAC1 mRNA splicing) and different growth conditions (i.e., semipermissive temperature vs. a transient shift to the restrictive condition) than did Inada and Guthrie (2004). At 30° and 37°, the sec14-3^ts cki1Δ double mutant exhibits high levels of INO1 expression and inositol prototrophy and, in fact, overproduces and excretes inositol into the growth medium (Patton-Vogt et al. 1997; Chang et al. 2002). In contrast, the sec14-3^ts single mutant exhibits inositol auxotrophy and lowered INO1 expression at the semipermissive temperature of 33°. The reduction in INO1 expression in sec14^ts-3 under semipermissive growth conditions was similar to that in other Sec⁻ strains studied here, suggesting that lowered INO1 expression is a general response to impairment of secretory function. The elevated INO1 expression that we reported previously in the sec14^ts-3 cki1Δ strain, and other sec14 strains carrying bypass suppressors affecting the CDP-choline pathway for phosphatidylcholine synthesis, is presumably a consequence of the specific mechanism of suppression (Patton-Vogt et al. 1997; Henry and Patton-Vogt 1998; Chang et al. 2002).

Regardless of the explanation for the general lack of correlation between UPR activation and INO1 expression in Sec⁻ mutants, it is clear that INO1 expression is reduced in Sec⁻ mutants experiencing secretory stress. The inhibitory effect of secretory stress on transcription of INO1 is similar to the effect of secretory stress on transcription of rRNA and ribosomal protein genes. Mizuta and Warner (1994) showed that the function of the entire secretory pathway is essential for ribosomal synthesis. Nierras and Warner (1999) subsequently demonstrated that rRNA transcription and ribosomal protein synthesis are slowed in cells undergoing secretory stress and that this response is not transduced by the UPR. Rather, they concluded that the effect on rRNA and ribosomal proteins genes in cells with secretory defects may be controlled by a mechanism involving the protein kinase C signal transduction pathway related to membrane stress. In similar fashion, under conditions of secretory stress, signals other than those generated by the UPR may take precedence in controlling expression of INO1 and other UAS_INO-containing genes. INO1 expression is known to be influenced by growth phase and nutrient availability Griac and Henry (1999) and the glucose response pathway has also been shown to influence the levels of INO1 expression (Ouyang et al. 1999; Shirra and Arndt 1999; Shirra et al. 2001). For these reasons, we are currently examining the relative roles of several other signal transduction pathways in transducing signals from the secretory pathway to the regulatory apparatus controlling INO1 transcription.

We are indebted to Elizabeth Jones, Randy Schekman, Chris Kaiser, Jeff Brodsky, and Ruth Collins for strains; Kazutoshi Mori for plasmids bearing HAC1 and HAC1ⁱ; and Chris Kaiser for the plasmid carrying SEC13. This report is taken in part from the Ph.D. Thesis of H.J.C. (Carnegie Mellon University, 2001). This work was supported by National Institutes of Health grant GM-19629 to S.A.H.