Purification and characterization of an RNA polymerase II phosphatase from yeast.

RNA polymerase (RNAP) II is subject to extensive phosphorylation on the heptapeptide repeats of the C-terminal domain (CTD) of the largest subunit. An activity that is required for the dephosphorylation of yeast RNAP II in vitro has been purified from a yeast whole cell extract by >30,000-fold. The yeast CTD phosphatase activity copurified with two bands with apparent molecular masses of 100 and 103 kDa. The properties of the yeast CTD phosphatase are similar to those of a previously characterized CTD phosphatase from HeLa cells. These properties include stimulation by the general transcription factor IIF (TFIIF), competitive inhibition by RNAP II, magnesium dependence, and resistance to okadaic acid. Both the HeLa and yeast CTD phosphatases are highly specific for their cognate polymerases. Neither phosphatase functions upon the polymerase molecule from the other species, even though the heptapeptide repeats of the CTDs in yeast RNAP II and mammalian RNAP II are essentially identical. The activity of the highly purified CTD phosphatase is stimulated >300-fold by a partially purified fraction of TFIIF. Recombinant TFIIF did not substitute for the TFIIF fraction, indicating that an additional factor present in the TFIIF fraction is required for CTD phosphatase activity. These results show that yeast contains a CTD phosphatase activity similar to that of mammalian cells that is likely composed of at least two components, one of which is 100 and/or 103 kDa.

. The association of the SRB complex and other factors with RNAP II forms what has been termed the holoenzyme. This complex is thought to be recruited to promoters in vivo and to mediate activated transcription (9).
The CTD is subject to extensive phosphorylation in vivo, with a stoichiometry of approximately one phosphate per repeat (1). Phosphorylation predominantly occurs on serine residues, with small amounts on threonine and tyrosine residues. The unphosphorylated form of RNAP II is designated RNAP IIA, whereas the phosphorylated form is designated RNAP IIO. A variety of studies have shown that RNAP IIA and RNAP IIO have distinct functions during transcription and that each round of transcription is associated with the reversible phosphorylation of the CTD. RNAP IIA preferentially assembles into a preinitiation complex with transcription factors on promoter DNA (10), whereas elongation of the transcript is almost exclusively catalyzed by RNAP IIO (11). Phosphorylation of RNAP IIA occurs sometime during the initiation of transcription and is thought to trigger the release of RNAP II from the preinitiation complex (12). RNAP IIO is presumably dephosphorylated at or after termination of transcription to regenerate RNAP IIA for the next round of transcription. CTD phosphorylation appears to be an important mechanism in regulating RNAP II activity in vivo, and changes in the phosphorylation state of RNAP II are associated with major changes in the pattern of transcription. For example, serum stimulation of quiescent cells, heat shock, and viral infection all dramatically alter cellular transcription and also result in a change in the phosphorylation state of RNAP II (13)(14)(15).
Many different protein kinases have been characterized that can phosphorylate the CTD in vitro. Two yeast protein kinases, CTK1 and KIN28, have been shown to be important for phosphorylation of RNAP II in vivo. Disruption of the CTK1 gene results in a large decrease in RNAP II phosphorylation in vivo (16). The null mutant ctk1 cells, although viable, exhibit slow growth and other phenotypes, but it is not known what role CTK1 has in transcription. The KIN28 kinase is found associated with the general transcription factor IIH, which assembles into the preinitiation complex (17). The KIN28 gene is essential, and loss of the KIN28 kinase activity results in a dramatic decrease in RNAP II phosphorylation and transcriptional activity (18).
Much less is known about CTD phosphatases. A CTD phosphatase activity has been purified from HeLa cells that is highly specific for RNAP II and dephosphorylates the CTD processively (19). Two proteins with apparent molecular masses of 205 and 150 kDa copurify with the HeLa CTD phosphatase activity, although it is not clear which contains the catalytic activity. In contrast to CTD kinases, HeLa CTD phosphatase does not use recombinant CTD as a substrate (20). Furthermore, CTD phosphatase is competitively inhibited by RNAP IIB, a form of RNAP II that lacks the CTD. These and other results suggested that a docking site exists on RNAP II that CTD phosphatase must first bind before it can gain access to the CTD. HeLa CTD phosphatase is also stimulated 5-fold by TFIIF, and the stimulation can be inhibited by TFIIB. The minimal region of TFIIF sufficient to stimulate CTD phosphatase is the C-terminal 160 amino acids of the RAP74 subunit (20).
Serine/threonine protein phosphatases have been classified into four families, PP1, PP2A, PP2B, and PP2C, based on biochemical properties and amino acid sequence (21). PP1, PP2A, and PP2B are all inhibited by okadaic acid, although with differing sensitivities. PP2B requires calcium ions for activity, and PP2C requires magnesium ions. HeLa CTD phosphatase has been classified as a type 2C phosphatase based on its requirement for magnesium ions and its resistance to okadaic acid. This report describes the purification and characterization of a type 2C phosphatase from yeast specific for the CTD of RNAP II. Polyacrylamide Gel Electrophoresis-SDS-PAGE was carried out according to the method of Laemmli (22). Native polyacrylamide gel electrophoresis was carried out on 6% polyacrylamide gels with 50 mM Tris acetate, pH 7.8, 20% glycerol, 0.1 mM EDTA, and 10 mM potassium acetate. The running buffer was the same as the gel buffer except no glycerol was added. Native gels were prerun for 2 h at 8 V/cm and then run for 4.5 h at 20 V/cm at 4°C. To recover proteins from the native polyacrylamide gel, the gel was sliced into 5-mm portions and mashed, and the gel fragments were soaked overnight in 4 volumes of buffer F with 20 mM potassium acetate on a rotating platform at 4°C. Polyacrylamide gels were stained with either Coomassie Blue or silver (23). Polyacrylamide gels were also prestained with Coomassie Blue before silver staining to increase sensitivity (24).

Materials-Radiolabeled
Preparation of [ 32 P]RNA Polymerase II-[ 32 P]RNAP II was prepared by incubating either yeast or calf thymus RNAP IIA with a partially purified TFIIH-associated CTD kinase and 2.5 M [␥-32 P]ATP (750 Ci) in buffer E for 1 h at 30°C. The 32 P-labeled RNAP II was purified on DE52 as described previously (10). Calf thymus RNAP II and yeast RNAP II radiolabeled with 32 P to similar specific activities. The radiolabeled band corresponding to the largest subunit of RNAP II was identified by the following criteria. (i) It comigrates on SDS-PAGE with the silver-stained largest subunit of RNAP II; (ii) the presence of the band is dependent on the addition of RNAP II to the kinase reaction; and (iii) the band undergoes a mobility shift on SDS-PAGE when extensively phosphorylated, characteristic of the largest subunit of RNAP II. Each mole of RNAP II contained ϳ0.4 mol of 32 P. Yeast RNAP IIA lacking subunits 4 and 7 (25) was used in preparing [ 32 P]RNAP II. No differences were observed in using this enzyme as a substrate in CTD phosphatase assays compared with wild-type RNAP II or RNAP II lacking subunit 9 (26), either in the presence or absence of the SP 0.48 fraction. 2 Purification of Proteins-TFIIH-associated CTD kinase was partially purified from a yeast whole cell extract using Bio-Rex 70, phosphocellulose, and Ni 2ϩ -nitrilotriacetic acid-agarose chromatography as described previously (27). CTD kinase activity was assayed according to Payne and Dahmus (28). The stimulatory fraction, containing TFIIF and used in CTD phosphatase assays, was prepared from material that bound to SP-Sepharose during the CTD phosphatase purification (see below). The SP-Sepharose column was developed with a 300-ml gradient of 0.05-1.2 M potassium acetate in buffer C, with the stimulatory activity eluting at 0.48 M potassium acetate (SP 0.48 fraction). HeLa CTD phosphatase was purified as described previously (19).
Protein Quantitation-Protein concentrations were determined using the protein dye assay (Bio-Rad) with bovine serum albumin as the standard.
Glycerol Gradient Sedimentation-Glycerol gradients were prepared in buffer C with 40 mM potassium acetate and glycerol concentrations from 15 to 35%. Thyroglobulin (669 kDa), catalase (240 kDa), and bovine serum albumin (66 kDa) were used as molecular mass markers. The gradients were centrifuged at 150,000 ϫ g in an SW 41 rotor for 16 h at 4°C.
Renaturation of Proteins-Proteins purified by SDS-PAGE were renatured according to the method of Conaway et al. (29).
CTD Phosphatase Assay-Samples of CTD phosphatase were dialyzed against buffer C containing 0.025% Tween 20 and 10 mM potassium acetate. The samples were incubated with [ 32 P]RNAP II and 6 g of the SP 0.48 fraction containing TFIIF in buffer F in a final volume of 20 l. The reactions were incubated at 30°C for 30 min unless otherwise indicated. Reactions were stopped by the addition of 6 ϫ Laemmli buffer (22), heated at 100°C for 5 min, and electrophoresed on an SDS-5% polyacrylamide gel. The gel was dried, and the band corresponding to the largest subunit of RNAP II was quantitated using a PhosphorImager (Molecular Dynamics, Inc.). The assay was linear until 40% of the label was removed. The amount of [ 32 P]RNAP II in the assay was ϳ2 fmol (0.1 nM). One unit of CTD phosphatase activity releases 1 pmol of 32 P/min. CTD phosphatase activity was also detected using an assay that specifically measures released 32 P (30).
Purification of CTD Phosphatase-The yeast strain YPH/TFB1.6HIS was grown at 30°C to mid-log phase in 200 liters of 1% Bacto-yeast extract, 2% Bacto-peptone, and 2% glucose; harvested by centrifugation (1-kg yield); washed in cold distilled water; and resuspended in 0.5 volume of 3 ϫ lysis buffer. All steps were carried out at 4°C. A whole cell extract was prepared as described previously (31), with the following modifications. Polyethyleneimine was omitted since this resulted in loss of CTD phosphatase activity, and cells were disrupted in a bead beater (Biospec Products, Inc.) with 8 ϫ 1-min bursts interrupted with 2 min of cooling on ice. The extract was dialyzed against buffer A containing 0.05 M potassium acetate and stored as four equal aliquots at Ϫ80°C. Each aliquot was diluted with buffer B to a protein concentration of 5 mg/ml, adjusted to 0.15 M potassium acetate, and loaded onto a DE52 column (25 ϫ 5 cm) equilibrated in buffer B with 0.15 M potassium acetate. The DE52 column was washed with 1 column volume of buffer B with 0.15 M potassium acetate and step-eluted with 2 column volumes of buffer B containing 0.25 M potassium acetate followed by 2 column volumes of buffer B with 0.45 M potassium acetate. CTD phosphatase activity eluted in the 0.45 M potassium acetate fraction and was dialyzed against buffer C containing 0.025 M potassium acetate, adjusted to 0.05 M potassium acetate, and loaded onto SP-Sepharose (15 ϫ 2.5 cm) and Q-Sepharose (12 ϫ 2.5 cm) columns connected in tandem and equilibrated in buffer C containing 0.05 M potassium acetate. After the sample was loaded, the columns were washed with 2 column volumes of buffer C with 0.05 M potassium acetate. The SP-Sepharose column was disconnected, and the Q-Sepharose column was developed with a 300-ml gradient of 0.05-1.2 M potassium acetate in buffer C. CTD phosphatase activity eluted at 0.86 M potassium acetate. The Q-Sepharose fractions containing CTD phosphatase activity were pooled; dialyzed against buffer C containing 10 mM potassium phosphate, pH 8.0, and 50 mM potassium acetate; and loaded onto blue Sepharose CL-6B (8 ϫ 1.5 cm) and Bio-Gel HTP (8 ϫ

Identification of a Yeast CTD Phosphatase Activity-Yeast
CTD phosphatase activity was measured by monitoring the removal of 32 P radiolabel from the largest subunit of RNAP II. To prepare the phosphatase substrate, yeast RNAP IIA was phosphorylated using a partially purified yeast TFIIH-associated CTD kinase in the presence of [␥-32 P]ATP. In addition to the largest subunit of RNAP II, a number of other proteins in the CTD kinase preparation were radiolabeled and served as controls for the specificity of phosphatase activity (Fig. 1). Previous studies showed that HeLa CTD phosphatase is okadaic acid-resistant and dependent on magnesium ions for activity, thereby classifying it as a type 2C phosphatase (19). Phosphatase assays were therefore performed in the presence of 0.5 M okadaic acid, sufficient to inhibit type 1 and 2A protein phosphatase activities, and with magnesium ions to activate type 2C phosphatases (32). 2-Glycerophosphate, a substrate of nonspecific acid and alkaline phosphatases, was included to competitively inhibit these activities. In addition, samples of yeast CTD phosphatase were dialyzed against buffer with 0.1 mM EDTA, thereby inhibiting the calcium-dependent type 2B phosphatases. A yeast whole cell extract was fractionated on a DE52 column, and an aliquot of the DE 0.45 fraction was incubated with the 32 P-labeled RNAP II for various times and analyzed by SDS-PAGE and autoradiography (Fig. 1A). The DE 0.45 fraction contained an activity that specifically removed 32 P radiolabel from the largest subunit of RNAP II in a time-dependent manner. In contrast, a Bio-Rex 70 chromatography fraction derived from the same yeast whole cell extract contained phosphatase activities that removed radiolabel from all the 32 P-labeled proteins except RNAP II, indicating that the largest subunit of RNAP II was relatively resistant to dephosphorylation by other type 2C phosphatases (Fig. 1B). The apparent phosphatase activity in the DE 0.45 fraction was not due to proteolysis of the CTD of RNAP II since no smaller peptides were generated even following a 2-h incubation with 50 times more enzyme. Furthermore, the release of 32 P and not 32 P-peptides was detected using a 32 P release assay (data not shown) (30).
Purification of Yeast CTD Phosphatase-A summary of the purification of a CTD phosphatase activity is presented in Table I. The DEAE column step separated the CTD phosphatase activity from almost all of the nonspecific phosphatase activity, nucleic acids, and Ͼ90% of the protein. The SP-Sepharose column step removed the general transcription factors and a large proportion of the RNAP II. After this step, the CTD phosphatase activity was highly dependent on the addition of the SP 0.48 fraction that contained TFIIF. The CTD phosphatase activity recovered from the Q-Sepharose column was stimulated 10-fold by the SP 0.48 fraction. The level of stimulation increased after several subsequent column steps, resulting in fractions from the Mono Q column step being stimulated Ͼ300fold by the SP 0.48 fraction. No phosphatase activity was detected in the SP 0.48 fraction. The Bio-Gel HTP column fractionated the remaining RNAP II from the CTD phosphatase activity. The CTD phosphatase enzyme has a high affinity for the anion exchange resins DEAE and Q-Sepharose, suggesting that it is a highly acidic protein. Two bands with apparent molecular masses of 100 and 103 kDa comigrated with CTD phosphatase activity on the Bio-Gel HTP, phenyl-Superose, Superose 12, and Mono Q columns in CTD phosphatase assays and on SDS-PAGE (data not shown) (Fig. 2). The 100-and 103-kDa proteins stained very poorly with silver; thus, polyacrylamide gels were prestained with Coomassie Blue to increase the sensitivity of detection with silver. CTD phosphatase activity eluted with an apparent molecular mass of 125 FIG. 1. Identification of a yeast CTD phosphatase activity. A, a sample of DE52-purified CTD phosphatase was incubated at 30°C with [ 32 P]RNAP II in a 120-l reaction in buffer F. Aliquots (20 l) were removed at 0, 5, 10, 20, and 40 min (lanes 1-5) and analyzed as described under "Experimental Procedures." B, a whole cell extract was fractionated on Bio-Rex 70 as described under "Experimental Procedures" for the CTD kinase purification. The fraction that eluted with 0.65 M potassium acetate (BR 0.65 fraction) was assayed for phosphatase activity with [ 32 P]RNAP II as described under "Experimental Procedures," except 5 M okadaic acid was added and the reaction was incubated for 2 h. Lane 1, no sample; lane 2, BR 0.65 fraction. kDa on Superose 12, but sedimented with an apparent molecular mass of only 61 kDa in glycerol gradients, suggesting that it has a monomeric and likely an elongated structure (data not shown). The glycerol gradient fractions were also examined by SDS-PAGE and silver staining. Activity again comigrated with the 100-and 103-kDa bands. The Mono Q-purified CTD phosphatase analyzed by SDS-PAGE and Coomassie Blue staining showed three other proteins with apparent molecular masses of Ͼ300, 178, and 66 kDa in addition to the 100-and 103-kDa bands (Fig. 2B). However, the Ͼ300-, 178-, and 66-kDa bands did not comigrate with CTD phosphatase activity on the Superose 12 or phenyl-Superose column.
To further investigate which protein contained the CTD phosphatase activity, an aliquot of the peak fraction from the Mono Q column was fractionated on a 6% native polyacrylamide gel. The Ͼ300-, 178-, 100/103-, and 66-kDa proteins were separated into four bands on the native polyacrylamide gel, with the 100-and 103-kDa bands comigrating together.
The gel was sliced, and protein from the gel slices was eluted and analyzed by SDS-PAGE as well as assayed for CTD phosphatase activity. CTD phosphatase activity only comigrated with the 100-and 103-kDa bands (Fig. 3). It remained possible that other poorly staining proteins might be present that could contain the CTD phosphatase activity. To test this, a sample of the Mono Q-purified CTD phosphatase was separated by SDS-PAGE. The 178-, 100/103-, and 66-kDa bands were excised, and the protein was extracted, renatured, and analyzed in a CTD phosphatase assay and again by SDS-PAGE. CTD phosphatase activity was detected only with the renatured 100/103-kDa proteins (Fig. 4). The Mono Q-purified CTD phosphatase preparation contained ϳ20 g of the 100/103-kDa proteins from 1 kg of starting material.
Characterization of the CTD Phosphatase Activity-CTD phosphatase was inactive in the absence of divalent cations and had 50% maximal activity with 1 mM MgCl 2 (Fig. 5). CTD phosphatase activity was also supported with CaCl 2 or MnCl 2 , but not ZnCl 2 (data not shown) (Fig. 5). This effect was surprising since other PP2C enzymes do not use calcium ions. In addition, HeLa CTD phosphatase does not function in the presence of calcium ions. The effect of calcium ions on CTD phosphatase activity in the presence of suboptimal amounts of magnesium ions was additive. CTD phosphatase activity was fully active in reactions containing 50 mM potassium acetate; however, Ͼ90% of the activity was inhibited in reactions containing 250 mM potassium acetate (data not shown). In contrast, other phosphatase activities present in the whole cell extract were only slightly inhibited with 400 mM potassium acetate. In keeping with the PP2C categorization of the CTD phosphatase, the protein phosphatase inhibitors okadaic acid and vanadate had no effect on activity at concentrations up to 10 M and 1 mM, respectively (data not shown).
Yeast CTD phosphatase appears to be specific for dephosphorylating RNAP II (Fig. 1). To further investigate the specificity, yeast and HeLa CTD phosphatases were tested with yeast [ 32 P]RNAP II and mammalian [ 32 P]RNAP II. Yeast CTD phosphatase dephosphorylated yeast RNAP II, but not calf thymus RNAP II (Fig. 6A, lanes 1-4); conversely, HeLa CTD phosphatase dephosphorylated calf thymus RNAP II, but not yeast RNAP II (lanes 5-8). Furthermore, yeast TFIIF and not human TFIIF stimulated yeast CTD phosphatase (Fig. 6B). In addition, yeast CTD phosphatase did not dephosphorylate calf thymus RNAP II in the presence of human TFIIF (data not shown).
Previous studies showed that HeLa CTD phosphatase is competitively inhibited by RNAP II (20). The addition of a 200-fold excess of unlabeled yeast RNAP II over yeast [ 32 P]RNAP II to a CTD phosphatase assay resulted in inhibition of CTD phosphatase activity (Fig. 6C). A 100-fold excess of unlabeled yeast RNAP II over [ 32 P]RNAP II resulted in a 50% inhibition of CTD phosphatase activity (data not shown). In  contrast, the addition of up to a 400-fold excess of calf thymus RNAP II over yeast [ 32 P]RNAP II did not inhibit yeast CTD phosphatase activity (data not shown) (Fig. 6C). The addition of an equal number of moles of yeast TFIIF did not overcome the inhibition by yeast RNAP II (data not shown).
As mentioned above, after the SP-Sepharose column step, detection of CTD phosphatase activity was highly dependent on the addition of the SP 0.48 fraction that contained TFIIF. The activity of the CTD phosphatase recovered from the Mono Q column was stimulated Ͼ300-fold by the addition of the SP 0.48 fraction. Substituting recombinant yeast TFIIF (Tfg1 and Tfg2) or recombinant Tfg1 (33) for the SP 0.48 fraction gave no stimulation of activity (data not shown). Recombinant TFIIF did stimulate CTD phosphatase activity 3-fold when added to the whole cell extract or the DE 0.45 fraction and 4.5-fold when added to the Q-Sepharose fraction (data not shown) (Fig. 6B). DISCUSSION Yeast contains a protein phosphatase activity that is highly specific for dephosphorylating the largest subunit of yeast RNAP II. Many of the properties of the yeast CTD phosphatase are similar to those of the HeLa enzyme (19). These similarities include specificity for RNAP II as a substrate, stimulation of activity by the general transcription factor IIF, inhibition by excess RNAP II, resistance to the phosphatase inhibitor okadaic acid, requirement for magnesium ions for activity, and inhibition by high concentrations of salt (19). The magnesium requirement and resistance to okadaic acid would characterize the yeast CTD phosphatase, like the HeLa enzyme, as a type 2C phosphatase. However, the yeast CTD phosphatase activity is also supported by calcium ions, a property not seen before in the type 2C class of enzymes (21). The HeLa CTD phosphatase cannot use calcium ions to support activity, but can use zinc, which did not support activity with the yeast enzyme. Other magnesium-dependent phosphatases are unaffected by calcium, inhibited (34), or stimulated (35). Millimolar amounts of calcium are required for significant activity of the yeast CTD phosphatase, and calcium does not support a level of activity higher than with magnesium alone. Thus, the enzyme is unlikely to be regulated by calcium in vivo. However, this unusual property could be used to distinguish it from other phosphata- Samples of the Mono Q-purified yeast CTD phosphatase were assayed as described under "Experimental Procedures" in the absence and presence of either calcium or magnesium ions. The CTD phosphatase activity was quantitated using a PhosphorImager. The values were calculated as a percent of maximum activity and are plotted as a function of divalent ion concentration. ses in whole cell extracts.
There appears to be a strict species specificity between each phosphatase and its cognate polymerase. The yeast and HeLa CTD phosphatases failed to dephosphorylate RNAP II from the other species. This species specificity is surprising since the heptapeptide repeats of the CTDs of yeast RNAP II and mammalian RNAP II are essentially identical. Previous studies with the HeLa CTD phosphatase indicated that there is a docking site on RNAP II distinct from the CTD where CTD phosphatase must interact before dephosphorylating the CTD (20). The species specificity is likely due to divergent protein structures at the docking site between the yeast and mammalian RNAP II molecules. Work directed toward identifying the subunit(s) of RNAP II important for interacting with CTD phosphatase is underway. Yeast RNAP II lacking either subunits 4 and 7 (25) or subunit 9 (26) is dephosphorylated with equal efficiency compared with wild-type RNAP II. 2 There is an additional level of species specificity in CTD phosphatase function and that concerns the stimulation by TFIIF. The yeast and HeLa CTD phosphatases are stimulated only by their cognate TFIIF molecules, i.e. human TFIIF cannot stimulate yeast CTD phosphatase activity with either yeast or mammalian RNAP II. The HeLa CTD phosphatase has been shown to bind to human TFIIF and RAP74 deletion constructs immobilized on Ni 2ϩ -nitrilotriacetic acid-agarose beads, and the binding interaction requires the C-terminal 160 amino acids of RAP74. 3 Thus, at least two protein-protein interactions are likely to be important for efficient phosphatase activity, phosphatase with RNAP II and phosphatase with TFIIF. Human TFIIF functions with Drosophila RNAP II in an in vitro transcription elongation system (36), and TFIIF from Schizosaccharomyces pombe and that from Saccharomyces cerevisiae are functionally interchangeable in a heterologous in vitro transcription system (37). However, Saccharomyces TFIIF does not function with mammalian RNAP II and transcription fac-tors in an in vitro transcription assay. 4 The highly purified yeast CTD phosphatase activity is stimulated Ͼ300-fold by an SP-Sepharose fraction from the CTD phosphatase purification that contains TFIIF. Highly purified HeLa CTD phosphatase is stimulated only 5-fold by recombinant human TFIIF (20). Surprisingly, recombinant yeast TFIIF did not stimulate highly purified yeast CTD phosphatase, but did stimulate partially purified fractions of CTD phosphatase. These results suggest that an additional essential factor for CTD phosphatase activity is fractionated from the 100/103-kDa component at the SP-Sepharose column step in the purification. If mammalian CTD phosphatase requires a similar factor, then it must remain stably associated since the HeLa enzyme showed no dramatic decline in activity after many purification steps and was not dependent on other fractions for activity (19). This difference in purified complexes might be analogous to the differential stability seen with TATA-binding protein-TATA-binding protein-associated factor complexes, which are stable in mammalian extracts, but dissociate during purification in yeast (38).
SDS-PAGE analysis of the highly purified yeast CTD phosphatase shows two bands with apparent molecular masses of 100 and 103 kDa. Glycerol gradient sedimentation and gel filtration analysis indicate a monomeric structure with an elongated shape. It is not clear if the two proteins represent two different forms of CTD phosphatase or if one of them is an unrelated protein. Microsequences from three peptides derived from the 100/103-kDa proteins match a single yeast gene of previously unknown function, predicting a highly acidic protein with a molecular mass of 83 kDa. 5 The yeast genome sequence data base contains five genes with homology to type 2C phosphatases with predicted molecular masses ranging from 32 to 64 kDa. None of these has detectable similarity to the 83-kDa protein. Since serine/threonine protein phosphatases are a highly conserved family of enzymes, the lack of detectable homology of the gene encoding the 83-kDa protein to other phosphatases presents two possibilities. The 83-kDa protein may represent a new specific class of protein phosphatases. Alternatively, this protein may have a regulatory role in CTD phosphatase function. The isolation and characterization of CTD phosphatase from yeast now allow a combined biochemical and genetic analysis to distinguish these possibilities. In addition, the function of this enzyme during transcription can now be examined in detail.