Affinity Purification of Mammalian RNA Polymerase I

Overlapping cDNA clones encoding the two largest subunits of rat RNA polymerase I, designated A194 and A127, were isolated from a Reuber hepatoma cDNA library. Analyses of the deduced amino acid sequences revealed that A194 and A127 are the homologues of yeast A190 and A135 and have homology to the β′ and β subunits of Escherichia coli RNA polymerase I. Antibodies raised against the recombinant A194 and A127 proteins recognized single proteins of approximately 190 and 120 kDa on Western blots of total cellular proteins of mammalian origin. N1S1 cell lines expressing recombinant His-tagged A194 and FLAG-tagged A127 proteins were isolated. These proteins were incorporated into functional RNA polymerase I complexes, and active enzyme, containing FLAG-tagged A127, could be immunopurified to approximately 80% homogeneity in a single chromatographic step over an anti-FLAG affinity column. Immunoprecipitation of A194 from 32P metabolically labeled cells with anti-A194 antiserum demonstrated that this subunit is a phosphoprotein. Incubation of the FLAG affinity-purified RNA polymerase I complex with [γ-32P]ATP resulted in autophosphorylation of the A194 subunit of RPI, indicating the presence of associated kinase(s). One of these kinases was demonstrated to be CK2, a serine/threonine protein kinase implicated in the regulation of cell growth and proliferation.

The nuclei of eukaryotic cells contain three classes of RNA polymerase enzymes, RNA polymerase I, II, and III. RNA polymerase I (RPI) 1 (1) is the central component of the machinery that transcribes the ribosomal RNA genes (rDNA). As such, it is responsible for 40 -60% of the RNA synthesis in the cell. RPI is a large, complex enzyme with a M r ranging between 500,000 and 600,000 (1). Yeast RPI consists of 14 subunits (2). The exact subunit composition of the mammalian enzyme has not been established yet.
Molecular and immunological studies of the two largest subunits of the three nuclear RNA polymerases revealed that these subunits are related to the ␤Ј and ␤ subunits of the bacterial RNA polymerase (3)(4)(5)(6)(7)(8). Biochemical studies suggest that the two largest subunits are likely to be involved in template binding, but the domains of the two proteins involved in this binding have not been defined (reviewed in Ref. 1). While both subunits can be cross-linked to nascent RNA chains (9 -11), the second largest subunit was identified, by affinity labeling techniques (12), as the subunit responsible for nucleotide binding (10). This subunit has also been shown to influence the selection of the transcription initiation site (1). Biochemical and genetic analyses demonstrated that the largest subunit plays a key role in chain elongation (13)(14)(15). For example, resistance to the drug ␣-amanitin, which interferes with chain elongation, maps to the largest subunit (16). With the exception of yeast RPI, our knowledge of the metazoan forms of RPI is very limited. Detailed analysis of the structure and function of the enzyme is required if we are to fully understand the mechanisms by which rDNA transcription is directed and/or regulated.
This problem is exemplified by the studies on the regulation of rDNA transcription in response to stimuli that acutely downregulate transcription. When some mammalian cell lines are subjected to various conditions that inhibit cell growth or protein synthesis, ribosomal RNA synthesis ceases with a half-life of approximately 30 min (17)(18)(19)(20). In vitro transcription assays have demonstrated that the RPI isolated from these cells is unable to initiate specific transcription, although it can "transcribe" a nonspecific template such as calf thymus DNA or poly(dA-dT) (17, 19 -21). Two laboratories have reported the purification of an RPI-associated factor that can restore the ability of the enzyme to initiate specifically (22,23). However, the subunit compositions of these factors, referred to as TF1C or TIF-1A, differ. Thus, it is not clear if they are the same factor, although both are tightly associated with the polymerase (22,23). More recently, three polymerase-associated factors, termed PAF53, PAF51, and PAF49 with respect to the sizes of the proteins, have been identified (24). One of these factors, PAF53, coimmunoprecipitates with RPI and can interact directly with the auxiliary rDNA transcription factor UBF, in vitro. Antibodies to PAF53 block specific transcription from the rDNA promoter, suggesting that this factor may be involved in the formation of the initiation complex. Immunolocalization studies indicate that PAF53 is associated with RNA polymerase I in logarithmically growing NIH3T3 but not quiescent cells (24). Thus, in many respects, the properties of PAF53 are similar to those of TF1C/TIF-A (24). Clearly, eluci-dation of the relationship of these factors to one another in vivo and to RNA polymerase itself requires that the structure of the polymerase be determined.
Recent studies have successfully used immunoaffinity techniques to purify RNA polymerase II (RPII) complexes (25)(26)(27). These approaches are significantly more rapid than conventional biochemical purification schemes and have allowed for the identification of a wide range of closely associated suppressors and co-activators of RPII (25)(26)(27). It follows that similar immunoaffinity purification schemes specific for RPI should be of use in identifying coregulators of RPI transcription. Accordingly, we have set a goal to obtain molecular and immunological tools sufficient to affinity-purify RPI and its associated factors.
In this paper we report the cloning of the two largest subunits of rat RPI, designated A194 and A127, respectively, based on their predicted sizes of 194 and 127 kDa. Antibodies raised to domains of these subunits expressed in E. coli recognized proteins of apparent masses 190 and 120 kDa, respectively, in mammalian tissues. Affinity purification experiments using extracts from cell lines expressing His-tagged A194 or FLAGtagged A127 subunits demonstrated that the recombinant proteins assembled with the endogenous RPI subunits and that the affinity-purified RPI was capable of RNA synthesis. Furthermore, we have used the FLAG affinity purification scheme to demonstrate that CK2, a serine/threonine protein kinase implicated in the regulation of cell growth and proliferation (28 -35), co-affinity-purifies with the RPI complex. In vitro kinase assays indicate that the major proteins phosphorylated by the copurifying CK2 include the 194-kDa subunit and an unidentified 51-kDa polypeptide.

MATERIALS AND METHODS
Cloning of the Rat RNA Polymerase I Homologue of the ␤Ј Subunit of the Bacterial RNA Polymerase-A partial clone of the Xenopus laevis RPI ␤Ј subunit 2 was used to screen a Reuber hepatoma cDNA library in LambdaZap (CLONTECH) by in situ hybridization (36,37). Bluescript plasmids containing inserts were isolated from positive LambdaZap clones by in vivo excision using the appropriate helper phage. The DNA sequences of the cDNAs were determined by chain termination sequencing (37), and an open reading frame of 5151 nucleotides was identified. Analysis of the deduced amino acid sequence was carried out using the routines provided in DNA* and DNASIS. Pairwise and multiple alignments were carried out using the method of Clustal. Multiple alignments were carried out with a Gap penalty set to 10. The settings for pairwise alignment were Ktuple ϭ 1, Gap penalty ϭ 3, Window ϭ 5 to optimize the alignments. The carboxyl-terminal domain of the ␤Ј subunits of yeast and mouse RPII were deleted from the sequences for pairwise alignment.
Cloning of the Rat RNA Polymerase I Homologue of the ␤ Subunit of the Bacterial RNA Polymerase-The initial fragment of the rat RPI ␤ subunit was generated by nested reverse transcription-PCR using N1S1 cell mRNA and degenerate oligonucleotides. Reverse transcription of at N1S1 cell mRNA was carried out using random hexamers and Moloney murine leukemia virus reverse transcriptase under standard conditions. PCR was then performed using degenerate oligonucleotide primers based on the amino acid sequences MYQCQM and GEMERD under standard conditions, followed by a second round of PCR using degenerate oligonucleotide primers based on the amino acid sequences GKQTMG and RLRHMV. The initial PCR product was used to screen a Reuber hepatoma cDNA library in LambdaZap (CLONTECH) yielding two clones, pB15-2 and pB19 -1. Bluescript plasmids containing the inserts were isolated, and the DNA sequences were determined by chain termination sequencing (37). The amino-terminal end of rat RNA polymerase I ␤ was obtained by 5Ј-RACE using rat liver 5Ј-RACE-Ready cDNA (CLONTECH) and nested primers based on sequences 230 -290 nucleotides downstream of the 5Ј-end of clone PB15-2. The products of the 5Ј-RACE reaction were cloned and sequenced, and in combination with constructs PB15-1 and pB19 -1, an open reading frame of 3462 nucleotides was identified. Analysis of the deduced amino acid sequence was carried out using the routines proved in DNA* and DNASIS as described above.
Antibody Preparation-The deduced amino acid sequences of the clones for the putative A194 and A127 subunits were compared with those of the other eukaryotic ␤Ј and ␤ homologues. One domain of the putative A127 protein, amino acids 710 -1007 and two domains of the putative A194 protein, amino acids 231-429 and 1272-1491, were identified that shared very little identity with the sequence of the homologous subunits of either RPII or RPIII. The nucleotides coding for these three domains were amplified with specific PCR primers containing the appropriate restriction endonuclease sites and cloned into pET19 (Novagen). The PCR primer for the 5Ј-end of each cDNA coded for a His 6 leader. Positive clones were identified by in situ hybridization and were sequenced. Recombinant protein was induced with isopropyl-1-thio-␤-D-galactopyranoside and purified using Ni ϩ affinity resins as recommended by the supplier (Qiagen). These proteins were used to immunize rabbits using RIBI adjuvant at a dose and schedule recommended by the manufacturer (RIBI). The antisera are referred to as either anti-A127 710 -1007 or anti-A194 231-429 and anti-A194 1272-1491 .
Immunoblotting and Silver Staining-Protein determinations were performed using the Bio-Rad Bradford assay kit with bovine serum albumin as the protein standard. Western blots were carried out as described previously (39). A127 and A194, A40, PAF53, and CK2 (Upstate Biotechnology, Inc.) were detected by incubating the filters with a 1:5000 dilution of the appropriate antibody, followed by incubation with horseradish peroxidase-conjugated anti-rabbit antibodies (Amersham Corp.). FLAG-A127 was detected by incubating the filters with a 1:2000 dilution of the M2-anti-FLAG antibody (Eastman Kodak Co.) followed by incubation with horseradish peroxidase-conjugated anti-mouse antibodies (Amersham). Immunoreactive proteins were visualized by the Enhanced Chemiluminescence (ECL) method (Amersham). The molecular sizes of the immunodetected proteins were verified by comparison to the migration of prestained protein markers (Bio-Rad) electrophoresed in parallel lanes. Silver staining was performed with 1-2 mg of anti-FLAG affinity-purified RPI complex according to standard protocols (37).
Preparation of S100 Fractions and Nuclear Extracts-S100 fractions were prepared from cultured N1S1 cells by standard procedures (41). Nuclear extracts of rat Novikoff hepatoma ascites cells were prepared and fractionated as described previously (42,43). The extracts were frozen in liquid N2 and stored at Ϫ80°C.
In Vitro Transcription-Random and specific transcription by RPI was carried out using nuclear and S100 extracts from N1S1 cells essentially as described previously (43)(44)(45). To determine the effect of anti-A194 antibodies on specific transcription from the rDNA promoter, N1S1 nuclear extracts were incubated with either the affinity-purified anti-A194 231-429 antibodies or preimmune antibodies immobilized on Affi-Gel HZ resin (Bio-Rad) for 2 h at 4°C. After the incubation, the samples were clarified by centrifugation at 13,000 rpm for 10 s, and the supernatants were transferred to clean microcentrifuge tubes. Subsequent transcription was carried out with 10 l of the antibody-treated nuclear extracts and 1 g of EcoRI linearized template DNA, pU5.1E/X (42), which contains the rat 45 S rDNA promoter (Ϫ286 to ϩ630). The in vitro synthesized RNA was purified and analyzed by urea-polyacrylamide gel electrophoresis and autoradiography (41).
The ability of immunopurified FLAG-tagged RPI to carry out nonspecific RNA synthesis was carried out using S-100 extracts fractionated by anti-FLAG affinity resin chromatography. Nonspecific transcription reactions contained 25 l of the starting S100 extracts or anti-FLAG affinity column eluate and nicked calf thymus DNA as template as described previously (44). The radioactivity incorporated was determined by liquid scintillation spectrophotometry.
Construction of A194 and A127 Expression Plasmids-polBЈUN-TRHIS was generated by PCR from the entire coding region of the rat A194 using a primer pair that introduced an additional 30 nucleotides complementary to the coding region of a His 10 tag immediately before the stop codon in the 3Ј-end of the A194 coding region. The PCR products were cloned directly into the pCR3.1 vector (Invitrogen), which drives mammalian expression of A194 under the control of the cytomegalovirus promoter. pBFLAG39 was generated by PCR from the complete coding region for rat A127 using a primer pair that introduced 24 nucleotides coding for the FLAG peptide between the start ATG and the second codon of the 5Ј-end of the A127 cDNA. The PCR products were cloned into the XhoI site of pCDNA3 (Invitrogen), which uses the cytomegalovirus promoter to drive expression in mammalian cells. The orientations of all inserts were confirmed by sequencing.
Preparation of Cell Lines Expressing His-tagged A194 and FLAGtagged A127-N1S1 cell lines expressing His-tagged A194 or FLAGtagged A127 were prepared by standard techniques. Briefly, 2 ϫ 10 6 N1S1 cells were transfected with 2 g of polBЈUNTRHIS or pBFLAG39 using DMRIE-C (Life Technologies, Inc.) according to the manufacturer's instructions. After 5 h, the media were replaced with RPMI media containing 5% fetal calf serum and 300 g/ml G418. After 10 days, the cells were diluted in 96-well plates. After 2 weeks with regular media and antibiotic changes the cell lines were tested for expression of His-tagged A194 or FLAG-tagged A127 by Western analysis. Cell lines were routinely maintained in spinner flasks containing RPMI and 5% horse serum with G418 at 300 g/ml. S100 extracts were prepared from N1S1 cells as described above.
Ligand and Immunoaffinity Purification of RNA Polymerase I-Anti-A194 IgG was affinity-purified against an A194 peptide, residues 231-429, and then affinity-purified IgG (1.25 mg) was coupled to 0.5 ml of packed Affi-Gel HZ resin according to the manufacturer's instructions. Control resin was prepared by coupling an equivalent amount of rabbit IgG purified from preimmune antiserum to 0.5 ml of Affi-Gel HZ resin. Before use, the resin was equilibrated in 10 column volumes of C/20 buffer (43) containing 150 mM NaCl (C-20/150). The resins were tumbled for 2 h with either 0.5 ml of S100 or nuclear extracts, after which they were packed into columns and washed with 10 column volumes of C-20/150. Bound proteins were then eluted with 0.1 M glycine, pH 2.8, and neutralized by the addition of 1 M Tris-HCl, pH 8.0. The eluted fractions as well as the flow-through were analyzed by immunoblot analysis as described above.
RPI complexes containing FLAG-tagged A127 were immunopurified using 0.2 ml of packed anti-FLAG resin (2.5 mg/ml, ICI/Kodak) equilibrated in 150 mM C-20 (C-20/150) and 0.1% Nonidet P-40. The resin was tumbled for 2 h with 0.5 ml of S100 fraction, packed into columns, and washed with 10 column volumes of C-20/150 containing 0.1% Nonidet P-40. Bound proteins were then eluted with C-20/150 containing 0.5 mg/ml FLAG peptide. The eluted fractions as well as the flow-through were analyzed by immunoblot analysis as described above. To control for proteins that might nonspecifically bind to the FLAG affinity resin, S100 extracts from nontransfected control N1S1 cells were subjected to affinity purification in parallel with reactions containing tagged proteins.
RPI complexes containing His-tagged A194 were purified using 0.2 ml of packed Talon metal affinity resin (Stratagene) equilibrated in C-20/150 containing 0.1% Nonidet P-40 and 20 mM imidazole. The resin was tumbled for 2 h with 0.5 ml of S100 fraction, packed into columns, and washed with 10 column volumes of C-20/150 containing 0.1% Nonidet P-40 and 20 mM imidazole. Bound proteins were then eluted with 0.1 M EDTA, pH 8.0. The eluted fractions as well as the flow-through were analyzed by immunoblot analysis as described above. To control for proteins that might nonspecifically bind to the Talon affinity resin, S100 extracts from nontransfected control N1S1 cells were subjected to affinity purification in parallel with extracts containing tagged proteins.
In Vivo Phosphorylation of A194 -Analysis of in vivo phosphorylation of the A194 subunit of RPI was carried out by immunoprecipitation of [ 32 P]orthophosphate-labeled A194 protein from CHO cell extracts using polyclonal anti-A194 231-429 antiserum. Exponentially growing CHO cells, maintained in DMEM and 10% fetal bovine serum, were metabolically labeled with [ 32 P]orthophosphate (1 mCi/60-mm dish) for 24 h (37). For metabolic labeling of serum-starved CHO cells, the media of exponentially growing CHO cells were replaced with DMEM containing 0.5% fetal bovine serum and [ 32 P]orthophosphate (1 mCi/60-mm dish), and the cells were incubated for an additional 24 h. After labeling, the cells were washed three times in phosphate-buffered saline and scraped directly into 500 l of modified radioimmune precipitation buffer (20 mM Tris-HCl, pH 8.0, 137 mM NaCl, 10% glycerol, 2 mM EDTA, 0.5% Nonidet P-40, 0.3% SDS, 0.5% deoxycholate) containing appropriate protease inhibitors. 50 l of affinity-purified anti-A194 231-429 antiserum coupled to 50 l of protein A-agarose beads was added to the lysates, and then the mixture was incubated for 2 h at 4°C. After four 1-ml washes in modified radioimmune precipitation buffer, the beads were boiled for 10 min in the presence of 2 ϫ Laemmli sample buffer. Phosphorylated A194 was resolved by SDS-PAGE and visualized by autoradiography. Autoradiograms were analyzed using a Molecular Dynamics PhosphorImager.
Protein Kinase Assays-Protein kinase assays were performed with 5 l of FLAG-tagged RPI affinity-purified as described above. Reactions were carried out in a total volume of 50 l containing 20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovana-date, 1 mM dithiothreitol, 12.5 mM MgCl 2 , 90 mM ATP, and 150 Ci of [␥-32 P]ATP (6000 Ci/mmol). In some cases, recombinant CK2 (Upstate Biotechnology), heparin, and GTP were added to the reaction buffer at the concentrations indicated. Reactions were carried out at 30°C for 20 min and terminated by the addition of 10 l of 6 ϫ Laemmli sample buffer.
Preparation of Figures-Autoradiograms obtained from Western, Northern, and transcription assays were scanned and converted to 8-bit TIFF bitmap files using a laser densitometer and ImageQuant software (Molecular Dynamics). The TIFF bitmap files were subsequently imported into Corel Draw!, where labels and headings were attached. Completed figures were output to paper on a Tektronix Phaser 440 printer.

RESULTS
Isolation of cDNA Clones of the Rat RNA Polymerase I ␤Ј and ␤ Subunits-To isolate cDNA clones of the largest subunit of rat RPI (A194), a rat Reuber hepatoma cDNA library was screened at low stringency with a 32 P-labeled partial cDNA clone of the equivalent subunit of X. laevis RPI. 2 Five rat cDNA clones were isolated and sequenced on both strands. Two overlapping clones containing an open reading frame of 5151 nucleotides were used to prepare a construct comprising the entire coding region of rat A194 (pBSrPolFL). Hybridization of the 32 P-labeled pBsrPolFL cDNA insert to 10 g of poly(A ϩ ) RNA isolated from N1S1 cells identified a discrete RNA species of approximately 6.2 kilobases, which compares favorably with the size of the cloned A194 cDNA (results not shown). To clone the second largest subunit of rat RPI, degenerate oligonucleotides were designed based on the amino acid sequence of highly conserved regions identified from multiple alignments of other eukaryotic RPI ␤ homologues. These primers were used in a reverse transcription-PCR reaction of N1S1 cell RNA, and the resulting DNA fragments were used to probe a LambdaZap rat Reuber hepatoma cDNA library. The longest clone obtained, PB15-2, lacked the NH 2 -terminal coding region including the ATG codon. 5Ј-RACE was carried out to clone the 5Ј-end of the coding region of the ␤ subunit of rat RPI. The sequence of the longest RACE product obtained was cloned into PCRII (Invitrogen) and used in conjunction with PB15-2 to prepare a construct consisting of the entire coding region of the rat A127 mRNA (pBFL). Hybridization of 32 P-labeled pBFL cDNA insert to 10 g of poly(A ϩ ) RNA isolated from N1S1 cells identified a discrete RNA species of approximately 3.9 kilobases, which compares favorably with the size of the cloned A127 cDNA (results not shown).
Sequence Alignments of Rat A194 and A127 with Eukaryotic Homologues-Alignment of the amino acid sequences derived from the A194 and A127 cDNA clones (Fig. 1, A and B) with the equivalent subunits of RNA polymerase I, polymerase II, and polymerase III from other organisms demonstrated that we had cloned the ␤Ј and ␤ subunits of rat RNA polymerase I. The putative rat A194 was more closely related to the 190-kDa subunit of RNA polymerase I of Saccharomyces cerevisiae (32.3% similar) than it was related to the ␤Ј subunit of mouse RNA polymerase II (17.8%). In fact, the ␤Ј subunits of yeast RNA polymerase I and yeast RNA polymerase II were only 17.4% similar. Similarly, the putative rat A127 was more closely related to the equivalent RNA polymerase I subunits of yeast (40.5% similar) and Drosophila (36.4% similar) than it was to those of RNA polymerase II (23.4 and 21.9% similar, respectively) or RNA polymerase III (21.2 and 24.0% similar, respectively). These findings support the conclusion that the cloned A194 and A127 cDNAs are the rat RP1 homologues of the prokaryotic ␤Ј and ␤ subunits.
Immunological Demonstration of the Identity of the Putative A194 and A127 Clones-Two polyclonal antibodies, A194 231-429 and A194 1272-1491 were raised to discrete regions in the amino-terminal (amino acids 231-429) and carboxyl-terminal (amino acids 1272-1491) domains of A194. One polyclonal antibody A127 710 -1007 was raised to a discrete region in the NH 2 -terminal domain of A127 (amino acids 710 -1007). These antibodies were used for immunocytochemical, Western, and immunoprecipitation experiments. These regions were selected because they shared very little amino acid identity (Ͻ20%) with the equivalent subunits of other mammalian nuclear RNA polymerases II and III.
Immunoblots of nuclear extracts of N1S1 cells with either anti-A127 710 -1007 ( Fig. 2A, lane 2) or anti-A194 231-429 ( Fig. 2A,  lane 4) and anti-A194 1272-1491 (data not shown) antisera demonstrated single bands of approximately 190 and 120 kDa, respectively. The sizes of these immunoreactive proteins are consistent with the molecular mass of 194 and 127 kDa predicted from the deduced amino acid sequences. Further, the size of the immunoreactive bands reduced the possibility that the antibodies recognized the homologous subunits of either RPII or RPIII, which have reported molecular masses of 220 and 160 kDa, respectively, for the ␤Ј subunits and 150 and 128 kDa, respectively, for the ␤ subunits (47)(48)(49)(50)(51)(52).
A second line of evidence that the immunoreactive proteins are the two largest subunits of RNA polymerase I comes from the correspondence of the results of Western blot analysis using the anti-A194 and anti-127 antisera and the biochemical fractionation of nuclear and whole cell extracts. We have previously reported that the fractionation of nuclear extracts by chromatography on DEAE-Sephadex results in a fraction, re-ferred to as DE-175 (43), that contains RNA polymerase I. Western analysis of the fractions generated by DEAE-Sephadex column chromatography with the anti-A194 and A127 antisera confirmed those results and demonstrated that only the DE-175 fraction contained immunoreactive material (Fig. 2B).
Immunopurification of the RNA Polymerase I Enzyme-We examined the possibility that our antibodies to the A194 subunit would be capable of precipitating the "intact" RPI enzyme. Affinity-purified anti-A194 231-429 antibodies were immobilized on Affi-Gel Hz (Bio-Rad) to yield an anti-A194 affinity resin. Incubation of the anti-A194 affinity resin with transcriptionally competent rat nuclear extracts completely inhibited the ability of the extract to transcribe from the rat rDNA promoter (Fig. 3A, lane 2). In contrast, nuclear extracts incubated with preimmune IgG affinity resin still contained active RPI as demonstrated by specific transcription (Fig. 3A, lane 3). These results suggest that the anti-A194 affinity resin was capable of interacting with the A194 subunit in the context of intact, i.e. native, RPI rather than only interacting with free or denatured A194 subunit.
To extend this observation, we attempted to demonstrate that other subunits of RPI coimmunopurified with the A194 subunit. Nuclear extracts of Novikoff hepatoma ascites cells were applied to the anti-A194 affinity resin and to a control column containing preimmune IgG. After washes with loading buffer and 0.15 M NaCl, the columns were eluted with glycine buffer (pH 2.8), and the eluted fractions were analyzed by  (Fig. 3B). Immunoblotting with the A194 231-429 antibody demonstrated that the A194 subunit had been selectively retained by the anti-A194 231-429 affinity column and eluted by the low pH glycine buffer. In contrast, A194 did not bind to the preimmune IgG affinity column. The identity of the band as A194 was confirmed by stripping the A194 231-429 antibody from the membrane and reprobing with the anti-A194 1272-1491 antibody (Fig. 3C). Importantly, reprobing of the same blot with anti-A127 and anti-AC40 antibodies (53) demonstrated that the 127-and 40-kDa subunits of RPI (A127 and A40, respectively) were copurified with the A194 subunit (Fig. 3C). Taken together, these data strongly support the transcription experiments and suggest that under the conditions used in these experiments, the anti-A194 affinity matrix is capable of immunopurifying intact RPI.

SDS-PAGE and Western analysis
UBF Is Not Associated with RNA Polymerase I in Vivo-Two recent studies have reported a protein-protein interaction between RPI and the rDNA transcription factor UBF (24,54). Interestingly, in the studies described here, UBF was not selectively retained on the anti-A194 affinity column, since Western analysis of proteins eluted by high salt or glycine failed to demonstrate UBF in those fractions. UBF was only detected in the column flow-through that was generated in the presence of 150 mM NaCl (Fig. 3C, bottom). Similarly, when we examined immunoprecipitates generated with an anti-UBF affinity resin, we were unable to detect any A194 in the eluate regardless of the conditions used (55).
One possible interpretation of the above experiments is that the anti-A194 and anti-UBF antibodies compete in some way with UBF for binding to A194. For example, UBF may bind to (or within) the domain that the anti-A194 231-429 antiserum recognizes. To test this hypothesis, we examined whether we could copurify UBF with A194 from cell extracts expressing a His-tagged version of A194. A mammalian expression construct containing the entire coding region of A194 tagged with 10 histidine residues at the C-terminal end was prepared (pol-BЈUNTRHIS). The His 10 tag allows purification of the A194 protein using metal affinity chromatography. S100 extracts from either control N1S1 cells or N1S1 cells that were expressing polBЈUNTRHIS were applied separately to preequilibrated Talon affinity resin. After washes with C-20/150, the columns  3. Immunoaffinity purification of RPI using affinity-purified anti-A194 antibodies. A, treatment of a nuclear extract with the anti-A194 antiserum inhibits rDNA transcription. A nuclear extract (lane 1) was treated with either affinity-purified rabbit anti-A194 231-429 antibodies coupled to Affi-Gel Hz resin (lane 2) or with preimmune antibodies coupled to Affi-Gel Hz resin (lane 3) and then analyzed for its ability to specifically transcribe rat rDNA as described under "Materials and Methods." Trans., the specific transcript. Int. Std., an internal standard added to standardize for the recovery of nucleic acids. B, nuclear extracts of N1S1 cells were fractionated by chromatography over columns of affinity-purified rabbit anti-A194 231-429 antibodies coupled to Affi-Gel HZ or preimmune antibodies coupled to Affi-Gel HZ, as indicated, and the protein compositions of the various fractions were analyzed with the antiserum raised to A194 231-429 . C, the proteins that were eluted in each fraction from a column of affinity-purified rabbit anti-A194 231-429 antibodies coupled to Affi-Gel HZ were subjected to SDS-PAGE and Western analysis with the indicated antisera as described under "Materials and Methods." were eluted with 100 mM EDTA, pH 8.0, and the eluted fractions were analyzed by SDS-PAGE and Western analysis (Fig.  4). Immunoblotting with the A194 231-429 antibody demonstrated that the His-tagged A194 subunit was selectively retained by the Talon affinity resin and eluted with 100 mM EDTA (Fig. 4, lanes 3-6). Reprobing of the same blot with anti-A127, anti-AC40, and anti-PAF53 antibodies (24) demonstrated that both the A127 and A40 subunits of RPI and PAF53 co-eluted in the same fractions as did the A194 subunit (Fig. 4). In agreement with the above results (Fig. 3C), we did not observe coelution of the accessory transcription factor UBF with the A194, A127, or A40 subunits of polymerase. Comparison of the sensitivity of the antisera to the RPI subunits with that of the anti-UBF antiserum demonstrated that one molecule of UBF/100 molecules of RPI would have been detectable. In control experiments, untagged endogenous A194 did not bind to the Talon affinity column under the same conditions as used above (Fig. 4, lanes 9 -12). Thus, transfected recombinant His-tagged A194 incorporated into the RPI complex in mammalian cells, and, at least under the conditions used for these experiments, UBF is not associated with this complex.
Isolation of an Active RNA Polymerase I Complex Containing FLAG-tagged A127-Definitive evidence that affinity-purified RPI contains at least the minimal number of the core subunits required for catalytic activity requires demonstration that the eluted RPI complex can support transcription. To address this question, we used anti-FLAG peptide affinity resin, since this allows for elution of FLAG-tagged proteins under gentle conditions. A mammalian expression construct, pBFLAG39, containing the entire coding region of A127 tagged with the FLAG epitope at the NH 2 terminus was prepared, and a clone, N1S1C3, that stably expressed FLAG-A127 was isolated. S100 extracts derived from N1S1 cells or N1S1C3 cells were applied separately to preequilibrated anti-FLAG affinity columns. After washes with C-20/150, the columns were eluted with 0.5 mg/ml FLAG peptide in C-20/150, and the eluted fractions were analyzed for the presence of different polypeptides by SDS-PAGE followed by Western blot analysis or silver staining (Figs. 5, A and B). Immunoblotting with the A127 antibody demonstrated that the FLAG-tagged A127 subunit was selectively retained by the anti-FLAG resin and eluted with 0.5 mg/ml FLAG peptide (Fig. 5A, lanes 3-5). The identity of the band as FLAG-tagged A127 was confirmed by stripping the A127 antibody from the membrane and reprobing with the anti-FLAG antibody (Fig. 5A). Screening of the same blot with anti-A194, anti-AC40 (53), and anti-PAF53 antibodies (24) demonstrated that both the A194 and A40 subunits of RPI and PAF53 co-eluted in the same fractions as did the A127 subunit (Fig. 5A). The accessory transcription factor UBF did not coelute with the A127 subunit under any conditions. In control experiments, untagged endogenous A127 was not eluted from the anti-FLAG resin by the FLAG peptide (Fig. 5A, lanes 8 -10).
Analysis of the polypeptides present in the silver-stained gel indicates the presence of the core subunits of RPI enzyme including A194 and A127, AC40 and A27, and a cluster of polypeptides between 54 and 49 kDa that correspond to the polymerase-associated factors (24), PAF53, PAF51, and PAF49 (Fig. 5B, lane 3). This subunit composition is similar to that recently reported by Hanada et. al. (24), and Song et. al. (53), using conventional biochemical purification techniques. In addition, two polypeptides at 65 and 60 kDa are present that may correspond to the 65-and 60.5-kDa polypeptides previously observed to copurify with purified rodent RPI (reviewed in Ref.  53). The polypeptide migrating at 67 kDa (denoted by an asterisk) is a contaminant protein not associated with RPI, since it also present in FLAG affinity-purified extracts from cells not expressing FLAG-A127 (Fig. 5B, lane 4).
We determined if the RPI eluted from the FLAG-affinity resin could synthesize RNA. In five separate experiments, using three different S-100 extracts of N1S1C3 cells, 16.8% of the RPI activity applied to the immunoaffinity resin was recovered in the eluate obtained with the FLAG peptide (Table I). These results indicate that approximately one in six polymerase molecules contained FLAG-tagged A127 subunits. In contrast, when S100 extracts derived from N1S1 cells were applied to the anti-FLAG affinity resin, 0.8% of the activity was eluted with FLAG peptide. Moreover, our preliminary studies indicate that the FLAG-purified RPI is capable of initiating transcription from the rDNA promoter in the presence of partially purified SL-1 (results not shown). Considered together, these results demonstrate that active FLAG-tagged RPI enzyme can be extensively purified from whole cell extracts (S100) in a single step using anti-FLAG affinity resin.
CK2 Coimmunopurifies with RNA Polymerase I and Phosphorylates the A194 Subunit-Immunoprecipitation of RPI from [ 32 P]orthophosphate-labeled exponentially growing CHO cells using the anti-A194 231-429 antibody demonstrated that the A194 subunit of RPI is a phosphoprotein (Fig. 6A, lane 1). Moreover, at least in CHO cells, the amount of phosphorylated A194 in the cell appears to correlate with the rate of cellular growth, since serum starvation of the cells for 24 h resulted in an 80% decrease in the level of phosphorylated A194 recovered in the immunoprecipitate (Fig. 6A, lane 2).
The above findings led us to examine whether we could successfully use our FLAG affinity purification scheme to determine if a protein kinase activity might copurify with the RPI complex. Fractions containing the FLAG affinity-purified RPI were pooled, and aliquots (5 l) were incubated with [␥ 32 P]ATP at 30°C for 20 min. After incubation, the samples were separated by SDS-PAGE, transferred to nylon membranes, and subjected to autoradiography (Fig. 6B). Two of the polypeptides (190 and 51 kDa) in the affinity-purified RPI were efficiently phosphorylated, indicating the presence of associated kinase(s) (Fig. 6B, lanes 1 and 2). The polypeptide that migrated at 190 kDa was determined to be the A194 subunit of RPI by Western blot analysis (Fig. 6B, Western Blots). The identity of the 51-kDa phosphoprotein is unknown.
Previous studies have provided evidence that CK2, a protein kinase implicated in cell cycle control and the regulation of growth, might be involved in the regulation of rDNA transcription and may be associated with RPI (56 -60). Accordingly, we examined whether the kinase activity that copurified with the FLAG affinity-purified RPI might be CK2. One particular characteristic of CK2 is that it is strongly inhibited by low concentrations of heparin that do not affect most other protein kinases (62,63). The copurified kinase in the present experiments was almost completely inhibited by 0.2 g/ml heparin (Fig. 6B, lane  3), which compares favorably with the I 50 of 0.05 g/ml heparin reported for purified CK2 (64). A second hallmark of CK2 is that it can utilize GTP almost as well as ATP as the phosphate donor (65). As shown in Fig. 6B, lane 4 6. The A194 subunit of RPI is a phosphoprotein and is phosphorylated by CK2. A, A194 was immunoprecipitated from exponentially growing (lane 1) or serum-starved (lane 2) CHO cells that had been continuously labeled with medium containing [ 32 P]orthophosphate for 24 h as described under "Materials and Methods." The cells were harvested into modified radioimmune precipitation buffer, and A194 was immunoprecipitated using anti-A194 antisera bound to protein A-agarose beads. The radioactive proteins that bound to the beads were resolved by SDS-PAGE and visualized by autoradiography. Each lane contains A194 immunoprecipitated from equivalent numbers of cells (1 ϫ 10 6 ). B, FLAG affinity-purified RPI (5 l) was incubated with 1 l of [␥-32 P]ATP (6000 Ci/mmol) in kinase buffer alone (kinase assay, right, lanes 1 and 2) or kinase buffer containing heparin (0. purified RPI (Fig. 6B, lane 5). Importantly, the proteins phosphorylated by the exogenous CK2 were identical to those phosphorylated by the kinase that copurified with RPI (Fig. 6B,  compare lanes 1 and 2 with lane 5).
Consistent with our observation that the majority of the RPI-associated kinase activity was due to CK2, Western blot analysis of the anti-FLAG affinity-purified RPI with antibodies that recognize the ␣/␣Ј subunits of CK2 (Upstate Biotechnology) demonstrated that CK2 copurified in the same fractions as did the FLAG-tagged RPI (Fig. 7A, lanes 1-6). In contrast, CK2 was not observed in the eluates from the control column (Fig.  7A, lanes 7-12). To determine how tight the association between CK2 and RPI might be, four parallel FLAG affinity purifications were performed in which the FLAG affinity columns were washed with buffer containing 200, 400, 600, or 800 mM NaCl before elution with the FLAG peptide. As shown in Fig. 7B (bottom), CK2 remained tightly bound to RPI in the presence of salt as high as 600 mM NaCl. Above 600 mM NaCl, CK2 began to be dissociated from RPI, although a significant amount remained even in the presence of 800 mM NaCl. Salt concentrations above 600 mM also resulted in dissociation of at least a 50% of the core subunits of RPI from the FLAG affinity column (Fig. 7B, top) and loss of 50 -75% of RPI activity.
Quantitative Western analysis was performed to determine the molar ratio of CK2 to the RPI core enzyme. As shown in Fig.  7C and Table II, the molar ratio of CK2 to A194 in the FLAG affinity-purified RPI complex was 0.15:1, indicating that only one in six molecules of RNA polymerase I contain CK2. In the same preparation, PAF53 exhibited a molar ratio of 0.5:1 with respect to RPI.
Taken together, these results strongly suggest that the major kinase activity that copurifies with the RPI under the conditions used in these experiments is CK2. This enzyme appears to be tightly associated with a subset of the RPI in the cell. The major RPI substrates for CK2 in vitro are the 194-kDa subunit of RPI and an associated, but as yet unidentified, 51-kDa polypeptide. DISCUSSION Considerable progress has been made in the purification and/or cloning of some of the components that contribute to the formation of the RPI initiation complex on the mammalian rDNA promoter. For example, these would include UBF and SL-1 (reviewed in Refs. 66 and 67) and more recently enhancer 1 binding factor/Ku and core promoter binding factor/upstream stimulatory factor 1 (68 -72). However, other factors such as TF1C/TIF-1A and the RPI subunits remain largely uncharacterized. Related to this is the identification of a group of proteins, e.g. PAF53, that may be involved in the initiation of specific transcription (24). These proteins are not considered to be core subunits of RPI, although they are closely associated with the enzyme. The in vivo interactions among the various RPI subunits/PAF53 and between RPI and the associated transacting factors such as SL-1, UBF, enhancer 1 binding factor/Ku, and core promoter binding factor are as yet largely uncharacterized. Definition of the nature of these interactions should provide important clues to the mechanism(s) by which cells regulate rDNA transcription and subsequent cellular growth.
As an initial step toward investigating these interactions, we report here the cloning and expression of the two largest subunits of mammalian RPI. Moreover, we demonstrate that stably expressed, epitope-tagged versions of these subunits can be used to rapidly and reproducibly immunopurify RPI complexes. Finally, we have used this purification scheme to demonstrate that the protein kinase CK2 is a submolar component of the RPI complex in vivo and that CK2 efficiently phosphorylates the 194-kDa subunit of this enzyme in vitro.
A194 and A127 Are Homologues of the Bacterial RNA Polymerase ␤Ј and ␤ Subunits-Amino acid sequence data and multiple alignments with the corresponding subunits from other organisms confirmed that the cloned A194 and A127 are the rat FIG. 7. CK2 coaffinity-purifies with RPI. A, S-100 extracts of N1S1C3 cells, which had been stably transformed with a pCR3.1 driving the expression of a FLAG-tagged A127 subunit, and control N1S1 cells were fractionated by chromatography over FLAG affinity resin. The columns were washed and eluted with FLAG peptide (lanes 3-6 and 9 -12) as described under "Materials and Methods." The proteins present in each fraction were subjected to SDS-PAGE and Western analysis with the indicated antisera as described under "Materials and Methods." B, CK2 remains associated with RPI under "high salt" conditions. Nuclear extracts of N1S1C3 cells were fractionated by chromatography over FLAG affinity resin in four separate experiments (lanes 1-4) as described above (Fig. 7A). Prior to elution with the FLAG peptide, the columns were washed with buffer containing either 200 mM NaCl (lane 1), 400 mM NaCl (lane 2), 600 mM (lane 3), or 800 mM (lane 4) NaCl. The fractions containing the FLAG-eluted proteins from each column were pooled and subjected to SDS-PAGE and Western analysis with the indicated antisera as described under "Materials and Methods." C, quantitation of the relative amounts of CK2 associated with affinity purified RPI. Increasing amounts of FLAG affinity-purified RPI (5, 25, and 50 l, lanes 1-3) and standard amounts of A194, PAF53, and CK2 (␣/␣Ј subunits) (1, 5, 10, 25, 50, 100, 250, and 500 ng, lanes 4 -11) were subjected to SDS-PAGE and Western analysis with the indicated antisera as described under "Materials and Methods." homologues of the ␤Ј and ␤ subunits of bacterial RNA polymerases. The majority of the identical amino acid residues were located in approximately eight domains that appear to be strictly conserved among the largest subunits of the RNA polymerase enzymes (1). These domains are thought to represent important functional sites of the enzyme (reviewed in Ref. 1). For example, the putative zinc ion binding domain, a CX 2 CX 6 -12CXGHXGX 24 -37 CX 2 C consensus sequence (reviewed in Ref. 1), which is highly conserved among the ␤Ј subunits of the three eukaryotic RNA polymerases (36), is located between amino acids 64 and 107 of the rat A194 protein (Fig. 1A). Rat A127 contains all of the highly conserved domains common to the ␤ subunits of all three RNA polymerases that are believed to be important for the functions of this subunit (reviewed in Ref. 1). The two lysine residues shown to be critical for nucleotide binding and viability in yeast are conserved in A127 (residues K882 and K890), as they are in all of the RNA polymerase ␤ subunits. In addition, a putative zinc finger, CXXC(X-X) 24 CXXC (domain J) found in all of the RNA polymerase ␤ subunits is also found in A127, residues 1070 -1101 (Fig. 1B). It has been proposed that the zinc finger mediates binding of the ␤ and ␤Ј subunits (1). Considered together, these findings support previous studies suggesting that, while RPI appears to have diverged from RPII and RPIII early in evolution, strong pressures have been exerted to conserve essential domains involved in catalytic processes (1).
Additional confirmation of the identity of the putative A194 and A127 subunits was obtained from immunological data. Polyclonal antibodies raised against fragments of A194 and A127 recognized single polypeptides of approximately 190 and 120 kDa, respectively, on immunoblots of both whole cell and nuclear extracts of N1S1 cells. The sizes of these immunoreactive proteins are in good agreement with the molecular masses (194 and 127 kDa, respectively) predicted by the deduced amino acid sequences. Moreover, when nuclear extracts were fractionated by ion exchange chromatography, the anti-A194 and anti-A127 antibodies only recognized proteins in the fractions that exhibited RPI activity. Immunocytochemistry provided the third line of evidence. Immunofluorescent staining of CHO cells with anti-RPI antisera demonstrated immunoreactivity that was restricted exclusively to the nucleoli, as determined by co-staining with an antibody specific for fibrillarin (45), a nucleolar protein (results not shown).
Immunopurification of RNA Polymerase I Using Affinitypurified Anti-A194 Antisera-Past purification protocols have resulted in preparations of vertebrate RPI consisting of two large polypeptides and as few as three or four smaller subunits (46,74,75). However, more recent studies indicate that purified murine RPI may consist of at least 11 subunits and two or three associated factors (24,53,76). Even two recent publications (53,54) report RPI preparations with different polypeptide compositions. Such differences serve to highlight the problems associated with biochemical purification schemes that require multiple ion exchange and molecular sizing fractionation steps. To circumvent some of the potential difficulties, we have investigated the use of monospecific antibodies directed against the wild type A194 subunit or affinity resins directed toward tagged versions of A194 and A127 subunits of mammalian RPI.
Affinity-purified anti-RPA194 231-429 antibodies inhibited both random transcription from calf thymus DNA as well as specific transcription from the rDNA promoter. These results suggested that the antibody was capable of immunoprecipitating not only free A194 subunits but also those incorporated into functional RPI enzyme. Consistent with the ability of anti-A194 antibodies to precipitate intact polymerase, Western analysis of glycine eluates from an anti-A194 affinity column demonstrated that both the A127 and A40 subunits of RPI eluted in the same fractions as did A194. Similar results were obtained using antibodies raised to a domain consisting of amino acids 1272-1491 of A194. These results demonstrate that we can immunopurify RPI enzyme using antibodies directed toward the A194 subunit of this complex.
Affinity-purified RNA Polymerase I Enzyme Contains Polymerase I-associated Proteins-RPI can exist in at least two forms. One form is able to initiate both nonspecifically on calf thymus DNA and specifically from the rDNA promoter (RPIB) (76,77). The second form of RPI can only catalyze nonspecific transcription (RPIA) (76,77, and references therein). The difference in the properties of these two forms of polymerase has been attributed to either modifications of polymerase I itself or alterations in the factors that associate with the enzyme. For example, when mammalian cell lines are subjected to various conditions that inhibit cell growth or protein synthesis, ribosomal RNA synthesis rapidly ceases (17)(18)(19)(20)77). Subsequent in vitro transcription studies demonstrated that the RPI isolated from those cells was unable to initiate specific transcription, although it could "transcribe" a nonspecific template such as calf thymus DNA or poly(dA-dT) (17, 19 -21). Two laboratories have reported the purification of an RPI-associated factor that can restore the ability of the enzyme to initiate specifically (22,23). However, the subunit compositions of these factors, referred to as TF1C or TIF-1A, differ. Since neither factor has been cloned, it is not clear if they are the same. More recently, three RPI-associated factors termed PAF53, PAF51, and PAF49 have been identified and cloned. The available data suggest that while PAF53 is not a core subunit of RPI, it is a closely associated but dissociable factor involved in formation of the initiation complex. The activity of PAF53 is reminiscent of TF1C/TIF-1A, and it has been suggested that PAF53 may represent one of the subunits of this complex (24).
In addition to PAF53, two recent reports indicate that RPI may also interact with the auxiliary rDNA transcription factor UBF. In one study, the UBF-RPI interaction was reported to be mediated by the 62-kDa subunit of murine RPI (54). However, in a second study, the purified murine RPI did not contain a 62-kDa subunit (53). This second study demonstrated interactions of UBF with the 180-kDa (␤Ј), 114-kDa (␤), and 44-kDa subunits of RP I as well as PAF53 in far Western assays (24). Accordingly, to further clarify some of these discrepancies, we examined whether factors such as PAF53 and UBF would copurify with RPI in our affinity purification experiments.
In our studies, we found that PAF53 co-eluted in the same fractions as did A194, A127, and A40 when RPI was purified from S100 extracts using affinity-purified anti-A194 antibodies. Similar results were obtained when RPI was purified from S100 extracts expressing His-tagged A194 or FLAG-tagged A127 using appropriate affinity matrices. These findings support the observation (24) that PAF53 is closely associated with solubilized RPI in exponentially growing cells. In marked contrast, we found no evidence of an association between solubi- lized RPI and UBF when antibodies directed toward A194 were used to immunopurify the whole enzyme. In a complementary experiment, A194 did not coelute with UBF from an anti-UBF affinity matrix (55). It is not readily apparent why our results differ from those reported previously (24,54). It is possible that the binding of anti-A194 antibodies to RPI is mutually exclusive with the binding of UBF to the polymerase. However, this explanation seems unlikely since 1) UBF was not found to be associated with RPI that had been affinity-purified on a His affinity column from S100 extracts expressing a His-tagged version of A194, and 2) UBF was not co-eluted with RPI that had been purified on an anti-FLAG affinity column from extracts expressing FLAG-tagged A127. It is possible that the interaction between UBF and RPI is labile and was disrupted either during the preparation of the nuclear extracts or during the incubations with the affinity resins. However, we have been able to demonstrate interactions between either UBF and SL-1 or UBF and Rb using similar extracts under similar conditions (55,78). Interestingly, in agreement with previous in vitro studies (24), we have also been able to demonstrate an association of UBF with the two largest subunits of RPI using in vitro overlay assays (results not shown). Thus, while UBF may be capable of directly associating with RPI when assayed in vitro in overlay assays and protein affinity matrix chromatography, it is unlikely that such an association occurs with solubilized RPI in vivo.
Stable Expression of His-tagged A194 and FLAG-tagged A127 in N1S1 Cells-One of the goals of our studies includes the identification of regions of RPI that are required for basic enzymatic activity as well as those domains that interact with the various associated factors that contribute to the formation of the initiated complex. To do this will ultimately require the in vivo assembly of RPI with subunits containing deletion/ substitutions in domains thought to be important to those functions. As a first step in this process, we have created N1S1 cell lines stably expressing His-tagged A194 or FLAG-tagged A127 and have examined whether these recombinant subunits would assemble into functional RPI molecules.
Western analysis of nuclear extracts of cell lines expressing either His-tagged A194 or FLAG-tagged A127 demonstrated significant accumulation of the respective protein products in the cell nuclei. RPI complexes were affinity-purified from these cell lines using either His or anti-FLAG affinity resins. Western analysis of the proteins eluted from the His affinity column demonstrated that both the A127 and A40 subunits of RPI and PAF53 eluted in the same fractions as did the wild type and His-tagged A194 subunits. Similarly, analysis of the proteins eluted from an anti-FLAG affinity column with FLAG peptide demonstrated that both the A194 and A40 subunits of RPI and PAF53 coeluted in the same fractions as did the wild type and FLAG-tagged A127 subunits. Moreover, elution of the FLAG affinity-purified RPI with the FLAG peptide resulted in a preparation that was capable of transcribing nonspecifically on calf thymus DNA. These results clearly indicate that stably expressed, recombinant A194 and A127 subunits will assemble into functional RPI molecules in vivo and demonstrate the potential utility of this approach for investigating mechanisms regulating RPI activity.
Silver staining of the proteins eluted from an anti-FLAG affinity column indicated that the RPI was at least 80% pure. The polypeptide composition was remarkably similar to that observed in recent studies in which RPI was biochemically purified (24,53). For example, in addition to the core subunits of RPI such as A194, A127, AC40, and A27, we also observed a cluster of polypeptides corresponding to the PAFs (PAF53, PAF51, and PAF49). Two major polypeptides of 65 and 60 kDa were also observable in the silver-stained FLAG-purified RPI. These polypeptides were not present in the RPI preparations of Hanada et al. (24) and Song et al. (53), although other groups have reported polypeptides in that range (46,54,57,76). It is possible that these polypeptides represent factors that are "loosely" associated with RPI. If so, then their presence in RPI preparations would probably be highly dependent on the conditions used during the purification process. It would be expected that a rapid affinity purification scheme with minimal manipulations, such as that described herein, is more likely to result in an "intact" RPI complex containing such associated factors than a traditional multistep biochemical purification scheme.
The Protein Kinase, CK2, Coaffinity-purifies with RNA Polymerase I and Phosphorylates the 194-kDa Subunit-In contrast to RPI, the regulation of RPII activity by phosphorylation has been the focus of extensive study. The largest subunit of RPII (B220) contains at its carboxyl-terminal domain 52 repeats of the heptapeptide YSPTSPS and variations thereof (reviewed in Ref. 1). The serines of the carboxyl-terminal domain are the target of extensive phosphorylation by a number of protein kinases that coaffinity-purify with RPII complexes (reviewed in Ref. 79). We have shown here that the largest subunit of RPI (A194) is also a phosphoprotein. However, since A194 lacks the carboxyl-terminal domain motif characteristic of B220, it is unlikely that this subunit would be subject to regulation by the same kinases as B220. Indeed, in these studies we have demonstrated that the major kinase activity that coimmunoprecipitated with RPI could be attributed to the protein kinase CK2. Interestingly, both the A194 and A127 subunits contain multiple CK2 consensus sites (Ser(Thr)-X-X-acidic). However, in these studies only the 194-kDa subunit was phosphorylated by the coaffinity-purifying CK2.
A number of previous groups have shown that CK2 copurifies with RPI during biochemical purification of the enzyme (56,57,61). However, studies by Dahmus (61) suggested that CK2 was merely a contaminant of RPI preparations, since it could be separated from RPI by rechromatography on DEAE-Sephadex. In contrast, Jacob and co-workers (56,57) were unable to remove the CK2 activity from their RPI preparations. Moreover, they demonstrated that the addition of CK2 to RPI containing limited quantities of CK2 resulted in a 5-fold stimulation in RNA synthesis (56). The results described herein clearly demonstrate that CK2 copurifies with immunoaffinity-purified RPI. Moreover, a significant amount of CK2 remains associated with the affinity-purified RPI even under high salt conditions (800 mM NaCl). Thus, the observation that CK2 tightly associates with RPI regardless of the purification protocol (i.e. biochemical versus immunoaffinity) leads us to suspect that CK2 is a true, albeit substoichiometric, component of mammalian RPI complexes rather than a contaminant. Additional studies are required to 1) demonstrate that CK2 can physically interact with RPI when assayed in vitro in overlay assays or protein affinity matrix chromatography and 2) demonstrate that CK2 phosphorylates RPI in vivo.
Interestingly, there appears to be a marked discrepancy between the previous studies of Rose et al. (57) and our experiments regarding which subunits of RPI are phosphorylated by CK2. In our studies, the major polypeptides phosphorylated were the 194-kDa subunit and to a lessor extent an associated 51-kDa polypeptide. In the studies of Rose et al. (57), the major substrates were a 120-kDa polypeptide (assumed to be the second largest subunit of RPI) and a 62-kDa polypeptide. The reason for these differences is not readily apparent. It is possible, for instance, that Rose et al. (57) were in fact observing phosphorylation of the 194-kDa subunit but that proteolysis prior to SDS-PAGE resulted in this subunit being cleaved into a number of smaller polypeptides. Although this explanation is not entirely satisfactory, it is difficult to speculate further on the conflicting results. Nevertheless, these two sets of data confirm that CK2 is associated with RPI and support the hypothesis that this protein kinase may be an important regulator of RPI activity in vivo.
Finally, the observation that the core RPI contains substoichiometric amounts of associated factors such as PAF53 and CK2 suggests that RPI complexes containing different subsets of cofactors must exist in the nucleolus. For example, the RPI complex purified in this study contained A194, PAF53, and CK2 in molar ratios of 1:0.55:0.15. Thus, in this preparation CK2 is associated with only 15% of the RPI, whereas PAF53 is associated with 55% of the RPI. In fact, it is possible that a small percentage of RPI may exist as a preassembled RPI "holoenzyme" containing associated repressors and activators of rDNA transcription in much the same manner as has been demonstrated for RPII (25-27, 80 -83). The affinity purification scheme for RPI described here requires a minimal number of manipulations and therefore should be of considerable use in identifying the components of these RPI complexes and in determining their physiological relevance.