Proc. Natl. Acad. Sci. USAVol. 90, pp. 4027-4031, May 1993Biochemistry

The TFIIIB-assembling subunit of yeast transcription factor TFIIIChas both tetratricopeptide repeats and basic helix-loop-helix motifs

(TFC4/TATA-binding protein/tRNA)

CHRISTIAN MARCK*, OLIVIER LEFEBVRE, CHRISTOPHE CARLES, MICHEL RIVA, NATHALIE CHAUSSIVERT,ANNY RUET, AND ANDRE SENTENACService de Biochimie et de Gdndtique Moldculaire, D6partement de Biologie Cellulaire et Moldculaire, Centre d'Etudes de Saclay, F 91191 Gif-sur-YvetteCedex, France

Communicated by E. Peter Geiduschek, January 25, 1993 (received for review December 21, 1992)

ABSTRACT The multisubunit yeast transcription factorIIIC (TFIIIC; also called x) can undergo considerable confor-mational changes upon binding to the A and B blocks of tRNAgenes. After binding to DNA encoding tRNA (tDNA), TFIIICacts as an assembly factor to recruit an initiation factor,TFIIIB, via its x131 subunit. We have cloned the gene encodingthe x131 subunit and named it TFC4. This gene is unique,essential for cell viability, and encodes a Mr 120,153 protein.Epitope-tagging and mobility-shift assays indicated the pres-ence of a single x131 subunit in TFlIIC-tDNA complexes. 7131contains two sequence motifs, accounting for nearly one-half ofthe protein mass, that may provide a molecular explanation forthe properties of TFMIIC-tDNA complex. A series of 11 copiesof the tetratricopeptide repeat motif may account for theflexibility and interaction properties of TFIIIC. A motifakin tothe basic helix-oop-helix motif of MyoD suggests the directinvolvement of 7131 in promoting DNA binding of TFIIIB.

Transcription factor IIIC (TFIIIC) is a large protein complexthat is responsible for the first step of tRNA gene (tDNA)activation in eukaryotes. TFIIIC recognizes the two intra-genic promoter elements, the A and B blocks, with remark-able flexibility since it can accommodate variable spacingsand therefore variable relative helical orientations of thesetwo elements (1, 2). The polypeptide composition of TFIIIChas been investigated by several groups (3-5). At least fourpolypeptides of 138, 131, 95, and 60 kDa comigrate with theTFIIIC-tDNA complexes. The 138- and 95-kDa subunits canbe specifically crosslinked to tDNA by UV irradiation ofTFIIIC-tDNA complexes (3). Two additional polypeptidechains of 91 and 50 kDa, consistently found in affinity-purified factor, are also potential subunits of TFIIIC.

After binding to tDNA, or to the TFIIIA-5S gene complex,TFIIIC acts as an assembly factor to promote binding ofTFIIIB to an upstream position. TFIIIB by itself is unable todetectably bind DNA (6-9). Once bound, however, TFIIIBbecomes tightly associated with DNA and TFIIIB-tDNAcomplexes, artificially deprived of TFIIIC by heparin treat-ment, can direct multiple rounds of transcription by RNApolymerase III (7). Bartholomew et al. (5, 10) have succeededin identifying and mapping to specific DNA positions thecomponents of TFIIIB and TFIIIC that associate with theSaccharomyces cerevisiae SUP4 tRNATYr gene. The 138- and95-kDa subunits of TFIIIC, were mapped to the B and Ablocks, respectively, whereas two polypeptides (90 and 70kDa) from TFIIIB were crosslinked on opposite sides of theDNA helix in a 6-bp region located 35 bases upstream of theSUP4 tRNATYr gene transcriptional start site (10). TheTFIIIC 131-kDa subunit covered a very long stretch ofDNA,

since it could be crosslinked both upstream of the startsite-i.e., close to the 70-kDa polypeptide of TFIIIB-andalso in between the A and B blocks, together with the TFIIIC60-kDa subunit. Binding of TFIIIB enhanced crosslinking ofthe 131-kDa subunit, indicating that TFIIIB brings the 131-kDa subunit closer to upstream DNA and suggesting that thissubunit is the TFIIIB assembling subunit of TFIIIC (10).

Recently, it has become clear that TBP (TATA-bindingprotein), originally thought to be necessary only for class IIgene transcription, is an essential component of class I, II,and III transcription systems (11-15). Two lines of resultshave clarified the role of TBP in class III gene transcription.First, the gene encoding the 70-kDa subunit of TFIIIB, PCF4(16), TDS4 (17), or BRFI (18), has been cloned as anallele-specific high-copy suppressor of mutations in TBP,suggesting an association between these two proteins. Sec-ond, TBP has been shown to be required for the stableassembly of TFIIIB on tRNA genes (19, 20), to be part of theheparin-resistant initiation complex (17, 19, 20), and to berecruited by the 70-kDa subunit after its assembly on theTFIIIC-tDNA complex (19). Therefore, the role ofTFIIIC isto assemble the initiation complex TFIIIB (that comprises the90- and 70-kDa subunits and TBP)-TFIIIC-tDNA.The genes encoding the 95-kDa subunit (21, 22) and the

138-kDa subunit (23) of TFIIIC have been cloned and foundto be essential for cell viability. In this work, we report thecloning of TFC4, the gene encoding the 131-kDa subunit ofTFIIIC, which is the presumed TFIIIB assembling subunit.t

MATERIALS AND METHODSPurification of TFIIIC. Affinity purification of TFIIIC has

been described (21). Several preparations of the factor werepooled and loaded on a single well of an SDS/polyacrylamidegel. The polypeptide of 131 kDa was excised from the gel andpeptide sequences were obtained as described (23).

Cloning of TFC4. Four peptide sequences were obtainedfrom tryptic digestion: (i) (Gly)-(Ala)-Met-Tyr-Asn-Pro-Tyr-Gly-Ala-(Asn)-Ile-Leu-(Arg), (ii) (Gln)-Tyr-Asn-Ile-Pro-Ile-Asp-Ile-(Lys), (iii) (Ala)-(Ala)-Leu-Ala-(Trp)-Val-?-Tyr-(Arg), (iv) (Leu)-(Ala)-Glu-Gly-Asp-Ser-Val-Phe-Glu-Gly-Pro-Leu-Met-Thr-(Gly) (parentheses denote ambiguities).Six fully degenerate 20-bp oligonucleotides that correspondto peptides 1, 2, and 4 were designed and six parallel PCRexperiments were performed using a GeneAmp amplificationkit in a Perkin-Elmer DNA thermal cycler using 30 cycles (1min at 92°C, 1 min at 55°C, 1 min at 72°C extended for 5 sec

Abbreviations: TFIIIB, -C, transcription factors IIIB, -C, respec-tively; tDNA, DNA encoding tRNA; TBP, TATA-binding protein;bHLH, basic helix-loop-helix; TPR, tetratricopeptide repeat.*To whom reprint requests should be addressed.tThe sequence reported in this paper has been deposited in theGenBank data base (accession no. L12722).

4027

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 90 (1993)

at each cycle). Amplification of yeast genomic DNA usingoligonucleotides 5'-TAYAAYATHCCNATHGAYAT-3'(coding for Tyr-Gln-Ile-Pro-Ile-Asp-Ile; peptide 2, direct) and5'-GTCATNARNGGNCCYTCRAA-3' (coding for Phe-Glu-Gly-Pro-Leu-Met-Thr; peptide 4, reverse) yielded a singleamplified DNA band that revealed an uninterrupted readingframe (H is A, C, or T; N is A, C, G, or T; R is A or G; Yis C or T). The AEMBL3a yeast genomic DNA library wasthen screened by hybridization with this band.

Sequencing of TFC4. A 3354-bp BamHI and a 3159-bpEcoRI fragment from the same phage, both hybridizing to thePCR probe, were subcloned and sequenced. Double-strandDNA sequencing was performed with synthetic oligonucle-otides to walk along the DNA sequence. Both strands weretotally sequenced over the 4798-bp BamHI/EcoRI fragmentusing the Pharmacia T7 sequencing kit. All four trypticpeptides are found in this sequence, with only two errors.Data bank searches were performed using the CITI2 com-puting facilities (24).

Disruption of TFC4. A 4024-bp Acc I/BstBI fragmentcontaining TFC4 was inserted at the EcoRV site of plasmidBluescript SK yielding plasmid pCK14 (Fig. 1). A 2016-bp StyI/Afl II fragment of this plasmid was replaced by the 1778-bpBamHI/BamHI fragment of plasmid pSZ63 (25) carrying theyeast HIS3 selectable marker. In this construction, namedpCK15, 65% of the TFC4 gene coding sequence was deleted(amino acids 170-839). Yeast strain CMY214 (a/a trpl-Alltrpl-Al, his3A200/his3A200, ura3-52/ura3-52, ade2-101/ade2-101, lys2-801/lys2-801, canl/CANI) (26) was trans-formed with a 3850-bp Apa I/Sma I fragment from pCK15carrying the disrupted TFC4 gene. Genomic DNA from fourHIS' transformant colonies was prepared and digested withEcoRI and Eag I. A Southern blot (data not shown) using a521-bp PCR fragment (positions 3029-3549 in plasmidpCK14) as a probe confirmed that the TFC4 gene wasdisrupted in three of these colonies. One of the transform-ants, YCK104, was retained and sporulated. Twenty-threeasci were dissected and a 2/2 segregation of spore viabilitywas observed for all asci. All spores, except those from asingle ascus, were confirmed to be haploid cells.Tagging of 7131 and Purification of Wild-Type and Tagged

Proteins. The double-stranded oligonucleotide 5'-TAC-

Genomic

BE Ac E HpE HYEI 1 I1F 1^{~~pCKI4(wild)

pCK14 (wil(l)pCK16 (tagged)

K E EGHYE

CCATACGATGTTCCAGATTACGCT-3' encoding theTyr-Pro-Tyr-Asp-Val-Pro-Asp-Tyr-Ala epitope from the in-fluenza hemagglutinin protein (27) was inserted at the uniqueHpa I site of plasmid pCK14 (amino acid 25 of T131), givingplasmid pCK16. A Kpn I/Spe I fragment of either pCK14(bearing the wild-type TFC4 gene) or pCK16 (bearing thetagged copy) was inserted into the centromeric plasmidpUN75 (28), giving plasmids pCK17 and pCK18. Diploidyeast strain YCK104 was transformed with either plasmidand both strains were induced to sporulate. HIS+ sporescould germinate and grew normally, showing that thetfc4::HIS3 disruption could be complemented by either thewild-type copy (haploid strain YCK107) or the epitope-tagged copy (haploid strain YCK109) of TFC4 with similarefficiency. Small scale (30 g of cells) TFIIIC purification andgel-retardation assays were performed as described (3, 21).

RESULTS

Cloning and Sequence Analysis of TFC4. The genes encod-ing the 95-kDa subunit (21) and the 138-kDa subunit (23) ofTFIIIC have been cloned in our laboratory based on proteinmicrosequence data. We have followed the same strategy toisolate the gene encoding the 131-kDa polypeptide (seeMaterials and Methods). In accordance with the previouslyused notation (21, 23), we have called this gene TFC4 [it is thefourth transcription factor gene of class C (III) transcriptionsystem to be cloned] and named its product T131. A singleopen reading frame of 3075 bp encodes a protein of 1025amino acids (Fig. 2) with a molecular weight of 120,153, inreasonable agreement with the 131-kDa value estimated fromSDS/PAGE experiments (4). Sequencing downstream ofTFC4 has shown that it lies 2.5 kb upstream of RME1 (29).

Disruption of TFC4 and Tagging of 7131. A large centralfragment ofthe TFC4 coding region was replaced by the HIS3gene deleting 8 of the 11 TPR units (units 2-9) and the wholebasic helix-loop-helix (bHLH) region (see below). Southernblot experiments (data not shown) confirmed the disruptionof one allele in diploid transformants and also indicated thatTFC4 is a unique gene. Spores bearing the disrupted genewere able to germinate and grow for only four or fivegenerations. This phenotype is the same as that caused by

TFC4

TFC4

HB AH

ILIBs HpAE

111

HB AH

III

pCK15 HAp E E*GHY H H A GAH E Sm

FIG. 1. Schematic representation of a 4798-bp fragment of the yeast S. cerevisiae genome encompassing the TFC4 gene and of the insertsof plasmids pCK14, pCK15, and pCK16. Open bar to the left of the TFC4 gene indicates the start of an antiparallel open reading frame. Shadedarea is the TFC4 region amplified in the PCR experiment. Solid bars denote the 3354-bp BamHI and 3159-bp EcoRI fragments subcloned forsequencing. A 1325-bp Eag I/EcoRI band of the disrupted allele (dashed line) and the 3159-bp EcoRI band of the wild-type allele hybridizedwith a PCR probe (hatched box) in the Southern blotting experiment. A, Afl II; Ac, Acc I; Ap, Apa I; B, BamHI; Bs, BstBI; E, EcoRI; G, EagI; H, HindIll; Hp, Hpa I; K, Kpn I; Sm, Sma I; Sp, Spe I; Y, Sty I. For the sake of clarity, not all sites are indicated in pCK14, pCK15, andpCK16.

4028 Biochemistry: Marck et al.

E Sp

Imm

Proc. Natl. Acad. Sci. USA 90 (1993) 4029

MAAGKLKKEQ QNQSAERESA DTGKVNDEDE EHLYGNIDDY KHLIQDEEYD DEDVPHDLQL SEDEYNSERDGEISEDDEED FMNAIREASN FKVKKKKKND KGKSYGRQRK ERVLDPEVAQ LLSQANEAFV RNDLQVAERLNFAAYETLGD IYQLQGRLND CCNSWFLAAH LNASDWEFWK IVAILSADLD HVRQAIYCFS RVISLNPMEWKKTGQLARAL DGFQRLYMYN PYDANILREL AILYVDYDRI EDSIELYMKV FNANVERREA ILAALENALDEDADEKEPLE QDEDRQMFPD INWKKIDAKY KCIPFDWSSL NILAELFLKL AVSEVDGIKT IKKCARWIQRPDDSEFDNRR FKNSTFDSLL AAEKEKSXNLBVRLGL LRLNTDNLVE ALNHFQCLYD ETFSDVADLYEKYKEAIDFF TPLLSLEEWR TTDVFKPLAR CYKEIESYET AKEFYELAIK SEPDDLDIRV SLAEVYYRLNDVVEMRKHQV DETLHRISNE KSSNDTSDIS SKPLLEDSKF RTFRKKKRTP YDAERERIER ERRITAKVVDNSGLNEAKQA SIWINTVSEL VDIFSSVKNF FMKSRSRKFV GILRRTKKFN TELDFQIERL SKLAEGDSVFLTSATELRGL SYEQWFELFM ELSLVIAKYQ SVEDGLSVVE TAQEVNVFFQ DPERVKMMKF VKLAIVLQMDGLLNQFQFNR KVLQVFMYSL CRGPSSLNIL SSTIQQKFFL RQLKAFDSCR YNTEVNGQAS ITNKEVYNPNYAVLLYSSRG FLSALQYLTR LEEDIPDDPM VNLLMGLSHI HRAMQRLTAQ RHFQIFHGLR YLYRYHKIRKADYNLGRAFH LIGLVSIAIE YYNRVLENYD DGKLKKHAAY NSIIIYQQSG NVELADHLME KYLSI

SSLLAEFSDY 80FNEVIKKDAR 160ESIYRRSMLY 240SSDEESAAEG 320RESQTFWDHV 400FEAATALTRA 480DPETFKHMLV 560KYEKMKKFEL 640EGPLMEERVT 720DEEELAENLR 800KKSSPYLYYI 880SLYTDLEKQE 960

1025

* *4o -- -Acidic Basic TPR 1-5 Acidic TPR 6-9 bH L H TPR 10 TPR 11

H-rich

FIG. 2. (Upper) Amino acid sequence of 7131, the protein encoded by TFC4. Peptides obtained from microsequencing are underlined. (Lower)Schematic representation of T131. Arrow indicates the tagging site and asterisks indicate the two putative nuclear localization signals. Horizontalbar shows extent of protein sequence removed by tfc4::HIS3 gene disruption.

disruption of TFCJ (21) and TFC3 (23), which encode the 9s5and T138 subunits, respectively. To prove that TFC4 is a genecoding for one of the TFIIIC subunits, we inserted theinfluenza hemagglutinin epitope (27) at a unique Hpa I site atresidue 25 in the hydrophilic N-terminal region. After sporu-

lation, haploid cells harboring the disrupted genomic copy

and the tagged copy of TFC4 on a centromeric plasmid grewnormally. Affinity-purified T131-tagged TFIIIC was then pre-

pared (21) and TFIIIC-tDNA complexes were analyzed bygel retardation (Fig. 3). Addition of increasing amounts of12CA5 anti-epitope antibodies converted progressively theT131-tagged complex into a slower-migrating form, while thecomplex formed with wild-type TFIIIC remained unaffectedby the antibodies. The supershift induced by the binding ofthe 12CA5 antibody showed that T131 is part ofTFIIIC and thesingle bandshift observed (even with higher antibody con-

WILD TYPE TAG-TFC4I

0 40 80 500 0 40 80 320 IgG ng

-T

FIG. 3. TFC4 encodes a component of TFIIIC. TFIIIC was

purified from two haploid strains harboring a disrupted TFC4 geneand transformed with a centromeric plasmid expressing either wild-type TFIIIC or T31-tagged TFIIIC. TFIIIC-tDNA complexes wereformed using a 32P-labeled DNA fragment containing afinity-purified TFIIIC and a tRNAGlu gene for 10 min at 25'C and thenunderwent reaction with increasing amounts of the 12CA5 mono-

clonal antibody directed against the epitope tag for 1 h at 25'C.Complexes were analyzed by gel-retardation assay and revealed byautoradiography. T, Position of TFIIIC-tDNA complexes; r-IgG,position of TFIIIC-tDNA complexes bound by antibodies.

centrations; data not shown) suggested that only one mole-cule of 7131 is present in TFIIIC.Sequence Analysis of x131. The overall sequence of T131 is

rich in aromatic residues (11.0o) and contains 9 cysteines.Two acidic regions are found at positions 9-97 (37% acid) and310-334 (58% acid) as well as two basic regions at positions102-122 (57% basic) and 599-637 (44% basic) (Fig. 2). Bothbasic regions include a stretch of five amino acids in a row,at positions 104 (KKKKK) and 604 (RKKKR) that could bepart of nuclear localization signals. A short region (positions919-946), at the end of T131, includes 5 histidine residues thatcould be involved in zinc chelation.A search of the GenPro data bank (version 73) revealed a

slight similarity between the sequences of T131, bimA (30), andSSN6/CYC8 (31, 32) proteins. This prompted us to look forthe presence of tetratricopeptide repeats (TPRs) (33) in theT131 sequence. Through visual inspection and use of a com-puter procedure based on the knowledge of numerous TPRunits (C.M., unpublished data), we have observed that T131

contains 11 TPR motifs (Fig. 4). A first block is made of fiveperfectly repeated TPR units (units 1-5), a second blockcomprises 4 units (units 6-9), and two units (units 10 and 11)are isolated. Such features are commonly found in TPRproteins (33-35).

1 5 10 15 20 25 30aealfgLGhiYeklgdl.kAldaFqk.AllldPnn

1 vaqLlsqAneafvrndlqvAerlFneVikkdarn (128-2 faaYetLGdiYqlqgrindccnsWflAahlnasd3 wefwkiVAilsadldhvrqAiyclsrVislnPme4 wesiyrrSmlYkktgqlarAldgFqrlymynPyd5 aniLreLAilYvdydriedSielYmkVfnanver -297)6 idirvrLGllrlntdnlveAlnhFqclydetfsd v (4327 adlYfeaAtaLtraekykeAidfFtpllsleewr t8 tdvFkpLArcYkeiesyetAkefYeLAiksePd9 ldirvaL&evrYyrlndpet fkhmLvdVvemrkhq -569)

10 pylYyiyAvlLyssrgflsAlqyLtrleediPdd (875-908)11 qeadynLGraFhliglvsiAieyYnrVlenydd,g (959-992)

FIG. 4. (Upper) Consensus sequence of the TPR unit as definedby Sikorski et al. (33). (Lower) The 11 TPR units of T131. Capitalboldface letters denote the four conserved residues that form the"hole" (residues 4, 7, 8, and 11), the three residues that form the"knob" (residues 20, 24, and 27), and the unique proline residueoften found at position 32. Lowercase boldface letters denote otherresidues that fit the TPR consensus sequence. The foliowing equiv-alence based on the examination of "100 TPR units (C.M., unpub-lished data) is used; (A, G, S, V), (E, D), (K, R), (I, L, M, V), and(L, F, Y, H, W). Aspartate (D) residues often found in positions 33and 34 of TPR units are underlined.

181

161241321401481561641721801881961

..:,4

Biochemistry: Marck et al.

Proc. Natl. Acad. Sci. USA 90 (1993)

<---- basic ----> <- helix-1 -> <------- loop --------> <----- helix-2 ---->Consensus ---------RER-R--K Vn--F--Lr--Ip Kv-iLr-Av-yI--L---L-

+ + + + 650 660 + + + + + +620 630 nsglneakqasiwintvselvdifssvknffm 690

T131 frKKKrTp ydaeReRieRERRitaK VvdkYekMkkfel ksrsrk fvgiLrrTkkfnteLdfqle*=*= = = *=** * = = *= =*= =- - =* *+ +

MyoD acKRKtTn adRRKaatmRERRRlsK VneaFetLkrcts snpnqrlp KveiLrnAiryIegLqalLrTBP klasrk yariiqkigfaakftdfkiqAS-CT5 gpsviRRnaRERnRvKq VnngFsqLrqhIp aaviadlsngrrgigpgankkls KvstLkmAveyIrrLqkvLhE-spl m5 hylKvKKpllERqRRaR MnkcLdtLktlVa efqgddailrmd KaemLeaAlvfMrkqvvkqqE-spl m7 qyRKvmKpllEXKRRaR InkcLdeLkdlMa ecvaqtgdakfe KadiLevTvqhLrkLkeskkc-myc nvKRRthnvlERqRRne LkrsFfaLrdqIp elennekap KvviLkkAtayIlsVqaeeqCBF1 XqRKdshKevERRRRen IntaInvLsdlLp vress KaaiLarAaeyIqkLketde

FIG. 5. Comparison of the T131 bHLH motif with MyoD (37), TBP (38) and other bHLH proteins, AS-CT5 (39), E-spl m5, and E-spl m7 (40),c-myc (41) and CBF1 (49). The bHLH consensus is derived from that defined by Murre et al. (36) and from the set of sequences presented here.+ Signs indicate the 3/4 repetitivity of the amphipathic a-helices (36); these positions, located on the same side of the helix, are generallyoccupied by hydrophobic residues (shown here in capital boldface letters). Other conserved residues are shown in lowercase boldface letters.For the sake of compactness, part of the loop of T131 is written one line above helix 1 and helix 2. Sequence similarities of T131 and MyoD arerated according to the Dayhoffpairwise coefficients: *, >4; =, 4; +, 3; -, 2; blank, <2 (42). Boldface letters in TBP sequence indicate similarityto T131.

The second basic region around residue 600 in fact con-stitutes the basic region of a bHLH motif (36). Most of theconserved amino acids of the bHLH pattern are present inT131, in particular the conserved hydrophobic residues mak-ing up the two helices and the basic residues that precedehelix 1 (Fig. 5). Interestingly, among all bHLH proteins, theT13, bHLH motif most closely resembles that of MyoD (36,37). T131 and MyoD share inside the bHLH motif (except theloop), and in the 6 residues immediately upstream, a total of29 identical or evolutionarily conserved (Dayhoff pairwisecoefficient : 2) of 61 residues. The predicted loop extendsfrom position 641 to position 678 and is longer than in otherdescribed bHLH motifs: 39 residues instead of 23 in As-CT5(39). We have also found sequence similarity between helix2 of the T131 bHLH and yeast TBP (Fig. 5). Assuming thesame three-dimensional structure for yeast and ArabidopsisTBPs, this similarity lies in the H2 a-helix of TBP, which isbelieved to be the target of TBP-associating factors (38).

DISCUSSIONWe describe the cloning of TFC4, the gene encoding the131-kDa subunit of TFIIIC that is thought to be responsiblefor assembly of TFIIIB upstream of the transcription startsite. The encoded protein, r131, is shown to be essential invivo and to be part of TFIIIC-tDNA complexes. The dis-covery of TPR units (Fig. 4) and of a bHLH motif (Fig. 5) inT131 sheds light on the molecular structures underlying theintriguing properties of TFIIIC.Each TPR unit is made up of an underlying 34-residue

pattern, of which 8 positions appear to be especially con-served (33). These repeated motifs are believed to exist as"snap helix," where repeats are arranged as a-helices,punctuated by proline-induced turns, with "knob" and"hole" acting as associating motifs located on opposite sidesof the helix (33, 43). What specific functions could TPR unitsconfer on T131? A survey ofthe various TPR proteins indicatesthat the TPR motif can serve in different contexts, whichsuggests a role in protein-protein interactions (see ref. 35 forreview). We speculate that the TPR of T131 could be involvedin the multiple conformational changes that TFIIIC under-goes: (i) When observed by scanning transmission electronmicroscopy, TFIIIC factor alone appeared as two tightlyassociated globular domains. However, upon binding to largetRNA genes, these two domains became clearly separatedand were mapped over the distant A and B blocks (44). (ii)Partial proteolysis ofTFIIIC-tDNA complexes generated theprotease-resistant T domain that retained B block bindingspecificity. Remarkably, this TB-tDNA complex resisteddissociation by single-stranded DNA, while TFIIIC-tDNAcomplexes were fully dissociated (45). This intriguing obser-

vation implied that dissociation of rA due to competitorsingle-stranded DNA (46) caused, in turn, the release of TB,which suggested the existence of a hinge region susceptibleto proteolysis linking the TA and TB domains. (iii) TFIIICundergoes a conformational change upon binding to TFIIIB.Upon addition of TFIIIB to TFIIIC-tDNA complexes, thecrosslinking of T131 to DNA upstream of the transcriptionstart site was enhanced, despite competition of the twofactors for the photochemical reaction, indicating that T131reached closer to DNA (10).TPR units have some structural properties that may ac-

count for the properties of yeast TFIIIC. Limited proteolysisof a TPR-containing protein generates a ladder ofpolypeptidebands at 3- to 4-kDa intervals, suggesting a selective cleavageat the proline-induced turn that punctuates the C-terminalend of each TPR unit (43). Some TPR units could constitutethe postulated protease-sensitive hinge region. In the snaphelix model, each TPR unit (34 residues) forms an a-helix of-50 A. The A-B block distance extends from 31 to 93 bases(i.e., an increase of 220 A) in natural yeast tRNA genes.Several stacked TPR units-for example, TPR 1-9--couldconstitute a hinge and "unsnap" upon binding of TFIIIC toDNA and extend up to 220 A. This hypothesis would corre-late with the finding that T131 extends downstream between Aand B blocks (10). Finally, as protein-protein interactionmotifs, some TPR repeats could be involved in subunit-subunit interactions-with T95, for example-and participatein TFIIIB assembly that also generates considerable confor-mation changes revealed by DNA crosslinking (5, 10). Inter-estingly, another repeating motif, the ankyrin repeat (43residues), has been shown to serve as a molecular tetherbetween subunits a and P ofGA binding protein transcriptionfactor (47).The bHLH motif is a dimerization/DNA binding motif first

identified by Murre et al. (36). Through the formation ofvarious heterodimers, the bHLH transcription factors canmediate a large number of regulatory and differentiationpathways (see ref. 48 for a review; see also refs. listed in thelegend of Fig. 5). The presence of a bHLH motif in T131therefore calls for an interacting, probably DNA binding,partner. No other subunit of TFIIIC already sequencedharbors a bHLH motif. Instead, a likely partner could beTFIIIB since it was shown that, of all the components ofTFIIIC, only T131 can be crosslinked to DNA upstream of thetranscriptional start site (10). Since the 70-kDa component ofTFIIIB can interact by itself with TFIIIC-tDNA complexes,it is a likely candidate to interact with T131. The presence ofa conserved sequence in TBP and the bHLH motif of T131suggests an interaction between the two proteins or a com-mon target. More work, such as generation of point muta-tions in the bHLH region, will be necessary to assess whether

4030 Biochemistry: Marck et al.

Proc. Natl. Acad. Sci. USA 90 (1993) 4031

this motif constitutes the docking part of T131 since mostbHLH proteins found up to now bind to the conserved coresequence CANNTG and no such conserved motif is foundupstream of tRNA genes.

We would like to thank Jean Gagnon (Service de Biologie Struc-turale, Centre d'Etudes de Grenoble) for peptide microsequencing,Catherine Doira (Centre d'Etudes de Saclay) for synthesis of nu-merous oligonucleotides, Genevieve Dujardin (Centre de GenetiqueMoleculaire, Gif-sur-Yvette) for the gift of yeast genomic DNA, JeanPierre Abastado (Institut Pasteur, Paris) and Yveline Frobert (Ser-vice de Pharmacologie et d'Immunologie, Centre d'Etudes deSaclay) for the gift of the 12CA5 antibody, and Robert Swanson(Centre d'Etudes de Saclay) for antibody purification. We thank CarlMann and Cathy Jackson (Centre d'Etudes de Saclay) for reviewingthe manuscript.

1. Geiduschek, E. P. & Tocchini-Valentini, G. P. (1988) Annu.Rev. Biochem. 57, 873-914.

2. Camier, S., Baker, R. E. & Sentenac, A. (1990) Nucleic AcidsRes. 18, 4571-4578.

3. Gabrielsen, 0. S., Marzouki, N., Ruet, A., Sentenac, A. &Fromageot, P. (1989) J. Biol. Chem. 264, 7505-7511.

4. Parsons, M. C. & Weil, P. A. (1990) J. Biol. Chem. 265,5095-5103.

5. Bartholomew, B., Kassavetis, G. A., Braun, B. R. & Gei-duschek, E. P. (1990) EMBO J. 9, 2197-2205.

6. Klekamp, M. S. & Weil, P. A. (1986) J. Biol. Chem. 261,2819-2827.

7. Kassavetis, G. A., Riggs, D. L., Negri, R., Nguyen, L. H. &Geiduschek, E. P. (1989) Mol. Cell. Biol. 9, 2551-2566.

8. Kassavetis, G. A., Braun, B. R., Nguyen, L. H. & Gei-duschek, P. (1990) Cell 60, 235-245.

9. Kassavetis, G. A., Bartholomew, B., Blanco, J. A., Johnson,T. E. & Geiduschek, P. (1991) Proc. Natl. Acad. Sci. USA 88,7308-7312.

10. Bartholomew, B., Kassavetis, G. & Geiduschek, P. (1991) Mol.Cell. Biol. 11, 5181-5189.

11. Margottin, F., Dujardin, G., Gerard, M., Egly, J.-M., Huet, J.& Sentenac, A. (1991) Science 251, 424-426.

12. Comai, L., Tanese, N. & Tjian, R. (1992) Cell 68, 965-976.13. Cormack, B. P. & Struhl, K. (1992) Cell 69, 685-696.14. Schultz, M. C., Reeder, R. H. & Hahn, S. (1992) Cell 69,

697-702.15. White, R. J., Jackson, S. P. & Rigby, P. W. J. (1992) Proc.

Natl. Acad. Sci. USA 89, 1949-1953.16. Lopez-De-Leon, A., Librizzi, M., Puglia, K. & Willis, I. M.

(1992) Cell 71, 211-220.17. Buratowski, S. & Zhou, H. (1992) Cell 71, 221-230.18. Colbert, T. & Hahn, S. (1992) Genes Dev. 6, 1940-1949.19. Kassavetis, G. A., Joazeiro, C. A. P., Pisano, M., Gei-

duschek, E. P., Hahn, S. & Blanco, J. A. (1992) Cell 71,1056-1064.

20. Huet, J. & Sentenac, A. (1992) Nucleic Acids Res. 20, 6451-6454.

21. Swanson, R. N., Conesa, C., Lefebvre, O., Carles, C., Ruet,

A., Quemeneur, E., Gagnon, J. & Sentenac, A. (1991) Proc.Natl. Acad. Sci. USA 88, 4887-4891.

22. Parsons, M. C. & Weil, P. A. (1992) J. Biol. Chem. 267,2894-2901.

23. Lefebvre, O., Carles, C., Conesa, C., Swanson, R. N., Bouet,F., Riva, M. & Sentenac, A. (1992) Proc. Natl. Acad. Sci. USA89, 10512-10516.

24. Dessen, P., Fondrat, C., Valencien, C. & Mugnier, C. (1990)Comput. Appl. Biosci. 6, 355-356.

25. Orr-Weaver, T. L., Szostak, J. W. & Rothstein, R. J. (1981)Proc. Natl. Acad. Sci. USA 78, 6354-6358.

26. Mann, C., Buhler, J.-M., Treich, I. & Sentenac, A. (1987) Cell48, 627-637.

27. Kolodziej, P. A. & Young, R. A. (1991) Methods Enzymol. 194,508-519.

28. Elledge, S. J. & Davis, R. W. (1988) Gene 70, 303-312.29. Covitz, P. A., Herskovitz, I. & Mitchell, A. P. (1991) Genes

Dev. 5, 1982-1989.30. O'Donnel, K. L., Osmani, A. H., Osmani, S. A. & Morris,

N. R. (1991) J. Cell Sci. 99, 711-719.31. Schultz, J. & Carlson, M. (1987) Mol. Cell. Biol. 7, 3637-3645.32. Trumbly, R. J. (1988) Genes 73, 97-111.33. Sikorski, R. S., Boguski, M. S., Goebl, M. & Hieter, P. (1990)

Cell 60, 307-317.34. Boguski, M. S., Sikorski, R. S., Hieter, P. & Goebl, M. (1990)

Nature (London) 346, 114.35. Goebl, M. & Yanagida, M. (1991) Trends Biochem. Sci. 16,

173-177.36. Murre, C., McCaw, P. S. & Baltimore, D. (1989) Cell 56,

777-783.37. Davis, R. L., Weintraub, H. & Lassar, A. B. (1987) Cell 51,

987-1000.38. Nikolov, D. B., Hu, S.-H., Lin, J., Gasch, A., Hohhmann, A.,

Horikoshi, M., Chua, N.-H., Roeder, R. G. & Burley, S. K.(1992) Nature (London) 360, 40-46.

39. Villares, R. & Cabrera, C. V. (1987) Cell 50, 415-424.40. Klanbt, C., Knust, E., Tietze, K. & Campos-Ortega, J. A.

(1989) EMBO J. 8, 203-210.41. Colby, W. W., Chen, E. Y., Smith, D. H. & Levinson, A. D.

(1983) Nature (London) 301, 722-725.42. Dayhoff, M. 0. (1978) Atlas ofProtein Sequence and Structure

(Natl. Biomed. Res. Found., Washington, DC), Vol. 5, Suppl.3.

43. Hirano, T., Kinoshita, N., Morikawa, K. & Yanagida, M.(1990) Cell 60, 319-328.

44. Schultz, P., Marzouki, N., Marck, C., Ruet, A., Oudet, P. &Sentenac, A. (1989) EMBO J. 8, 3815-3824.

45. Marzouki, N., Camier, S., Ruet, A., Moenne, A. & Sentenac,A. (1986) Nature (London) 323, 176-178.

46. Stillman, D. J., Caspers, P. & Geiduschek, E. P. (1985) Cell 40,311-317.

47. Thompson, C. C., Brown, T. A. & McKnight, S. L. (1991)Science 253, 762-768.

48. Lamb, P. & McKnight, S. L. (1991) Trends Biochem. Sci. 16,417-422.

49. Cai, M. & Davis, R. W. (1990) Cell 61, 437-446.

Biochemistry: Marck et al.