Specific Components of the SAGA Complex Are Required for Gcn4- and Gcr1-Mediated Activation of the his4-912 Promoter in Saccharomyces cerevisiae

Corresponding author: Fred Winston, Department of Genetics, Harvard Medical School, 200 Longwood Ave., Boston, MA 02115., winston{at}rascal.med.harvard.edu (E-mail)

Communicating editor: M. CARLSON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutations selected as suppressors of Ty or solo insertion mutations in Saccharomyces cerevisiae have identified several genes, SPT3, SPT7, SPT8, and SPT20, that encode components of the SAGA complex. However, the mechanism by which SAGA activates transcription of specific RNA polymerase II-dependent genes is unknown. We have conducted a fine-structure mutagenesis of one widely used SAGA-dependent promoter, the element of his4-912, to identify sequence elements important for its promoter activity. Our analysis has characterized three regions necessary for full promoter activity and accurate start site selection: an upstream activating sequence, a TATA region, and an initiator region. In addition, we have shown that factors present at the adjacent UAS_HIS4 (Gcn4, Bas1, and Pho2) also activate the promoter in his4-912. Our results suggest a model in which the promoter in his4-912 is primarily activated by two factors: Gcr1 acting at the UAS and Gcn4 acting at the UAS_HIS4. Finally, we tested whether activation by either of these factors is dependent on components of the SAGA complex. Our results demonstrate that Spt3 and Spt20 are required for full promoter activity, but that Gcn5, another member of SAGA, is not required. Spt3 appears to be partially required for activation of his4-912 by both Gcr1 and Gcn4. Thus, our work suggests that SAGA exerts a large effect on promoter activity through a combination of smaller effects on multiple factors.

TRANSCRIPTION initiation in eukaryotes is a complex process requiring the coordination of multiple DNA-protein and protein-protein interactions. RNA polymerase II-dependent promoters in Saccharomyces cerevisiae contain combinations of four elements: upstream activating sequences (UAS), upstream repressing sequences (URS), TATA elements, and sites of transcriptional initiation (for review, see STRUHL 1995 ). UAS and URS elements are binding sites for gene-specific activators and repressors, respectively, and usually determine a promoter's regulation in response to environmental stimuli. TATA elements are the site of binding for the TATA-binding protein (TBP) and for the assembly of the general transcription initiation factors. In S. cerevisiae, transcription often initiates at multiple sites within a window of 30–120 nucleotides 3' of the TATA element, although the sequence determinants of start site selection are poorly understood. The binding of transcription factors to the appropriate promoter elements facilitates the assembly of RNA polymerase II into a preinitiation complex.

In eukaryotes, transcription by RNA polymerase II is often regulated by multiple activators, repressors, TATA elements, and transcription initiation sites within the same promoter. Promoters in larger eukaryotes are often complex, spanning up to several thousand nucleotides and containing many promoter elements (for example, see THANOS and MANIATIS 1995 ). In contrast, RNA polymerase II-dependent promoters in S. cerevisiae are usually compact, consisting of only a few hundred nucleotides and containing only a few promoter elements, facts that have made yeast an excellent system for the study of transcription initiation in vivo. However, one class of complex promoters that exists in S. cerevisiae are the composite promoters generated by insertions of Ty1 and Ty2 retrotransposons into the 5' regions of other genes (for review, see WINSTON 1992 ). Because the Ty promoter resides in its 5' long terminal repeat (), the presence of solo elements often alters transcription from promoters into which they insert due to the introduction of additional promoter elements. One such insertion mutation is the well-characterized allele his4-912, in which a Ty1 element lies between the UAS and TATA elements of HIS4 (FARABAUGH and FINK 1980 ). Strains containing his4-912 are His^- because transcription initiates at the Ty mRNA start site within the , instead of at the HIS4 start site, resulting in a nonfunctional -HIS4 mRNA transcript (SILVERMAN and FINK 1984 ; HIRSCHMAN et al. 1988 ). The presence of multiple promoter elements makes his4-912 a good model for complex promoters such as those found in larger eukaryotes.

The UAS_HIS4 region, located ~25 bp upstream of the in his4-912, has been extensively characterized. This region, which contains binding sites for the activators Rap1, Bas1, Pho2 (Bas2/Grf10), and Gcn4 (ARNDT and FINK 1986 ; ARNDT et al. 1987 ; TICE-BALDWIN et al. 1989 ; DEVLIN et al. 1991 ), regulates HIS4 via two different pathways (ARNDT et al. 1987 ). The basal pathway is constitutively active and requires the binding of both Bas1 and Pho2 to their adjacent sites within the UAS_HIS4 (ARNDT et al. 1987 ; TICE-BALDWIN et al. 1989 ). The general amino acid control pathway is regulated by amino acid levels and is mediated by the binding of the strong transcriptional activator Gcn4 (HINNEBUSCH et al. 1985 ; HOPE and STRUHL 1985 ; ARNDT and FINK 1986 ). HIS4 transcription via both pathways is dependent on the transcriptional activator/repressor protein Rap1, which is believed to allow binding of the other factors by maintaining an accessible chromatin structure (DEVLIN et al. 1991 ). Only combinations of mutations in genes encoding components of both the constitutive and general amino acid control pathways, for example, both bas1 and gcn4, abolish promoter activity and confer a His^- phenotype (ARNDT et al. 1987 ). In addition to activating the HIS4 promoter, there is some evidence that Gcn4 at the UAS_HIS4 can activate the promoter of his4-912 (SILVERMAN and FINK 1984 ), although activation by UAS_HIS4 factors has not been extensively characterized.

In contrast to the HIS4 promoter, relatively few sequence elements important for promoter activity have been identified. Two previous deletion analyses demonstrated that the first 100 nucleotides of the are dispensable for full promoter activity (LIAO et al. 1987 ; FULTON et al. 1988 ). However, the two studies reached different conclusions regarding the importance of the UAS element. One study found no sequences upstream of the TATA region to be necessary for expression of a Ty1 element (FULTON et al. 1988 ), while a second study identified a UAS region important for promoter activity in a Ty2 element (LIAO et al. 1987 ). The activity of the Ty2 UAS is dependent on the transcriptional activator Gcr1, which can bind the UAS in vitro (TURKEL et al. 1997 ). The promoter of his4-912 also contains two consensus TATA elements that bind TBP with equal affinity in vitro (ARNDT et al. 1992 ). Although several studies have shown that mutations in the TATA region significantly decrease promoter activity (LIAO et al. 1987 ; CONEY and ROEDER 1988 ; FULTON et al. 1988 ; HIRSCHMAN et al. 1988 ; ARNDT et al. 1994 ), the relative contributions of the two consensus TATA elements present in many sequences have not been analyzed. Finally, the transcription start sites of both Ty1 and his4-912 have been mapped to the same single nucleotide (ELDER et al. 1983 ; SILVERMAN and FINK 1984 ), although no sequence determinants for this start site selection have been identified. Thus, previous studies have identified a region containing two consensus TATA elements and possibly a UAS element as being required for full promoter activity, although other regions of the have not been examined by a fine-structure analysis.

Despite the fact that sequence elements within the have not been thoroughly characterized, genetic analysis has identified a large number of trans-acting factors required for promoter function (for review, see WINSTON 1992 ). Many of these factors were originally identified by mutations that cause an Spt^- (Suppressor of Ty) phenotype, that is, the ability to suppress transcription of full-length Ty elements or solo insertion alleles such as his4-912. One class of these suppressors includes mutations in SPT3, SPT7, SPT8, and SPT20. Mutations in these genes share a set of mutant phenotypes with certain mutations in SPT15, which encodes the TATA-binding protein, and cause transcriptional defects in a subset of RNA polymerase II-dependent promoters (WINSTON 1992 ). Recently, these Spt proteins were shown to be components of the 1.8-MD SAGA (Spt/Ada/Gcn5/acetylase) complex (GRANT et al. 1997 ). Although SAGA was originally purified on the basis of its ability to acetylate histones in a nucleosomal template (GRANT et al. 1997 ), genetic evidence suggests that this large complex contains a variety of additional functions (HORIUCHI et al. 1997 ; ROBERTS and WINSTON 1997 ). However, the mechanism by which large transcription complexes such as SAGA are targeted to, and function at, specific subclasses of RNA polymerase II-dependent promoters is not yet understood.

In addition to components of the SAGA complex, genetic analysis has demonstrated that several other important classes of transcription factors are required for normal expression of his4-912. Components of RNA polymerase II (HEKMATPANAH and YOUNG 1991 ); the Srb/mediator complex, Gal11/Spt13 (FASSLER and WINSTON 1988 , FASSLER and WINSTON 1989 ), Rgr1 (JIANG and STILLMAN 1992 ), and Sin4 (JIANG and STILLMAN 1992 ); and the general transcription apparatus, TBP, (EISENMANN et al. 1989 ), TFIIA (MADISON and WINSTON 1997 ), and Mot1 (MADISON and WINSTON 1997 ) are all required for expression from the promoter of his4-912. Expression of his4-912 also depends on many factors that influence chromatin structure, including histones (CLARK-ADAMS et al. 1988 ; PRELICH and WINSTON 1993 ; SMITH et al. 1996 ; SANTISTEBAN et al. 1997 ); factors that regulate histone gene expression (SHERWOOD and OSLEY 1991 ; XU et al. 1992 ); Spt4, Spt5, and Spt6 (WINSTON et al. 1984A ); and the Snf/Swi complex (HAPPEL et al. 1991 ). The promoter of his4-912 also appears to be sensitive to epigenetic regulation (JIANG and STILLMAN 1996 ). Thus, further characterization of his4-912 may provide new information on the function of these factors and complexes at other promoters.

In this study, we have characterized the cis-acting elements necessary for the promoter activity of his4-912 by fine-structure mutagenesis of a Ty-lacZ fusion. We have analyzed the roles of several sequences in the : a UAS, a weak URS, two consensus TATA elements, and a region surrounding the transcription start site. We have also analyzed the requirement for transcriptional activators that act at the UAS_HIS4 (Bas1, Pho2, and Gcn4) and the UAS (Gcr1 and a currently unidentified factor) and the requirement for the SAGA complex. Our analysis suggests that the his4-912 promoter is activated primarily by two factors: Gcr1 at the UAS and Gcn4 at the UAS_HIS4. In addition, our results demonstrate that this activation is dependent on some but not all components of the SAGA complex. Surprisingly, one SAGA component, Spt3, appears to be partially required for the activity of a number of his4-912 promoter factors, including Gcr1 and Gcn4. Our work suggests a model in which specific SAGA components exert large effects on the promoter as a sum of smaller effects on multiple factors.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains, genetic methods, and media:
All yeast strains used in this study are isogenic and were originally derived from a GAL2 S288C derivative (WINSTON et al. 1995 ). ß-Galactosidase assays and primer extension analyses were performed on two to six independent strains of each relevant genotype, with minor variations in mating type and auxotrophic markers. A representative strain of each genotype is presented in Table 1. A complete strain list is available upon request. The spt3-202 (WINSTON and MINEHART 1986 ), spt20100::URA3 (ROBERTS and WINSTON 1996 ), gcn5::HIS3 (ROBERTS and WINSTON 1997 ), gcn4::LEU2 (DRYSDALE et al. 1995 ), bas1-2 (ARNDT et al. 1987 ), gcr1::LEU2 (BAKER 1986 ), and pho2::LEU2 (BRAZAS and STILLMAN 1993 ) mutations have all been described previously. Construction of the Ty91244-lacZ alleles is described below. Yeast strains were transformed by the lithium acetate procedure (ELBLE 1992 ). Standard methods of mating, sporulation, and tetrad analysis were used (ROSE et al. 1990 ). Rich (YPD), minimal (SD), synthetic complete (SC), 5-fluoro-orotic acid (5-FOA), and sporulation media were prepared as described previously (ROSE et al. 1990 ).

Construction of promoter mutations:
To facilitate the analysis of a large number of promoter mutations in a variety of mutant backgrounds, promoter mutations were constructed in a Ty-lacZ fusion. The structure of this reporter gene, Ty91244-lacZ, has been described previously (WINSTON et al. 1987 ) and contains the entire 334 bp of Ty1-912 and 54 bp of the epsilon region fused in-frame to the Escherichia coli lacZ gene. The Ty91244-lacZ fusion will be referred to as his4-912-lacZ and the sequence itself as 912. The promoter fusion does not contain the HIS4 TATA region or transcription start site, and thus all his4-912-lacZ expression is a result of promoter function. In the present study, each of the his4-912-lacZ alleles is integrated at the HIS4 locus, such that the position of the relative to the UAS_HIS4 is the same as the in his4-912 (FARABAUGH and FINK 1980 ).

To identify sequences important for 912 promoter activity, a series of small, clustered base pair substitutions was constructed between 912 nucleotides 90 and 250 (Figure 1 and Table 2). Each mutation alters the sequence of 6–11 consecutive base pairs and creates a restriction site that is unique in 912. In addition to the mutations listed in Table 2, three double mutants were constructed. The double UAS mutant contains both mutations 12 and 13 (Table 2). The double 13/18 mutant contains both mutations 13 and 18 (Table 2). The double TATA mutant was constructed in a previous study and alters the sequence of the TATA region from (5' TATAAACATATAAA 3') to (5' TGTAGACACTGCAG 3'), where the 3' TATA is replaced by a PstI site (ARNDT et al. 1994 ).

View larger version (22K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Relative positions of the mutations created in this study. The structure of the his4-912-lacZ reporter gene is shown, including the positions of the UASHIS4, which contains Rap1-, Pho2-, Bas1-, and Gcn4-binding sites, the consensus Gcr1 site, the consensus TATA elements, and the site of transcription initiation. The mutagenized region corresponds to nucleotides 90 through 250.

View this table: [in this window] [in a new window]   Table 2. sequence changes introduced by the promoter mutations used in this study

Mutation 23 and the double TATA (17/18) mutations were constructed in a previous study (ARNDT et al. 1994 ) and subcloned from pKA41 derivatives that contain the region III and IV TATA mutations, respectively (ARNDT et al. 1994 ). The remaining mutations were constructed by site-directed mutagenesis using methods of KUNKEL 1985 and HO et al. 1989 or the QuickChange mutagenesis kit (Stratagene, La Jolla, CA). The presence of the correct base-pair changes as well as the absence of secondary mutations was confirmed by DNA sequence analysis. To reduce the amount of DNA sequencing, a 320-bp PacI-BseRI fragment containing each of the 912 mutations created in pLG39 was subcloned into the same sites of an unmutagenized his4-912-lacZ in pAD1. Since mutation 25 abolishes the BseRI site, a larger region of the original pLG39 derivative was sequenced and used for one-step gene replacement technique. Each his4-912-lacZ allele was stably integrated into the genome of FY1022 at the HIS4 locus by one- or two-step gene replacement technique (ROSE et al. 1990 ). The structure of the reporter gene and integration at the HIS4 locus were confirmed by Southern blot analysis (SOUTHERN 1975 ). The presence of each promoter mutation was confirmed by restriction digest of PCR products generated from each integrated his4-912-lacZ allele.

Plasmids:
The plasmid pLG40 contains a mutant allele of his4-912, his4-912::URA3-lacZ, which was integrated at the HIS4 locus to facilitate the selection of one- and two-step integrants of his4-912-lacZ mutant alleles. pLG40 was derived from pKA40 (ARNDT et al. 1994 ) in which the promoter region of his4-912 (sequences -403 to +156 relative to the HIS4 transcription start site) is replaced by the yeast URA3 gene. To create pLG40, pKA40 was linearized with NheI and filled with Klenow enzyme. A 3-kb BamHI fragment containing the E. coli lacZ gene was isolated from pMC1871 (generously provided by Dr. Malcolm Casadaban), filled with Klenow enzyme, and subcloned into the filled NheI site of pKA40.

Unless noted, each of the 912 promoter mutations was constructed in pLG39, which contains a fragment of the his4-912-lacZ fusion sufficient for site-directed mutagenesis and integration at the HIS4 locus via his4-912::URA3-lacZ. To create pLG39, pRS306 (SIKORSKI and HIETER 1989 ) was linearized with BamHI, filled with Klenow enzyme, and digested with SacI. pLG39 was constructed by subcloning a SacI-HpaI fragment from pFW82 (WINSTON et al. 1987 ), which contains a portion of the his4-912-lacZ fusion, into this pRS306 vector. To provide an additional marker, TRP1, for screening two-step integration candidates, the plasmid pAD1 was constructed by subcloning the 1.5-kb SacI-ClaI fragment from pLG39 into pRS304 (SIKORSKI and HIETER 1989 ).

The plasmids pAD11 through pAD24 contain the his4-912-lacZ alleles 11–24 in pAD1, respectively. pAD25 contains the his4-912-lacZ allele 25 in pLG39. The plasmids pAD28, pAD29, and pAD30 contain the his4-912-lacZ double UAS, double TATA, and double 13/18 alleles in pAD1, respectively.

The plasmid pBM947 is a URA3 derivative of pHR307a (LIU et al. 1993 ) that lacks the 0.8-kb NcoI fragment containing TRP1. Briefly, this plasmid contains HIS3 under the transcriptional control of a GAL1 promoter from which the UAS has been deleted and replaced by a multiple cloning site. The plasmid pAD9 was created by subcloning annealed oligonucleotides that contained the UAS sequence, sites 12–14 (Table 2), flanked by BamHI-compatible ends into the pBM947 BamHI site. pAD9 contains three tandem insertions of the UAS sequence, with the first two copies inserted in the correct orientation and the third inserted in the reverse orientation, with respect to the orientation of the promoter.

The plasmid B238 is derived from plasmid p164 (HINNEBUSCH 1985 ) and contains a constitutively expressed GCN4 allele in which the four upstream regulatory ORFs have been abolished by point mutations (MUELLER and HINNEBUSCH 1986 ).

ß-Galactosidase assays:
Cells were grown to 1–2 x 10⁷ cells/ml in SD media supplemented with the appropriate amino acids. Crude extracts were prepared and assayed as described previously (ROSE et al. 1990 ). ß-Galactosidase units, normalized to total protein concentration, are calculated as described in ROSE et al. 1990 . Protein concentrations were determined by Bradford protein assay (Bio-Rad, Richmond, CA). Because of the large number of promoter mutations examined in this study, ß-galactosidase assays were performed on groups of strains as indicated in figure and table legends. Each group consisted of two to six strains per genotype and was assayed three to six times. The average ß-galactosidase units for each genotype are presented plus and minus the standard error, which was <10% in most cases. For reasons not understood, the absolute ß-galactosidase values sometimes varied between experiments, although the relative differences between mutants were highly reproducible. For this reason, two controls were included in every experiment. First, a negative control strain containing the disrupted his4-912::URA3-lacZ allele produced 0–3 units of ß-galactosidase activity (A. M. DUDLEY and F. WINSTON, unpublished data). Also, the same set of wild-type strains was included in each group. Thus, the differences between the values of these wild-type strains represent the experiment-to-experiment variability of the absolute ß-galactosidase values. All relevant comparisons were performed within the same experiments.

RNA isolation and primer extension analysis:
Cells were grown to 1–2 x 10⁷ cells/ml in SD media supplemented with the appropriate amino acids. Total RNA was isolated by the hot-phenol method (AUSUBEL et al. 1988 ). Two oligonucleotides, lacZ (5'GGGTAACGCCAGGGTTTTCCCAGTCACGACGTTGTA 3') and U6.48-72 (5'GCAGGGGAACTGCTGATCATCTCT 3'), were 5' end labeled with [-³²P]ATP using T4 polynucleotide kinase (New England Biolabs, Beverly, MA) as described previously (AUSUBEL et al. 1988 ). Primer extension reactions were performed on 20 µg of total RNA combined with 1.2 ng labeled lacZ oligo, 1 ng labeled U6 oligo, and 10 ng unlabeled U6 oligo as described previously (PRELICH and WINSTON 1993 ). The lacZ oligo is used to detect Ty-lacZ mRNA and produces a 212-nucleotide product that corresponds to the Ty transcription initiation site (ELDER et al. 1983 ). Extension of U6.48-72 produces a 72-nucleotide U6-specific product (BROW and GUTHRIE 1988 ) that serves as a normalization control.

Electrophoretic mobility shift assays:
Yeast nuclear extracts were prepared from FY114 as described previously (HULL et al. 1995 ). The following oligonucleotides were annealed to generate probe and competitor DNA: short wild type (AD14 5' GATCCGTCATCTAAATTAGTG 3'; AD15 5' GATCCACTAATTTAGATGACG 3'), long wild type (AD6 5'GATCCGTCATCTAAATTAGTGGAAGCTGAAACGCAAGGG 3'; AD75' GATCCCTTGCGTTTCAGCTTCCACTAATTTAGATGACG3'), mutant 12 (AD10 5'GATCCCTCGGTACCAATAGTGGAAGCTGAAACGCAAGGG 3'; AD11 5'GATCCCTTGCGTTTCAGCTTCCACTATTGGTACCGAGG 3'), and mutant 13 (AD12 5'GATCCGTCATCTAAATGAGAATTCGTTGAAACGCAAGGG 3'; AD13 5'GATCCCTTGCGTTTCAACGAATTCTCATTTAGATGACG 3'). Pairs of oligonucleotides were designed to recreate a double-stranded DNA molecule containing the appropriate wild-type or mutant UAS sequence flanked by BamHI-compatible ends when annealed. Underlined regions indicate the presence of the appropriate mutations. The double-stranded probe was labeled by filling the BamHI-compatible ends of the annealed oligonucleotides with Klenow enzyme in the presence of [-³²P]dGTP. Cold competitor substrates were generated in the same manner, except in the presence of cold nucleotides.

Mobility shift assays were carried out in the following DNA-binding buffer: 20 mM Hepes, 10 mM NaCl, 1 mM MgCl₂, 10% glycerol, and 0.1% Nonidet P-40. A total of 20 µl of binding reactions contained 1x binding buffer, 5 mM dithiothreitol, 0.5 µg poly[dG-dC], 100 µg bovine serum albumin, ~1 ng ³²P-labeled probe DNA, 10 µg yeast nuclear extract, and unlabeled competitor DNA, where indicated. Reaction mixtures containing all components, except the ³²P-labeled probe and unlabeled competitor DNA, were assembled on ice and incubated at room temperature for 10 min. Labeled probe DNA and unlabeled competitor DNA were added, and the reaction proceeded at room temperature for an additional 20 min. Reactions were loaded onto a 2.5% glycerol/6% polyacrylamide gel (60:1 cross-linking) containing 1x Tris-glycine (50 mM Tris, 380 mM glycine, 2 mM EDTA, pH 8.5) and electrophoresed against 1x Tris-glycine buffer at 100 V. Following electrophoresis, gels were vacuum dried onto Whatman filter paper and exposed to Kodak XAR X-OMAT autoradiographic film.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutant analysis defines regions of the necessary for full promoter activity and accurate initiation:
To study promoter elements in sequences, we chose to study the of a widely used and well-characterized insertion mutation, his4-912. For our studies, the sequence of his4-912 was fused in frame to the E. coli lacZ gene (MATERIALS AND METHODS; WINSTON et al. 1987 ). Because lacZ is fused to the , not to the HIS4 TATA or transcription initiation site, his4-912-lacZ expression is a result of promoter function. This fusion produces moderately high levels of lacZ mRNA in an SPT-dependent manner (WINSTON et al. 1987 ) and thus serves as an accurate reporter of his4-912 promoter activity. To identify sequences important for 912 promoter function, we constructed a series of small, clustered base pair substitution mutations (Figure 1) in a region of 912 previously shown to be important in other elements (LIAO et al. 1987 ; FULTON et al. 1988 ).

The effect of each 912 mutation on promoter function was measured by ß-galactosidase assay (Figure 2). This analysis revealed that mutations in two regions of 912 decreased ß-galactosidase activity. The first region contains a UAS element, defined by mutations 12, 13, and (to a lesser extent) 14. Mutations 12 and 13 both decreased expression to ~35% of wild type, while the 12/13 double UAS mutant decreased expression to 25% of wild type. While this work was in progress, an identical UAS sequence of a different was shown to bind the transcription factor Gcr1 (TURKEL et al. 1997 ). The Gcr1-binding site corresponds to site 13 in our studies (Figure 1). The second region of decreased activity contains two consensus TATA elements that are abolished by mutations 17 and 18. Interestingly, these elements do not contribute equally to 912 promoter activity. Mutation 17 in the 5' TATA only decreased promoter activity to 87% of the wild-type level, while mutation 18 in the 3' TATA decreased promoter activity to 24% of the wild-type level. The double TATA mutant decreased expression to 14% of the wild-type level. Thus, as was shown for other Ty elements (LIAO et al. 1987 ; CONEY and ROEDER 1988 ), the UAS and TATA regions of 912 are both necessary for full promoter activity. Our results suggest that the 912 UAS may contain two elements, defined by sites 12 and 13, and that the 3' TATA element is required for most of the 912 TATA activity.

View larger version (48K): [in this window] [in a new window] [Download PPT slide]   Figure 2. ß-Galactosidase activity of the promoter mutants. The ß-galactosidase activity of mutations 11–25, double UAS, and double TATA are graphed ± standard error. The left y-axis presents the scale in ß-galactosidase units and the right y-axis presents the scale as percent activity relative to the wild-type his4-912-lacZ promoter (black bar).

Mutations in two other regions of 912 caused weak effects on expression. One region, defined by mutations 15 and 16, may contain a weak negative site. Mutations 15 and 16 caused modest but reproducible increases in ß-galactosidase activity (Figure 2). This region contains an in vitro binding site for the Mot3 protein (MADISON et al. 1998 ), and genetic evidence suggests that Mot3 weakly represses the 912 promoter via this site (MADISON et al. 1998 ). A second region, defined by mutations 22 and 23, may contain a weak positive site.

To determine whether ß-galactosidase values correlated with his4-912-lacZ mRNA levels and to examine the effects of the promoter mutations on the position of transcription initiation, we analyzed his4-912-lacZ mRNA in each 912 promoter mutant by primer extension analysis (Figure 3). Results from the measurement of mRNA levels agreed with the ß-galactosidase assays, showing the same relative effects across the promoter region. Mutations within the UAS, 12 and 13, decreased lacZ mRNA expression to 50% of wild type, while the 12/13 double mutant decreased expression to 30% of wild type. Mutations 15 and 16, which contain a Mot3-binding site, produced a 150–200% increase in lacZ mRNA levels. Mutations that abolish the strong TATA element, site 18, decreased lacZ mRNA levels to 40% of wild type, while the double TATA mutation decreased expression to 20% of wild type. Thus, the relative effects of the 912 promoter mutations detected by ß-galactosidase assay correlate well with lacZ mRNA levels, although the degree of the effects is slightly less.

View larger version (45K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Primer extension analysis of the promoter mutants. Primer extension reactions were performed on RNA isolated from strains containing his4-912-lacZ promoter mutations. A lacZ primer detects his4-912-lacZ transcripts, and the U6 primer detects the U6 snRNA as a normalization control. A his4-912-lacZ sequencing ladder, generated using the lacZ oligo, was run in parallel with the primer extension products to map transcription start sites. The positions of the correct Ty mRNA ( + 1), the three novel transcripts seen in mutants 24 and 25 (start 2–4), and the U6 snRNA are indicated by arrows. This figure is a composite of different exposures of the same gel and was produced with Adobe Photoshop and a Fujix Pictography 3000 printer.

Two mutations, 24 and 25, altered transcriptional start site selection (Figure 3). Mutation 25 spans the nucleotide at which transcription initiates (ELDER et al. 1983 ) and changes the initiation base from a G to an A. Mutation 24 spans a region 5' of the initiation site (Table 2). Both mutations cause initiation to occur at three new sites (at or near positions +16, +25, and +35) in addition to the normal initiation site (+1). The correct transcription initiation site and the three downstream sites of initiation are utilized with relatively equal frequency in both promoter mutants, although mutation 25 shows a greater effect. Mutations in this region do not significantly decrease ß-galactosidase activity (Figure 2), presumably because all four mRNA species initiate upstream of the Ty translation initiation codon and encode functional ß-galactosidase. These two mutations define a 20-bp region surrounding the transcription initiation site that is necessary for accurate start-site selection.

Evidence for the activity of a second activator binding to the 912 UAS:
Our analysis of the 912 UAS indicated that both sites 12 and 13 contain UAS activity. Recent work has demonstrated that Gcr1 binds to the UAS over the region defined by site 13 (TURKEL et al. 1997 ). We have confirmed Gcr1-binding to site 13 using a recombinant Gcr1-binding domain peptide (generous gift of Dr. Henry Baker) in electrophoretic mobility shift assays (A. M. DUDLEY and F. WINSTON, unpublished results).

Since mutations in site 12 affect UAS activity but this site is distinct from the Gcr1-binding site, we also assayed binding to site 12. As shown in Figure 4, a DNA-binding activity in yeast nuclear extracts was able to gel shift a short DNA probe that contains site 12 and only 3 bp of site 13 (Figure 4B, lane 2). The binding was competed by unlabeled, double-stranded oligonucleotides containing site 12 sequence (Figure 4B, lanes 3–5 and 9–14) but not by oligonucleotides in which site 12 was mutated (Figure 4B, lanes 6–8), demonstrating that the DNA-binding activity is specific for site 12. Two pieces of evidence strongly suggest that this mobility-shifted complex is not a result of binding by Gcr1. First, the short, wild-type probe used in this assay does not contain the consensus Gcr1-binding site, and a recombinant Gcr1-binding domain peptide (generous gift of Dr. Henry Baker) is unable to bind this probe (A. M. DUDLEY and F. WINSTON, unpublished results). Second, the gel shift was completed by a double-stranded, unlabeled oligonucleotide containing a site 13 mutation, which destroys the Gcr1-binding site (Figure 4B, lanes 9–11). Thus, genetic and biochemical evidence support the hypothesis that a factor in addition to Gcr1 is able to bind and activate the UAS.

View larger version (106K): [in this window] [in a new window] [Download PPT slide]   Figure 4. A factor in yeast nuclear extracts is able to bind site 12. (A) Schematic diagram of probes and competitors used for electrophoretic mobility shift analysis. The short, wild-type probe contains the entire sequence of site 12 and 3 bp of site 13. The long, wild-type competitor contains the entire sequence of sites 12, 13, and 14. Mutant 12 and mutant 13 competitors are identical to the long, wild-type competitor except that they contain mutant 12 and 13 base pair substitutions, respectively. The Gcr1-consensus-binding site is underlined. All probes and competitors are double-stranded and were created by annealing single-stranded oligonucleotides (MATERIALS AND METHODS). (B) Electrophoretic mobility shift analysis was performed using 10 ng of yeast nuclear extract and 0.8 ng of the short wild-type probe. All lanes contain the labeled probe. Lanes 2–14 contain extract; lanes 3–5 contain 10-, 50-, and 100-fold excess cold wild-type long competitor; lanes 6–8 contain 10-, 50-, and 100-fold excess cold 12 mutant competitor; lanes 9–11 contain 10-, 50-, and 100-fold excess cold 13 mutant competitor; lanes 12–14 contain 10-, 50-, and 100-fold excess cold wild-type short competitor. The positions of free probe and the specific UAS complex are indicated by arrows. This figure was produced with Adobe Photoshop and a Fujix Pictography 3000 printer.

Upstream activators at both the UAS_HIS4 and the UAS contribute to 912 promoter activity:
Previous results suggested that Gcn4, acting at the adjacent UAS_HIS4, exerts significant control over transcription from the initiation site at his4-912 (SILVERMAN and FINK 1984 ). To confirm this aspect of his4-912 regulation and to test whether other elements of the UAS_HIS4 controlled his4-912-lacZ, we measured the effects of bas1, pho2, and gcn4 mutations on its expression. Expression from the his4-912-lacZ fusion decreased to 23% of wild type in a gcn4 mutant and to ~50% of wild type in both bas1 and pho2 mutants (Table 3). These results are similar to those obtained previously for a HIS4-lacZ fusion tested under similar growth conditions (ARNDT et al. 1987 ). Thus, factors bound at the UAS_HIS4 activate the promoter in a manner similar to their regulation of HIS4.

View this table: [in this window] [in a new window]   Table 3. Factors at the UASHIS4 and the UAS contribute to his4-912-lacZ activity

To compare the relative contributions of transcriptional activators at the UAS_HIS4 and the UAS, we measured his4-912-lacZ expression in double mutants that affect both UAS elements (Table 3). In combination with either a bas1 or pho2 mutation, the 912 site 12 or 13 mutations caused reductions in ß-galactosidase activity equal to those predicted for the double mutant combinations, ~18% of the wild-type activity. Similarly, a gcn4 site 12 double mutant exhibited close to the expected reduction in promoter activity. In contrast, the combination of gcn4 with a mutation in the Gcr1-binding site (site 13) produced a greater-than-expected decrease in expression, to only 2% of the wild-type activity. The distinct behaviors of site 12 and site 13 mutations in a gcn4 mutant support the hypothesis that the two elements are bound by different activators. Taken together, these results suggest that, among the various activators known to bind either the UAS_HIS4 or the UAS, significant 912 promoter function requires either Gcr1 or Gcn4. Only in the presence of at least one of these factors can other activators at either the UAS_HIS4 or the UAS contribute additional activity.

To test if Gcn4 activates via the UAS as well as the UAS_HIS4, we examined the effect of a gcn4 on the activity of the UAS in a heterologous context. Multiple copies of the UAS sequence were cloned 5' to a heterologous reporter gene as described in MATERIALS AND METHODS. As expected, the UAS was able to activate expression of the HIS3 reporter in a wild-type strain; however, this activation was abolished by a gcr1 mutation (Figure 5), consistent with previous studies (TURKEL et al. 1997 ). Moreover, the UAS was still able to activate in a gcn4 strain (Figure 5). The Gcn4 independence of the UAS strongly suggests that the effects observed in double mutant combinations between gcn4 and UAS mutants (Table 3) are not solely caused by the loss of Gcn4 activation, but rather by the combined loss of multiple activators.

View larger version (57K): [in this window] [in a new window] [Download PPT slide]   Figure 5. The UAS is GCR1 dependent and GCN4 independent. Wild-type, gcr1, and gcn4 strains were transformed with plasmids containing a HIS3 reporter gene, either without a UAS (no UAS) or with three copies of the UAS sequence. Approximately 2 x 105 cells were spotted onto solid media containing histidine (SC-Ura) and lacking histidine (SC-Ura-His) and the plates were incubated at 30°. The media contained 2% glycerol/2% ethanol as carbon sources to allow the growth of the gcr1 strains, which grow poorly on glucose.

To test the relative contribution of Gcn4 activation under conditions that mimic amino acid starvation, Gcn4 was constitutively expressed. To do this, we transformed cells containing either the wild-type his4-912-lacZ promoter or the double UAS (12/13) mutant with a plasmid that carries a GCN4 constitutive mutation (Table 4). Because this allele of GCN4 is no longer under its normal translational repression, it is highly expressed regardless of amino acid abundance (MUELLER and HINNEBUSCH 1986 ). Constitutive expression of GCN4 increases his4-912-lacZ activity approximately twofold in both the presence and absence of UAS activity (Table 4). Similarly, constitutive GCN4 expression increases his4-912-lacZ activity approximately twofold in an spt3 mutant (Table 4). These results demonstrate that high levels of GCN4 expression are not sufficient to overcome the defects caused by a UAS or spt3 mutation.

View this table: [in this window] [in a new window]   Table 4. Analysis of the contribution of Spt3 and the UAS under conditions of constitutive Gcn4 expression

Interactions of mutations in the two 912 TATA regions with mutations that affect HIS4 and 912 UAS activity:
To assess the roles of the two 912 TATA elements with respect to activation by Gcn4 and Gcr1, we examined a set of double mutants that affect both UAS and TATA function. First, we examined the effect of mutations in the 912 TATA elements in combination with a gcn4 mutation (Table 5). Similar to the severe defect seen when both Gcn4 and Gcr1 activation were abolished, expression decreased to 2% of wild type in a 3' TATA (18) gcn4 double mutant. This result suggests a model in which virtually all the activation by Gcr1 occurs via the 3' TATA; that is, the gcn4 site 18 double mutant mimics loss of both Gcn4 and Gcr1. The fact that the weak defect observed for the 5' TATA mutation (17) is the same in GCN4 and gcn4 backgrounds suggests that this TATA element may be responsive to activation by Gcn4. However, the weak activity of this TATA in the presence of a functional 3' TATA element makes this analysis difficult. These results suggest that 912 requires the activity of either Gcn4 or TATA 18 to have any significant activity.

View this table: [in this window] [in a new window]   Table 5. Analysis of the contribution of Gcn4 at the UASHIS4 and Gcr1 at the UAS in TATA mutants

To test the requirement of the 3' TATA (18) in Gcn4 activation, we constructed a his4-912-lacZ derivative containing mutations in the Gcr1-binding site (mutation 13) and in the 3' TATA (mutation 18). The decrease in the site 13/18 double mutant was less severe than in the gcn4 site 18 double mutant, 9% vs. 2% of wild-type activity (Table 5). These results suggest that in the absence of Gcr1 and the 3' TATA, Gcn4 is able to weakly activate the 912 promoter, presumably through the 5' TATA. Taken together, these results suggest that Gcr1 activates via TATA 18 and Gcn4 activates via both TATA sequences.

Expression of his4-912-lacZ is dependent on specific classes of SAGA components:
Transcription of full-length Ty elements as well as many solo insertion alleles is strongly dependent on certain Spt proteins (WINSTON et al. 1987 ) that are components of the SAGA complex (GRANT et al. 1997 ). SAGA contains a histone acetyltransferase activity that is dependent upon one SAGA component, Gcn5 (BROWNELL et al. 1996 ; GRANT et al. 1997 ). To learn more about SAGA control of the promoter, we have analyzed his4-912-lacZ expression in three different classes of SAGA mutants, spt20, spt3, and gcn5 (Table 6). Recent work has shown that spt20 likely abolishes all SAGA activities, while spt3 and gcn5 each affect distinct subsets of SAGA activities (GRANT et al. 1997 ; HORIUCHI et al. 1997 ; ROBERTS and WINSTON 1997 ). As expected (WINSTON et al. 1984B ; ROBERTS and WINSTON 1996 ), spt3 and spt20 mutations caused decreased levels of his4-912-lacZ expression. However, gcn5 caused no defect in his4-912-lacZ expression. These results support the hypothesis that SAGA contains multiple functions that may be required at different promoters to modulate their expression. For example, the Spt3-dependent SAGA activity is required for his4-912-lacZ expresssion, whereas the Gcn5 histone acetyltransferase activity is not required.

View this table: [in this window] [in a new window]   Table 6. Expression of his4-912-lacZ in three classes of SAGA mutations

Several elements of the his4-912 promoter are partially Spt3 dependent:
The Spt3-dependent function of SAGA is strongly required at some but not all RNA polymerase II-dependent promoters in S. cerevisiae. However, the factors or promoter elements that confer Spt3-dependence on these promoters are unknown. Unexpectedly, our analysis of promoter elements within 912 did not identify any single element that could account for the large decrease in expression observed in an spt3 mutant. On the basis of this analysis, we reasoned that Spt3 activity might be exerted over a number of promoter elements, or even the entire promoter region. If the function of a promoter element was completely dependent upon Spt3, we would expect that in an spt3 mutant, a mutation in that promoter element would not further reduce expression. To determine whether the activities of important 912 promoter elements were dependent on Spt3, we measured the activity of a set of 912 promoter mutants in an spt3 background (Table 7). This analysis shows that, while none of the promoter elements tested were completely Spt3 dependent, several promoter elements appeared partially dependent. The strongest example of this partial dependence occurred with the Gcr1-binding site. In an spt3 background, a mutation in the Gcr1-binding site (13) in the 912 promoter caused a significantly weaker defect than in an SPT3⁺ background. Similar effects were seen for several other elements, including the strong 3' TATA box (18), Gcn4, and the unidentified protein that binds to site 12 of the 912 UAS (Table 7). We also assayed the 13/18 mutation in an spt3 background and demonstrated a strong but still partial dependence upon Spt3. Taken together, these results support a model in which Spt3 activity at the 912 promoter occurs via partial effects at a number of important promoter elements.

View this table: [in this window] [in a new window]   Table 7. Analysis of Spt3-dependence of expression of his4-912

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study we have conducted a genetic analysis of a Ty insertion mutation, his4-912, to identify sites and factors important for promoter function in vivo. Past studies and this work have shown that several activators act at the two UAS elements present in this complex promoter. However, our analysis has demonstrated that two of these activators, Gcr1 acting at the UAS, and Gcn4 acting at the UAS_HIS4, are the primary activators required for 912 expression. In addition, we found that multiple his4-912 promoter elements are partially dependent upon Spt3, a component of the SAGA complex. Thus, the Spt3-dependent SAGA activity appears to exert large effects on the promoter through a combination of partial effects on several promoter elements.

Identification of 912 sequences important for promoter function:
The region identified as a UAS in 912 is identical to the UAS identified previously in a Ty2 element in sequence and position (LIAO et al. 1987 ). Our results have provided evidence for two distinct sites within this UAS, which we have called sites 12 and 13. Our analysis of site 13 is consistent with a previous study that demonstrated binding of Gcr1 to this region of the UAS (TURKEL et al. 1997 ). Although we have not yet identified the factor that binds to site 12, the site 12 sequence hints that this factor might be a homeodomain protein. The site 12 sequence TAAATTA contains AATT and TAAT sequences often seen in sites bound by homeodomain proteins (WOLBERGER 1996 ). We have tested null mutations in several genes encoding homeodomain proteins for an effect on his4-912-lacZ expression. There was no decrease in his4-912-lacZ expression for any of the mutations tested in the genes MATa1, MAT2, PHO2, CUP9, and YGL096 (A. M. DUDLEY and F. WINSTON, unpublished results). In addition, gel shift experiments have suggested that the factor binding to site 12 does not bind cooperatively with Gcr1 (A. M. DUDLEY and F. WINSTON, unpublished data). More analysis is required to understand fully the role of site 12 and the factor that binds to it.

Mutations in the two 912 consensus TATA elements also decreased expression from his4-912-lacZ. Although many studies have demonstrated the importance of this TATA region for activity (LIAO et al. 1987 ; CONEY and ROEDER 1988 ; FULTON et al. 1988 ; HIRSCHMAN et al. 1988 ; ARNDT et al. 1994 ), the relative contribution of each TATA element was unknown prior to this study. Surprisingly, the activity of these identical TATA elements is not equivalent; mutation of the 5' TATA produces a very modest decrease in expression while mutation of the 3' TATA causes a decrease of severalfold. Two previous studies have examined the utilization of identical TATA elements introduced into the S. cerevisiae genes HIS4 (NAGAWA and FINK 1985 ) and CYC1 (LI and SHERMAN 1991 ). In both cases, only the 5' TATA element was functional. In contrast, the 3' TATA is the most active in the 912 promoter. Several possible mechanisms could explain this differential TATA selection. The function of specific activators could increase the activity of a particular TATA element. Certain activators may prefer specific TATA elements based on position or some other mechanism. Alternatively, the flanking sequence surrounding a consensus TATA element may affect its utilization in vivo (for example, see HARBURY and STRUHL 1989 ). There is evidence that sequences flanking the TATA in 912 are important for promoter function (HIRSCHMAN et al. 1988 ). Although the mechanism for 912 TATA preference is still unknown, our results suggest that at some promoters the mechanism may be more complex than merely choosing the best consensus TATA located closest to the UAS.

Our analysis identified a region of the required for accurate transcription initiation. Mutations in this initiator element alter the position of initiation without affecting the overall level of expression. The role of initiator elements is not well understood in S. cerevisiae. Surveys of the transcription start site of a number of genes and analysis of the effects of distance from the TATA element have led to a model in which transcription usually initiates from a site resembling a loose consensus sequence around ~30 to 120 nucleotides from the TATA element (CHEN and STRUHL 1985 ; HAHN et al. 1985 ; NAGAWA and FINK 1985 ; HEALY et al. 1987 ; RUDOLPH and HINNEN 1987 ; MAICAS and FRIESEN 1990 ; LI and SHERMAN 1991 ). Start site selection appears to be intimately linked to the general transcription apparatus since mutations in the genes encoding TFIIB (PINTO et al. 1992 ), Rbp1 (BERROTERAN et al. 1994 ), and Rpb9 (HULL et al. 1995 ) alter the transcriptional initiation patterns of several genes. We speculate that Ty1 may have evolved a strong initiator element because a single initiation site is important for some aspect of its life cycle, such as translation, mRNA packaging, or reverse transcription.

Transcription from his4-912 is controlled by promoter elements of both the and the UAS_HIS4:
Ty promoter function in his4-912 is determined by the activity of a variety of UAS and TATA promoter elements and is therefore a good model for the complex promoters found in larger eukaroytes. Our results have led to a model in which the 912 promoter is primarily activated by Gcr1 and Gcn4 (Figure 6). In this model, Gcn4 at the UAS_HIS4, Gcr1 at the UAS, and TBP at the 3' TATA are all crucial for promoter function. Other factors at the UAS_HIS4 (Bas1 and Pho2), the UAS (an unidentified factor at site 12), and the 5' TATA (TBP) increase expression from 912 as long as Gcr1 or Gcn4 is present. Thus, the ability of the promoter to respond to the activity of transcription factors positioned near it as a result of transposition allows activation of the promoter via -specific and heterologous factors.

View larger version (21K): [in this window] [in a new window] [Download PPT slide]   Figure 6. A model for his4-912 promoter function. The drawing shows the relative positions of several factors known to bind to the his4-912 promoter region: Bas1, Pho2, and Gcn4 binding at the UASHIS4; Gcr1 and a currently unidentified factor binding at the UAS; and TBP binding at the two consensus TATA elements. The transcription start site is indicated by the arrow. In this model, Gcr1 activates via the 3' TATA element and Gcn4 activates via both TATA sequences. Many of these important promoter elements are partially dependent on the Spt3-dependent activity of the SAGA complex, resulting in a large Spt3-dependent effect on the promoter as the sum of partial effects at a number of promoter elements.

How does SAGA function at a promoter?
Our results have demonstrated that the Spt3-dependent activity of SAGA is required for 912 promoter function and that the Gcn5-dependent histone acetyltransferase activity of SAGA is not required. These results are consistent with previously observed Spt^- phenotypes of spt3 and spt20 mutants and the Spt⁺ phenotype of a gcn5 mutant (WINSTON et al. 1984B ; ROBERTS and WINSTON 1996 , ROBERTS and WINSTON 1997 ). The results are also consistent with previous models that suggest that functions encoded by different components of SAGA may be required at different promoters (HORIUCHI et al. 1997 ; ROBERTS and WINSTON 1997 ).

Our analysis of the effect of an spt3 mutation in combination with mutations in important promoter elements suggests that activation by both Gcn4 at the UAS_HIS4 and Gcr1 at the UAS is partially Spt3 dependent. Our results provide the first evidence that Spt3 may exert large effects on the transcriptional activity of a promoter by a combination of smaller effects at multiple promoter elements. The relatively weak SPT3 dependence of the 912 TATA elements was surprising as previous results suggested that Spt3 interacts with TBP (EISENMANN et al. 1992 ). However, Spt3 may affect the activity of specific TATA elements either by direct interaction with TBP or by controlling communication between TBP and activators.

Our results are consistent with two obvious models for SAGA function at 912: a transcription-factor-binding model and a transcription-factor-activation model. In the factor-binding model, activators such as Gcr1 and Gcn4 would partially require the Spt3-dependent SAGA activity to bind their cognate sites. If this model is correct, these activators must be able to bind their sites, albeit to a lesser degree, in an spt3, because the activities of Gcr1 and Gcn4 are not completely Spt3 dependent. One parameter that could affect the binding of multiple proteins over a region of this length is the chromatin structure (for review, see KINGSTON et al. 1996 ; STEGER and WORKMAN 1996 ). However, it is unlikely that the nucleosome acetylation activity of SAGA facilitates the binding of these factors because a gcn5, which completely abolishes the histone acetylase activity of SAGA (26), has no effect on his4-912 transcription. In the factor-activity model, activators such as Gcr1 and Gcn4 would not require SAGA to bind their sites, but would at least partially require the Spt3-dependent SAGA activity for a subsequent step in transcriptional activation. This model is consistent with previous models proposed for the function of several classes of SAGA components, including Spt proteins (EISENMANN et al. 1992 ), Ada proteins (BERGER et al. 1992 ), and Gcn5 (GEORGAKOPOULOS and THIREOS 1992 ). In the context of this model, our results suggest that Spt3 and other components of the SAGA complex would mediate the interaction of multiple activators with the transcription machinery. Experiments designed to distinguish between these two models will greatly increase our understanding of SAGA function, and the detailed analysis of SAGA-dependent promoters will provide important tools for such studies.

	&