Isolation and characterization of TAF25, an essential yeast gene that encodes an RNA polymerase II-specific TATA-binding protein-associated factor.

We describe the cloning and analysis of TAF25, a previously uncharacterized yeast gene that encodes a yeast TATA-binding protein-associated factor or yTAF of Mr = 25,000. The gene encoding yTAF25 is a single copy essential gene, and the protein sequence deduced from TAF25 exhibits sequence similarity to a metazoan hTAFII. The results from immunological studies confirm that yTAF25 is a subunit of a large multiprotein TATA-binding protein-yeast TATA-binding protein-associated factor complex that contains a subset of the total number of the yTAFs present in yeast cell extracts. Both genetic and biochemical analyses demonstrate that yTAF25 can interact directly with itself. Transcriptional data show that the activity of the multiprotein complex containing yTAF25 is RNA polymerase II-specific, thus indicating that TAF25 encodes a bona fide yeast RNA polymerase II TAF. Hence the protein encoded by TAF25 has been termed yTAFII25.

We describe the cloning and analysis of TAF25, a previously uncharacterized yeast gene that encodes a yeast TATA-binding protein-associated factor or yTAF of M r ‫؍‬ 25,000. The gene encoding yTAF25 is a single copy essential gene, and the protein sequence deduced from TAF25 exhibits sequence similarity to a metazoan hTAF II . The results from immunological studies confirm that yTAF25 is a subunit of a large multiprotein TATA-binding protein-yeast TATA-binding protein-associated factor complex that contains a subset of the total number of the yTAFs present in yeast cell extracts. Both genetic and biochemical analyses demonstrate that yTAF25 can interact directly with itself. Transcriptional data show that the activity of the multiprotein complex containing yTAF25 is RNA polymerase II-specific, thus indicating that TAF25 encodes a bona fide yeast RNA polymerase II TAF. Hence the protein encoded by TAF25 has been termed yTAF II 25.
The TATA-binding protein (TBP) 1 is required for transcription by all three eukaryotic nuclear DNA-dependent RNA polymerases (Refs. 1-4; reviewed in Refs. 5 and 6), where it plays an essential but as of yet incompletely understood role in transcription. In all the systems studied thus far (Drosophila, human, and yeast) TBP exists as a component of multisubunit complexes comprised of TBP and TBP-associated factors (TAFs) (1,(7)(8)(9)(10)(11)(12), (see Ref. 13 and 14 for recent reviews). The paradigm that has emerged from these studies is that the polymerase specificity of a particular TBP⅐TAF complex is determined by the collection of TAFs associated with TBP and not by the TBP molecule itself. Attempts are now being made to define the total number of TAFs, the number and composition of discrete TBP⅐TAF complexes, and the minimal TBP-TAF complements for each RNA polymerase. Through these efforts several distinct TBP⅐TAF complexes have been identified and characterized. The RNA polymerase I-specific TBP⅐TAF complex termed SL1 contains three TBP-associated proteins or TAF I s (1,15,16), whereas the RNA polymerase II-specific TBP⅐TAF complex termed TFIID contains eight to twelve TAF II s (10,(17)(18)(19)(20). Similarly, the RNA polymerase III TBP⅐TAF complex, TFIIIB, contains two TAF III s (21)(22)(23) at least in yeast. A fourth TBP⅐TAF complex, termed the SNAP complex or SNAP c , appears to be involved in directing Small nuclear RNA gene transcription by RNA polymerases II and III and consists of at least three TAFs (24). The various TBP⅐TAF complexes (SL1, TFIID, TFIIIB, and SNAP c ) subserve key regulatory functions in the transcription process at the level of transcription initiation. In addition to these initiation-competent TBP⅐TAF complexes, there are additional TAFs that appear to negatively modulate transcription, perhaps by sequestering TBP thereby affecting its ability to form initiation competent TBP⅐TAF complexes (25)(26)(27)(28).
In recent years our studies have been directed toward defining TBP⅐TAF complexes in the yeast Saccharomyces cerevisiae. We previously reported the use of anti-TBP antibodies for immunoaffinity chromatography to purify TBP⅐TAF complexes from yeast whole cell extracts (WCE) (29). The TBP⅐TAF complexes in this TAF fraction reproducibly contains at least nine polypeptides ranging in size from M r ϭ 170,000 to M r ϭ 25,000. In the process of characterizing the proteins in this TAF fraction, we have determined that these TAFs are derived from at least three distinct TBP⅐TAF complexes. The transcription factor TFIIIB, a TBP⅐TAF complex specific to the function of RNA polymerase III, was the first complex we identified in this fraction (29). Yeast TFIIIB has been shown to contain TBP and two yTAF III s, an M r ϭ 70,000 subunit termed BЈ (or, alternatively, Brf1p, Tds4p, or Pcf4p) (30 -32) and an M r ϭ 90,000 subunit termed BЉ (22,33). Both BЈ (Brf1p) and BЉ are present in our yTAF preparation as demonstrated by the fact that we can reconstitute specific tRNA gene transcription in vitro using purified RNA polymerase III, TFIIIC, and the TAF fraction as a source of TFIIIB (29). The second TBP⅐TAF complex we characterized consists minimally of yTAF170 and TBP (25). yTAF170 is encoded by MOT1, which was originally identified in a genetic screen performed by Thorner and colleagues (34) as a gene encoding a repressor of transcription of numerous genes. Mot1p/yTAF170 has also been independently identified and characterized by Auble and Hahn as ADI, an ATP-dependent inhibitor of TBP binding to DNA (26). This action of Mot1p appears to be polymerase II-specific (35). Finally, we (11) and Reese et al. (12) recently described a multisubunit TFIID-like complex in yeast. Our work (11) demonstrated that the yTAFs of M r ϭ 150, 130, 90, 60, 40, 30, and 25 kDa, which are present in our TAF fraction, appear to exist in a single TBP⅐TAF complex with many of the biochemical and genetic hallmarks of metazoan TFIID.
In this report we describe the cloning and characterization of TAF25 a previously uncharacterized yeast gene that encodes a yeast TAF protein that exhibits an apparent M r of 25,000. As was observed for other yeast TAFs (11,12), we found that the deduced amino acid sequence of yTAF25 exhibits significant similarity to the deduced amino acid sequence of a metazoan TAF, in this case hTAF II 30 (17). We show here that TAF25 is a single copy essential gene by gene disruption experiments, and through both genetic and biochemical analyses we demonstrate that yTAF25 can interact with itself. Coimmunoprecipitation and transcriptional studies confirm that yTAF25 is in fact a bona fide TBP-associated factor, that it is a subunit of a multiprotein TBP⅐TAF complex, and that its activity is RNA polymerase II-specific. By these criteria the TAF25-encoded protein has been termed yTAF II 25. Documenting the RNA polymerase II specificity of the yTAF II 25-containing TBP⅐TAF complex is further evidence that yTAF II 25 is part of a yeast TFIID complex with a minimal TAF II subunit composition of yTAF II 150, yTAF II 130, yTAF II 90, yTAF II 60, yTAF II 40, yTAF II 30, and yTAF II 25.

yTAF25 Purification and Protein Sequencing
Yeast TAF proteins were purified by immunochromatography on columns containing anti-TBP IgG covalently attached to protein A-Sepharose and separated by preparative SDS-PAGE as described (25,29). Purified yTAFs were transferred to a nitrocellulose membrane, and the band corresponding to yTAF25 was excised and cleaved with trypsin. The resulting tryptic peptides were purified by HPLC and sequenced. The sequences of the peptides were: peptide 1, (K/R)QLLQGQ-QQPGVQQIXQQQQQ; peptide 2, (K/R)VVLTVNDLSSAVAE(Y); peptide 3, (K/R)FVSDIAKDAYEY; peptide 4, (K/R)LLALATQK; and peptide 5, (K/R)EAVVDD. BLAST searches (41) were used to scan data bases for protein sequence homologies, and Geneworks software (Intelligenetics Inc.) was used to generate and format sequence alignments.

TAF25 Gene Cloning
The gene encoding yTAF25 was cloned via a PCR strategy based upon the amino acid sequence of a yTAF25 tryptic peptide. Two degenerate primers with HindIII (upstream) and XbaI (downstream) restriction endonuclease recognition sites (upstream, GAGAAAGCTTGT(G/A/ T/C)GT(G/A/T/C)CT(G/A/T/C)AC(G/A/T/C)GT; HindIII site underlined) and (downstream, GAGATCTAGA(A/G)TA(T/C)TC(G/A/T/C)GC(G/A/T/ C)GC(G/A/T/C)GC; XbaI site underlined) were generated from yTAF25 peptide 2 (see above). In a first round of PCR the upstream oligo was used together with an oligo (TAATACGACTCACTATAG) complementary to the T7 promoter site just downstream of the insertion site of the cDNA into vector pRS316-Gal1-cDNA (42) using total library DNA as the template. The product of this first PCR reaction was used as the template along with both degenerate oligonucleotides as primers in a second PCR reaction. The products of this PCR reaction were subse-quently digested with HindIII and XbaI, and the correct length product (67 base pairs) was cloned into HindIII/XbaI-digested pBSIIKS ϩ (Stratagene) for DNA sequence determination. From the determined nondegenerate sequence, an oligonucleotide was designed (GTGAACGATCT-CAGTAGCGC) that was then used to screen a yeast genomic DNA library (ATCC 77164) to obtain full-length plasmid clones of the gene encoding yTAF25. The inserts of these plasmids were sequenced using standard dideoxy sequencing methods.

Construction of a Yeast Strain Harboring a Chromosomal Null Mutation of TAF25
Diploid strain SEY6210.5 was used as the parental strain for the disruption of TAF25. PCR oligos were designed to contain 50 nucleotides of TAF25 sequence both upstream and downstream of the TAF25 ORF. Each PCR oligo also contained 20 nucleotides of sequence complementary to TRP1 sequences (upstream, CGGTCGAAAAAA-GAAAAGGAG, and downstream, GCAAGTGCACAAACAATACTT) at their 3Ј ends. These oligos were used in a large scale PCR reaction templated with YDp-W DNA (43), a plasmid that contains TRP1 sequences. The resulting amplification product thus contained 50 base pairs of TAF25 sequences flanking an intact 818-base pair TRP1 gene. This DNA fragment (taf25⌬1::TRP1) was gel-purified and used to generate a TAF25 null allele by transforming strain SEY6210.5 to tryptophan prototrophy; the resulting Trp ϩ strain was called yEK2. Proper integration of the taf25⌬1::TRP1 disrupting fragment at the TAF25 locus was verified by genomic Southern blotting and PCR. A yeast expression plasmid encoding yTAF25 was constructed to serve as a covering plasmid for the chromosomal taf25⌬1::TRP1 null allele. This covering plasmid was constructed by cloning a BstYI/EcoRV genomic fragment containing TAF25 into BstYI/EcoRV-digested plasmid pRS416 (Stratagene) that had been cut with BamHI and EcoRV. The resulting plasmid, which also carries URA3, was termed pRS416-TAF25. This plasmid was then used to transform yEK2 to uracil prototrophy to produce yeast strain yEK2/pRS416-TAF25. yEK2/ pRS416-TAF25 was then sporulated, and the resulting tetrads were dissected. These spores were germinated and subjected to phenotypic testing, and one of the resulting spore clones derived from this dissection was termed yEK16, which has the relevant genotype taf 25⌬1::TRP1/pRS416-TAF25.

Bacterial Expression and Purification of yTAF25
The TAF25 ORF was amplified using Pfu DNA polymerase (Stratagene) in PCR reactions (oligo sequences, upstream, GAGAGACTC-GAGGATTTTGAGGAAGATTACGAT, XhoI site underlined, and downstream, GAGAGAGTCGACCTAACGATAAAAGTCTGGGC, SalI site underlined) templated with cloned genomic TAF25 sequences. The PCR product was digested with XhoI and SalI, gel purified, and then cloned into each of the following three XhoI-digested vectors: pET-His, pETHisK (47), and pGEX-KG (48) to allow for expression in bacteria. Fusion proteins were purified under native conditions using either nickel-nitrilotriacetic acid (Qiagen) resin for His 6 -TAF25 or glutathione-agarose (Sigma) for GST-TAF25.

TAF25 Protein-Protein Interaction Studies
The plasmids pBTM116 (herein referred to as pLexA) and pVP16 (kindly provided by Stanley M. Hollenberg) were utilized to create the desired "bait" (pLexAϪ) and "prey" (pVP16Ϫ) plasmids for two-hybrid protein-protein interaction analyses (38). Oligonucleotides with flanking BamHI restriction endonuclease recognition sites (underlined below) were designed to allow PCR amplification of either the yTAF25 ORF (oligo sequences, upstream, GAGAGGATCCATTTTGAGGAAGA-TTACGATGCGGA, and downstream, GCGCGGATCCTTATCACTAA-CGATAAAAGTCTGGGCG) or a portion of the human estrogen receptor (hER) encoding hER amino acids 179 -341 (upstream, GAGAGGATC-CTCGGGAAAGGAGACTCGCTA, and downstream, GAGAGGATC-CGAAGCTTCACTGAAGGGGT (49). Pfu DNA polymerase (Stratagene) was used in all PCR reactions. PCR products were digested with BamHI and ligated into BamHI-digested pLexA or pVP16 vectors. The structures of all the clones were verified by DNA sequencing. Yeast strain L40 was used as the reporter strain for testing two-hybrid interactions. L40 cells were transformed with various combinations of bait and prey plasmids (see below) using plasmid-borne prototrophic markers to select transformants. At least 10 individual transformants from each strain containing a bait and prey plasmid pair were streaked together as a patch on a plate with the appropriate medium selection. After growth overnight at 30°C each patch of cells was used to make a glycerol stock, which was utilized to inoculate medium for two-hybrid analyses. Synthetic complete medium (36) minus the relevant components was utilized to maintain appropriate selection during these procedures.
To perform the two-hybrid protein-protein interaction analyses, 5-ml cultures of the desired strain were inoculated from a glycerol stock into selective medium. These cultures were grown to an A 600 ϭ ϳ0.8/ml, and the cells were then used for growth plating assays (data not shown) and for determination of ␤-galactosidase expression levels as described previously (50,51).

GST Pull-down Experiments
GST-TAF25 and GST resins were prepared from bacterial lysates derived from isopropyl-1-thio-␤-D-galactopyranoside-induced E. coli cells harboring plasmids expressing either GST or GST-TAF25 as described elsewhere (48). These resins were stored in column buffer (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 ⅐7H 2 0, 1.4 mM KH 2 PO 4 ; phosphate-buffered saline), 0.1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, and 10 mM DTT) containing 0.02% azide at 4°C until used. The amount of GST or GST-TAF25 protein bound to the resin was estimated by boiling a fixed amount of resin in SDS sample buffer and fractionating the sample by SDS-PAGE along with known amounts of a protein standard followed by Coomassie Blue staining. 32 P-Labeled HMK-TAF25 was prepared for protein-protein interaction studies as follows. His 6 HMK-tagged yTAF25 was purified from bacterial lysates using nickel-nitrilotriacetic acid resin and a 5-500 mM imidazole gradient for elution. His 6 -HMK-tagged yTAF25 peak fractions were identified and pooled, and His 6 -HMK-yTAF25 protein concentration was estimated via SDS-PAGE/Coomassie Blue staining as described above. Labeling and desalting of 32 P-labeled HMK-yTAF25 were then performed as described (52).
The GST binding reactions were performed in two steps. During an initial preincubation step resin containing approximately 5 g of GST or GST-TAF25 was suspended in 75 l of binding buffer (300 mM KOAc, 20 mM KHPO 4 , pH 6.8, 7 mM MgCl 2 , 10% glycerol, 0.1 mM phenylmethylsulfonyl fluoride, 2 mM Benzamidine, 1 mM DTT, 0.5 mg/ml E. coli lysate protein, and 0.1% Nonidet P-40). The beads were allowed to incubate at 4°C for 10 min. In step two, 25 l of 32 P-labeled TAF25 probe (range, 0 -500 ng) in probe buffer (25 mM Hepes, pH 7.6, 12.5 mM MgCl 2 , 20% glycerol, 100 mM KCl, 1 mg/ml bovine serum albumin, 1 mM DTT, and 0.1% Nonidet P-40) was added to the reactions. Binding was allowed to proceed by incubating these reactions on a tiltboard for 1 h at 4°C. The resin from each reaction was then recovered by centrifugation, and the beads were washed three times with 1 ml of binding buffer. After the final wash the supernatant was removed, 15 l of 2 ϫ SDS sample buffer was added to the beads, and the samples were mixed, heated to 100°C for 3 min, and then fractionated on a denaturing 12% polyacrylamide gel. The gels were then dried and exposed to x-ray film. Quantitation of binding was performed by exposing the dried gels to phosphorimaging screens.
Competition studies were performed in order to document the specificity of the TAF25-TAF25 interaction. Reactions were prepared in binding buffer containing 5 g of GST-TAF25, 80 ng of 32 P-labeled HMK-TAF25 probe and varying quantities of unlabeled His 6 -TAF25 (range, 30 -8000 ng) as a competitor. The total amount of protein added to each reaction was kept constant (8000 ng) by the addition of appropriate amounts of bovine serum albumin. Binding and quantitation was performed as detailed above.

Antibodies and Immunological Methods
Polyclonal antibodies directed against yTAF25 were raised in rabbits by Bethyl Labs, Inc. (Montgomery, TX) using purified His 6 TAF25 pro-duced in E. coli as antigen. Total immunoglobulins were purified from rabbit serum using ammonium sulfate precipitation and DEAE-cellulose chromatography (53). Anti-yTAF25 antibodies were affinity purified from IgG preparations on a column containing His 6 TAF25 coupled to cyanogen bromide-activated Sepharose 4B (Sigma). Anti-influenza virus hemaglutinin (anti-HA) monoclonal antibody, 12CA5, which reacts with the HA-epitope (YPYDVPDA), was purchased from Boehringer Mannheim.

Immunodepletion Experiments
Monoclonal anti-HA antibody 12CA5 was cross-linked to protein A-Sepharose (2 g of anti-HA mAb/l Sepharose) using demethylpimilimidate (53). This cross-linked immunoaffinity matrix was then used to deplete HA 3 -TAF25 and associated proteins from a WCE made from yeast strain yEK20, which expresses HA 3 -TAF25 as the only form of yTAF25. Prior to incubating the resin with the extract, 600-l aliquots of the resin were incubated with buffer alone (BA200 buffer: 20 mM Hepes-KOH, pH 7.6, 10% glycerol, 100 g/ml bovine serum albumin, 1 mM DTT, 200 mM KOAc, 2 mM benzamidine, and 0.1 mM phenylmethylsulfonyl fluoride) or a 100-fold molar excess of a peptide recognized by the antibody (sequence, GGYPYDVPDYAGGYPYDVPDYAGGYPYD-VPDYAGG, HA epitope underlined) in BA200 buffer for 1 h at room temperature in a total volume of 1.5 ml. Following this preincubation/ blocking step the resin was washed six times with 1 ml of BA200 buffer. Next, 250 l of yEK20 WCE (100 g of protein/l) was added to approximately 200 l of each of these treated resins, and the reactions were placed on a tiltboard at 4°C overnight to allow mAb-antigen binding. The supernatants were then harvested by centrifugation, aliquoted, and frozen at Ϫ70°C until assayed for TAF25 content and transcriptional activity.

Immunoprecipitation and Immunoblotting
Immunoprecipitation and immunoblotting were performed with yeast WCE as detailed (25) with the following modifications. Antibody and WCE were allowed to incubate overnight on a tiltboard rather than for 1 h at 4°C. Where indicated, blots were subjected to quantitation by exposing the developed blot to a Bio-Rad high intensity imaging screen, which was then analyzed using the Bio-Rad Imager and Molecular Analyst software.

In Vitro Transcription Assays
RNA Polymerase I Assay-WCE (approximately 400 g of total protein) was added to a reaction mixture mix in a total volume of 24 l. Specific DNA template (pYR12-5 (54)) was added and incubated for 5 min at room temperature. rNTPs were then added, bringing the final reaction volume to 30 l. Final conditions for the transcription reaction were as follows: 10 mM Hepes-KOH, pH 7.6, 10% glycerol, 90 mM K ϩ (as either KCl or KOAc), 10 g/ml ␣-amanitin, 5 units/reaction RNasin (Promega), 10 mM MgCl 2 , 2 mM DTT, 5 g/ml pYR12-5, and 500 M each of all four ribonucleotides. Transcription was allowed to proceed for 30 min at room temperature. The reaction was stopped by the addition of 200 l of Stop Mix which contained 0.3 M NaOAc, pH 5.5, 0.5% SDS, 10 mM EDTA, 10 g/reaction E. coli tRNA, and 1000 cpm of a 300-nucleotide end-labeled DNA fragment included as a recovery control. The reactions were extracted sequentially with two volumes of a 1:1 mixture of phenol and chloroform and two volumes of chloroform. Nucleic acids were then ethanol precipitated along with ϳ500,000 cpm (10 fmol) of a 32 P-end-labeled oligonucleotide complementary to sequences ϩ25 to ϩ50 of the rRNA primary transcript (sequence, CTC-GAGGGTCTTGAGGCTCAGAATCG). Primer extension reactions were then performed as described (55), and the amount of appropriate length (50 nucleotides) extended product produced was quantitated by Phosphorimager using the Molecular Analyst software (Bio-Rad), and the yield of extension product was normalized to the yield of the internal recovery control.
RNA Polymerase II and RNA Polymerase III Assays-Both RNA polymerase II- (56) and RNA polymerase III-specific (57) transcription assays were carried out as described previously; Gcn4p was purified as described (58) through the ammonium sulfate precipitation step. Specific transcript production was quantitated using a PhosphorImager as described above.

RESULTS AND DISCUSSION
Cloning of the Yeast Gene Encoding yTAF25-As shown in Fig. 1, yTAF25 is the smallest readily visualized TAF within our yeast TAF fraction, which was prepared via immunoaffinity chromatography using affinity purified rabbit anti-TBP IgG (see also Refs. 11,25,29,and 59). In order to begin the molecular characterization of yTAF25, the TAF fraction was fractionated by SDS-PAGE and electrotransferred to nitrocellulose, and the M r ϭ 25,000 polypeptide was visualized with Ponceau S. The band corresponding to yTAF25 was excised and digested with trypsin, and five of the resulting HPLC-purified tryptic peptides were subjected to amino acid sequencing. Initial data base searches using these derived sequences and the BLAST search algorithms (41) revealed that these sequences were novel.
In order to clone the gene encoding yTAF25, the amino acid sequence of one of the peptides ((K/R)VVLTVNDLSSAVAEY) was used to design degenerate oligonucleotide primer pairs for use in the polymerase chain reaction for amplification of the DNA encoding this peptide. Total yeast cDNA library DNA (42) was used as the template for this PCR reaction. As detailed under "Materials and Methods," two separate rounds of PCR amplification had to be performed prior to obtaining the correct length product, probably due to the degeneracy of the oligonucleotides used as primers. The correct length PCR product was cloned, and the nucleotide sequence of the insert was determined. This nondegenerate DNA sequence was then synthesized, labeled, and used as a probe to screen a yeast genomic library (ATCC 77164). One of the positive clones obtained was selected for sequencing and an ORF of 618 nucleotides was found within the region of DNA sequenced (Fig. 2). As shown by the underlined portions of Fig. 2, this ORF encoded all five of the peptides generated from the M r ϭ 25,000 polypeptide that we had sequenced suggesting that this DNA indeed encoded yTAF25. To substantiate this fact we prepared polyclonal antibodies against protein expressed from the 206-amino acidlong putative yTAF protein that had been produced in E. coli (see "Materials and Methods"). Affinity purified IgG prepared from this serum with the purified recombinant protein was able to detect yTAF25 in the TAF fraction (data not shown). These data corroborate that we had in fact cloned the gene encoding yTAF25. This gene has been termed TAF25 for TBPassociated factor 25,000. The predicted molecular weight of yTAF25 is M r ϭ 23,018, which is quite similar to its apparent M r observed in our SDS-PAGE analyses. The protein is some-what acidic, exhibiting a calculated pI of 4.2, and it is also relatively glutamine-rich particularly in the C-terminal portion of the molecule. Recent BLAST searches with TAF25 as the query sequence revealed that it is identical to an uncharacterized ORF found on S. cerevisiae chromosome IV (GenBank accession number S50913).
Characterization of TAF25-When the deduced amino acid sequence of yTAF25 was analyzed, no striking similarities were found between yTAF25 sequences and previously characterized conserved protein structural or functional motifs. However, because the C-terminal region of yTAF25 is somewhat glutamine-rich, there was some similarity observed to other glutamine-rich proteins. While this work was in progress Jacq et al. reported the cloning and characterization of a 30-kDa human TAF II termed hTAF II 30 (17). Direct sequence comparisons of yTAF25 to this similarly sized human TAF protein showed that there are two regions, yTAF25 residues 82-128 and residues 180 -198, which show high sequence similarity (59.6 and 52.6% identity, respectively) to portions of hTAF II 30 (Fig. 3). However, overall sequence identity between yTAF25 and hTAF II 30 is only 27%. The finding of yet another yTAF that is similar at the amino acid sequence level to a metazoan TFIID subunit lends strong additional support to our previously stated hypothesis (11) that the cloned yTAF genes encode subunits of a yeast TFIID multiprotein complex (see also Ref. 12).
TAF25 Is a Single Copy Essential Yeast Gene-As a prelude to gene disruption experiments, we characterized TAF25 genomic DNA and RNA species to rule out the potential for multiple yTAF25-encoding genes. Genomic DNA blots (not shown) indicated that TAF25 is present at a single copy per haploid genome, whereas RNA blots showed that only a single 0.73-kilobase RNA species anneals with TAF25-derived probes (not shown). In order to test whether this apparently single copy gene was required for viability, we created the diploid yeast strain yEK2, which contained one wild type and one null mutant copy of TAF25 (i.e., strain yEK2, genotype TAF25/taf25⌬1::TRP1). When yeast strain yEK2 was sporulated and the resulting tetrads were dissected, viability segregated 2 ϩ :2 Ϫ and all viable spores were Trp Ϫ , suggesting that TAF25 is an essential gene. When yEK2 was subsequently transformed to Ura ϩ with the URA3-marked plasmid pRS416-TAF25 and this strain was subjected to sporulation and the resulting tetrads dissected, all four spores were viable, all Trp ϩ spores were Ura ϩ , and the Trp ϩ phenotype segregated 2 ϩ :2 Ϫ as expected. These genetic and biochemical analyses show that TAF25, like the previously characterized yeast TAF-encoding genes (TAF170, TAF150, TAF130, TAF90, and TAF60 (11,25), is an essential gene indicating an important role for these gene products in yeast cell physiology.
yTAF25 Is a Bona Fide yTAF-One definitive method of demonstrating that two proteins are associated with one another in a specific complex is to show that the two proteins coimmunoprecipitate. We wanted to test if yTAF25 truly was associated with TBP in crude yeast protein fractions and decided to use coimmunoprecipitation methods to address this question. The first immunoprecipitation experiment we performed utilized anti-HA antibody and WCE derived from a yeast strain expressing HA 3 -TBP. Following incubation of the anti-HA antibody with this WCE, the resulting immunoprecipitate was recovered, fractionated by SDS-PAGE, and electrotransferred to a PVDF membrane. We then determined if yTAF25 coimmunoprecipitated with precipitated TBP by blotting the membrane with anti-TAF25 IgG. WCEs derived from yeast strains expressing either HA 3 -TAF25 or the nontagged wild type strain were assessed similarly, serving as positive and negative controls, respectively. The results of this assay are shown in the left half of Fig. 4. The slower mobility of the HA 3 -TAF25 relative to native yTAF25 is due to the addition of the HA 3 -tag at the N terminus of yTAF25. These data show that yTAF25 and TBP are associated, because coprecipitating yTAF25 was readily detected in the TBP immunoprecipitate. We next performed the converse experiment utilizing anti-TAF25 IgG for the immunoprecipitation step and anti-HA mAb for the immunodetection step. The results of this analysis, shown in the right half of Fig. 4, show that HA 3 -TBP is again readily detected in the yTAF25 immunoprecipitate, indicating once more that TBP and yTAF25 are associated. Taken together, these experiments clearly demonstrated that TBP and yTAF25 are complexed, thus confirming that TAF25 encodes a bona fide yeast TBP-associated factor or TAF.
yTAF25 Is Present in a Multiprotein Complex with TFIIDspecific TAFs-We extended our coimmunoprecipitation analyses to determine which if any of the other TAFs in our TAF fraction are directly or indirectly associated with yTAF25. Our previous biochemical studies indicated that yTAF25 is part of a complex consisting minimally of yTAFs 150, 130, 90, 60, 40, 30, 25, and TBP. To address this question directly using the immunological reagents developed in the current study, we again used coimmunoprecipitation techniques to ask whether the 25-kDa yTAF polypeptide is also in a complex with other known TAFs. We were aided in these studies by the availability of our collection of different strains of yeasts, which all separately express epitope-tagged versions of either TBP or one of the other known yeast TAFs (i.e. yTAFs 170 (25), 150, 130, 90, 60, (11), and 25 (this report) or Brf1p. 2 Each of these eight yeast strains express a specific TAF or TBP containing three copies of the influenza virus HA epitope tag appended to their N termini.
In the experiment depicted in Fig. 5 (top panel), immunoprecipitates were formed using WCEs prepared from each of these strains and the anti-HA mAb. An immunoblot of these immunoprecipitates shows that the epitope tagged TAFs or TBP are all readily detected using anti-HA mAb. Thus for each strain tested the only reacting species detected in the anti-HA immunoblot is the corresponding single, correct sized epitope-tagged polypeptide that the strain expresses. In Fig. 5 (lower panel) are presented the results of the immunoprecipitation experiment designed to determine which of the other yTAFs are in the yTAF25-TBP complex. In this experiment immunoprecipitates were again formed using anti-HA mAb, but in this case, the SDS-PAGE fractionated immunoprecipitates were probed with anti-TAF25 IgG. As can be seen in Fig. 5 (lower   In the most similar region between these two TAFs (yTAF residues 82-128 and hTAF30 residues 125-170), these molecules display ϳ60% identity. FIG. 4. yTAF25 is a bona fide TAF. Left panel, yTAF25 immunoprecipitates with TBP. WCEs from yeast strains that express wild type (WT) proteins, HA 3 -TBP, and HA 3 -TAF25 (yEK20; see "Materials and Methods"), respectively, were subjected to immunoprecipitation with anti-HA mAb 12CA5. The immunoprecipitated proteins were fractionated by SDS-PAGE, electrotransferred to a PVDF membrane, and blotted for the presence of yTAF25 using affinity purified anti-yTAF25 polyclonal IgG. The arrows and labels indicate yTAF25 and the more slowly migrating HA 3 -TAF25. Right panel, TBP immunoprecipitates with yTAF25. WCEs from the same strains used for the left panel were subjected to immunoprecipitation with anti-TAF25 IgG immunoprecipitates were harvested, fractionated by SDS-PAGE, electrotransferred to a PVDF membrane, and blotted with anti-HA mAb 12CA5. HA 3 -TBP and HA 3 -TAF25 are indicated by the arrows and labels. lanes [1][2][3][4][5], yTAF25 is detected in the immunoprecipitates formed using anti-HA mAb and WCEs derived from yeast strains expressing HA 3  Both of these TAFs are components of distinct non-TFIID, TBP⅐TAF-containing complexes (21,25). If the converse experiment is performed, i.e. immunoprecipitating yTAF25 (and associated polypeptides) from WCEs with polyclonal anti-yTAF25 antibodies and using anti-HA mAb for HA 3 -TAF and/or HA 3 -TBP immunodetection, the expected result is obtained; that is, yTAF25 is found to be associated with yTAFs 150, 130, 90, 60, 30, and TBP but not with either Mot1p or Brf1p (data not shown). These data support our previous conclusion that a yeast TFIID complex consists minimally of yTAFs 150, 130, 90, 60, 40, 30, and 25 and yTBP (11).
yTAF25 Is Able to Interact with Itself in Vivo-One of the features of hTAF II 30 described by Jacq et al. is that this protein is able to specifically interact with the hER (17). These workers determined that hTAF II 30 interacts directly with hER amino acids 300 -330, which lie within the AF-2 activation domain of the hER (60 -62). Interestingly, this precise region (hER amino acids 300 -330 termed the hER AF-2a domain) can function as a potent hormone-independent constitutive activator of transcription in yeast (63). We initiated studies to see if the structural similarity noted between yTAF25 and hTAF II 30 extended to the functional level. To accomplish this we asked whether we could show that yTAF25 could also bind to hER AF-2a sequences. For this purpose we used the variant of the yeast two-hybrid screening system described by Hollenberg and colleagues (38).
Using this yeast two-hybrid approach, we were not able to detect any indication of an interaction between the hER AF-2a region and yTAF25 (data not shown), which suggests that the structural similarity we noted between yTAF25 and hTAF II 30 may not extend to the functional level, although to definitively address this point will require additional experimentation. However, we did notice that yTAF25 could apparently interact with itself. The ␤-galactosidase assay data presented in Table  I demonstrate that yTAF25 interacts with itself in vivo. The LexA-hER strain (strain 1), which contains both a DNA binding domain (from LexA) and a potent activation domain (the hER2a domain (63)) within a single fusion protein, as expected, gave the highest ␤-galactosidase values, almost 69 units of activity. The helix-loop-helix interacting pair of proteins, MyoD and Daughterless, present in strain 2 had moderate ␤-galactosidase levels (ϳ3 units). This level of reporter expression is FIG. 5. yTAF25 coimmunoprecipitates with TFIID-specific TAFs. Top panel, HA 3 -TBP or HA 3 -TAF moieties are immunoprecipitable from WCEs prepared from yeast strains expressing the relevant HA 3 -tagged proteins. WCEs were prepared from yeast strains expressing the indicated HA-tagged versions of TBP or TAF. HA 3 -tagged proteins present in the WCEs were immunoprecipitated using anti-HA mAb 12CA5, and the immunoprecipitates were fractionated by SDS-PAGE, blotted to a PVDF membrane, and proteins containing the HA 3 epitope were detected by blotting with anti-HA mAb 12CA5. The arrows and labels indicate the various epitope tagged proteins. Bottom panel, yTAF25 is only detectable in anti-HA mAb immunoprecipitates derived from strains expressing HA-tagged proteins that are putative TFIIDspecific TAFs. Immunoprecipitates were prepared using anti-HA antibody from the WCEs used in the top panel, and the immunoprecipitates were fractionated via SDS-PAGE and transferred to a membrane, but in this case the blots were probed with polyclonal anti-yTAF25 IgG. The arrows and associated labels indicate the presence of either native TAF25 or the more slowly migrating HA 3 -TAF25.
FIG. 6. GST pull-down experiments show that yTAF25 interacts directly with itself. Glutathione beads containing approximately 5 g of either GST-TAF25 or GST alone were incubated with the indicated amounts of 32 P-labeled HMK-His 6 -TAF25 and washed as described under "Materials and Methods". Resin-bound proteins were separated by 12% SDS-PAGE and visualized by exposing the dried gel to x-ray film. Alternate lanes show the 32 P-yTAF25 protein that bound to GST-TAF25 (odd-numbered lanes) or to the GST resins (even-numbered lanes). The arrow indicates the labeled yTAF25 protein bound to the resin. The faster migrating band most likely represents a 32 P-yTAF25 degradation product. The results of the binding assay were quantitated by PhosphorImager and are presented graphically as an inset in the graph. Only the predominant, slower migrating 32 P-labeled yTAF25 species was quantitated.  7. Immunodepletion demonstrates that a yTAF25-containing complex is required for both basal and activated RNA polymerase II transcription but not for RNA polymerase I or RNA polymerase III transcription. A, immunodepletion was able to remove most of the yTAF25 from a yeast whole cell extract. A yeast WCE expressing HA 3 TAF25 as the only form of yTAF25 was immunodepleted of yTAF25 and associated proteins using anti-HA mAb 12CA5 cross-linked to protein A-Sepharose as described under "Materials and Methods". Prior to incubation with WCE, the resin-bound mAb was preincubated with either buffer alone (DEPLETED) or a 100-fold molar excess of specific peptide recognized by the antibody (MOCK DEPLETED). To determine the extent of immunodepletion, the indicated quantities of supernatant resulting after harvesting the immunoprecipitates by centrifugation were fractionated by SDS-PAGE, transferred to PVDF, and immunoblotted using polyclonal anti-TAF25 IgG for detection. The signal on the immunoblot representing yTAF25 is illustrated by the label and the arrow. The relative amount of yTAF25 remaining in the supernatant as compared with untreated WCE was quantitated using a Bio-Rad Imager and is indicated by the numbers below the blot. B, specific RNA polymerase I activity is unaffected by depletion of yTAF25-containing complexes. As described under "Materials and Methods," the variably treated WCEs were tested in an in vitro RNA polymerase I-specific transcription assay. The results of the assay were scored by primer extension analysis. Following the transcription assay a constant number of cpm of a 300-nucleotide 32 P-labeled fragment was included in all the subsequent steps of the analysis as a recovery control. The autoradiograms pictured represent the expected 50-nucleotide extension product for the rRNA primary transcript and the 300-nucleotide recovery control. The results were quantitated by imaging and are graphically depicted (bottom). After subtracting for background and normalizing to the recovery control, the PhosphorImager units detected for each extract were as follows: Control, 2335 units; Depleted, 2278 units; and Mock depleted, 2554 units. C, specific transcription by RNA polymerase III is not affected by depletion of yTAF25-containing complexes from WCE. The two WCEs were assayed for specific RNA polymerase III transcription using a tRNA Leu3 gene as the template. The autoradiogram shown reveals the products of the transcription assay with primary, partially processed, and mature tRNA Leu3  considerably lower than that of strain 5, which contains the LexA-TAF25 and TAF25-VP16 fusion proteins (ϳ38 units of activity). The LexA-TAF25 strain (strain 3) had ␤-galactosidase levels approaching those of the MyoD-Daughterless pair (ϳ2 units of activity versus ϳ3 units). This activity can either be attributed to the presence of a "cryptic" activation domain within the yTAF moiety of the LexA-TAF25 fusion or to the potential formation of TFIID seeded by the local high concentration of DNA bound LexA-yTAF25 at the promoter of the lexA-driven transcription unit (i.e., lexAop 8 -lacZ). Such a phenomenon could be similar to that described for DNA tethered Brf1p, TBP, or Gal11p (64 -67). If TFIID recruitment/formation is the explanation for the low but significant levels of ␤-galactosidase expression in the yeast cells expressing LexA-TAF25, then it is possible that direct TAF25-TAF25 interactions could play a significant role in this process.
yTAF25 Directly Homomultimerizes in Vitro-Although the two-hybrid results described above suggested that yTAF25 could interact with itself in vivo, these assays do not prove that this interaction is a direct one. To test this idea we performed in vitro protein-protein interaction studies using purified proteins. As detailed under "Materials and Methods," GST pulldown experiments were performed utilizing GST or GST TAF25 beads and 32 P-labeled yTAF25. Following incubation of the probe 32 P-TAF25 with both matrices, resin-bound 32 P-labeled yTAF25 was monitored by SDS-PAGE and autoradiography. An autoradiogram resulting from these binding reactions is presented in Fig. 6 along with a graphical representation of this experimental data obtained by Phosphorimager analysis. As can be seen from the graph, dose-dependent yet saturable binding of HMK-His 6 -TAF25 to the GST-TAF25 beads was observed with little or no binding of the labeled protein to the GST resin. The specificity of this binding was assessed by performing competition experiments where increasing amounts of unlabeled His 6 TAF25 (range of cold competitor, 30 -8000 ng) were included in binding reactions containing a constant amount of GST-TAF25 resin and 32 P-labeled yTAF25. These experiments revealed that unlabeled His 6 TAF25 protein could effectively compete with the 32 P-TAF25 protein by essentially eliminating labeled protein binding to the GST-TAF25 resin (data not shown), whereas a comparable amount of a control protein, bovine serum albumin, had no effect upon the binding of 32 P-labeled TAF25 to GST-TAF25. It should also be noted that all of these GST pull-down binding studies were performed in the presence of a large excess of nonspecific E. coli lysate proteins (see "Materials and Methods"). These data clearly indicate that the binding of yTAF25 to itself is both direct and specific.
The physiological relevance of yTAF25-yTAF25 interactions remains to be determined. It is possible that yTAF25 could be present in two copies in the yTFIID complex, or alternatively this protein could facilitate interactions between two TFIID molecules. The exact stoichiometry of the various components of the yTAF25-containing yeast TFIID complex or for that matter the exact stoichiometry of TAF subunits in metazoan TFIIDs remains to be established.
yTAF25-containing Complexes Are Specific to the Function of RNA Polymerase II-With the exception of the SNAP complex isolated from HeLa cells, which appears to be required for transcription of small nuclear RNA genes by both RNA polymerases II and III (24), all of the other initiation factor TAF⅐TBP complexes thus far characterized appear to be dedicated to transcription by a single RNA polymerase. With this in mind our next goal was to examine the RNA polymerase specificity of yTAF25-containing TBP⅐TAF complexes. Before initiating these studies, our prediction was that yTAF25 would be part of the RNA polymerase II-specific TFIID complex. This hypothesis was based upon two observations. First, there are striking homologies between the sequences of the yeast TAFs (including yTAF25) of the yTAF⅐TBP core complexes previously described (11,12) and metazoan TFIID subunits. Secondly we (11) and others (12) have shown that yeast TAF fractions, prepared using two quite distinct affinity chromatographic approaches, could specifically mediate activation by upstream activators in vitro, indicating that the proteins in the yeast TBP⅐TAF fraction possess TFIID-like coactivator activity. Taken together these findings suggested but did not prove that the yTAF25-containing TAF⅐TBP complex represents a yeast TFIID complex and therefore would be RNA polymerase II-specific.
The strategy that we took in order to test this idea was to immunodeplete a yeast WCE of yTAF25 and associated proteins and then use in vitro transcription assays to measure the ability of the depleted extracts to support specifically initiated transcription by each of the three distinct DNA-dependent RNA polymerases. If the yTAF25-containing complex is uniquely involved in a TFIID complex, then only RNA polymerase II transcription should be affected by the depletion process. As shown in Fig. 7A treatment of yeast WCE with the anti-HA mAb 12CA5 removed ϳ70% of the HA-tagged yTAF25 from the WCE. This depletion was specific because control, "mock depleted" anti-HA mAb (i.e.. 12CA5 mAb preincubated with the HA peptide prior to mixing with WCE) failed to remove yTAF25 from the WCE (compare left and center lanes, Fig. 7A). When this depleted extract was tested for its ability to support transcription by RNA polymerase I (Fig. 7B) or RNA polymerase III (Fig. 7C), only minimal effects were observed, because in both cases the depleted extract was as competent for transcription as was the mock depleted extract. Marked effects, however, were observed when the depleted extract was tested for either basal or Gcn4p-activated RNA polymerase II transcription. Basal transcription mediated by the depleted WCE was reduced to ϳ10% of the mock depleted control (Fig. 7D,  left), whereas Gcn4p-activated transcription was reduced to 15-20% of the control (Fig. 7D, right). This data clearly demonstrated that the yTAF25-containing TBP⅐TAF complex was RNA polymerase II-specific, and thus yTAF25 was designated yTAF II 25 to indicate this fact. Furthermore these data argue that the yTAF II 25-containing TBP⅐TAF complex represents the major if not the only form of yTFIID, because if the TAF II 25containing TFIID complex represented only a subset of the total cellular yTFIID, then one would not expect to see such a large decrease in basal RNA polymerase II-specific transcription upon its depletion from the WCE. However, the possibility that other isoforms of TFIID may exist in yeast cannot be eliminated. Apparent multiple forms of TFIID have been reported in the human system (17).
The addition of purified TBP to these depleted and mock depleted WCEs results in the recovery of both the basal transcription signal and surprisingly the signal from Gcn4p activated transcription (not shown). There are several possible explanations for this result. First, because we were only able to deplete ϳ70% of yTAF II 25 and associated polypeptides from the WCE, the addition of TBP could lead to the reformation of a functional TFIID complex, particularly if yTAF II polypeptides are in excess relative to TBP. Such reformed TFIID could then function in both basal and activated transcription. However, recent quantitative immunoblotting experiments in the our lab 3 indicate that TBP, TAF II s, and TAF III s are roughly stoichiometric in yeast cells, a result consistent with functional studies of TBP content in vivo performed by others (68 -70). Second, the transactivation event observed with the depleted WCE upon TBP addition could theoretically be fundamentally different in mechanism from that observed prior to depletion or in the mock depleted WCE. Two groups recently identified and characterized "holoenzyme" forms of yeast RNA polymerase II, which are competent for mediating transactivation in vitro (71,72) and in vivo (73). Holoenzyme preparations can support Gcn4p (the activator used in our studies)-mediated transactivation in vitro in reconstituted transcription systems dependent on TBP (72). Because neither TBP nor TAFs appear to be components of the multisubunit RNA polymerase II holoenzyme (11,59,71,72) and because the amount of this form of enzyme, which is competent for transcriptional activation, would be unchanged in our depleted WCE, transactivation events could still occur in our "depleted" WCEs if TBP is resupplied. Finally, extensive genetic and biochemical analyses have implicated various coactivator or adaptor molecules in RNA polymerase II transactivation events. Two such adaptor molecules that have been well characterized are Ada2p and Sug1p (74 -76). These proteins can both directly interact with the activation domains of transcriptional regulatory proteins and with TBP, thereby potentially obviating any requirements for TAFs for transducing signals between DNA bound transactivator proteins and the basal transcription machinery. Clearly such a scenario would suggest that once again, our depleted extract could still support a Gcn4p-mediated transactivation event if supplemented with TBP. Final resolution of this conundrum must await additional characterization of yeast adaptors, holoenzyme, and TAFs.
In this report we have described the identification, cloning, and characterization of a gene termed TAF25 that encodes a novel yeast TBP-associated factor, yTAF II 25. The single copy TAF25 gene is essential for vegetative growth, a property this gene shares with other yeast TAF-encoding genes. Further we have demonstrated that this protein is a bona fide TAF and is a subunit of a yeast TFIID multiprotein TBP⅐TAF complex. Additionally, we have found that the deduced amino acid sequence of yTAF II 25 bears significant sequence similarity to a human TAF, though it is still not clear if yTAF II 25 and hTAF II 30 are functionally related. Through the work described in this report, we have added to the body of evidence supporting the existence of a discrete yeast TFIID complex by directly demonstrating that the yTAF25-containing TBP⅐TAF complex is specific to RNA polymerase II function. Additional work will be needed to document the exact composition of the yeast TFIID complex and to determine if it contains only TBP and yTAF II s 150, 130, 90, 60, 40, 30, and 25. Cloning all of the genes encoding these yTAF II s will allow us to reconstitute this complex from its component parts to further study the exact role of TFIID in transcription by RNA polymerase II.