¹Department of Cell Signaling, DNAX Research Institute of Molecular and Cellular Biology, 901 California Avenue, Palo Alto, California 94304, USA

²Department of Chemistry, Stanford University, Stanford, California 94305-5080, USA

³Department of Molecular Biology and Genetics, Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, Maryland 21209, USA

^aAuthor for correspondence

^bCurrent address: Department of Cellular and Molecular Pharma-cology, University of California at San Francisco, 513 Parnassus Avenue, San Francisco, California 94143-0450, USA

Phosphorylation of the carboxyl-terminal domain (CTD) of RNA polymerase II is important for basal transcriptional processes in vivo and for cell viability. Several kinases, including certain cyclin-dependent kinases, can phosphorylate this substrate in vitro. It has been proposed that differential CTD phosphorylation by different kinases may regulate distinct transcriptional processes. We have found that two of these kinases, cyclin C/CDK8 and cyclin H/CDK7/p36, can specifically phosphorylate distinct residues in recombinant CTD substrates. This difference in specificity may be largely due to their varying ability to phosphorylate lysine-substituted heptapeptide repeats within the CTD, since they phosphorylate the same residue in CTD consensus heptapeptide repeats. Furthermore, this substrate specificity is reflected in vivo where cyclin C/CDK8 and cyclin H/CDK7/p36 can differentially phosphorylate an endogenous RNA polymerase II substrate. Several small-molecule kinase inhibitors have different specificities for these related kinases, indicating that these enzymes have diverse active-site conformations. These results suggest that cyclin C/CDK8 and cyclin H/CDK7/p36 are physically distinct enzymes that may have unique roles in transcriptional regulation mediated by their phosphorylation of specific sites on RNA polymerase II.

cyclin C; CDK8; CTD; RNA polymerase II

Introduction

The carboxyl-terminal domain (CTD) of the largest subunit of eukaryotic RNA polymerase II is a highly conserved, unusual structure unique to this enzyme (Allison et al., 1985; Corden et al., 1985). It consists of multiple, near-perfect tandem repeats of the sequence Tyr-Ser-Pro-Thr-Ser-Pro-Ser, and a minimum number of these CTD heptapeptide repeats is required for RNA polymerase II activity in vivo and for cell viability (Allison et al., 1988; Bartolomei et al., 1988; Nonet et al., 1987; Zehring et al., 1988). RNA polymerase II becomes heavily phosphorylated during the course of transcription, and this phosphorylation mostly occurs on the CTD (Cadena and Dahmus, 1987; Dahmus, 1981). Furthermore, mutational analysis has demonstrated that potential phosphorylation sites on the CTD are also essential in vivo (West and Corden, 1995).

CTD phosphorylation appears to play multiple roles in transcriptional regulation (Corden, 1993; Dahmus, 1996; Shilatifard et al., 1997; Steinmetz, 1997). Phosphorylation of the CTD is thought to be required for both promoter clearance and efficient elongation (Akoulitchev et al., 1995; Arias et al., 1991; Lee and Greenleaf, 1997; Marshall et al., 1996; O'Brien et al., 1994; Song, 1996; Yankulov et al., 1995). Various experiments have suggested that CTD phosphorylation may be important for modulation of gene-specific transcription (Arias et al., 1991; O'Brien et al., 1994; Song, 1996). Phosphorylation of the CTD in mitosis may be involved in general inhibition of transcription (Gebara et al., 1997). It has recently been demonstrated that the hyperphosphorylated CTD physically and functionally interacts with proteins required for splicing and polyadenylation, suggesting that it may also coordinate transcription with mRNA processing (Bregman et al., 1995; Du and Warren, 1997; Kim et al., 1997; McCracken et al., 1997; Mortillaro et al., 1996; Steinmetz, 1997; Yuryev et al., 1996).

Several kinases and phosphatases can modify the phosphorylation state of the CTD in vitro. One of the best-characterized CTD kinases is cyclin H/CDK7/p36, which, like its yeast homolog KIN28/CCL1, is a component of the general transcription factor TFIIH (Cismowski et al., 1995; Feaver et al., 1991, 1994; Lu et al., 1992; Roy et al., 1994; Serizawa et al., 1995; Shiekhattar et al., 1995). Cyclin C/CDK8 can also phosphorylate the CTD and co-purifies with a transcriptionally active RNA polymerase II holoenzyme (Maldonado et al., 1996; Rickert et al., 1996). The homologs of cyclin C and CDK8 in Saccharomyces cerevisiae, SRB10 and SRB11, genetically interact with the CTD and also co-purify with an RNA polymerase II holoenzyme (Liao et al., 1995). Other cyclin-dependent kinases are capable of phosphorylating the CTD in vitro, including cyclin B/CDC2, cyclin T/CDK9, cyclin K and the yeast kinase CTDK-1 (Cisek and Corden, 1989; Lee and Greenleaf, 1989; Marshall and Price, 1995; Zhu et al., 1997; Edwards et al., 1998; Wei et al., 1998; Peng et al., 1998). In addition, both MAP kinases (Bellier et al., 1997; Dubois et al., 1994b; Venetianer et al., 1995) and the tyrosine kinases c-Abl and Arg (Baskaran et al., 1997; 1993) can phosphorylate the CTD in vitro. CTD phosphatases have been purified from both human and yeast cells, and CTD dephosphorylation may also be important for transcriptional regulation (Chambers and Dahmus, 1994; Gruenwald and Heitz, 1993).

The CTD has been found to exist in multiple phosphorylation states (Dubois et al., 1997; Lee and Greenleaf, 1991; Payne and Dahmus, 1993; Rice et al., 1994), suggesting that different kinases may differentially phosphorylate the CTD and thus play different roles in regulating RNA polymerase II activity. Unfortunately, it is difficult to assess the functional significance of CTD phosphorylation by different kinases, as the phosphorylated CTD is not required for in vitro transcriptional assays using highly purified components (Jiang and Gralla, 1994; Mäkelä et al., 1995; Serizawa et al., 1993a). It will thus be necessary to use in vivo assays to determine the functional differences between the various CTD kinases. In this paper we have examined the difference in phosphorylation of specific sites on the CTD by the closely related kinases cyclin C/CDK8 and cyclin H/CDK7/p36, as well as by cyclin B/CDC2. We have also investigated the ability of small-molecule kinase inhibitors to block CTD phosphorylation by these kinases. Both of these analyses will provide important tools for determining the specific function of CTD phosphorylation by these different CTD kinases.

Results

Purified cyclin C/CDK8, cyclin H/CDK7/p36 and cyclin B/CDC2 enzymes differentially phosphorylate recombinant CTD substrates

Both cyclin C/CDK8 and cyclin H/CDK7/p36 have been implicated in transcriptional regulation. Although cyclin C/CDK8 is associated with a RNA polymerase II holoenzyme and cyclin H/CDK7/p36 with TFIIH, these kinases share a common substrate, the CTD of RNA polymerase II. To determine whether these kinases might have distinct functions, we were interested to see if they caused identical or different phosphorylation patterns on RNA polymerase II.

To determine if these kinases can inherently recognize different sites on the CTD, mapping studies were performed in vitro with purified recombinant enzymes and substrates. Three constructs based on the mouse CTD sequence were created and used to express and purify fusion proteins of glutathione S transferase (GST) with repeats 1 - 29 (GST - CTD1 - 29), repeats 30 - 53 (GST - CTD30 - 53), or repeats 1 - 53 (GST - CTD1 - 53). These recombinant proteins, as represented diagrammatically in Figure 1a, were then used as substrates in kinase assays with purified cyclin C/CDK8, cyclin H/CDK7/p36, and cyclin B/CDC2. All three kinases were able to phosphorylate GST - CTD1 - 53 to a similar extent. Figure 1b shows the level of phosphate incorporated by each substrate relative to that by GST - CTD1 - 53 in the presence of each kinase. Cyclin C/CDK8 phosphorylated all three substrates fairly equally. Cyclin H/CDK7/p36 phosphorylated GST - CTD1 - 53 and GST - CTD1 - 29 to a similar extent, but in contrast to cyclin C/CDK8 displayed a fourfold preference for GST - CTD30 - 53 as a substrate. Cyclin B/CDC2 also phosphorylated GST - CTD30 - 53 preferentially, and surprisingly was barely able to phosphorylate GST - CTD1 - 29 under these experimental conditions. These differences indicate that these kinases can differently phosphorylate different portions of the CTD sequence.

Both cyclin C/CDK8 and cyclin H/CDK7/p36 phosphorylate Ser-5 in the CTD consensus repeat

Our results suggest that cyclin C/CDK8 and cyclin H/CDK7/p36 have preferences for specific phosphorylation sites within the CTD. To analyse these differences further, these kinases were examined to determine their specificity for phosphorylation sites within the CTD consensus repeat sequence. Both of the CDK consensus phosphorylation sites in the CTD repeat, Ser-2 and Ser-5, have been demonstrated to be essential in yeast (West and Corden, 1995). Furthermore, genetic suppression experiments indicate a functional difference between these two serines in vivo (Yuryev and Corden, 1996).

To determine which specific residues in the consensus sequence are phosphorylated by cyclin C/CDK8, kinase assays were performed using recombinant CTD fusion proteins. These proteins are fusions of glutathione S-transferase (GST) with between 14 and 16 repeats of either the consensus sequence (WT = YSPTSPS), or with alanine substituted for serine at position two or five (A2 = YAPTSPS, A5 = YSPTAPS). The CTD kinases were immunoprecipitated from CEM cell lysate using specific antibodies to cyclin C (-C pep), cyclin H (-cyc H) and CDC2 (-CDC2) as indicated in Figure 2. An immunoprecipitation using normal rabbit serum (NR) was included as a negative control. Figure 2 shows that cyclin C/CDK8 could only phosphorylate the wild type and A2 fusion proteins, but not the A5 fusion protein, indicating that this kinase specifically targets the Ser-5 residue in CTD consensus repeats. As previously demonstrated, cyclin H/CDK7/p36 also phosphorylated only the serine at position 5 in the consensus repeat (Roy et al., 1994), whilst CDC2 was able to phosphorylate both Ser-2 and Ser-5 (Gebara et al., 1997). The same result was also seen using purified recombinant cyclin/CDK complexes (C/CDK8 and H/CDK7/p36) as shown in Figure 2, showing that this specificity is not altered by the presence of other proteins in the immune complexes.

Cyclin C/CDK8, cyclin H/CDK7/p36, and cyclin B/CDC2 have different specificities towards non-consensus CTD repeats

Although both cyclin C/CDK8 and cyclin H/CDK7/p36 phosphorylate the same residue on the CTD consensus sequence, it was possible that they could differentially recognize various non-consensus repeats. Analysis of the CTD amino acid sequence reveals that the non-consensus repeats (approximately 60% of the CTD) are clustered near the end of the primary CTD structure, and that most of the substitutions occur at position seven. This position is +2 from the residue specifically phosphorylated by both cyclin C/CDK8 and cyclin H/CDK7/p36, suggesting that this could be a major determinant of specificity for these kinases.

To examine the specificity of these CTD kinases for these non-consensus repeats, the three most common substitutions at this position were chosen: lysine, threonine, and asparagine. Peptides were synthesized with four repeats of the consensus heptapeptide, either unsubstituted or with K, T, or N substituted for serine at position seven: (YSPTSPX)₄, where X = S, K, T, or N (Research Genetics). These peptides were then tested as substrates for the CTD kinases cyclin C/CDK8, cyclin H/CDK7/p36, or cyclin B/CDC2. As can be seen in Figure 3a, cyclin C/CDK8 could phosphorylate the consensus (WT), N-substituted, K-substituted and T-substituted peptides all to a similar extent. In contrast, phosphorylation of the K-substituted peptide by cyclin H/CDK7/p36 was at least fourfold greater than the other peptides which were otherwise recognized at comparable levels. Cyclin B/CDC2 displayed the most dramatic substrate preference, phosphorylating the K-substituted peptide over 20-fold better than either the consensus (WT), N-substituted, or T-substituted peptides. These results are shown graphically in Figure 3b. The presence of this basic residue in position seven of these CTD repeats has a significant effect on the specificity of these kinases, strongly suggesting structural differences in the active sites of these enzymes. The selectivity of other kinases for optimal peptide substrates has been correlated with the known structures of their catalytic sites (Songyang et al., 1994).

Substrate length dependence of CTD phosphorylation by cyclin C/CDK8 and cyclin H/CDK7/p36

The difference in phosphorylation of recombinant CTD substrates by these kinases may result from recognition of the primary sequence of the CTD or of a folded conformation. To examine how the secondary structure of the CTD affects its ability to be phosphorylated by these kinases, CTD substrates of varying lengths were analysed. If the kinase recognizes each heptapeptide repeat equally, it should phosphorylate these substrates in proportion to the number of repeats present. If, however, the kinase recognizes some secondary structure present with multiple repeats (Cagas and Corden, 1995) then it may differentially phosphorylate longer and shorter substrates.

To analyse the length dependence of CTD phosphorylation, DNA constructs were created for expression and purification of GST-fusion proteins with 4, 6, 8 or 10 repeats of the CTD consensus sequence: GST-(YSPTSPS)_4-10. These recombinant proteins were then assayed as substrates for purified cyclin C/CDK8 or cyclin H/CDK7/p36. Figure 4a shows that both cyclin C/CDK8 and cyclin H/CDK7/p36 cause greater incorporation of phosphate onto the longer substrates. However, cyclin C/CDK8 routinely phosphorylated the shorter substrates significantly more efficiently than did cyclin H/CDK7/p36. The quantification of these results is shown in Figure 4b. If these kinases phosphorylated each repeat on these substrates equally, with twice as many phosphates on GST-(YSPTSPS)₈ as on GST-(YSPTSPS)₄, a plot of kinase activity versus number of repeats on this graph would be expected to have a slope of 2. The plot of cyclin C/CDK8 CTD phosphorylation has a slope of 0.99, while the plot of cyclin H/CDK7/p36 phosphorylation has a slope of 2.7. Similar results are obtained with cyclin C/CDK8 and cyclin H/CDK7/p36 immunoprecipitated from cell lysates (data not shown). Cyclin C/CDK8 thus preferentially phosphorylates shorter CTD-like substrates, and cyclin H/CDK7/p36 preferentially phosphorylates longer CTD-like substrates in vitro. Phosphorylation of these substrates by cyclin B/CDC2 indicates that substrate recognition by this kinase is not significantly length-dependent (data not shown). These results suggest that cyclin C/CDK8 and cyclin H/CDK7/p36 may differentially recognize secondary structure in the CTD.

Cyclin C/CDK8 and cyclin H/CDK7/p36 differentially phosphorylate an endogenous RNA polymerase II substrate

Our results suggest that there is a specificity difference between cyclin C/CDK8 and cyclin H/CDK7/p36 for different recombinant CTD substrates. To determine whether this difference is also observed in vivo, the endogenous RNA polymerase II substrate phosphorylated by cyclin C/CDK8 or cyclin H/CDK7/p36 immune complexes was compared by partial proteolytic mapping. CEM cell lysates were immunoprecipitated with -cyclin C, -cyclin H, or normal rabbit antisera. The immunoprecipitations were then incubated with [-³²P]ATP to allow phosphorylation of associated polypeptides. The kinase reactions were denatured, then reimmunoprecipitated with either an antibody to RNA polymerase II or an isotype control antibody. These reactions were electrophoresed and autoradiographed to visualize the phosphorylated RNA polymerase II band (Figure 5a, lanes 4 and 6), as previously identified in (Rickert et al., 1996). To proteolytically fingerprint these proteins, similar RNA polymerase II bands from -cyclin C, -CDK8, -cyclin H, or -CDK7 immunoprecipitations were excised from SDS-polyacrylamide gels and digested with various dilutions of Staphylococcus aureus V8 protease (Cleveland et al., 1977) as shown in Figure 5b.

Figure 5b shows that the pattern of digested ³²P-labeled fragments obtained from the phosphorylated endogenous RNA polymerase II substrate from cyclin C/CDK8 or cyclin H/CDK7/p36 immunoprecipitates is different. Cyclin H/CDK7/p36-associated RNA polymerase II has two phosphorylated digestion products approximately 50 - 60 kD which are absent from the cyclin C/CDK8-associated RNA polymerase II. In addition, cyclin C/CDK8-associated RNA polymerase II gives rise to more small labeled digestion products (6 - 15 kD) than are seen with cyclin H/CDK7/p36-associated RNA polymerase II. These differences in phosphorylated digestion patterns indicate that RNA polymerase II assumes a different phosphorylation state in the presence of each kinase. As an internal control, the digestions were performed with phosphorylated RNA polymerase II from cyclin and CDK immunoprecipitates separately. The patterns from each cyclin match those from their respective kinase partner.

The different patterns observed are consistent with the inherent specificity of these kinases for certain phosphorylation sites on the CTD as shown by previous experiments. However we can not rule out the influence of other factors in vivo, such as the existence of pre-existing phosphates on RNA polymerase II, different accessibilities of these kinases for this substrate, or maybe even the presence of other proteins (e.g. kinases) present in cyclin C/CDK8 and cyclin H/CDK7/p36 immune complexes.

Differential inhibition of cyclin C/CDK8 and cyclin H/CDK7/p36 by ATP analogs

To study the functional consequences of differential CTD phosphorylation by the different CTD kinases we evaluated a number of chemical kinase inhibitors in search of a selective compound that could specifically inhibit one enzyme but not the other. The compounds H7 (1-(5-isoquinolinoylsulfonyl)-3-methylpiperazine), H8 (N-[2-(methylamino)-ethyl]-5-isoquinoline sulfonamide), and DRB (5,6-dichloro-1--M-ribofuranosylbenzimidazole) can inhibit transcriptional processes, possibly through their ability to inhibit CTD phosphorylation (Dubois et al., 1994a; Sehgal et al., 1976; Serizawa et al., 1993a). These ATP analogs have been described to inhibit various CTD kinases, including the TFIIH-associated kinase, cyclin H/CDK7/p36 (Cisek and Corden, 1991; Payne and Dahmus, 1993; Serizawa et al., 1993b; Stevens and Maupin, 1989; Yankulov et al., 1995).

To determine whether H7, H8 and DRB could also inhibit cyclin C/CDK8-associated kinase activity, kinase assays were performed with purified recombinant cyclin C/CDK8 or cyclin H/CDK7/p36 in the presence of various concentrations of these inhibitors. The purified kinase was pre-incubated with the inhibitor for 10 min, and then the CTD peptide substrate (YSPTSPS)₄, [-³²P]ATP, and non-labeled ATP were added and the reactions were incubated for an additional 30 min. Kinase activity was quantified as shown in Figure 6, and each IC₅₀ value and its estimated uncertainty were calculated as the mean±s.d. from three independent experiments (Table 1).

Figure 6 and Table 1 show that H7, H8, and DRB can all inhibit cyclin C/CDK8 kinase activity towards the CTD peptide. Interestingly, H7 and H8 have different specificities towards cyclin C/CDK8 and cyclin H/CDK7/p36. H7 preferentially inhibited cyclin C/CDK8 (IC₅₀»amp;4 M versus 40 M for cyclin H/CDK7/p36), and H8 preferentially inhibited cyclin H/CDK7/p36 (IC₅₀»amp;6 M versus 50 M for cyclin C/CDK8). DRB, on the other hand, inhibited both cyclin C/CDK8 and cyclin H/CDK7/p36 similarly (IC₅₀»amp;20 M). Similar results were obtained with cyclin C and cyclin H complexes immunoprecipitated from human cell lysates (data not shown). The different specificities of these inhibitors for these CTD kinases suggests that the enzymes' active sites have different conformations. However the differences are unlikely to be of sufficient magnitude to provide specific inhibition in an intact cell, so we tested other ATP analogs that have been discovered to be highly specific for cyclin-dependent kinases.

These CDK inhibitors can easily block cell growth and may be useful as potential anti-tumor agents and for biochemical and functional analysis of CDKs (Meijer, 1996). Two of these inhibitors, flavopiridol [(+) cis-2-(2-chlorophenyl)-5,7-dihydroxy-8-(3-hydroxy-1-methyl-4-piperidinyl)-4H-benzopyran-4-one] (Losiewicz et al., 1994; Meijer, 1996) and olomoucine [2-(2-hydroxyethylamino)-6-benzylamino-9-methylpurine], were analysed to determine if they could inhibit cyclin C/CDK8-associated kinase activity. Inhibition assays were performed as described above for the transcriptional inhibitors with cyclin C/CDK8 and cyclin H/CDK7/p36. As a positive control, cyclin B/CDC2 CTD kinase activity was also monitored in the presence of these inhibitors.

As shown in Figure 7 flavopiridol was a potent inhibitor of both cyclin C/CDK8 (IC₅₀»amp;20 nM) and cyclin B/CDC2 (IC₅₀»amp;10 nM) and, to a lesser extent, cyclin H/CDK7/p36 (IC₅₀»amp;100 nM) (see also Table 2). This compound has previously been found to specifically inhibit CDC2, CDK2 and CDK4 at fairly low concentrations (IC₅₀»amp;0.4 M for cyclin B/CDC2 using histone H1 as a substrate). Interestingly, cyclin B/CDC2 phosphorylation of the CTD peptide was much more sensitive to flavopiridol than was its phosphorylation of histone H1.

Olomoucine strongly inhibited cyclin H/CDK7/p36 (IC₅₀»amp;10 M) and cyclin B/CDC2 (IC₅₀´4 M) but was much less effective towards cyclin C/CDK8 (IC₅₀»amp;90 M) (see Figure 7 and Table 2). This inhibitor has been previously described as being highly specific for CDC2, CDK2 and CDK5 (IC₅₀»amp;7 M for cyclin B/CDC2 using Histone H1 as a substrate) and can also cause cell cycle arrest at both G1/S and G2/M (Abraham et al., 1995; Glab et al., 1994; Misteli and Warren, 1995; Vesely et al., 1994). However, an isomer of olomoucine called isoolomoucine [2-(2-hydroxyethylamino)-6-benzylamino-7-methylpurine], which had little inhibitory activity towards cyclin B/CDC2 (IC₅₀>0.5 mM) or cyclin H/CDK7/p36 (IC₅₀>1 mM), was able to inhibit cyclin C/CDK8 as well as olomoucine (IC₅₀»amp;90 M). Isoolomoucine has not been previously found to inhibit CDC2, CDK2 or CDK5 (IC₅₀>500 M for cyclin B/CDC2 using histone H1 as a substrate) (Vesely et al., 1994). All three kinases were resistant to the inhibitor solvent, DMSO, up to a concentration of 1% (v/v), so this isoolomoucine-specific inhibition is not due to a solvent effect. Cyclin C, cyclin H and CDC2 complexes immunoprecipitated from human cell lysates are similarly inhibited by these molecules (data not shown).

Whilst evaluation of these small chemical inhibitors suggested that selectivity between the CTD kinases may ultimately be achievable, existing compounds do not present enough specificity to be useful in evaluating the differential roles of the CTD kinases.

Discussion

Differential phosphorylation of the carboxyl-terminal domain (CTD) of RNA polymerase II by different CTD kinases may be important for differential regulation of CTD function. Cyclin C/CDK8 and cyclin H/CDK7/p36 are both CTD kinases, found in distinct complexes associated with the transcriptional machinery (Maldonado et al., 1996; Rickert et al., 1996; Roy et al., 1994; Serizawa et al., 1995; Shiekhattar et al., 1995). It has been demonstrated here that these kinases can differentially phosphorylate the CTD of RNA polymerase II, both as an exogenous recombinant substrate and as an endogenous substrate.

Cyclin C/CDK8 and cyclin H/CDK7/p36 both phosphorylate the same residue in the CTD consensus sequence (Ser-5). There is a difference in their ability to phosphorylate CTD substrates of varying lengths, suggesting that these kinases may differentially recognize secondary structure in the substrate. However, the major differences seen in phosphorylation of recombinant CTD substrates by these kinases and by cyclin B/CDC2 may be attributed to their different ability to phosphorylate CTD non-consensus sequences with a lysine substituted for serine at position seven, with both cyclin H/CDK7/p36 and cyclin B/CDC2 preferentially recognizing this substituted repeat. The appearance of most of the lysine-substituted non-consensus repeats in the last third of the CTD is consistent with the results observed with the three GST - CTD fusion proteins. The differences in phosphorylation of the recombinant CTD substrate are less dramatic than those with the lysine-substituted peptide, presumably due to the presence of non-lysine-substituted repeats, but are still significant. It would be interesting to examine the effects of removing or adding such non-consensus repeats to the CTD in an in vivo assay (Bartolomei et al., 1988). In addition, analysing the effects of CTD phosphorylation on Plasmodium berghei RNA polymerase II, which has a high number of lysine substitutions at the same position (Giesecke et al., 1991), may provide insight into the roles of cyclin C/CDK8 and cyclin H/CDK7/p36.

The different specificity of the ATP analogs to inhibit CTD phosphorylation by cyclin-dependent kinases provides further evidence that these kinases may have a different conformation in or near their ATP-binding sites. The crystal structure of CDK2 complexed with olomoucine shows that the purine ring of olomoucine binds CDK2 in the same location as ATP, but with the purine ring oriented differently from that of ATP (Schulze-Gahmen et al., 1995). Although the purine N7 atom of olomoucine makes various contacts with CDK2, the identical inhibition profiles of isoolomoucine and olomoucine for cyclin C/CDK8 suggest that neither N7 nor N9, the locations of the differently substituted methyl groups, makes significant contacts to this kinase. Olomoucine is thus likely to bind cyclin C/CDK8 in a different orientation than it does the other CDKs. The different structures of these kinase active sites may explain their different specificities towards CTD substrates. Furthermore, cyclin H/CDK7/p36, but not cyclin C/CDK8, can phosphorylate GST - CDK2 and Histone H1 (Rickert, 1997), showing a different substrate specificity between these two kinases. The ability of H7, H8, and DRB to inhibit both of these kinases indicates that more selective compounds will need to be identified for further dissection of the functions of CTD kinases in transcription regulation in vivo.

The ability of these cyclin-dependent kinases to cause different phosphorylation patterns on the CTD does not require the presence of other proteins or pre-existing phosphates to confer specificity. However, these conditions may modulate these kinases' phosphorylation pattern in vivo. Phosphorylation of the full length CTD construct, GST-CTD1-53, by immunoprecipitated cyclin H/CDK7/p36 and cyclin B/CDC2 was 2 - 3-fold more efficient than by purified enzymes (data not shown), suggesting that specificity may be affected by other proteins in the complexes or by post-translational modifications. In addition, the V8 digestion pattern of an endogenous RNA polymerase II substrate appears more complex than the simple pattern produced from a recombinant CTD substrate (data not shown). This suggests that there may be other phosphorylation sites on RNA polymerase II not found in the CTD itself, as is the case with Trypanosoma brucei RNA polymerase II, which becomes phosphorylated despite a lack of heptapeptide repeats (Chapman and Agabian, 1994).

The experiments presented here demonstrate that cyclin C/CDK8 and cyclin H/CDK7/p36 can differentially phosphorylate their common substrate, the carboxyl-terminal domain of RNA polymerase II. These kinases are found in separate complexes associated with transcriptional proteins, suggesting that they may have different roles (Maldonado et al., 1996; Rickert et al., 1996). These findings are in support of previous genetic evidence that the yeast homologs of cyclin C/CDK8 and cyclin H/CDK7/p36, SRB10/SRB11 and KIN28/CCL1, respectively, have distinct functions (Kuchin et al., 1995; Liao et al., 1995; Simon et al., 1986; Valay et al., 1993). Together, these results suggest that these kinases play different roles in regulating the functions of the CTD and RNA polymerase II.

Materials and methods

Materials

[-³²P]ATP (> 5000 Ci/mmol) was obtained from Amersham. H7 (Sigma) and H8 (Seikagaku Corporation) were used from 10 mM stock solutions in H₂O. DRB (Sigma) was used from a 10 mM stock solution in ethanol. Olomoucine and isoolomoucine were both obtained from Calbiochem and used from 25 mM stock solutions in dimethylsulfoxide (DMSO). Flavopiridol was obtained from the National Cancer Institute and was used from a 10 mM stock solution in DMSO. V8 protease was obtained from Worthington. The anti-cyclin C and anti-CDK8 antibodies -C pep and -CDK8 pep have been previously described (Rickert et al., 1996). Cyclin H and CDK7 antibodies were obtained from Santa Cruz Biotechnology, Inc.

Recombinant proteins from Sf9 cells

cDNAs expressing human cyclin C and human CDK8 were cloned into a pVL1393 baculovirus transfer vector (PharMingen). CDK8 was epitope-tagged (EEEEYMPME) at its amino-terminus for affinity purification. The vectors were transfected into Ready-Plaque Sf9 cells (Novagen) and recombinant baculovirus was isolated and amplified according to the manufacturer's directions. Cyclin B (GST-fusion protein) and CDC2 (Hemagglutinin (HA)-tagged) baculoviruses were similarly created. Cyclin H (6´His-tagged), CDK7 (HA-tagged), and p36^MAT1 baculoviruses were gifts from DO Morgan (University of California, San Francisco). Sf9 cells were infected with these baculoviruses as described (Gruenwald and Heitz, 1993). For purified enzymes, cells were co-infected with cyclin C and Glu-CDK8 together or His-cyclin H, HA-CDK7, and p36 together. Cells were singly infected with either GST-cyclin B or HA-CDC2.

Cells were harvested and washed in 50 mM HEPES pH 7.0, 150 mM NaCl by centrifuging at 1000 r.p.m. 4°C for 5 min. The cells were resuspended in 5 ml E1A lysis buffer (50 mM HEPES pH 7.0, 150 mM NaCl, 0.1% NP-40, 5 g/ml aprotinin, 5 mg/ml leupeptin, 5 g/ml pepstatin, and 125 g/ml Pefabloc^Ò SC) per gram of cells and rocked at 4°C for 30 min. The mixture was then homogenized with 20 strokes in a Dounce homogenizer and centrifuged at 10 000 r.p.m. 4°C for 20 min to remove cell debris. Cyclin B and CDC2 lysates were mixed at room temperature for 20 min. The lysates were then purified on -Glu (Grussenmeyer et al., 1985) or 12CA5 (Field et al., 1988) coupled to protein G Sepharose^Ò, or glutathione-agarose as appropriate, eluting with two column volumes of 100 g/ml antigenic Glu peptide or antigenic hemagglutinin peptide in phosphate-buffered saline, or two column volumes of 20 mM reduced glutathione in 50 mM Tris pH 8.0. The peptides were EEEEYMPME (Glu) and YPYDVPDYA (HA) and were obtained from Research Genetics. Purified enzymes were assayed for purity and for activity (Rickert, 1997).

Construction of GST-CTD fusion proteins

GST-CTD1 - 53, GST-CTD1 - 29, and GST-CTD30 - 53 were derived by PCR from cDNA encoding the mouse CTD, using oligonucleotide primers specific for the sequences encoding the repeats indicated. The PCR products were subcloned into pGEX-6P (Pharmacia Biotech) for expression and purification as GST-fusion proteins (Sambrook et al., 1989). The GST-CTD fusion proteins WT, A2, and A5 were subcloned from constructs described in (West and Corden, 1995). The GST-(YSPTSPS)_n, n=4, 6, 8, or 10, were created as described in (West and Corden, 1995) and were subcloned into a GST vector as described in (Rickert et al., 1996).

Kinase assays

Kinase assays and reimmunoprecipitations were performed as previously described (Rickert et al., 1996). For assays using purified enzymes, 0.5 g of purified enzyme (cyclin C/CDK8, cyclin H/CDK7/p36, or cyclin B/CDC2) was used. 1 g of CTD peptides (YSPTSPX)₄, GST-CTD1 - 53, GST-CTD1 - 29, or GST-CTD30 - 53 was used as a substrate. For the substrate length dependence, 100 pmol of GST-(YSPTSPS)_n, n=4, 6, 8, or 10, was added as a substrate. Results were normalised against mock reaction controls.

Inhibition assays

Purified enzymes or immunoprecipitations from cells, prepared as described above or in (Rickert et al., 1996), were added to kinase buffer containing inhibitors at various concentrations to a total volume of 19 l. This mixture was incubated at 30°C for 10 min. Subsequently, a 1 l reaction mix containing 1 g of CTD peptide, 2 Ci [-³²P]ATP, and 20 M ATP (final concentration 1 M) was added to give a final reaction volume of 20 l. The reactions were incubated at 30°C for 30 min, stopped, and analysed as usual. Kinase activity was quantified using a Molecular Dynamics PhosphorImager and calculated as a percentage of activity in the absence of any inhibitors. Inhibition curves were graphed and fit using CA-Cricket Graph III 1.5.

Partial proteolytic mapping

For V8 mapping of phosphorylated RNA polymerase II, immunoprecipitation-kinase assays were performed as described previously (Rickert et al., 1996) in triplicate using 1 mg of CEM whole cell lysate per assay. The kinase reactions were electrophoresed on 6% SDS-polyacrylamide gels, which were dried and autoradiographed. The RNA polymerase II bands (identified through co-migration of reimmunoprecipitated RNA polymerase II) were excised and used for partial proteolytic mapping with V8 protease as described previously (Cleveland et al., 1977).

Acknowledgements

We thank David Morgan for his generous gift of cyclin H, CDK7, and p36^MAT1-expressing baculoviruses. We are also grateful to F Shanahan for assistance with tissue culture. We thank J Bolen, D Mahony, W Seghezzi, F Shanahan and N Solvason for critical comments on the manuscript and W Huestis for helpful advice. DNAX Research Institute of Molecular and Cellular Biology is supported by the Schering - Plough Corporation.

Abraham RT, Acquarone M, Andersen A, Asensi A, Belle R, Berger F, Bergounioux C, Brunn G, Buquet-Fagot C, Fagot D, Glab N, Goudeau H, Goudeau M, Guerrier P, Houghton P, Hendriks H, Kloareg B, Lippai M, Marie D, Maro B, Meijer L, Mester J, Mulner-Lorillon O, Poulet SA, Schierenberg E, Schutte B, Vaulot D and Verlhac MH. (1995). Biol. Cell 83, 105-120. MEDLINE

Akoulitchev S, Mäkelä TP, Weinberg RA and Reinberg D. (1995). Nature 377, 557-560. MEDLINE

Allison LA, Moyle M, Shales M and Ingles CJ. (1985). Cell 42, 599-610. MEDLINE

Allison LA, Wong JK-C, Fitzpatrick VD, Moyle M and Ingles CJ. (1988). Mol. Cell. Biol. 8, 321-329. MEDLINE

Arias JA, Peterson SR and Dynan WS. (1991). J. Biol. Chem. 266, 8055-8061. MEDLINE

Bartolomei MS, Halden NF, Cullen CR and Corden JL. (1988). Mol. Cell. Biol. 8, 330-339. MEDLINE

Baskaran R, Chiang GG, Mysliwiec T, Kruh GD and Wang JYJ. (1997). J. Biol. Chem. 272, 18905-18909. MEDLINE

Baskaran R, Dahmus ME and Wang JY. (1993). Proc. Natl. Acad. Sci. USA 90, 11167-11171. MEDLINE

Bellier S, Dubois MF, Nishida E, Almouzni G and Bensaude O. (1997). Mol. Cell. Biol. 17, 1434-1440. MEDLINE

Bregman DB, Du L, van der Zee S and Warren SL. (1995). J. Cell Biol. 129, 287-298. MEDLINE

Cadena DL and Dahmus ME. (1987). J. Biol. Chem. 262, 12468-12474. MEDLINE

Cagas PM and Corden JL. (1995). Proteins 21, 149160.

Chambers RS and Dahmus ME. (1994). J. Biol. Chem. 269, 26243-26248. MEDLINE

Chapman AB and Agabian N. (1994). J. Biol. Chem. 269, 4754-4760. MEDLINE

Cisek LJ and Corden JL. (1989). Nature 339, 679-684. MEDLINE

Cisek LJ and Corden JL. (1991). Methods Enz. 200, 301-325.

Cismowski MJ, Laff GM, Solomon MJ and Reed SI. (1995). Mol. Cell. Biol. 15, 2983-2992. MEDLINE

Cleveland DW, Fischer SG, Kirschner MW and Laemmli UK. (1977). J. Biol. Chem. 252, 1102-1106. MEDLINE

Corden JL. (1993). Curr. Opin. Genet. Dev. 3, 213-218. MEDLINE

Corden JL, Cadena DL, Ahearn Jr JM and Dahmus ME. (1985). Proc. Natl. Acad. Sci. USA 82, 7934-7938. MEDLINE

Dahmus ME. (1981). J. Biol. Chem. 256, 3332-3339. MEDLINE

Dahmus ME. (1996). J. Biol. Chem. 271, 1900919012.

Du L and Warren SL. (1997). J. Cell Biol. 136, 5-18. . MEDLINE

Dubois M-F, Nguyen VT, Bellier S and Bensaude O. (1994a). J. Biol. Chem. 269, 13331-13336. MEDLINE

Dubois M-F, Nguyen VT, Dahmus ME, Pagès, G., Pouysségur J and Bensaude O. (1994b). EMBO J. 13, 4787-4797. MEDLINE

Dubois M-F, Vincent M, Vigneron M, Adamczewski J, Egly J-M and Bensaude O. (1997). Nucleic Acids Res. 25, 694-700. Article MEDLINE

Edwards MC, Wong C and Elledge SJ. (1998). Mol. Cell. Biol. 18, 4291-4300. MEDLINE

Feaver WJ, Gileadi O, Li Y and Kornberg RD. (1991). Cell 67, 1223-1230. MEDLINE

Feaver WJ, Svejstrup JQ, Henry NL and Kornberg RD. (1994). Cell 79, 1103-1109. MEDLINE

Field J, Nikawa J-I, Broek D, MacDonald B, Rodgers L, Wilson IA, Lerner RA and Wigler M. (1988). Mol. Cell. Biol. 8, 2159-2165. MEDLINE

Gebara MM, Sayre MH and Corden JL. (1997). J. Cell. Biochem. 64, 390-402. MEDLINE

Giesecke H, Barale JC, Langsley G and Cornelissen AW. (1991). Biochem. Biophys. Res. Commun. 180, 1350-1355. MEDLINE

Glab N, Labidi B, Qin LX, Trehin C, Bergounioux C and Meijer L. (1994). FEBS Lett. 353, 207-211. MEDLINE

Gruenwald S and Heitz J. (1993). Baculovirus Expression Vector System: Procedures & Methods Manual, 2nd edn. PharMingen, San Diego.,

Grussenmeyer T, Scheidtmann KH, Hutchinson MA, Eckhart W and Walter G. (1985). Proc. Natl. Acad. Sci. USA 82, 7952-7954. MEDLINE

Jiang Y and Gralla JD. (1994). Nucleic Acids Res. 22, 4958-4962. MEDLINE

Kim E, Du L, Bregman DB and Warren SL. (1997). J. Cell Biol. 136, 19-28. MEDLINE

Kuchin S, Yeghiayan P and Carlson M. (1995). Proc. Natl. Acad. Sci. USA 92, 4006-4010. MEDLINE

Lee JM and Greenleaf AL. (1989). Proc. Natl. Acad. Sci. USA 86, 3624-3628. MEDLINE

Lee JM and Greenleaf AL. (1991). Gene Expr. 1, 149-167. MEDLINE

Lee JM and Grenleaf AL. (1997). J. Biol. Chem. 272, 10990-10993. MEDLINE

Liao SM, Zhang J, Jeffery DA, Koleske AJ, Thompson CM, Chao DM, Viljoen M, van Vuuren HJ and Young RA. (1995). Nature 374, 193-196. MEDLINE

Losiewicz MD, Carlson BA, Kaur G, Sausville EA and Worland PJ. (1994). Biochem. Biophys. Res. Commun. 201, 589-595. Article MEDLINE

Lu H, Zawel L, Fisher L, Egly J-M and Reinberg D. (1992). Nature 358, 641-645. MEDLINE

Mäkelä TP, Parvin JD, Kim J, Huber LJ, Sharp PA and Weinberg RA. (1995). Proc. Natl. Acad. Sci. USA 92, 5174-5178. MEDLINE

Maldonado E, Shiekhattar R, Sheldon M, Cho H, Drapkin R, Rickert P, Lees E, Anderson CW, Linn S and Reinberg D. (1996). Nature 381, 86-89. MEDLINE

Marshall NF, Peng J, Xie Z and Price DH. (1996). J. Biol. Chem. 271, 27176-27183. MEDLINE

Marshall NF and Price DH. (1995). J. Biol. Chem. 270, 12335-12338. MEDLINE

McCracken S, Fong N, Yankulov K, Ballantyne S, Pan G, Greenblatt J, Patterson SD, Wickens M and Bentley DL. (1997). Nature 387, 357-361. MEDLINE

Meijer L. (1996). Trends Cell Biol. 6, 393-397.

Misteli T and Warren G. (1995). J. Cell Sci. 108, 2715-2727. MEDLINE

Mortillaro MJ, Blencowe BJ, Wei X, Nakayasu H, Du L, Warren SL, Sharp PA and Berezney R. (1996). Proc. Natl. Acad. Sci. USA 93, 8253-8257. MEDLINE

Nonet M, Sweetser D and Young RA. (1987). Cell 50, 909-915. MEDLINE

O'Brien T, Hardin S, Greenleaf A and Lis JT. (1994). Nature 370, 75-77. MEDLINE

Payne JM and Dahmus ME. (1993). J. Biol. Chem. 268, 80-87. MEDLINE

Peng J, Zhu Y, Milton JT and Price DH. (1998). Genes Dev. 12, 755-762. MEDLINE

Rice SA, Long MC, Lam V and Spencer CA. (1994). J. Virol. 68, 988-1001. MEDLINE

Rickert P. (1997). Functional and Biochemical Characterization of the cyclin C/CDK8 kinase. PhD Thesis. Stanford University, California.,

Rickert P, Seghezzi W, Shanahan F, Cho H and Lees E. (1996). Oncogene 12, 2631-2640. MEDLINE

Roy R, Adamczewski JP, Seroz T, Vermeulen W, Tasan J-P, Schaeffer L, Nigg EA, Joeijmakers JHJ and Egly J-M. (1994). Cell 79, 1093-1101. MEDLINE

Sambrook J, Fritsch EF and Maniatis T. Molecular Cloniong: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Plainview, NY.,

Schulze-Gahmen U, Brandsen J, Jones JH, Morgan DO, Meijer L, Vesely J and Kim S-H. (1995). Proteins 22, 378-391. MEDLINE

Sehgal PB, Darnell Jr JE and Tamm I. (1976). Cell 9, 473-480. MEDLINE

Serizawa H, Conaway JW and Conaway RC. (1993a). Nature 363, 371-374. MEDLINE

Serizawa H, Conaway RC and Conaway JW. (1993b). J. Biol. Chem. 268, 17300-17308. MEDLINE

Serizawa H, Mäkelä TP, Conaway JW, Conaway RC, Weinberg RA and Young RA. (1995). Nature 374, 280-282. MEDLINE

Shiekhattar R, Mermelstein F, Fisher RP, Drapkin R, Dynlacht B, Wessling HC, Morgan DO and Reinberg D. (1995). Nature 374, 283-287. MEDLINE

Shilatifard A, Conaway JW and Conaway RC. (1997). Curr. Opin. Genet. Dev. 7, 199-204. MEDLINE

Simon M, Séraphin B and Faye G. (1986). EMBO J. 5, 2697-2701. MEDLINE

Song CZ. (1996). Biochem. Biophys. Res. Commun. 229, 810-816. MEDLINE

Songyang Z, Blechner S, Hoagland N, Hoekstra MF, Piwnica-Worms H and Cantley LC. (1994). Curr. Biol. 4, 973-982. MEDLINE

Steinmetz EJ. (1997). Cell 89, 491-494. MEDLINE

Stevens A and Maupin MK. (1989). Biochem. Biophys. Res. Commun. 159, 508-515. MEDLINE

Valay JG, Simon M and Faye G. (1993). J. Mol. Biol. 234, 307-310. MEDLINE

Venetianer A, Dubois MF, Nguyen VT, Bellier S, Seo SJ and Bensaude O. (1995). Eur. J. Biochem. 233, 83-92. MEDLINE

Vesley J, Havlicek L, Strnad M, Blow JJ, Donella-Deana A, Pinna L, Letham DS, Kato J-Y, Detivaud L, Leclerc S and Meijer L. (1994). Eur. J. Biochem. 224, 771-786. MEDLINE

Wei P, Garber ME, Fang SM, Fischer WH and Jones KA. (1998). Cell 92, 451-462. MEDLINE

West ML and Corden JL. (1995). Genetics 140, 1223-1233. MEDLINE

Yankulov K, Yamashita K, Roy R, Egly JM and Bentley DL. (1995). J. Biol. Chem. 270, 23922-23925. MEDLINE

Yuryev A and Corden JL. (1996). Genetics 143, 661-671. MEDLINE

Yuryev A, Patturajan M, Litingtung Y, Joshi RV, Gentile C, Gebara M and Corden JL. (1996). Proc. Natl. Acad. Sci. USA 93, 6975-6980. Article MEDLINE

Zehring WA, Lee JM, Weeks JR, Jokerst RS and Greenleaf AL. (1988). Proc. Natl. Acad. Sci. USA 85, 3698-3702. MEDLINE

Zhu Y, Pe'ery T, Peng J, Ramanathan Y, Marshall N, Marshall T, Amendt B, Mathews MB and Price DH. (1997). Genes Dev. 11, 2622-2632. MEDLINE

Figure 1 Differential phosphorylation of recombinant CTD substrates by cyclin C/CDK8, cyclin H/CDK7/p36, and cyclin B/CDC2. (a) Schematic of recombinant CTD substrates, shown by repeat number. The consensus repeats are shown as white boxes, non-consensus repeats with lysine substitutions at position as hatched boxes, and other non-consensus repeats as dark gray boxes. (b) Phosphorylation of GST-CTD1 - 53, GST-CTD1 - 29, and GST-CTD30 - 53 by purified, recombinant cyclin C/CDK8, cyclin H/CDK7/p36 and cyclin B/CDC2. Reactions were electrophoresed on an 8% SDS-polyacrylamide gel and quantified on a Molecular Dynamics Storm PhosphorImager. The phosphorylation of each substrate relative to phosphorylation of GST-CTD1 - 53 by each kinase was calculated as mean±s.d. from two independent experiments. Data were normalized against mock reactions using SF9 extract alone

Figure 2 Phosphorylation of Ser-5 in the CTD consensus repeat by cyclin C/CDK8 and cyclin H/CDK7/p36. CEM whole cell lysates (~300 g total protein) were immunoprecipitated with -C pep, -cyc H (Santa Cruz) or -CDC2 antibodies. Normal rabbit sera (NR) was used as a negative control. Recombinant cyclin C/CDK8 and cyclin H/CDK7/p36 were purified from SF9 cells. Kinase reactions were performed with 1 g of WT (GST-[YSPTSPS]₁₆), A2 (GST-[YAPTSPS]₁₄), A5 (GST-[YSPTAPS]₁₅), or no substrate added. Reactions were electrophoresed on a 10% SDS-polyacrylamide gel which was subsequently Coomassie stained, dried, and autoradiographed

Figure 3 Phosphorylation of non-consensus CTD repeats by recombinant, purified cyclin C/CDK8, cyclin H/CDK7/p36, and cyclin B/CDC2. WT = (YSPTSPS)₄, K = (YSPTSPK)₄, N = (YSPTSPN)₄ and T = (YSPTSPT)₄. (a) Reactions were electrophoresed on an 8 - 20% gradient SDS-polyacrylamide gel and autoradiographed. (b) Kinase reactions were quantified on a Molecular Dynamics Storm PhosphorImager. The phosphorylation of each peptide relative to phosphorylation of the WT peptide by each kinase was calculated as mean±s.d. from three independent experiments. Data were normalized against mock reactions using SF9 extract alone

Figure 4 Phosphorylation of GST-(YSPTSPS)_{4 - 10} by purified, recombinant cyclin C/CDK8 and cyclin H/CDK7/p36. (a) Reactions were electrophoresed on a 10% SDS-polyacrylamide gel which was Coomassie stained, dried, and autoradiographed. (b) Kinase reactions with cyclin C/CDK8 () and cyclin H/CDK7/p36 () were quantified on a Molecular Dynamics Storm PhosphorImager. The relative amount of phosphate incorporated by each substrate was calculated as mean±s.d. from four independent experiments. Slopes were calculated from a linear curve fit in Cricket Graph

Figure 5 V8 proteolytic mapping of phosphorylated RNA polymerase II. (a) CEM whole cell lysate (~1.5 mg) was immunoprecipitated with -C pep, -cyclin H and normal rabbit (NR) as a negative control and assayed for kinase activity. The immune complex kinase reactions were denatured and reimmunoprecipitated with either an -RNA polymerase II monoclonal antibody (Promega) or PAb 419 as a negative control. Samples were electrophoresed on a 6% SDS-polyacrylamide gel and visualized by autoradiography. (b) Kinase assays were performed as above with -C pep, -CDK8 pep, -H (Santa Cruz) or -CDK7 (Santa Cruz) antisera and electrophoresed on a 6% SDS-polyacrylamide gel. The phosphorylated RNA polymerase II from each reaction was excised from the gel and compared by V8 partial proteolytic mapping on an 8 - 20% gradient SDS-polyacrylamide gel

Figure 6 Inhibition of cyclin C/CDK8 () and cyclin H/CDK7/p36 () kinase activity by H7, H8 and DRB. Recombinant, purified cyclin C/CDK8 and cyclin H/CDK7/p36 enzymes were pre-incubated with the indicated concentration of inhibitor. Kinase reactions were then performed in the presence of the inhibitor using the CTD peptide (YSPTSPS)₄. Kinase reactions were electrophoresed on 14% SDS-polyacrylamide gels and quantified on a Molecular Dynamics PhosphorImager

Figure 7 Inhibition of cyclin C/CDK8 (), cyclin H/CDK7/p36 () and cyclin B/CDC2 (x) kinase activity by flavopiridol, olomoucine, and isoolomoucine. Recombinant, purified cyclin C/CDK8, cyclin H/CDK7/p36, and cyclin B/CDC2 enzymes were pre-incubated with the indicated concentration of inhibitor. Kinase reactions were then performed in the presence of the inhibitor using the CTD peptide (YSPTSPS)₄. Kinase reactions were electrophoresed on 14% SDS-polyacrylamide gels and quantified on a Molecular Dynamics PhosphorImager

Table 1 Table 1

Table 2 Table 2