Evidence for the Involvement of the Glc7-Reg1 Phosphatase and the Snf1-Snf4 Kinase in the Regulation of INO1 Transcription in Saccharomyces cerevisiae

Corresponding author: Karen M. Arndt, Department of Biological Sciences, University of Pittsburgh, 269 Crawford Hall, Pittsburgh, PA 15260., arndt{at}vms.cis.pitt.edu (E-mail)

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Binding of the TATA-binding protein (TBP) to the promoter is a pivotal step in RNA polymerase II transcription. To identify factors that regulate TBP, we selected for suppressors of a TBP mutant that exhibits promoter-specific defects in activated transcription in vivo and severely reduced affinity for TATA boxes in vitro. Dominant mutations in SNF4 and recessive mutations in REG1, OPI1, and RTF2 were isolated that specifically suppress the inositol auxotrophy of the TBP mutant strains. OPI1 encodes a repressor of INO1 transcription. REG1 and SNF4 encode regulators of the Glc7 phosphatase and Snf1 kinase, respectively, and have well-studied roles in glucose repression. In two-hybrid assays, one SNF4 mutation enhances the interaction between Snf4 and Snf1. Suppression of the TBP mutant by our reg1 and SNF4 mutations appears unrelated to glucose repression, since these mutations do not alleviate repression of SUC2, and glucose levels have little effect on INO1 transcription. Moreover, mutations in TUP1, SSN6, and GLC7, but not HXK2 and MIG1, can cause suppression. Our data suggest that association of TBP with the TATA box may be regulated, directly or indirectly, by a substrate of Snf1. Analysis of INO1 transcription in various mutant strains suggests that this substrate is distinct from Opi1.

CELL growth and differentiation depend upon accurate gene expression in response to signals from the environment. These signals must be transduced through the cell, and many stimuli ultimately effect activation or repression of transcription. Regulation of transcription requires interactions between sequence-specific activators and repressors, coactivators and corepressors, and the RNA polymerase II general transcription factors. Previous work supports two primary models in explaining assembly of the RNA polymerase II preinitiation complex in response to transcriptional activators. According to both models, promoters are first recognized by the general factor TFIID, which consists of the TATA-binding protein (TBP) and TBP-associated factors (reviewed in BURLEY and ROEDER 1996 ). One model argues that after TFIID binding to the promoter, the other components of the preinitiation complex assemble in a stepwise manner (BURATOWSKI et al. 1989 ). In the second model, binding of TFIID to the promoter is followed by the recruitment of a protein complex termed the RNA polymerase II holoenzyme (reviewed in PTASHNE and GANN 1997 ).

Using genetic and biochemical approaches, several groups have investigated the regulation of TBP-TATA complex formation. Both in vitro and in vivo, binding of TBP to the TATA box has been shown to be an important rate-limiting step in transcription (reviewed in STARGELL and STRUHL 1996 ). Direct interactions between TBP and certain gene-specific transcriptional activators and repressors have been reported (for examples see STRINGER et al. 1990 ; HORIKOSHI et al. 1991 , HORIKOSHI et al. 1995 ; EMILI et al. 1994 ; MELCHER and JOHNSTON 1995 ; UM et al. 1995 ; ZHANG et al. 1996 and references therein). In addition, genetic studies in Saccharomyces cerevisiae have identified several proteins that more generally affect TATA box binding by TBP, including Mot1, Spt3, Rtf1, and the Not proteins (AUBLE et al. 1994 ; COLLART 1996 ; MADISON and WINSTON 1997 ; STOLINSKI et al. 1997 ). However, the mechanisms by which these factors modulate the activity of TBP in vivo are not well understood. Therefore, studies of TBP mutants that are impaired in their response to certain activators may provide insights into the significant problem of promoter-specific regulation. We have identified such a class of TBP mutants and determined that these mutants have severely reduced affinity for DNA in vitro (ARNDT et al. 1995 ). The identification of these and similar TBP mutants (KIM et al. 1994 ; LEE and STRUHL 1995 ) has further established the importance of TATA box binding as a regulatory step in transcription.

While TBP and the other general transcription factors have been extensively studied for their interactions with activator proteins and promoter DNA in vitro, the complex regulatory circuitry used by cells to modulate the activity or assembly of the preinitiation complex in response to environmental cues is less well understood. Many of our existing insights into this important problem have come from studies in yeast, where transcriptional responses to signals such as glucose and inositol availability have been well documented.

In the presence of high glucose levels, many genes in yeast are repressed, including those encoding proteins needed to metabolize other carbon sources (reviewed in GANCEDO 1998 ; JOHNSTON and CARLSON 1992 ; TRUMBLY 1992 ; RONNE 1995 ). Genetic analyses of glucose repression, using regulation of SUC2 expression as a paradigm, have identified a complex network of negative and positive regulatory proteins. Current results favor a model in which high glucose levels, which are monitored by the HXK2 gene product, hexokinase PII, signal the Glc7 phosphatase to negatively regulate the activity of the Snf1 serine-threonine kinase (JOHNSTON and CARLSON 1992 ). Important components of this pathway are the REG1 and SNF4 gene products, which bind to and control the Glc7 phosphatase and the Snf1 kinase, respectively, in a positive manner (TU and CARLSON 1995 ; JIANG and CARLSON 1996 ). During derepression of SUC2 in low-glucose conditions, the model argues that the Snf1-Snf4 kinase reduces the activity of repressor complexes comprised of Ssn6(Cyc8), Tup1, and the sequence-specific binding proteins Mig1 and Mig2 (TREITEL and CARLSON 1995 ; LUTFIYYA and JOHNSTON 1996 ). In this way, the Snf1-Snf4 kinase, in conjunction with the Swi-Snf chromatin-remodeling complex, may counter a repressive chromatin state maintained by Ssn6 and Tup1 (EDMONDSON et al. 1996 ; GAVIN and SIMPSON 1997 ). In addition, the kinase may stimulate function of the RNA polymerase II holoenzyme (CARLSON 1997 ).

Another well-studied signaling pathway in yeast affects transcription in response to inositol availability. In the presence of high inositol levels, transcription of the INO1 gene, which encodes inositol-1-phosphate synthase, is strongly repressed by the OPI1 and UME6 gene products (KLIG et al. 1985 ; JACKSON and LOPES 1996 ). Opi1 contains a putative leucine zipper adjacent to a basic domain, a motif that has been postulated to bind DNA (WHITE et al. 1991 ); however, direct DNA binding by Opi1 has not been demonstrated (S. A. HENRY, personal communication). Ume6 binds to URS1 elements in several promoters, including INO1, and recruits a protein complex containing Sin3 and the Rpd3 histone deacetylase (KADOSH and STRUHL 1997 ; KASTEN et al. 1997 ). In response to low inositol levels, the Ino2 and Ino4 basic helix-loop-helix proteins activate INO1 transcription (AMBROZIAK and HENRY 1994 ). The proteins that transmit the inositol signal to these transcriptional regulatory factors are not completely known, partly because of the complexity of overlapping pathways that impinge on INO1 expression (HENRY and PATTON-VOGT 1998 ). However, snf1 strains are inositol auxotrophs, suggesting that the Snf1-Snf4 kinase may be involved in INO1 regulation (HIRSCHHORN et al. 1992 ). Furthermore, mutations that affect components of the RNA polymerase II holoenzyme, the Swi-Snf complex, and the SAGA histone acetyltransferase complex cause inositol auxotrophy (reviewed in HENRY and PATTON-VOGT 1998 ), suggesting that the inositol signal ultimately influences preinitiation complex formation and chromatin structure.

In earlier work, we reported the isolation of TBP mutants that cause inositol auxotrophy as well as a defect in galactose metabolism (ARNDT et al. 1995 ). These Ino^- and Gal^- phenotypes correlate strongly with severely impaired induction of the INO1, GAL1, and GAL10 genes (ARNDT et al. 1995 ). The ability of these TBP mutants to sustain growth of yeast strains indicates that their effects are gene specific. Indeed, transcription of certain other induced and constitutively expressed genes is relatively unaffected in the mutant strains (ARNDT et al. 1995 ). In vitro, the TBP mutants are severely defective for TATA box binding. By selecting for genetic suppressors of a TBP mutant in this class, TBP-P109A, we have sought to identify factors that regulate the formation or stability of the TBP-TATA complex in a promoter-specific manner. Here, we report the identification of four genes that, when mutated, significantly restore INO1 transcription to the TBP mutant strain. The finding that two of these genes are REG1 and SNF4 implicates proteins previously described for their roles in glucose repression in the control of INO1 transcription. Our results are consistent with a model in which the Snf1 kinase regulates preinitiation complex assembly at the INO1 promoter, perhaps by facilitating the TBP-TATA interaction.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic methods and media:
Rich (YPD), minimal (SD), synthetic complete (SC), 5-fluoroorotic acid (5-FOA), as well as presporulation and sporulation media were prepared as described (ROSE et al. 1990 ). Auxotrophic requirements were scored on SD media supplemented with the appropriate nutrients or SC media lacking the appropriate nutrients. Canavanine tests were performed using SC media lacking arginine and containing 60 mg/liter canavanine sulfate (SC-Arg+CAN, SCHILD et al. 1981 ). 2-Deoxyglucose media contained YEP (1% yeast extract, 2% Bacto-peptone), 2% sucrose, 200 µg/ml 2-deoxyglucose (Sigma, St. Louis), and 1 µg/ml antimycin A (Sigma), which was added to simulate anaerobic conditions. Media lacking inositol (-Ino) were prepared as described (SHERMAN et al. 1981 ); control media (+Ino) contained 200 µM inositol. Solid -Ino media contained adenine, uracil, and the amino acids histidine, lysine, leucine, and tryptophan (SD-Ino). Because we noticed that adenine caused a small induction of INO1 transcription (M. K. SHIRRA and K. M. ARNDT, unpublished observations), liquid -Ino media did not contain adenine. For growth of spt15-328 hxk2 and spt15-328 reg1 double mutants, solid -Ino media contained adenine, uracil, and all 20 amino acids (SC-Ino). Transformation of yeast cells was performed using the lithium acetate procedure (ITO et al. 1983 ). Plasmids were recovered from yeast as described (ARNDT et al. 1995 ) and transformed into Escherichia coli strain MH1 (HALL et al. 1984 ) for propagation. Plasmids were transformed into E. coli strain DH5 for sequencing.

Yeast strains:
With the exception of MCY2616 (TU and CARLSON 1994 ), Y153 (DURFEE et al. 1993 ), and KA20, all strains are congenic with FY2, a GAL2⁺ ura3-52 derivative of S288C (WINSTON et al. 1995 ; Table 1). Strains were constructed by standard methods for tetrad analysis (ROSE et al. 1990 ) or one-step gene replacement (ROTHSTEIN 1983 ). Strains containing the suppressor mutations in an SPT15⁺ background were identified from the nonparental ditypes of crosses performed to determine if the suppressor mutations were linked to SPT15⁺ (see below) and were confirmed by reconstruction of the suppression. reg1 strains and hxk2 strains were constructed by transforming KY494 with the 4.2-kb XbaI-EcoRI fragment from pUCsrn1::URA3 (TUNG et al. 1992 ) or with the 3.0-kb HindIII-PvuII fragment from pMR226 (gift from K.-D. Entian), respectively, and sporulating the resulting transformants. tup1 strains were constructed by transforming KY496 with the 3.6-kb SacI-SpeI fragment from pMC134 (gift from H. Ronne) and sporulating the resulting transformants. KA20 and KY499 were obtained as spores from matings between KY214 and MCY2616 or FY1179, respectively. ssn6 strains and opi1 strains were constructed by PCR-mediated, one-step gene disruptions (AUSUBEL et al. 1998 ) in KY494 followed by sporulation and dissection. Oligonucleotides used to amplify ssn6::HIS3 were as follows: 5'-GCTATAAGCCTTTAGACTAGTACTACAACTACAACAGCAACTGTGCGGTATTTCACACCG-3' and 5'-TGATTATAAATTAGTAGATTAATTTTTTGAATGCAAACTTAGATTGTACTGAGAGTGCAC-3'. Oligonucleotides used to amplify opi1::HIS3 were as follows: 5'-TGTTTACAGTGCTGATTAAAGCGTGTGTATCAGGACAGTGCTGTGCGGTATTTCACACCG-3' and 5'-TTACTGGTGGTAATGCATGAAAGACCTCAATCTGTCTCGGAGATTGTACTGAGAGTGCAC-3'. Disruptions of REG1, TUP1, SSN6, and OPI1 were confirmed by Southern blotting (AUSUBEL et al. 1998 ). Disruption of HXK2 or the presence of glc7-T152K in strains was confirmed by resistance to 2-deoxyglucose (LOBO and MAITRA 1977 ; NEIGEBORN and CARLSON 1987 ). Because ssn6 strains are also temperature sensitive, the SPT15 genotype of KY529 and KY530 was determined by backcrossing to strains containing SPT15⁺ SSN6⁺. The trp1 genotype of KA20 was determined by PCR analysis. KY503 was created by transforming KY108 with pPS52 (see below), which had been linearized with EcoRI. Insertion of an additional copy of SNF4, marked by URA3, was confirmed by Southern blotting (AUSUBEL et al. 1998 ).

Isolation of extragenic suppressors of spt15-328:
To reduce the likelihood of recovering true revertants of spt15-328 (see RESULTS), we created derivatives of KY214 and KY484 that carried pPS5, a 2µ plasmid containing spt15-328 and URA3. Sixteen individual colonies from each strain were patched onto SC-Ura media and replica plated to media lacking inositol and uracil. Patches were mutagenized with UV radiation of 0–1500 µJ/cm² in a Stratalinker (Stratagene, La Jolla, CA). No more than one Ino⁺ candidate was purified from each patch to ensure that all suppressor candidates were independently derived. Because strains containing spt15-328 can become polyploid (K. M. ARNDT, unpublished observation), we monitored whether the suppressor candidate strains were haploid using a modified canavanine test (SCHILD et al. 1981 ). Patches of candidate strains were replica plated onto SC-Arg+ CAN media, UV irradiated with 9000 µJ/cm² in a Stratalinker, and allowed to grow at 30° for 4–5 days. Patches originating from haploid strains are much more likely to contain canavanine-resistant papillations. After purification, ploidy analysis, and passage on 5-FOA, 14 haploid strains were isolated that contained suppressors of the Ino^- phenotype of spt15-328 in the absence of pPS5.

To determine whether the suppressor mutations were extragenic to SPT15, strains containing spt15-328 with the suppressor mutations were crossed to KY87 or FY630, and the resulting tetrads were analyzed for their pattern of growth on SD-Ino media. None of the 14 candidates were linked to SPT15. To determine whether the mutations were dominant or recessive and if the suppression phenotype was caused by a single mutation, double-mutant strains were crossed to either KY214 or KY484. To compensate for the sporulation defect of spt15-328 homozygous diploids (K. M. ARNDT, unpublished observation), the KY214 and KY484 parents were first transformed with pDE28-6, a CEN/ARS plasmid containing URA3 and SPT15⁺ (EISENMANN et al. 1989 ). Diploids resulting from these crosses were patched and replica plated to 5-FOA media to remove pDE28-6. Recessive mutations were scored as an inability of the diploids to grow on SD-Ino media. Diploids still containing pDE28-6 were allowed to sporulate, and a 2:2 Ino⁺:Ino^- phenotype in the resulting tetrads, after passage on 5-FOA media, indicated that suppression was caused by a mutation in a single gene. Complementation and linkage analysis among the candidates showed that these suppressors comprised four linkage groups: three represented by only recessive mutations and one represented by only dominant mutations.

Cloning of suppressor genes:
A YCp50-based yeast genomic library (ROSE et al. 1987 ) was transformed into strains KY486 and KY489, which contain recessive suppressor mutations, and transformants were screened for complementation of the Ino⁺ phenotype. Four plasmids complemented the Ino⁺ phenotype of KY486, and two plasmids complemented the Ino⁺ phenotype of KY489. For each complementing plasmid, the sequence of ~100 bp of genomic DNA was determined, and a search of the Saccharomyces Genome Database (http://genome-www.stanford.edu/Saccharomyces/) was performed. The four plasmids that complemented KY486 contained overlapping sequences on chromosome IV (inserts numbered per the Saccharomyces Genome Database): pPS24 (497819–504852), pPS25 (498118–514935), pPS26 (497274–507787), and pPS53 (494534–509421). Both plasmids that complemented KY489 contained overlapping sequences on chromosome VIII (inserts numbered per the Saccharomyces Genome Database): pPS32 (58894–68065) and pPS33 (61812–69897). Subcloning (see below) identified the complementing genes for KY486 and KY489 to be REG1 and OPI1, respectively. Analysis of tetrads sporulated from a cross between KY534 and KY487 showed that OPI1 is the correct gene.

To identify the dominantly acting suppressor in KY485, a plasmid library of KY485 genomic DNA in vector pRS316 (SIKORSKI and HIETER 1989 ) was constructed following the protocol of THOMPSON et al. 1993 . One plasmid conferred an Ino⁺ phenotype when transformed into KY214. The plasmid, pPS34, contained genomic sequences from 285282 to 294950 of chromosome VII. Subcloning (see below) localized the complementing activity to the SNF4 gene. Analysis of tetrads sporulated from a cross between KY485 and KY501 showed that SNF4 is the correct gene.

To identify the SNF4 mutation in KY504, the 7.8-kb BglII-AvrII fragment from pPS34 was transformed into KY504 for gap repair (ORR-WEAVER et al. 1983 ). The recovered plasmid, pPS60, contains the SNF4-313 gene, which encodes a substitution of aspartic acid for glycine at position 145 of Snf4 (Snf4-G145D).

Plasmids:
Standard molecular techniques were used for plasmid constructions (AUSUBEL et al. 1998 ). pKA86 is a derivative of pKA75 (ARNDT et al. 1995 ), which contains the spt15-328 gene in place of SPT15⁺. The 2.4-kb XhoI (from the polylinker)-BamHI fragment from pKA86, which contains spt15-328 sequences, was subcloned into the same sites in pRS426 (SIKORSKI and HIETER 1989 ) to create pPS5. To identify the gene encoding the suppressor in KY486, a 5.0-kb EcoRI fragment from pPS26, which contains the entire open reading frame (ORF) of REG1, was subcloned into the EcoRI site in pRS316 (SIKORSKI and HIETER 1989 ) to create pPS27. Interestingly, one of the original complementing plasmids, pPS25, lacked the 80 C-terminal amino acids of Reg1. To eliminate the involvement of a second potential ORF, YDR030C, which is also encoded in pPS27, a 1.1-kb EcoRI-ClaI fragment from pPS24 that encodes only YDR030C was subcloned into the same sites in pRS316 to create pPS29. pPS29 did not encode complementing activity. To identify the gene encoding the suppressor in KY489, a 1.9-kb XhoI fragment from pPS32 containing OPI1 was subcloned into the XhoI site in pRS316 to create pPS31. To determine the location of the mutation in SNF4, 1.8-kb HindIII-NcoI (blunted) fragments from pPS34 and pFE27-2, which contain wild-type SNF4 sequences (CELENZA et al. 1989 ), were subcloned into pRS316 that was digested with HindIII and XbaI (blunted) to create pPS47 and pPS48, respectively. The 0.8-kb XhoI (from the polylinker)-EcoRI fragment from pPS48, which encodes the promoter and N-terminal sequences of wild-type Snf4, was substituted for the corresponding sequences in pPS47 to create pPS49. Phenotypic analysis showed that pPS49 still contained the SNF4-204 mutation. Sequencing confirmed the location of the mutation (see RESULTS). pPS52 was constructed by subcloning the 1.8-kb KpnI-SacI fragment containing SNF4 sequences from pPS48 into the corresponding sites in pRS306 (SIKORSKI and HIETER 1989 ).

For the two-hybrid analysis, pGBT9 and pGAD424 were obtained from Clontech (Palo Alto, CA); pGBT9-SNF1 was a gift from M. C. Schmidt (TILLMAN et al. 1995 ). To facilitate subsequent subcloning, the 1.9-kb HindIII fragment from pNI12 (FIELDS and SONG 1989 ), which encodes the entire Snf4-Gal4 activation domain fusion protein, was inserted at the HindIII site in pRS425 (SIKORSKI and HIETER 1989 ) to create pPS50. A 1.0-kb SalI-AvrII fragment from pPS50, which contains the N terminus of SNF4, was substituted with the related sequence from pPS49 to create pPS51, which encoded Snf4-N177Y fused to the Gal4 activation domain. Most plasmid sequences are available upon request.

Northern hybridization analysis:
Cells were grown at 30° to a density of 1–2 x 10⁷ cells/ml in the appropriate media and induced as described in the figure legends (see also RESULTS). Isolation of RNA and Northern analyses was performed as described (ARNDT et al. 1995 ). Hybridization probes for INO1, TUB2, and GAL1-GAL10 were prepared from pJH310 (HIRSCH and HENRY 1986 ), pYST138 (SOM et al. 1988 ), and pGAL1-GAL10, respectively, by nick translation (Boehringer Mannheim, Indianapolis). pGAL1-GAL10 contains the EcoRI fragment 4812 (ST. JOHN and DAVIS 1981 ) subcloned into the same site in pUC18. Quantitation was performed using a FUJIX BAS2000 phosphorimager with MacBAS version 2.4 software.

Invertase assays:
Cells were grown at 30° to mid-log phase in YPD. For derepression of SUC2, 10 ml of cells was pelleted, washed twice with an equal volume of water, and resuspended in YEP + 0.05% glucose. Cells were allowed to grow for an additional 165 min at 30°. OD₆₀₀ was determined for both repressed and derepressed samples. Invertase assays were performed in duplicate on three isolates of each strain as described (GOLDSTEIN and LAMPEN 1975 ; BU and SCHMIDT 1998 ). Average values are reported. Standard errors were <17%.

Two-hybrid analysis:
Plasmids were transformed into Y153 (DURFEE et al. 1993 ) and selected on SC-Trp-Leu media containing 2% glucose. For quantitative ß-galactosidase assays, cells were grown to a density of 1–2 x 10⁷ cells/ml in either SC-Trp-Leu containing 2% glucose or SC-Trp-Leu containing 2% galactose, 2% glycerol, 2% ethanol, and 0.05% glucose. Extract preparations, ß-galactosidase assays, and unit calculations were performed as described (MILLER 1972 ; ROSE and BOTSTEIN 1983 ). Values represent the average of two experiments in which three transformants for each plasmid were assayed at two different extract concentrations. Standard errors were <15%.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of extragenic suppressors of a TBP mutant defective in activated transcription.
We previously reported the identification of the TBP mutant TBP-P109A, which is encoded by the spt15-328 gene. This mutant TBP exhibits promoter-specific defects in activated transcription in vivo and greatly reduced affinity for TATA boxes in vitro (ARNDT et al. 1995 ). We exploited the Ino^- phenotype of spt15-328 strains to identify, through genetic analysis, factors that may regulate TBP function. We reasoned that recessive suppressor mutations might uncover factors that negatively regulate the response of TBP to a particular activator or the binding of TBP to certain promoters, while dominant suppressor mutations might identify factors that assist these functions of TBP.

In our initial selections for suppressors of spt15-328, we repeatedly isolated intragenic suppressors that exhibited essentially wild-type phenotypes. Therefore, we modified our approach to enhance our ability to detect extragenic suppressors. In particular, we noticed that overexpression of spt15-328 has a slight dominant-negative effect in SPT15⁺ strains, causing a partial Ino^- phenotype compared with strains that do not express spt15-328 (M. K. SHIRRA and K. M. ARNDT, unpublished observations). Based on this finding, we conducted a selection for Ino⁺ suppressors using strains that expressed spt15-328 from both a 2µ plasmid (pPS5) and the endogenous chromosomal copy (see MATERIALS and METHODS for details). Twelve recessive and two dominant extragenic suppressors of the Ino^- phenotype of spt15-328, but no intragenic suppressors, were recovered. These suppressors comprise four linkage groups: three that correspond to previously identified genes (see below), and one that has not been cloned. We named the unidentified gene RTF2, for Restores TBP Function. RTF1 was reported previously as a suppressor of a different spt15 allele (STOLINSKI et al. 1997 ). All suppressor mutations restore growth to spt15-328 strains on -Ino media (Figure 1). However, they do not significantly affect the temperature sensitivity or Gal^- phenotype of spt15-328 strains. None of the suppressor mutations nor spt15-328 itself cause an Spt^- phenotype.

View larger version (95K): In this window In a new window Download PPT slide   Figure 1. Suppression of the Ino- phenotype of spt15-328. Representatives of four linkage groups that restored growth to KY214 in the absence of inositol were purified on YPD plates and replica plated to SD-Ino media or SD-Ino media supplemented with inositol. Photographs were taken after 2 days of growth at 30°. Strains used were as follows: SPT15+ (FY630), spt15-328 (KY214), SNF4-204 spt15-328 (KY485), reg1-230 spt15-328 (KY486), rtf2-315 spt15-328 (KY488), and opi1-319 spt15-328 (KY489).

For the recessive mutations, genes were cloned by complementation of the Ino⁺ phenotype of strains containing spt15-328 and individual suppressor mutations. Subcloning and DNA sequence analysis of complementing plasmids (see MATERIALS AND METHODS) showed that the smallest DNA fragment able to complement the suppressor mutation in KY486 included REG1, a gene primarily studied for its role during glucose repression (MATSUMOTO et al. 1983 ; NIEDERACHER and ENTIAN 1991 ; reviewed in JOHNSTON and CARLSON 1992 ; GANCEDO 1998 ). Demonstrating that we have cloned the correct gene, tetrad analysis showed that the KY486 suppressor mutation, like REG1, is tightly linked to TRP1 (90 parental ditypes, 0 nonparental ditypes, and 7 tetratypes among 97 complete tetrads). Complementation and linkage tests demonstrated that seven different suppressor strains contain recessive mutations in REG1. We found that the suppressor mutation in KY489 can be complemented by a plasmid carrying the OPI1 gene, a previously identified negative regulator of INO1 transcription (GREENBERG et al. 1982B ; WHITE et al. 1991 ). The strong Opi^- phenotype of KY489 suggested that the suppressor mutation in this strain is in OPI1, and this has been confirmed by linkage analysis (see MATERIALS AND METHODS). Four suppressor strains contain recessive mutations in OPI1. Repeated efforts to isolate a complementing clone for rtf2-315, the sole member of the RTF2 complementation group, have been unsuccessful, in part because this mutation causes a partially dominant phenotype. Furthermore, in screens to reveal additional phenotypes that might be useful in cloning RTF2, neither SPT15⁺ rtf2-315 nor spt15-328 rtf2-315 strains exhibited any differences relative to RTF2⁺ strains.

To clone the dominant suppressor mutation in KY485, a genomic library was constructed from this strain and transformed into an spt15-328 mutant. One plasmid that suppressed the Ino^- phenotype of KY214 was isolated. Subcloning and DNA sequencing of this plasmid, followed by linkage analysis (see MATERIALS AND METHODS), showed that the dominant suppressor mutation in KY485 lies in the gene for Snf4, a component of the Snf1 kinase complex that is required for derepression of glucose-repressible genes (SCHULLER and ENTIAN 1988 ; CELENZA et al. 1989 ; reviewed in JOHNSTON and CARLSON 1992 ; GANCEDO 1998 ). Moreover, sequencing of the mutant SNF4 gene showed that the mutation (SNF4-204) encodes a substitution of tryptophan for asparagine at position 177 (Snf4-N177Y). This asparagine is conserved between S. cerevisiae and Schizosaccharomyces pombe SNF4 homologs (GenBank accession no. 2130248), although it is not strictly conserved in homologs of other species (WILSON et al. 1994 ; GAO et al. 1996 ; PIOSIK et al. 1996 ; WOODS et al. 1996 ). Gap rescue followed by DNA sequence analysis showed that SNF4-313, a second isolate from this linkage group, encodes Snf4-G145D (see MATERIALS AND METHODS). This glycine is not conserved in homologs of other species, and no further characterization of this mutant protein has been performed.

Further evidence that the reg1, opi1, rtf2, and SNF4 mutations are involved in transcriptional control was obtained from their suppression of another activation-defective TBP mutant encoded by spt15-341 (ARNDT et al. 1995 ). Similar to spt15-328 strains, spt15-341 strains are inositol auxotrophs and inefficiently use galactose as a carbon source (ARNDT et al. 1995 ). In addition, the mutant TBP encoded by spt15-341 is severely impaired in TATA box binding in vitro (ARNDT et al. 1995 ). Using genetic crosses, we found that members of each of the four linkage groups could also suppress the Ino^- phenotype of spt15-341 (data not shown), demonstrating that suppression is not specific to the spt15-328 allele.

Suppressor mutations restore transcription of INO1 in TBP mutant strains:
To determine if the Ino⁺ phenotype of the suppressed strains correlated with an increase in INO1 transcription, we performed Northern analyses on the double-mutant strains (Figure 2A). A low concentration of inositol was used in the derepression media to allow partial derepression of INO1 while permitting growth of strains severely defective in INO1 transcription (GREENBERG et al. 1982A ; HIRSCH and HENRY 1986 ). Cells were induced for 10 hr, a time sufficient for maximal induction of INO1 transcription in the wild-type background (K. M. ARNDT, unpublished results). Under these conditions, all suppressor mutations restored transcription of INO1, from 10- to 30-fold, in the spt15-328 background (Figure 2A). These results suggest that the suppressor mutations affect the ability of the mutant TBP to support transcription initiation at the INO1 promoter.

View larger version (194K): In this window In a new window Download PPT slide   Figure 2. Suppressor mutations restore transcription of INO1 but do not greatly affect transcription of GAL1 or GAL10 in strains containing spt15-328. (A) Northern analysis of INO1 transcription. Repressed RNA samples (R) were obtained from cells grown in -Ino media supplemented with 200 µM inositol. Derepressed RNA samples (DR) were obtained from cells that were washed, resuspended in -Ino media supplemented with 10 µM inositol, and harvested after incubation at 30° for an additional 10 hr. Strains used were as follows: FY630 (lanes 1 and 2), KY214 (lanes 3 and 4), KY485 (lanes 5 and 6), KY486 (lanes 7 and 8), KY488 (lanes 9 and 10), and KY487 (lanes 11 and 12). (B and C) Northern analysis of GAL1 and GAL10 transcription. (B) Uninduced RNA samples (U) were obtained from cells grown in SC media containing 2% raffinose. Cells were induced (I) for GAL1 and GAL10 transcription by the addition of galactose to a final concentration of 5% and incubation of the culture for an additional 1.5 hr at 30°. Strains used were as follows: FY630 (lanes 1 and 2), KY214 (lanes 3 and 4), KY484 (lanes 5 and 6), KY486 (lanes 7 and 8), KY488 (lanes 9 and 10), and KY489 (lanes 11 and 12). (C) Uninduced RNA samples (0 min) were obtained from cells grown in SC media containing 3% glycerol and 2% potassium lactate (pH 5.7). Cells were induced for GAL1 and GAL10 transcription by the addition of galactose to a final concentration of 5%. A portion of the culture was harvested at the indicated times after the addition of galactose. Strains used were as follows: FY630 (lanes 1–4), KY214 (lanes 5–8), and KY485 (lanes 9–12). In each experiment, the filter from the top panel was reprobed for TUB2 mRNA. Quantitation is presented as the percentage of induced mRNA levels in the wild-type control, normalized to TUB2. Results from representative experiments are shown.

Although the suppressor mutations do not significantly affect the Gal^- phenotype of spt15-328 strains, we asked whether any subtle effects on GAL gene transcription could be detected by Northern analysis (Figure 2, B and C). Strains were grown in media (2% raffinose or 3% glycerol/2% lactate) that are nonrepressing and noninducing for GAL gene expression. Galactose was added subsequently to activate transcription of GAL1 and GAL10. [spt15-328 SNF4-204 double mutants are unable to grow in media containing raffinose as the sole carbon source (M. K. SHIRRA and K. M. ARNDT, unpublished results), necessitating the use of the alternative glycerol/lactate media.] No more than a twofold effect on GAL1 or GAL10 transcription was observed for any of the suppressor mutations in spt15-328 strains.

To examine the transcriptional effects of the suppressor mutations in a wild-type TBP background, Northern analysis of INO1 transcription was performed with SPT15⁺ strains containing the suppressors (Figure 3). Because SPT15⁺ strains will grow in the absence of inositol, induction of INO1 in these strains was achieved without the addition of low levels of inositol and was monitored for 4 hr; we have seen that maximal derepression of INO1 transcription in the absence of inositol occurs in 3–4 hr (for example, see Figure 4, lanes 2–5). Although strains containing SNF4-204 or reg1-230 show slightly reduced levels of INO1 transcription in derepressing conditions, rtf2-315 and opi1-319 have little effect on derepressed levels of INO1 mRNA. As seen for other opi1 mutants (GREENBERG et al. 1982A ; WHITE et al. 1991 ; ASHBURNER and LOPES 1995 ), the INO1 gene is expressed in an SPT15⁺opi1-319 strain in the presence of inositol (Figure 3, lane 9). In contrast, spt15-328 strains that contain the opi1-312 mutation, which is phenotypically similar to the opi1-319 mutation, do not express INO1 under repressing conditions (Figure 2A, lane 11, and data not shown). Taken together with the in vitro DNA-binding defect of the TBP mutant, these results argue for an additional OPI1-independent mechanism of inositol-mediated repression that may involve TATA box accessibility. As we describe below, this OPI1-independent repression appears to be regulated, at least in part, by the Snf1-Snf4 kinase pathway.

View larger version (72K): In this window In a new window Download PPT slide   Figure 3. In strains containing SPT15+, the suppressor mutations have little effect on derepressed levels of INO1 transcription. Northern analysis of INO1 transcription is shown. Repressed RNA samples (R) were obtained from cells grown in -Ino media supplemented with 200 µM inositol. Derepressed RNA samples (DR) were obtained from cells that were pelleted, washed, resuspended in -Ino media, and harvested after incubation at 30° for an additional 4 hr. The filter in the top panel was reprobed for TUB2 mRNA. Quantitation is presented as the percent of derepressed INO1 mRNA levels in the wild-type control, normalized to TUB2. Strains used were as follows: FY630 (lanes 1 and 2), KY491 (lanes 3 and 4), KY490 (lanes 5 and 6), KY493 (lanes 7 and 8), and KY492 (lanes 9 and 10). Results from a representative experiment are shown.

View larger version (39K): In this window In a new window Download PPT slide   Figure 4. INO1 transcription is not greatly affected by glucose levels. (A) Northern analysis of INO1 transcription. SPT15 (FY630) cells were grown in -Ino media supplemented with 200 µM inositol (lane 1). Portions of the culture were centrifuged, and cell pellets were washed and resuspended in -Ino media containing 2% glucose (lanes 2–5), -Ino media containing 0.05% glucose and 200 µM inositol (lanes 6–9), or -Ino media containing 0.05% glucose (lanes 10–13). Samples were harvested at the indicated times after incubation at 30°. The filter in the top panel was reprobed for TUB2 mRNA as a normalization control. (B) Quantitation of Northern analysis shown in A. Normalized INO1 mRNA levels are shown relative to the lowest level detected (lane 7), which was arbitrarily set to 1. Results from a representative experiment are shown.

The SNF4-204 mutation alters the interaction of Snf4 with Snf1:
As an initial characterization of the mutation in SNF4-204, we examined the ability of the SNF4-204 gene product, Snf4-N117Y, to interact with Snf1 in the two-hybrid system. The interaction between Snf1 and Snf4, which has been well documented with this assay (FIELDS and SONG 1989 ), is regulated by glucose levels such that low-glucose conditions significantly enhance formation of the Snf1-Snf4 complex (JIANG and CARLSON 1996 ). Therefore, we performed a two-hybrid analysis of Snf1 and Snf4-N177Y in media containing 2 or 0.05% glucose (Table 2). Strikingly, unlike Snf1 and wild-type Snf4, the interaction between Snf1 and Snf4-N177Y, as measured in ß-galactosidase units, was readily detected in the presence of high glucose. Furthermore, under low-glucose conditions, ß-galactosidase levels were twofold greater for the Snf1-Snf4-N177Y interaction pair than for Snf1 and wild-type Snf4. These results indicate that the dominant SNF4-204 mutation increases the affinity of Snf4 for Snf1 and renders the interaction independent of glucose levels.

Analysis of INO1 transcription in response to glucose:
The identification of opi1 in our selection for spt15 suppressors was not surprising. We expected to uncover negative regulators of INO1 transcription, since mutations in these factors might relieve a block in transcriptional activation to which the mutant TBP is particularly sensitive. However, the identification of mutations in REG1 and SNF4 was unanticipated. Although these genes have been implicated in a number of biological processes, they have been most extensively studied for their roles in glucose regulation of gene expression. The isolation of recessive alleles in REG1 and dominant alleles in SNF4 in our selection may indicate that these mutations function by relieving glucose repression of INO1 transcription. To test this idea, we have examined the effect of glucose levels on INO1 expression.

In the case of glucose-repressed genes such as SUC2, derepression of transcription occurs when the levels of glucose are reduced. To determine if a similar mechanism is involved in the regulation of INO1, we examined the levels of INO1 transcription by Northern analysis in high (2%) and low (0.05%) glucose (Figure 4). We found that INO1 transcription in low-glucose conditions is reproducibly elevated only twofold relative to high-glucose conditions 1 and 2 hr after shifting cells from repressing (high inositol) to inducing (no inositol) media (Figure 4, compare lanes 2 and 10 and lanes 3 and 11). However, the maximal levels of INO1 mRNA are unaffected by the glucose concentration (Figure 4, compare lanes 4 and 12). Note that in low-glucose conditions, INO1 transcription levels drop between the 3- and 4-hr time points, presumably because of the depletion of the carbon source (HENRY and PATTON-VOGT 1998 ). These data also show that in the presence of inositol, growth in low glucose is not sufficient to derepress INO1 (Figure 4, lanes 6–9).

Although these results indicate that glucose levels do not greatly affect INO1 expression in SPT15⁺ cells, we investigated whether a defect in the glucose repression pathway could be responsible for suppression of the Ino^- phenotype in spt15-328 strains. In addition, we wanted to compare any effect of glucose derepression to the effect of a suppressor mutation on INO1 transcription in an spt15-328 strain (Figure 5). We chose reg1-230 for this comparison because it showed a small but reproducible effect on GAL1 transcription in spt15-328 strains and, therefore, might be impaired in glucose repression. To mimic the growth conditions typically used to study glucose repression, we used a scheme similar to that used in Figure 4. Clearly, the small derepressing effect of low glucose on INO1 transcription in the spt15-328 background (less than twofold effect; compare Figure 5, lanes 5 and 9) is substantially less than the degree of suppression by reg1-230 under either low- or high-glucose conditions. These results suggest that suppression of spt15-328 by mutations in REG1 and SNF4 may not arise from the release of INO1 transcription from glucose repression.

View larger version (39K): In this window In a new window Download PPT slide   Figure 5. INO1 transcription in spt15-328 strains is not greatly affected by glucose levels. Northern analysis of INO1 transcription is shown. KY214 (lanes 1–9) or KY486 (lanes 10–18) cells were grown in -Ino media supplemented with 200 µM inositol (lanes 1 and 10). Portions of the culture were centrifuged, and cell pellets were washed and resuspended in -Ino media containing 2% glucose (lanes 2–5 and 11–14) or -Ino media containing 0.05% glucose (lanes 6–9 and 15–18). Samples were harvested at the indicated times after incubation at 30°. All lanes are from the same autoradiogram, but they have been rearranged for clarity of presentation. (B) Quantitation of Northern analysis shown in A. These values represent INO1 mRNA levels that have not been normalized to TUB2 mRNA levels, because we consistently noticed a decrease in TUB2 mRNA levels in spt15-328 reg1-230 double mutants grown in low glucose. However, ribosomal RNA levels in these samples were approximately equivalent, as determined by ethidium bromide staining. INO1 mRNA levels are shown relative to the lowest level detected (lane 10), which was arbitrarily set to 1. Results from a representative experiment are shown.

reg1-230 and SNF4-204 strains still exhibit glucose repression:
To evaluate whether the reg1 and SNF4 mutations we isolated are indeed defective in glucose repression, we examined the expression of a gene known to be regulated primarily through this pathway, SUC2. We measured the activity of invertase, the SUC2 gene product, in strains containing wild-type TBP and either reg1-230 or SNF4-204 (Table 3). As a control, we constructed a reg1 null allele in our strain background. In agreement with previous results (TU and CARLSON 1995 ; FREDERICK and TATCHELL 1996 ), a reg1 mutation caused derepression of SUC2 in the presence of high-glucose concentrations. Strikingly, our suppressor mutations did not relieve glucose repression of SUC2, even though the SNF4-204 product interacts with Snf1 in the presence of glucose (Table 2). These data also show that the reg1-230 mutation is not equivalent to a null allele. The SNF4-204 strain may be somewhat defective in SUC2 derepression, in agreement with the inability of spt15-328 SNF4-204 strains to grow in raffinose media. Importantly, independent of any effect on derepression, strains containing the reg1-230 and SNF4-204 alleles are still capable of repressing SUC2 expression under high-glucose conditions, suggesting that these mutations do not generally alleviate glucose repression.

Suppression of spt15-328 can be achieved by mutations in some but not all genes implicated in glucose repression:
To determine if any defect in the glucose-repression pathway could suppress spt15-328, we constructed double mutants between spt15-328 and null mutations in REG1, MIG1, TUP1, SSN6, and HXK2. Because a deletion of GLC7 is lethal (CLOTET et al. 1991 ; TU and CARLSON 1994 ), we introduced the glc7-T152K mutation into the spt15-328 background. The glc7-T152K mutation partially relieves glucose repression of gene expression (NEIGEBORN and CARLSON 1987 ) and diminishes the interaction of Glc7 with Reg1 (TU and CARLSON 1995 ). Double-mutant strains were examined for their ability to grow in the absence of inositol (Figure 6). To varying degrees, glc7-T152K, tup1, ssn6, and reg1 restore growth of spt15-328 on -Ino media. Preliminary Northern analyses showed that this suppression is occurring at the level of INO1 transcription (M. K. SHIRRA and K. M. ARNDT, unpublished observations). However, null alleles of MIG1 and HXK2 do not significantly suppress the Ino^- phenotype conferred by spt15-328. Together with the previous data, these findings suggest that suppression of the TBP mutant by reg1-230 and SNF4-204 is unrelated to the glucose-repression pathway and may represent a distinct role for these genes in INO1 transcription.

View larger version (77K): In this window In a new window Download PPT slide   Figure 6. The Ino- phenotype conferred by spt15-328 is suppressed by mutations in some but not all genes involved in glucose repression. Cells were grown in YPD to saturation, washed, and diluted to 1 x 108 cells/ml. A total of 3 µl of 10-fold serial dilutions (A and C) or 3-fold serial dilutions (B) were spotted onto solid media as indicated. (C) Because spt15-328 hxk2 and spt15-328 reg1 double mutants are unable to grow on minimal media, we supplemented our standard -Ino media (SD-Ino) with all 20 amino acids (SC-Ino) to examine suppression in these strains. Photographs were taken after (A and C) 3 days or (B) 5 days of growth at 30°. Strains used were as follows: FY630 (WT), KY214 (spt15-328), KY485 (SNF4-204), KY499 (mig1), KA20 (glc7-T152K), KY500 (tup1), KY529 (ssn6-a), KY530 (ssn6-b), KY502 (hxk2), and KY501 (reg1).

REG1 and SNF4 regulate INO1 transcription in a manner independent of OPI1:
To begin to address the mechanism by which the Snf1 kinase regulates INO1 transcription, we examined the genetic interactions between SNF1 and OPI1. snf1 mutants are inositol auxotrophs, while opi1 mutants overproduce inositol. We performed a genetic cross between KY531 and FY1193 to examine the epistatic relationship between these two genes. If snf1 mutations are unaffected by opi1 mutations, we would expect 2:2 segregation of the inositol auxotrophy in the resulting tetrads. Instead, we found 5 tetrads that show 2:2 segregation of Ino⁺:Ino^- phenotypes, 4 tetrads that show 4:0 segregation, and 24 tetrads that show 3:1 segregation. Among these tetrads were spores that could not efficiently use raffinose as the sole carbon source (Snf^- phenotype) but could grow on media lacking inositol. These data show that the opi1-319 mutation is epistatic to a snf1 mutation. This has been confirmed by Northern blot analysis, in which we found that INO1 is transcribed under repressing conditions in a snf1 opi1-319 strain (data not shown). These results could indicate that Snf1 and Opi1 function in the same pathway to regulate INO1 transcription. Alternatively, the opi1 mutation may phenotypically bypass the effect of the snf1 mutation.

To test the hypothesis that Snf1 operates through Opi1, we asked whether the degree of spt15-328 suppression caused by an opi1 mutation is affected by our reg1 and SNF4 mutations. If Reg1 and Snf4 modulate INO1 transcription solely through an effect on Opi1, then reg1 and SNF4 mutations, when combined with an opi1 mutation, should not increase the level of spt15-328 suppression relative to an opi1 mutation alone. To rule out any effect of Reg1 or Snf4 on Opi1 activity, we performed this analysis with an opi1 mutation. As shown in Figure 7, the opi1 mutation, like our original opi1-312 suppressor mutation, significantly restores INO1 transcription in the TBP mutant strain (lanes 1–4). Unlike the opi1-312 allele, however, the opi1 mutation renders INO1 transcription partially derepressed in the mutant TBP background in the presence of high levels of inositol (Figure 7, compare lanes 3 and 4). Introduction of the reg1-230 and SNF4-204 mutations into the spt15-328 opi1 background leads to further increases in INO1 transcription in both repressing and derepressing conditions (Figure 7, lanes 3–8). These results argue that the Snf1 kinase and the Opi1 repressor operate through different pathways to regulate INO1 transcription. In addition, the residual, Opi1-independent repression observed in high-inositol conditions (Figure 7, compare lanes 3 and 4) is alleviated by the reg1 and SNF4 mutations (compare lanes 3, 5, and 7), suggesting that the Snf1 kinase pathway counters a repressive mechanism that is inositol mediated but Opi1 independent. Finally, because the opi1 mutations derepress INO1 transcription to a greater extent in SPT15⁺ strains (Figure 3 and data not shown) compared with spt15-328 strains (Figure 2A and Figure 7), the mutant TBP appears to be more sensitive than wild-type TBP to this additional layer of INO1 repression.

View larger version (60K): In this window In a new window Download PPT slide   Figure 7. In opi1 strains, reg1 and SNF4 mutations elevate the level of spt15-328 suppression. Northern analysis of INO1 transcription is shown. Repressed RNA samples (R) were obtained from cells grown in -Ino media supplemented with 200 µM inositol. Derepressed RNA samples (DR) were obtained from cells that were washed, resuspended in -Ino media supplemented with 10 µM inositol, and harvested after incubation at 30° for an additional 10 hr. Strains used were as follows: KY214 (lanes 1 and 2), KY535 (lanes 3 and 4), KY536 (lanes 5 and 6), KY537 (lanes 7 and 8), KY532 (lanes 9 and 10), and KY533 (lanes 11 and 12). The filter from the top panel was reprobed for TUB2 mRNA. Quantitation is presented as INO1 transcription levels relative to lane 3, after normalization to TUB2. Results from representative experiments are shown.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To identify factors that regulate TBP function in vivo, we performed a genetic selection for suppressors of the spt15-328 gene product. We specifically searched for extragenic suppressors of the Ino^- phenotype of this TBP mutant because INO1 transcription is particularly sensitive to mutations that affect components of the RNA polymerase II transcription machinery (NONET and YOUNG 1989 ) and chromatin remodeling factors (GANSHEROFF et al. 1995 ; ROBERTS and WINSTON 1996 ; GRANT et al. 1997 ). In this way, we identified four genes that directly or indirectly affect TBP function at the INO1 promoter: OPI1, REG1, SNF4, and RTF2.

Identification of suppressor mutations in OPI1:
One model to explain the transcriptional properties of the TBP mutant argues that promoter-specific factors negatively control TATA box accessibility at the most highly affected genes. Therefore, we expected to isolate mutations in genes, such as OPI1, that encode repressors of INO1 transcription. Although the biochemical activity of Opi1 remains elusive, our genetic results suggest that Opi1 operates, at least in part, by impairing TBP function at the INO1 promoter.

The identification of an opi1 mutation in our selection suggested the possibility that disruption of any negative regulator of INO1 transcription could suppress the Ino^- phenotype of the TBP mutant. Other negative regulators of INO1 transcription include UME6 (JACKSON and LOPES 1996 ); HHF1, which encodes histone H4 (SANTISTEBAN et al. 1997 ); and SIN3 (HUDAK et al. 1994 ). We constructed deletions of SIN3 and RPD3 in our genetic background and tested their ability to suppress spt15-328. Instead of suppression, we found that double-mutant strains containing spt15-328 and either sin3 or rpd3 are inviable (M. K. SHIRRA and K. M. ARNDT, unpublished observations). While this synthetic lethality has intriguing implications for the functions of SIN3 and RPD3, it prevented an analysis of INO1 transcription in the double-mutant strains. Further tests of the specificity of suppression by opi1 mutations will require the use of mutations in genes that are less pleiotropic.

Identification of suppressor mutations in REG1 and SNF4:
In addition to the well-established importance of REG1 and SNF4 in glucose repression, various genetic results have indicated an involvement of these genes in other biological processes, such as RNA processing (PEARSON et al. 1982 ; TUNG et al. 1992 ; MADDOCK et al. 1994 ), glycogen accumulation (TU and CARLSON 1995 ; HUANG et al. 1996 ), and sporulation (CELENZA et al. 1989 ). In our study, we have uncovered a role for REG1 in the regulation of INO1 transcription. The inositol auxotrophy of snf1 mutant strains previously suggested a requirement for the Snf1-Snf4 complex in INO1 induction (HIRSCHHORN et al. 1992 ). In the accompanying article a mutation in REG1 was also identified in a search for suppressors of a mutant Ino4 transactivator, and a reg1 strain was shown to constitutively express INO1 (OUYANG et al. 1999 ).

Several results suggest that the functions of REG1 and SNF4 in INO1 regulation may be unrelated to their roles in glucose repression. First, glucose levels do not affect the maximal, induced levels of INO1 transcription and do not bypass the normal induction signal for this gene. Second, our reg1 and SNF4 mutations are not generally defective in glucose repression, as indicated by the high level of repression seen at SUC2. Third, mutations in MIG1 and HXK2, two other genes with well-known roles in glucose repression, do not suppress our TBP mutant. Fourth, we have found that unlike the singly mutated strains, spt15-328 reg1 double-mutant strains grow extremely slowly on rich media and are unable to grow on minimal media and that spt15-328 reg1-230 SNF4-204 triple mutants are inviable (M. K. SHIRRA and K. M. ARNDT, unpublished observations). Such synthetic growth defects suggest more global roles for REG1 and SNF4 in gene regulation. In agreement with this interpretation, others have noted functions for REG1 that are apparently distinct from its involvement in glucose repression (TUNG et al. 1992 ; FREDERICK and TATCHELL 1996 ; HUANG et al. 1996 ).

An alternative explanation for our results is that INO1 transcription is subject to a modest level of glucose repression (i.e., twofold), and that alleviation of this repression by our reg1 and SNF4 alleles is sufficient to overcome the transcriptional defect of the TBP mutant. At promoters that are more strongly repressed by glucose, such as SUC2, our reg1 and SNF4 mutations may be too weak to relieve repression. Because hxk2 and mig1 mutations do not relieve repression of all glucose-repressed genes (MARYKWAS and FOX 1989 ; MOEHLE and JONES 1990 ; LUNDIN et al. 1994 ), our genetic results cannot completely eliminate the possibility that glucose levels are regulating the INO1 promoter. Importantly, independent of the actual signal that is transduced by the Reg1-Glc7 phosphatase and the Snf1-Snf4 kinase, our data strongly suggest that this pathway directly or indirectly regulates the activity of TBP at the INO1 promoter.

Connections to the RNA polymerase II holoenzyme and chromatin:
At glucose-repressed promoters, the Snf1-Snf4 kinase regulates phosphorylation of Mig1 (TREITEL et al. 1998 ), negating the effects of the tethered Ssn6-Tup1 corepressor complex (TREITEL and CARLSON 1995 ; LUTFIYYA and JOHNSTON 1996 ). Our observation that tup1 and ssn6 mutations can moderately suppress the Ino^- phenotype conferred by spt15-328 suggests that a similar mechanism is operating at INO1. Since a mig1 does not suppress the Ino^- phenotype of our TBP mutant, we postulate that some other protein tethers the Ssn6-Tup1 complex to the INO1 promoter. Consistent with reports that Opi1 lacks DNA binding activity (S. A. HENRY, personal communication), our genetic results suggest that this protein is unlikely to be Opi1. In addition we have found that tup1 mutants do not have a strong Opi1^- phenotype (M. K. SHIRRA and K. M. ARNDT, unpublished observations), providing further evidence that Tup1 does not act through Opi1.

Based on previous results, two principal mechanisms can be proposed to explain suppression of the activation- and DNA-binding-defective TBP mutants by mutations in REG1, SNF4, SSN6, and TUP1. The promoter-specific effects of the TBP mutants suggest that TATA box accessibility may be more constrained at some promoters, such as INO1, than at others. Nucleosome positioning may be critical for this distinction. A combination of genetic, molecular, and biochemical data strongly support a role for the Ssn6-Tup1 complex in regulating chromatin structure (COOPER et al. 1994 ; EDMONDSON et al. 1996 ; GAVIN and SIMPSON 1997 ). Although one analysis of INO1 chromatin structure did not show any gross differences in the presence or absence of inositol (SANTISTEBAN et al. 1997 ), it would be interesting to examine nucleosome positioning at this promoter in our mutant strains.

Alternatively, our suppressor mutations may affect the RNA polymerase II holoenzyme, enabling it to compensate for a defective TBP. In support of this idea, truncations of the heavily phosphorylated C-terminal domain of the large subunit of RNA polymerase II result in inositol auxotrophy (NONET and YOUNG 1989 ). Selections for suppressors of these C-terminal domain truncations have identified a class of SRB genes that appear to play negative roles in gene regulation (CARLSON 1997 ). Interestingly, mutations in these same SRB genes suppress the effects of an snf1 mutation at SUC2 (SONG et al. 1996 ). Whether mutations in these holoenzyme components can restore INO1 transcription to our TBP mutant strains remains to be determined.

What is the target of the Snf1 kinase for INO1 regulation?
Our results imply that activation of the Snf1 kinase, either by inactivating Reg1 or by stimulating the Snf1-Snf4 interaction, is responsible for suppression of our TBP mutant. Thus, phosphorylation of some target appears to bypass the need for a completely functional TBP at the INO1 promoter. The identity of this target is unknown. We found that an opi1 mutation is epistatic to snf1 for INO1 transcription, suggesting that OPI1 acts downstream of the kinase. However, Northern analysis on double- and triple-mutant strains (Figure 7) revealed that the Snf1-Snf4 kinase does not act solely through Opi1 to suppress our TBP mutant. In the accompanying article, OUYANG et al. 1999 showed that a reg1 mutation suppresses the inositol auxotrophy of ino4-8, but not ino4 or ino2, mutant strains. This necessity for residual Ino4 function may indicate that the Ino2-Ino4 complex is a target of the Snf1-Snf4 kinase. In addition, components or regulators of chromatin or the RNA polymerase II transcription machinery may be substrates for the Snf1 kinase at the INO1 promoter.

In summary, by searching for suppressors of an activation-defective TBP mutant, we have implicated a pathway that includes the Reg1-Glc7 phosphatase and the Snf1-Snf4 kinase in transcription initiation at the highly regulated INO1 promoter. Together with our previous results, our current findings support a model in which the formation or stability of the TBP-TATA complex at the INO1 promoter, and perhaps at other promoters, may be regulated by a substrate of the Snf1 kinase.

	ACKNOWLEDGMENTS

We thank M. Carlson, K.-D. Entian, A. K. Hopper, H. Ronne, and M. C. Schmidt for their gifts of strains and plasmids; S. A. Henry and M. Ruiz-Noriega for sharing unpublished data; P. J. Costa and C. Thompson for help constructing the genomic library used to identify SNF4-204; K. Roinick for help with DNA