The Saccharomyces cerevisiae RuvB-like Protein, Tih2p, Is Required for Cell Cycle Progression and RNA Polymerase II-directed Transcription*

Two highly conserved RuvB-like putative DNA helicases, p47/TIP49b and p50/TIP49a, have been identified in the eukaryotes. Here, we study the function of Saccharomyces cerevisiae TIH2, which corresponds to mammalian p47/TIP49b. Tih2p is required for vegetative cell growth and localizes in the nucleus. Immunoprecipitation analysis revealed that Tih2p tightly interacts with Tih1p, the counterpart of mammalian p50/TIP49a, which has been shown to interact with the TATA-binding protein and the RNA polymerase II holoenzyme complex. Furthermore, the mutational study of the Walker A motif, which is required for nucleotide binding and hydrolysis, showed that this motif plays indispensable roles in the function of Tih2p. When a temperature-sensitive tih2 mutant,tih2–160, was incubated at the nonpermissive temperature, cells were rapidly arrested in the G1 phase. Northern blot analysis revealed that Tih2p is required for transcription of G1 cyclin and of several ribosomal protein genes. The similarities between the mutant phenotypes of tih2–160 and those of taf145 mutants suggest a role for TIH2in the regulation of RNA polymerase II-directed transcription.

One of the common post-translational modifications of nuclear and cytosolic proteins in eukaryotes is the addition of N-acetylglucosamine residues O-linked (O-GlcNAc) 1 to serine and threonine residues. A number of physiologically or structurally important proteins have thus far been shown to contain O-GlcNAc, including the largest subunit of RNA polymerase II (RNAP II) (1), transcription factors such as Sp1 and hepatocyte nuclear factor 1 (2-4), nuclear pore proteins (5,6) and chromatin-associated proteins (7). To study nuclear factors involved in nuclear transport and other nuclear events, we previously per-formed an in vitro binding assay of a rat liver nuclear matrix fraction using a wheat germ agglutinin affinity column (8).
Several O-GlcNAc-containing proteins such as nucleoporins as well as nonglycosylated proteins like importin ␤ were isolated (9). Two RuvB-like proteins, p50 (9) and p47 (10), were also isolated by this method probably because of their interaction with O-GlcNAc-bearing proteins. RuvB is a prokaryotic DNA helicase, and its helicase activity and DNA binding affinity are enhanced by interaction with RuvA (11,12). These two factors form a large motor protein complex to promote branch migration at Holliday junctions at the late stages of homologous recombination. The p50/p47 and RuvB proteins share highly conserved Walker domains (A and B), indicative of proteins that bind nucleotide triphosphates (10,13). Based on this, p50 and p47 were proposed to be the eukaryotic homologues of the RuvB DNA helicase. Indeed, p50 and p47 are present in a wide range of eukaryotes ranging from yeast to mammals, suggesting that this basic helicase activity may be conserved among the eukaryotes.
In addition to recombination, transient unwinding of the DNA duplex is required in many cellular processes such as replication, transcription, and DNA repair. To accommodate these different aspects of DNA metabolism, cells typically encode multiple helicases, occasionally found in functionally distinct complexes. To date, at least 10 different DNA helicases have been identified in Saccharomyces cerevisiae (14 -17). However, only a few DNA helicase genes are essential for viability, such as RAD3, SSL2, and DNA2. Functional analyses of RAD3, SSL2, and DNA2 indicate that they are indeed required for fundamental cellular functions. Rad3p and Ssl2p are essential components of TFIIH, which is required for transcriptional initiation and transcription-coupled repair (15,16), whereas Dna2p is required for DNA replication (17).
Recently, several groups have independently reported the identification of p50. First, a 49-kDa protein isolated from rat liver nuclear extracts was found to interact with TATA-binding protein (TBP) (13). The 49-kDa protein, referred to as TIP49a (TBP-interacting protein) was identical to p50 and was enriched in the testes. The recombinant TIP49a is an ATP-dependent DNA helicase (18). Meanwhile, a human homologue of p50, RUVBL1, was shown to co-purify with RNAP II holoenzyme, and the yeast counterpart was reported to be essential for viability (19). Furthermore, in a search for proteins that interact with ␤-catenin, two novel human proteins, 52-and 44-kDa, were detected (20). The 52-kDa protein corresponding to p50 (TIP49a, RUVBL1) was designated Pontin 52 and shown to bind to TBP directly. In addition, p50 was also isolated as NMP238, a ubiquitously occurring nuclear matrix protein (21). On the other hand, little has been described thus far for p47. In a very recent study, a new RuvB-like DNA helicase termed TIP49b was demonstrated to associate with TIP49a (22). The protein sequence between p47 (mouse) and TIP49b (rat) varies by only 2 residues.
Although the above reports indicate that p50 interacts with TBP, RNAP II holoenzyme, or both, p50 and p47 may also provide helicase enzymatic activity in recombination events.
The key question of what critical function(s) p50 provides in vivo that makes it essential for viability is not yet understood, but it is important to note that studies to date link p50 primarily to the machinery of gene transcription. Given the similarities shared by recombination, replication, and transcription events at the DNA helicase level, it is not surprising that p50 may participate in multiple cellular events. In this study, we undertook a genetic approach to study the in vivo role of Tih2p, the less well studied yeast homologue of p47. Similar to the yeast homologue of p50, TIH1 (TIP49a homologue), TIH2 (TIP49b homologue) is also an essential gene. Whereas the sequence homology of Tih2p to RuvB suggests a role in recombination and DNA repair, our results indicate that Tih2p is essential for proper cell cycle progression and for the selective transcription of a subset of genes. The participation of Tih2p in these functions may explain its requirement for viability.

EXPERIMENTAL PROCEDURES
Yeast Strains, Media, and Plasmid Constructions-Yeast strains used in this paper are listed in Table I. All strains were grown in rich medium (YPD) or selective medium (synthetic complete, synthetic dextrose, synthetic galactose) supplemented with the appropriate nutrients. Genetic manipulation of yeast cells was performed as described previously (23). The pRS vector series was described previously (24). Details of primers used in this study are available on request. The TIH2 knockout plasmid, pDisTIH2, was constructed by PCR amplification of a pair of DNA fragments flanking the TIH2 ORF, creating appropriate restriction sites at each end, and by subsequent ligation of these fragments into pRS303. To replace the entire TIH2 ORF with HIS3, pDis-TIH2 was linearized with BamHI and transformed into a diploid strain, CRDFY (FY background, Table I, FY strains are generous gifts from F. Winston, Harvard Medical School, Boston, MA). Disruption was verified by Southern blotting. The TIH2 locus on an EcoRI-ClaI fragment was amplified by PCR of genomic DNA and ligated into pRS316 and pRS314 to obtain pRS316-TIH2 and pRS314-TIH2, respectively. To create pRS425-ADH1p-p47, the ORF of p47 was PCR-amplified from a pBluecript II SK(Ϫ) plasmid containing p47 (10) and subsequently ligated into HindIII/NotI-digested pRS425-ADH1p (25). Plasmid pRS425-ADH1p-p50 was created similarly except that p50 was PCRamplified from the plasmid p50(3A-5) (9). To create pRS314-ADH1p-S65TGFP-TIH2, TIH2 was amplified by PCR (ORF minus start codon) and ligated into EcoRI/NotI-digested pRS314-ADH1p-S65TGFPnucleoplasmin, constructed previously by ligating an XhoI/NotI ADH1p-S65TGFP-nucleoplasmin fragment isolated from pAGN5 (26) into pRS314. To create K81R-TIH2, fusion PCR was performed (27). The final PCR product was digested with EcoRI/ClaI and ligated into pRS314 yielding pRS314-K81R-TIH2. Plasmid pRS426-GAL1p-K81R-TIH2 was created by PCR amplification of the ORF of K81R-TIH2 from pRS314-K81R-TIH2 and subsequent subcloning into EcoRI/NotI-digested pRS426-GAL1p. Plasmid pRS426-GAL1p was created by amplification of GAL1p (template: pYES2, InVitrogen) and subsequent subcloning into XhoI/HindIII-digested pRS426. Plasmid pRS426-GAL1p-TIH2 was constructed similarly as pRS426-GAL1p-K81R-TIH2, except that pRS316-TIH2 was used as the TIH2 ORF template in the PCR step. To express an N-terminally double hemagglutinin-tagged Tih2p fusion protein from its authentic promoter (pRS314-HA-TIH2), fusion PCR was performed (27). The final fusion PCR products were subcloned as an EcoRI/ClaI-digested fragment into pRS314. To tag ProtA to the N terminus of Tih1p, a fragment containing NOP1p-ProtA from pUN100-ProtA-NUP133 (28) was isolated as a SalI/EcoRI fragment and subcloned into pRS314 yielding pRS314-NOP1p-ProtA. A fragment containing TIH1 (ORF minus start codon) was PCR-amplified from genome DNA with the attachment of EcoRI/NotI sites and subsequently inserted into pRS314-NOP1p-ProtA to obtain PRS314-NOP1p-ProtA-TIH1.
Isolation of Temperature-sensitive (ts) TIH2 Alleles-A DNA frag-ment (nucleotides Ϫ817 to 1701) of TIH2 (ORF of TIH2 corresponds to nucleotides 1-1416) was randomly mutagenized by PCR, using a 1:5 ratio of dATP to the other dNTPs (29). The PCR product was digested with EcoRI and ClaI and subcloned into pRS314 to generate a pRS314based TIH2 mutant plasmid library. The mutant plasmid library was introduced into the parental strain CRPA1. Transformants were plated on 5-FOA at 23°C to counterselect against the TIH2,URA3 plasmid. A total of ϳ500 colonies were isolated and streaked on YPD. After 3-5 days of incubation at 23°C, they were replica plated onto two YPD plates and incubated at either 23 or 37°C. After 2-3 days of incubation at the respective temperatures, plates were compared, and strains that were able to grow at 23°C but not 37°C were isolated. Plasmids harboring the mutagenized TIH2 genes were recovered from these putative ts strains and re-introduced into the parental strain (CRPA1). Temperature sensitivity was reconfirmed by plasmid shuffling. The sequences of some of the mutant alleles were determined. RNA Preparation and Analysis-Cells grown in YPD medium to the early exponential phase at 23°C (A 600 Յ 0.2), were harvested at the appropriate times after temperature shift by centrifugation, washed with medium, frozen in liquid nitrogen, and stored at Ϫ80°C. Total RNA from cells was isolated by hot phenol extraction (30). Northern blotting was performed as described (31). Twelve g of each RNA sample were subjected to electrophoresis in a 1.1% formaldehyde/agarose gel, followed by transfer to a nylon membrane (Hybond-Nϩ, Amersham Pharmacia Biotech). All probes used for hybridization were generated by PCR amplification using genomic DNA as template and labeled using a random primer labeling kit (TaKaRa). Blots were stripped for reprobing by boiling and cooling to room temperature in 0.1% SDS. For slot blot analysis, 2 g of RNA from each sample were used. Experiments were performed essentially as described (32).
Whole-cell Extract Preparation and Immunoblot Analysis-Protein extracts from yeast strains grown in YPD liquid culture prepared essentially as described (33) were separated on SDS-polyacrylamide gels, transferred to nitrocellulose membranes and probed with either anti-yTAF II 145 antibody or anti-TBP antibody or anti-yTAF II 61 or anti-yTAF II 90 antibodies. Detection of the antibody signal was accomplished with the Amersham Pharmacia Biotech ECL Western detection kit.
IgG Pull-down, in Vivo Labeling, and Immunoprecipitation-Extracts were prepared by pelleting 10 ml of cells (A 600 ϳ0.5) cultured in synthetic complete, washing once with lysis buffer (50 mM Tris-HCl, pH 8.0, 5 mM EDTA), and resuspending again in 200 l of lysis buffer supplemented by protease inhibitors. Composition of protease inhibitors was 1 mM phenylmethylsufonyl fluoride and 10 g/ml each of pepstatin, leupeptin, and aprotinin. Glass beads (200 l) were added, and cells were lysed by several cycles of vortexing for 30 s followed by a 1-min incubation on ice. Extracts were clarified by two consecutive centrifugations (14,000 ϫ g for 5 min). Extract protein concentration was determined using BCA protein assay (Pierce) and found to be typically 2-4 mg/ml. Extracts (170 l) were preincubated in 4 volumes of immunoprecipitation buffer (1.25% Triton X-100, 180 mM NaCl, 6 mM EDTA, 60 mM Tris-HCl (pH 8.0), 6% skim milk) containing 5 l of Sepharose for 30 min at 4°C. The supernatant was collected, and 5 l of immunoprecipitation buffer-pre-equilibrated IgG-Sepharose was added and incubated on a rotator for 1 h at 4°C. Beads were then washed extensively with wash buffer (50 mM Tris-HCl (pH 8.0), 1% Triton X-100, 150 mM NaCl, 5 mM EDTA), and 10 l of SDS sampling buffer was added; the sample was boiled for 5 min and then loaded on a 10% SDS-polyacrylamide gel. To detect HA-Tih2p or ProtA-Tih1p, anti-HA 12CA5 monoclonal antibody or rabbit IgG was used, respectively.
To label cells with [ 35 S]methionine/cysteine (Express protein labeling mix; NEN Life Science Products) 5 ϫ 10 7 exponentially grown cells in methionine/cysteine-free synthetic complete medium were resuspended in 1 ml of methionine/cysteine-free synthetic complete medium and labeled with 1 mCi of [ 35 S]methionine/cysteine for 20 min at 30°C. Cell extracts were prepared using glass beads. Immunoprecipitation was performed as described (34) using anti-HA 12CA5 monoclonal antibody as the first antibody. Extracts were fractionated using ammonium sulfate (0.4 g/ml) prior to immunoprecipitation.
Budding Morphology-To determine the budding index, log phase cells growing at 23°C in YPD medium were temperature-shifted to 37°C, and aliquots were removed at the appropriate time points. These were sonicated briefly and fixed with 3.7% formaldehyde in phosphatebuffered saline (100 mM NaCl, 80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , pH 7.3), and the number of single, small budded and large budded cells were determined microscopically.
DNA Flow Cytometry-1 ϫ 10 7 cells growing exponentially in YPD at 23°C were temperature-shifted to 37°C and harvested at various time intervals. They were fixed in 70% ethanol overnight at 4°C and washed twice in 50 mM sodium citrate (pH 7.4). These fixed cells were resuspended in the same buffer (1 ml) containing 0.25 mg/ml RNase A and incubated for 1 h at 50°C. One mg of proteinase K was added, and incubation was continued at 50°C for another hour. Cells were washed once and resuspended in 1 ml of the same buffer plus 16 g/ml propidium iodide. DNA content was measured using a fluorescence-activated cell sorting system (FACScan, Becton Dickinson).

RESULTS
TIH2 Is an Essential Gene-In our previous studies, two eukaryotic homologues of the bacterial RuvB DNA helicase, p50 and p47, were isolated. Searching through the S. cerevisiae Genome Data base (Stanford University), two ORFs designated as YDR190C and YPL235W were identified as the yeast homologues of p50 and p47, respectively. Sequence identities are as high as 70% between p50 and YDR190C and 69% between p47 and YPL235W (Fig. 1A). On the other hand, sequence identities between p50 and p47 and between YDR190C and YPL235W are only 43 and 40%, respectively. Therefore, we propose classifying these eukaryotic RuvB-like proteins into two subfamilies, i.e. the p50 and the p47 subfamily (Fig. 1A). It was of great interest to us that different eukaryotes would encode two highly conserved subfamilies of RuvB-like proteins that are homologous to each other and of similar molecular size (50 and 51 kDa). To investigate the functions of these RuvB-like proteins, we undertook a genetic study of the less well studied p47 yeast homologue, YPL235W. For simplicity, we will refer to YDR190C (p50 homologue) as TIH1 for TIP49a homologue (see Introduction) and YPL235W as TIH2.
Because TIH1 had been shown to be an essential gene (9,19,22), we first asked if TIH2 is also required for viability. The TIH2 gene was disrupted as described under "Experimental Procedures." A CEN/ARS plasmid containing the wild-type TIH2 gene (URA3, TIH2) was introduced into the TIH2/tih2::HIS3 diploid strain. One HIS3 URA3 progeny obtained from the tetrad analysis, designated CRPA1 (Table I), was transformed with either a pRS314-based plasmid (CEN/ARS) containing wildtype TIH2 (TRP1, TIH2) or a control vector (pRS314) and was counter selected on 5-FOA medium. Only transformants containing the (TRP1, TIH2) plasmid were able to grow on 5-FOA medium, indicating that TIH2 is indeed required for mitotic vegetative growth (Fig. 1B).
To determine if Tih2p could be functionally replaced by the highly conserved mammalian p47, we constructed a multicopy plasmid overexpressing p47 from a yeast constitutive promoter (ADH1p). We found out that neither p47 nor p50 was able to rescue the tih2 ⌬ disruptant, whereas a similar control vector overexpressing the yeast TIH2 gene did so as well as complementation with the endogenously expressed gene (Fig. 1B). Thus the yeast TIH2 gene cannot be complemented by p47, similar to the results obtained with p50, which was unable to replace the function of TIH1 in a tih1⌬ disruptant (19, 22) (data not shown).
Tih2p Is Localized to the Nucleus-To probe the cellular location of Tih2p, a construct consisting of Tih2p with the green fluorescent protein fused to its N terminus (GFP-Tih2p) was inserted into an ADH1p expression vector. The parental null allele strain CRPA1 was transformed with this GFP-TIH2 fusion and tested for growth on 5-FOA medium. Transformants were found to grow normally, indicating that GFP-TIH2 was fully functional (Fig. 2A). Observation using a fluorescent microscope showed that this fusion protein was localized primarily to the nucleus (Fig. 2B), consistent with the fact that p47 was isolated from rat liver nuclear extract (10). Faint GFP fluorescence was also detected in the cytosol, probably because of overexpression. The cytosolic fluorescence was not caused by GFP-Tih2p degradation, because GFP-Tih2p proteolysis was not observed (data not shown). Similar results were obtained with a yeast strain expressing a functional HA-tagged TIH2 (HA-Tih2p) under the control of its own promoter (data not shown).
The Walker A Domain of Tih2p Is Essential for Its Function-The only recognizable functional motifs in Tih2p are the so-called Walker A and B motifs, which are involved in ATP binding/hydrolysis and are the region of Tih2p most homologous to the prokaryotic helicase, RuvB (9,10,35). To investigate whether the helicase domain of Tih2p contributes to its essential physiological role, the invariant lysine at position 81 in the major GX 4 GKT nucleotide-binding loop was changed to arginine (K81R) by site-directed mutagenesis (Fig. 3A).
A pRS314-based plasmid harboring the mutated gene K81R-TIH2 (TRP1, K81R-TIH2) or wild-type TIH2 (TRP1, TIH2) was introduced into the parental null allele strain CRPA1. When counter selected on 5-FOA medium, transformants harboring the plasmid (TRP1, K81R-TIH2) failed to grow at different temperatures (Fig. 3B, only plates incubated at 30°C are shown). This indicated that K81R-TIH2 alone is unable to support the growth of tih2 ⌬ cells. Therefore, we concluded that the Walker A motif is essential for the function of Tih2p that contributes to cell viability.
cantly retarded when grown on synthetic medium containing galactose as the sole carbon source (Fig. 3C). This result implies that the function of Tih2p may require the formation of self-complexes or complexes with certain cellular components in vivo.

FIG. 3. The Walker A domain of Tih2p is required for its essential function in vivo.
A, the position and the sequence of Walker A and B domains of Tih2p is shown. The highly conserved lysine residue (Lys 81 ) in the Walker A domain was mutated by site-directed mutagenesis to arginine yielding K81R-TIH2. B, strain CRPA1 was supertransformed with TRP1 CEN/ARS plasmids (pRS314) carrying the wildtype TIH2 gene or the K81R-TIH2 gene. The ability of the K81R-TIH2 allele to complement the TIH2 chromosomal deletion was tested as described in Fig. 1B. C, overexpression of K81R-TIH2 mutant protein conferred a dominant negative growth defect. A strain (CRH6) recovered from the 5-FOA medium as described in B was transformed with a URA3 2 plasmid (pRS426) containing the wild-type gene, no gene, or the K81R-TIH2 gene expressed from the GAL1 promoter. The resultant strains were grown to log phase in synthetic glucose medium lacking tryptophan and uracil and spotted in 10-fold serial dilutions (10 1 -10 4 ) on synthetic medium plates containing glucose (synthetic dextrose, left) or galactose (synthetic galactose, to induce expression from the GAL1 promoter, right) as the sole carbon source.
pressed under the constitutive nucleolar protein promoter (NOP1p) (36) was functional (data not shown). ProtA-Tih1p was detected by immunoblotting using rabbit IgG when introduced into a tih2 ⌬ strain harboring plasmid (TRP1, TIH2) (Fig. 4B, lanes 1 and 2). Cell lysates from tih2 ⌬ strains ex-pressing both ProtA-Tih1p (URA3, ProtA-TIH1) and either nontagged Tih2p (TRP1, TIH2) or HA-tagged Tih2p (TRP1, HA-TIH2) were first precipitated using IgG-Sepharose followed by immunoblotting using anti-HA antibody (lanes 6 and 8). A specific band corresponding to HA-Tih2p was detected only in the strain expressing both tagged TIH2 (HA-Tih2p) and tagged TIH1 (ProtA-Tih1p) (compare the band positions in lanes 4 and 8). This result indicated that HA-Tih2p was selectively recovered by IgG-Sepharose precipitation from the strain also expressing ProtA-Tih1p. A tih2 ⌬ strain expressing ProtA-Tih2p as the sole copy of TIH2 was also constructed, and a pull-down of the lysates by IgG-Sepharose followed by silver staining revealed two major and several as yet unidentified minor bands. Microsequencing of the peptide making up the major band adjoining ProtA-Tih2p revealed that this protein was Tih1p (data not shown). Therefore, both Tih1p and Tih2p may interact directly to form a complex in yeast cells.
Growth Characteristics of a Temperature-sensitive Mutant, tih2-160 -To determine the cellular role(s) of Tih2p, we isolated several ts tih2 alleles. On plates, all of the mutants exhibited a similar growth defect at the nonpermissive temperature of 37°C. However, in liquid media their growth at 37°C varied. We focused our analysis on one of the ts mutants, tih2-160, because it exhibited not only a very stringent ts growth on plates (no growth above 30°C) but also a relatively rapid cessation of growth at 37°C in liquid media (arrested within 4 -5 h). Growth at 23°C is also slower in this strain, but no other particular defects could be observed (Fig. 5A). Interestingly, tih2-160 cells incubated on plates at 37°C regained viability and the capacity for growth when shifted down to 23°C but not when shifted down to 30°C (Fig. 5B). The amount of tih2-160 mutant protein detected in the cell was relatively unchanged even after 4 h of incubation at 37°C suggesting that the mutant phenotype was because of loss of function and not degradation of the protein (data not shown).
Reversible G 1 Arrest-When tih2-160 cells were shifted to 37°C, we noticed that there was an accumulation of unbudded cells, indicative of a G 1 phase cell cycle arrest (Fig. 6, B and C, upper panel). To delineate this cell cycle block more precisely, the arrested cells were analyzed using a fluorescence-activated cell sorter. As expected, tih2-160 cells showed a dramatic and rapid increase in the proportion of cells with a 1N DNA content (Fig. 6A). Nearly 70% of the cells were arrested within 3-4 h as unbudded cells, characteristic of the G 1 phase, whereas the budding pattern of isogenic wild-type cells grown at 37°C stayed relatively unchanged throughout the course (Fig. 6B). Consistent with the recovery of growth (Fig. 5B), the cell cycle block was released when the tih2-160 cells were shifted back to 23°C (Fig. 6A) and the proportion of large budded cells increased to around 34% of the population, whereas that of unbudded cells dropped to ϳ40% (Fig. 6, B and C). This result indicated that Tih2p function is indeed required for proper cell cycle progression. The reversibility of this cell cycle arrest with temperature is consistent with the idea that the cell cycle arrest in tih2-160 cells is because of inactivation of tih2-160p at 30°C and higher.
Inactivation of tih2-160p Abolishes Transcription of G 1 Cyclin and Ribosomal Proteins-One explanation for the cell cycle arrest observed in tih2-160 cells is that Tih2p is required for the transcription of specific genes involved in the progression beyond the G 1 phase. To test this hypothesis, tih2-160 cells and isogenic wild-type cells were analyzed for transcription of G 1 cyclin. RNA was isolated at 0, 1, 2, and 4 h following incubation at the nonpermissive temperature (37°C), and Northern blot analysis was performed. The result showed that in tih2-160 cells, transcription of the G 1 cyclin CLN2 decreased dramati-  1-4). Extracts of strains with the indicated genotypes were subjected to immunoprecipitation using IgG-Sepharose and detected using an anti-HA antibody (lanes [5][6][7][8]. Closed triangles, Ͻ, and Ͻ Ͻ indicate the position of HA-Tih2p, the IgG heavy chain, and ProtA-Tih1p, respectively. ProtA-Tih1p was detected (lanes 6 and 8) because of the cross-reaction of anti-HA antibody to the IgG domain. cally within 1 h, whereas transcription of CLN3, which is not cell cycle regulated, was unaffected (Fig. 7A). Transcription of these genes was unaffected following temperature shift in the wild-type strain. Therefore, the G 1 arrest observed in tih2-160 cells could be explained by a specific defect in the transcription of G 1 cyclins such as CLN2. However, we could not exclude the possibility that the loss of G 1 cyclin transcripts was a secondary effect of the cell cycle block. To examine if the transcriptional defect could be an indirect consequence of the G 1 arrest, we analyzed the transcription of cell cycle-independent and constitutively expressed genes other than CLN3 such as the small ribosomal subunit protein genes. As a positive control showing decreased transcription, we used a strain mutant for the largest RNA polymerase II subunit, rpb1-1 (37). We observed the small ribosomal subunit protein genes RPS5, RPS30, and RPS26b were expressed at dramatically lower levels in the tih2-160 mutant, and this decrease was comparable to that observed in the rpb1-1 mutant. In addition, the gene encoding the large ribosomal subunit protein, RPL32, was significantly affected. In contrast, transcription of the ADH1 and actin (ACT) genes were unaffected even after 4 h of incubation at the nonpermissive temperature. Therefore, we concluded that Tih2p may be required for the transcription of a subset of genes, including those necessary for proper cell cycle progression. Taken together, these results are similar to those observed with taf145 mutants. yTAF II 145 is the yeast homologue of the higher eukaryotic protein TAF II 250 and is the only yTAF II known to contact TBP directly (38). Incubation of a temperature-sensitive yeast taf145 mutant for 4 h at its nonpermissive temperature results in G 1 arrest, indicated by the accumulation of unbudded cells and in down-regulation of G 1 cyclin and ribosomal protein genes transcription (32,39). However, yTAF II 145 was found to be otherwise dispensable for global RNA polymerase II-directed transcription (32,39). To investigate if Tih2p is required for global transcription, total RNA was prepared from tih2-160, an isogenic wild-type strain, and from the positive control RNAP II mutant, rpb1-1, and subjected to slot blot analysis using 32 P-labeled oligo(dT) as a probe. As expected, inactivation of RNAP II resulted in a rapid loss of RNAP II-mediated transcription, whereas inactivation of tih2-160p by incubation at the nonpermissive temperature for 4 h had no significant effect on total poly(A)ϩ mRNA synthesis (Fig. 7B). Therefore, we concluded that global RNA polymerase II-directed transcription is unaffected in the tih2-160 mutant, similar to the result in taf145 mutants.
Immunoblot Analysis of yTAF II 145, TBP, and Other yTAF II s Following Temperature Shift Inactivation of Tih2p-The mammalian p50/Pontin52/TIP49a was previously shown to interact with the TBP (13,20), whereas the yeast homologue, Tih1p, was shown here to complex with Tih2p. Therefore, it is quite likely that p50 and p47 in mammalian cells and Tih1p and Tih2p in yeast form complexes that interact with their respective TBPs. If these interactions contribute to the stability of TBP, then the transcriptional defect observed upon Tih2p inactivation could be due in part to the disruption of these complexes and destabilization of TBP. Alternatively, another interpretation of the similarities between the tih2-160 and taf145 mutant phenotypes is that the defect conferred by tih2-160 mutation leads to ytaf II 145 degradation. To examine such possibilities, whole-cell extracts were prepared from tih2-160 cells incubated for 0, 1, 2, and 4 h at the nonpermissive temperature and subjected to immunoblot analysis using several antibodies. In addition, to examine the integrity of the TFIID complex in tih2-160 mutant, antibodies against yTAF II 61 and yTAF II 90, which are present in both TFIID and Spt-Ada-Gcn5-acetyltransferase complex, were also tested (40).
In both wild-type and tih2-160 cells, TBP, yTAF II 145, yTAF II 61, and yTAF II 90 were detected at substantial levels even after 4 h of incubation at the nonpermissive temperature (Fig. 7C). This result argues against the possibility that inactivation of Tih2p leads to degradation of components of the TFIID or Spt-Ada-Gcn5-acetyltransferase complex and argues for a primary role of Tih2p in the transcription of these genes. DISCUSSION Two novel RuvB-like proteins, p50 and p47, were isolated from a rat liver nuclear extract using a wheat germ agglutinin-Sepharose affinity column (8). These proteins have highly homologous counterparts in a wide range of eukaryotes and can be classified into two subfamilies, i.e. the p50 and the p47 subfamily (Fig. 1A). The yeast counterpart of the p50 gene, designated TIH1 in this study, has been previously shown to be an essential gene. Here, we showed that the yeast counterpart of the p47 gene, TIH2, is also required for viability (Fig. 1B). Immunofluoresence studies showed that Tih2p is localized pri- Strains with a tih2⌬::HIS3 chromosomal background and transformed with either the wild-type TIH2 or mutant tih2-160 gene carried on a TRP1 CEN/ARS plasmid were grown in YPD medium and absorbance at 600 nm (A 600 ) was recorded over time. Squares represent samples that were temperature shifted from 23°C to 37°C, and the time of temperature shift is shown by arrow. Circles represent samples maintained at 23°C throughout the course. B, strains described in A were cultured at 23°C in YPD medium and streaked onto YPD plates. After incubation at 37°C for 2 days, incubation temperature was shifted to 23°C (top) or 30°C (bottom) for the indicated period. marily in the nucleus, consistent with the fact that it was isolated from a rat liver nuclear extract and suggesting that it participates in nuclear events (Fig. 2). To study the function of Tih2p in detail, we created several ts mutants and further examined the characteristics of one of them, tih2-160, which exhibited a relatively rapid cessation of growth at the nonpermissive temperature (Fig. 5).
Homology to RuvB Suggests a Role for Tih2p in DNA Repair-The ruvB mutants were first isolated as UV-and mitomycin C-sensitive mutants (41). However, the tih2 mutants we have isolated exhibited only a very modest sensitivity to UV radiation, mitomycin C, or methyl methanesulfonate compared with other mutants known to be defective in DNA repair (data not shown). We reasoned that the function of Tih2p in DNA repair may be redundant. For example, the prokaryotic RecG gene encodes a Holliday junction-specific helicase and can partially replace the function of Ruv (42). Recombination and DNA repair in a recG mutant are only mildly affected but become severely defective when recG is combined with mutations in the ruv genes (43). If similar redundancy exists in the eukaryotic system, mutations in other genes may be required in order for the defect in DNA repair in tih2 mutants to be revealed. In this context, creating double mutants of TIH2 and its homologous gene, TIH1, could be interesting because they interact with each other (Fig. 4) and could be functionally synergistic.
Another observation that suggests a role for Tih2p in DNA repair is the fact that the tih2-160 mutant undergoes G 1 arrest. It is possible that Tih2p is required for repairing DNA damage associated with replication, such as double strand breaks. Inactivation of tih2-160p may result in the accumulation of double strand breaks in replicated DNA, thereby activating cell cycle checkpoint genes leading to a G 1 arrest. We tested whether the G 1 arrest observed in tih2-160 cells requires the activation of the checkpoint control gene RAD9 (44). We created a double rad9⌬ tih2-160 mutant and found that lethality was not enhanced compared with the tih2-160 single mutant, and the arrest was not Rad9-dependent (data not shown). Moreover, the double mutant exhibited reversibility of growth arrest similar to that observed in the tih2-160 single mutant (data not shown, Fig. 5) indicating good viability. Therefore, a role of Tih2p repairing DNA damage associated with replication cannot be observed in this study.
A Regulatory Role for Tih2p in RNAP II-mediated Transcription-The ability of the yeast cell to progress through the mitotic cell division cycle requires the coordinate and highly regulated transcription of several genes (45). The G 1 cell cycle arrest observed at the nonpermissive temperature in tih2-160 mutant may be because of temperature-dependent conformational changes in tih2-160p that impair its ability to interact with gene-specific transcription factors necessary for cell cycle progression. Our hypothesis that Tih2p plays a role in transcription is supported by our preliminary observation that Tih2p was coimmunoprecipitated with yeast TBP and by several recent reports identifying the p50 family members, TIP49a, Pontin 52, and RUVBL1, which interact with general transcription factors TBP (with TIP49a and Pontin52) and RNAP II holoenzyme (with RUVBL1) (13,19,20). Furthermore, it is known that transcription factors such as Sp1, hepatocyte nuclear factor 1, and RNAP II holoenzyme are modified by O-GlcNAc and can be purified via wheat germ agglutinin-Sepharose column chromatography (1)(2)(3)(4). Therefore, these transcriptional machinery components may mediate the binding of p50/p47 to the wheat germ agglutinin column. Several yeast TAFs have been shown to be required for progression through the cell cycle: mutation of TAF145 leads to G 1 arrest (see below), functional inactivation of yTAF II 90 by ts mutations or depletion leads to arrest at the G 2 /M phase of the cell cycle (31), and ts mutants of TSM1/TAF150 arrest at the nonpermissive temperature as large-budded cells with a 2N DNA content (46). Northern blot analysis of several RNA transcripts in tih2-160 cells suggests that transcription of G 1 cyclin and ribosomal protein genes are down-regulated at the nonpermissive temperature. Other genes like actin, alcohol dehydrogenase, and the constitutive G 1 cyclin CLN3 (Fig. 7) as well as those transcribed by RNA polymerase I and III (data not shown) are unaffected. Because this down-regulation occurs even when the levels of other TAF proteins and TBP are normal, we propose that Tih2p has a primary role in the regulation of transcription.
Tih2p May Function Similar to yTAF II 145 to Regulate Transcription-yTAF II 145 is the only yTAF II known to contact TBP directly (38), and its higher eukaryotic homologue dTAF230/ hTAF250 is required for reconstitution of TFIID activity in vitro (47). However, yTAF II 145 is dispensable for global RNA polymerase II-directed transcription (32). Given the phenotypic similarities between taf145 and tih2 mutants, it is likely that Tih2p, or maybe a complex containing both Tih1p and Tih2p, is required for a mechanism of transcriptional regulation similar to that involving yTAF II 145. yTAF II 145 is required to mediate stimulation of RNAP II by transcriptional activators, leading to the initiation of transcription (48). Similarly, Tih2p (Tih1p-Tih2p complex) may also act as a bridge for TBP and genespecific activators to activate transcription. Consistent with this idea is the observation that Pontin52/p50 bridges ␤-catenin/lymphocyte enhancer factor-1 and TBP to form a multiprotein complex that is believed to transcriptionally activate target genes of Wnt/Wg signaling (20).
Functional Comparison to Other Helicases-Recently, p50 and p47, isolated as TIP49a and TIP49b, have been shown to be ATP-dependent DNA helicases of opposite polarity (18,22), and they exist in the same complex. This observation is reminiscent of the TFIIH complex, which also contains two DNA helicases, ERCC3(XPB/Ssl2p) and ERCC2(XPD/Rad3p), mediating opposite polarity of unwinding (49). Both SSL2 and RAD3 are essential genes and are required for gene transcription as well as DNA repair. Mutation of the Walker A domain causes lethality in ssl2⌬ cells but not in rad3⌬ cells (50,51). Together with results obtained in vitro using purified mammalian proteins (52), it appears that transcription initiation requires the helicase activity of Ssl2p but not that of Rad3p. On the other hand, both helicase activities are essential for DNA repair. The fact that a K81R mutation in Tih2p generates a lethal allele indicates that the Walker A domain of Tih2p performs an essential cellular function. A similar mutation introduced into the Walker A domain of Tih1p, K84R-Tih1p, also generates a lethal phenotype (data not shown). Taken together with the effect of Tih2p on transcription, these results suggest that Tih1p and Tih2p may be a novel class of helicase involved in transcriptional regulation via ATP utilization.
Recently, the mini-chromosome maintenance proteins, which are ATP-dependent helicases and are required for initiation of DNA replication (53), were reported to co-purify with RNA polymerase II and general trancription factors in high molecular weight holoenzyme complexes isolated from Xenopus oocytes and HeLa cells (54). The result suggests that minichromosome maintenance proteins function as components of the RNA polymerase II transcriptional apparatus. In yeast and mammalian cells, mini-chromosome maintenance proteins function as replication licensing factors but are far more abundant than the replication origins to which they bind (55). This abundance of mini-chromosome maintenance proteins may be now partly explained by their additional function in transcription. In yeast, abundance of Tih1p is about twice as much as Mcm3p (56). Although our data here suggest a relationship between Tih2p and the transcriptional machinery, other cellular functions of Tih1p, Tih2p, and hence the Tih1p-Tih2p complex are likely to be revealed by further biochemical and genetic analysis or interaction screening. FIG. 7. G 1 cyclin and ribosomal protein gene transcription, but not global RNA polymerase II-directed transcription, was defective in tih2-160 cells. A, total cellular RNA was isolated from TIH2, tih2-160, and an RNA polymerase II ts strain, rpb1-1, at the indicated times after a shift to 37°C. rpb1-1 served as a positive control producing little or no polymerase II transcripts. Total RNA of each strain (12 g) was used for Northern blotting analysis of genes as indicated on the left. B, 32 P-labeled oligo(dT) probe was hybridized to 2 g of total slot-blotted RNA prepared from strains indicated on the left. C, immunoblot analysis of whole-cell extracts prepared from TIH2 and tih2-160 strains for TBP, yTAF II 145, yTAF II 61, and yTAF II 90. advice; Richard Young (University of Tront) for a yeast mutant strain rpb1-1; Akiko Kobayashi and Yoshihio Tsukihashi (NAIST) for the preparation for anti-TBP and anti-TAF II antibodies; Kazumi Maekawa for excellent technical assistance; and Marc Lamphier for critical reading of the manuscript.