Three Human RNA Polymerase III- Specific Subunits Form a Subcomplex with a Selective Function in Specific Transcription Initiation

Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three human RNA polymerase III- specific subunits form a subcomplex with a selective function in specific transcription initiation

Zhengxin Wang and Robert G. Roeder ~ The Rockefeller University, Laboratory of Biochemistry and Molecular Biology, New York, New York 10021 USA

Transcription by RNA polymerase III involves recruitment of the polymerase by template-bound accessory factors, followed by initiation, elongation, and termination steps. An immunopurification approach has been used to demonstrate that human RNA Pol III is composed of 16 subunits, some of which are apparently modified in HeLa cells. Partial denaturing conditions and sucrose gradient sedimentation at high salt result in the dissociation of a subcomplex that includes hRPC32, hRPC39, and hRPC62. Cognate cDNAs were isolated and shown to encode three subunits that are specific to RNA Pol III and homologous to three yeast subunits. The human RNA Pol III core lacking the subcomplex functions in transcription elongation and termination following nonspecific initiation on a tailed template, but fails to show promoter-dependent transcription initiation in conjunction with accessory factors. The capability for specific transcription initiation can be restored either by the natural subcomplex or by a stable subcomplex composed of recombinant hRPC32, h RPC39, and hRPC62 polypeptides. One component (hRPC39) of this subcomplex interacts physically with both hTBP and hTFIIIB90, two subunits of human RNA Pol III transcription initiation factor IIIB. These data strongly suggest that the hRPC32-hRPC39-hRPC62 subcomplex directs RNA Pol III binding to the TFIIIB-DNA complex via the interactions between TFIIIB and hRPC39. [Xey Words: Transcription; RNA polymerase III; transcription initiation] Keceived January 31, 1997; revised version accepted March 27, 1997.

I a contrast to prokaryotes, in which a single RNA poly- dispensable for yeast cell growth, the 14 cloned yeast raerase is responsible for the synthesis of all types of RNA Pol III subunits are all essential for yeast cell vi- cellular RNAs, eukaryotic cells contain three different ability. The two largest subunits are structurally and f3rms of RNA polymerase (RNA polymerases I, II, and functionally related to the two largest subunits of yeast III) that specifically transcribe different classes of genes RNA Pol I and II, as well as the ~' and ~ subunits of in conjunction with distinct sets of accessory transcrip- Escherichia coli RNA polymerase, and are involved in tion initiation factors (for review, see Roeder 1996a). basic RNA Pol III functions such as nucleotide binding t',NA Pol I transcribes rRNA genes, RNA Pol II tran- and interactions with both the DNA template and na- scribes all protein-coding genes and several snRNA scent RNA (Memet et al. 1988; Treich et al. 1992; Dieci ~:enes, and RNA Pol III transcribes genes that produce et al. 1995). Two subunits (yRPAC40 and yRPAC19) small structural RNAs, including 5S and tRNAs, U6 and common to yeast RNA Pol I and III have some sequence 7SK RNAs, and adenovirus virus-associated (VA) RNAs. similarity to the ~ subunit of E. coli RNA polymerase Saccharomyces cerevisiae RNA polymerases have and to yRPB45 of RNA Pol II. In view of the strong evo- been relatively well characterized, both biochemically lutionary conservation in prokaryotic and eukaryotic ,~.nd genetically (Gabrielsen and Sentenac 1991; Young RNA polymerases, subunits yRPAC40, yRPAC19, and ]991). S. cerevisiae RNA Pol III is composed of 16 sub- yRPB45 likely play a role in subunit assembly (Lalo et al. units with sizes ranging from 10 to 160 kD (Gabrielsen 1993). Five small subunits (the common subunits) are and Sentenac 1991). Recently, genes for 14 subunits have identical in S. cerevisiae RNA Pol I, Pol II, and Pol III and been isolated (Gabrielsen and Sentenac 1991; Sadhale are essential components of all three nuclear RNA poly- and Woychik 1994). In contrast to the situation for yeast merases (Gabrielsen and Sentenac 1991; Young 1991). RNA Pol I or II, each of which has some subunits that are Their sequences have not yet provided clues to their functions; however, it has been proposed that the com- 1Corresponding author. mon subunits could be involved in nuclear localization, E-MAIL [email protected];FAX (212) 327-7949. transcription efficiency, or the coordinate regulation of

GENES & DEVELOPMENT 11:1315-1326 © 1997 by Cold Spring Harbor LaboratoryPress ISSN 0890-9369/97 $5.00 1315 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Wang and Roeder

RNA synthesis (Young 1991). In addition, subunits dard fractionation on phosphocellulose (P11), and the de- yRPC82, yRPC53, yRPC34, and yRPC31 are specific to rived P11 0.35 M KC1 fraction, containing most of the yeast RNA Pol III and have no counterparts in yeast RNA Pol III activity, was incubated with agarose beads RNA Pol I or Pol II. An RNA Pol III with a genetic mu- containing an immobilized monoclonal antibody di- tation in the amino-terminal zinc-binding domain of rected against the FLAG epitope (M2 agarose). After yRPC160, the largest subunit of RNA Pol III, simulta- washing, the bound RNA Pol III was eluted with a FLAG neously loses three subunits (yRPC31, yRPC34, and peptide. Analysis by gradient SDS-PAGE revealed that yRPC82) upon heat inactivation and ion-exchange chro- the immunopurified human RNA Pol III contains 15 matography (Werner et al. 1992). In addition, one polypeptides of 155, 135, 82, 62, 53, 40, 39, 36, 32, 29, 20, (yRPC34) of these subunits has been mapped to a more 18, 15, 12, and 10 kD (Fig. 1A). These results are consis- upstream position on the promoter (Bartholomew et al. tent with our previously published subunit composition 1993) and interacts with a subunit (yTFIIIB70) of yeast of human RNA Pol III (Wang and Roeder 1996). Under TFIIIB (Werner et al. 1993; Khoo et al. 1994). These re- the same conditions, RNA Pol III can also be immuno- sults raised the possibility that these subunits form a purified directly from nuclear extract made from the subcomplex, within RNA Pol III, that plays a role in the same cell line. The purity, however, is not as high as that recognition of the preinitiation complex. The observation of the enzyme purified from the P11 0.35 M KC1 fraction that a conditional mutation in the yRPC31 subunit (partial (data not shown). Because of the structural complexity of deletion of the carboxyl terminus) impairs transcription human RNA Pol III, two-dimensional gel electrophoresis initiation, but not transcription elongation and termina- was used to further analyze its polypeptide composition tion, supports this hypothesis (Thuillier et al. 1995). (Fig. 1B; diagrammed in 1C). One band migrating at the In comparison to S. cerevisiae RNA Pol III, human 39-kD position in one-dimensional gel electrophoresis RNA Pol III is poorly characterized in terms of subunit (Fig. 1A) was further resolved into two polypeptides, now composition and function. Only one cDNA, that encod- designated hRPC39 and hRPC38, by two-dimensional ing the 53-kD subunit of human RNA Pol III, has been gel electrophoresis (Fig. 1B, C). In addition, two-dimen- cloned (Ittmann et al. 1993). This subunit shows a low sional gel electrophoresis analysis revealed apparent sequence similarity with yeast RPC53 and does not modifications of several subunits (hRPC39, hRPC38, and complement RPC53 mutations in yeast (Man et al. hRPC62) because multiple spots existed for each of these 1992). We showed previously that 15 polypeptides coe- subunits (Fig. 1B). The smaller subunits (hRPC10, lute with the human RNA Pol III activity during con- hRPC12, hRPC15, hRPC18, and hRPC20) were nega- ventional purification (Wang and Roeder 1996). In this tively stained by silver and show up as white spots on study we have further investigated the subunit compo- the two-dimensional gel (Fig. 1B). sition of human RNA Pol III and the mechanism directing RNA Pol III binding to the TFIIIB-DNA complex. We Subunits hRPC32, hRPC39/38, and hRPC62 show that an affinity purified human RNA Pol III is com- selectively dissociate from RNA Pol III posed of at least 16 subunits and that three or four of these form a subcomplex that can be dissociated from a RNA Pol III subunits are tightly associated and remain residual core RNA Pol III. The core enzyme lacking the intact during conventional chromatographic purification three subunits is normal in terms of transcription elon- and immunoprecipitation processes. It was reported, gation and termination, but incapable of accurate tran- however, that specific subunits of yeast RNA polymer- scription initiation. We show further that the natural ases (yRPC31, yRPC34, and yRPC82 of RNA Pol III; subcomplex, as well as a recombinant three-subunit sub- yRPB32, and yRPB17 of RNA Pol II; yRPA39.5 and complex, interacts physically with components of initia- yRPA49 of RNA Pol I) selectively dissociate from RNA tion factor TFIIIB and restores transcription initiation of polymerases under some conditions (Edwards et al. 1991; the core RNA Pol III lacking the three subunits, thus Werner et al. 1992; Smid et al. 1995). These observations suggesting that this interaction is a determinant that di- suggested that different subunits may play different roles rects RNA Pol III to its cognate genes. in maintaining polymerase structure and function. To see whether specific subunits of human RNA Pol III could be selectively dissociated from the core enzyme, Results different concentrations of urea or sarkosyl was used to wash the RNA Pol III immobilized on M2-agarose beads. Immunopurification and subunit composition of The washed RNA Pol III was then eluted from the beads human RNA Pol HI with FLAG peptide and analyzed by SDS-PAGE (Fig. 2A). We showed previously that 15 polypeptides copurified, Consistent with the concept of a tight structure for hu- through a number of conventional chromatography man RNA Pol III, 2 M urea (Fig. 2A, lane 3 vs. lane 2) or steps, with the human RNA Pol III activity (Wang and 0.3% Sarkosyl (data not shown) washes apparently did Roeder 1996). To simplify the purification procedure, we not change the subunit composition of RNA Pol III. established a cell line (BN51) that constitutively ex- However, a 4 M urea wash selectively removed subunits presses a FLAG epitope-tagged 53-kD subunit of human hRPC32, hRPC39/38, and hRPC62 from the others, RNA Pol III (for details, see Materials and Methods). An which remained associated (Fig. 2A, lane 4 vs. lane 2). A S100 fraction from this cell line was subjected to stan- 0.5% Sarkosyl wash also selectively dissociated the

1316 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three RNA Pol III subunits

B C

~--1S5 pa pH ~--135 V

-- 82 --1.155 -- 135 -- 62 82 ~ 62

--411 04o -- 39/38 --36 -- 32 32 0 O~

M 2o~. 1 o f,18 15 ~. ,°_~.12

lo0

Figure 1. Subunit composition of immunopurified human RNA Pol III from HeLa cells. (A) The immunopurified human RNA Pol III was separated by a gradient (4%-18 %) SDS-PAGE. The gel was cut into two parts and the proteins on each part were transferred onto PVDF membranes and stained with Ponceau S. The molecular sizes (kilodaltons) of the subunits are indicated at the right. The band indicated with an asterisk represents the junction of the two parts of the gel. (B) Two-dimensional electrophoretic pattern of the immunopurified human RNA Pol III. Immunopurified RNA Pol-III (107al) was applied to the basic end of the isoelectrofocusing tube gel. First-dimension electrophoresis was performed in a gel containing 3.5 % polyacrylamide cross-linked with 0.3 % bis diacrylamide, 9 M urea and 2% ampholytes (1 part Bio-lyte 3/10 and 2 parts Bio-lyte 5/7). Cathode solution (upper reservoir): 100 mM NaOH. Anode solution (lower reservoir): 10 mM HgPO 4. Second dimension electrophoresis was performed in a gradient (4%-18%) SDS polyacrylamide gel. Protein spots were visualized by silver staining. (C) Schematic representation of the pattern of the subunits of human RNA polymerase III in two-dimensional gel electrophoresis. The subunit spots stained positive by silver are denoted by solid circles or ovals and those that stained negative are denoted by broken circles. same subunits (data not shown). When immunopurified hRPC62) again were clearly separated from the others RNA Pol III was subjected to an overnight sucrose gra- (Fig. 2B). Moreover, the dissociated subunits cosedi- dient centrifugation step (for details, see Materials and mented in fractions 7-10 whereas the others cosedi- Methods), the same subunits (hRPC32, hRPC39/38, and mented in fractions 4-6. These results indicate that the

B Figure 2. Selective dissociation of sub- ~ Fr M 1 2 3 4 5 6 7 8 9 10 units hRPC32, hRPC39/38, and hRPC62 from human RNA Pol III. (A) Dissociation by 4 M urea treatment. Gradient (4%-18 %) SDS-PAGE of the immunopurified RNA IZ o o o 220 Pol III in the absence of (lane 2), or follow- 220 -- ing treatment with 2 M (lane 3) or 4 M (lane 116 -- 97-- 4) urea as described in Materials and Meth- 66 -- -- RPC62 116 ods. The gel was stained with silver. The 45 -- 97 subunits (hRPC32, hRPC39/38, and -- RPC39138 hRPC62) selectively released from the -- RPC62 core RNA Pol III are indicated at the right. 32-- -- RPC32 (Lane 1 ) Protein molecular weight markers (Bio-Rad). (B) Dissociation by sucrose gra- 21- dient sedimentation. Sucrose gradient cen- -- RPC39/38 trifugation fractions (10 ~11 were subjected to gradient (4%-18%) SDS-PAGE. The gel -- RPC32 1234 was stained with silver. The core RNA Pol III is in fractions 4 to 6 and the dissociated subunits (hRPC32, hRPC39/38, and hRPC62, indicated at the right) are in fractions 7-10. Lane 1 is the protein molecular 1 2 3 4 5 6 7 8 9 10 11 markers (Bio-Rad).

GENES & DEVELOPMENT 1317 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Wang and Roeder dissociable subunits form a stable subcomplex whereas III (Fig. 3A, lane 3 vs. lane 2). Promoter-directed tran- the others form a residual core enzyme, and that the scription by RNA Pol III involves accurate initiation, corresponding complexes can be separated under par- efficient elongation and correct termination. To distin- tially denaturing conditions and by sucrose gradient guish which step is impaired with the residual core en- sedimentation. Because the two smallest subunits of zyme, a tailed template was used to measure transcrip- RNA Pol III were not resolved in these analyses, we can- tion elongation and termination by RNA Pol III in the not exclude the possibility that they are also dissociated absence of other accessory factors. After initiation at the from the core enzyme. single-stranded terminus, RNA Pol III should transcribe the template strand of the VA1 gene and encounter the two strong tandem VA1 termination signals located near RNA Pol III lacking hRPC32, hRPC39/38, and the 3' end of the gene (Fig. 3C). In this case, transcription hRPC62 is defective in promoter-directed transcription by intact RNA Pol III generates three major transcripts initiation that result from termination at the first (designated T1) The function of the core RNA Pol III lacking the sub- and the second (designated T2) termination sites and complex was first assayed on the adenovirus VA1 tem- from readthrough to the end of the VA1 DNA fragment plate. Figure 3A shows that the intact immunopurified (designated RT). The termination pattern produced by RNA Pol III supported accurate transcription from the the residual core RNA Pol III was similar to that pro- VA1 template in the presence of purified human TFIIIB, duced by the intact RNA Pol III (Fig. 3B, lane 12 vs. lane TFIIIC1, and TFIIIC2 (lane 2), whereas equimolar 6). A kinetic analysis also failed to reveal any differences amounts of the core RNA Pol III lacking hRPC32, between the two enzymes (Fig. 3B, lanes 7-12 vs. lanes hRPC39/38, and hRPC62, obtained either by the 4 M 1-6). urea wash (lane 4) or by sucrose sedimentation (fraction These observations indicate that the dissociated sub- 4; lane 5), were clearly defective under these conditions. units do not play detectable roles in RNA Pol III tran- The residual activities (Fig. 3A, lanes 4,5) account for scription elongation and termination. The core RNA Pol 10% of that produced by the intact RNA Pol Ill (lane 2) III lacking these subunits, however, was overall threefold and may reflect the incomplete separation of the disso- less active in transcription when compared with an ciated subunits (Fig. 2B, lane 5). Consistent with the equimolar amount of intact RNA Pol III (Fig. 3A, lane 12 SDS-PAGE analysis (Fig. 2A, lane 3), the RNA Pol III vs. lane 6). The reason for this reduced transcription washed with 2 M urea supports accurate transcription at level is unknown but could reflect different levels of ini- a level similar to that observed with untreated RNA Pol tiation on the tailed template. In addition, for intact

Figure 3. Functional analysis of the core B ,..,, RNA Pol III. (A) Promoter-dependent tran- A g pol III core pol III (SG #4) scription on the VA1 template. Reactions 0 15,, "m" 30" 1' 5' 25' I I5" 10" 30" 1' 5' 25' ] contained the VA1 template, TFIIIB, TFIIIC1, TFIIIC2, and either 48 fmoles of ~ o the immunopurified RNA Pol III (lane 2), 0 RT 48 fmoles of the enzyme treated with 2 M --T2 urea (lane 3), 48 fmoles of the core enzyme T1 obtained by 4 M urea treatment (lane 4), or VA1 48 fmoles of the core enzyme from sucrose gradient centrifugation (lane 5). (B) Kinetic analysis of promoter-independent tran- 1 2 3 4 5 scription. Reactions contained the tailed template (C) and either 48 fmoles of the immunopurified RNA Pol III (lanes 1-6) or C 48 fmoles of the core enzyme (lanes 7-12). ,t,tT1 1"2 Numbers at the top of each lane indicate CnCCCACGT r,//# 1 v,1 RNA ~ene + 156 ~I'/~/] reaction time in seconds and minutes. The arrows indicate transcripts terminated at T1 ,~ (197) the first termination site (designated Tll, T2 ~,~ (236) the second termination site (designated RT v~ (259) T2), or the end of the template (designated RT). (C) The deoxycytidine monophos- phate (C)-tailed PstI-EcoRI fragment. The single-stranded tail of cytosine residues is 1 2 3 4 5 6 7 8 9 10 11 12 indicated at the 3' end of the PstI site by Cn, where n is the number of residues added. Positioned downstream of the tailed end is the 156-nucleotide VA1 RNA gene denoted by the shaded area. The hatched areas both 3' and 5' to the gene denote regions of VA1 RNA gene-flanking sequences extending from -31 to +203 with respect to the VA1 transcription start site. The two tandem termination sites at the end of the VA1 gene are shown as T1 and T2. Transcripts synthesized from this template by RNA Pol Ill are indicated with their lengths given in nucleotides.

1318 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three RNA Pol III subunits

RNA Pol III, the relative levels of transcripts terminated The cDNA for hRPC32 (GenBank accession no. U93868) at the first termination site (T1), at the second termina- contains an open reading frame encoding a protein of 233 tion site (T2), and at the end of the template (RT) were amino acids with a predicted mass of 27 kD and a pI of 17%, 58%, and 25%, respectively (Fig. 3B, lanes 5,6). For 4.54. The carboxy-terminal acidic tail of hRPC32 is a the core enzyme lacking the dissociated subunits, how- characteristic of the yeast RPC31 subunit and is essen- ever, these distributions were 24%, 35%, and 42%, re- tial for yeast cell viability (Mosrin et al. 1990). Human spectively (Fig. 3B, lanes 11,12), mainly reflecting more RPC32 shows a low, but significant, sequence similarity readthrough and less termination at T2. These results to the yeast RPC31 subunit (25 % identity and 51% simi- suggest that there are subtle differences between intact larity) (Mosrin et al. 1990). The cDNA for hRPC39 (Gen- RNA Pol III and the core enzyme lacking hRPC32, Bank accession no. U93869) has an open reading frame hRPC39/38, and hRPC62 on the tailed template. encoding a protein of 317 amino acids with a predicted mass of 36 kD and a pI of 6.35. The protein sequence of hRPC39 shows 27% identity and 50% similarity to the Cloning of cDNAs encoding three subunits of human yeast RPC34 subunit (Stettler et al. 1992). Antibodies RNA Pol HI raised against corresponding bacterially-expressed pro- Database searches showed that two peptide sequences tein reacted with RPC39, but not RPC38, in immunoblot derived from the human RPC62 subunit (see Materials analysis of RNA Pol III subunits resolved in 2-D gel (data and Methods) match an EST sequence (GenBank acces- not shown). The cDNA for hRPC62 (GenBank accession sion no. H19022) for which the cDNA was obtained. By no. U93867) has an open reading frame encoding a pro- use of probes designed according to peptide sequences tein of 533 amino acids with a predicted mass of 60 kD derived from subunits hRPC32 and hRPC39, correspond- and a pI of 8.35. The overall sequence of hRPC62 shows ing cDNAs were obtained by screening a human HeLa a relatively low sequence similarity to the yeast RPC82 cell cDNA library. The deduced protein sequences for subunit (20% identity and 42% similarity)(Chiannilkul- hRPC32, hRPC39, and hRPC62 are shown in Figure 4. chai et al. 1992). On the basis of their sequence similari-

hRPC32 1 MAGNKGRGRAA ...... YTF~SKGEKLPDVVLKPPPLFPDT 41 C I.: :1 :t:: :.:..:.11 .... :t.:.l. . :.:. hRPC62 3 QAEIKLCSLLLQEHFGEIVEKIGVHLIRTGSOPLR.VIAHDTGTSLDQVK 51 yRPC31 1 MSSYRGGSRGGGSNYMSNLPFGLGYGDVGKNHITEFPSIPLPINGPITNK 50 .::: I:. I:..1:11 ...: . I: i. .:I ::.. .I .:1 II yRPC82 43 NPDLFLYKELVKAHLGERAASVIGMLVALGRLSVRELVEKIDGMDVDSVK 92 hRPC32 42 D ...... YKPVPLKTGEGEEYMLALKQELRETMKRMPYFI.ET 77 • I..I: II. . :: .tl ... ll.I :t :. hRPC62 52 KALCVLVQHNLVSY ...... QVHKRGVVEYEAQCSRVLRIVRYPRYIYTT 95 yRPC31 51 ERSLAVKYINFGKTVKDGPFYTGSMS..LIIDQQENSKSGKRKPNIILDE 98 ..I. I.f . I.I ..I • I:.:. ::l .... ] yRPC82 93 TTLVSLTQLRCVKYLQETAISGKKTTYYYYNEEGIHILLYSGLIIDEIIT 142 hRPC32 78 PEERQDIERYSKRYMKVYQEEWIPDWRRLPREMMPRNKCKKAGPKPKK,. 125 .:,.::llltl.:l:l . : I :.. :::1.: :, I ..ll hRPC62 96 KTLYSDTGEL ...... IVEELLLNGKLTMCTVVKKV.ADRLTET ...... 132 yRPC31 99 DDTNDGIERYSDKYLKKRKIGISIDDHPYNLNLFPNELYNVMGINKKKLL 148 • .1•:1 I1:::: I.I1: • :..I .I.:. f yRPC82 143 QMRVNDEEEHKQLVAEIVQNVISLGSLTVEDYLSSVTSDSMKYTISSLFV 192 hRPC32 126 ...... AKDAGKGTPLTNTEDVLKKMVELEKRGDGEKSDEENEEKEGSK 168 hRPC62 133 ...... MEDGKTMDYXEVSNTF~FVQRCPSVP ...... 164 yRPC31 149 AISKFNNADDV~TGTGLQDENIGLSMLAKLKELAED...VDDASTGDGAA 195 : : .,:,i ,:,: :. I :.1: .... yRPC82 193 QLCEMGYLIQISKLHYTPIEDLWQFLYEK]{YKNIPRNSPLSDLKKRSQAK 242 hRPC32 169 EKSKEGDDDDDDDAAEQEEYDEEEQEEENDYINSYF.,EDGDDFGADVMT 216 • II.l:::lll I.:: I It:l::::: :.It :1:11:1.: hRPC62 165 ...... TTENSDPGPPPPAPTLVINEKDMYLVPKLSLIGKGKR 201 yRPC31 196 KGSKTGEGEDDDLADDDFEEDEDEEDDDDYNAEKYFNNGDDDDYGDE... 242 ..I I: ...I ..I I ...: I. .I. : II:l yRPC82 243 MNAKTDFAKIINKPNELSQILTVDPKTSLRIVKPTVSLTINLDRFMKGRR 292 hRPC32 217 TWMRQPIRHEIF 228 :I .. I hRPC62 202 RRSSDEDAAGEPKAKRPKYTTDN ...... KEPIPDDGIYWQANLD 240 yRPC31 243 ...EDPNEEAAF 251 .: : I ...... :.. . T:f:•:.l::.:.: . yRPC82 293 SKQLINLAKTRVGSVTAQVYKIALRLTEQKSPKIRDPLTQTGLLQDLEEA 342

hRPC39 1 MGEVKVKVOPPDADPVEIENRIIELCHOFPHGITDQVIQNEMPQYRSPAA 50 hRPC62 241 RFRQH ...... FRDQAIVSAVANRMI~TSSEIVRTMLRMSEIT 277 t::: ..... :: .:..•: ::..:. :t:l :l.:t .... : I. I.. .:.. ::. :l :1 : I .... yRPC34 1 MSGMIENGLQLSDNAKTLHSQM•.MSKGIGALFTQQELQKQMGIGSLTDL 48 yRPC82 343 KSFQDEAELVEEKTPGLTFNAIDLARHLPAELDLRGSLLSRKPSDNKKRS 392

hRPC39 51 GSSINRLLSMGQLDLLRSNTGLLYRIKDSQNAGKMKGSDNQEKLVYQIIE i00 hRPC62 278 TSSSAP ...... FTQPLFSNEIFRSLPVGYNI ...... 303 I :. II... :.1:: I.:1 :. . :1.1 .... :1 Ill .11 .l••l: I. I :...: :lJ. : :. yRPC34 49 MSIVQELLDKNLIKLVKQNDELKFQGVLESEAQKKATMSAEEALVYSYIE 98 yRPC82 393 GSNAAASLPSKKLKTEDGFVIPALPAAVSKSLQESGDTQEEDEEEEDLDA 442

hRPC39 i01 DAGNKGIWSRDIRYKSNLPLTEINKILKNLESKKLIKAVKSVAASKKKVY 150 hRPC62 304 ...... SKQVLDQYLTLLADDPLEFVGKSGDSGGGMYVINLHKALRSLAT 347 • .l•.lll[:.l: :.11. : I.II.Itl.:.:l.llll ..:1:1 I ::: .I.:lJ.. :.1: .:. .f:l : . I :. I . yRPC34 99 ASGREGIWSKTIKARTNLHQHVVLKCLKSLESQRYV-KSVKSVKFPTRKIY 148 yRPC82 443 DTEDPHSASLINSHLKILASSNFPFL...NETKPGVYYVPYSKLMPVLKS 489

hRPC39 151 MLYNLQPDRSVTGGAWYSDODFESEFVEVLNQQCFKFLQSKA ...... ET 194 hRPC62 348 ATLESVVQERFGSRCARICRLVLQKKHIEQKQVEDFAMILQ.EAKDMLYK 396 Ill.Ill. .:111:1:.1.::: ll:: I .::1: ... • ..i I:.. :I... I::I : :.i :.:i ::. I:: : : :. I • yRPC34 149 MLYSLQPSVDITGGPWFTDGELDIEFINSLLTIVWRFISENTFPNGFKNF 198 yRPC82 490 SVYEYVIASTLGPSAMRLSRCIRDNKLVSEKIINST~EKDIRSTLAS 539

hRPC39 195 ARESKQNPMIQRN.SSFASSHEVWKYICELGISKVELSMEDIETILNTLI 243 hRPC62 397 MLSGNFMSLQEIPKTPDHAPSRTFYLYTVNILSAARMLFDRCYKSIANLI 446 ..:.1.1.:...I ..:...:1::.:1.. .:..IJI. .:1 -: :.1: ::• i :.:II:I:I:I:.:11..:I: .... :: : .... :III: yRPC34 199 ENGPKKNVFYAPNVKNYSTTQEILEFITAAQVANVELTPSNIRSLCEVLV 248 yRPC82 540 LIRYNSVEIQEVPRTADRSASRAVFLFRCKETHSYNFMRQNLEWNMANLL 589

hRPC39 244 YDGKVEMTIIACKRRHSWQCRWTHETVQGSQSNHPSHRFGPGHPVDSAPV 293 hRPC62 447 ERRQFETKENKRLLEKSQRVEAI IASMQATGAEEAQLQE I EEMITAPERQ 496 II:l:l .... I :1 I1.: I I: t.1.1 .:... ::: ..II. 11.1..I : : :.: .... :I.:: .f:...l • yRPC34 249 YDDKLEKVTHDC ...... YRVTLESIL..QMNQ ..... GEGEPEAGNKA 284 yRPC82 590 FKKEKLKQENSTLLKKANRDD.. VKGRENELLLPSELNQL, KMVNERELN 636

hRPC39 294 FDDCHE ...... GGEISPSNCIYMTEW 314 hRPC62 497 QLETLKRNVNKLDASEIQ 514 ::I ,I .:. ..:.:1:.11 :. I.I :• ::. ::. yRPC34 285 LEDEEEFSIFNYFKMFPASKHDKEVVYFDEW 315 yRPC82 637 LFARLSRLL SLWEVFQMA 654 Figure 4. The predicted amino acid sequences and sequence alignments with the yeast counterparts of the three RNA Pol III-specific subunits (hRPC32, hRPC39, and hRPC62). The peptide sequences obtained by direct microsequence analyses are underlined• (A) The predicted amino acid sequence of human RPC32 and alignment with yeast RPC31 (Mosrin et al. 1990). (B) The predicted amino acid sequence of human RPC39 and alignment with yeast RPC34 (Stettler et al. 1992). (C) The predicted amino acid sequence of human RPC62 and alignment with yeast RPC82 (Chiannilkulchai et al. 1992).

GENES & DEVELOPMENT 1319 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Wang and Roeder ties, dissociation from the core enzyme and interactions plus the fact that the three subunits dissociate together with TFIIIB (see below), it is probable that these three from the enzyme, makes it likely that this subcomplex subunits are the human counterparts of the yeast RNA also exists within RNA Pol III. Pol III-specific subunits yRPC31, yRPC34, and yRPC82. These represent three of a total of four RNA Pol III-spe- Subunit hRPC39 of the subcomplex interacts with cific subunits in yeast. Because the leucine repeats close human TFIIIB to the carboxy-terminal end of yRPC82 are not conserved in hRPC62, this argues against the possible in- RNA Pol III is recruited onto the template through involvement of the leucine repeats in interactions with the teractions with template-bound TFIIIB (Kassavetis et al. other RNA Pol III subunits as was suggested previously 1990). We tested whether these interactions could occur (Chiannilkulchai et al. 1992). independently of the template by a coimmunoprecipita- tion assay. Figure 6A (lanes 1-6) shows that antibodies against the RPC82 subunit of human RNA Pol III can RNA Pol III-specific subunits hRPC32, hRPC39, and specifically coimmunoprecipitate RNA Pol III (detected hRPC62 form a subcomplex by the anti-hRPC82 antibody) and human TFIIIB (de- Consistent with the demonstrations that yeast subunits tected by anti-hTBP and anti-hTFIIIB90 antibodies) from RPC31, RPC34, and RPC82 can interact with each other HeLa cell nuclear extract (lanes 1-6). Figure 6A (lanes and be dissociated from the core enzyme (Valenzuela et 7-14) also shows that antibodies against human TFIIIB90 al. 1979; Werner et al. 1992, 1993), the human counter- can specifically coimmunoprecipitate both TFIIIB (de- parts hRPC31, hRPC39, and hRPC62 were also simulta- tected by anti-hTFIIIB90 antibodies) and RNA Pol III (de- neously dissociated from human RNA Pol III (Fig. 2) and tected by anti-hRPC82, -hRPC20, and -hRPC53 antibod- cosedimented during sucrose gradient analysis (Fig. 2B). ies). The persistence of these interactions in the presence To further test the hypothesis that these three subunits of Chromomycin A3, an antibiotic that has high affinity form a subcomplex, the corresponding cDNAs were ex- for double-stranded DNA and strongly inhibits DNA- pressed in bacteria and Sf9 cells. A subcomplex was as- directed RNA synthesis both in vivo and in vitro (Keniry sembled from the recombinant proteins by incubation of et al. 1993), indicates that the interactions are indepen- immobilized FLAG-tagged hRPC39 with excess dent of DNA (data not shown). We further investigated amounts of histidine-tagged hRPC32 and hRPC62. The which subunit is involved in these interactions. It was washed and FLAG peptide-eluted complex contained an shown previously that the yeast RPC34 subunit inter- -1:1:1 molar ratio of the three subunits as determined by acts with yeast TFIIIB70 (Werner et al. 1993; Khoo et al. SDS-PAGE and Coomassie blue staining with bovine se- 1994). Because human RPC39 and TFIIIB90 are counter- rum albumin as a standard (Fig. 5, lanes 3,4). When ana- parts of yeast RPC34 and yeast TFIIIB70 (Fig. 4B; Wang lyzed by chromatography on a Superose 12 column and Roeder 1995), respectively, we first investigated (Smart System, Phamarcia), this recombinant subcom- whether interactions between these two human proteins plex eluted as an 160-kD entity (data not shown). These can be observed by glutathione S-transferase (GST)-fu- results show directly that these subunits form a stable sion protein pull-down assays. Figure 6B shows interac- subcomplex in the absence of the other subunits. This, tions of TFIIIB (detected by anti-TFIIIB90 antibodies) with GST-hRPC39 (lane 4) but not with GST (lane 2), GST-hRPC32 (lane 3), or GST-hRPC62 (lane 5). To know which subunits of TFIIIB and which domains of RPC39 are involved in these interactions, we used recombinant hTFIIIB90, hTBP, and truncated hRPC39 mutants fused 97 ' to GST (Fig. 6B, lanes 6-11) to carry out the same experi- 66 .... -- ~ -- RPC62 ment. Additional data in Figure 6B indicate that both hTFIIIB90 (lanes 12-14) and hTBP (lanes 19 and 20) in-

45-- ~ ~ --RPC39 teract specifically with hRPC39. The domain of hRPC39 that interacts with hTFIIIB90 is located between amino --RPC32 acid residues 84 to 317 (lanes 12-18) and the domain that 32~ : : interacts with TBP between residues 179 to 317 (lanes 19-24). These results are consistent with yeast studies 21 : showing that the carboxyl terminus of yRPC34 is re-

1234 quired for interactions with yTFIIIB70 (Khoo et al. 1994). Previous studies failed to detect an interaction of yeast Figure 5. Three recombinant RNA Pol III-specific subunits RPC34 with yeast TBP by the yeast two-hybrid assay form a subcomplex in vitro. Recombinant hRPC62 (lane 2) and (Werner et al. 1993). A possible explanation for this dif- different amounts (lane 3, 1.5 pmole; lane 4, 3 pmole) of the subcomplex formed from the bacterially expressed hRPC39 and ference is that the interaction domain on yeast TATA- baculovirus expressed hRPC32 and hRPC62 were analyzed by binding protein (TBP) is masked by TBP-interacting pro- SDS-PAGE (10% gel). The gel was stained with Coomassie Bril- teins in vivo. The functional relevance of these interac- liant blue R-250. Lane 1 shows protein standard molecular tions needs to be further investigated through analysis of weight markers (Bio-Rad). point mutations which affect these interactions.

1320 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three RNA Pal III subunits

Figure 6. hRPC39 interacts with hTFIIIB90 and TRIIB90 immunoprecipitate hTBP. (A) Human TFIIIB and RNA Pol III interact (~ RPC82 immunoprecipitate | in HeLa nuclear extract. RNA Pol III was immu- I I P I P i P I P I P I P I P I 116 -- noprecipitated from HeLa nuclear extract by binding to a polyclonal anti-RPC82 preimmune (lanes 116 .... [ 80-- m t 80-- D 1,3,5) or immune (lanes 2,4,6) antibody resin as 49-- described previously (Wang and Roeder 1995). Af- 4, ter washing, the resin was boiled in SDS sample buffer and analyzed by Western blot with polyclonal antibodies against hRPC82 (lanes 1,2), hT- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 FIIIB90 (lanes 3,4) or hTBP (lanes 5,6). Human I I L I I I t I I I I I I I TFIIIB90 was immunoprecipitated from HeLa Blot: (x RPC82 ~ TFIIIBg0 c~ TBP a TFIIIB90 ¢¢ RPC82 RPC20 ~RPC53 nuclear extract by binding to a polyclonal anti- TFIIIB90 preimmune (lanes 7, 9,11,13) or immune (lanes 8,10,12,14) resin. The resin was boiled in m SDS sample buffer and analyzed by western blot with antibodies against hTFIIIB90 (lanes 7,8), hRPC82 (lanes 9,10), hRPC20 (lanes 11,i2) or ~¢qID o(Jo hRPC53 (lanes 13,14). For the western blot assays, o. o. o. rr rr n- identical samples were electrophoresed in lanes '~' 1,3, and 5 (~RPC82 preimmune immunoprecipi- --~~ tate), in lanes 2,4, and 6 (c~RPC82 immunoprecipi- tare), lanes 7, 9, 11, and 13 (~TFIIIB90 preimmune m -~- TFIIIBg0 immunoprecipitate) and in lanes 8, 10, 12, and 14 (o~TFIIIB90 immunoprecipitate). After blotting, regions of the membrane containing sequential pairs of lanes were separated and incubated with the in- 12345 6 7 8 91011 dicated antibodies. After processing and autoradi- ography, the separated regions were realigned to their original gel positions. The sizes of the immu- noreactive bands correspond exactly to the sizes of the proteins to which the antibodies were raised. In lane 8, the uppermost band corresponds to the intact full-length TFIIIB90, whereas the other bands correspond to breakdown products that are -9,- TFilIB90 e ~,, e --~-TBP not present in the TFIIIB that is immunoprecipitated in association with the RNA Pol III in nuclear extract (lane 4). (B) hRPC39 directly inter- 12131415161718 192021 222324 acts with hTFIIIB90 and hTBP. GST-fusion full- length hRPC32, hRPC39, hRPC62, or truncated hRPC39 proteins were expressed in bacteria and immobilized on glutathione-agarose beads. Protein amounts were normalized by SDS-PAGE and stained with Coomassie Brilliant blue R-250 (lanes 6-11 and data not shown for GST-RPC32 and GST-RPC62). Beads containing GST or GST-fusion proteins were incubated with the TFIIIB-containing P11 B fraction of HeLa nuclear extract (lanes 1-5), Sf9 cell extract containing recombinant hTFIIIB90 (lanes 12-18) or bacterial lysate containing recombinant hTBP (lanes 19-24) for 1-3 hr at 4°C. The beads washed five times with BC150-0.05% NP-40, then boiled in the SDS sample buffer for Western blot analysis with antibodies against hTFIIIB90 (lanes i-5; 12-18) or hTBP (lanes 19-24). IP denotes 10% of material used for the pull-down assays. GST-RPC39 contain full-length hRPC39 whereas other fusion proteins contain various hRPC39 deletion proteins with the residues present indicated by numbers.

The three subunit subcomplex is selectively required purified natural subcomplex (lane 5) and the recombi- for gene-specific transcription initiation by RNA nant subcomplex {lane 7) can restore the capability for Pol HI gene-specific transcription to the core enzyme lacking the three subunits. Under the same conditions, neither Human RPC39 physically interacts with hTFIIIB and is a the natural or recombinant subcomplex (lanes 4 and 6) part of the subcomplex containing RNA Pol III-specific nor the core RNA Pol III (lane 3) can independently sup- subunits hRPC32, hRPC39/38, and hRPC62. Hence, we port gene-specific transcription (Fig. 7A). The recombi- attempted to correct the transcription defect in the core nant subcomplex was twofold less active, on a molar RNA Pol III lacking these subunits (sucrose fraction 4) by basis, compared with the natural purified subcomplex incubation with an equimolar amount of either the natu- (Fig. 7A, lane 7 vs. lane 5). This could reflect partial dis- ral subunit subcomplex (sucrose gradient fraction 9) or a assembly of the recombinant subcomplex, the lack of subcomplex reconstituted with recombinant hRPC32, modifications of recombinant hRPC39 and hRPC62 (the hRPC39, and hRPC62. Figure 7A shows that both the analysis of Fig. 1B indicates in vivo modifications), or the

GENES & DEVELOPMENT 1321 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Wang and Roeder

ral subcomplex, possibly due to post-translational modifications or the presence of other dissociated subunit(s) in the latter. o

- = _= i (~I m Discussion = • Accurate transcription by RNA Pol III is a result of specific interactions between promoter control elements, multiple transcription initiation factors (minimally + ÷ il i¸¸¸: ii : i!i ¸:¸ TFIIIB and TFIIIC), and RNA Pol III. Concomitant with progress in defining the structure and function of the ! RT mammalian initiation factors (for review, see Wang and T2 Roeder 1996), we have turned our attention to an analy- e T1 sis of RNA Pol III itself. Here, we report a simple affinity v,,---. purification method that has allowed us to define a 16- subunit human RNA Pol III complex. Further studies

1 2 3 4 5 6 7 i i~!~i ¸ ;i ~i i : :ii have revealed the presence of posttranslational modifications in some subunits, the reversible dissociation of RNA Pol III into an initiation-defective core RNA Pol III and a three-subunit subcomplex that interacts with the initiation factor TFIIIB. These studies have revealed the evolutionary conservation, from yeast to human, of RNA Pol III subunit structures and recruitment mecha- nisms, and provide a further basis for more detailed studies of the evolutionarily more diverse RNA Pol III factors in human. 1 2 3 4 5 6 Figure 7. The hRPC32-hRPC39-hRPC62 subcomplex restores Structure and reversible dissociation of human RNA promoter-dependent transcription initiation to the core RNA PoI III Pol III that is otherwise competent for transcription elongation and termination. (A) Transcription of the natural VA1 template. The human RNA Pol III described here forms a tight (B) transcription of the C-tailed template VA1. Reactions con- complex of 16 subunits that can withstand partially de- tained the tailed template and 48 fmoles of the RNA Pol III naturing conditions (2 M urea or 0.3% Sarkosyl), and is components indicated at the top of the lanes: Pol III, intact RNA similar in complexity to the 16-subunit yeast RNA Pol polymerase III; core Pol III, core RNA polymerase III (sucrose III (Gabrielsen and Sentenac 1991). Under more harsh gradient fraction 4); SG #9, natural subcomplex (sucrose gradi- conditions (4 a urea or 0.5% Sarkosyl) or during sucrose ent fraction 9); r32/39/62, subcomplex of recombinant hRPC32, gradient sedimentation, however, RNA Pol III is disso- hRPC39, and hRPC62. ciated into a stable core and a subcomplex comprised mainly of three subunits (hRPC32, hRPC39, and hRPC62) that are conserved in yeast (yRPC31, yRPC34, lack of other dissociated subunits (hRPC38 or possibly and yRPC68, respectively) and specific to RNA Pol III. the two smallest RNA Pol III subunits, see above) in the These results are consistent with the reported dissocia- recombinant complex. As shown as Figure 7B, however, tion of the corresponding subunits from yeast RNA Pol neither the natural nor the recombinant subcomplex had III during native polyacrylamide gel electrophoresis (Va- any major effects on elongation or termination on the lenzuela et al. 1976) and during ion exchange chroma- tailed template when added alone (lanes 3, 5) or together tography at high temperature of a mutant form of the (lanes 4, 6) with the core RNA Pol III. The addition of enzyme (Werner et al. 1992). Whereas the presence of a these subcomplexes also did not change the overall tran- stable subcomplex of yeast subunits was inferred from scription level of the core enzyme (Fig. 7B, lanes 4 and 6 both genetic and two-hybrid interaction assays (Werner vs. lane 2). However, the distribution of transcripts pro- et al. 1993), our studies provide the first direct evidence duced by the natural subcomplex plus the core RNA Pol for such a stable complex. The physically separated core III (28% T1, 42% T2, and 30% RT) was closer to that RNA Pol III lacking the three RNA Pol III-specific sub- generated by intact RNA Pol III (17% T1, 58% T2, and units effects normal levels of elongation and termination 25% RT) than that of the core enzyme alone (24% T1, (following nonspecific promoter-independent initiation 35% T2, and 42% RT). In toto, these results strongly on a tailed template) but is defective in promoter-depen- suggest that the three subunit subcomplex plays a direct dent transcription in a system reconstituted with general role in accurate transcription initiation, most likely initiation factors TFIIIB and TFIIIC. We also show the through the indicated physical interactions with TFIIIB. ability to reconstitute a fully functional RNA Pol III Nonetheless, they leave open the possibility that the re- from the residual core RNA Pol III and either the natural combinant subcomplex may not be identical to the natu- subcomplex or a subcomplex reconstituted with recom-

1322 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three RNA Pol III subunits binant hRPC32, hRPC39, and hRPC62. This provides 1993; Khoo et al. 1994), however, we show that the clear evidence that the loss of promoter-dependent tran- RPC39 subunit of human RNA Pol III (homolog of scription is caused by dissociation of the three cloned yRPC34), but not hRPC32 or hRPC62, also interacts di- RNA Pol III-specific subunits and not to dissociation of rectly with human TFIIIB90 (homolog of yTFIIIB70). another subunit (e.g., hRPC38 or one of the two smallest Thus, there is an evolutionary conservation of RNA Pol subunits) or another associated initiation factor present III-TFIIIB interactions that are likely involved in RNA in substoichiometric levels. Pol III recruitment by promoter-bound TFIIIB. Our dem- Similar phenomena of selective dissociation of RNA onstration of a novel interaction between hRPC39 and polymerase-specific subunits have also been observed for TBP, however, suggests that secondary contacts of RNA Pol I and Pol II. In the case of yeast RNA Pol I, the hRPC39 with TFIIIB may also be important for recruit- selective dissociation of subunits yRPA43 and yR- ment of human RNA Pol III. Possibly related, the ability PABC23 by disruption of the yRPA14 gene results in an of a Gal4-yTFIIIB70 fusion protein to serve as an activa- enzyme that is inactive in a nonspecific in vitro tran- tor, but not as an initiation factor, in promoting transcription assay (Smid et al. 1995). In the case of yeast scription from templates containing a Gal4 binding site, RNA Pol II, subunits yRPB32 and yRPB17 appear to form but lacking the B box, suggests that efficient initiation a subcomplex that can be dissociated under partially de- complex formation might also involve RNA Pol III con- naturing conditions (Edwards et al. 1991). The enzyme tacts with another component of TFIIIB or with TFIIIC lacking the yRPB32-yRPB17 subcomplex is fully active (Marsolier et al. 1994). in promoter-independent initiation and elongation in As discussed in the preceding section, there is indirect vitro, but, in activator-stimulated promoter-dependent evidence from studies in yeast and direct evidence from transcription, shows a reduced activity that is partially studies presented here that the TFIIIB-interacting sub- restored by the purified subcomplex. These results sug- unit yRPC34/hRPC39 is part of a stable subcomplex that gest a role for the yeast yRPB32-yRPB17 subcomplex in also contains yRPC31/hRPC32 and yRPC82/hRPC62. the efficiency of promoter-dependent transcription. The Our demonstration of the selective loss of promoter-de- fact that yRPB32 and yRPB 17 are not essential for yeast pendent initiation (but not elongation or termination) viability under some conditions suggests either that this capabilities upon dissociation of this subcomplex, and role is not critical or that there is a compensation by the regain of this function following readdition of a sub- other factors. Possibly related, a different yeast RNA Pol complex of the three recombinant subunits, provide II-specific subunit (yRPB13) was reported to be necessary strong evidence for the proposed function of hRPC39/ for accurate start-site selection (Furter-Graves et al. yRPC34. Although the functions of the other two sub- 1994; Hull et al. 1995). units of the subcomplex are not clear, they could be in- On the basis of these observations, it is reasonable to volved either in recruitment and stabilization of the sub- propose that RNA Pol I, Pol II, and Pol III are each com- complex within RNA Pol III or in stimulating catalytic posed of a core enzyme containing (1) the two largest functions of other subunits during initiation. Relevant to subunits, each containing sequences highly conserved these possibilities, a conditional mutation in yRPC31 between RNA Pol I, Pol II, and Pol III, (2) either yRPB45 was reported to selectively inhibit specific transcription or the related yRPAC40 and yRPAC19 (or their human initiation, but not elongation and termination, and the counterparts), and (3) the five common subunits. The conditional phenotype of this yRPC31 mutant was sup- RNA polymerase type-specific subunits, including those pressed by overexpression of the yRPC 160 subunit (Thu- described here, may then assemble on the core enzymes illier et al. 1995). The later observation, suggesting direct and dissociate under specific conditions. interactions between yRPC31 and yRPC160, is consistent with the demonstration that a conditional mutation in yRPC160 can facilitate chromatographic dissociation Role of the hRPC32-hRPC39-hRPC62 subcomplex in of the three-subunit subcomplex (Werner et al. 1992). Of transcription initiation by human RNA Pol III note, our ability to reassemble an initiation competent In both yeast and human, transcription initiation on sub- RNA Pol III with recombinant hRPC32, hRPC39, and class 3 genes (including tRNA and VA RNA genes) in- hRPC62 will allow us to directly test the function of volves recognition of core promoter elements (A and B these individual proteins and the involvement of specific boxes) by TFIIIC, followed by the sequential recruitment domains. of TFIIIB and RNA Pol III (for review, see Gabrielsen and Sentenac 1991). In yeast, residual TFIIIB-DNA com- Evolutionary conservation of RNA Pol III subunits plexes from which TFIIIC has been stripped were shown to direct multiple rounds of transcription initiation by Comparative sequence analyses revealed that the RNA RNA Pol III, indicating a role for direct interactions be- Pol III-specific subunits are not as well conserved between RNA Pol III and TFIIIB in this process (Kassavetis tween yeast and human (<30% identity) (Fig. 4; Ittmann et al. 1990). It has not been possible to show a similar et al. 1993) as are the RNA Pol II-specific subunits (30%- phenomenon in the human system. Consistent with the 70% identity) (Young 1991; McKune et al. 1995). Con- direct demonstration of an interaction of the yeast sistent with this, two of the human RNA Pol II-specific RPC34 subunit of RNA Pol III with the TFIIB-related subunits can fully substitute for their yeast counterparts component (yTFIIIB70) of yeast TFIIIB (Werner et al. in yeast viability tests (McKune et al. 1995), whereas

GENES & DEVELOPMENT 1323 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Wang and Roeder neither human RPC53 (Mann et al. 1992) nor human Purification of transcription factors RPC39 (J.C. Andrau, Z. Wang, R. Roeder, and M. Werner, TFIIIB (HPLC SP-SPW column fraction), TFIIIC1 (FPLC Mono Q unpubl.) can do so. In addition, whereas at least some of fraction), and TFIIIC2 (B box oligonucleotide affinity column the yeast TFIIIB and TFIIIC subunits have counterparts fractionl were purified as described previously (Wang and in the corresponding human factors (Wang and Roeder Roeder 1996). 1996; Y. Hsieh, Z. Wang, and R. Roeder, unpubl.), the overall sequence conservation (<30% identity) is gener- In vitro transcription assay ally less than that (30%-70% identity) reported for the RNA Pol II general initiation factors {for review, see To make the tailed template, 25 ~g of pVA1 (Wang and Roeder Roeder 1996b). However, while there is virtually a 1:1 1995) was cleaved at the unique PstI site and added to a 100qal reaction mixture containing 0.2 M potassium cacodylate, 1 mM correspondence between the general RNA Pol II initia- CoC12, 2 mM ~-mercaptoethanol, and 50 mM dCTP. After addition factors from yeast and human (Roeder 1996b), there tion of 31.25 units of terminal deoxynucleotidyltransferase are substantial variations between the subunits of yeast (Boehringer Mannheim), the mixture was incubated at 37°C for and human RNA Pol III factors. Thus, whereas yeast 45 min and the reaction was terminated by addition of 100 ~l of TFIIIC is composed of a six-subunit complex, human SETS (150 mM NaC1, 5 mM EDTA, 0.5% SDS, 50 mM Tris-HC1 TFIIIC is composed of a five-subunit TFIIIC2 complex at pH 8.0). After the sample was extracted twice with phenol- with only two yeast TFIIIC-related subunits (Lagna et al. chloroform (1:1), precipitated with ethanol, and washed with 1994; L'Etoile et al. 1994; Sinn et al. 1995; Y. Hsieh, R. 70% ethanol, the template was cleaved at the unique EcoRI site. Kovelman, and R. Roeder, unpubl.) and a large, but less The tailed template, comprised of a 260-bp PstI-EcoRI fragment well-characterized, TFIIIC1 complex (Wang and Roeder with an average tail length of 70-80 dCTP residues, was purified by agarose gel electrophoresis. This fragment contains the VA1 1996). Along with variable factor requirements for U6/ RNA gene positioned downstream of the tailed end, so that 7SK RNA versus 5S RNA/tRNA gene transcription in RNA Pol III initiating at the tailed end transcribes the coding human (Wang and Roeder 1995; Yoon et al. 1995; Mital strand of the VA1 RNA gene. et al. 1996; Teichman and Seifart 1996) but not in yeast For promoter-dependent transcription, reaction mixtures in a (Burnol et al. 1993), these observations further indicate final volume of 25 ~1 contained 200 ng of pVA1 template, 60 mM that the RNA Pol III general transcriptional machinery is KC1, 6 mM MgC12, 2 mM DTT, 8% glycerol, 10 mM HEPES at pH evolutionarily less conserved than the RNA Pol II gen- 7.9, 0.6 mM ATP, CTP, and UTP, 0.025 mM GTP with 2.5 ~Ci eral transcriptional machinery. Hence, one may expect of [~-32P]GTP, and RNA Pol III and transcription factors as in- both conserved (as shown here for the hRPC39 interac- dicated. Reactions were allowed to proceed for 60 min before tions) and nonconserved interactions of human RNA Pol being stopped by addition of 25 ~1 of the stop solution (0.2 M NaC1, 30 mM EDTA, 1% SDS, 100 mg/ml yeast tRNA1. Labeled III subunits with the general transcriptional machinery. RNA products were extracted with 50 ~1 of phenol chloroform (1:1 ), precipitated with ethanol, and resolved on a 8 % polyacryl- Materials and methods amide/7 M urea gel. RNA bands were visualized by autoradiog- raphy or quantified with a PhosphorImager (Molecular Dynam- Cell line preparation and fractionation ics). For tailed template assays, conditions were the same except A retrovirus-mediated gene transfer method lChiang et al. 19931 that 50 ng of the tailed template (in place of pVA1) was incu- was used to establish a stably transfected HeLa $3 cell line bated with RNA Pol III at 25°C for 15 min before addition of (BN51) expressing a FLAG-tagged 53-kD subunit (Ittmann et al. ribonucleoside triphosphates and salts. The reaction mixture 1993). Nuclear extract and S 100 fractions were prepared as de- was then incubated for an additional 30 min at 25°C before scribed (Dignam et al. 1983). processing.

Immu~opurification of human RNA Pol III Gene cloning and sequence a~alysis One milliliter of the S100 Pll 0.35 M KC1 fraction (6 mg/ml The immunopurified RNA Pol III from 600 ml of the S100 P11 protein) was adjusted to 300 mM KC1-0.1% NP-40 by addition of 0.35 fraction (3600 mg protein) was separated on a 4%-18% 3 M KC1 and 10% NP-40, and incubated with 20 ~al of anti-FLAG gradient SDS-polyacrylamide gel. The proteins were transferred M2-agarose beads (IBI/Kodak) at 4°C for 3-6 hr by rotation. onto a PVDF membrane and stained with Ponseau S. Internal After five 1-ml washes with BC300-0.1% NP-40, proteins were peptides were microsequenced as described (Wang and Roeder eluted from the beads by incubation at 4°C for 30 rain with 20 1995). Two peptide sequences from hRPC62 matched an EST ~1 of BC100-0.1% NP-40 plus 0.2 mg/ml FLAG peptide. In some sequence (GenBank ID 19022). cDNA clones encoding hRPC32 cases, the immobilized RNA Pol III (on M2-agarose beads) was and hRPC39 were obtained by screening a HeLa cDNA library washed twice (1 ml each time) with 2 or 4 M urea in BC100 with probes derived from the corresponding peptide sequences. before elution with FLAG peptide. DNA sequences were obtained by the dideoxynucleotide chain- termination method (U. S. Biochemical). Protein sequences were aligned with GCG program (Genetics Computer Group, Sucrose gradient sedimentation Inc.). Immunopurified human RNA Pol III (0.2 ml) was loaded onto a 4.0-ml, 5%-20% sucrose gradient containing 0.5 M KC1 in BC Protein expression and subcomplex assembly buffer (20 mM HEPES at pH 7.9, 20% glycerol, 0.5 mM EDTA, 1 mM DTT, 0.5 mM PMSF) and centrifuged for 24 hr at 4°C, 56,000 Histidine- or FLAG-tagged hRPC32, hRPC39, and hRPC62 were rpm in an SW60 rotor (Beckman). Fractions (0.2 ml) were col- expressed in bacteria through the pET15b vector (Novagen). lected from the bottom of the tube. Histidine-tagged hRPC32 and hRPC62 were also expressed in

1324 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three RNA Pol III subunits

Sf9 cells via the baculovirus pVAL1393 vector (PharMingen t. and cofactors involved in transcription by RNA polymerases Recombinant proteins were purified through Ni ~÷ agarose (Qia- II and III. EMBO J. 12: 2749-2762. gen) or M2-agarose (IBI/Kodak). The three-subunit subcomplex Chiannilkulchai, N., R. Stalder, M. Riva, C. Carles, M. Werner, was assembled by incubation of 20 ~ag of FLAG-tagged hRPC39 and A. Sentenac. 1992. RPC82 encodes the highly conserved, immobilized on 50 ~al of the M2-agarose beads with Sf9 cell third-largest subunit of RNA polymerase C (III) from Sac- lysate containing 200 ~g of the His-tagged hRPC32 and 200 ~g charomyces cerevisiae. Mol. Cell. Biol. 12: 4433-4440. of the His-tagged hRPC62 at 4°C for 1 hr in BC300-0.1% NP-40. Dieci, G., S. Hermann-Le Denmat, E. Lukhtanov, P. Thuriaux, The beads were washed five times ( 1 ml each time) with BC300- M. Werner, and A. Sentenac. 1995. A universally conserved 0.1% NP-40, and the subcomplex was eluted with 50 ~1 of region of the largest subunit participates in the active site of BC100-0.2 mg/ml of FLAG peptide. The GST-fusion full- RNA Pol III. EMBO J. 14: 3766-3776. length and truncated hRPC39 clones were constructed by sub- Dignam, J.D., R.M. Lebovitz, and R.G. Roeder. 1983. Accurate cloning corresponding PCR fragments into the pGEM-7(+) vec- transcription initiation by RNA polymerase II in a soluble tor and expressed in bacteria (Wang and Roeder 1995). extract from isolated mammalian nuclei. Nucleic Acids Res. ll: 1475-1489. Edwards, A.M., C.M. Kane, R.A. Young, and R.D. Kornberg. Protein-protein interaction assays 1991. Two dissociable subunits of yeast RNA polymerase II Recombinant GST fusion protein (1 ~g) was immobilized on 20 stimulate the initiation of transcription at a promoter in ~1 of glutathione agarose beads (Pharmacia) and incubated with vitro. J. Biol. Chem. 266: 71-75. bacterial or Sf9 cell lysate containing 0.1 ~g of the other recom- Furter-Graves, E.M., B.D. Hall, and R. Furter. 1994. Role of a binant protein in a final volume of 100 ~l containing 150 mM small RNA pol II subunit in TATA to transcription start site KC1 and 0.05% NP-40. The beads were washed five times (1 ml spacing. Nucleic Acids Res. 22: 4932-4936. each time) with the incubation buffer, boiled in 20 lal of the SDS Gabrielsen, O.S. and A. Sentenac. 1991. RNA polymerase III (C) gel sample buffer and analyzed by western blot analysis. and its transcription factors. Trends Biochem. Sci. 16" 412- 416. Hull, M.W., K. McKune, and N.A. Woychik. 1995. RNA poly- Antibody preparation merase II subunit RPB9 is required for accurate start site Polyclonal antibodies were generated by injection of rabbits selection. Genes & Dev. 9:481-490. with bacterially expressed and affinity-purified (Ni 2+ agarose) Ittmann, M., J. Ali, A. Greco, and C. Basilico. 1993. The gene full-length hRPC20, hRPC39, and hRPC53 and fragments of complementing a temperature-sensitive cell cycle mutant of hRPC82 (residues 1-427) and hTFIIIB90 (residues 262-486). BHK cells is the human homologue of the yeast RPC53 gene, Full-length cDNA clones for hRPC82 and hRPC20 were ob- which encodes a subunit of RNA polymerase C (III). Cell tained on the basis of direct peptide sequence analyses. Anti- Growth Differ. 4:503-511. bodies against hRPC82, hRPC53, hRPC39, and hRPC20 reacted Kassavetis, G.A., B.R. Braun, L.H. Nguyen, and E.P. Geidus- specifically with single polypeptides of the expected size in pu- chek. 1990. S. cerevisiae TFIIIB is the transcription initiation rified RNA Pol III and showed no reactivity with any polypep- factor proper of RNA polymerase III, while TFIIIA and tides in purified TFIIIB and TFIIIC (data not shown). TFIIIC are assembly factors. Cell 60" 235-245. Keniry, M.A., D.L. Banville, P.M. Simmonds, and R. Sharer. 1993. Nuclear magnetic resonace comparison of the binding Acknowledgments sites of Mithramycin and Chromomycin on the self-comple- mentary oligonucleotide d(ACCCGGGT)2. J. Mol. Biol. We thank J. Fu for help in DNA sequence analysis and GST- 231: 753-767. fusion protein pull-down assays and M.J. Ittmann for the Khoo, B., B. Brophy, and S.P. Jackson. 1994. Conserved func- hRPC53 cDNA clone. We also thank Y. Tao for helpful discus- tional domains of the RNA polymerase III general transcrip- sions, data bank searching and protein sequence alignments. tion factor BRF. Genes & Dev. 8: 2879-2890. Protein sequence analysis was provided by the Rockefeller Uni- Lagna, G., R. Kovelman, J. Sukegawa, and R.G. Roeder. 1994. versity Protein Sequencing Facility, which is supported in part Cloning and characterization of an evolutionarily divergent by National Institutes of Health (NIH) shared instrumentation DNA-binding subunit of mammalian TFIIIC. Mol. Cell. Biol. grants and funds provided by the U.S. Army and Navy for pur- 14: 3053-3064. chase of equipment. This work was supported by a grant Lalo, D., C. Carles, A. Sentenac, and P. Thuriaux. 1993. Inter- (CA42567) from the NIH to R.G.R. actions between three common subunits of yeast RNA poly- The publication costs of this article were defrayed in part by merases I and III. Proc. Natl. Acad. Sci. 90" 5524-5528. payment of page charges. This article must therefore be hereby L'Etoile, M.D., M.L. Fahnestock, Y. Shen, R. Aebersold, and A. marked "advertisement" in acccordance with 18 USC section Berk. 1994. Human transcription factor IIIC box B binding 1734 solely to indicate this fact. subunit. Proc. Natl. Acad. Sci. 91: 1652-1656. Mann, C., J.Y. Miouin, N. Chiannilkulchi, I. Teich, J.M. Buhler, References and A. Sentenac. 1992. RPC53 encodes a subunit of Saccha- romyces cerevisiae RNA polymerase C (III) whose inactiva- Bartholomew, B., D. Durkovich, G.A. Kassavetis, and E.P. Gei- tion leads to a predominantly G1 arrest. Mol. Cell. Biol. duschek. 1993. Orientation and topography of RNA Pol III in 12:4314-4326. transcription complexes. MoI. Cell. Biol. 13: 942-952. Marsolier, M.C., N. Chaussivert, O. Lefebvre, C. Conesa, M. Burnol, A.F., F. Margottin, J. Huet, G. Almouzni, M.N. Prioteau, Werner, and A. Sentenac. 1994. Directing transcription of an M. Mechali, and A. Sentenac. 1993. TFIIIC relieves repres- RNA polymerase III gene via GAL4 sites. Proc. Natl. Acad. sion of U6 snRNA transcription by chromatin. Nature Sci. 91:11938-11942. 362: 475-477. McKune, K., P.A. Moore, M.W. Hull, and N.A. Woychik. 1995. Chiang, C.M., H. Ge, Z. Wang, A. Hoffmann, and R.G. Roeder. Six human RNA polymerase subunits functionally substitute 1993. Unique TATA-binding protein-containing complexes for their yeast counterparts. Mol. Cell. Biol. 15" 6895-6900.

GENES & DEVELOPMENT 1325 Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Wang and Roeder

Memet, S., W. Saurin, and A. Sentenac. 1988. RNA polymerases (PTF) is a multisubunit complex required for transcription of B and C are more closely related to each other than to RNA both RNA polymerase II- and RNA polymerase III-depen- polymerase A. J. Biol. Chem. 263: 10048-10051. dent small nuclear RNA genes. Mol. Cell. Biol. 15:2019- Mital, R., R. Kobayashi, and N. Hernandez. 1996. RNA poly- 2027. merase III transcription from the human U6 and adenovirus Young, R.A. 1991. RNA polymerase II. Annu. Rev. Biochem. type 2 VA1 promoters has different requirements for human 60: 689-715. BRF, a subunit of human TFIIIB. Mol. Cell. Biol. 16: 7031- 7042. Mosrin, C., M. Riva, M. Beltrame, E. Cassar, A. Sentenac, and P. Thuriaux. 1990. The RPC31 gene of Saccharomyces cerevisiae encodes a subunit of RNA polymerase C (III) with an acidic tail. Mol. Cell. Biol. 10: 4737-4743. Roeder, R.G. 1996a. Nuclear RNA polymerases: Role of general initiation factors and cofactors in eukaryotic transcription. Methods Enzymol. 273: 165-171. • 1996b. The role of general initiation factors in transcription by RNA polymerase II. Trends Biochem. Sci. 21: 327- 335. Sadhale, P.P., and N.A. Woychik. 1994. C25, an essential RNA polymerase III subunit related to the RNA polymerase II subunit RPB7. Mol. Cell. Biol. 14: 6164-6170. Sinn, E, Z. Wang, R. Kovelman, and R.G. Roeder. 1995. Cloning and characterization of a TFIIIC2 subunit (TFIIICB) whose presence correlates with activation of RNA polymerase III- mediated transcription by adenovirus E1A expression and serum factors. Genes & Dev. 9: 675-685. Smid, A., M. Riva, F. Bouet, and A. Sentenac. 1995. The association of three subunits with yeast RNA polymerase is sta- bilized by A14. J. Biol. Chem. 270: 13534-13540. Stettler, S., S. Mariotte, M. Riva, A. Sentenac, and P. Thuriaux. 1992. An essential and specific subunit of RNA polymerase III (C) is encoded by gene RPC34 in Saccharomyces cerevisiae. J. Biol. Chem. 267: 21390-21395. Teichmann, M. and K.M. Seifart. 1995. Physical separation of two different forms of hTFIIIB activity in the transcription of the U6 or the VA1 gene in vitro. EMBO J. 14: 5974-5983. Thuillier, V., S. Stettler, A. Sentenac, P. Thuriaux, and M. Werner. 1995. A mutation in the RPC31 subunit of Saccha- romyces cerevisiae RNA polymerase III affects transcription initiation. EMBO J. 14: 351-359. Treich, I., C. Carles, A. Sentenac, and M. Riva. 1992. Determi- nation of lysine residues affinity labeled in the active site of yeast RNA polymerase II (B) by mutagenesis. Nucleic Acids Res. 20: 4721-4725. Valenzuela, P., G.L. Hager, F. Weinberg, and w.J. Rutter. 1976. Molecular structure of yeast RNA polymerase III: Demon- stration of the tripartite transcriptive system in lower eu- karyotes. Proc. Natl. Acad. Sci. 73: 1024-1028. Wang, Z. and R.G. Roeder. 1995. Structure and function of a human transcription factor TFIIIB subunit that is evolutionarily conserved and contains both TFIIB- and high-mobility- group protein 2-related domains. Proc. Natl. Acad. Sci. 92: 7026-7030. ~. 1996. TFIIIC1 acts through a downstream region to sta- bilize TFIIIC2 binding to RNA polymerase III promoters. Mol. Cell. Biol. 16: 6841-6850. Werner, M., S. Hermann-Le Denmat, I. Treich, A. Sentenac, and P. Thuriaux. 1992. Effect of mutations in a zinc-binding domain of yeast RNA polymerase C (III) on enzyme function and subunit association. Mol. Cell. Biol. 12: 1087-1095. Werner, M., N. Chaussivert, I.M. Willis, and A. Sentenac. 1993. Interaction between a complex of RNA polymerase 1II subunits and the 70-kDa component of transcription factor IIIB. J. Biol. Chem. 268: 20721-20724. Yoon, J., S. Murphy, L. Bai, Z. Wang, and R.G. Roeder. 1995. Proximal sequence element-binding transcription factor

1326 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press

Three human RNA polymerase III-specific subunits form a subcomplex with a selective function in specific transcription initiation.

Z Wang and R G Roeder

Genes Dev. 1997, 11: Access the most recent version at doi:10.1101/gad.11.10.1315

References This article cites 39 articles, 25 of which can be accessed free at: http://genesdev.cshlp.org/content/11/10/1315.full.html#ref-list-1

License

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top Service right corner of the article or click here.