Genetic and Biochemical Analyses of Yeast TATA-binding Protein Mutants*

We have taken a combined genetic and biochemical approach to study TATA-binding protein (TBP) struc-ture-function relationships. Using site-directed muta- genesis coupled with a screen for conditional lethal growth, we have isolated a number of temperature- sensitive TBP alleles in the region of amino acid positions 188, 189, and 190. Conditional growth is not a result of increased TBP turnover as most of the mutant proteins are stable in vivo as evidenced by immunoblot detection of TBP steady-state levels. DNA binding as- says reveal that mutations at position 188 do not affect DNA binding activity of these mutants, even at high temperatures. Utilizing whole cell extracts which con- tain mutant TBPs in in vitro transcription experiments, we confirm that TBP is required for transcrip- tion by all three nuclear polymerases. However, certain of our TBP mutants are only compromised for RNA polymerase I1 transcription. The regulated expression of eukaryotic class I1 genes of interactions between factors bound at enhancers or upstream activating sequences and the RNA polymerase I1 preinitiation complex formed typically around the TATA box DNA (for (19901, Greenblatt Roeder Pugh and Tjian (1992)). The nature

We have taken a combined genetic and biochemical approach to study TATA-binding protein (TBP) structure-function relationships. Using site-directed mutagenesis coupled with a screen for conditional lethal growth, we have isolated a number of temperaturesensitive TBP alleles in the region of amino acid positions 188, 189, and 190. Conditional growth is not a result of increased TBP turnover as most of the mutant proteins are stable in vivo as evidenced by immunoblot detection of TBP steady-state levels. DNA binding assays reveal that mutations at position 188 do not affect DNA binding activity of these mutants, even at high temperatures. Utilizing whole cell extracts which contain mutant TBPs in in vitro transcription experiments, we confirm that TBP is required for transcription by all three nuclear polymerases. However, certain of our TBP mutants are only compromised for RNA polymerase I1 transcription.
The regulated expression of eukaryotic class I1 genes is thought to occur as the result of interactions between factors bound at enhancers or upstream activating sequences and the RNA polymerase I1 preinitiation complex formed typically around the TATA box DNA element (for reviews see Ptashne and Gann (19901, Greenblatt (1991), Roeder (1991), Pugh and Tjian (1992)). The nature of these interactions is not well understood but may involve direct protein-protein contacts between transcriptional activators and selected components of the general transcription machinery. In studies designed to identify the immediate target(s) of the herpes simplex virus VP16 activation domain, it was shown that either transcription factor IIB (TFIIB)' or TBP was selectively retained when nuclear extracts were chromatographed on columns containing immobilized VP16 (Stringer et al., 1990;. These protein-protein interactions were presumed specific since VP16 mutants which fail to activate transcription efficiently in viuo exhibited diminished binding to either TFIIB or TBP in uitro (Ingles et al., 1991;. Still, these direct contacts probably do not represent the sole mechanism for the modulation of transcription initiation as there is a requirement for factors other than the basal factors (for recent reviews, see Sawadogo and Sentenac (1990), Roeder (1991), Weinmann (1992)) to achieve maximal activated transcription in vitro. These additional factors have been termed mediators or coactivators and may function as a bridge between transactivators bound at upstream activating sequences and their targets in the basal transcription complex (perhaps TFIIB and TFIJD -TBP) by making specific proteinprotein contacts (Berger et al., 1990;Kelleher et al., 1990;Pugh and Tjian, 1990;Dynlacht et al., 1991;Flanagan et al., 1991;Tanese et al., 1991).
The gene encoding TFIID or TATA-binding protein (TBP) was first cloned from Saccharomyces cerevisiae (Cavallini et al., 1989;Hahn et al., 1989;Horikoshi et al., 1989;Schmidt et al., 1989) and was found to be identical to SPT15 (Eisenmann et al., 1989), a suppressor of Ty insertion mutations of HIS4. This single copy essential gene encodes a 27-kDa protein.
Initial structure-function studies on yeast TBP revealed that the nonconserved amino terminus was dispensable for both in vitro  and in uiuo function in most yeast strains (Cormack et al., 1991;Gill and Tjian, 1991;Poon et al., 1991;Reddy and Hahn, 1991;Zhou et al., 1991), suggesting that all essential activities of TBP (specific DNA binding, the ability to associate with other general transcription components, and the ability to respond to or interact with upstream activators and/or coactivators) were localized to the conserved carboxyl 180 amino acids of the molecule. Unfortunately, conventional deletion analyses failed to identify precisely individual functional domains within the carboxyl-terminal portion of TBP as any deletion within this conserved region inactivated the protein Poon et al., 1991). However, results from point mutagenesis studies suggest that TBP contains a bipartite DNA binding domain and that this domain is localized to the direct repeats (Reddy and Hahn, 1991;Yamamoto et al., 1992;Arndt et al., 1992). A similarly detailed mutagenesis study indicates that the basic region of yeast TBP appears to be involved in direct binding to TFIIA (Buratowski and Zhou, 1992). The basic region of human TBP has been implicated in direct 5005 binding to the viral activators Ela and Zta (Lee et al., 1991;Lieberman and Berk, 1991).
TO identify other important functional domains within yeast TBP we have taken a combined genetic and biochemical approach. From a previous study, we found that a mutation from leucine (L) to lysine (K) at TBP amino acid position 189 which had no effect on DNA binding or basal transcription in a HeLa nuclear extract in vitro (Yamamoto et al., 1992) totally abolished function in vivo.* This discrepancy between the in vitro and in vivo analyses of TBP function led us to focus our attention on the region encompassing position 189, and we devised a genetic scheme to test whether or not this portion of TBP was functionally important. Our goals in this study were 2-fold. First, we wanted to see whether our site-directed mutagenesis scheme could be used to generate temperature-sensitive TBP alleles efficiently and second, to see whether these mutant TBP alleles could be manifested as functional defects. Using our saturation mutagenesis method we were able to isolate conditional lethal mutations within this region of the gene encoding TBP. Furthermore, our data indicate that we have isolated TBP mutants that bind DNA efficiently but which fail to promote basal transcription by RNA polymerase 11.

MATERIALS AND METHODS
Yeast Strains and Media-All yeast strains used in this study are described in Table I. Yeast cells were propagated in YPA medium (2% yeast extract, 1% Bacto-Peptone, and 40 pg/ml adenine sulfate) or synthetic complete medium supplemented with the appropriate amino acids and carbon sources. Strains YTW22 and YSDl3 were described previously (Poon et al., 1991). Plasmids were introduced into yeast using the lithium acetate transformation procedure. All yeast manipulations were performed as described (Guthrie and Fink, 1991).
In Vitro Mutagenesis and Isolation of Mutants-Uracil containing single-stranded DNA was produced from plasmid pTFIID-WT (Poon et al., 1991) in Escherichia coli strain CJ236 (dut-, ung-). Mutagenic oligonucleotides were designed to contain random nucleotides at designated codon positions (see below). Each mutagenic oligonucleotide was treated with T4 polynucleotide kinase and used separately in in vitro mutagenesis reactions. The resulting mutated DNAs were used to transform E. coli strain DH5a (dut+,ung') to select for the mutated plasmids. Each pool of mutant DNA was then purified and used to transform YTW22 to His+. Transformation plates were incubated at 30 "C for 2 days, and then transformants were replica plated onto synthetic complete medium plus 2% glucose plates containing 1 mg/ml5-fluoroorotic acid (5-FOA) (Boeke et al., 1987). The replica plates were subsequently incubated at 12, 25, 30, and 35 "C for 3-6 days (10-16 days for the 12 "C set). Cells that grew at 25 or 30 but not 12 or 35 "C were isolated and retested for the conditional lethal phenotype. About 4,000 colonies were screened per codon position mutagenized. Mutated plasmids were recovered from the mutant strains, passaged through E. coli, and were sequenced using the dideoxy method to determine the mutation responsible for the conditional lethal phenotype and to ensure that only a single mutation arose from this mutagenesis scheme. To ensure that the TBP mutant alleles were responsible for the conditional growth phenotype, the recovered plasmids were transfected into YTW22 followed by 5-FOA selection. The mutagenic oligonucleotides used in this study were as follows (boldface N denotes any of the four nucleotides, A, G, T, or 0 . Immunoblot Analysis-In vivo steady-state levels of wild-type and mutant TBPs were determined as follows. Plasmids containing the TBP mutant alleles were recovered from the temperature-sensitive strains and passaged through E. coli. The purified plasmids were then used to transform strain YSD13 to His+ individually. Transformants were restreaked to isolate single colonies which were then used to inoculate 100 ml of synthetic complete medium plus 2% glucose cultures. These cultures were incubated overnight with shaking at 35 "C to mid-log phase. Cells were harvested and washed with cold H20. 10 Am units of cells were suspended in 150 pl of RIPA buffer (50 mM Tris-C1, pH 7.9, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS) supplemented with 0.1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 2 pg/ml leupeptin, and 1 @g/ml pepstatin. Extracts were prepared by vortexing the cell suspension in the presence of glass beads (six pulses of 20-s duration each). 50 pl of extract (about 150 pg of total protein) was fractionated by SDS-PAGE. TBP was detected by immunoblotting as described previously (Poon et at., 1991) except that 3-amino-9-ethylcarbazole was used as the developing reagent instead of 4-chloro-1-naphthol.
Purification of Recombinant TBP-TBP mutant alleles were subcloned, as polymerase chain reaction fragments, into the NdeI and BamHI sites of the T7 polymerase expression vector PET-3a (Studier et al., 1990). For the polymerase chain reaction the upstream primer was5'-GCGCGGAATTCATATGGCCGATGAGGAACGTTTAA ACGAG-3' (the NdeI site is underlined; nucleotides corresponding to TBP sequences +1 to +20 are indicated in bold). The downstream primer was 5'-GCGCGGGATCCTATTATCACATTTTTCTAAA TTCACTTAGCAC-3' (the BamHI site is underlined nucleotides corresponding to TBP sequences +697 to +723 are indicated in bold).
Coordinates are relative to the position of the initiating ATG codon.
The resulting constructs were sequenced in their entirety. The constructs were transformed into E. coli strain BL21(DE3) (Studier et al., 1990), and recombinant proteins were prepared as follows. Cells were grown at 30 "C in 100-ml cultures of TBGM9 (1% tryptone, 0.5% NaCI, 0.5% glucose, 1 mM MgS04, 0.1% NH4CI, 0.3% KH2P04, 0.6% Na2HPOJ plus 100 pg/ml ampicillin to an Asm of 0.6 after which isopropyl-1-thio-P-D-galactopyranoside was added to 0.4 mM. The cultures were incubated at 30 "C for an additional 3 h. The induced cells were harvested, washed in cold buffer containing 20 mM Tris-C1, pH 7.9, and 200 mM NaCl and resuspended in 4 ml of LB buffer (20 mM Tris-C1, pH 7.9, 10% glycerol plus 20 mM 2-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 2 pg/ml pepstatin, 2 pg/ml leupeptin) containing 200 mM NaCl (LB/200). Cell lysates were prepared by sonication. After removal of the cellular debris by centrifugation of the cell lysate, the supernatant was applied directly to heparin-Sepharose (1-ml resin/100 Am units of cells) preequilibrated with LB/200. The resin was washed with LB/200, and the bound protein was batch eluted with LB buffer containing 1.5 M NaCl. The final salt concentration of the eluate was about 600 mM NaCI, and the protein concentration was estimated to be approximately 1 mg/ml using the Bradford protein assay. The resulting TBP was estimated by SDS-PAGE and Coomassie Blue staining to be 50% pure. For use in in vitro transcription assays, TBP was dialyzed against whole cell extract dialysis buffer (20 mM HEPES, pH 7.9, 20% glycerol, 10 mM MgS04, 10 mM EGTA, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 50 mM KOAc. Mobility Shift Assays-The probe used in mobility shift assays contained the Ad2 major late promoter TATA box and was prepared by first annealing the following two oligonucleotides.

3'-ACAAGGACTTGCCGCCGATATTTTCCGCCACCGCCGCGCA-5'
OLIGONUCLEOTIDES 6 AND 7 Six pmol of the resulting template-primer was filled in using 15 units of E. coli Klenow DNA polymerase I in a 25-pl reaction containing 2 p~ [ w~~P ]~A T P (3,000 Ci/mmol); a 40 p M concentration each of dGTP, dCTP, dTTP; 10 mM dithiothreitol; 50 mM Tris-C1, pH 7.5; and 10 mM MgC12. The elongation reaction was allowed to proceed on ice for 45 min at which time an additional 4 nmol of all four dXTPs and 5 units of Klenow polymerase were added, and incubation on ice was continued for 15 min. The reaction was stopped by heating at 65 'C for 15 min, and unincorporated nucleotides were removed by gel filtration on Sephadex G-50. Typical probe specific activities obtained were 10,000 cpm/fmol. DNA binding assays were performed in 20-pI reactions containing 12.5 fmol of probe, 40 ng of recombinant TBP, 50 mM NaCl, 20 mM Y K S l 8 9 -1 1 Y K S 1 9 0 -1 0 Y K S l 9 0 -1 1  HEPES, pH 7.9,5 mM MgC12, 10% glycerol, 1 mM dithiothreitol, 100 ng of poly(dG.dC), and 50 pg/ml bovine serum albumin. Samples were incubated at either 30 or 35 'C for 30 min and subsequently fractionated on a 6% polyacrylamide gel cast and run in a buffer consisting of 12.5 mM Tris, 95 mM glycine, and 5 mM MgClz at either 25 or 35 "C. Equivalent DNA binding profiles were obtained regardless of the temperature at which binding reactions or electrophoresis was performed (data not shown). Whole Cell Extract Preparation and in Vitro Transcription Assays-Cells were grown in YPA plus 2% glucose at 25 "C to mid-log phase.
Pol I Transcription-Transcription reactions were performed according to Schultz et al. (1991). Each reaction contained approximately 250 pg of WCE and 5 pg/ml pYrllA as template which was linearized with SmaI. Where appropriate, wild-type recombinant TBP (amounts are indicated in the figure legends) was added to WCE and preincubated at 30 "C prior to the addition of template or nucleotides. It was necessary to linearize the template with SmaI to dissociate the T7 promoter (which is upstream of the SmaI site) from the 35s rRNA promoter (which is downstream of the SmaI site) because our recombinant TBP preparations contained residual T7 polymerase which gave rise to transcripts originating at the T7 promoter (data not shown).
Pol II Transcription-Transcription reactions contained 50 mM HEPES, pH 7.9, 120 mM potassium glutamate, 10% glycerol, 0.75% w/v polyethylene glycol 4,000, a 400 p~ concentration each of ATP and CTP, 2 p M UTP, 0.1 pM [CY-~~P]UTP (650 Ci/mmol), 15 mM MgAc, 5 mM EGTA, 2.5 mM dithiothreitol, 10 units of RNasin (Promega), 10 pg/ml creatine kinase, 30 mM creatine phosphate, and 20 pg/ml pGAL4GC-as template. Each reaction contained approximately 400 pg (optimal for our conditions) of whole cell extract protein; where appropriate, recombinant TBP (amounts are indicated in the figure legends) was added to WCE and template and preincubated at 25 "C prior to the addition of nucleotides. Nucleotides were added, and the transcription reaction mixtures were incubated at 25 'C for 30 min. 120 pl of RNase T1 in buffer containing 240 mM NaC1,8 mM HEPES, pH 7.9,O.a mM EDTA, and 15 units of RNase T1 (Boerhinger Mannheim) was then added, and the reactions were incubated for an additional 10 min at 25 "C. RNase digestion was terminated by the addition of 4 p1 of 20% (w/v) SDS and 4 p1 of 10 mg/ml proteinase K and incubation at 65 "C for 20 min. RNA was precipitated by the addition of 160 pl of 5 M NH4Ac and 200 p1 of isopropyl alcohol, harvested, washed with 70% ethanol, dried, and resuspended in 95% formamide, 10 mM EDTA, pH 7.9, 0.1% each bromphenol blue and xylene cyanol. RNA was fractionated on a 5% polyacrylamide TBE-buffered gel containing 7 M urea and visualized by autoradiography.
Pol III Transcription-Transcription reactions were performed according to Klekamp and Weil (1986). Each reaction contained approximately 250 pg of WCE and either 25 pg/ml plasmid pYLEU3 or plasmid pUC5S as template (Wang and Weil, 1989). Where appropriate, recombinant TBP (amounts are indicated in the figure legends) was added to WCE and preincubated at 25 "C prior to the addition of template and nucleotides.

RESULTS
Isolation of Temperature-sensitive Mutants-To generate mutant alleles of the TBP gene, a site-directed mutagenesis scheme was employed (Fig. lA). Synthetic oligonucleotides, designed to contain random nucleotides at codon positions corresponding to amino acid residues 187-191 of TBP ( Fig.  I B ) , were used individually, in conjunction with uracil-containing single-stranded DNA derived from pTFIID-WT (HZS3, CENIARS; Poon et al., 1991), for in vitro mutagenesis reactions. The subsequent pools of site-specifically mutated DNA were isolated and purified from E. coli and used to transform YTW22 (see Table I for genotype), a haploid yeast strain that contains a chromosomal deletion of the gene encoding TBP. Because the TBP gene is single copy and essential, YTW22 also contains PURA-TFIID ( Uracil containing single-stranded DNA was derived from pTFIID-WT (HZS3, CENIARS; Poon et al. (1991)) by propagation in E. coli strain CJ236 (dut-, ung-), Haploid yeast strain YTW22 contains a chromosomal deletion of the gene encoding T B P but remains viable because it contains a copy of the TBP open reading frame expressed from the PGK promoter on a URA3 marked plasmid. Each pool of mutagenized DNA was used to transform YTW22 to His'. The transformants were allowed to grow a t 30 "C for 2 days and were then replica-plated onto medium containing 5-FOA to select for cells that had lost the URA3 plasmid. Each set of replica plates was incubated a t 12, 25,30, and 35 "C to test for conditional growth phenotypes associated with mutation of the gene encoding TBP. In this schematic (lower) open circles represent colonies exhibiting conditional lethal growth, and dark circles represent wild-type growth. Panel B, diagrammatic representation of TBP. The short open rectangle represents the amino terminus of yeast T B P (amino acid residues 2-61), which is dispensable for vegetative yeast growth. The arrows represent the direct repeats (amino acid residues 68-127 and 158-218, respectively). The region targeted for mutagenesis in this study lies within the second direct repeat and is shown (PELFP; amino acid residues 187-191). Panel C, growth of the temperature-sensitive strains used in this study is shown. A key indicating the form of T B P present within each strain is indicated at the top; numbers indicate the amino acid position mutated. The letters are one-letter amino acid codes indicating wild-type (first letter) and mutant (second letter) residues. Cells were streaked onto YPA plus 2% glucose and incubated a t either 25 "C (left) or 35 "C (right) for 2-3 days. Poon et al., 1991), a plasmid that contains the TBP gene and thus complements the chromosomal TBP null allele.

ARS;
After growth on the appropriate medium, transformants were replica plated onto medium containing 5-FOA (Boeke et al., 1987) to select for cells that have lost the URA3 plasmid, PURA-TFIID. Replica plates were incubated at 12, 25, 30, and 35 "C to identify cells that fail to grow at either the high (35 "C) or low (12 "C) temperatures relative to the control plates (25 and 30 "C). Putative temperature-sensitive colonies were isolated and retested. Approximately 4,000 initial transformants were screened for each codon position targeted. Both temperature-sensitive and cold-sensitive mutants were isolated using this approach, but only the temperature-sensitive mutants that failed to grow a t 35 "C were used in this study.
Interestingly, no conditional lethal mutants were isolated from the 187P position (the number refers to the position, and the letter refers to the amino acid residue), and only one mutant was isolated from, respectively, the 189L and the 191P positions, suggesting that mutations at these positions are either well tolerated or lethal. The various amino acid substitutions at position 188 (see Table 11) were diverse, ranging from hydrophobic (isoleucine) to aromatic (tyrosine) to basic (histidine) to neutral (asparagine). Eight mutants were isolated from the 190F position, but all of these mutations were changes to either threonine or glutamine. Although all mutant strains grew at the permissive temperature (Fig.  lC), their growth rates were somewhat slower than wild-type on all media (rich or synthetic) and carbon sources (glucose, glycerol, lactate, or galactose) tested.
Mutant Forms of T B P Are Stable at the Nonpermissive Temperature in Vivo-The temperature-sensitive phenotypes observed in the mutant strains could be the result of rapid proteolytic degradation rather than altered activity of the mutant TBPs. Therefore, we designed an experiment that would directly assess the in uiuo stability of wild-type and mutant forms of TBP and thus allow us, using immunoblot analyses, to examine the steady-state levels of the mutant TBPs from cells grown at the nonpermissive temperature ( Fig. 2A). The plasmids containing the mutant TBP alleles were recovered from the temperature-sensitive strains, passaged through E. coli, and used to transform YSD13, a strain derived from YTW22 via 5-FOA selection. YSD13 contains a fully functional but amino-terminally deleted form of TBP (AN-TBP; amino acid residues 2-60 deleted; Poon et al. (1991)). The use of this isogenic host strain which contains the fully functional AN form of TBP has two advantages. First, AN-TBP completely complements the chromosomal TBP null allele, hence YSD13 derivatives grow at any desired temperature and yet still produce mutant TBPs which are expressed from another plasmid. Second, operationally our polyclonal antiserum reacts only with the amino-terminal portion of TBP (we estimate a differential reactivity, amino terminus/carboxyl terminus, of greater than 1,000-fold). Thus in immunoblot analyses using this antiserum we will detect only the temperature-sensitive TBP mutants and not the AN form of TBP. The results of the immunoblot analysis are shown in Fig.   2B. Recombinant TBP, expressed in and purified from E. coli, was included as a positive control. As expected, no full-length 27-kDa TBP was detected in the extract produced from YSD13 cells which contain just the vector alone (lane marked "insert). An immunoreactive 27-kDa polypeptide is detected, however, from YSD13 cells expressing wild-type TBP (see lane marked wildtype), as well as from cells expressing various mutant forms of TBP. These results reveal that, with the exception of mutants 188EI and 191PQ (which were unstable hence not studied further), all mutant TBPs are stable at the nonpermissive temperature and further, that steady-state levels are essentially the same as wild-type, thus demonstrating that the observed phenotype is not a reflection of protein instability.
Effects of Various Mutations on TBP DNA Binding Actiuity-TBP is required for the expression of genes transcribed by RNA polymerase 11. Any perturbation in TBP activity, whether it be the ability to bind DNA, interact with general transcription factors, or interact with upstream activators, could conceivably result in altered patterns of transcription of genes which require TBP for expression. Clearly, a defect in any of these functions of TBP could have resulted from our mutagenesis scheme. Furthermore, it is not expected that particular defects in these TBP activities would have predictable phenotypes. Either a defect in DNA binding or a defect in the interaction between TBP and TFIIA, for example, may give rise to the same conditional growth phenotype.
To determine the effects of the various mutations on TBP DNA binding activity, recombinant wild-type and mutant TBPs were produced and purified from E. coli and assayed for their ability to bind DNA specifically in vitro. The gel mobility shift assay was used to examine the TATA box binding properties of the various mutant forms of TBP (Fig.  3), and the reactions were incubated and electrophoresed at either 30 or 35 "C to determine whether DNA binding was temperature-dependent. The results from these mobility shift assays indicated that none of the mutations at position 188 (we were unable to produce sufficient amounts of purified 188EH protein) affected DNA binding at either temperature, whereas mutations at position 190 appear to abolish DNA binding. Therefore, this test for specific DNA binding revealed that TBP mutants with amino acid substitutions at position 188 (EN, EY, EP) bound DNA with essentially wild-type affinity regardless of temperature ( Fig. 3 and additional titration experiments, data not shown), whereas TBP mutants with amino acid substitutions at position 189 (LP) and 190 (FT, FQ) exhibited a great reduction in their ability to bind DNA. The altered migration of the 188 mutant TBP.TATA box complexes has been documented for other TBP mutants (Yamamoto et al., 1992) and presumably do not represent any global structural changes and do not correlate with any specific functional defect in vitro (Yamamoto et al., 1992)

or in uivo.2 TBP Mutations Have Pleiotropic Effects on Transcription by All Three RNA Polymerases in Vitro-
We next wanted to test whether our various TBP mutants productively interacted with all the components of the pol I1 basal transcription machinery. It has been demonstrated that TBP directly interacts with the general initiation factors TFIIA and TFIIB and perhaps other components of the general machinery Ranish and Hahn, 1991;Roy et al., 1991;Safer et al., 1991;Usuda et al., 1991;Koleske et al., 1992;Usheva et al., 1992). To test whether our TBP mutants were functional in pol I1 basal transcription, WCE (Woontner et al., 1991) were prepared from yeast cells containing wild-type 3OoC 35OC

FIG. 3. Effects of TBP mutations on DNA binding.
Mutant forms of TBP expressed in and purified from E. coli were tested for TATA box DNA binding activity using the mobility shift assay. 40 ng of purified TBP was mixed with a 40-base pair radiolabeled probe containing the Ad2 major late promoter TATA box. The reactions shown here were incubated a t 30 or 35 "C for 30 min and electrophoresed on a 6% polyacrylamide gel at room temperature or 35 "C, respectively. The slightly faster migration of mutants 188EN, 188EY, and 188EP is presumed to be the result of the indicated amino acid substitution. Such altered mobilities caused by single amino acid substitutions have been observed previously (Yamamoto et al., 1992). or selected mutant forms of TBP. For simplification, we chose one mutant strain representing each amino acid position to study in further detail (e.g. strains containing TBP mutants 188EN,189LP,and 190FQ,respectively). Specific transcription was analyzed using the plasmid pGAL4CG- (Lue et al., 1989) as template. This template contains a single GAL4 binding site upstream of the CYCI promoter fused to the Gcassette (Sawadogo and Roeder, 1985). In the absence of any exogenously added transcriptional activator, this template will direct only basal transcription. As shown in Fig. 4A, none of the mutant extracts supports basal transcription by pol I1 as efficiently as the wild-type control, suggesting that the mutations somehow affect the ability of the mutant TBPs to promote basal transcription. This apparent defect could be explained by the fact that mutations at positions 189 and 190 severely decrease DNA binding. Hence, unable to bind DNA efficiently, such a TBP mutant would also be unable to promote pol I1 initiation complex formation. This argument could be applied to the 189LP and 190FQ mutant TBP extracts but would seem very unlikely for the 188EN extract since mutations at position 188 do not affect DNA binding. Presumably a defect in another TBP functionality is responsible for the apparent inability of the 188EN extract to promote pol 11-catalyzed basal transcription.
Recently, it was shown that not only is TBP required for pol I1 transcription but that it is also required for transcription by pol I and I11 (Cormack and Struhl, 1992;Schultz et al., 1992;White et al., 1992). We therefore decided to test whether the above described mutant TBPs could efficiently promote pol I or pol I11 transcription in vitro. As shown in Fig. 4B, each extract was tested for its ability to specifically transcribe pol I (35 S rRNA gene promoter) and pol I11 genes (a tRNA gene and the 5 S rRNA gene). Unlike the results seen with the pol I1 assay, the 188EN extract was able to support both pol I and pol I11 transcription. Although the pol I and pol I11 transcriptional levels for the 188EN extract are somewhat lower than that of the wild-type extract, significant amounts of transcripts were produced, and thus this result indicates that this extract is not generally deficient in pol I or pol I11 transcription factors. Moreover, these pol I and pol I11 transcription signals are in sharp contrast to the pol I1 assay where there was absolutely no detectable transcription above background for the 188EN extract. Interestingly, the 189LP extract which contains a non-DNA-binding form of TBP was unable to support either pol I or pol I11 transcription, suggesting that either the specific mutation may affect the ability of this mutant form of TBP to promote transcription by all three polymerases or that the extract is nonfunctional for trivial reasons. Lastly, the 190FQ extract, which was unable to efficiently promote pol I transcription, was able to promote pol I11 transcription, albeit at a level lower than for the wildtype extract, despite the fact that this extract contains a non-DNA-binding form of TBP. This result suggests that DNA binding is not essential for TBP to participate in pol I11 transcription of the tRNA and 5 S genes. However, since we assayed only binding to the Ad2 major late promoter TATA box, we cannot rule out that TBP-DNA interactions may still play a role in pol I11 transcription.
Transcriptional Levels May Be a Reflection of TBP Levels in WCE-Because each extract was prepared individually, the relative transcription levels seen between extracts could possibly be a result of differential solubilization or stability of TBP during extract preparation. To ensure that each extract contained comparable amounts of endogenous TBP, a quantitative immunoblot was performed on these extracts as shown in Fig. 5. Increasing amounts of each extract were fractionated by SDS-PAGE, electroblotted to an Immobilon polyvinylidene difluoride membrane, and probed with anti-TBP antiserum as detailed under "Materials and Methods." The results show that the 188EN and 189LP extracts'contain somewhat less TBP (approximately 3-and 4-fold less, respectively) than the wild-type extract and that the 190FQ extract contained slightly more TBP than the wild-type extract (approximately 2-fold more). The lower levels of TBP in the 188EN and 189LP mutant extracts may explain in part the overall lower levels of transcription from these extracts. In fact, the %fold lower level of TBP in the 188EN extract may reasonably account for the approximate 4-fold lower pol I transcriptional activity and the approximate 5-fold lower pol I11 transcriptional activity but does not explain the absolute absence of pol I1 transcriptional activity in this extract.
Transcriptional Defects Are TBP-dependent-We next wanted to see whether the addition of recombinant wild-type TBP could restore or stimulate specific transcription by any or all RNA polymerases in these extracts. A stimulation of transcription in this case could either mean that a particular extract is really limiting in TBP or that the endogenous mutant TBP is functionally unable to promote transcription. In the experiment shown in Fig. 6, the amount of the various extracts was held constant, and increasing amounts of recombinant wild-type TBP were added to each extract. The results indicate that the addition of recombinant wild-type TBP had no stimulatory effect on transcription in the wild-type extract. In fact, high amounts of recombinant wild-type TBP had a slight inhibitory effect on both pol I1 and pol I11 transcription, presumably because of sequestration of an essential factor(s) by excess TBP. From various immunoblot experiments ( Fig.  5 and data not shown) we estimate that in these T B P addition experiments, we are adding no more than a 5-fold molar excess of recombinant T B P relative to the amount of endogenous TBP in these extracts (see also Fig. 6 legend).
In the reactions using extract 188EN, the addition of recombinant T B P restored pol I1 transcription but had no effect on either pol I or pol I11 transcription. Similar to the case with the wild-type extract, high amounts of recombinant TBP had a slight inhibitory effect on pol I11 transcription in the 188EN extract. The fact that only pol I1 transcription is stimulated by the addition of recombinant TBP suggests that the 188EN extract is specifically deficient in the ability to promote pol I1 transcription. In the case of the 189LP extract, the addition of recombinant T B P resulted in a large stimulation of both pol I1 and pol I11 transcription, whereas in the case of the 190FQ extract, the addition of recombinant TBP resulted only in a slight stimulation in pol I1 and pol I11 activity, with the stimulation of pol I11 activity greater than pol I1 activity.
We were unable to stimulate pol I transcription in any of the extracts by the addition of recombinant TBP. Recently, Schultz et al. (1992) noticed a requirement for a high temperature preincubation of extract with TBP to restore pol I transcription in extracts deficient in TBP activity. With our extracts, this pretreatment step had no effect on the ability of exogenous TBP to restore pol I transcription, even with preincubations a t 30 or 37 "C. Therefore we have not ruled out the possibility that the mutant TBP in the 188EN extract could also be defective in pol I transcription. However, we think this is unlikely given the fact that both pol I and pol I11 transcriptional activities are comparable in this extract and that the addition of recombinant TBP had no effect on pol I11 activity but greatly stimulated pol I1 activity. Lastly, the addition of recombinant wild-type TBP to the mutant extracts restored pol I1 and pol I11 transcription only to a fraction of the levels seen with the wild-type extract, It is possible that the mutant extracts may contain lower levels of the other general transcription factors, perhaps at levels comparable to the lower T B P levels.

Pol I Z Specificity at Amino Acid Position 188-The results
presented thus far indicate that the TBP mutation at position 188 from glutamic acid to asparagine specifically affects the ability of this factor to promote transcription by RNA pol 11. We next wanted to see whether the other mutations at this position also would give rise to pol I1 specific defects. Rather than make WCE from the other mutant strains, we decided to use the complementation assay described above to compare directly the transcriptional activities between the individual mutants. Our 189LP extract described above served as our TBP-deficient extract since transcription can be restored by the addition of recombinant TBP, at least for pol I1 and pol I11 genes. The results of such a complementation assay are shown in Fig. 7. Consistent with results shown earlier, all 188 mutants were able to function in pol I11 transcription, but none of them could support pol I1 transcription. Taken together these results indicate that mutational analyses of TBP amino acid 188 had allowed us to identify a pol 11-specific domain of this important transcription factor.

DISCUSSION
TBP has a large number of biochemically defined activities: specific DNA binding (Davison et al., 1983;Nakajima et al., 1988); interactions with general transcription factors (eg. TFIIA and TFIIB) Ranish and Hahn, 1991;Roy et al., 1991;Safer et al., 1991;Usuda et al., 1991); interactions with TBP-associated factors (Pugh and Tjian, 1990;Dynlacht et al., 1991;Tanese et al., 1991;Timmers and Sharp, 1991); direct interactions with upstream activators (Stringer et al., 1990;Ingles et al., 1991;Lee et al., 1991;  The dash indicates that no TBP was added to these reactions. Pol I1 transcription was directed from the CYCl promoter in plasmid pGAL4CG-while pol 111 transcription was directed from the tRNAbu3 promoter in pYLEU3 or the 5 S promoter in pUC5S. Lieberman and Berk, 1991); interactions with coactivators/ mediators (Berger et al., 1990;Kelleher et al., 1990;Pugh and Tjian, 1990;Flanagan et al., 1991; interactions with negative regulators (Meisterernst and Roeder, 1991;Inostroza et al., 1992); interactions with RNA polymerase I1 (Koleske et al., 1992;Usheva et al., 1992); and interactions with RNA polym-erase I-and RNA polymerase 111-specific transcription components (Lobo et al., 1991;Margottin et al., 1991;Simmen et al., 1991;Pugh and Tjian, 1992;Comai et al., 1992). Conceivably, any mutation within TBP that compromised or disrupted any of these interactions could result in a conditional growth phenotype. We have used a combined genetic and biochemical approach to analyze the structure-function relationships of S. cereuisiae TBP. The availability of a readily assayable phenotype (conditional lethal growth) served as a valuable tool in this study. Using our mutagenesis protocol, we were able to isolate a large number of conditional lethal mutations within the TBP region targeted. The functional importance of this exact region of TBP had not previously been observed. Our initial goal was to use yeast genetics in a screen to identify novel functional domains within TBP which were specifically required for RNA polymerase I1 transcription. The data presented here indicate that we have indeed identified some mutations that specifically affect pol I1 transcription and others that cause more generalized functional defects. We believe that the pol 11-specific defect lies within a failure of the TBP 188 mutants to promote the formation of a pol I1 initiation complex. Results from the analyses of mRNAs that were isolated from the mutant strains used in this study revealed no alteration in start site selection of a number of pol 11-transcribed genes.* Also we should point out that the yeast and HeLa in uitro transcription systems may exhibit some subtle functional differences; our recombinant 188 mutants are able to complement a heat-treated HeLa nuclear extract for basal pol I1 transcription (data not shown). This result is in sharp contrast to the results we obtained from the yeast in uitro complementation assays presented in this report.
TBP and Basal Transcription-It seems incredible that such a small protein (27 kDa) can apparently make so many specific protein-protein interactions, regardless of whether these interactions are transient or stable. The association of TFIIA and TFIIB with TBP or more precisely, TFIID, facilitates or is required for the subsequent recruitment of RNA polymerase I1 to the promoter to form a complete preinitiation complex (Fire et al., 1984;Reinberg and Roeder, 1987;Buratowski et al., 1989). However, the exact mechanism(s) of initiation complex formation is not yet fully understood. For instance, TFIID may or may not require TFIIA for stable association with the TATA box (Reinberg and Roeder, 1987;Van Dyke et al., 1988;Buratowski et al., 1989;Maldonado et al., 1990;Cortes et al., 1992;Killeen et al., 1992). TFIIA may function either early or late in the course of initiation complex formation. In fact, the requirement for certain basal transcription factors may be gene-specific (Parvin et al., 1992).
In the time period since the cloning of the gene encoding TBP, certain of its functional domains have been identified the TBP domain involved in interacting with TFIIA has been mapped to the basic region of TBP, amino acid residues 133-156 (Buratowski and Zhou, 1992), and SPT3 has been shown to contact TBP residues clustered at position 174 (Eisenmann et al., 1992). Since our TBP mutants at position 188 are defective in pol I1 transcription, it is conceivable that this position is required for making specific protein-protein contacts with other general transcription factors. Possible obvious candidates include TFIIA, TFIIB, SRB2, and the carboxyl-terminal domain of pol 11. We are continuing to take a combined genetic and biochemical approach to elucidate what the exact molecular defect is in our TBP 188 mutants.