MOLECULAR AND CELLULAR BIOLOGY, May 1992, p. 2372-23820270-7306/92/052372-11$02.00/0Copyright © 1992, American Society for Microbiology

Biochemical and Genetic Characterization of a Yeast TFIID MutantThat Alters Transcription In Vivo and DNA Binding In VitroKAREN M. ARNDT, STEPHANIE L. RICUPERO, DAVID M. EISENMANN, AND FRED WINSTON*

Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115

Received 14 January 1992/Accepted 26 February 1992

A mutation in the gene that encodes Saccharomyces cerevisiae TFIID (SPT15), which was isolated in a

selection for mutations that alter transcription in vivo, changes a single amino acid in a highly conserved regionof the second direct repeat in TFIID. Among eight independent sptl5 mutations, seven cause this same aminoacid change, Leu-205 to Phe. The mutant TFIID protein (L205F) binds with greater affinity than that ofwild-type TFIID to at least two nonconsensus TATA sites in vitro, showing that the mutant protein has alteredDNA binding specificity. Site-directed mutations that change Leu-205 to five different amino acids cause fivedifferent phenotypes, demonstrating the importance of this amino acid in vivo. Virtually identical phenotypeswere observed when the same amino acid changes were made at the analogous position, Leu-114, in the firstrepeat of TFIID. Analysis of these mutations and additional mutations in the most conserved regions of therepeats, in conjunction with our DNA binding results, suggests that these regions of the repeats play equivalentroles in TFIID function, possibly in TATA box recognition.

The eukaryotic general transcription factor TFIID plays a

central role in transcription initiation. Binding of TFIID tothe TATA box of RNA polymerase II-dependent promotersis the first step in the assembly of a complex that containsRNA polymerase II and at least four other general transcrip-tion factors (TFIIA, TFIIB, TFIIE, and TFIIF) (5, 6, 60).This initiation complex is sufficient to support basal levels oftranscription in vitro (11, 32, 44, 45, 49, 50). In addition,previous studies have shown that TFIID is important invivo; in yeast cells, it is essential for growth and for normaltranscription (7, 13).TFIID has been studied in vitro in some detail, and it has

been shown to be capable of several interactions: binding toDNA, interactions with other general factors, and interac-tions with transcription regulatory factors (for a review, see

reference 42). Since it is the first step in the assembly ofthe general initiation complex, binding of TFIID to DNAis thought to be a likely target for the regulation of tran-scription initiation. In support of this idea, several studieshave suggested that TFIID interacts with a number ofspecific transcription activator proteins (21, 22, 25, 30, 51,56). In addition, another class of transcription regulatoryproteins, variously termed coactivators, mediators, or adap-tors, may be required to relay a signal from some specifictranscription activators to the general initiation complex viaan interaction with TFIID (3, 28, 41). Finally, other resultssuggest that TFIID may prevent the formation of inactivechromatin over a promoter (69). These in vitro studiesindicate that activators may directly facilitate binding ofTFIID to the promoter to overcome repression by nucleo-somes (68, 70, 71).Genes that encode TFIID have been isolated from a

number of species, and a comparison of the predicted aminoacid sequences shows a very high degree of conservation (80to 90% identity) in the carboxy-terminal 180 amino acids ofthese proteins (for a review, see reference 42). Within thiscarboxy-terminal region there are two imperfect repeats ofapproximately 60 amino acids that are 30% identical (7, 18,

* Corresponding author.

34, 57). Deletion analysis of yeast TFIID has demonstratedthat the carboxy-terminal region is essential for growth invivo (40) and transcription and DNA binding in vitro (24, 43).Analysis of dominant negative mutations that change aminoacids in the direct repeats of TFIID and that abolish DNAbinding suggested that the two repeats form a bipartite DNAbinding domain (43). In contrast to the carboxy termini, theamino-terminal regions of TFIID proteins from differentspecies show no striking similarity. For human and Droso-phila melanogaster TFIID, the amino-terminal regions havebeen implicated as targets for the action of regulatoryproteins (39, 41). In yeast strains, the requirement for thisregion in normal transcription remains unclear; however,strains that encode only the carboxy-terminal region ofTFIID are viable (10, 14, 40, 43, 72).We have previously reported the isolation of mutations in

the Saccharomyces cerevisiae SPTJ5 gene, which encodesTFIID (13). These mutations were identified as suppressors

of insertion mutations in the 5' regions of the HIS4 and LYS2genes caused by the yeast retrotransposon Ty or its solo longterminal repeat derivative (5). These insertion mutationsinhibit adjacent gene expression by abolishing or otherwisealtering transcription (for a review, see reference 4). Previ-ous work showed that mutations in several SPT (suppressorof Ty) genes, including at least one mutation in SPT15,suppress these insertion mutations by restoring functionaltranscription (for reviews, see references 4 and 62).To begin to correlate the transcriptional changes observed

in sptl5 mutants in vivo with the biochemical defects of themutant TFIID gene products in vitro, we have used bio-chemical and genetic approaches to study the defects causedby one particular sptl5 mutation. Here we report that thissptl5 mutation, sptl5-122, changes an amino acid in a highlyconserved region of the second repeat in TFIID. This aminoacid change leads to altered DNA binding in vitro and alteredtranscription initiation in vivo. In addition, analysis of mu-

tations causing other related amino acid changes stronglysuggests that the highly conserved region of both repeatsparticipates in the same aspect of a critical TFIID function,possibly DNA recognition.

2372

Vol. 12, No. 5

YEAST TFIID MUTANTS 2373

MATERIALS AND METHODS

Plasmids. Plasmids were constructed and isolated fromEscherichia coli strains by using standard methods (2). DNAtransformations into E. coli HB101 were performed asdescribed previously (2). For use in DNA sequencing, plas-mids were isolated by the alkaline lysis method. Isolationwas followed by equilibrium sedimentation in CsCl gradi-ents. Restriction enzymes and T4 DNA ligase were pur-chased from New England BioLabs (Beverly, Mass.) andBoehringer Mannheim Biochemicals (Indianapolis, Ind.).One plasmid (pDE38-9) used for the gapped rescue of

sptlS mutant alleles contains a 6.6-kb ClaI-SacI restrictionfragment from the plasmid pFW213 (13) subcloned into theClaI-SacI sites of pRS316 (CEN6 ARSH4 URA3) (54). Asecond plasmid (pDE59-1) used for gapped rescue containsthe same ClaI-SacI fragment cloned into pRS314 (CEN6ARSH4 TRPI) (54). The gapped rescue of the sptl5-122 gene(described later) from strain FW969 with plasmid pDE38-9yielded the plasmid p969-1. The 6.6-kb ClaI-SacI restrictionfragment from p969-1 was subcloned into the ClaI-SacI sitesof the yeast-integrating plasmid pRS306 (54) to create theplasmid pKA13.

Plasmids for the expression of the wild-type and mutantSPT1S genes in E. coli were derivatives of the T7 RNApolymerase expression vector, pET3b (47). The 2.4-kbEcoRI-BamHI fragment from pDE32-1 (13) containingSPT15 and the 2.4-kb EcoRI-BamHI fragment from p969-1containing sptl5-122 (L205F) were subcloned into the poly-linker of M13mpl9. Oligonucleotide (oligo)-directed muta-genesis (29) (Bio-Rad Laboratories In Vitro MutagenesisKit) was used to introduce an NdeI restriction site at theATG start codon of each gene. The 1,143-bp NdeI-BamHIrestriction fragment from each gene was then subcloned intothe NdeI-BamHI sites of pET3b to generate plasmids pKA9(SPT15) or pKA10 (sptl5-122).To simplify the construction of plasmids for analysis of the

TFIID direct repeats, a Sall restriction site was inserted 24bp 5' to the ATG start codon of SPT15 by oligo-directedmutagenesis. Plasmids containing this Sall site and theSPT15 gene within the 2.4-kb EcoRI-BamHI fragment fromthe SPT15 locus fully complement sptl5 mutations. A 2.4-kbEcoRI-BamHI fragment containing the SPT15 gene and theupstream SalI site was subcloned into the EcoRI-BamHIsites of plasmid pFW4 (a YIp5 derivative with the SalI sitedestroyed) (38, 63) to create pKA12. To introduce specificmutations into the first repeat of SPT15, the 536-bp Sall-XbaI fragment was subcloned into M13mpl8 and subjectedto oligo-directed mutagenesis. To introduce specific muta-tions into the second repeat of SPT1S, the 628-bp XbaI-BamHI fragment of SPT15 was subcloned into M13mpl8 andsubjected to oligo-directed mutagenesis. The presence of thedesired mutation, as well as the absence of any undesiredchanges, was confirmed by DNA sequence analysis. Muta-genized SalI-XbaI (first repeat) or XbaI-BamHI (secondrepeat) restriction fragments were then subcloned intopKA12 to replace the wild-type SPT15 sequence, generatingyeast-integrating plasmids containing the sptl5 mutationsthat encode the amino acid changes listed in Table 2.Two different null alleles of SPT15 were used in this study.

Plasmid pDE51-5 contains the spt1SAlOl::LEU2 allele andwas constructed by deleting the 161-bp XbaI-HindIII frag-ment of SPT15 in pDE25-2 (13) and replacing it with a 2.2-kbXbaI-HindIII fragment containing the S. cerevisiae LEU2gene. Plasmid pKA23 contains the sptlSA102::LEU2 alleleand was constructed by deleting the 778-bp Spel-Hindlll

fragment of pKA12 and inserting a 2.2-kb XbaI-HindIIILEU2 fragment. The deletion-insertion in pKA23 removesall but the last 15 codons of the SPT1S gene.Gap repair and sequencing. For the gapped rescue (37) of

sptl5 mutations, plasmid pDE38-9 was digested with EcoRIor plasmid pDE59-1 was digested with EcoRI and BamHI todelete an -2.4-kb fragment containing the SPTI5 openreading frame. Vector fragments containing SPT15 flankingsequences were transformed into sptlS mutant strains:pDE38-9 for strains FW969 and FW971 and pDE59-1 forstrains FY174, FY177, FY184, FY200, and FY259. PlasmidDNA was recovered from Ura+ or Trp+ transformants andtransformed into E. coli for propagation (19). The isolation ofmutant sptlS genes was confirmed by transformation of therecovered plasmids into an sptl5 strain (FW1259 or FY255).For each sptl5 mutation, both strands of the sptl5 openreading frame were sequenced by using synthetic primers.

Strains, media, and Northern (RNA) hybridization analysis.Rich (YPD), minimal (SD), synthetic complete (SC), andsporulation media were prepared as described previously byRose et al. (46). Presporulation medium (GNA) was pre-pared as described previously by Swanson et al. (59). Thesuppression of 8 insertion mutations at HIS4 and LYS2 andother nutrient auxotrophies were scored on SD media sup-plemented with the appropriate amino acids or SC medialacking the appropriate amino acid. Yeast transformantswere selected on SC media lacking the appropriate aminoacid, and ura3 strains were selected on SC media containing5-fluoro-orotic acid (5-FOA) prepared as described previ-ously (46).The S. cerevisiae strains used in this study are derived

from S288C (A Tot gal2) and are listed in Table 1. Strainswere constructed by using standard genetic methods formating, sporulation, and tetrad analysis (46) or by one- ortwo-step gene replacements (48, 52). Yeast strains weretransformed by the lithium acetate procedure (26).The sptlS-122 mutation was recombined into strains

FY167 and FY546 by two-step gene replacements withplasmid pKA13. His' recombinants (FY474 and FY475)were verified as correct by genetic crosses with SPT15strains, by complementation with a CEN ARS plasmidcontaining the SPT1S gene (pDE73-21), and by Southernhybridization analysis. Diploid strains FY548 and FY418contain null alleles of the SPT1S gene. The 3.3-kb SpeI-BamHI restriction fragment from plasmid pDE51-5 whichcontains the sptlSAlOl::LEU2 allele was recombined intostrain FY547 by transformation and selection for Leu+transformants. The sptl5A1O2::LEU2 allele was recombinedinto strain FY547 by transformation with the 2.9-kb SnaBI-BamHI restriction fragment from pKA23 and selection forLeu+ transformants. Replacement of one of the wild-typeSPT1S alleles in FY547 with sptlSAlOl::LEU2 or sptlSA102::LEU2 was confirmed by tetrad analysis (2:2 for viabil-ity) and by Southern hybridization analysis.

Analysis of site-directed mutations in the direct repeats ofTFIID was performed as follows. Yeast-integrating plasmids(marked by URA3) containing the mutations were linearizedwith the restriction enzyme SpeI and transformed into strainFY548 (plasmids containing SPT15, L114F, G119V,G21OV, L205A, L205I, or L205K) or linearized with SnaBIand transformed into strain FY418 (plasmids containingSPT15+, L114F, L114A, L114I, L114K, G119P, G210P,K127A, K218A, K127R, K218R, or L205D). The plasmidswere digested with SpeI or SnaBI to target integration toSPT15. Strains in which the plasmid had integrated adjacentto the sptlSA allele were identified after sporulation and

VOL. 12, 1992

2374 ARNDT ET AL.

TABLE 1. Yeast strains

Strain Genotypea

FW969 MATa sptlS-122 his4-9178 lys2-173R2 ura3-52FW971 ML4Ta sptl5-135 his4-9178 lys2-173R2 ura3-52FW1259 MATa sptlS-21 his4-9178 lys2-173R2 trpl Al ura3-52FY167 MATa his4-9178 lys2-173R2 leu2A&l ura3-52 trplAlFY174 MATa sptlS-6 his4-9178 trplA63 leu2AlFY177 MA4Ta sptl5-8 his4-9178 trplA63 leu2AJFY184 MATa sptlS-24 his4-9178 trplA63 leu2AJFY200 MATa sptlS-3 his4-9178 lys2-173R2 ura3-52 trplA63FY255 MA4Ta sptl5-21 his4-9178 lys2-173R2 leu2AO ura3-52

trplAlFY259 ALTa sptlS-7 his4-9178 lys2-173R2 leu2AJ ura3-52

trplA63FY384 ALTa lys2-173R2 leu2Al ura3-52 trplAlFY418 A4Ta/MATa sptlSAlO2::LEU2ISPTJS his4-9178/his4-

9178 lys2-173R2/lys2-17R2 trplAl/trplAl leu2Al/leu2Al ura3-52Iura3-52

FY474 AMTa sptlS-122 his4-9178 lys2-173R2 leu2Al ura3-52trplAl

FY475 MATa sptlS-122 his4-9128 lys2-173R2 trplA63 ura3-52

FY499 AL4Ta sptl5-305::sptlSAl0l his4-9178 lys2-173R2Ieu2Al ura3-52 trplAl

FY503 MATa sptl5-301 his4-9178 lys2-173R2 leu2Al ura3-52trplAl

FY504 MATa sptlS-303 his4-9178 lys2-173R2 leu2Al ura3-52trplAl

FY505 AL4Ta sptlS-304 his4-9178 lys2-173R2 leu2Al ura3-52trplAl

FY508 MATa sptl5-122 lys2-173R2 ura3-52FY532 ALTa sptlS-311::sptlSAlO2 his4-9178 lys2-173R2

leu2Al ura3-52 trplAlFY540 MATa sptl5-309 his4-9178 lys2-173R2 leu2Al ura3-52

trplAlFY541 MATa sptlS-310 his4-9178 lys2-173R2 leu2Al ura3-52

trplAlFY546 MATa his4-9128 lys2-173R2 trplA63 ura3-52FY547 AL4Ta/MA Ta his4-9178/his4-9178 Iys2-173R21lys2-

173R2 trplAl/trplAl leu2Al1/leu2Al ura3-52/ura3-52FY548 AL4Ta/MATa sptlSAl01::LEU2/SPT1S his4-9178/his4-

9178 Iys2-173R2/lys2-173R2 trplAll/trplAl leu2AJ/leu2Al ura3-52/ura3-52

a The amino acid substitutions in TFIID encoded by the various sptl5alleles are listed in Materials and Methods.

tetrad analysis of several transformants. These strains gaverise to haploid segregants in which the Ura+ and Leu+phenotypes cosegregated.For each site-directed mutation, an original diploid inte-

grant or a haploid segregant (for nonlethal mutations) inwhich the plasmid recombined next to the sptl5 null locuswas analyzed in two ways, Southern hybridization analysisand DNA sequencing of the mutation after recloning fromthe genome, to verify the presence of the desired mutation(64). We found it necessary to confirm the presence of thesite-directed mutation because occasionally different trans-formants with sptlS mutant plasmids unexpectedly gave riseto haploid segregants with a wild-type phenotype. Sequenceanalysis of plasmids recovered from such wild-type segre-gants revealed that these strains had in fact lost the desiredmutation, presumably by recombination or gene conversionwith the SPT1S copy in the diploid or with the adjacentsptlSAlOl::LEU2 locus.

For those mutations that did not cause lethality, thephenotype conferred by the mutation was also determined inthe absence of the adjacent sptlS null locus after selection on

5-FOA. The site-directed mutations that encode amino acidchanges in TFIID are as follows: sptl5-301 for L114F,sptl5-302 for G119V, sptl5-303 for L205A, sptl5-304 forL205I, sptl5-305 for L205K, sptl5-306 for G21OV, sptlS-309for L114A, sptl5-310 for L1141, sptlS-311 for L114K, sptl5-312 for G119P, sptl5-313 for K127A, sptlS-314 for K127R,sptl5-315 for L205D, sptl5-316 for G210P, sptlS-317 forK218A, and sptl5-318 for K218R.Northern hybridization analysis was performed as de-

scribed previously by Swanson et al. (59).Expression and purification of TFIID. E. coli BL21(DE3)/

pLysS (58) was transformed with plasmids pKA9 andpKA10. Transformed strains were grown at 30°C in LB brothcontaining 25 ,ug of ampicillin per ml and 30 ,ug of chloram-phenicol per ml to an optical density at 600 nm of 0.4. IPTG(isopropyl-,-D-thiogalactopyranoside) was added to a finalconcentration of 0.4 mM, and incubation at 30°C was con-tinued for 90 min. Proteins were purified by using the methodof Schmidt et al. (53) as modified by Workman et al. (71)through a heparin-Sepharose CL-2B column (Pharmacia).The buffer used throughout the purification contained 20 mMHEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonicacid)-KOH (pH 7.9), 20% (vol/vol) glycerol, 1 mM EDTA, 1mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluorideand the appropriate concentration of KCl. The lysis bufferand the final dialysis buffer contained 0.1 M KCl. Extractswere loaded on DEAE-cellulose (Whatman DE52) and hep-arin-Sepharose columns in purification buffer containing 0.1M KCI. TFIID was eluted from the heparin-Sepharosecolumn as described previously by Workman et al. (71).Aliquots of the final preparations were quick-frozen in liquidnitrogen, stored at -70°C, and used immediately after thaw-ing. Samples were not refrozen, since this treatment signif-icantly reduced the DNA binding activity of the TFIID-L205F protein. The estimated purities of the proteins were50% for wild-type TFIID and 20% for TFIID-L205F. Therelative amounts of TFIID protein in the final preparationswere determined by quantitative analysis of Coomassie-stained sodium dodecyl sulfate-polyacrylamide gels. Weestimate the relative error in our measurements to be lessthan 20%.DNase I protection. An end-labelled DNA probe encoding

the adenovirus major late TATA box was prepared fromplasmid pRW (5). Plasmid DNA was restricted with EcoRIand HinclI. The top strand of the fragment was radiolabelledby filling in the EcoRI site with Kienow enzyme (BoehringerMannheim) and [a-32P]dATP (Amersham). An end-labelledprobe containing the S. cerevisiae his4-9128 promoter region(-387 to +97, relative to the HIS4 transcription start site)was prepared from plasmid pKAL. Plasmid pKA1 containsthe 1.9-kb Sall restriction fragment of YCp701 (17) insertedat the SalI site of pUC118 (61). To radiolabel the top strandof the 484-bp his4-9128 fragment, pKA1 was digested withSpeI, treated with calf intestinal alkaline phosphatase(Boehringer Mannheim), and then phosphorylated with T4polynucleotide kinase (Boehringer Mannheim) and [y_32p]ATP (New England Nuclear). The phosphorylated plasmidwas then digested with HaeIII. Probes were purified byelectrophoresis on 5% polyacrylamide gels and electro-eluted. The concentration of each probe was determined byan ethidium bromide spot test with DNA standards of knownamounts.DNase I protection assays were performed essentially as

described previously by Buratowski et al. (5) by using -2 ngof adenovirus major late promoter probe or 4 to 8 ng ofhis4-9128 probe in a 50-,ul reaction mixture. DNase I diges-

MOL. CELL. BIOL.

YEAST TFIID MUTANTS 2375

68

1111 : : : :1IZ ID R&VI

II :1 1 11158

spt

FIG. 1. Amino acid change caused by spt15-1two direct repeats of S. cerevisiae TFIID are sharrow indicates the L205F substitution causedmutation. Asterisks mark amino acids that weredirected mutations in SPT15, as listed in Tableindicate amino acid identities, and two dots indichanges as previously defined (18).

tions were stopped by the addition of an estop mix (27) to his4-9128 reaction mixtuivolume of a modified stop mix (contains 250salmon sperm DNA per ml instead of tRNA;adenovirus major late promoter reaction mireaction mixtures were extracted once with <of phenol-chloroform-isoamyl alcohol (25:2itated with ethanol. Dried pellets were suspformamide loading buffer-0.1 N NaOH (2:13 min at 90°C. Protection reactions withmajor late probe and those with the his4-9analyzed by electrophoresis on 8% polyaciurea sequencing gels and 6% polyacrylamsequencing gels, respectively. Gels were drieto autoradiography.

RESULTSAn amino acid change in a highly conserv

region of TFIID affects transcription in vivo.pendent mutations in the SPT15 gene, whcerevisiae TFIID, have been isolated by sepressors of the 8 insertion mutation his4-917of these mutations cause very similar mutanvivo; strains that contain these mutations stthe insertion mutations his4-9178 and lys2defects in mating and sporulation, and grow30°C, and many are moderately temperatugrowth at 37°C. These phenotypes suggest Ision of many genes is altered in these muanalysis demonstrated that at least one s

sptlS-21, causes transcriptional changes in,%To determine whether the sptl5 mutatior

and whether they cause amino acid changes icarboxy-terminal region of TFIID, we cloneceight independent alleles (see Materials andmarkably, of the eight independent mutati4seven of them, typified by the mutation sptlidentical amino acid change, Leu-205 toreferred to as L205F). This amino acid chanthighly conserved region of the second direct(Fig. 1). Of the seven mutations causing thesix change codon 205 from TTA to lTlT achanges codon 205 from ITA to TTC. Alcarboxy-terminal domains of the TATA-bindmany species demonstrates that the leucine athe S. cerevisiae sequence has been conserexamined (see reference 42 for referencemutation that was sequenced, sptl5-21 (13), cacid change elsewhere in the second repedescribed in a separate report (12).To determine whether the L205F amino

* * *-t127 TFIID affects transcription in vivo, we examined transcrip-11 111 1:11 1 1 tion of two different insertion mutations, his4-9125 and

* * 218 his4-9178. Although these two insertion mutations differ inthe orientation and site of insertion of the 8 sequence in the

15-122 (L25F) HIS4 promoter region, they both confer a His- phenotype122 mutation. The by causing similar effects on transcription of the HIS4 gene.1own aligned. The In SPT15+ strains carrying either his4-9128 or his4-9176, a

by the sptd5-122 longer-than-normal HIS4 message is synthesized because ofchanged by site- transcription initiation within the 8 element instead of at the2. Vertical lines normal HIS4 initiation site (55, 65). This long HIS4 tran-

icate conservative script is nonfunctional, presumably because it encodestranslational start and stop codons 5' to the actual HIS4AUG. To test whether the sptlS-122 mutation alters tran-scription initiation at these loci in vivo, Northern hybridiza-

-qual volume of tion analysis was performed on RNA isolated from strainsres or an equal containing the wild-type HIS4 gene, his4-9125, or his4-9178I,g of sonicated (Fig. 2A). Although the sptlS-122 mutation has little or noas the carrier) to effect on transcription from the wild-type HIS4 promoter,xtures. Stopped for both his4-9128 and his4-9178 it causes a shift in HIS4an equal volume mRNA size from the longer message found in SPT15+4:1) and precip- strains to a message indistinguishable in size from wild-typeended in 6 ,u1 of HIS4 mRNA (diagrammed in Fig. 2B). Primer extensionand heated for analysis of mRNA from both of these strains demonstratedthe adenovirus that this shift in transcript size is due to a shift in the site of'128 probe were transcription initiation from the 8 to the normal HIS4 startrylamide-8.3 M site (1). These results, in conjunction with the pleiotropicvide-8.3 M urea nature of sptlS-122, indicate that the L205F amino acid-d and subjected change in TFIID affects transcription initiation of many

genes in vivo.The TFIID-L205F mutant protein has altered DNA binding

properties. To identify the function of TFIID that is alteredby the L205F amino acid change, we have characterized the

ed and repeated wild-type (TFIID) and mutant (TFIID-L205F) proteins inFourteen inde- vitro. To do this, TFIID and TFIID-L205F were expressed

Lich encodes S. in E. coli and partially purified as described in Materials andlection for sup- Methods. The abilities of the wild-type and mutant TFIID% (63, 66). Most proteins to support basal-level transcription in vitro wereIt phenotypes in assayed in a HeLa cell nuclear extract depleted of endoge-rongly suppress nous TFIID activity by mild heat treatment (36). The wild-'-1 73R2, exhibit type TFIID protein is approximately twofold more active forv very slowly at transcription from the adenovirus major late promoter thanIre sensitive for the TFIID-L205F protein in this system (1). This differencethat the expres- in transcription activity may reflect altered binding by theitants. Previous mutant TFIID protein to the adenovirus major late TATAsptJ5 mutation, box, as described in a later section.vivo (13). The DNA binding properties of TFIID and TFIID-L205Fns are clustered were analyzed by DNase I protection experiments for twoin the conserved different cases, the his4-9128 5' region, which contains both1 and sequenced the 8 and HIS4 promoters, and the well-characterized ade-Methods). Re- novirus major late promoter. For the his4-9128 promoter

ons sequenced, region, titration experiments with increasing amounts of!5-122, cause an TFIID and TFIID-L205F revealed that four sites are pro-Phe (hereafter tected from DNase I cleavage but that the wild-type and

ge is in the most mutant proteins differ with respect to their relative affinitiesrepeat of TFIID for the four sites. Sites I and IV, corresponding to the HIS4L205F change, and 8 TATA boxes, respectively, are bound with compara-Lnd the seventh ble affinity by both wild-type TFIID and TFIID-L205Flignment of the mutant protein (Fig. 3A) (1). However, the TFIID-L205Fling factors from protein binds to two additional sites in the 8 (sites II and III)It position 205 in with two- to fourfold greater affinity than wild-type TFIIDved in all cases (Fig. 3A). The position of each site of protection within thes). The eighth his4-9128 5' region is diagrammed in Fig. 3B. Binding ofcauses an amino TFIID-L205F to site II also generates three hypersensitive-at and will be sites at positions -112, -120, and --134 (relative to the

HIS4 transcription start site) that are less pronounced or notacid change in evident when wild-type TFIID binds to this site. Additional

VOL. 12, 1992

2376 ARNDT ET AL.

A.

HIS-4 hi4-jt2 bS9-I97&0Q 0 CD

> LL > U. > 1UX) Ln 6 on 6 LO

C- N CNa: -j a. ajR -j

HIS4 -> a 1g

TUB2 -- V _ _ _ _

B.

his4-9 123

TATAUAS _ TATA HIS4

SPT15 +

spt 15- 122

his4-91 73

TATAUAS TATA HIS4

SPT15 +

sptl5- 122

FIG. 2. Transcriptional effects of sptl5-122 mutation in vivo.The sptl5-122 mutation, which encodes the L205F substitution inTFIID, alters transcription initiation at two 8 insertion mutations atHIS4. (A) Northern hybridization analysis was performed on totalRNA from SPT15 (wild-type) and sptlS-122 (L205F) strains contain-ing the wild-type HIS4 gene, his4-912&, or his4-9178. Strains FY384,FY508, FY546, FY475, FY167, and FY474 were used. Lanes 1 and2 contain 2 ,ug of total RNA, while lanes 3 to 6 contain 10 ,ug of totalRNA. The filter in the upper panel was probed for HIS4 mRNA. Theposition of the wild-type-length HIS4 message is indicated by an

arrow. Faintly hybridizing, lower-molecular-weight transcripts are

due to rRNA. RNA levels were normalized by probing for TUB2mRNA (lower panel) and by measuring rRNA levels by ethidiumbromide staining. (B) Schematic diagrams of the insertion mutationshis4-9128 and his4-9178 and the shift in the transcription start siteobserved in sptl5-122 mutants. The 8 elements are boxed; filledarrows indicate the direction of transcription of the normal Tymessage. The diagram is not drawn to scale.

minor differences in protection between sites II and III arealso visible. These observations may indicate that the twoproteins assume different conformations when bound tothese sites. Increased binding of the mutant TFIID protein tosites II and III in the his4-9128 5' region suggests that theL205F amino acid change in TFIID leads to altered orrelaxed specificity for TATA sites, since these sites containsequences that resemble TATA boxes but that deviate fromthe consensus TATAAA sequence by at least 1 bp (Fig. 3C).The increased binding by TFIID-L205F to sites II and III

has been reproducibly observed by using three differentpreparations of TFIID and TFIID-L205F purified from bac-teria by three different procedures (1). Furthermore, thealtered binding by TFIID-L205F cannot be attributed to anycontaminating proteins in the final preparation, since ex-tracts prepared from E. coli strains that contain the pET3bexpression vector (47) lacking the SPT15 gene do not bind tothe his4-9128 fragment, nor do these extracts affect thebinding of TFIID (1).

In addition to the altered DNA binding specificity ob-served with the his4-9128 promoter fragment, the TFIID-L205F protein exhibits some qualitative difference in bindingto the adenovirus major late TATA box. At this promoter,DNase I protection by wild-type TFIID extends from - 19 to-35 relative to the start site of transcription, while protec-tion by TFIID-L205F extends from -22 to weak protectionat -39 (Fig. 4). Thus, relative to binding by wild-typeTFIID, binding by TFIID-L205F appears to be shiftedapproximately 3 bp away from the site of transcriptioninitiation. In addition, the mutant TFIID has reduced affinityfor the adenovirus major late TATA box. Although TFIID-L205F has decreased affinity for this TATA box, it hasnearly wild-type affinity for the 8 TATA box. Since both ofthese sites have the consensus TATA sequence, TATAAA,this difference is likely to be due to flanking sequences.The direct repeats of TFIID function similarly in vivo. Since

the L205F amino acid change affects transcription and DNAbinding and since Leu-205 is in a very highly conservedregion of TFIID, we made additional changes within thisregion of each repeat and assayed their phenotypic andtranscriptional effects in vivo. In particular, we used oligo-directed mutagenesis to introduce specific amino acidchanges in TFIID at position 205, as well as at positions 210and 218 in the second repeat and at the equivalent positionsin the first repeat (amino acids 114, 119, and 127). Thesemutations were integrated into the genome, and their effectswere analyzed in haploid strains in the absence of any otherfunctional SPT15 gene (see Materials and Methods).For the first set of mutations, we changed Leu-205 to four

other amino acids with different size and/or charge proper-ties: Ala, Ile, Lys, and Asp. Strains containing these foursubstitutions at Leu-205 have four different phenotypes;L205A is Spt-, although the phenotype is somewhat weakerthan that conferred by L205F; L205I is wild type; L205K isSpt+ but extremely temperature sensitive for growth; andL205D is inviable (Fig. 5 and Table 2). Including the initialL205F change, then, five distinct phenotypes are conferredby five different amino acid substitutions at Leu-205; thisresult strongly suggests that some important aspect of TFIIDfunction is sensitive to changes at this position.

Since most changes which we examined at position 205allowed viability but caused distinct mutant phenotypes,these findings provided us with a sensitive assay for testingthe functional similarity of the direct repeats. When theanalogous amino acid in the first repeat, Leu-114, waschanged to Phe, Ala, Ile, or Lys, phenotypes virtually

MOL. CELL. BIOL.

YEAST TFIID MUTANTS 2377

_IwSOE

-I-P_ j~-m d

U-~~~~~~O

-- immi;mm

m mm 00 m mmt M- -So-as goa dw 40 aw

W -..4-.

a ____4A :* _

2 3 4 5 6 7 8 9 10 11 12 13 14 15

B.

IV III 'I

91l26 1- H4 _

- 50 bp

1. HIS4 TATA

5' CTGTGTATATAATAGATA3'3' GACACATATATTATCTAT5'

identical to those caused by the same substitutions at Leu-205 were observed (Fig. 5 and Table 2). For example, L114Fcauses a strong Spt- phenotype, slow growth, and moderatetemperature sensitivity for growth. However, some slightdifferences in phenotypes were observed for analogouschanges in the two repeats. In one case, strains containingthe L114F substitution have a slightly weaker Spt- pheno-type and grow more rapidly than strains containing theL205F amino acid change. The weaker Spt- phenotypecaused by L114F provides the best explanation for whymutations that encode this amino acid change were notisolated in our mutant selections. In addition, strains thatcontain the L114K substitution are unable to grow onminimal media and grow more slowly on rich media thanstrains that contain the L205K substitution.To test further the functional equivalence of the two

repeats, site-directed mutations at two other positions withineach repeat were made and analyzed. These changes weredesigned to confer different structural or charge properties tothese regions of the protein. Glycine residues at positions119 and 210 were changed to either valine or proline. Both

IV changes, whether made in the first or the second directrepeat, cause inviability under all growth conditions tested(Table 2). Substitution of the carboxy-terminal lysine of eachrepeat (positions 127 and 218) with either arginine or alanineresults in a wild-type phenotype (Table 2). Thus, for everyamino acid change examined, virtually identical phenotypeswere observed when the substitutions were made at analo-gous positions in the first or second repeat of TFIID. Thesefindings strongly suggest that the highly conserved regions ofthe two direct repeats of TFIID have nearly equivalent butnot redundant functions in vivo.

In addition to causing similar Spt and growth phenotypes,the same amino acid substitutions at Leu-114 and Leu-205cause similar transcriptional effects in vivo. Northern hy-bridization analysis was performed on total RNA fromstrains containing the his4-9178 allele and site-directed mu-tations in SPTJ5 (Fig. 6). A strong correlation between theSpt phenotype and the transcriptional effect at his4-9178 wasobserved. Strains that suppress his4-9178 (the Phe and Alasubstitutions) have a significant amount of wild-type-lengthHIS4 message, while strains that are phenotypically wildtype (the Ile substitutions) have only 8-initiated transcripts.

11. nonconsensus TATA

5' AATATTATAGCCTTTATCCAACAATGAAT3'3' TTATAATATCGGAAATAGTTGTTACCTTA5'

l1. nonconsensus TATA5' CCCTTTTATGGATTCCTA3'3' GGGAAAATACCTAAGGAT5'

IV. 8 TATA

5' AATGAATATAAACATATAAAATGATGATAATAATATTTATAGAATTGTGTA3'3'TTACTTATATTTGTATATTTTACTACTATTATTATAAATATCTTAACACAT5'

FIG. 3. DNase I protection analysis of TFIID and TFIID-L205Fbinding at the his4-9128 5' region. (A) DNase I protection analysiswas performed with increasing amounts of partially purified wild-type TFIID and TFIID-L205F mutant proteins. TFIID storagebuffer (lanes 4 and 10), TFIID-L205F mutant protein (lanes 5 to 9),or wild-type TFIID protein (lanes 11 to 15) was incubated with aprobe from the his4-9128 5' region prior to treatment with DNase Ias described in Materials and Methods. Reaction mixtures containedthe following amounts of TFIID-L205F or TFIID protein: lanes 5and 11, 2.5 ng; lanes 6 and 12, 5 ng; lanes 7 and 13, 10 ng; lanes 8 and14, 20 ng; and lanes 9 and 15, 40 ng. Lanes 1 to 3 contain productsof sequencing reactions (33). Brackets denote regions of protection

from DNase I cleavage. The asterisk indicates an additional regionof protection 3' to the HIS4 TATA protection that is seen only withwild-type TFIID and that likely reflects binding of a second mole-cule. Dots denote hypersensitive sites. The vertical line withinregion IV marks the position of the 8 TATA sequence. The largeregion of protection over the 8 TATA box is most likely due to thebinding of multiple TFIID molecules in this region. (B) Schematicdiagram of the his4-9128 5' region. For each site of protection listedin panel C, the position and extent of protection are indicated bybars. Relative to the HIS4 transcription start site at + 1, thefollowing sequences are protected: site I, -50 to -67; site II, -135to -163; site III, -202 to -219; and site IV, -232 to -282. Arrowsmark the HIS4 and 8 transcription start sites at + 1 and -194,respectively. The diagram is drawn to scale. (C) Sequences pro-tected from DNase I cleavage by binding of TFIID or TFIID-L205F,as indicated in panel A. For site I, the HIS4 TATA box that has beenshown to be functional in vivo (35) is in boldface. For site IV, the 8TATA region, there are two TATA consensus sequences (shown inboldface) within the region shown to be functional (31). The under-lined sequence in site IV is an additional consensus TATA box inthis region. For sites II and III, sequences that deviate from theTATAAA consensus sequence at one position are in boldface.

A.A

G G

Sp

C.

0 TFIID-L205F 10 TFIIDc c0 aT r_ c

soI* qw

VOL. 12, 1992

1

2378 ARNDT ET AL.

TF11D-L205FTFIID

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

FIG. 4. DNase I protection analysis of TFIID and TFIID-L205F

binding to the adenovirus major late TATA box. DNase I protection

analysis was performed with increasing amounts of partially purifiedTFIID and TFIID-L205F proteins. TFIID storage buffer (lanes 4 and

11), wild-type TFIID protein (lanes 5 to 10), or TFIID-L205F mutant

protein (lanes 12 to 17) was incubated with an end-labelled fragment

containing the adenovirus 2 major late TATA box prior to DNase I

cleavage as described in Materials and Methods. Reaction mixtures

contained the following amounts of TFIID or TFIID-L205F: lanes 5

and 12, 5 ng; lanes 6 and 13, 10 ng; lanes 7 and 14, 20 ng; lanes 8 and

15, 40 ng; lanes 9 and 16, 80 ng; and lanes 10 and 17, 160 ng. Lanes

1 to 3 contain the products of sequencing reactions. Brackets denote

the extents of protection by the two proteins.

The lysine changes at positions 114 and 205 result in a more

complex pattern of HIS4 transcripts, most likely due to

nonspecific transcription initiation or to increased levels of

RNA degradation. The relatively minor differences observed

either in different RNA preparations from the same strain

(compare Fig. 2, lane 6, and Fig. 6, lane 3) or between strains

are due to unknown factors that cause a small degree of

variability between experiments. The relative amounts of

TFIID protein in strains analyzed in Fig. 5 and 6 were

determined by Western blotting (immunoblotting) and were

found to be equivalent (1). Therefore, the phenotypic and

transcriptional effects resulting from amino acid changes at

positions 114 and 205 are due to altered functions, and not

altered levels, of the mutant TFIID proteins.

DISCUSSION

Previous work has shown that selection for suppressors ofTy long terminal repeat (b) insertion mutations at the HIS4

and LYS2 genes results in the isolation of mutations in

several genes, including the SPT15 gene, which encodes

yeast TFIID (13). We have determined that at least seven

independent sptl5 mutations isolated in this way cause the

same single amino acid change (L205F) in a highly conserved

region of the second direct repeat of TFIID. This amino acid

change leads to altered transcription initiation at tinsertionmutations in vivo. In addition, the pleiotropic effects of this

mutation indicate that transcription of many other genes isaffected. In vitro studies comparing the wild-type and mu-

tant (L205F) TFIID proteins demonstrate that the mutant

protein has altered DNA binding and suggests that the

transcriptional alterations observed in the sptl5 mutant mayarise from altered TATA box recognition in vivo. The

suggestion that TFIID-L205F is a gain-of-function mutant

with altered DNA recognition is also supported by our

observation that sptl5-122 is partially dominant for causingaltered transcription in vivo (1).The isolation of a mutation in the most conserved region of

a direct repeat of SPT15 has allowed us to address questionsconcerning the functional similarity of these regions of theTFIID repeats. Different amino acid substitutions at position205 and nearby in the second TFIID repeat cause a broadspectrum of mutant phenotypes that cannot simply be due toloss of function. This same spectrum is conferred by theanalogous amino acid changes in the first TFIID repeat.Therefore, these findings provide strong evidence that thehighly conserved regions of the two repeats play equivalentroles in a common function of the protein.

In addition, the results from our mutational studies of theTFIID repeats argue against the importance of the previ-ously reported weak homology between TFIID and regions2.3 and 2.4 of bacterial sigma factors (20, 23). A regionanalogous to the reported sigma factor homology region(amino acids 181 to 211 in the second TFIID repeat) is notpresent in the first TFIID repeat because of sequencedifferences between the repeats. However, amino acid sub-stitutions at positions 205 and 210 of TFIID and, in the firstrepeat, at positions 114 and 119 strongly suggest that theseregions perform very similar functions. Moreover, theL205K substitution, which improves the similarity betweenTFIID and virtually all known sigma factors (16), severelyimpairs TFIID function in vivo.Our in vitro studies with purified wild-type and mutant

(L205F) TFIID proteins suggest that the direct repeats ofTFIID play a role in TATA box recognition and binding.Relative to binding by wild-type TFIID, TFIID-L205F ex-hibits both qualitative and quantitative differences in DNAbinding. At the adenovirus major late TATA box, the mutantTFIID protein has a pattern of DNase I protection that isaltered compared with that of wild-type TFIID, implying apossible conformation difference between the two proteins.The pattern of binding of the TFIID-L205F mutant protein tothe his4-9126 region (increased affinity for nonconsensusTATA boxes at sites II and III) (Fig. 3) suggests that themutant protein has a relaxed specificity for TATA sites,since it binds with comparable affinity to a number ofTATA-like sequences, including two that are only weaklyrecognized by wild-type TFIID. Previous experiments haveshown that wild-type TFIID has little or no function at theseparticular TATA box variants (8, 67). Additional experi-ments will be needed in order to determine the preferredsequence requirements, if any, of TFIID-L205F.These results extend previous structure-function studies

that demonstrated that certain dominant-negative mutationsand small internal deletions within the TFIID repeats com-pletely eliminate DNA binding (24, 43). Some of the aminoacid changes that abolish DNA binding (43) are just 2positions away from the L205F change which we havestudied in this work. In contrast to these earlier studies thatfocused primarily on loss-of-function mutations in SPT15,the genetic selection that yielded the sptl5-122 mutation,which encodes the L205F change, required viability andaltered gene expression and, therefore, resulted in the iso-lation of a mutation that altered but did not eliminate DNAbinding. Since loss of function could be caused by a varietyof nonspecific effects, such as the disruption of proteinstructure or improper localization, the gain-of-functionchange, L205F, provides more compelling evidence that thisregion of the protein is important either directly or indirectlyfor DNA recognition. Furthermore, our mutant analysissuggests that, in addition to this region of the second TFIID

MOL. CELL. BIOL.

YEAST TFIID MUTANTS 2379

A. complete^ . . ... ...;_

mn;;

C. YPD 300C

I __~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-

B. -histidine

III._

D. YPD 370C

FIG. 5. Mutant phenotypes of strains containing amino acid substitutions at positions 205 and 114. Strains encoding wild-type or differentmutant TFIID proteins were grown in liquid cultures and spotted onto plates. For each strain, the leftmost spot in the series of four containsan aliquot of a saturated culture (4 x 108 cells per ml) and the next three spots contain aliquots of 10-fold serial dilutions of the starting culture.Cells were transferred from liquid cultures with a multipronged frogger device. The following strains were analyzed: FY167 (wild type),FY474 (L205F), FY503 (L114F), FY504 (L205A), FY540 (L114A), FY505 (L205I), FY541 (L1141), FY499 (L205K), and FY532 (L114K). Acomparison of panels A and B demonstrates the ability of the strains to suppress the his4-9178 insertion mutation (the Spt- phenotype). Inpanels A and B, the plates were incubated at 30°C. A comparison of panels C and D demonstrates the temperature sensitivity of the strainson YPD media. For all plates, growth after 3 days of incubation is presented.

repeat, the same region of the first repeat is also importantfor DNA recognition. We attempted to address this issue byexamining the DNA binding properties of TFIID-L114Fpurified from E. coli. However, this protein is extremelyunstable in vitro, and we have been unable to obtain activeprotein for DNA binding studies.Assuming that TFIID-L205F also has altered DNA bind-

ing properties in vivo, and on the basis of the many types ofinteractions that TFIID makes with other molecules, severalmodels can be proposed to explain the transcriptional effectsof this amino acid change in vivo. Furthermore, alteredbinding at different promoters could cause transcriptionalchanges by different mechanisms. For one case studied,his4-9125, the 5' region contains two promoters, the 8promoter and the HIS4 promoter (Fig. 2B and 3B). Previousanalysis demonstrated that transcription initiates primarilyat the 8 initiation site in SPTJ5+ his4-9128 strains. Additionalstudies of cis-acting mutations at or near the 8 and HIS4

TATA boxes that caused a shift in initiation to the wild-typeHIS4 site suggested that the choice of initiation sites wasdecided by a competition between the two promoters (17).Other experiments suggested that chromatin structure mayplay a role in mediating this competition (9). In our presentwork, we have shown that in sptlS-122 his4-9128 strains, thewild-type-length HIS4 transcript is the predominant mes-sage. Therefore, in some way, TFIID-L205F alters thecompetition between the 8 and HIS4 TATA regions, result-ing in a change in transcription initiation.The simplest model for how a mutant TFIID could alter

transcription initiation at his4-9128 is one in which theTFIID-L205F mutant protein preferentially recognizes theHIS4 TATA box over the 8 TATA box. However, this modelis not strongly supported by our data. In vitro bindingstudies have shown that wild-type TFIID and TFIID-L205Fhave approximately equal affinities for both TATA regions,indicating that altered transcription at his4-9128 is not

wild-typeL205F LII4FL205A L 114AL2051 L1141L205K Li 14K

VOL. 12, 1992

2380 ARNDT ET AL.

TABLE 2. Phenotypes conferred by site-directedmutations in SPT15

Phenotype for:Aminoacid Growth Growth

change Spt on 30°C on 37°C GrowthYPD YPD onS

Wild type + + + +L114F - +/- +/-- +1-L205F - -/+ --l+ -l+L114A -/+ + + +L205A -/+ + + +L1141 + + + +L2051 + + + +L114K + --/+ - -L205K + --/+ - --/+K127A + + + +K218A + + + +K127R + + + +K218R + + + +L205D NA - - -G119V NA - - -G210V NA - - -G119P NA - - -G210P NA - - -

a The Spt phenotype indicates the degree of suppression of the insertionmutations his4-9178 and lys2-173R2. Spt-, strong suppression of 8 insertionmutations. Minimal media (SD) contained all necessary nutrient supplementsto compensate for the known auxotrophies of each strain. +, wild-typegrowth; +/-, +/--, -/+, and --/+ intermediate levels of growth; -, nogrowth; NA, not applicable.

caused by preferred binding of wild-type or mutant TFIID toone of these TATA sites over the other.

In considering this model, we cannot rule out two possi-bilities not answered by our in vitro binding assays. First,the mutant TFIID, when bound to the HIS4 TATA, mayhave an altered conformation that increases its transcrip-tional activity. However, we have seen no evidence of this inseveral DNase I protection experiments in which the HIS4TATA region was well resolved (1). Second, in vivo, theTFIID-L205F protein may actually have increased affinityfor all TATA and near-TATA sequences. This possibility issupported by the observation that TFIID-L205F is unstablein a variety of storage conditions with respect to binding atall TATA sequences examined (see Materials and Methods);therefore, it seems likely that our preparations of TFIID-L205F contain fewer active molecules than our preparationsof wild-type TFIID and that the observed differences inbinding between mutant and wild-type proteins may beunderestimated.

In alternative models, TFIID-L205F may alter the compe-tition between the 8 and HIS4 TATA sites at his4-9128 bybinding to the nonconsensus TATA site II, site III, or both.This binding could allow the formation of functional tran-scription initiation complexes at these sites that woulddirectly contribute to transcription from the HIS4 initiationsite. However, the distance (150 bp) between the HIS4initiation site and the closest of these TATA-like sequences,site II, is somewhat greater than that normally found in S.cerevisiae (15, 35). Alternatively, binding of TFIID-L205F tothese nonconsensus TATA sites could increase initiation atHIS4 indirectly, by altering chromatin structure throughoutthe region. In support of the latter model, in vitro studieshave shown that TFIID can compete with nucleosomes forDNA binding (69), and in vivo studies have shown thatmutations in the genes that encode histones H2A and H2B

his4-91 760)

UL LL. 4 e 4'A-6 Ln -t In v in v n

04 Y- 04 N~ r- N rz-i -A -j -i -j -i -j ~~~~-j

HIS4 -- on! m so Jmw

wI

*.i- ",

TUB2 - qwwq-pa-wqpqPp

FIG. 6. Transcriptional effects of site-directed mutations inSPT15. Northern hybridization analysis was performed on totalRNA from strains containing the his4-917b insertion mutation andthe indicated amino acid change in TFIID. The same strains thatwere analyzed in Fig. 5 were used in this experiment. Lanes 2 to 10contain 10 ,g of RNA. Lane 1 is a control lane that contains 2 ,ug ofRNA from an SPTJ5 HIS4 strain (FY384) to indicate the position ofwild-type HIS4 mRNA. The filter in the upper panel was probed forHIS4 mRNA. RNA levels were normalized by probing for TUB2mRNA (lower panel) and by measuring rRNA by ethidium bromidestaining.

(HTAI and HTB1) cause transcriptional effects at his4-9128virtually identical to those observed in sptl5-122 mutants (9).

In summary, we have shown that a particular amino acidchange in a highly conserved region of TFIID causes alteredDNA binding. Our results also strongly suggest that thisregion in each of the two direct repeats of TFIID playsimportant and equivalent roles in TFIID function, possiblyin TATA box recognition. Distinguishing among possiblemechanisms of suppression of his4-9128 by sptl5-122 shouldincrease our understanding of the factors that govern TFIIDfunction in vivo. In addition, isolation of both extragenic andintragenic suppressors of sptl5-122 and related mutations,coupled with further biochemical analysis of TFIID mutants,should lead to a greater understanding of the componentsthat regulate the binding specificity of TFIID in vivo.

ACKNOWLEDGMENTSWe are extremely grateful to Ian Taylor and Bob Kingston for

advice on TFIID purification, to Rick Wobbe for assistance in thepreparation of HeLa cell nuclear extracts, to Steve Buratowski andMartin Schmidt for anti-TFIID antisera, and to Mark Fleming forsynthesizing the oligonucleotides used in this study. We also thankBob Kingston and Rick Wobbe for many helpful discussions andGreg Prelich for critical reading of the manuscript.

This work was supported by grants from the National Institutes ofHealth (GM32967 and GM45720) and from the Stroh BrewingCompany, all to F.W., and by a postdoctoral fellowship from theHelen Hay Whitney Foundation to K.M.A.

ADDENDUM IN PROOF

Strubin and Struhl (Cell 68:721-732) have recently de-scribed a TFIID double mutant with altered DNA-bindingproperties. The mutant that they identified contains the

MOL. CELL. BIOL.

YEAST TFIID MUTANTS 2381

amino acid changes 1194F and L205V, both of which arenecessary for the observed altered specificity.

REFERENCES1. Arndt, K. M., and F. Winston. Unpublished data.2. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G.

Seidman, J. A. Smith, and K. Struhl. 1988. Current protocols inmolecular biology. Greene Publishing Associates and WileyInterscience, New York.

3. Berger, S. L., W. D. Cress, A. Cress, S. J. Triezenberg, and L.Guarente. 1990. Selective inhibition of activated but not basaltranscription by the acidic activation domain of VP16: evidencefor transcriptional adaptors. Cell 61:1199-1208.

4. Boeke, J. D., and S. B. Sandmeyer. 1991. Yeast transposableelements, p. 193-261. In J. R. Broach, J. R. Pringle, and E. W.Jones (ed.), The molecular and cellular biology of the yeastSaccharomyces, vol. 1: genome dynamics, protein synthesis,and energetics. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, N.Y.

5. Buratowski, S., S. Hahn, L. Guarente, and P. A. Sharp. 1989.Five intermediate complexes in transcription initiation by RNApolymerase II. Cell 56:549-561.

6. Buratowski, S., M. Sopta, J. Greenblatt, and P. A. Sharp. 1991.RNA polymerase II-associated proteins are required for a DNAconformation change in the transcription initiation complex.Proc. Natl. Acad. Sci. USA 88:7509-7513.

7. Cavallini, B., I. Faus, H. Matthes, J. M. Chipoulet, B. Winsor,J. M. Egly, and P. Chambon. 1989. Cloning of the gene encodingthe yeast protein BTF1Y, which can substitute for the humanTATA box-binding factor. Proc. Natl. Acad. Sci. USA 86:9803-9807.

8. Chen, W., and K. Struhl. 1988. Saturation mutagenesis of ayeast his3 "TATA element": genetic evidence for a specificTATA-binding protein. Proc. Natl. Acad. Sci. USA 85:2691-2695.

9. Clark-Adams, C. D., D. Norris, M. A. Osley, J. S. Fassler, andF. Winston. 1988. Changes in histone gene dosage alter tran-scription in yeast. Genes Dev. 2:150-159.

10. Cormack, B. P., M. Strubin, A. S. Ponticelli, and K. Struhl.1991. Functional differences between yeast and human TFIIDare localized to the highly conserved region. Cell 65:341-348.

11. Dignam, J. D., P. L. Martin, B. S. Shastry, and R. G. Roeder.1983. Eukaryotic gene transcription with purified components.Methods Enzymol. 104:582-598.

12. Eisenmann, D. M., K. M. Arndt, S. L. Ricupero, J. W. Rooney,and F. Winston. Submitted for publication.

13. Eisenmann, D. M., C. Dollard, and F. Winston. 1989. SPTIS, thegene encoding the yeast TATA binding factor TFIID, is re-quired for normal transcription initiation in vivo. Cell 58:1183-1191.

14. Gill, G., and R. Tjian. 1991. A highly conserved domain ofTFIID displays species specificity in vivo. Cell 65:333-340.

15. Hahn, S., E. T. Hoar, and L. Guarente. 1985. Each of three"TATA elements" specifies a subset of the transcription initi-ation sites at the CYC-1 promoter of Saccharomyces cerevisiae.Proc. Natl. Acad. Sci. USA 82:8562-8566.

16. Helmann, J. D., and M. J. Chamberlin. 1988. Structure andfunction of bacterial sigma factors. Annu. Rev. Biochem. 57:839-872.

17. Hirschman, J. E., K. J. Durbin, and F. Winston. 1988. Geneticevidence for promoter competition in Saccharomyces cerevi-siae. Mol. Cell. Biol. 8:4608-4615.

18. HoeUmakers, J. H. J. 1990. Cryptic initiation sequence re-vealed. Nature (London) 343:417-418.

19. Hoffman, C. S., and F. Winston. 1987. A ten-minute DNApreparation from yeast efficiently releases autonomous plasmidsfor transformation of Escherichia coli. Gene 57:267-272.

20. Hoffmann, A., E. Sinn, T. Yamamoto, J. Wang, A. Roy, M.Horikoshi, and R. G. Roeder. 1990. Highly conserved coredomain and unique N terminus with presumptive regulatorymotifs in a human TATA factor (TFIID). Nature (London)346:387-390.

21. Horikoshi, M., M. F. Carey, H. Kakidani, and R. G. Roeder.

1988. Mechanism of action of a yeast activator: direct effect ofGAL4 derivatives on mammalian TFIID-promoter interactions.Cell 54:665-669.

22. Horikoshi, M., T. Hai, Y.-S. Lin, M. R. Green, and R. G.Roeder. 1988. Transcription factor ATF interacts with theTATA factor to facilitate establishment of a preinitiation com-plex. Cell 54:1033-1042.

23. Horikoshi, M., C. K. Wang, H. Fujii, J. A. Cromlish, P. A. Weil,and R G. Roeder. 1989. Cloning and structure of a yeast geneencoding a general transcription initiation factor TFIID thatbinds to the TATA box. Nature (London) 341:299-303.

24. Horikoshi, M., T. Yamamoto, Y. Ohkuma, P. A. Weil, and R. G.Roeder. 1990. Analysis of structure-function relationships ofyeast TATA box binding factor TFIID. Cell 61:1171-1178.

25. Horikoshi, N., K. Maguire, A. Kralli, E. Maldonado, D. Rein-berg, and R. Weinmann. 1991. Direct interaction betweenadenovirus EIA protein and the TATA box binding transcrip-tion factor IID. Proc. Natl. Acad. Sci. USA 88:5124-5128.

26. Ito, H., Y. Fukuda, K. Murata, and A. Kimura. 1983. Transfor-mation of intact yeast cells treated with alkali cations. J.Bacteriol. 153:163-168.

27. Jones, K. A., K R. Yamamoto, and R. Tjian. 1985. Two distincttranscription factors bind to the HSV thymidine kinase pro-moter in vitro. Cell 42:559-572.

28. Kelleher, R. J., III, P. M. Flanagan, and R. D. Kornberg. 1990.A novel mediator between activator proteins and the RNApolymerase II transcription apparatus. Cell 61:1209-1215.

29. Kunkel, T. A., J. D. Roberts, and R. A. Zabour. 1987. Rapid andefficient site-specific mutagenesis without phenotypic selection.Methods Enzymol. 154:367-382.

30. Lee, W. S., C. C. Kao, G. 0. Bryant, X. Liu, and A. J. Berk.1991. Adenovirus EIA activation domain binds the basic repeatin the TATA box transcription factor. Cell 67:365-376.

31. Liao, X., J. J. Clare, and P. J. Farabaugh. 1987. The upstreamactivation site of a Ty2 element of yeast is necessary but notsufficient to promote maximal transcription of the element.Proc. Natl. Acad. Sci. USA 84:8520-8524.

32. Matsui, T., J. Segall, P. A. Weil, and R. G. Roeder. 1980.Multiple factors required for accurate initiation of transcriptionby purified RNA polymerase II. J. Biol. Chem. 255:11992-11996.

33. Maxam, A. M., and W. Gilbert. 1980. Sequencing end-labeledDNA with base-specific chemical cleavages. Methods Enzymol.65:499-560.

34. Nagai, K. 1990. Cryptic initiation sequence revealed. Nature(London) 343:418.

35. Nagawa, F., and G. R. FinL 1985. The relationship between the"TATA" sequence and transcription initiation sites at the HIS4gene of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA82:8557-8561.

36. Nakajima, N., M. Horikoshi, and R. G. Roeder. 1988. Factorsinvolved in specific transcription by mammalian RNA polymer-ase II: purification, genetic specificity, and TATA box-promoterinteractions of TFIID. Mol. Cell. Biol. 8:4028-4040.

37. Orr-Weaver, T., J. W. Szostak, and R. Rothstein. 1983. Geneticapplications of yeast transformation with linear and gappedplasmids. Methods Enzymol. 101:228-245.

38. Parent, S. A., C. M. Fenimore, and K. A. Bostian. 1985. Vectorsystems for the expression, analysis and cloning of DNAsequences in S. cerevisiae. Yeast 1:83-138.

39. Peterson, M. G., N. Tanese, B. F. Pugh, and R. Tjian. 1990.Functional domains and upstream activation properties ofcloned human TATA binding protein. Science 248:1625-1630.

40. Poon, D., S. Schroeder, C. K. Wang, T. Yamamoto, M. Horiko-shi, R. G. Roeder, and P. A. Weil. 1991. The conservedcarboxy-terminal domain of Saccharomyces cerevisiae TFIID issufficient to support normal cell growth. Mol. Cell. Biol. 11:4809-4821.

41. Pugh, B. F., and R. Tjian. 1990. Mechanism of transcriptionalactivation by Spl: evidence for coactivators. Cell 61:1187-1197.

42. Pugh, B. F., and R. Tjian. 1992. Diverse transcriptional func-tions of the multisubunit eukaryotic TFIID complex. J. Biol.Chem. 267:679-682.

VOL. 12, 1992

2382 ARNDT ET AL.

43. Reddy, P., and S. Hahn. 1991. Dominant negative mutations inyeast TFIID define a bipartite DNA-binding region. Cell 65:349-357.

44. Reinberg, D., M. Horikoshi, and R. G. Roeder. 1987. Factorsinvolved in specific transcription in mammalian RNA polymer-ase II: functional analysis of initiation factors IIA and IID andidentification of a new factor operating at sequences down-stream of the initiation site. J. Biol. Chem. 262:3322-3330.

45. Reinberg, D., and R. G. Roeder. 1987. Factors involved inspecific transcription by mammalian RNA polymerase II: puri-fication and functional analysis of initiation factors IIB and IIE.J. Biol. Cherti. 262:3310-3321.

46. Rose, M. D., F. Winston, and P. Hieter. 1990. Methods in yeastgenetics: a laboratory course manual. Cold Spring HarborLaboratory Press, Cold Spring Harbor, N.Y.

47. Rosenberg, A. H., B. N. Lade, D. Chui, S.-W. Lin, J. J. Dunn,and F. W. Studier. 1987. Vectors for selective expression ofcloned DNAs by T7 RNA polymerase. Gene 56:125-135.

48. Rothstein, R. J. 1983. One-step gene disruption in yeast. Meth-ods Enzymol. 101:202-211.

49. Samuels, M., A. Fire, and P. A. Sharp. 1982. Purification andcharacterization of a specific RNA polymerase II transcriptionfactor. J. Biol. Chem. 261:2003-2013.

50. Sawadogo, M., and R. G. Roeder. 1985. Factors involved inspecific transcription by human RNA polymerase II: analysis bya rapid and quantitative in vitro assay. Proc. Natl. Acad. Sci.USA 82:4394-4398.

51. Sawadogo, M., and R. G. Roeder. 1985. Interaction of a gene-specific transcription factor with the adenovirus major latepromoter upstream of the TATA box region. Cell 43:165-175.

52. Scherer, S., and R. W. Davis. 1979. Replacement of chromo-some segments with altered DNA sequences constructed invitro. Proc. Natl. Acad. Sci. USA 76:4951-4955.

53. Schmidt, M. C., C. C. Kao, R. Pei, and A. J. Berk. 1989. YeastTATA-box transcription factor gene. Proc. Natl. Acad. Sci.USA 86:7785-7789.

54. Sikorski, R. S., and P. Hieter. 1989. A system of shuttle vectorsand yeast host strains designed for efficient manipulation ofDNA in Saccharomyces cerevisiae. Genetics 122:19-27.

55. Silverman, S. J., and G. R. Fink. 1984. Effects of Ty insertionson HIS4 transcription in Saccharomyces cerevisiae. Mol. Cell.Biol. 4:1246-1251.

56. Stringer, K. F., C. J. Ingles, and J. Greenblatt. 1990. Direct andselective binding of an acidic transcriptional activation domainto the TATA-box factor TFIID. Nature (London) 345:783-786.

57. Stucka, R., and H. Feldmann. 1990. An element of symmetry inyeast TATA-box binding protein transcription factor IID. FEBSLett. 261:223-225.

58. Studier, F. W., and B. A. Moffatt. 1986. Use of bacteriophage T7

RNA polymerase to direct selective high-level expression ofcloned genes. J. Mol. Biol. 189:113-130.

59. Swanson, M. S., E. A. Malone, and F. Winston. 1991. SPT5, anessential gene important for normal transcription in Saccharo-myces cerevisiae, encodes an acidic nuclear protein with acarboxy-terminal repeat. Mol. Cell. Biol. 11:3009-3019.

60. Van Dyke, M. W., R. G. Roeder, and M. Sawadogo. 1988.Physical analysis of transcription preinitiation complex assem-bly on a class II gene promoter. Science 241:1335-1338.

61. Vieira, J., and J. Messing. 1987. Production of single-strandedplasmid DNA. Methods Enzymol. 153:3-11.

62. Winston, F. Analysis of SPT genes: a genetic approach towardsanalysis of TFIID, histones and other transcription factors ofyeast. In S. McKnight and K. R. Yamamoto (ed.), Transcrip-tional regulation, in press. Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, N.Y.

63. Winston, F. Unpublished data.64. Winston, F., F. Chumley, and G. R. Fink. 1983. Eviction and

transplacement of mutant genes in yeast. Methods Enzymol.101:211-228.

65. Winston, F., K. J. Durbin, and G. R. Fink. 1984. The SPT3 geneis required for normal transcription of Ty elements in S.cerevisiae. Cell 39:675-682.

66. Winston, F., C. Dollard, E. A. Malone, J. Clare, J. G. Kapakos,P. Farabaugh, and P. L. Minehart. 1987. Three genes arerequired for trans-activation of Ty transcription in yeast. Ge-netics 115:649-656.

67. Wobbe, C. R., and K. Struhl. 1990. Yeast and human TATA-binding proteins have nearly identical DNA sequence require-ments for transcription in vitro. Mol. Cell. Biol. 10:3859-3867.

68. Workman, J. L., S. M. Abmayr, W. A. Cromlish, and R. G.Roeder. 1988. Transcriptional regulation by the immediate earlyprotein of pseudorabies virus during in vitro nucleosome assem-bly. Cell 55:211-219.

69. Workman, J. L., and R. G. Roeder. 1987. Binding of transcrip-tion factor TFIID to the major late promoter during in vitronucleosome assembly potentiates subsequent initiation by RNApolymerase II. Cell 51:613-622.

70. Workman, J. L., R. G. Roeder, and R. E. Kingston. 1990. Anupstream transcription factor, USF (MLTF), facilitates theformation of preinitiation complexes during in vitro chromatinassembly. EMBO J. 9:1299-1308.

71. Workman, J. L., I. C. A. Taylor, and R. E. Kingston. 1991.Activation domains of stably bound GAL4 derivatives alleviaterepression of promoters by nucleosomes. Cell 64:533-544.

72. Zhou, Q., M. C. Schmidt, and A. J. Berk. 1991. Requirement foracidic amino acid residues immediately N-terminal to the con-served domain of Saccharomyces cerevisiae TFIID. EMBO J.10:1843-1852.

MOL. CELL. BIOL.