ABSTRACT
The multi-protein complex SL1, containing TBP, which is essential for RNA polymerase I catalyzed transcription, has been analyzed in fission yeast. It was immunopurified based on association of component subunits with epitope-tagged TBP. To enable this analysis, a strain of Schizosaccharomyces pombe was created where the only functional TBP coding sequences were those of FLAG-TBP. RNA polymerase I transcription components were fractionated from this strain and the TBP-associated polypeptides were subsequently immunopurified together with the epitope- tagged TBP. An assessment of the activity of this candidate SL1 complex was undertaken cross-species. This fission yeast TBP-containing complex displays two activities in redirecting transcriptional initiation of an S.pombe rDNA gene promoter cross-species in Saccharomyces cerevisiae transcription reactions: it both blocks an incorrect transcriptional start site at +7 and directs initiation at the correct site for S.pombe rRNA synthesis. This complex is essential for accurate initiation of the S.pombe rRNA gene: rRNA synthesis is reconstituted when this S.pombe TBP-containing complex is combined with a S.pombe fraction immunodepleted of TBP.
Activation of transcription of eukaryotic rRNA genes entails recognition of the promoter by an essential initiation factor, termed SL1 (1 ,2 ) or, alternatively, TIF-IB (3 -5 ), factor D (6 ,7 ) or Rib1 (8 ). This factor is critical in directing association of the catalytic enzyme, RNA polymerase I, with an SL1-rDNA gene promoter complex for initiation of the pre-~37-45S rRNA (9 ,10 ). In addition, SL1 confers species specificity to transcriptional initiation of eukaryotic rRNA genes, as is evident when even closely related species do not have the capability of directing correct transcriptional initiation of the other species' rRNA genes (reviewed in 7 ). The subunit structure of the SL1/TIF-IB factor has been determined in human and mouse and consists of TBP and three TBP-associated factors (TAFIs): TAFI110 (human), TAFI95 (mouse) and TAFI63 and TAFI48 (human and mouse; 1 ,11 ). However, it is not known whether an SL1 complex consisting of TBP and associated subunits is universal in eukaryotes and, if so, whether the subunit composition and mechanism of interaction with species-specific rDNA promoters varies (see for example 12 ). The first report of a multi-subunit complex required for rRNA synthesis in the yeast S.cerevisiae indicated that TBP was not a stably associated subunit (13 ).
TBP plays a central role in transcription catalyzed by all three nuclear RNA polymerases (14 -16 ), yet forms specific multi- subunit complexes that differ for each of the three polymerases (reviewed in 16 ). The polymerase II complex, TFIID, bears seven or more TBP-associated factors (TAFIIs; 17 -19 ), while the polymerase III complex, TFIIIB, bears two (20 -23 ). While the mouse and human SL1/TF-IB factors each contain TBP and three TAFIs, evidence suggests that the Acanthamoeba polymerase I essential initiation factor, TIF-IB, consists of TBP and four associated polypeptides (13 ). The candidates for SL1 subunits in bakers yeast, including Rrn6p, Rrn7p and a 66 kDa polypeptide, co-purified in a complex which did not contain TBP (13 ), although TBP was shown to fractionate with the initiation factor in early stages of purification (24 ). However, recent analyses revealed that TBP did associate with this polymerase I `core factor' complex and that Rrn11p was the 66 kDa polypeptide (25 ,26 ).
An exploration of the composition and activity of the essential initiation factor for rRNA synthesis was undertaken in fission yeast. To this end, the S.pombe tbp + gene (27 ,28 ) was disrupted and a strain of S.pombe created whose sole functional TBP was an epitope-tagged version. A complex was immunopurified that displayed SL1-like activity: it directed correct initiation of the S.pombe rRNA minigene cross-species in S.cerevisiae and repressed incorrect initiation. Reconstitution of S.pombe rRNA synthesis using homologous S.pombe factors was dependent on this complex.
A clone containing the genomic tbp+ gene (with an ~10 kb genomic insert in pDB248; kindly provided by Dr Alexander Hoffmann; 27 ) was partially digested with XbaI and a 2.6 kb fragment was gel isolated and subcloned into the pBS SK+ XbaI site (corresponding to positions 12 376-14 991; DDBJ/EMBL/GenBank accession no. Z66525). Following partial digestion of the resultant plasmid, p2.6tbp+, with HindIII, a 5.4 kb fragment was gel isolated which lacked 162 bp between the two HindIII sites of tbp+ (see Fig. 1 A). This was ligated to a 1.8 kb HindIII fragment containing the wild-type S.pombe ura4+ coding sequences, which was isolated from pREP2 (29 ,30 ). The resultant plasmid, p[Delta]tbp::ura4, bearing the disrupted tbp gene, was linearized with XbaI and the insert was transformed into S.pombe as described below.
A 3.3 kb XbaI fragment containing [Delta]tbp::ura4 sequences was gel isolated from p[Delta]tbp::ura4 (see Fig. 1 A) and 1 [mu]g was used to transform S.pombe SP826 (h+/h+ leu l-32//leu1-32 ura-4- D18/ura4-18 ade6-216/ade6-210; kindly sent by Dr Dave Frendewey). Ura+ transformants were selected (30 ) and Southern analysis was performed to assess whether a chromosomal allele of tbp+ was disrupted in the Ura+ transformants. Genomic DNA was extracted from 10 ml culture using the glass beads method as described (31 ). One fifth of the extracted nucleic acids from individual S.pombe Ura+ transformants was subjected to Southern analysis (31 ). Following fractionation on a 1% agarose-1* TBE gel and transfer to nitrocellulose, the DNA was cross-linked to the membrane using a GN Gene linker (BioRad) and hybridized in 50% formamide, 6* SSC, 0.1% SDS, 5* Denhardt's reagent, 50 mM Na3PO4, pH 6.5, 50 [mu]g/ml single-stranded calf thymus DNA and ~106 c.p.m. genomic tbp+ probe at 42oC overnight. As seen in Figure 1 B, an ~7.2 kb EcoRI fragment contains the TBP coding sequences in the parental diploid strain (lane 4), while one-step gene disruption of this diploid results in production of an ~8.8 kb EcoRI fragment containing tbp sequences (lane 3), due to the ura4+ coding sequences present in the disrupted copy, in addition to the 7.2 kb EcoRI fragment. Digestion of the genomic DNAs with EcoRI and HindIII liberates the same sized genomic fragment from both the parental and the diploid strains carrying the disrupted tbp allele, ~1.4 kb in size, as expected (Fig. 1 B, lanes 1 and 2).
Aliquots of 100 ng 1 kb XbaI fragment containing the 5' two-thirds of the S.pombe tbp+ gene (see Fig. 1 A) were labeled by the random priming method in the presence of dA,G and TTPs and 40 [mu]Ci [[alpha]32P]dCTP (NEN; >3000 Ci/mmol) (31 ). Random hexamers were obtained from Pharmacia.
The S.pombe tbp+ cDNA sequences were amplified utilizing a primer designed to insert coding sequences for the eight amino acid FLAGtm epitope tag (Kodak/IBI) following the initiating methionine at the N-terminus of tbp+. This position was shown to be neutral for insertion of an epitope tag in the human TBP coding sequences (33 ). To this end, the S.pombe TBP coding sequences [from the TFIID cDNA clone (27 ), kindly sent by Drs Alexander Hoffman and M.Horikoshi] were amplified using as forward primer 5'-GCCATATGGATTACAAAGACGATGACGACAAGGATTTCGCTTTACC, encoding MDYKDDDDKDFAL, and a vector-specific reverse primer. The FLAGtm tag consists of the eight amino acids DYKDDDDK (Kodak/IBI). The PCR products were treated with T4 DNA polymerase to convert the ends to blunt ends and ligated to SalI linkers. Following digestion with SalI and preparative isolation of the fragment, it was subsequently ligated into the SalI site of pBluescript SK+. Due to frequent deletions of the TBP coding sequence, this procedure and screening of E.coli Ampr transformants had to be repeated multiple times until a correct, full-length clone was isolated, pFLAG-S.p.TBP. The FLAG-TBP insert was released by digestion with NdeI and BamHI and ligated into the NdeI and BamHI sites of the S.pombe/E.coli shuttle vector pRep1 (29 ; kindly sent by Dr Kinsey Maundrell), creating pRep1/FLAG-S.p.TBP.
PlasmidpRep1/FLAG-S.p.TBP was introduced into a diploid S.pombe strain bearing a disrupted tbp+ allele (tbp/[Delta]tbp::ura4) and Leu+ transformants were selected. Following sporulation and selective killing of non-sporulating diploid cells, Leu+, Ura+, Ade- haploids were isolated. This procedure (suggested by Dr Henry Levin, NIH) involved treating ~1 ml diploid cells (~107 cells/ml) overnight with 20 [mu]l 1:10-fold dilution of glusulase; removing the glusulase and incubating the cells for 30 min in 30% ethanol before plating onto solid medium lacking leucine and uracil. Southern analysis revealed that these haploids contained the disrupted [Delta]tbp::ura4 chromosomal allele and the extrachromosomal FLAG-TBP coding sequences (data not shown). The resultant strain is Sp[Delta]TBP (h+ leu l-32 ura-4-D18 ade6-[Delta]tbp::ura4 pRep1/FLAG-S.p.TBP). The extrachromosomal plasmid pRep1/FLAG-S.p.TBP did not segregate during growth in rich medium, as expected, since it carried the only viable TBP coding sequences.
Sp[Delta]TBP was grown in thiamine-deficient EMM medium, to ensure maximal expression of FLAG-TBP (30 ), with constant vigorous shaking at 30oC and cells were collected while in mid logarithmic growth phase. S-100 was prepared from 40 l cells as described (32 ). Aliquots of 195 mg S-100 (total protein concentration; S-100 was made from 50 g pelleted and frozen cells) were adjusted to ammonium sulfate 60% saturation, centrifuged (15 000 g, 4oC, 15 min) and suspended in TA buffer (20 mM Tris-acetate, pH 7.5) as described (25 ). Following 4 h dialysis, the solution was diluted to 20-30 mg/ml and centrifuged at 10 000 g (25 ). The pellet was suspended in 0.2 ml buffer [20 mM HEPES-KOH, pH 7.9, 50 mM KCl, 5 mM EGTA, 0.05 mM EDTA, 2.5 mM dithiothreitol (DTT), 20% glycerol, 0.2 mM phenylmethylsulfonyl fluoride (PMSF)]. The suspended pellet (35 mg/ml) and the supernatant (1.5 ml with a protein concentration of 59 mg/ml) were stored at -75oC.
A sample of 31 mg protein (the supernatant following the second centrifugation, above) was adjusted to 1* loading buffer [25 mM HEPES, pH 7.9, 0.2 mM EDTA, 5 mM MgCl2, 20% glycerol (v/v), 1 mM PMSF, 1 mM DTT], 0.1 M KCl and loaded onto a 5 ml Pharmacia HiTrap-Q column pre-equilibrated with 0.1 M KCl, 1* loading buffer. This was followed by step elution of fractions at 1* loading buffer with 0.175, 0.35 , 0.7 and 1.0 M KCl, at a flow rate of 1 ml/min (controlled by a Pharmacia Gradifrac System). All steps were repeated with a second batch of ~30 mg protein.
The peak fractions for RNA polymerase I transcription components were pooled from two trials (~10 mg total protein). The transcriptional activity was assessed as described below (see also Fig. 4 A). The KCl concentration was adjusted to 0.15 M and the samples were loaded onto a 1 ml anti-FLAGtm M2 affinity gel (Kodak/IBI; binding capacity 25 nmol FLAG protein/ml gel), as recommended. The flow-through was collected and re-loaded; the column was washed three times with phosphate-buffered saline, pH 7.4. Elution of FLAG-TBP and associated polypeptides was conducted using 1 ml elution buffer with increasing concentrations of FLAG peptide (50, 100, 125 and 200 ng/ml), followed by 1 ml 0.1 M glycine-HCl, pH 3.0. The fraction containing unbound polypeptides was demonstrated to lack TBP by Western analysis (data not shown) and served as the fraction immunodepleted of TBP (see Fig. 7 A and B).
An aliquot of 20 [mu]g protein from S-100 extract, or fractions as indicated, was fractionated by 15% SDS-PAGE and transferred to nitrocellulose using a BioRad SD Semi-Dry Transfer Cell. The anti-FLAG antibody (anti-FLAGtm M2 monoclonal antibody; Kodak/IBI) was added at 1:300 dilution, following standard protocols (31 ,33 ). Detection utilized enhanced chemiluminiscent reagents from Amersham and included sheep horseradish peroxidase-linked anti-mouse Ig whole antibody as the secondary antibody.
The SDS-PAGE minigel containing fractionated TBP and associated polypeptides (Fig. 5 A and B) was immersed in 50 ml 1:5000 dilution of SYPRO orange protein stain (BioRad) in 7.5% (v/v) acetic acid and stained and photographed under UV light (34 ).
p-243:XH (Fig. 3 ) has been described previously (32 ). p5'[Delta]-243/3'[Delta]+31 is a similar template containing promoter sequences from -243 to +31, but in the cloning vector pBS (32 ); it was used in the transcriptional assays shown in Figure 6 A-C. Template p5'[Delta]-243/3'[Delta]+89 (32 ) was used in the transcription assays shown in Figure 7 (at a final concentration of 0.5 [mu]g/ml). In vitro transcription reactions were 40 [mu]l and contained 10 [mu]l extract (~50 [mu]g protein) or indicated amounts of fractionated RNA polymerase I transcription components, template DNA (0.025-1.25 [mu]g/ml, as indicated), 20 mM HEPES-KOH, pH 7.9, 70-90 mM KCl, 10 mM MgCl2, 5 mM EGTA, pH 7.9, 0.05 mM EDTA, pH 7.9, 10% glycerol, 1.3 mM DTT, 100-500 [mu]M each of the four ribonucleoside triphosphates (Pharmacia) and 10 [mu]g/ml [alpha]-amanitin (Sigma) and were incubated for 45 min at 26oC (32 ). Transcription in an S.cerevisiae S-100 extract (32 ,35 ) was performed as above, except that ~1 ng FLAG-TBP complex was added, with the final concentration of transcription buffer adjusted to be the same as above. The reconstitution assays (shown in Fig. 7 ) utilized 5 [mu]l S.pombe fraction depleted of TBP and/or 1.0 [mu]l (~1 ng) FLAG-TBP complex. The RNA was isolated and S1 analysis conducted as described (32 ). S1-protected fragments were resolved by electrophoresis on 4% acrylamide- 9 M urea gels; the size markers were 5'-32P-labeled HpaII fragments derived from pBR322. The 5'-end-labeled probe used to detect transcription supported by template p-243:XH was prepared by labeling at the unique XbaI site (+340) on the template strand and converting the DNA to single-stranded as described (32 ). For preparation of the probe used to detect transcription supported by p5'[Delta]-243/3'[Delta]+31, 5'-end-labeling was at position +77 on the template strand (at a unique XhoI site). The probe used to detect transcription supported by p5'-243/3'[Delta]+89 probe was labeled at a unique XhoI site at +135 on the template strand. The initiation site for in vitro S.pombe rRNA synthesis was shown to be the same in vitro as in vivo (32 ; see also Fig. 6 C). S1-protected fragments representing rRNAs initiated in vitro in S.cerevisiae or S.pombe S-100 extracts were electrophoresed on sequencing gels next to Maxam-Gilbert sequencing reactions for mapping the start site (see Fig. 6 C; 31 ).
Schizosaccharomyces pombe S-100 extract made from wild-type strain 972 (h-; kindly sent by Dr H.Levin) was used in control transcription reactions (32 ). The diploid strain used for disruption of tbp+ was S.pombe SP826 h+/h+ leu l-32 ura-4-D18/ura4-18 ade6-216/ade6-210 (kindly sent by Dr Dave Frendewey, NYU Medical Center). The bacterial strains used included: XL1-Blue [endA1, hsdR17 (rk-, mk+), supE44, thi-1, [lambda]-, recA1, gyrA96, relA1, lac, (F', proAB, lacIqZ[Delta]M15, Tn10, (tetr)] and SUREtm [mcrA, [Delta](mcrBC-hsdRMS-mrr)171, endA1, supE44, thi-a, [lambda]-, gyrA96, relA1, lac, recB, recJ, sbcC, umuC::Tn5, (kanr), uvrC, (F', proAB, lacIqZ[Delta]M15, Tn10, (tetr)] (Stratagene). The S.cerevisiae S-100 extract was made from S.cerevisiae W303 [MAT[alpha], ade2-1, his3-11,15, leu-23,112, trp1-1, ura3-1, can1-100 (R.Rothstein)]. Media used included EMM (31 ) and SC medium (32 ). Bacterial cells were transformed by electroporation using a BioRad Gene Pulser (32 ). The lithium acetate transformation method was used for introduction of plasmid DNAs into yeast (31 ,32 ).
Our strategy to determine whether the essential initiation factor for rRNA synthesis in fission yeast was stably associated with TBP was to introduce a tagged copy of TBP into S.pombe. This would facilitate detection and localization of TBP during fractionation of RNA polymerase I transcription components and enable immunoaffinity purification of associated factors. To ensure that all TBP coding sequences were epitope tagged, the chromosomal copy of the tbp+ gene was inactivated in a diploid strain using one-step gene disruption (36 ). Such a strain would also facilitate analysis of interactions of the essential initiation factor for RNA polymerase I, SL1, with other RNA polymerase I transcription factors and with the regulatory regions of the S.pombe rRNA gene.
A plasmid was constructed that contained a disrupted copy of the S.pombe TBP coding sequences, named p[Delta]tbp::ura4 (see Fig. 1 A and Materials and Methods for details). One-step gene disruption of the chromosomal tbp+ allele was conducted in diploid strain SP826 of S.pombe, since TBP is an essential gene. Southern analysis confirmed that gene replacement was successful and that a diploid strain was constructed containing one wild-type and one disrupted allele of TBP (see Fig. 1 B). A plasmid bearing an epitope-tagged version of tbp+ cDNA,pRep1/FLAG-S.p.TBP, was constructed and introduced into this [Delta]tbp::ura4/tbp+ diploid strain of S.pombe (see Materials and Methods for details on construction of the FLAGtm epitope-tagged TBP). To ensure high levels of expression, the TBP coding sequences were placed under the control of the nmt promoter (29 ,30 ).
The essential initiation factor for polymerase I catalyzed transcription was fractionated from the resultant haploid strain of S.pombe, based on its presence in transcriptionally active fractions and on affinity purification via the epitope-tagged TBP. The fractionation scheme for purification of the essential initiation factor for rRNA synthesis is outlined in Figure 2 . Polypeptides that precipitated at 60% ammonium sulfate were collected and dialyzed, as described in Riggs et al. (24 ). However, the S.pombe RNA polymerase I transcription components behaved differently from those of S.cerevisiae, where required RNA polymerase I transcription components formed a sedimentable complex following dialysis of the suspended ammonium sulfate precipitated polypeptides (24 ). In the case of the S.pombe RNA polymerase I transcription factors, they were largely present in the `low salt supernatant' (Fig. 3 , lane 2), although a fraction did form a sedimentable complex (Fig. 3 , lane 3).
Formation of the complex assembly of factors required to direct correct initiation of eukaryotic rRNA genes involves association of the essential initiation factor SL1 (also called TIF-IB, Rib1 and factor D; 1 -8 ) at an early step in this process (7 ,41 ). This association is promoted by UBF in vertebrates (8 ,41 -43 ), by an enhancer binding factor in Acanthamoeba (44 ) and apparently by an upstream activating factor, UAF (45 ), in S.cerevisiae. An rDNA transcriptional stimulatory activity of S.pombe forms a stable complex with the rDNA promoter and may also promote association of SL1 (Chen,L., Zhao,A., Liu,Z., Boukghalter,B. and Pape,L., submitted for publication).
While TBP is a component of the essential initiation complex for all three nuclear RNA polymerases in yeast (14 ,15 ), its association with the essential initiation factor for rRNA synthesis initially appeared less stable in the yeast S.cerevisiae (13 ) than was the case for mammalian SL1 complexes (4 ; TIF-IB; 4 ). The TFIID initiation factor for RNA polymerase II catalyzed transcription was initially isolated as the TBP monomer from yeast (46 ), but both TFIIIB and TFIID were later shown to consist of multiple subunits (47 -48 ), akin to the analogous complexes in higher eukaryotes (17 ).
In S.cerevisiae, three of the subunits of an essential transcription factor for rRNA synthesis are Rrn6p, Rrn7p (13 ) and Rrn11p (25 ,26 ). Very recent results demonstrate that these subunits associate with TBP (25 ,26 ). In this paper, we have shown that a fission yeast complex can be immunopurified from active RNA polymerase I transcription components consisting of a tagged TBP and TBP-associated polypeptides. Furthermore, this complex is capable of repressing an incorrect transcriptional start site on a S.pombe rDNA promoter and promoting the correct start cross-species. It is of interest that the yeast, human and mouse subunits of the essential initiation factor for rRNA synthesis show a similar polypeptide profile. Comparison of the profile of polypeptides co-eluting with TBP and with the profile of polypeptides in the human and mouse SL1 complex (1 ,4 ,11 ), as well as in the multi-subunit initiation complex for S.cerevisiae (13 ,25 ,26 ), suggests that the three TBP-associated polypeptides that may be the S.pombe SL1 TAFIs are the ~64, 74 and ~101 kDa polypeptides, however, assignment awaits sequence determination of these polypeptides. Polypeptides present that are not bona fide TAFIs could be contaminating polypeptides that are detected upon affinity purification of TBP-associated factors (48 ) and yeast TFIID (49 ) or subunits of other polymerase-TBP complexes or subunits of a polymerase I UAF-like complex (50 ).
It remains to be determined what the primary sequence of the S.pombe SL1 TAFs are and whether they share homology with human TAFI110, TAFI63 and TAFI48 (1 ). The subunits of the essential initiation factor for rRNA synthesis in S.cerevisiae, Rrn6p, Rrn7p and Rrn11p (p66) (13 ,25 ,26 ), are unrelated in primary sequence to the mammalian SL1 subunits and efforts to isolate coding sequences for the S.pombe subunits utilizing heterologous mammalian or S.cerevisiae probes have been unsuccessful, suggesting that their primary sequences may also vary significantly from other SL1/core factor subunits.
The association of the essential RNA polymerase I initiation factor with the rDNA core promoter region is critical for rRNA synthesis, but stimulatory factors are required to stabilize this interaction (8 ,42 -46 ,51 ). We have found that S.pombe SL1 can form a weak complex with the S.pombe rDNA promoter (data not shown). Figure 6 D shows a comparison of the core rDNA promoter sequences of S.pombe (32 ) with those of S.cerevisiae (52 ,53 ) and may explain why a transcription start, albeit aberrant, is seen in cross-species transcription of an S.pombe rDNA promoter with S.cerevisiae polymerase I transcription components (Fig. 6 A). Conserved regions extending between -26 and -14 and between -10 and -3 may direct basal level cross-species initiation dependent on this core rDNA promoter, but at an altered initiation site (54 ). Addition of the putative S.pombe SL1 complex results in correct recognition of and association with its own species promoter to direct initiation at the natural start site.
It has not been possible to identify homologous TAFI-encoding genomic sequences in S.pombe as of yet by searching S.pombe sequence databases. This may be due to sequence heterogeneity for all of the TAFIs or simply that the genomic region encoding the TAFIs has not been sequenced. While the S.cerevisiae TAFIIs are highly homologous to their human counterparts (49 ), none of the subunits of the S.cerevisiae essential RNA polymerase I transcription factor, Rrn6p, Rrn7p or Rrn11p (13 ,25 ,26 ), show any defining homology to the TAFI110 and TAFI48 or TAFI63 polypeptides (1 ,55 ). This lends further evidence to differences in factors and mechanisms involved in species-specific rDNA promoter activation. The identity of the interactions directing species-specific RNA polymerase I transcriptional initiation will further our understanding of the evolution of species-specific cis-acting regulatory elements of eukaryotic rRNA genes and, in turn, of the corresponding RNA polymerase I transcriptional machinery that correctly transcribes only its target genes, in both a polymerase class- and a species-specific manner.
This work was supported by NSF grant MCB-9219220 to L.K.P. and in part by NYU Research Challenge Fund grants. NSF is also thanked for its support of the computing resources through grant BIR-9318128. We thank Drs Alexander Hoffman, Masami Horikoshi and Robert Roeder for S.pombe TBP genomic and cDNA clones and suggestions and we thank Drs Henry Levin and Dave Frendewey for useful suggestions and discussions.
*To whom correspondence should be addressed. Tel: +1 212 998 8444; Fax: +1 212 260 7905; Email: papel01@mcrcr.med.nyu.edu
+Present address: Department of Microbiology, Ohio State University, Columbus, OH 43210, USA
REFERENCES
Return