Characterization of the CTD phosphatase Fcp1 from fission yeast. Preferential dephosphorylation of serine 2 versus serine 5.

The C-terminal domain (CTD) of RNA polymerase II undergoes extensive phosphorylation and dephosphorylation at positions Ser2 and Ser5 during the transcription cycle. A single CTD phosphatase, Fcp1, has been identified in yeast and metazoans. Here we conducted a biochemical characterization of Fcp1 from the fission yeast Schizosaccharomyces pombe. The 723-amino acid Fcp1 protein was expressed at high levels in bacteria. Recombinant Fcp1 catalyzed the metal-dependent hydrolysis of para-nitrophenyl phosphate with a pH optimum of 5.5 (kcat = 2 s(-1); K(m) = 19 mm). Deletion analysis showed that 139- and 143-amino acid segments could be deleted from the N and C termini of Fcp1, respectively, without affecting phosphatase activity. A segment containing amino acids 487-580, deletion of which abolished activity, embraces a BRCT domain present in all known Fcp1 orthologs. Mutations of residues Asp170 and Asp172 abrogated Fcp1 phosphatase activity; the essential aspartates are located within a 170DXDXT172 motif that defines a superfamily of metal-dependent phosphotransferases. We exploited defined synthetic CTD phosphopeptide substrates to show for the first time that: (i) Fcp1 CTD phosphatase activity is not confined to native polymerase II and (ii) Fcp1 displays an inherent preference for a particular CTD phosphorylation array. Using equivalent concentrations (25 microm) of CTD peptides of identical amino acid sequence and phosphoserine content, which differed only in the positions of phosphoserine within the heptad, we found that Fcp1 was 10-fold more active in dephosphorylating Ser2-PO4 than Ser5-PO4.

The C-terminal domain (CTD) 1 of the largest subunit of RNA polymerase II (pol II) serves as a landing pad for macromolecular assemblies that regulate mRNA synthesis and processing (1). The CTD is composed of a tandemly repeated heptad motif (consensus sequence YSPTSPS). The mammalian pol II CTD has 52 heptad repeats, the fission yeast Schizosaccharomyces pombe CTD has 29 repeats, and the budding yeast Saccharomyces cerevisiae CTD has 26 -27 copies. The CTD undergoes a cycle of extensive phosphorylation and dephosphorylation at positions Ser 5 and Ser 2 , which is coordinated with the transcription cycle (2). The CTD phosphorylation state is also re-sponsive to developmental cues (3). Multiple CTD kinases are present in eukaryotic cells, e.g. budding yeast has four CTD kinases (4). In contrast, only one CTD phosphatase, Fcp1, has been identified thus far (5)(6)(7)(8)(9). Fcp1 is essential for cell viability in budding and fission yeast (7,10).
The potential complexity of the CTD serine phosphorylation array comprises 2 2n different structures, where n is the number of heptad repeats. Thus, there is ample opportunity to fine tune the putative effector functions of the CTD by remodeling the phosphorylation array, a process that would entail changing the balance between the activities of the CTD kinases and phosphatases at either some or all of the serine phosphorylation sites. The functional effects of varying the phosphorylation array are seen clearly in the case of the mammalian mRNA capping enzyme, Mce1. Binding of Mce1 to defined CTD phosphopeptides containing phosphoserine at position 5 stimulates the catalytic activity of the guanylyltransferase domain, and the extent of stimulation increases with the number of Ser 5 -PO 4 heptad repeats (11). Mce1 also binds to the CTD phosphorylated on Ser 2 , but this interaction has no effect on guanylyltransferase activity (11). Thus, the total number of CTD phosphates and their position in the heptad determine allosteric effects on an mRNA processing enzyme. In vivo studies confirm that recruitment of the capping apparatus to the pol II elongation complex requires the action of the TFIIH-associated CTD kinase Kin28, which specifically phosphorylates Ser 5 (12,13). Although it has been suggested that dephosphorylation of Ser 5 during transcription elongation facilitates recycling of the capping apparatus, there are differing views as to whether Fcp1 is the enzyme responsible for that dephosphorylation step (12)(13)(14).
Fcp1 was purified initially from human cells and from the budding yeast S. cerevisiae (5,6). Genes and cDNAs encoding Fcp1 were subsequently identified in fungi, humans, and Xenopus (2,(7)(8)(9)(10). Early biochemical studies established that Fcp1 phosphatase activity depends on a divalent cation cofactor (5,6). Fcp1 action on pol II is typically measured by conversion of the phosphorylated IIo isoform of the large subunit to the IIa form that migrates more rapidly during SDS-PAGE (5). The large subunit is detected by Western blotting or by prior labeling of the pol II substrate with 32 P. Studies of the yeast phosphatase employed the TFIIH-associated kinase to 32 P-label the CTD, presumably exclusively at Ser 5 (6,15). The fact that the yeast Fcp1 preparation effected the release of 32 P from this pol II substrate would suggest that Fcp1 is capable of dephosphorylating Ser 5 .
Whereas human Fcp1 was able to dephosphorylate either free pol IIo or pol IIo in the context of the transcription elongation complex (9,16,17), the enzyme was reportedly unable to dephosphorylate recombinant CTD fusion proteins or the purified IIo subunit (9,18). The inference from these data was that binding to a non-CTD site on pol II was required for CTD phosphatase function. Several groups have demonstrated the presence of Fcp1 in a complex with pol II (7,9); Kimura et al. (10) recently identified a direct contact between S. pombe Fcp1 and the Rpb4 subunit of pol II. Fcp1 also binds to the RAP74 subunit of TFIIF, which results in a stimulation of CTD phosphatase activity (6,18). Recombinant S. cerevisiae Fcp1 per se was capable of dephosphorylating pol II that was 32 P-labeled in vitro with the TFIIH kinase; however, the reaction required stoichiometric concentrations of Fcp1 (15). Recombinant S. cerevisiae Fcp1 also catalyzed dephosphorylation of p-nitrophenyl phosphate (pNØP), albeit at high substrate concentrations and with a low turnover number (15).
Here we conducted a biochemical characterization of the Fcp1 ortholog of the fission yeast S. pombe. The S. pombe Fcp1 cDNA encodes a predicted 723-amino acid polypeptide related to the Fcp1 proteins from S. cerevisiae, humans, and other metazoan species; the extent of sequence conservation is highest in a putative catalytic module spanning residues 140 -326 of S. pombe Fcp1 (Fig. 1). Our interest in the fission yeast enzyme stems from our prior work on the interactions of the S. pombe capping enzymes with the phosphorylated CTD (19). We anticipated that knowledge of the enzymology of CTD dephosphorylation in S. pombe would illuminate the potential influence of Fcp1 on the association of the capping enzymes with the pol II elongation complex.
We show that S. pombe Fcp1 can be expressed at high levels in bacteria as a soluble catalytically active phosphatase and purified conveniently by affinity chromatography. We characterize its activity using both pNØP and synthetic CTD phosphopeptides as substrates. Fcp1 displays a distinctive acidic pH optimum at 5.5, and there is considerably less activity at pH 7.5-8.0, which are the conditions used in all previously reported studies of Fcp1 activity. We find that whereas purified Fcp1 readily dephosphorylates CTD phosphopeptides consisting of four heptad repeats containing only Ser 2 -PO 4 , it is an order of magnitude less active on a CTD substrate containing only Ser 5 -PO 4 . Thus, Fcp1 activity is not confined to native pol II, and it displays an intrinsic preference for action at certain phosphorylation arrays. Preferential action of Fcp1 at Ser 2 -PO 4 is consistent with the recent finding that conditional mutations in yeast Fcp1 lead to increased levels of CTD Ser 2 phosphorylation in vivo (14).

EXPERIMENTAL PROCEDURES
Recombinant S. pombe Fcp1-S. pombe Fcp1 was produced in Escherichia coli as a His 10 -tagged fusion as follows. The fcp1 ϩ coding sequence (GenBank TM accession AL390814) was PCR-amplified from a S. pombe cDNA library using primers designed to introduce an NcoI site at the translation start codon, a glycine codon following the start codon, and a BamHI site 3Ј of the stop codon. The PCR product was restricted with NcoI and BamHI and then inserted between the NcoI and BamHI sites of the yeast vector pYN132 (CEN TRP1). After sequencing the fcp1 ϩ insert to confirm that no other coding changes were introduced compared the sequence in GenBank TM , the insert was excised with NcoI and BamHI and inserted between the NcoI and BamHI sites of a customized bacterial expression vector pET16m (a derivative of pET16b) so as to fuse the Fcp1 protein in-frame with an N-terminal 14-amino acid leader peptide (MGHHHHHHHHHHSA). The pET-Fcp1 plasmid was transformed into E. coli BL21(DE3)-RIL (Stratagene). A 500-ml culture was grown at 37°C in LB medium containing 0.1 mg/ml ampicillin and 50 g/ml chloramphenicol until the A 600 reached 0.6. The culture was adjusted to 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside and 2% ethanol, and incubation was continued for 20 h at 17°C. The cells were harvested by centrifugation and stored at Ϫ80°C. All of the subsequent procedures were performed at 4°C. Thawed bacteria were resuspended in 25 ml of buffer A (50 mM Tris-HCl, pH 7.5, 200 mM NaCl, 10% glycerol). Phenylmethylsulfonyl fluoride and lysozyme were added to final concentrations of 500 M and 100 g/ml, respectively. After incubation on ice for 30 min, Triton X-100 was added to a final concentration of 0.1%, and the lysate was sonicated to reduce viscosity. Insoluble material was removed by centrifugation for 45 min at 18,000 rpm in a Sorvall SS34 rotor. The soluble extract was mixed for 30 min with 4 ml of Ni ϩ2 -nitrilotriacetic acid-agarose (Qiagen) that had been equilibrated with buffer A containing 0.1% Triton X-100. The slurry was poured into a column, and the resin was washed serially with 20-ml aliquots of buffer A containing 0.1% Triton X-100 plus 5 or 10 mM imidazole. Fcp1 was then step-eluted with 250 mM imidazole in buffer A. The enzyme preparation (15-20 mg of protein) was dialyzed against buffer containing 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1 mM dithiothreitol, 10% glycerol, 0.01% Triton X-100 and stored at Ϫ80°C. The protein concentration was determined using the Bio-Rad dye binding method with bovine serum albumin as the standard.
Mutagenesis of Fcp1-Amino acid substitution mutations D170A, D170N, D170E, D172A, D172N, and D172E (and diagnostic restriction sites) were introduced into the fcp1 ϩ cDNA by the two-stage overlap extension method (20). pET-Fcp1 was used as the template for the first stage amplification. The mutated full-length cDNAs generated in the second stage amplification were digested with NcoI and BamHI and then inserted into pET16m. The inserts of the resulting pET-Fcp1* plasmids were sequenced completely to confirm the desired mutations and exclude the acquisition of unwanted changes during amplification or cloning. The pET-Fcp1* plasmids were introduced into E. coli BL21(DE3)-RIL, and the mutant Fcp1 proteins were purified from soluble bacterial lysates using the same protocol described for the wild-type Fcp1.
N-terminal deletion mutants were constructed by PCR amplification with mutagenic sense-strand primers that introduced an NcoI restriction site and a methionine codon in lieu of the codon for Gly 106 or Gly 139 . C-terminal deletion mutants were constructed by PCR amplification with mutagenic antisense-strand primers that introduced a stop codon in lieu of the codons for Pro 581 , Lys 487 , or Glu 400 and a BamHI site 3Ј of the new stop codon. The PCR products were digested with NcoI and BamHI and then inserted into pET16m. The inserts of the resulting pET-Fcp1⌬ plasmids were sequenced completely to confirm the desired mutations and exclude the acquisition of unwanted changes. The pET-Fcp1⌬ plasmids were introduced into E. coli BL21(DE3)-RIL, and the truncated Fcp1 proteins were purified from soluble bacterial lysates as described for the wild-type Fcp1.
Velocity Sedimentation-An aliquot (50 g) of the wild-type Fcp1 preparation was mixed with catalase (45 g), bovine serum albumin (45 g), and cytochrome c (45 g), and the mixture was applied to a 4.8-ml 15-30% glycerol gradient containing 50 mM Tris-HCl (pH 7.4), 0.1 M NaCl, 1 mM EDTA, 2 mM dithiothreitol, 0.05% Triton X-100. The gradient was centrifuged in a SW50 rotor at 50,000 rpm for 15 h at 4°C. Fractions (0.2 ml) were collected from the bottom of the tube. Aliquots (20 l) of odd-numbered gradient fractions were analyzed by SDS-PAGE. Aliquots (5 l) of every fraction were assayed for hydrolysis of pNØP.
Phosphatase Assay-Reaction mixtures (100 l) containing 50 mM Tris acetate (pH 5.5), 10 mM MgCl 2 , 10 mM pNØP, and Fcp1 as specified were incubated for 30 min at 37°C. The reactions were quenched by adding 900 l of 1 M sodium carbonate. Release of p-nitrophenol (pNØ) was determined by measuring A 410 and extrapolating the value to a pNØ standard curve.
CTD Phosphatase Assay-N-terminal biotinylated CTD phosphopeptides composed of four tandem YSPTSPS repeats containing phosphoserine at either position 2, position 5, or positions 2 and 5 of each repeat were synthesized and purified as described previously (11,21). CTD phosphatase reaction mixtures (25 l) containing 50 mM Tris acetate (pH 5.5), 10 mM MgCl 2 , 25 M CTD peptide, and Fcp1 as specified were incubated for 60 min at 37°C. The reactions were quenched by adding 0.5 or 1 ml of malachite green reagent (BIOMOL green reagent, purchased from BIOMOL Research Laboratories, Plymouth Meeting, PA). Phosphate release was determined by measuring A 620 and extrapolating the value to a phosphate standard curve.

Phosphatase Activity of Recombinant S. pombe Fcp1-The S.
pombe open reading frame encoding Fcp1 was PCR-amplified from a cDNA library and cloned into a T7 RNA polymerasebased bacterial expression vector so as to fuse the Fcp1 protein to a 14-amino acid N-terminal leader peptide containing 10 tandem histidines. The expression plasmid was introduced into E. coli BL21(DE3), a strain that contains the T7 RNA polymerase gene under the control of a lacUV5 promoter. A prominent 105-kDa polypeptide was detectable by SDS-PAGE in whole cell extracts of isopropyl-1-thio-␤-D-galactopyranosideinduced bacteria grown at 17°C (not shown). This polypeptide was not present when bacteria containing the pET vector alone were induced with isopropyl-1-thio-␤-D-galactopyranoside. After centrifugal separation of the crude lysate, the Fcp1 protein was recovered in the soluble supernatant fraction and was the single most abundant polypeptide in the soluble fraction (not shown). The His tag facilitated rapid purification of the recombinant Fcp1 by adsorption to an immobilized nickel resin and subsequent elution with buffer containing imidazole. SDS-PAGE analysis of the peak imidazole eluate fraction showed that the preparation was highly enriched with respect to the 105-kDa Fcp1 polypeptide ( Fig. 2A). The yield of Fcp1 was 30 -40 mg of protein/liter of induced bacterial culture. The apparent size of S. pombe Fcp1 as gauged by SDS-PAGE was greater than the value of 84 kDa predicted for the His-tagged gene product. Aberrant slow electrophoretic mobility of Fcp1 has been noted for native Fcp1 purified from yeast or human cells and for recombinant versions of Fcp1 (7, 9, 15), including S. pombe Fcp1 (10).
Initial characterization of the phosphatase activity of the purified recombinant Fcp1 protein was performed using 10 mM pNØP as a substrate. The pNØ reaction product was detected via its absorbance at 410 nM. We found that Fcp1 catalyzed the conversion of pNØP to pNØ and that the extent of the reaction was directly proportional to the concentration of the recombinant protein (Fig. 2B). Optimal reaction conditions were delineated via systematic variations of the pH and the concentrations of divalent cation cofactors and the pNØP substrate. The first notable finding was that the Fcp1 phosphatase activity displayed a bell-shaped pH profile with an acidic optimum at pH 5.0 -5.5 (Fig. 3A). Phosphatase activity was virtually nil at lower pH (pH Յ 4.5) and also declined sharply as the pH was increased toward neutrality. The activity of Fcp1 at pH 7.5 was 10% of the activity at pH 5.5.
Hydrolysis of pNØP by S. pombe Fcp1 required a divalent cation cofactor. Magnesium supported optimal activity at 5-10 mM concentration; at lower concentrations (Յ1.25 mM) activity was directly proportional to the input magnesium (Fig. 3B). Manganese and cobalt were also capable of satisfying the divalent cation requirement (Fig. 3C). The concentrations of manganese and cobalt that supported optimal activity (0.12-1 mM cobalt and 0.06 -1 mM manganese) were lower than the optimal effector concentrations for magnesium. The activity levels in 1 mM cobalt and 1 mM manganese were 60 and 35%, respectively, of the activity in 10 mM magnesium. In contrast, calcium was completely ineffective as a phosphatase cofactor up to 10 mM (Fig. 3B). Our finding that calcium could not activate Fcp1 is in agreement with the findings for the human CTD phosphatase (5). In contrast, S. cerevisiae CTD phosphatase can be activated by calcium (6).
The kinetic parameters were gauged by measuring the production of pNØ as a function of substrate concentration in the range of 5-40 mM pNØP. A double-reciprocal plot of the data fit well to a linear function (Fig. 3D). From these data, we calculated a K m of 19 mM pNØP and a k cat of 2 s Ϫ1 . The turnover number of S. pombe Fcp1 is 1000-fold higher than the k cat value of 2 ϫ 10 Ϫ3 s Ϫ1 reported for the hydrolysis of pNØP by recombinant S. cerevisiae Fcp1 (15). The K m value of the S. cerevisiae enzyme for pNØP (60 mM) is 3-fold higher than that of S. pombe Fcp1.
Increasing ionic strength exerted a modest inhibitory effect on the hydrolysis of pNØP by S. pombe Fcp1. The extents of pNØ formation in the presence of 125, 250, and 500 mM NaCl were 75, 48, and 33% of the activity in the absence of added salt (data not shown).
Sedimentation of S. pombe Fcp1 in a Glycerol Gradient-The native size of Fcp1 was gauged by zonal velocity sedimentation through a 15-30% glycerol gradient. Marker proteins catalase (248 kDa), bovine serum albumin (66 kDa), and cytochrome c (13 kDa) were included as internal standards. After centrifugation, the polypeptide compositions of the odd-numbered gradient fractions were analyzed by SDS-PAGE (Fig. 4A). The Fcp1 polypeptide (calculated size, 84 kDa) sedimented as a single discrete peak coincident with bovine serum albumin. The phosphatase activity profile paralleled exactly the sedimentation profile of the Fcp1 polypeptide (Fig. 4B). The peak of Fcp1 activity was one fraction lighter than the bovine serum albumin peak. Centrifugation of the recombinant Fcp1 protein in a parallel gradient without marker proteins resulted in an identical single-peak profile of phosphatase activity that coincided with the distribution of the Fcp1 polypeptide (not shown). These results suggest that S. pombe Fcp1 is a monomeric protein in solution, probably with an elongated shape.
Fcp1 Residues Asp 170 and Asp 172 Are Essential for Catalysis-Fcp1 proteins from diverse sources display extensive primary structure similarity across the region spanning S. pombe Fcp1 residues 140 -326 (Fig. 1). Within this so-called Fcp1 homology domain is a short conserved peptide motif ( 167 LIVDLDQTII 176 in S. pombe Fcp1) that corresponds to the signature sequence of a family of metal-dependent phosphohydrolases and phosphotransferases (22,23). The two aspartates in the DXDXT element were found to be essential for phosphoryl transfer by human phosphomannomutase and L-3-phosphoserine phosphatase (22,24). Several family members have been shown to act via an acyl-phosphoenzyme intermediate in which the phosphate is linked to the first aspartate in the DXDXT motif (25). Kobor et al. (15) recognized the presence of the DXDXT motif in S. cerevisiae Fcp1 and found that mutation of either aspartate abolished Fcp1 function.
To assess the role of the DXD motif in S. pombe Fcp1, we replaced Asp 170 and Asp 172 with alanine, asparagine, and glutamate and then purified the recombinant mutant proteins from bacteria ( Fig. 2A). The D170A, D170N, and D170E proteins were apparently inert in hydrolyzing pNØP at up to 5 g of input protein (Fig. 2B). Comparison with the titration profile of wild-type Fcp1 (which released 21 nmol of pNØ/g of protein in 30 min) indicated that the specific activities of the three Asp 170 mutants were Ͻ0.2% of the activity of wild-type Fcp1. The D172A, D172N, and D172E mutants were also catalytically defective, with their respective specific activities being 1.8, 4, and 1% of the wild-type value (Fig. 2B). We conclude that: (i) Asp 170 and Asp 172 are likely constituents of the active site of S. pombe Fcp1; (ii) catalysis is dependent on a carboxylate functional group at both positions; and (iii) there is a steric constraint on the main chain to carboxylate distance both positions that precludes functional substitution of either aspartate by glutamate.
Deletion Analysis Defines a Phosphatase Catalytic Domain-Two N-terminal deletion mutations Fcp1(107-723) and Fcp1-(140 -723) were designed to progressively truncate the Fcp1 protein up to the margins of the conserved region shown in Fig.  1. The N-terminal deletion mutants were expressed in bacteria as His-tagged fusions and purified from soluble lysates by nickel-agarose chromatography (Fig. 5A). The Fcp1(107-723) and Fcp1(140 -723) proteins migrated more rapidly during SDS-PAGE than the wild-type Fcp1, as expected, although their mobility was still anomalously slow compared with their calculated sizes of 72 and 66 kDa, respectively. Fcp1(107-723) and Fcp1(140 -723) both retained activity in the hydrolysis of pNØP to pNØ (Fig. 5B); their respective specific activities were 83 and 120% of the wild-type activity. We conclude that the N-terminal 140 amino acids of S. pombe Fcp1 are not required for catalysis.
Fcp1 Dephosphorylates a Defined CTD Phosphopeptide-A caveat inherent in the use of pol IIo as a substrate for CTD phosphatases is that the structure of the phosphorylation array is not known and is likely to consist of a heterogeneous mixture of different arrays. A rigorous treatment of the substrate spec-ificity of Fcp1 requires a defined CTD molecule in which the number and position of the phosphates is known and, if possible, subject to manipulation by the investigator. In the present study, we examined the activity of recombinant S. pombe Fcp1 with a synthetic 28-amino acid phosphopeptide (Fig. 6) consisting of four tandem repeats of the CTD heptad sequence (YSPTSPS) in which all Ser 2 and Ser 5 residues are Ser-PO 4 .
Fcp1 was incubated for 60 min at 37°C with 25 M of the Ser 2 -PO 4 /Ser 5 -PO 4 CTD peptide and 10 mM MgCl 2 . Release of inorganic phosphate from the CTD was measured colorimetrically using the malachite green method. We found that wildtype Fcp1 was active in dephosphorylating the synthetic CTD phosphopeptide; the extent of P i release was proportional to input Fcp1 in the range of 40 -620 ng (Fig. 6A). P i release plateaued at 2.5-5 g of input Fcp1, at which point ϳ60% of the input phosphoserine residues had been hydrolyzed. The specific activity on this substrate in the linear range of the Fcp1 titration curve was 7 nmol of P i release/g of protein in 60 min, which corresponds to a turnover number of ϳ0.16 s Ϫ1 . P i release was ablated when magnesium was omitted from the reaction mixture (data not shown). The activity of the recombinant Fcp1 preparation with the synthetic CTD phosphopeptide substrate was evidently intrinsic to Fcp1, insofar as the recombinant D170A mutant preparation catalyzed no detectable P i release up to 5 g of input protein (Fig. 6A).
Fcp1 activity on the Ser 2 -PO 4 /Ser 5 -PO 4 CTD peptide displayed a bell-shaped pH profile with an acidic optimum at pH 5.5 (Fig. 6B). As we observed earlier with Fcp1 activity on the pNØP substrate, its CTD phosphatase activity was also abol- ished at a pH level of Յ4.5. The CTD phosphatase activity diminished as the pH was increased toward neutrality. P i release from the CTD at pH 7.5 was ϳ20% of the activity at pH 5.5.
Fcp1 Is More Active on a CTD Ser 2 -PO 4 Substrate than a CTD Ser 5 -PO 4 Substrate-The preceding experiment using a doubly phosphorylated Ser 2 -PO 4 /Ser 5 -PO 4 peptide substrate showed clearly that Fcp1 is capable of dephosphorylating the CTD outside the context of pol II. To gauge whether Fcp1 has any intrinsic preference for the position of phosphoserine within the CTD heptad, we employed synthetic 28-amino acid peptides consisting of four tandem repeats phosphorylated exclusively at Ser 2 or Ser 5 of each heptad. Reaction of Fcp1 with 25 M of the CTD Ser 2 -PO 4 peptide resulted in P i release proportional to input Fcp1 (Fig. 7). Dephosphorylation of Ser 2 plateaued at 0.6 -5 g of input Fcp1, at which point 98% of the input phosphoserine residues had been hydrolyzed. The specific activity was 16 nmol of P i release/g of protein in 60 min, which corresponds to a turnover number of ϳ0.37 s Ϫ1 . The instructive finding was that Fcp1 was much less active in dephosphorylation of Ser 5 . The titration profile with 25 M of the CTD Ser 5 -PO 4 substrate displayed a clear shift to the right (Fig. 6). The specific activity of 1.6 nmol of P i release/g of protein in 60 min was 1 order of magnitude lower than the activity of Fcp1 in dephosphorylating the Ser 2 -PO 4 substrate. We conclude that S. pombe Fcp1 displays an intrinsic preference for dephosphorylation of Ser 2 versus Ser 5 of the CTD heptad.
The specificity of Fcp1 for the CTD peptide was underscored by an assay of the ability of Fcp1 to dephosphorylate free phosphoserine. A 60-min reaction of 10 g of Fcp1 with 2 mM phosphoserine (a level 20-fold higher than the total concentration of phosphoserine present in 25 M of the 28-mer CTD Ser 2 -PO 4 peptide) resulted in the release of 0.3 nmol of P i . The specific activity with 2 mM phosphoserine was 0.2% of the activity with 25 M CTD Ser 2 -PO 4 peptide. We surmise that the CTD itself and the position of the phosphoserine within the heptad are recognized by Fcp1. DISCUSSION Our studies of the physical and enzymatic properties of recombinant S. pombe Fcp1 complement and extend prior studies of CTD phosphatases from other sources; they also provide important new insights to the intrinsic ability of Fcp1 to dephosphorylate the isolated CTD (divorced from the body of pol II) and an inherent preference of Fcp1 for dephosphorylation of CTD Ser 2 -PO 4 versus CTD Ser 5 -PO 4 .
Fission yeast Fcp1 provides an excellent model system for biochemical and structural analysis, insofar as the soluble recombinant protein is expressed at very high levels in bacteria and is easily purified. We surmised via velocity sedimentation that S. pombe Fcp1 has a monomeric native structure, most likely with an asymmetric shape to account for the fact that it sediments slightly lighter than expected for a globular protein of 84 kDa. Our data agree with the findings of Chambers and Kane (6) for the S. cerevisiae CTD phosphatase, which they judged via sedimentation and gel filtration to be an elongated monomer. Human Fcp1 is a 961-amino acid polypeptide, but it chromatographs during gel filtration with an apparent size of ϳ300 kDa (9). It is not clear whether the native mammalian enzyme is a monomer or an oligomer of the Fcp1 polypeptide. The human Fcp1 polypeptide is 30% larger than the fungal Fcp1 polypeptides (723-732 amino acids), and the human protein contains two long internal structural modules that are missing from the fungal orthologs. Thus, it is conceivable that the segments unique to the human enzyme might either extend the conformation of a human Fcp1 monomer or else mediate its oligomerization.
One of our notable findings, with practical and mechanistic implications, is that Fcp1 activity displays an acidic pH optimum (at pH 5.5) for the hydrolysis of pNØP and the phosphorylated CTD. Practically, it is conceivable that prior studies of CTD phosphatases underestimated Fcp1 catalytic activity because the assays were performed at a pH level of Ն7.5. Mechanistically, the pH profile suggests the existence of at least two different functional groups at which the protonation state has a strong impact on phosphatase activity. We speculate that the sharp decrease in catalysis between pH 5.5/5.0 and pH 4.5 is a consequence of protonation of one or both of the essential Asp residues of the 170 DXD 172 active site motif. We entertain two possible explanations for the decline in activity as the pH is increased above neutrality: (i) activity requires a protonated functional group on the Fcp1 enzyme (e.g. a histidine, with a presumptive pK a of ϳ6.5) or (ii) Fcp1 is more active when the phosphate group of the substrate is in the monoanionic state (protonated) rather than the dianionic state; this transition would have a presumptive pK of ϳ6.5-7.0.
Our demonstration that the Asp 170 side chain of the 170 DXD 172 motif of S. pombe Fcp1 is strictly essential for catalysis is consistent with the mutational studies of the equivalent side chain (Asp 180 ) of S. cerevisiae Fcp1 (15). However, whereas we found that replacement of Asp 172 of S. pombe Fcp1 with either Asn or Glu caused a severe decrement in hydrolysis of pNØP, the activity of the corresponding S. cerevisiae Fcp1 mutant D182E with pNØP was 42% of the wild-type activity. The different residual activity of the glutamate mutants (1% for S. pombe versus 42% for S. cerevisiae) may reflect the different conditions used for the phosphatase assay. Alternatively, different enzymes in the DXD phosphatase family may vary in their sensitivity to changes in the main chain-to-carboxylate distance at the distal Asp of the motif. We found recently that both Asp residues of the DXD motif are strictly required for the activity of T4 polynucleotide 3Ј phosphatase, i.e. neither could be substituted functionally by glutamate (26).
Deletion analysis showed that large segments could be deleted from the N and C termini of S. pombe Fcp1 without loss of phosphatase activity. The apparent margins of the catalytic domain (amino acids 140 -580) are in accord with the results of in vivo deletion analysis of S. cerevisiae Fcp1 (27). In both fission and budding yeast, the BRCT domain is apparently critical for Fcp1 function along with the Fcp1 homology domain. The complete loss of S. pombe Fcp1 phosphatase activity in vitro with deletion of the BRCT domain implicates this segment either in ensuring the proper tertiary structure of the protein or in supplying one or more functional groups to the phosphatase active site.
Ser 5 and Ser 2 are both extensively phosphorylated in vivo, and various CTD serine kinases differ in their site preference (28 -30). If the enzymes that add phosphates to the CTD can discriminate position within the heptad, then it is reasonable to think that the enzymes that dephosphorylate the CTD might also have inherent site specificity. This issue had not been addressed previously. Here we found that, contrary to prior reports, Fcp1 was able to dephosphorylate an isolated CTD substrate divorced from the rest of pol II. Indeed, Fcp1 was capable of releasing a large molar excess of phosphate from the CTD peptide substrate during the in vitro reaction (ϳ1300-fold excess in the case of the CTD Ser 2 -PO 4 peptide). It is worth emphasizing that the CTD phosphatase activity was observed at 25 M concentration of the synthetic CTD phosphopeptide, i.e. nearly 3 orders of magnitude lower than the K m of Fcp1 for the nonspecific substrate pNØP. The inactivation of the CTD phosphatase by the D170A mutation dispelled any doubt that the CTD phosphatase function was intrinsic to Fcp1. Earlier studies of Fcp1 activity on the isolated CTD may not have been optimally sensitive because of the use of higher pH reaction conditions than those used presently and, perhaps more significantly, the use of a recombinant CTD fusion protein substrate that had been phosphorylated on Ser 5 (see below).
We exploited the synthetic CTD peptide substrates to show for the first time Fcp1 displays an inherent preference for a particular CTD phosphorylation array. Using equivalent concentrations of CTD peptides of identical amino acid sequence and phosphoserine content, which differed only in the positions of phosphoserine within the heptad, we found that Fcp1 was 10-fold more active in dephosphorylating Ser 2 -PO 4 than Ser 5 -PO 4 . Preferential action of Fcp1 at Ser 2 -PO 4 in vitro provides an explanation for the observation of Cho et al. (14) that conditional mutations of S. cerevisiae Fcp1 lead to increased levels of CTD Ser 2 phosphorylation in vivo at the semipermissive temperature. Their finding that Ser 5 phosphorylation was apparently not affected in the fcp1 mutants led to the suggestion that a phosphatase other than Fcp1 is likely to be responsible for dephosphorylating Ser 5 in vivo (14). Schroeder et al. (12) reported that Ser 5 phosphate levels did increase on transcribed genes in fcp1 mutants at restrictive temperature, and they inferred that Fcp1 is involved in dephosphorylating Ser 5 in vivo. Here we show that recombinant Fcp1 can dephosphorylate CTD Ser 5 -PO 4 , albeit less efficiently than Ser 2 -PO 4 . It is conceivable that the Ser 5 and Ser 2 dephosphorylation activities of conditional mutants of Fcp1 display differential sensitivity to thermal inactivation.
The finding that S. pombe CTD phosphatase has inherent specificity for phosphate position raises numerous interesting questions about the potential for regulation of Fcp1 action, e.g. via variations in the complexity of the phosphorylation array, changes in CTD structure or accessibility caused by proteins that interact with the CTD, and changes in Fcp1 catalytic activity or substrate specificity elicited by proteins (or other regulators) that interact with Fcp1. We anticipate that synthetic CTD phosphopeptides will provide powerful tools to dissect the structural requirements for substrate recognition and catalysis by Fcp1 and the effects of potential regulatory factors.