An essential component of a C-terminal domain phosphatase that interacts with transcription factor IIF in Saccharomyces cerevisiae

Abstract

One of the essential components of a phosphatase that specifically dephosphorylates the Saccharomyces cerevisiae RNA polymerase II (RPII) large subunit C-terminal domain (CTD) is a novel polypeptide encoded by an essential gene termed FCP1. The Fcp1 protein is localized to the nucleus, and it binds the largest subunit of the yeast general transcription factor IIF (Tfg1). In vitro, transcription factor IIF stimulates phosphatase activity in the presence of Fcp1 and a second complementing fraction. Two distinct regions of Fcp1 are capable of binding to Tfg1, but the C-terminal Tfg1 binding domain is dispensable for activity in vivo and in vitro. Sequence comparison reveals that residues 173–357 of Fcp1 correspond to an amino acid motif present in proteins of unknown function predicted in many organisms.

Promoter-dependent transcription by RNA polymerase II (RPII) requires six general transcription factors (reviewed in ref. 1). The form of RPII that binds preferentially to promoters is not extensively phosphorylated on the C-terminal heptapeptide-repeat domain (CTD) of its largest subunit (2–4). Instead, the CTD becomes phosphorylated during or shortly after initiation, and elongating RPII generally has a phosphorylated CTD (5, 6). CTD phosphorylation can be a regulated process: heat shock induction of the Drosophila hsp70 gene causes RPII, paused downstream of the promoter with an unphosphorylated CTD, to transcribe downstream sequences with a phosphorylated CTD (7). Activator-induced CTD phosphorylation also may facilitate RNA chain elongation (8–11). Prior to or following transcriptional termination, CTD dephosphorylation presumably must occur to regenerate enzyme capable of reinitiating transcription.

Transcription factor IIF (TFIIF) (RAP30/74) binds directly to RPII and helps assemble RPII into the preinitiation complex at a promoter (12–14). Elongating RPII also interacts with TFIIF, which stimulates its rate of chain elongation (15, 16). TFIIF also stimulates the activity of a CTD phosphatase purified from human cells (17, 18) and yeast (19). Two components are required for yeast CTD phosphatase activity (19), one of which is described here, encoded by an essential gene, and closely related to a human protein also necessary for human CTD phosphatase activity (J.A., H. Xiao, G. Pan, G. Dahmus, M.C., S. Zhang, R. G. Roeder, M. Dahmus, and J.G., unpublished work). That TFIIF can modulate the CTD phosphatase activity highlights the importance of a phosphorylation cycle for RPII in regulating the initiation and elongation reactions of transcription.

MATERIALS AND METHODS

Plasmid Construction and Preparation of Recombinant yFcp1.

Yeast growth and manipulations were performed essentially as described (20). Yeast were transformed by the lithium acetate method (21).

Plasmid JA739 was used to express in yeast the complete Fcp1 ORF under the control of the GAL10 promoter. A BglII-cut PCR fragment encoding yFcp1 (amplified using the primers 5′-CCCAGATCTATGACCACACAAATAAGGTCTCCC-3′ and 5′-CCCAGATCTGATACGGCATCTGAGCTGCTAATC-3′; the ATG and Stop codons are underlined) was inserted into the BamHI site of plasmid pYGal. PYGal is a derivative of pFL39 (TRP1 CEN6 ARS; 22) and carries the GAL10 promoter (inserted between the EcoRI and SmaI sites) and the PGK transcription termination signal (between the SphI and HindIII sites) (obtained from James Friesen, University of Toronto). Plasmid JA754 [expresses functional Fcp1 protein from the GAL10 promoter with the protein tagged at its N terminus with six histidine residues followed by the hemagglutinin (HA) epitope] contained a double-stranded oligonucleotide (5′-ATGGCCCATCATCATCATCATCATGCCATGGCTTACCCATACGATGTTCCAGATTACGCTG-3′ and its complement) inserted into the SmaI site of JA739 immediately upstream of the yFcp1 ORF. Plasmid JA756 encoding truncated Fcp1 protein lacking the C-terminal 66 amino acids, was constructed by replacing a SpeI-SalI yFCP1 fragment in JA754 with a SpeI-XhoI cut PCR product encoding amino acids 457–666 of Fcp1 with a synthetic TAG Stop codon downstream of codon 666. Plasmid JA798 encoding a truncated Fcp1 protein lacking the C-terminal 275 amino acids, contained the palindromic oligonucleotide 5′-CTAGCCTAGGCCTAGG-3′ encoding an in-frame stop codon (underlined) inserted into the SpeI site of plasmid JA754.

Plasmid JA782 encoding amino acids 1–732 of Fcp1 fused to the HA epitope was constructed by isolating an NcoI-HindIII yFCP1 fragment from plasmid JA754 and inserting it downstream of the T7 promoter, between the NcoI and HindIII sites of plasmid pET23-d (Novagen). Plasmid pJA785 encoding amino acids 1–666 of Fcp1 downstream of the T7 promoter was constructed by inserting a NcoI-HindIII fragment from pJA756 between the NcoI and HindIII sites of pET23d. Plasmid pJA801 encoding amino acids 1–457 of Fcp1 contained the palindromic oligonucleotide 5′-CTAGCCTAGGCCTAGG-3′ encoding an in-frame stop codon (underlined) inserted into the SpeI site of pJA782. A derivative of JA782 (T7-HA-yFcp1) expressing a polyhistidine- and HA-tagged Fcp1 protein (21 additional amino acids MHHHHHHMAYPYDVPDYAGGS) was constructed by inserting the following double-stranded oligonucleotide into the NcoI site of JA782 (5′-CATGCACCACCACCACCACCA-3′ and 5′-CATGTGGTGGTGGTGGTGGTG-3′; the initiator ATG codon is underlined).

Plasmids pDB3 and pDB4, expressing amino acids 1–732 and 1–626, respectively, from the endogenous FCP1 promoter, were constructed by cloning fragments of FCP1 into pDB2. Plasmid pDB2 contained the FCP1 promoter region (−303 to −10 relative to the first nucleotide in the start codon of the ORF) and the region downstream of the stop codon of the ORF (+4 to +295 relative to the first nucleotide in the stop codon) inserted into the EcoRI and SpeI sites, respectively, of the low copy plasmid pRS313 (23). The promoter region was amplified using PCR with primers: 5′-TCGGATCCAACGTCTAGATGCCGCAACC-3′ and 5′-CGCAGATCTGCTTGTGTTTACACGTTCGC-3′. The downstream region was amplified using primers: 5′-TCGGATCCCTCAGATGCCGTATCTTTCC-3′ and 5′-CGCAGATCTAGCAGCTAATGGCTCGAGAG-3′. Amplification was carried out with the American Type Culture Collection Cosmid 8021. The NcoI/XhoI fragment of pJA782 containing the full-length ORF of FCP1 was cloned into the SmaI site of pDB2 to make pDB3. The NcoI/NdeI fragment of pJA782 containing residues 1–626 of the FCP1 ORF was cloned into the SmaI site of pDB2 to make pDB4.

Plasmids pJA794 and pJA796 expressing, respectively, yFcp1(1–732) and yFcp1(1–666) from the endogenous yFCP1 promoter contained a 738-bp, NcoI-cut PCR fragment encoding the yFCP1 promoter and upstream region (amplified from yeast genomic DNA using primers: 5′-GGGGCCATGGTCCCATGTTTAGACAGATGCTCACC-3′ and 5′-GGGGCCATGGTCTGCTTTGCTTGTGTTTACACGTTCG-3′) inserted into the NcoI site of plasmids pJA754 and pJA756. Plasmid pJA804 expressing yFcp1(1–457) contained the palindromic oligonucleotide 5′-CTAGCCTAGGCCTAGG-3′ encoding an in-frame stop codon (underlined) inserted into the SpeI site of plasmid pJA794.

Plasmids used to express recombinant Fcp1 and its deletion derivatives were transformed into Escherichia coli BL21 (DE3), and the cells were induced according to Novagen’s instructions. Cells were harvested, washed in cold distilled water, resuspended in 50 mM Tris-acetate, pH 7.8/20% glycerol/0.5 M potassium acetate/5 mM imidazole/0.1% Triton X-100/5 mM 2-mercaptoethanol/1 mM phenylmethylsulfonyl fluoride, and the cells were lysed by sonication. The insoluble recombinant yFcp1 protein was recovered by centrifugation of the extract at 12,000 × g for 20 min and the pellet was resolubilized in the same buffer as above with 6 M urea. Insoluble material was removed by centrifugation and the supernatant was loaded onto a Ni-nitrilotriacetate (NTA)-agarose column (Qiagen, Chatsworth, CA). The column was washed with the same buffer containing 6 M urea and 20 mM imidazole, and the recombinant Fcp1 was eluted in the same buffer with 100 mM imidazole. Recombinant Fcp1 was renatured as described (24).

Plasmids encoding amino acids 458–732, 667–732, and 458–666 of Fcp1 fused to E. coli thioredoxin (TR) were constructed by inserting BamHI- and EcoRI-cut PCR fragments into the BamHI and EcoRI sites of pET32-a (Novagen). In each case, plasmid JA782 was used as template and the amplifying primers were as follows: 5′-GGGGGATCCGTTGATGACGATGATGAACTATAC-3′ and 5′-GGGGAATTCTAATCATCCAGCATATCCATCAA-3′ for amino acids 458–732 [TR-yFCP1 (458–732)]; 5′-GGGGG ATCCGACGAAAGTGATGACGAAAACAAC-3′ and 5′-GGGGAATTCTAATCATCCAGCATATCCATCAA-3′ for amino acids 667–732 [TR-yFCP1 (667–732)]; 5′-GGGGGATCCGTTGATGACGATGATGAACTATCA-3′ and 5′-GGGGAATTCTAGTCGTGGTCGTCATCTTCGTC-3′ for amino acids 458–666 [TR-yFCP1 (458–666)]. Stop codons are underlined. The human FCP1-expressing plasmid [hFCP1a(443–842)] encodes a 400 amino acid C-terminal fragment inserted into the BamHI site of pRSETC (Invitrogen), and its isolation and construction will be described elsewhere. Proteins were produced in vitro as described below. Plasmid pJA728 encoding the C-terminal 87 amino acids of yeast RAP74 (TFG1) fused to glutathione S-transferase (GST) was constructed by inserting a BglII-cut TFG1 PCR fragment (amplified from yeast genomic DNA using primers: 5′-CCCAGATCTCGGCATCGAATACAGTGCC TTCGCC-3′ and 5′-CCCAGATCTCTACTCTTTCTTTAATTCCATGTGGTC-3′; codon encoding Ala-649 and stop codon are underlined) into the BamHI site of pGEX-3X (Pharmacia). Plasmid pJA8 encoding the C-terminal 82 amino acids of human RAP74 fused to GST will be described elsewhere.

Protein Affinity Chromatography.

Columns (40 μl) containing GST, GST-yRAP74 (649–735), or GST-hRAP74 (436–517) immobilized on glutathione-Sepharose at a concentration of 3–4 mg/ml were prepared (25) using immobilized proteins from E. coli transformed with pGEX-3X, pJA728, or pJA8. Columns were equilibrated with 300 μl of ACB buffer (10 mM Hepes, pH 7.9/1 mM EDTA/1 mM DTT/10% glycerol) containing 0.1 M NaC1 and 5 mg/ml BSA, and next with 400 μl of ACB buffer containing 0.1 M NaC1 and 1 mg/ml BSA. Columns were loaded with 20 μl of rabbit reticulocyte lysate from the TNT transcription translation system (Promega) programmed with 0.4 μg of various plasmid DNAs [T7-HA-yFcp1, pJA785, pJA801, TR-yFCP1(458–732), TR-yFCP1(667–732), TR-yFCP1(458–666), or hFCP1a(443–842)], after the reticulocyte lysate had been diluted 10 times with ACB buffer containing 0.1 M NaC1 and 1 mg/ml BSA. After loading, columns were washed with 400 μl of ACB buffer containing 0.1 M NaC1. Bound proteins were eluted sequentially with 120 μl of ACB containing 1 M NaC1 and 120 μl of ACB buffer containing 1% SDS. Aliquots of the input proteins and of the NaC1 and SDS eluates were analyzed by SDS/PAGE and autoradiography.

Disruption of the Yeast FCP1 Gene.

Plasmid JA744 used to disrupt the endogenous FCP1 gene was constructed in two steps. In the first step, a BglII cut PCR fragment corresponding to the genomic region located immediately upstream of the FCP1 ORF (amplified from yeast genomic DNA using the following two oligonucleotides: 5′-CCCAGATCTGTCCCATGTTTAGACAGATGCTCACC-3′ and 5′-CC CAGATCTAATCTGCTTTGCTTGTGTTTACACGTTC-3′; the portion of each oligonucleotide that hybridizes to genomic DNA is underlined) was inserted upstream of the LEU2 gene into the BamHI site of plasmid JJ250 (26). The resulting plasmid was next modified by inserting a HindIII cut PCR fragment corresponding to sequences located downstream of the FCP1 ORF (amplified from yeast genomic DNA using the following two oligonucleotide primers: 5′-CCCCAAGCTTCAGCTCAGATGCCGTATCTTTCCACA-3′ and 5′-CCCCAAGCTTGAGCTCCTTAGAATGGAATCTATGATCGTGT-3′) downstream of the LEU2 gene into the HindIII site. The resulting plasmid, pJA744, carries a modified version of the FCP1 genomic locus in which the FCP1 ORF was replaced with the LEU2 gene. This disrupted allele of FCP1 was released from the plasmid backbone by digestion with SmaI and SacI and used to disrupt one of the FCP1 alleles in the diploid yeast strain W303 (Mata/Matα can1–100/can1–100 his3–11,15/his3–11,15 leu2–3,112/leu2–3,112 trp1–1/trp1–1 ura3/ura3 ade2–1/ade2–1) using a one-step gene disruption protocol (27). Disruption was confirmed by PCR (data not shown), creating strain JA830.

Indirect Immunofluorescence.

Cells were prepared for immunofluorescence (28, 29) with minor modifications. Fixed and washed cells were incubated 2 hr at room temperature with a 1:20 dilution of preabsorbed (30) MMS101R monoclonal anti-HA antiserum (Babco, Richmond, CA; provided by Kevin Madden, Yale University, New Haven, CT). After washing, preabsorbed (30) secondary antibody (CY3-conjugated goat anti-mouse antibody from Jackson ImmunoResearch, 1:40 dilution, provided by Kevin Madden) was applied. After washing, cells were stained with 4′,6-diamidino-2-phenylindole and observed by microscopy.

RESULTS

Identification of an Essential Component of a Yeast CTD Phosphatase.

Purification of CTD phosphatase from extracts of Saccharomyces cerevisiae led to the resolution of two component fractions essential for the phosphatase activity (19). One of these fractions contained yeast TFIIF, but could not be replaced with highly purified TFIIF. When the other essential component was purified to homogeneity, a polypeptide with an apparent molecular mass of 100/103 kDa during SDS/PAGE copurified with the activity (see Fig. 1A, lane 3). Excision of this polypeptide from the gel, followed by removal of the SDS and renaturation, led to recovery of the CTD phosphatase activity in combination with the TFIIF-containing fraction (ref. 19; see Fig. 1B, lane 3). As reported, in this phosphatase assay, the enzyme specifically removes phosphate from the CTD in the ³²P-labeled IIo form of the largest subunit of RPII, converting it to the dephosphorylated IIA form (19).

Figure 1.

Reconstitution of CTD phosphatase activity with recombinant yeast Fcp1. (A) Yeast Fcp1 purified from yeast whole cell extracts [yFCP1, Fraction 62, MonoQ Fraction (19)] and recombinant Fcp1 (ryFCP1) produced in E. coli and purified by Ni-NTA-garose chromatography were separated on a SDS/7.5% polyacrylamide gel and stained with Coomassie blue. An arrow indicates the position of ryFcp1 that contains 21 extra amino acids: six histidines and a HA tag. (B) Assay of yFcp1 and ryFcp1 for CTD phosphatase activity. The assay (19) measures removal of ³²P from the largest subunit of RPII. Lane 1, autoradiographic signal from the phosphorylated form of the largest subunit of yeast RPII. Lanes 2 and 3, signal remaining after treatment with either renatured recombinant (≈0.2 μg) or purified yeast (≈0.02 μg) Fcp1, respectively; these amounts of protein were saturating in the assay.

The 100/103-kDa polypeptide excised from the SDS gel was digested with the protease lysC and the resulting peptides were resolved by HPLC on an Applied Biosystems C8 Aquipore-300 column. Partial amino acid sequences were obtained for three well-resolved peptides. All three sequences (KFWYLVERASDTGDD, amino acids 38–52; KEFFA, amino acids 254–258; KLXXXLATAXE, amino acids 405–415) were present in a 732 amino acid ORF on chromosome XIII of S. cerevisiae. Expression of this yeast gene in E. coli along with an N-terminally fused influenza virus HA epitope and a polyhistidine tag led, as expected, to the production of a recombinant protein only slightly less mobile than the natural protein on an SDS/polyacrylamide gel (Fig. 1A, lanes 2, 3), and able to functionally replace the second component required for yeast CTD phosphatase activity (Fig. 1B, lanes 1, 2).

This novel yeast protein was similar in amino acid sequence for most of its length (32% identity and 54% similarity) to a human protein cloned on the basis of its interaction with human TFIIF (J.A., H. Xiao, G. Pan, G. Dahmus, M.C., S. Zhang, R. G. Roeder, M. Dahmus, and J.G., unpublished work). The yeast protein also associates with TFIIF as we describe below. Thus, this new yeast protein has been named Fcp1 for TFIIF-associating component of CTD phosphatase. This phosphatase has a type 2C activity that requires divalent cations and is resistant to okadaic acid and vanadate (19). Yet, the amino acid sequences of yeast and human Fcp1 (Fig. 2A) do not resemble any known phosphatases in or outside the PP2C family. Fcp1 also is unrelated to another phosphatase, Sit4, implicated in transcription control (31). Therefore, either Fcp1 is itself the phosphatase or it activates a CTD phosphatase present in the other fraction needed for activity. Curiously, the Fcp1 protein contains a BRCT motif found in many proteins involved in DNA repair (32) (Fig. 2B).

Figure 2.

Sequence and structure of yeast Fcp1. (A) Comparison of yeast and human Fcp1. Amino acid sequences of yeast Fcp1 (top line) and human FCP1a (J.A., H. Xiao, G. Pan, G. Dahmus, M.C., S. Zhang, R. G. Roeder, M. Dahmus, and J.G., unpublished work) (bottom line) were aligned using the sequence alignment program bestfit (Wisconsin sequence analysis package, GCG). Identical amino acids are indicated by a vertical line, highly similar and similar amino acids by a colon and a dot, respectively. (B) Regions of highest similarity and approximate locations of functional motifs in yeast and human Fcp1. Shaded areas are those most similar between yeast and human proteins. “FCP1 homology” refers to the region of similarity presented in C. (C) Multiple alignment of amino acid sequences homologous to yeast Fcp1 residues 173–362 identified by blast searches of GenBank. Bold letters: residues identical in all sequences. Shaded letters: regions of similarity. 1. S. cerevisiae FCP1, Z49704; 2. S. cerevisiae ORF, X90564; 3. S. cerevisiae ORF, U39205; 4. S. cerevisiae ORF, U10555; 5. S. cerevisiae ORF, Z73115; 6. Schizosaccharomyces pombe ORF, Z50142; 7. Brassica campestris EST, L46538; 8. Arabidopsis thaliana EST, T44887; 9. Brugia malayi EST, H48204; 10. Toxoplasma gondii EST, W35521; 11. Caenorhabditis elegans EST, U29536; 12. Mus musculus EST, W29399; 13. Homo sapiens EST, H24417; 14. Homo sapiens EST, H66914; 15. Homo sapiens EST, H95720; 16. Homo sapiens EST, H84869; 17. Homo sapiens FCP1.

Three regions of yeast and human Fcp1 exhibit the highest degrees of similarity: the N-terminal region (36% identity) corresponds to amino acids 124–352 of yeast Fcp1, the middle region (32% identity) to amino acids 425–593 of yeast Fcp1, and the C-terminal region (25% identity) to amino acids 610–728 of yeast Fcp1 (see Fig. 2B). The separation of these regions by long sequences missing in either the yeast or human protein suggests that each may be a functional domain. The C-terminal conserved region contains at least one binding site for TFIIF, and the central conserved region contains all or part of a second TFIIF-binding site (see below). The N-terminal conserved region has no known function, but most of it contains short sequence motifs related to sequences in 15 other ORFs of unknown function in nine species (see Fig. 2C).

Fcp1 was localized in cells of S. cerevisiae expressing an HA-tagged FCP1 gene from a low copy number plasmid. Immunofluorescence with an anti-HA mAb revealed that HA-Fcp1 (Upper Left, Fig. 3) colocalizes with DNA (4′,6-diamidino-2-phenylindole staining in Lower Left, Fig. 3) and is, therefore, a yeast nuclear protein.

Figure 3.

Fcp1 is localized to the nucleus. Cells from asynchronous cultures of strain JA830 (fcp1 ΔLEU2 pHA∷FCP1), which contains pJA794 producing HA-Fcp1 from the FCP1 promoter, and W303 (no HA tag) were fixed and stained with anti-HA mAb. Diffuse background staining was detected in non-HA containing W303 (“no tag”); this frame was overexposed intentionally to detect background fluorescence. HA-Fcp1-specific staining colocalized with the nucleus as revealed by 4′,6-diamidino-2-phenylindole staining (HA∷FCP1 panels). Both the 4′,6-diamidino-2-phenylindole and anti-HA images were superimposed onto Nomarski images of the stained cells.

Yeast Fcp1 Interacts with Tfg1, the Yeast Homologue of RAP74.

Yeast TFIIF stimulates activity of the yeast phosphatase (19), and human TFIIF stimulates activity of the human phosphatase (18). To test whether yeast Fcp1 and human FCP1a were likely to have similar functions, the binding of yeast Fcp1 to yeast TFIIF subunit Tfg1 (33, yeast RAP74) was examined (Fig. 4A). ³⁵S-labeled proteins containing all or various portions of yeast Fcp1 were made by translation in vitro and applied to microaffinity columns containing the C-terminal amino acids 649–735 of yeast RAP74 (33) fused to GST.

Figure 4.

Binding between Fcp1 and RAP74. (A) Various ³⁵S-labeled portions of yeast Fcp1 and human FCP1a, as indicated, were synthesized in rabbit reticulocyte lysates (input), applied to GST and GST-yRAP74 (649–735) microaffinity columns, and processed as described in Materials and Methods. Volumes corresponding to 5% of the input and 25% of each eluate were loaded on the gel. Only the region of each autoradiogram containing full-length in vitro synthesis products is shown. (B) Various ³⁵S-labeled portions of yeast Fcp1 and human FCP1a, as indicated, were chromatographed on GST and GST-hRAP74 (436–517) microcolumns. The columns were run and the eluates analyzed as described for A.

Indeed, yeast Fcp1 bound to the C-terminal region of yeast RAP74 and was eluted partly with salt and partly with SDS (Fig. 4A, line 1). Moreover, the C-terminal amino acids 667–732 of Fcp1, synthesized in vitro as a fusion protein with TR, were sufficient for this interaction (Fig. 4A, compare lines 5 and 7). However, amino acids 1–666 of yeast Fcp1 still bound RAP74, while amino acids 1–457 could not (Fig. 4A, lines 2, 3), implying there is a second RAP74-binding site in amino acids 458–666 of yeast Fcp1. When this region of Fcp1 was produced in vitro as a fusion protein with TR, it was also sufficient to bind RAP74 (Fig. 4A, line 6). Therefore, yeast Fcp1 has two RAP74-binding regions (Fig. 2B). Binding between this central region and RAP74 was somewhat weaker and more salt-sensitive than the interactions involving the C-terminal region of Fcp1 (Fig. 4 and data not shown).

The ability of TFIIF to stimulate CTD phosphatase is species specific (19). Yet, yeast Fcp1 can bind to the C-terminal 83 amino acids of human RAP74 (Fig. 4B, line 1) and this interaction depends on the RAP74-binding region of yeast Fcp1 (Fig. 4B, line 2). Human FCP1a also can bind to the C-terminal 87 amino acids of yeast RAP74 (Fig. 4A, line 8). Thus, the Fcp1-RAP74 interaction likely involves amino acids conserved between the yeast and human proteins.

The FCP1 Gene Is Essential for the Growth of S. cerevisiae.

RPII with a dephosphorylated CTD initiates transcription efficiently at many promoters in vitro (2, 3). Thus, dephosphorylation of the CTD prior to or following termination of transcription could be critical for gene expression in vivo. To test whether FCP1 is essential, the gene was disrupted by insertion of LEU2 into one of the FCP1 loci. Disruption was confirmed by PCR (data not shown). After meiosis and sporulation, dissection of asci revealed the 2:2 segregation pattern of viable and nonviable spores characteristic of an essential gene (Fig. 5A). Nonviable spores gave rise to microcolonies of 32–64 cells, indicating that there may be an excess of Fcp1 protein in a wild-type cell. As expected, the viable spores were leu2⁻ (data not shown). Viable LEU2⁺ spores were recovered when the disrupted diploid strain was transformed with a plasmid expressing FCP1 under the control of the GAL10 promoter, and grown in medium containing galactose (Fig. 5B). Therefore, FCP1 is an essential gene, and dephosphorylation of the RPII CTD is likely an essential function in S. cerevisiae. However, since FCP1 is only one component of CTD phosphatase, it remains formally possible that its essential nature is due to its regulating phosphorylation of an additional substrate in vivo.

Figure 5.

Yeast FCP1 is essential. (A) Diploid strain JA830 disrupted for one copy of FCP1 was induced to sporulate, and all spores of each tetrad were tested for growth. (B) Diploid JA830 was transformed with pJA754 to produce HA-tagged Fcp1 from the GAL10 promoter. Sporulation was induced, and spores were tested for growth on medium containing galactose. (C) Deletion constructs of Fcp1 were tested for complementation of a strain disrupted for FCP1 in vivo and for phosphatase activity in vitro. Amino acids in each construct and locations of RAP74 interacting regions are indicated. Low copy refers to cells heterozygous for the FCP1 disruption transformed with plasmids containing FCP1 under the control of the endogenous promoter (pJA794, pDB3, pJA796, pDB4, and pJA804); cells were sporulated and growth of haploids was tested on selective media. High copy refers to cells heterozygous for the FCP1 disruption transformed with plasmids containing FCP1 under control of the GAL10 promoter (pJA739, pJA756, and pJA798); cells were sporulated and growth of haploids was tested on medium containing galactose. (+) refers to complementation for viability. The in vitro reactions (19) were carried out with Fcp1 proteins overproduced in E. coli and purified as described in Experimental Procedures; (+) indicates that phosphatase activity was detected. nd, not done.

Truncations of Fcp1 That Lack a RAP74 Binding Region Are Still Functional.

When FCP1 was expressed from its own promoter on a low copy number plasmid, viability was rescued with full-length FCP1 and with deletions [FCP1 (1–666) and FCP1 (1–626)] lacking the C-terminal RAP74-binding site (Fig. 5C). However, a deletion lacking both RAP74-binding sites [FCP1 (1–457)] cannot rescue viability even when expressed at high copy (Fig. 5C). Both FCP1 (1–666) and FCP1 (1–626) also function in vitro to dephosphorylate the CTD (summarized in Fig. 5C). Because the proteins for the in vitro assays of recombinant Fcp1 are recovered following denaturation and renaturation, accurate, comparative activity determinations are not possible. However, cells harboring FCP1 (1–626) instead of wild-type FCP1, while viable, are cold sensitive (data not shown). Thus, this truncation likely functions less well in vivo than does wild-type Fcp1. The protein encoded by FCP1 (1–457) is insoluble and thus has not been tested in vitro for activity.

DISCUSSION

A yeast nuclear protein, Fcp1, has been identified that is necessary for CTD phosphatase activity. The importance of this phosphatase in vivo is highlighted by finding that FCP1 is an essential gene. It is not yet known whether Fcp1 is a phosphatase enzyme or an essential regulatory component of the enzyme. Fcp1 has no known phosphatase motifs, but phosphatases are a diverse group of proteins that have few primary sequence motifs in common (34). Therefore, the short, conserved amino acid sequences between residues 173–357 of yeast Fcp1 (see Fig. 2) may be the signature of a new family of protein phosphatases as similar sequences are found in human Fcp1 and in 15 other predicted proteins of unknown function from many organisms. Alternatively, these short, conserved motifs may define a domain used to interact with a target site in a particular phosphatase or family of phosphatases that remains to be identified. If that were the case, then Fcp1 could be a signaling molecule that couples the phosphatase enzyme and the RPII/TFIIF complex.

Yeast Fcp1 binds to part of the C-terminal evolutionarily conserved domain of RAP74. An adjacent portion of this domain binds RPII (35); perhaps RAP74 stabilizes the position of Fcp1 relative to the polymerase so that its associated phosphatase can dephosphorylate the CTD. The phosphatase activity is processive in removing phosphate groups from the CTD (18, 19), but TFIIF itself is not necessary for this processivity, at least with the human phosphatase (18). Thus, the stimulation of phosphatase activity seen in the presence of TFIIF likely reflects a stabilized binding interaction between the phosphatase and the polymerase.

Interaction between RAP74 and Fcp1 likely involves amino acids conserved in the yeast and human proteins, because yeast Fcp1 can bind human RAP74 and human FCP1a can bind yeast RAP74. Nevertheless, the requirements for enzyme activity are more complex than the simple binding assays would indicate, because neither the yeast nor the human phosphatase can function with RAP74 or RPII derived from the heterologous species (19). Moreover, S. cerevisiae does not survive when yeast FCP1 is replaced with a cDNA encoding human FCP1a (J.A., unpublished data).

The RPII recruited to many promoters has a dephosphorylated largest subunit (2, 3, 7), and thus an RPII complex specialized for initiation (36–39) might be expected to contain a CTD phosphatase. There is ample evidence that CTD phosphorylation is an important regulatory modification during the transition from preinitiation to efficient elongation (reviewed in refs. 4 and 40). Many transcription activators, including HIV-1 Tat, herpes simplex virus VP16, and human p53 and E2F1, can bind TFIIH (8, 9, 41) whose CDK7 subunit is a CTD kinase for RPII (reviewed in refs. 42 and 43). Moreover, the HIV-1 Tat protein, which stimulates elongation by RPII (44, 45), also stimulates this kinase activity (11, 46, 47), and phosphorylation of the CTD by TFIIH may be important for elongation by RPII in vivo (10, 40). A second CTD kinase, P-TEFb, also is required for HIV Tat transactivation (48, 49), and this kinase also seems essential for productive elongation (48, 50).

Further, the RAP74 subunit of mammalian TFIIF can be phosphorylated by TFIIH (51) or by the largest subunit of TFIID (52), and the phosphorylated form of TFIIF stimulates elongation by RPII in vitro (53). Perhaps the phosphorylation of RAP74 in the context of RPII holoenzyme leads to the dissociation of Fcp1, reducing or eliminating CTD phosphatase activity during elongation. Alternatively, the phosphorylation of RAP74 may simply inhibit the CTD phosphatase activity. As well, TFIIB is an inhibitor of the CTD phosphatase activity (18) and may prevent dephosphorylation of the CTD during the transition from the initiation to the elongation process. Activators like Tat, which stimulate elongation by RPII, may do so by both stimulating a CTD kinase (11, 46, 47) and inhibiting the CTD phosphatase, maintaining a highly phosphorylated form of the polymerase in the elongation complex that emerges from the promoter. The sensitivity of the polymerase in elongation complexes to the action of the phosphatase is yet to be tested, but regulating the phosphorylation state of the polymerase may be necessary for efficient elongation all along a transcription unit.

ABBREVIATIONS

RPII

RNA polymerase II

CTD

C-terminal domain of repeated heptapeptide sequence in the largest subunit of RNA polymerase II

TFIIF

transcription factor IIF

TR

thioredoxin

HA

hemagglutinin

GST

glutathione S-transferase