Cyclin-dependent Kinase Inhibitor p16INK4A Inhibits Phosphorylation of RNA Polymerase II by General Transcription Factor TFIIH*

The cell cycle is regulated by various protein kinases, including cyclin-dependent kinases (CDKs). D-type CDKs, CDK4, and CDK6, phosphorylate retinoblastoma protein and are believed to regulate through the G1 phase of the cell cycle. CDK inhibitor p16INK4A has been characterized as binding CDK4 and CDK6 and as inhibiting phosphorylation of retinoblastoma protein by these CDKs. Thus p16INK4A is implicated in regulating the cell cycle at the G1 phase. The largest subunit of RNA polymerase II (pol II) contains an essential C-terminal domain (CTD). General transcription factor TFIIH, which contains CDK7, phosphorylates the CTD in vitro. The CTD phosphorylation is shown to be involved in transcriptional regulationin vivo and in vitro. Phosphorylation of RNA pol II CTD by TFIIH is thought to play an important role in transcriptional regulation. Here we report that p16INK4Aassociates with RNA pol II CTD and TFIIH. p16INK4Ainhibited the CTD phosphorylation by TFIIH. These findings suggest that p16INK4A may regulate transcription via CTD phosphorylation in the cell cycle.

The cell cycle is regulated by various protein kinases, including cyclin-dependent kinases (CDKs). D-type CDKs, CDK4, and CDK6, phosphorylate retinoblastoma protein and are believed to regulate through the G 1 phase of the cell cycle. CDK inhibitor p16 INK4A has been characterized as binding CDK4 and CDK6 and as inhibiting phosphorylation of retinoblastoma protein by these CDKs. Thus p16 INK4A is implicated in regulating the cell cycle at the G 1 phase. The largest subunit of RNA polymerase II (pol II) contains an essential C-terminal domain (CTD). General transcription factor TFIIH, which contains CDK7, phosphorylates the CTD in vitro. The CTD phosphorylation is shown to be involved in transcriptional regulation in vivo and in vitro. Phosphorylation of RNA pol II CTD by TFIIH is thought to play an important role in transcriptional regulation. Here we report that p16 INK4A associates with RNA pol II CTD and TFIIH. p16 INK4A inhibited the CTD phosphorylation by TFIIH. These findings suggest that p16 INK4A may regulate transcription via CTD phosphorylation in the cell cycle.
Transcription by RNA polymerase II (pol II) 1 requires general transcription factors (1,2). These essential factors include TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH and form a preinitiation complex on a promoter DNA. TFIIH is composed of multiple subunits and associates with multiple enzyme activities, DNA-dependent ATPase, DNA helicase, and protein kinase (Refs. 3-7 and references therein). It has been suggested that DNA helicase by TFIIH is involved in promoter opening surrounding the transcriptional start site. CDK7 and cyclin H have been identified as subunits of TFIIH, and TFIIH phosphorylates the CTD of the largest subunit of RNA pol II in vitro. Genetic studies in yeast suggest that CTD phosphoryla-tion by TFIIH is essential for transcription in vivo (8 -11). It has been proposed that phosphorylation of the CTD by TFIIH regulates transcription.
The cell cycle is believed to be regulated by cyclin-dependent kinases (CDKs) (Refs. 12-14 and references therein). The activities of CDKs are precisely regulated through the cell cycle. Several inhibitory factors of CDKs are also known. Cdk inhibitor p16 INK4A has been identified as binding Cdk4 and inhibiting phosphorylation of retinoblastoma protein (pRb) (15). CDK4, a G 1 CDK, phosphorylates pRb and is important for the cell cycle progression at the G 1 phase in the cell cycle. Therefore p16 INK4A is thought to repress the cell cycle by inhibiting G 1 CDKs. Here, we report that p16 INK4A associates with RNA pol II CTD and TFIIH and that p16 INK4A inhibits the CTD phosphorylation by TFIIH. These data suggest that p16 INK4A might directly regulate transcription through the CTD phosphorylation.

EXPERIMENTAL PROCEDURES
Materials-Antibodies against the TFIIH p89 subunit and p16 INK4A , CDK2, and CDK4 were obtained from Santa Cruz Biotech. The TNT reticulocite lysate kit was obtained from Promega.
Assay of Protein Kinase-Except as indicated in the figure legend, protein kinase reactions were carried out using TFIIH and RNA pol II purified from calf thymus, as described by Serizawa et al. (3). Purification of calf thymus TFIIH and RNA pol II will be described elsewhere. Reaction mixtures (15 l) were incubated at 28°C for 60 min, and the phosphorylation reactions were stopped by the addition of an equal volume of SDS sample buffer containing 100 mM Tris-HCl (pH 6.8), 200 mM DTT, 4% SDS, 0.2% bromphenol blue, and 20% glycerol. Phosphoproteins were analyzed by SDS-PAGE.
Preparation of Insect Cell Lysates-Sf21 cells were infected by recombinant baculoviruses containing cDNA sequences of p16 INK4A , CDK4, and CDK2. The infected cells were harvested after 48 h, were washed with phosphate-buffered saline, and were burst in homogenization buffer using N 2 cavitation as described by Graber et al. (16). Cell lysates were centrifuged to remove debris.
Purification of His-p16 and GST Fusion Proteins from Escherichia coli Lysates-GST fusion proteins were purified using a glutathione-Sepharose column as described by Serizawa et al. (3,17). His-p16 was prepared from E. coli lysates under denaturing conditions and purified using nickel-nitrilotriacetic acid column (Qiagen) as described in Koh et al. (18). The purified His-p16 was dialyzed against buffer C (17) containing 50 mM KCl and centrifuged at 30,000 rpm for 30 min. The resulting supernatant was loaded onto DEAE-5PW (7.5 ϫ 75 mm) (TosoHaas) pre-equilibrated with buffer C containing 25 mM KCl. The column was washed using buffer C (17) containing 100 mM KCl and eluted at 1 ml/min using a 33-ml linear gradient from 100 to 500 mM KCl in buffer C (17). 1-ml fractions were collected.
Coprecipitation Experiments Using Recombinant GST Fusion Proteins and Glutathione-Sepharose Beads-Except as indicated in the figure legend, reaction mixtures were incubated at 27°C for 30 min under conditions similar to those of the CTD kinase assay in the absence of ATP and NdeI-digested pDN-AdML. The reactions were subsequently incubated with glutathione-Sepharose beads at 4°C for 60 min and were spun down in a centrifuge. The supernatant was removed. The precipitate was extensively washed using ELB (19) containing 1 mM DTT and 0.5 mg/ml BSA, and proteins in the precipitate were eluted using 1 mM glutathione. The precipitates and the supernatant were subjected to SDS-PAGE, and proteins were detected on Western blots or autoradiograms. Bands on Western blots were visualized by chemiluminescence using horseradish peroxidase-conjugated secondary antibody (Amersham).

RESULTS AND DISCUSSION
TFIIH has been reported to contain CDK7 and cyclin H and to phosphorylate RNA pol II CTD and CDK2 (3-6). In the present study, we investigated whether p16 INK4A inhibits phos-phorylation of RNA pol II CTD by TFIIH. A recombinant fusion protein of glutathione S-transferase and p16 INK4A (GST-p16) was prepared from E. coli lysates and purified using a glutathione affinity column. Phosphorylation of RNA pol II CTD was inhibited by the addition of purified GST-p16 (Fig. 1a), and phosphorylation of a GST-CTD fusion protein by TFIIH was also inhibited by the addition of the purified GST-p16 (Fig. 1b). GST was purified using a glutathione affinity column from E. coli lysates. Phosphorylation of RNA pol II CTD and GST-CTD was not inhibited by the addition of amounts of the purified GST equivalent to that of GST-p16. We tested whether purified GST-p16 inhibits phosphorylation of CDK2 by TFIIH. Fig. 1c shows that the purified GST-p16 did not inhibit phosphorylation of CDK2 by TFIIH. The protein amount of the purified GST-p16 used in the phosphorylation of GST-CDK2 is equivalent to that in the phosphorylation of GST-CTD and RNA pol II, and phosphorylation of GST-CDK2 was carried out under conditions equivalent to those of GST-CTD. GST-p16 used here inhibited phosphorylation of retinoblastoma protein by CDK4 (data not shown). These results indicate that purified GST-p16 is responsible for the CTD kinase inhibitory activity and furthermore that purified GST-p16 did not inhibit phosphorylation of CDK2 by TFIIH.
To further confirm the CTD kinase inhibitory activity by p16 INK4A , a recombinant p16 INK4A protein containing six histidine residues at the N-terminal (His-p16) was prepared from E. coli lysates and was purified using a nickel affinity column. The purified His-p16 was further purified using an anion exchange HPLC column. A protein in fractions of the HPLC column was analyzed by SDS-PAGE. Phosphorylation of GST-CTD was carried out in the presence of these column fractions. The SDS-PAGE and the CTD kinase assay revealed that inhibition of the CTD phosphorylation was cofractionated with His-p16 (Fig. 2), thus confirming that recombinant p16 INK4A inhibits the CTD kinase activity of TFIIH.
To determine the molar ratio of TFIIH, GST-CTD, and His-p16 in effective inhibition of CTD phosphorylation, ϳ0.01 pmol of TFIIH highly purified from rat liver, ϳ40 pmol of purified GST-CTD, and various amounts of purified His-p16 were added in CTD phosphorylation reactions (Fig. 3). Highly purified rat TFIIH, shown in Fig. 3a, was used in the CTD phosphorylation reactions. The CTD kinase inhibitory activity was saturated under the condition of ϳ20 pmol of His-p16. Therefore, the molar ratio of His-p16 and GST-CTD in effective inhibition of CTD phosphorylation is about 1:2, and that of His-p16 and TFIIH is about 2,000:1.
To determine whether p16 INK4A associates with TFIIH, purified GST-p16 was incubated with a TFIIH preparation. GST-p16 was precipitated using glutathione-Sepharose beads, and the FIG. 1. GST-p16 inhibits phosphorylation of RNA pol II CTD by TFIIH but not CDK2. a, phosphorylation of RNA pol II was carried out using RNA pol II (0.001 units) and TFIIH purified from calf thymus. GST-p16 (5 and 10 g) and GST (5 and 10 g) were added in the reactions of lanes 4, 5, 7, and 8, respectively. The arrowhead indicates the largest subunit of the RNA pol II phosphorylated by TFIIH. b, phosphorylation of GST-CTD was carried out using 10 g of GST-CTD and calf thymus TFIIH. GST-p16 (5 and 10 g) and GST (5 and 10 g) were added in the reactions of lanes 2, 3, 4, and 5, respectively. The arrowhead indicates GST-CTD phosphorylated by TFIIH. c, phosphorylation of GST-CDK2 was carried out using 10 g of CDK2 and calf thymus TFIIH. 5 and 10 g of GST-p16 were added in the reactions of lanes 2 and 3, respectively. The arrowhead indicates GST-CDK2 phosphorylated by TFIIH. Portions of this and subsequent figures were prepared from digital replicas of primary data scanned using LaCie Silverscanner IV. precipitates were analyzed on Western blots using an antibody against the p89 subunit (ERCC3) of TFIIH (Fig. 4a). The p89 subunit was detected in the precipitates on the Western blots (lane 1). In control experiments, the p89 subunit was not detected in the precipitates on the Western blots, when the purified GST was incubated with the TFIIH preparation and precipitated using glutathione-Sepharose beads (lane 2). These results suggest that TFIIH associates and coprecipitates with GST-p16.
Phosphorylation of the CTD by TFIIH was inhibited by the addition of the purified GST-p16, whereas phosphorylation of the CDK2 was not inhibited (Figs. 1 and 2). p16 INK4A has been reported not to bind CDK2 (Fig. 5b) (15). The possibility that p16 INK4A might associate with the RNA pol II, a protein substrate of the CTD kinase of TFIIH, was next examined. The purified GST-p16 was incubated with a fraction containing purified RNA pol II, and GST-p16 was precipitated using glutathione-Sepharose beads. The largest subunit of RNA pol II was detected in the precipitates on Western blots using a monoclonal antibody against RNA pol II CTD (Fig. 5a, lane 1).
In control experiments, the largest subunit of RNA pol II was not detected in the precipitates on Western blots using the monoclonal CTD antibody, when the purified GST fraction was incubated with the purified RNA pol II fraction and precipitated using glutathione-Sepharose beads (lane 2). These results suggest that RNA pol II associates and coprecipitates with GST-p16.
To determine whether p16 INK4A associates with the CTD, the purified GST-CTD fraction was incubated with insect cell lysates The reactions were subsequently incubated with glutathione-Sepharose beads and were spun down in a centrifuge. The supernatants (sup) and precipitates (ppt) were prepared and analyzed as described under "Experimental Procedures." The arrowhead indicates the largest subunit of RNA pol II detected on Western blots using a monoclonal antibody against CTD, 8WG16. b, immunocoprecipitation of GST-CTD using His-p16. Reaction mixtures (30 l) were incubated at 27°C for 30 min under conditions similar to those of the CTD kinase assay in the absence of ATP and NdeI-digested pDN-AdML and contained 50 g of GST-CTD and insect cell lysates containing His-p16, CDK2, and CDK4 overexpressed by the baculovirus system. An antibody against p16 INK4A and an excess amount of a synthetic peptide used to generate the anti-p16 INK4A antibody were added in the reactions as indicated in the figure, and the reactions were subsequently incubated at room temperature for 30 min. Protein A-Sepharose beads were added. The reactions were incubated at room temperature for 30 min and were spun down in a centrifuge. The supernatant was removed. The precipitate was extensively washed using ELB (19) containing 1 mM DTT and 0.5 mg/ml BSA, and proteins in the precipitate were eluted using 100 mM glycine-HCl (pH 3.0). The precipitates (ppt) and the supernatant (sup) were subjected to SDA-PAGE. GST-CTD, CDK2, and CKK4 were detected on Western blots using antibodies against these molecules by chemiluminescence using horseradish peroxidase-conjugated secondary antibody (Amersham). The arrowheads indicate these molecules. c, coprecipitation of His-p16 using GST-CTD. Reaction mixtures (30 l) contained insect cell lysates containing p16 INK4A overexpressed by the baculovirus system (lanes 1-6) or purified His-p16 (Fraction No. 10 in Fig. 2) (lanes 7-12) and were incubated in the presence of 50 g of GST-CTD or GST. The reactions were subsequently incubated with glutathione-Sepharose beads and were spun down in a centrifuge. The supernatants (sup) and precipitates (ppt) were prepared and analyzed as described under "Experimental Procedures." The arrowhead indicates p16 INK4A detected on Western blots using anti-p16 INK4A antibody.

p16 INK4A Inhibits the CTD Kinase of TFIIH
containing His-p16 overexpressed by the baculovirus expression system. The His-p16 was precipitated using an anti-p16 INK4A antibody and protein A-Sepharose beads (Fig. 5b). GST-CTD was detected in the precipitates on Western blots using a monoclonal antibody against RNA pol II CTD (lane 1). In control experiments, when an excess of a synthetic peptide that was used to generate the anti-p16 INK4A antibody was incubated in the precipitation reactions (lane 2), the amount of GST-CTD detected in the precipitates on the Western blots was traceable but was much smaller than that in lane 1. These data suggest that GST-CTD associates and coprecipitates with His-p16.
To confirm that GST-CTD and His-p16 are coprecipitated together, glutathione-Sepharose beads were used to precipitate GST-CTD, when GST-CTD was incubated with purified His-p16 or the insect cell lysates containing His-p16 overexpressed by the baculovirus system (Fig. 5c). His-p16 was detected in the precipitates on Western blots using an anti-p16 INK4A antibody (lanes 2 and 8). In control experiments, when GST was incubated with purified His-p16 or the insect cell lysates containing His-p16 overexpressed and precipitated using glutathione-Sepharose beads (lane 1 and 7), the amount of His-p16 detected in the precipitates on Western blots using the anti-p16 INK4A antibody was traceable but was much smaller than that in lanes 2 and 8, respectively. These data suggest that GST-CTD directly bound and coprecipitated with His-p16. p16 INK4A has been reported to bind and inhibit CDK4 and CDK6 and to regulate the cell cycle via the inhibition of these CDKs (15,18). CDK4 and CDK6 phosphorylate pRb, which forms a protein complex with DNA binding transcription factor E2F. The phosphorylation of pRb results in the release of E2F from the complex, which activates transcription by RNA pol II to express gene products that are required for the cell cycle (20). Thus, p16 INK4A is thought to regulate the cell cycle through the governance of a signaling pathway involved in the CDKs and RNA pol II transcription. RNA pol II CTD has been shown to be essential for cell viability and to regulate transcription in vivo (9 -11). Phosphorylation of the RNA pol II CTD by TFIIH is thought to be important for transcriptional regulation in vivo and in vitro (1,9,(21)(22)(23)(24). A hypophosphorylated form of RNA pol II preferentially forms a preinitiation complex, and the CTD of RNA pol II engaged in transcriptional elongation is highly phosphorylated. Thus phosphorylation of the RNA pol II CTD has been proposed to be a regulatory event in transcription initiation. The present study shows that p16 INK4A is capable of inhibiting phosphorylation of RNA pol II CTD by TFIIH, indicating an alternative possibility that p16 INK4A may directly regulate transcription via phosphorylation of RNA pol II CTD in the cell cycle.
The RNA pol II holoenzyme has been reported to be a ribosome-sized protein complex containing factors involved in regulating transcription, cell cycle, and DNA repair (25)(26)(27)(28). The precise subunit composition of the holoenzyme is still unclear. This work indicates that p16 INK4A is a CTD-binding protein. Therefore, it is possible that p16 INK4A may be a subunit of the holoenzyme. CDK8 and cyclin C appear to be subunits of the holoenzyme (28). Yeast homologues of these molecules are SRB10 and SRB11, and the CTD kinase by SRB10 was shown to be responsible for transcriptional activation (29). It is possible that p16 INK4A might also inhibit the CTD phosphorylation by CDK8, and furthermore, that p16 INK4A might inhibit transcriptional activation.
Mutations and deletions in the p16 INK4A gene have been identified from many types of tumor cells (Ref. 18 and references therein). Mutants of p16 INK4A derived from tumor cells have revealed defects in inhibiting CDK4 and in regulating cell cycle at the G 1 phase (18). These p16 INK4A mutants may also have defects in inhibiting the CTD kinase of TFIIH and in binding RNA pol II CTD and TFIIH, and these defects might be involved in tumorigenesis.
CDK7 and cyclin H have been shown to form a 175-kDa complex in vitro (19,30). This 175-kDa complex activates the histone H1 kinase activity of CDK2 by phosphorylation of the threonine activation site. The kinase capable of activating a CDK is called CDK-activating kinase, CAK. Most of the CAK activity in cell lysates fractionates in a 175-kDa complex (30). TFIIH has been reported to contain in vitro CAK activity and to be a complex of about 700 kDa (3). However, the function of the 175-kDa CAK complex and TFIIH in the cell cycle is not well understood. This work reveals that p16 INK4A binds the recombinant CDK7 and the purified TFIIH and that although p16 INK4A inhibits CTD kinase of TFIIH, it does not inhibit CDK2 phosphorylation by TFIIH. These results suggest that CAK complexes, including TFIIH and the 175-kDa complex, may play a role in the cell cycle via a signaling pathway regulated by p16 INK4A . However, p16 INK4A may not be involved in a pathway to activate CDK2 by CAK.