Interaction between Yeast RNA Polymerase III and Transcription Factor TFIIIC via ABC10α and τ131 Subunits*

Yeast TFIIIC mediates transcription of class III genes by promoting the assembly of a stable TFIIIB-DNA complex that is sufficient for RNA polymerase III recruitment and function. Unexpectedly, we found an interaction in vivo and in vitro between the TFIIIB-recruiting subunit of TFIIIC, τ131, and ABC10α, a small essential subunit common to the three forms of nuclear RNA polymerases. This interaction was mapped to the C-terminal region of ABC10α. A thermosensitive mutation in the C terminus region of ABC10α (rpc10-30) was found to be selectively suppressed by overexpression of a mutant form of τ131 (τ131-ΔTPR2) that lacks the second TPR repeat. Remarkably, therpc10-30 mutation weakened the ABC10α-τ131 interaction, and the suppressive mutation, τ131-ΔTPR2 increased the interaction between the two proteins in the two-hybrid assay. These results point to the potential importance of a functional contact between TFIIIC and RNA polymerase III.

In eukaryotic cells, RNA polymerase (Pol) 1 III is responsible for the transcription of genes encoding tRNAs, 5 S RNA, U6 RNA, and a number of small RNA species. In tRNA genes (tDNA), the internal promoter elements, the A and B blocks, are recognized by TFIIIC. DNA-bound TFIIIC then directs the assembly of TFIIIB that, in turn, is sufficient to recruit RNA polymerase III for multiple transcription cycles. The transcription of all yeast class III genes is a variation of this scheme (1).
S. cerevisiae RNA polymerase (Pol) III is a multisubunit complex comprising 17 polypeptides ranging from 162 to 7.7 kDa (18), five of which, ABC27, ABC23, ABC14.5, ABC10␣, and ABC10␤, are shared with Pol I and II. A labile triad of subunits, C34, C31, and C82, has been implicated in the recruitment of Pol III and in transcription initiation (19). A mutation in C31 subunit was found to specifically affect transcription initiation but not the catalytic properties of the enzyme (20). C34 was found to be localized the furthest upstream on tDNA in initiation complexes (21,22), and analysis of mutant Pol III showed that mutations in C34 that decreased its interaction with TFIIIB70/BRF1 affected Pol III recruitment and open complex formation (23). This triad of subunits has its counterpart in human Pol III. These subunits form a subcomplex that is required for transcription initiation (24). One (hRPC39) of these subunits, homologous to yC34, interacts physically with two components of hTFIIIB (hTBP and hTFIIIB90). More recently, a new essential subunit of yeast Pol III, C17, was also found to interact with C31 and TFIIIB70/BRF1 thus adding a new linkage to the TFIIIB⅐Pol III connection. 2 These findings suggest that the recruitment, correct positioning, and activation of Pol III is mediated by multiple contacts between the enzyme and TFIIIB components.
In this work we report genetic and biochemical evidence in favor of a direct contact between yeast Pol III and the assembly factor TFIIIC, namely between the common subunit ABC10␣ and the TFIIIB-assembling subunit of TFIIIC, 131. Supporting initial two-hybrid experiments, recombinant ABC10␣ was found to interact in vitro with 131. A thermosensitive mutation in the conserved C-terminal region of ABC10␣, that weakens this interaction, can be rescued by overexpression of a variant form of 131. These data suggest the existence of functional interactions between TFIIIC and Pol III.

EXPERIMENTAL PROCEDURES
Yeast Strains, Media, and Genetic Methods-The yeast strains used in this study were constructed by genetic techniques based on transformation of lithium acetate-treated cells with standard media and growth conditions (25). Yeast strains are as follows: YLR-01 (Mat a ura3-52 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Supported by fellowships from the French Ministère de l'Education Nationale, de la Recherche et de la Technologie, and from the Association pour la Recherche contre le Cancer.
§ Supported by a fellowship from the Istituto Pasteur Fondazione Cenci-Bolognetti.
Preparation of Recombinant ABC10␣ Protein-pRSETthio/RPC10 (kindly provided by J.-M. Buhler) was generated by subcloning the entire RPC10 coding sequence (obtained by polymerase chain reaction from genomic DNA) in the T7 polymerase expression vector pRSETA (Invitrogen) at a BamHI site. This construct produced a ABC10␣thioredoxin fusion protein, tagged with six histidines and T7-Tag TM at the N terminus of rABC10␣. Formation of inclusion bodies in the Escherichia coli cytoplasm was prevented by the thioredoxin moiety. E. coli strain BL21(DE3)(pLysS) was transformed with pRSETthio/ RPC10, and cultures were grown at 37°C up to an A 600 of 0.4. Then isopropyl-␤-thiogalactopyranoside was added (0.5 mM final concentration), and induced cultures were grown for 2 h at 30°C. rABC10␣ was purified under native conditions by chromatography on Ni 2ϩ -nitrilotriacetic acid-agarose as specified by the manufacturer (Qiagen) with minor modifications as follows. Bacteria were harvested by centrifugation and resuspended in binding buffer (5 mM imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.9, protease inhibitors (Roche Molecular Biochemicals)) and lysed by heat shock and treatment with lysozyme (0.1 mg/ml final). The lysate was centrifuged, and the protein extract was added to Ni 2ϩ -nitrilotriacetic acid-agarose beads equilibrated in the binding buffer. After 1 h at 4°C, the flow-through fraction was removed, and the resin was washed with binding buffer containing 60 mM imidazole. Bound proteins were eluted stepwise with elution buffer (1 M imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.5). Samples of eluates were analyzed by Western blotting with anti-T7-Tag TM antibodies (Novagen).
The BamHI-BamHI fragment of pACT⌬TPR2 (15) was cloned into pET28a (Novagen) to produce the mutant 35 S-131-⌬TPR2 protein (lacking amino acids 162-195) (15). These expression plasmids, pET131 and pET⌬TPR2, were linearized using AseI and XhoI, respectively. The genes were transcribed and translated in vitro with TNT Coupled Wheat Germ Extract Systems (Promega) in the presence of [ 35 S]methionine. Expression of 35 S-131 (150,000 cpm/ml) and 35 S-131-⌬TPR2 (100,000 cpm/ml) was verified by SDS-PAGE. Partially purified rABC10␣-thioredoxin fusion, purified recombinant thioredoxin (Promega), and a control protein extract from E. coli were subjected to SDS-PAGE and blotted onto nitrocellulose for far Western analysis (29). The filter-bound proteins were subjected to a denaturation/renaturation treatment according to the method of Papavassiliou and Bohmann (30). To visualize the binding of 35 S-131, the 35 S-labeled background had to be reduced by addition of 5% low fat milk to the probe. This process was not necessary when probing with 35 S-131-⌬TPR2 due to a stronger interaction of the mutant protein with ABC10␣. Full size rABC10␣ was revealed by anti-T7-Tag TM antibodies. Immune complexes were visualized using the ECL chemiluminescence kit (Amersham Pharmacia Biotech), and the bound 35 S-labeled polypeptides were revealed by autoradiography.
Sequence Searches-Sequence data for Candida albicans was obtained from the Stanford DNA Sequencing and Technology Center website. Sequencing of C. albicans was accomplished with the support of the NIDR and the Burroughs Wellcome Fund. The sequence of the ABC10␣ C. albicans ortholog was identified in the unpublished sequence con4 -2986 using the NCBI Blast server and the S. cerevisiae sequence as entry. The sequence of ABC10␣ ortholog in Arabidopsis thaliana has been disclosed using TblastN 2.0 (32) run on the NCBI Blast server and non-redundant DNA data base with the human ABC10␣ sequence as entry. This protein sequence has been tentatively reconstituted from genomic data (2 introns are introduced) (Gen-Bank TM accession number AB010072). The sequence of P. abyssi was obtained from the Genoscope web site. ABC10␣ orthologs of Archaeoglobus fulgidus (33), Pyrococcus horikoshii (34), P. abyssi, and Methanococcus jannaschii (35) and were disclosed using TblastN 2.0 run on the same server and data base as indicated above. The TPR plots in Fig.  5 display a function that indicates the fit to a TPR consensus sequence matrix extracted from 200 TPR units of S. cerevisiae proteins. 3 Peaks are localized at the center of the TPR units.

RESULTS
131 Interacts with the Shared RNA Polymerase Subunit ABC10␣-The interaction of 131 with subunits of the yeast RNA Pol III was explored using the two-hybrid assay. The 131 gene (TFC4/YGR047c), fused in frame with the GAL4 activation domain was challenged with the complementary fusions of 12 Pol III subunits, C160, C128, C82, C53, AC40, C31, AC19, ABC27, ABC23, ABC14.5, ABC10␣, and ABC10␤, fused with GAL4 DNA-binding domain. The C34 subunit was not tested since it behaves, by itself, as a strong transcriptional activator (19,36). The C25 subunit (YKL1/RPC7) (37) was not assayed. Two additional subunits, C17 4 and C11, 5 have been assayed independently and gave a negative two-hybrid interaction with 131. Of all the Pol III subunits tested with 131, only ABC10␣ (RPC10/YHR143wa) (38) gave a positive interaction response (Fig. 1A). The ␤-galactosidase activity level obtained for this interaction was similar to that observed with the 131-TFIIIB70/BRF1 interaction (Fig. 1A, see lanes 4 and 6) (15). An interaction was previously noted between C53 and a fragment of 131 (39). This interaction could not be detected using the entire 131 protein. Other components of the Pol III transcription system, TFIIIA, 138 and TBP, were also tested and gave negative results (not shown).
A number of deletion mutants of 131 were assayed in order to map the interaction domain. As shown in Fig. 1B, the ABC10␣-131 interaction could not be restricted to a given subdomain of 131. Interestingly, however, some deletion mutant forms of 131, 131-⌬TPR1, 131-⌬TPR2, and 131-⌬TPR3 (15) were found to interact more efficiently with ABC10␣ than the wild-type 131 protein. The ␤-galactosidase activity generated by the 131-⌬TPR2-ABC10␣ interaction was increased 3-fold relative to ABC10␣-131. The interaction of the same collection of 131 mutants with TFIIIB70/BRF1 (15) and TFIIIB90/BЉ (16) has been previously described. The results, summarized in Fig. 1B, show that the interaction of 131 variants with the three proteins was quantitatively and qualita-3 H. Dumay and C. Marck, unpublished observations. 4 M. L. Ferri, personal communication. 5 S. Chédin, personal communication.
tively different. First, the N-terminal part of 131 interacted specifically with TFIIIB70/BRF1. In contrast, the deletion of the first, second, or third TPR units, which increased the interaction with ABC10␣, decreased or did not affect the interaction with TFIIIB70/BRF1. Similarly the ⌬TPR1 and ⌬TPR3 mutations abrogated and decreased, respectively, the interaction of 131 with TFIIIB90/BЉ, whereas the ⌬TPR2 mutation strongly stimulated this interaction like in the case of ABC10␣. Altogether, these results give weight to the observed ABC10␣-131 interaction and suggest that a conformational change of 131 favors this interaction.
To confirm the two-hybrid results, a partially purified rABC10␣-thioredoxin fusion protein was subjected to SDS-PAGE, transferred to a membrane, denatured, renatured, and probed with 35 S-131 protein and then with antibodies directed to the T7-Tag epitope present at the N terminus of rABC10␣. As shown in Fig. 2, the 35 S-131 probe was specifically retained at the level of rABC10␣-thioredoxin fusion protein (lane 2) but not by the thioredoxin alone (lane 1). In addition, no signal was observed with a control E. coli protein extract (lane 3) or when the filter was incubated with another 35 S-labeled TFIIIC subunit, 95, used as a control (data no shown). A similar signal was obtained with the mutant protein 35 S-131-⌬TPR2 (lanes 4 -6).
131 and Mutant 131-⌬TPR2 Interact with the Conserved C-terminal Basic Region of ABC10␣-In order to map the domain of ABC10␣ interacting with 131, we performed twohybrid experiments with various mutant proteins (26) (Fig. 3, A  and B). We first tested three C-terminal deletions removing 3, 5, or 7 amino acids (mutants rpc10-14, rpc10-15, and rpc10-16, respectively) (26). These short deletions were previously shown to confer a lethal phenotype (26). The corresponding fusion proteins were normally expressed in vivo suggesting that the lethality did not arise from mutation-induced protein degradation (results not shown). Remarkably, all three deletions were found to abolish the two-hybrid interaction with 131. Double or single point mutations in the basic C-terminal part of ABC10␣ (mutants rpc10-30 and rpc10 -24) that led to a thermosensitive phenotype (26) also suppressed or weakened the FIG. 1. In vivo interaction of wild-type or mutant 131 proteins with ABC10␣. The two-hybrid system was used to monitor protein-protein interactions between 131 and ABC10␣. Transcriptional activation of the lacZ reporter gene was assayed by growing the transformed cells on selective medium and overlaying them with X-gal agar. A, RPC82, TFC4, and TFC4-⌬TPR2 were fused in frame with GAL4 activation domain sequence in pACT2 vector; RPC82, RPC10, and BRF1 were fused in frame with GAL4 DNA binding domain sequence in pAS2 vector. For each two-hybrid experiment the bait and prey are indicated by plus signs and three independent transformants are shown. ␤-Galactosidase dosages are indicated below cell patches; units are expressed in nanomoles of X-gal hydrolyzed per min and per mg of protein; three independent experiments were compiled for each quantification. Columns 1-3, negative controls; column 4, ABC10␣-131 interaction; column 5, ABC10␣-131-⌬TPR2 interaction; column 6, 131-TFIIIB70/BRF1 interaction used as a reference (15). B, two-hybrid interactions between ABC10␣ and 131 deletion mutants. Wild-type or mutant 131 proteins were fused to the GAL4 activation sequence domain in pACT2 vector; ABC10␣ was fused with the GAL4 DNA binding domain sequence in pAS2 vector. Arbitrary values are given for white (Ϫ) and for different degrees of blue coloration ((ϩ), ϩ, ϩϩ, and ϩϩϩ) of cell patches on X-gal plates (same representation as in Ref. 16  interaction with 131. On the other hand, a double mutation lying outside this region, rpc10-11, which also caused a thermosensitive phenotype (26), did not affect the two-hybrid interaction with 131. These data suggest that 131 interacts with the C-terminal part of ABC10␣ and point to a critical role of the conserved Arg-60 residue in this interaction.
Interestingly, as shown in Fig. 3C, the ⌬TPR2 mutation increased nearly 3-fold the interaction with the wild-type ABC10␣ as well as with the two mutant proteins rpc10-11 and rpc10-30. In fact, the decrease of interaction strength caused by the rpc10-30 mutation (about 2-fold) was more than compensated by using the ⌬TPR2 version of 131. These results confirmed that the mutant 131-⌬TPR2 protein interacted more strongly with ABC10␣ than with the wild-type protein.
131-⌬TPR2 Is an Allele-specific Suppressor of rpc10-30 Mutant-To assess the functional role of the ABC10␣-131 interaction, we tested whether 131 or its ⌬TPR2 version could rescue the ts phenotype of two ABC10␣ mutants that affected (rpc10-30) or did not affect (rpc10-11) the level of interaction with 131. First, overexpression of the wild-type protein 131 did not suppress these two mutations; in contrast, however, overexpression of the mutant 131-⌬TPR2 selectively suppressed the rpc10-30 mutation (Fig. 4). After 5-FOA induced loss of the high copy number plasmid harboring 131-⌬TPR2, no cell growth could be observed at the restrictive temperature, thus confirming the suppression by the ⌬TPR2 mutation. The other mutation, rpc10-11, that did not affect the interaction with 131 (see Fig. 3C) was not suppressed by 131-⌬TPR2. Therefore, the mutation 131-⌬TPR2 restored both the twohybrid interaction with rpc10-30 and the growth of the rpc10-30 mutant at restrictive temperatures. The fact that this suppression was not observed with the wild-type 131 protein could be explained in the light of the ␤-galactosidase induction level. Indeed, the level of interaction of the ABC10␣-131-⌬TPR2 couple was nearly 3-fold that of the ABC10␣-131. Note that the rpc10-30-131-⌬TPR2 interaction was also stronger than that of the two wild-type proteins (see Fig. 3C).
It should be noted that the ⌬TPR2 mutation has been previously reported to confer a lethal phenotype to yeast cells harboring a partially deleted copy of TFC4 (15). However, we found that in another genetic context, in which the 131 gene has been totally deleted, the same mutation turned out to be viable but conferred a thermosensitive phenotype (data not shown). It was intriguing that the rpc10-30 ts mutant could be rescued at non-permissive temperature by the overexpression of 131-⌬TPR2 which also caused a ts phenotype. Note, however, that this suppression experiment was performed in yeast cells harboring a wild-type copy of TFC4. When 131-⌬TPR2 was overexpressed in a wild-type context for both ABC10␣ and 131, no effect on the cell growth rate could be observed, indicating that the ts phenotype of 131-⌬TPR2 was not dominant (not shown).
Putative Archaeal Orthologs of ABC10␣ and 131-Among the five subunits common to the three nuclear RNA polymerases, ABC27, ABC23, and ABC10␤ have an archaeal counterpart, named H, K, and N (40). No archaeal ortholog has yet been described for the ABC10␣ subunit. As five complete archaeal genomes are available, it was of interest to search for a possible counterpart of the eukaryotic ABC10␣ subunit. By using the sequence of ABC10␣ of S. cerevisiae as entry and TblastN 2.0.8 (32), a small unannotated ORF, named AF0055, was identified in the genome of A. fulgidus (33). Using this ORF as a probe, a similar ORF was identified in P. horikoshii (34) and P. abyssi genomes. By using the P. abyssi ORF, a similar ORF was also found in the M. jannaschii genome (35). Remarkably, these four short ORFs are always found immediately 3Ј of the gene coding for L37A, a conserved ribosomal protein specific to archaea and eukarya. Archaeal operons containing RNA polymerase subunit genes often contain ribosomal protein genes (40). The P. horikoshii sequence had been reported as being homologous to an unspecified S. cerevisiae RNA polymerase subunit, however, with a wrongly estimated length making this ORF overlap the end of L37A protein (34). A closer examination of the Methanobacterium thermoautotrophicum genome (41) revealed a similar short ORF located 3Ј of the L37A protein gene but lacking an initiation codon. These five archaeal sequences are shown in Fig. 5, alongside with the ABC10␣ sequences of A. thaliana, Caenorhabditis elegans, Homo sapiens, C. albicans, Schizosaccharomyces pombe, and S. cerevisiae. These sequence comparisons strongly suggest the existence of an ortholog of ABC10␣ in archaea.
A protein (MJ0941) of the archaea M. jannaschii has been annotated as a putative subunit of transcription factor IIIC (35). This observation was intriguing and prompted us to reexamine the relationship of this protein to TFIIIC subunits. Indeed, the archaeal protein showed a clear sequence similarity to 131. However, 131 and its human counterpart are characterized by their high content in TPR motives clustered in three blocks of 5, 4, 1, and 1 TPR (Fig. 5B). As the archaeal ORF was made of a succession of 9 TPR motives, the similarity between the two proteins was essentially based on the presence of the TPR motives (Fig. 5B). Furthermore, the archaeal protein was much shorter than 131; it was a tandemly duplicated protein, and the same arrangement was not conserved in other archaeal  35 S-131-⌬TPR2 as indicated. The bound labeled probes were revealed by autoradiography (upper panels). The migration of molecular mass markers is indicated. The same membrane was incubated after autoradiography with antibody raised against the T7-Tag epitope fused at the N terminus of ABC10␣, and immune complexes were visualized using the ECL chemiluminescence kit (Amersham Pharmacia Biotech) (lower panels).
genomes. Therefore, we found no evidence for the presence of TFIIIC-related proteins in archaea. DISCUSSION We present biochemical and genetic evidence in favor of a functional contact between Pol III and its cognate assembly transcription factor TFIIIC via a direct interaction between 131 and a shared polymerase subunit ABC10␣. This interaction extends the role of TFIIIC beyond its known functions in promoter recognition and TFIIIB assembly.
The two-hybrid and far Western experiments clearly demonstrated the affinity of ABC10␣ for 131. A deletion mutant analysis could not restrict the interaction with ABC10␣ to a specific domain of 131. The binding of ABC10␣ might necessitate a cooperation between two or more domains of 131. The stronger interaction of the 131-⌬TPR1, -⌬TPR2, or -⌬TPR3 mutants over the wild-type protein suggests that a conformational change, facilitated by the absence of TPR motif 1, 2, or 3, exposes some important interacting domain. In two-hybrid experiments, 131-⌬TPR2 interacted more strongly with TFIIIB90/BЉ (16), which was not the case with TFIIIB70/BRF1 (15). Remarkably, in a random mutagenesis experiment carried over TPR units 1-8, the search for suppressors of an A block-down mutation yielded 10 mutants covering a 53-amino acid area extending over TPR1, -2, and -3 and centered on TPR2. One of these mutants, PCF1-2, was shown to activate Pol III transcription by increasing the recruitment of TFIIIB70/ BRF1 through a non-equilibrium binding mechanism (42). These results and those presented in this work converge to underscore the importance of the second TPR motif. It is possible that both the improved recruitment of TFIIIB70/BRF1 by PCF1-2 and the better interaction between 131 and ABC10␣ could stem from the same conformational change in 131 favored either by mutations that disrupt the second TPR structure or by the deletion of this whole TPR unit. In fact, drastic conformational changes are likely to occur in 131 during the TFIIIB assembly process (43).
The finding of an interaction between two proteins belonging to two different multiprotein complexes raises the question of its functional significance. The ABC10␣-131 interaction occurred via the C terminus domain of ABC10␣ and was affected by a thermosensitive mutation in that region, rpc10-30. Interestingly, the ts phenotype of the rpc10-30 mutant was suppressed by overexpression of the 131-⌬TPR2 protein as could be expected since the ⌬TPR2 mutant interacted more strongly with ABC10␣. As a matter of fact, we observed that a ts mutation (rpc10 -11) in another region of ABC10␣ did not impair  Fig. 5A. B, the phenotype of the ABC10␣ mutants is summarized (26) as follows: ϩ, wild type; Ϫ, lethal; ts, thermosensitive; Pol III Ϫ or Pol Ϫ , specific transcription defect in vivo (26). The level of two-hybrid interaction is indicated as in Fig. 1B ABC10␣ mutations were transformed with plasmids, pFLRPC10, pFL131, and pFL⌬TPR2 overexpressing ABC10␣, wild-type 131, and 131-⌬TPR2, respectively, as indicated. The empty vector pFL44 was used as a control. Overexpression of 131-⌬TPR2 allowed cell growth of the ABC10␣ rpc10-30 but not of rpc10-11 mutant. Four Pol III mutant strains were checked for their ability to be rescued by 131-⌬TPR2 as follows: MW670 (rpc160 -112), MW1029 (rpc160 -270), SC91 (rpc53-256/424), and D132-ID (rpc31-236). Transformants were streaked on YPD medium and grown at the permissive (30°C) or non-permissive (37°C) temperature for 4 days. the two-hybrid interaction with 131, whereas the thermosensitive mutant (rpc10-30) in the C-terminal region affected the interaction. Reciprocally, the overexpression of wild-type 131 or 131-⌬TPR2 was unable to suppress the rpc10-11 ts mutant that did not affect the ABC10␣-131 interaction. It is also important to note that the rpc10-30 mutant was specifically affected, in vivo, in Pol III transcription, whereas the nonrescuable rpc10-11 mutant was not Pol III-specific (26). ABC10␣ has been previously identified as a suppressor of tsv115, a ts mutation in the 138 subunit of TFIIIC (44) affecting TFIIIC-DNA binding and the assembly of the Pol III preinitiation complex (45). None of the other Pol III subunits tested at that time (C160, C128, C82, C53, AC40, C34, C31, AC19, and ABC10␤) were found to suppress the tsv115 mutation when overexpressed. ABC10␣ was suggested to be a critical subunit limiting the rate of Pol III assembly. In fact, a purified Pol III harboring the rpc10-30 mutation did not display any transcriptional defect in vitro, but the level of Pol III in mutant extracts was much decreased (26). As diploid cells with only one gene copy for ABC10␣ have a growth defect, ABC10␣ is indeed likely to affect a rate-limiting step in polymerase assembly (26). It remains that the Pol III-specific phenotype of rpc10-30 may be due in part to its deficiency in TFIIIC interaction since it is partially suppressed by 131-⌬TPR2.
A contact between yeast TFIIIC and Pol III is not implied in the sequential initiation complex assembly model where TFIIIC assembles TFIIIB which in turn recruits Pol III. Indeed a preassembled TFIIIB⅐DNA complex can direct accurate transcription by Pol III in the absence of TFIIIC (46). Nevertheless, the association of Pol III with TFIIIC in yeast extracts has been demonstrated by coimmunoprecipitation experiments (18). Therefore, the ABC10␣-131 interaction may be involved in the formation and/or stability of a Pol III holoenzyme. Alternatively, this interaction may facilitate the recruitment of Pol III by the TFIIIB⅐TFIIIC⅐DNA complex. An interaction of Pol III with TFIIIC is also likely to occur when the enzyme elongates through the TFIIIC-bound intragenic promoter. Therefore, Pol III may well engage in many interactions with TFIIIC which have not yet been discovered. Recently, Roeder and collaborators (47) have characterized two subunits of human TFIIIC that are clearly homologous to 131, with its characteristic TPR units distributed similarly over the sequence, and to 95 that is involved in A block binding. Most interestingly, one of these polypeptides, hTFIIIC63 (homologous to 95), was found to interact with a human Pol III subunit, hRPC62 (homologous to the yeast Pol III subunit C82). In addition to its recognized role as TFIIIB assembly factor and in relieving the repression by chromatin of class III gene transcription, the observed interac-  (49), H. sapiens (31), C. albicans, S. pombe (31), and S. cerevisiae (38). A putative ortholog of ABC10␣ was found in each of the five complete archaeal genomes available: A. fulgidus (33), P. horikoshii (34), P. abyssi (see "Experimental Procedures"), M. jannaschii (35), and M. thermoautotrophicum (41). In all five genomes, the ABC10␣ ortholog ORF is found immediately 3Ј of the L37A ribosomal protein gene (the distance from the stop codon of L37A is indicated at the left). Note that the sequence of M. thermoautotrophicum lacks an initiation codon. The position of the stop codons in the genomes and orientation of the ORFs are indicated at the right. Capital letters indicate amino acids conserved separately in eukaryotic or in archaeal sequences, and boldface capital bold letters indicate amino acids conserved in both eukaryotic and archaeal sequences (the sequence of M. thermoautotrophicum that departs from the other ones was not taken into account at some positions in the N-terminal region). The amino acids equivalence used are: D and E; I, L, and V; G and S; K and R; and F and Y. Gaps, indicated as dashes, were introduced to maximize homologies. # and ϩ denote residues strictly or partly conserved, respectively. B, comparison of 131 and its putative archaeal ortholog from M. jannaschii (ORF MJ0941). The presence of TPR motives is indicated by the peaks localized at the center of each TPR motif which are numbered from 1 to 11 in 131 sequence or 1 to 9 in MJ0941. tion of TFIIIC subunits with Pol III subunits suggests additional functions for TFIIIC.