The Cyclin H/Cdk7/Mat1 Kinase Activity Is Regulated by CK2 Phosphorylation of Cyclin H

Oncogene (2002) 21, 5031 – 5037 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc ORIGINAL PAPERS The cyclin H/cdk7/Mat1 kinase activity is regulated by CK2 phosphorylation of cyclin H

Eberhard Schneider1,2, Sabine Kartarius1, Norbert Schuster1,3 and Mathias Montenarh*,1

1Medical Biochemistry and Molecular Biology, University of the Saarland, Building 44, D-66424, Homburg, Germany

Cyclin dependent kinases are regulated by phosphoryla- an initiation of mitosis. A third phosphorylation at tion and dephosphorylation of the catalytic cdk subunits, Thr-160 is also essential for the activity of p34cdc2 and by assembly with specific cyclins and by specific inhibitor this phosphorylation is achieved by another cyclin molecules. Recently, it turned out that cyclins are also dependent kinase complex which is known as cdk phosphoproteins, which means that they are also activating kinase CAK. This kinase complex consists of potential targets for a regulation by phosphorylation three subunits: the regulatory subunit cyclin H, the and dephosphorylation. Here, we show that cyclin H was catalytic subunit cdk7 and a third protein factor, phosphorylated by protein kinase CK2. Like most other named Mat1. This third factor, which is unusual for CK2 substrates cyclin H was much better phosphorylated other cdk/cyclin-complexes is an assembly factor for by the CK2 holoenzyme than by the a-subunit alone. By the cyclin H/cdk7/Mat1 complex (Johnson and Walker, using point mutants derived from the cyclin H sequence 1999) and seems to be implicated in the regulation of we mapped the CK2 phosphorylation site at threonine substrate specificity. 315 at the C-terminal end of cyclin H. Phosphorylation In contrast to other cyclin dependent kinases levels of at this position had no influence on the assembly of the cyclin H, cdk7 and Mat1 subunits as well as the resulting cyclin H/cdk7/Mat1 complex. However, phosphorylation kinase activity do not greatly vary over the eukaryotic at amino acid 315 of cyclin H turned out to be critical cell cycle (Adamczewski et al., 1996; Fisher and for a full cyclin H/cdk7/Mat1 kinase activity when the Morgan, 1994, Fisher et al, 1995). The cyclin H/cdk7/ CTD peptide of RNA polymerase II or cdk2 was used as Mat1 complex is found either in a free state (Morgan, a substrate. 1995) or as a part of the transcription factor complex Oncogene (2002) 21, 5031 – 5037. doi:10.1038/sj.onc. TFIIH (Svejstrup et al., 1996). This TFIIH complex has 1205690 a central role in two processes: first as a basal factor in the regulation of the transcription of RNA polymerase Keywords: protein kinase CK2; cyclin dependent II and second as an essential component of the DNA kinases; phosphorylation regulation repair complex which is responsible for the nucleotide excision repair (Schaefer et al., 1993). Two classes of substrates for the cyclin H/cdk7/Mat1 complexes are Introduction known, namely downstream cyclin dependent kinases as well as components of the transcription machinery of Cell cycle progression is tightly regulated by cyclin the cell including the C-terminal domain (CTD) of the dependent kinases (cdks). The activity of these kinases large subunit of RNA polymerase II (Feaver et al., 1994) is controlled by phosphorylation and dephosphoryla- and the growth suppressor protein p53 (Ko et al., 1997; tion events, by binding of the regulatory subunits, Lu et al., 1997). Human cyclin H is a polypeptide of 323 namely the cyclins to the catalytic cdk-subunits, as well amino acids with a molecular weight of 38 kDa. When as by the interaction of inhibitor molecules with the cyclin H assembles with cdk7 the kinase is activated and cyclin/cdk-complexes (Kaldis, 1999). Cyclin dependent the activity is further enhanced by the third factor Mat1. kinases are phosphorylated on several sites including So far, only little is known about the regulation of cyclin two inhibitory and one activating site. The inhibitory H/cdk7/Mat1 activity. phosphorylation sites on human cdc2, Thr-14 and Tyr- We have recently shown that wild-type p53 binds to 15, keep the p34cdc2 kinase silent until dephosphoryla- cyclin H and this binding leads to a down-regulation of tion by cdc25C activates p34cdc2, which goes along with the kinase activity in vitro and in vivo (Schneider et al., 1998). p53 is a growth suppressor and a regulator of cell cycle progression. Upon DNA damage and after *Correspondence: M Montenarh; E-mail: [email protected] reduction of the nucleotide pool p53 transactivates a Current addresses: 2Phylos GmbH, Industrial Park Hoechst, Building number of different genes including the p21WAF1 gene. G 830, 65926 Frankfurt, Germany; E-mail: Eberhard.Schneider@x- WAF1 3 Elevated transcription of the p21 leads to an zillion.com; Ludwig Institute for Cancer Research, Box 595, 75124 WAF1 Uppsala, Sweden increase in the concentration of the p21 protein Received 17 January 2002; revised 12 May 2002; accepted 20 May which is an efficient inhibitor of cyclin dependent 2002 kinases (El-Deiry et al., 1993). Thus, p53 regulates CK2 phosphorylation of cyclin H E Schneider et al 5032 cyclin dependent kinases indirectly by transactivating strates the amount of cyclin H which was present in an inhibitor molecule for these kinases. An alternative each lane of Figure 1a. As shown in Figure 1a (lane 1) way to explain a p53 dependent cell cycle arrest cyclin H was efficiently phosphorylated by the CK2 independent of p21 might exist in form of the observed holoenzyme and weaker phosphorylated by the CK2 a- p53/CAK interaction. subunit alone (lane 3). Phosphorylation of cyclin H by Recently, it was shown that cyclin H is a the CK2 holoenzyme as well as the CK2 a-subunit was phosphoprotein and one of the kinases that phosphor- inhibited by heparin (lanes 2 and 4). This shows that ylate cyclin H was identified as cyclin C/cdk8. This cyclin H is in fact a substrate of CK2 and behaves very enzyme phosphorylates cyclin H at residues Ser-5 and similar to the majority of other CK2 substrates. Lanes Ser-304 both in vitro and in vivo (Akoulitchev et al., 5 and 6 are controls where cyclin H was incubated with 2000). These two phosphorylation sites are in the [g-32P]ATP in the absence of added CK2 in order to vicinity of N- and C-terminal helices which are demonstrate the absence of a protein kinase in the uniquely orientated in cyclin H as compared with cyclin H preparation. As a control for the correct other cyclins (Andersen et al., 1997). Phosphorylation separation of cyclin H by SDS polyacrylamide gel of cyclin H in the TFIIH complex results in inhibition electrophoresis we used bacterially expressed recombi- of its transcription activity whereas when Ser-5 and nant cyclin H which was phosphorylated with protein Ser-304 were substituted by alanine transcription was kinase A (heart muscle kinase, Sigma) within an not inhibited. Here, we show that cyclin H is a artificial PKA-site at the N-terminus of the cyclin H substrate for another protein kinase namely protein (lane 7). kinase CK2. Furthermore, we identified the phosphor- By scanning the polypeptide chain of cyclin H for ylation site for CK2 at position 315 on the polypeptide potential phosphorylation sites five different sequences chain of cyclin H and analysed the influence of this were found as putative CK2 phosphorylation sites phosphorylation on the activity of the cyclin H/cdk7/ (marked by P in Figure 2). Taking into consideration Mat1 complex. Using point mutants of cyclin H the consensus sequence for CK2 phosphorylation and (T315D and T315A) mimicking phosphorylated or the fact that the presence of additional Asp or Glu unphosphorylated cyclin H, respectively, we analysed residues at positions 73,+1,+2,+4,+5 or +7 the kinase activity of the cyclin H/cdk7/Mat1 complex. relative to the Ser or Thr residues which are Based on these data we concluded that the phosphor- phosphorylated improves phosphorylation efficacy ylation of cyclin H at Thr-315 by CK2 seems to be (Marin et al., 1994; Meggio et al., 1994) all points to necessary for full kinase activity of the cyclin H/cdk7/ T315 as the preferred target site for phosphorylation. Mat1 complex with respect to the phosphorylation of Threonine 315 is exceptional out of the five putative the CTD peptide derived from the C-terminal domain CK2 sites because it is immersed within an acidic of RNA polymerase II as well as the phosphorylation of cdk2.

Results

Human cyclin H is a phosphoprotein of 323 amino acids with a molecular weight of 38 kDa. Analysis of the primary sequence of cyclin H revealed the existence of several potential phosphorylation sites for protein kinase CK2. Protein kinase CK2 is a holoenzyme composed of two regulatory b-subunits and two catalytic a-ora’-subunits. Most of the CK2 substrates are efficiently phosphorylated only by the CK2 holoenzyme whereas a few substrates are phosphorylated by the catalytic a-subunit alone (for review see: Guerra and Issinger, 1999b). Therefore, we first analysed whether cyclin H is in vitro a substrate of CK2 and second whether cyclin H is phoshorylated by the a-subunit alone or by the holoenzyme composed of Figure 1 Cyclin H is phosphorylated by the catalytic a-subunit a-andb-subunits. Bacterially expressed recombinant of CK2 and by the holoenzyme. (a) Purified cyclin H was incu- cyclin H was purified and then incubated with the a- bated with the CK2 holoenzyme (1,2) consisting of a- and b-subunits or with the a-subunit of CK2 (3,4) and with [g-32P]ATP in subunit and [g-32P]ATP or with the holoenzyme and 32 (b), (c) the presence (even lanes) or absence (odd lanes) of [g- P]ATP in the absence or presence of heparin, an 0.5 mg heparin. Proteins were analysed on a 10% SDS polyacry- efficient inhibitor of CK2 kinase activity (Hathaway et lamide gel followed by autoradiography. As a control cyclin H 32 al., 1980). Phosphorylated products were analysed by was incubated with [g- P]ATP in the absence of added CK2 (5,6). As a further control cyclin H was phosphorylated with pro- SDS polyacrylamide gel electrophoresis of the whole tein kinase A (PKA) at an artificial PKA-site (7). (b) Coomassie reaction mixture and autoradiography. Figure 1b brilliant blue stained 10% SDS polyacrylamide gel with cyclin shows a Coomassie blue stained gel which demon- H used in the corresponding experiments of (a)

Oncogene CK2 phosphorylation of cyclin H E Schneider et al 5033

Figure 2 Amino acid sequence of cyclin H with putative CK2 phosphorylation sites. The cyclin H amino acid sequence was analysed for the presence of putative CK2 phosphorylation sites using the prosite software. ‘P’ marks putative CK2 phosphorylation sites

environment having additional acidic residues at positions 73,+1,+2,+3. In order to map the phosphorylation of cyclin H directly to threonine 315 we constructed two point mutants of cyclin H where amino acid threonine 315 was substituted by alanine (T315A) or by aspartic acid (T315D). These two mutants as well as full-length cyclin H were expressed in baculovirus infected Sf9 cells, purified and then equal amounts of the different proteins were incubated with the CK2 holoenzyme together with [g-32P]ATP. Proteins were analysed on an SDS polyacrylamide gel and incorporated radioactivity was detected by autoradiography. As shown in Figure 3a (lane 1) wild-type cyclin H is strongly phosphorylated by CK2 whereas both point mutants of cyclin H were not phosphorylated at all. Figure 3b shows Figure 3 Phosphorylation of wild-type cyclin H and cyclin H Coomassie blue stained protein bands for the corre- mutated at position 315 by protein kinase CK2. (a) Wild-type cy- sponding wild-type or mutant cyclin H which were clin H and cyclin H mutated at amino acid 315 to alanine or aspartic acid from insect cells infected with recombinant baculo used as substrates. Thus, we have to conclude that viruses were incubated with protein kinase CK2 and [g-32P]ATP. threonine 315 is indeed the CK2 phosphorylation site. As a control wild-type cyclin H was incubated with [g-32P]ATP In order to evaluate the functional consequence of but without the addition of CK2. Phosphorylated proteins were the phosphorylation of cyclin H by CK2 we analysed analysed on a 12.5% SDS polyacrylamide gel followed by autora- the assembly and activity of the cyclin H/cdk7/Mat1 diography. (b) Coomassie brilliant blue stained SDS polyacrylamide gel with wild-type or mutated cyclin H used in the complex where cyclin H was substituted by either of corresponding experiments of (a) the two phosphorylation mutants. The cyclin H mutant T315D mimics a permanent CK2 phosphorylation at position 315 whereas mutant T315A represents a non- phosphorylated form of cyclin H. We first controlled was stained with Coomassie blue. As shown in Figure whether the point mutants might have an influence on 4a for wild-type cyclin H as well as for the two cyclin the assembly of the cyclin H/cdk7/Mat1 complex. To H mutants we found trimeric cyclin H/cdk7/Mat1 produce cyclin H/cdk7/Mat1 kinase complexes we complexes indicating that the point mutations at infected insect cells with recombinant baculoviruses position 315 of cyclin H did not affect complex expressing Mat1, cdk7 and cyclin H in either wild-type formation. or one of the both mutated forms. Cyclin H/cdk7/ In the next experiment we analysed the kinase Mat1 complexes were immunoprecipitated from insect activity of the three different forms of cyclin H cells using GSH-sepharose to immobilize the GST- (cycHwt, cycHT315A or cycHT315D) in complex with tagged form of Mat1. Protein complexes were cdk7 and Mat1 (shown in Figure 4a) with respect to separated on an SDS polyacrylamide gel and the gel the phosphorylation of the CTD peptide substrate.

Oncogene CK2 phosphorylation of cyclin H E Schneider et al 5034 a

Figure 5 Activity assay for the cyclin H/cdk7/Mat1 kinase using cdk2 as a substrate. (a) Equal amounts of the three different cyclin H/cdk7/Mat1 kinase preparations were incubated with [g-32P]ATP and cdk2 which itself was catalytically inactive (GST-cdk2K33R). After phosphorylation cdk2 was analysed on a 15% SDS polyacrylamide gel, followed by autoradiography. (b) Coomassie brilliant blue stained gel of the cyclin H/cdk7/Mat1 complexes used for the phosphorylation reactions shown in (a) Figure 4 Assembly of the cyclin H/cdk7/Mat1 complex and activity assay for the cyclin H/cdk7/Mat1 kinase with the CTD peptide as substrate. (a) Wild-type cyclin H and the phosphorylation Discussion mutants cyclin HT315A and cyclin HT315D were recovered from triple infected insect cells by binding together with cdk7 to GST- tagged Mat1 by GSH Sepharose. One half of the aliquots of Cyclin dependent kinases are the main regulators of the the trimeric complexes was analysed on a 12.5% SDS polyacryla- eukaryotic cell cycle (Johnson and Walker, 1999). They mide gel followed by staining with Coomassie brilliant blue. (b) Equal amounts of the three different cyclin H/cdk7/Mat1 kinase are composed of the regulatory subunits, the cyclins, preparations were incubated with [g-32P]ATP and the synthetic and the catalytic subunits, called cdks. The cyclin H/ CTD peptide as a substrate, which consists of four heptarepeats cdk7 complex normally contains a third subunit of the large subunit of RNA polymerase II. The phosphorylated namely Mat1 which functions as an assembly factor substrate peptides were analysed on a 20% SDS polyacrylamide stabilizing the kinase complex and as a factor which gel, followed by autoradiography regulates the substrate specificity (Devault et al., 1995; Fisher et al., 1995; Tassan et al., 1995). Cyclin dependent kinases are in general regulated by the This peptide consists of 4 out of 28 repeats from the C- assembly of the regulatory cyclin subunits with the terminal domain (CTD) of the large subunit of RNA catalytic cdk subunits, by phosphorylation and depho- polymerase II. We found that the trimeric complex sphorylation of the cdks and by specific inhibitor containing the T315A mutant of cyclin H showed a molecules (for review see: Kaldis, 1999). We recently significantly reduced CTD phosphorylation activity identified p53 as a cyclin H binding protein and compared to the kinase complex containing wild-type furthermore we could show that this interaction led to or the T315D mutant of cyclin H (Figure 4b). The a down-regulation of the cyclin H/cdk7/Mat1 activity same results were obtained when the CTD peptide was with regard to the phosphorylation of the C-terminal exchanged by an alternative well established substrate domain of the large subunit of RNA polymerase II and for the cyclin H/cdk7/Mat1 kinase: namely the protein of the cdk2 kinase (Schneider et al., 1998). Since kinase cdk2 which was used in form of a GST fusion mutant p53His175 failed to down-regulate the cyclin H/ protein carrying a point mutation at position 33 cdk7/Mat1 activity these data indicated that this kind (K?R) responsible for a loss of catalytic activity of negative regulation of the kinase activity is (Figure 5a). The corresponding cyclin H/cdk7/Mat1 exclusively due to a functional wild-type p53 (Schnei- complexes used for the phosphorylation reactions are der et al., 1998). shown in Figure 5b. By repeating the experiment at Cyclins are also phosphoproteins and therefore least four times we always found a lower phosphoryla- potential targets for a regulation by phosphorylation tion (60 – 80%) of cdk2 in case of the cyclin HT315A/ and dephosphorylation. It was indeed recently shown cdk7/Mat1 complex than for cyclin Hwt/cdk7/Mat1 that cyclin H was phosphorylated by the cyclin C/cdk8 complex (set to 100%). These results led us to conclude kinase complex and this phosphorylation repressed that phosphorylation at threonine 315 by CK2 is both the ability of the cyclin H/cdk7/Mat1 kinase in essential for a full kinase activity of the cyclin H/cdk7/ the TFIIH complex to activate transcription and its Mat1 complex with respect to the phosphorylation of CTD kinase activity (Akoulitchev et al., 2000). In the the CTD peptide derived from the C-terminal domain present study we showed that cyclin H is also a of RNA polymerase II as well as to cdk2. substrate for another kinase namely protein kinase

Oncogene CK2 phosphorylation of cyclin H E Schneider et al 5035 CK2. Using deletion mutants (data not shown) as well cible reduction in the kinase activity of the cyclin H/ as point mutants we identified threonine 315 in the C- cdk7/Mat1 complex carrying the cyclin HT315A mutant terminus of cyclin H as the main phosphorylation site instead of wild-type cyclin H. In contrast, in the for protein kinase CK2. These results were further presence of the cyclin HT315D mutant we observed a confirmed by using synthetic peptides as a substrate for strong phosphorylation of cdk2. CK2. We found that only the peptide derived from We have previously shown that the growth suppres- position 311 – 320 of authentic cyclin H including the sor protein p53 plays a role as a direct inhibitor for the potential phosphorylation site at threonine 315 was cyclin H/cdk7/Mat1 complex (Schneider et al., 1998) as phosphorylated by CK2 (data not shown). well as for protein kinase CK2 (Schuster et al., 2001). Most of the substrates for CK2 are only efficiently Inhibition of CK2 was specific for wild-type p53 and phosphorylated by the holoenzyme composed of two for the holoenzyme consisting of two catalytic a- and/ regulatory b-subunits and two catalytic a-ora’- or a’- and two regulatory b-subunits. From a variety of subunits and only a few substrates are known to be different experiments there is ample evidence that CK2 better phosphorylated by the catalytic a-subunit alone is actively implicated in cell proliferation (for review (for review see: Guerra et al., 1999a). Consistent with see: Guerra and Issinger, 1999b; Pinna and Meggio, the majority of CK2 substrates we could show that 1997). Furthermore, it was shown previously that the cyclin H is more efficiently phosphorylated by the carboxy-terminal heptapeptide of the largest subunit of holoenzyme than by the catalytic a-subunit alone. This RNA polymerase serves as an excellent substrate for phosphorylation was efficiently reduced by heparin, a purified CK2 (Angiolillo et al., 1993) and also for potent inhibitor of CK2. cyclin C/cdk8 (Akoulitchev et al., 2000). Thus, CK2 Interestingly, the phosphorylation site at threonine can directly regulate RNA polymerase II by phosphor- 315 is in close vicinity of serine 304 which was shown ylation and indirectly by the activating phosphor- to be phosphorylated by cyclin C/cdk8 (Akoulitchev et ylation of the cyclin H/cdk7/Mat1 kinase complex. The al., 2000). There is an additional cyclin C/cdk8 present data together with previous results led to the phosphorylation site at serine 5 in the N-terminus of following hypothesis. Up-regulation of p53 in a cell cyclin H. It was further shown that phosphorylation of due to various stress situations can result in two effects cyclin H by cyclin C/cdk8 is critical for full kinase with respect to the activity of the cyclin H/cdk7/Mat1 activity and transcriptional activity of the TFIIH complex. First, p53 inhibits the cyclin H/cdk7/Mat1 associated cyclin H/cdk7/Mat1 kinase complex. The activity directly by binding to cyclin H. Second, p53 phosphorylation sites at Thr-315 and at Ser-304 are inhibits the activity of the CK2 holoenzyme thus highly conserved in cyclin H homologues in human, preventing the stimulating phosphorylation of cyclin H, rat, mice and Xenopus laevis (Labbe et al., 1994; which is necessary for its full activity. Both effects lead Makela et al., 1994; gene bank accession number to a down-regulation of the RNA polymerase II AAD46521). Thus, it was an intriguing question dependent transcription as well as to cell cycle arrest. whether CK2 phosphorylation at position 315 might influence cyclin H/cdk7/Mat1 activity, too. By using wild-type cyclin H and the T315D and T315A mutants Materials and methods of cyclin H mimicking permanent phosphorylation or a non-phosphorylated form of cyclin H, respectively, we Plasmids and recombinant baculoviruses found a strong reduction in the cyclin H/cdk7/Mat1 cDNAs of cyclin H, cdk7 and Mat1 were isolated by PCR kinase activity in the presence of the cyclin HT315A from HeLa and human testis cDNA libraries and verified by mutant. In contrast, in the presence of wild-type cyclin sequencing. All newly constructed plasmids used herein were H or the cyclin HT315D mutant we found a strong subcloned from vectors already described (Schneider et al., phosphorylation of the CTD substrate peptide derived 1998). For the bacterial expression of wild-type cyclin H we from the C-terminal domain of the large subunit of used the pQE30 vector (Qiagen) (Schneider et al., 1998). For RNA polymerase II. Interestingly, phospho-specific baculovirus expression cdk7 full length cDNA (Schneider et al., 1998), wild type and point mutants of cyclin H cDNA antibodies which were raised against a short peptide were cloned as PstI/HindIII or BamHI/HindIII fragments, spanning amino acids 311 – 319 (EEEWT(P)DDDL) of respectively, into pBacPAK- His1 thus also allowing the cyclin H were able to recognize baculovirus expressed purification by the N-terminal histidine tag. Mat1 was cloned wild-type cyclin H. In contrast, these antibodies only as GST fusion construct into pAcG1 (Pharmingen, San detected bacterially expressed cyclin H after phosphor- Diego, CA, USA). In an overlap extension PCR procedure ylation by CK2 (data not shown). Thus, wild-type the mutations ACT to GCA and ACT to GAC of codon 315 cyclin H/cdk7/Mat1, which we used for the measure- resulting in an exchange of threonine 315 to alanine or ment of the CTD phosphorylation activity, represents aspartic acid, respectively, were introduced by primer directed at least in part cyclin H in a Thr-315 phosphorylated mutagenesis. Subsequently, mutants were also cloned into form. pBacPAK-His1. All constructs were verified by sequencing. When we used cdk2 instead of the CTD peptide as a substrate for the cyclin H/cdk7/Mat1 complex we Expression and purification of proteins produced in Sf9 cells obtained comparable results. Wild-type cyclin H in Sf9 (Spodoptera frugiperda) cells were grown in Sf900 complex with cdk7 and Mat1 were able to efficiently medium (Gibco, Life Technologies) with 10% foetal calf phosphorylate cdk2. There was a slight but reprodu- serum (FCS) at 278C. For the construction of baculoviruses

Oncogene CK2 phosphorylation of cyclin H E Schneider et al 5036 expressing human proteins the coding cDNAs were cloned (130 mM Tris/HCl, pH 6.8, 0.02% bromophenolblue, 10 mM into baculo transfer vectors (pBacPAK – constructs, pacG1 2-mercaptoethanol, 20% glycerin, 4% SDS) and boiling. – Mat1) and then transferred into baculovirus DNA by Phosphorylated proteins were separated in a 12.5% SDS recombination with Baculogold-DNA as recommended by polyacrylamide gel and subsequently visualized by autoradio- the manufacturer (Pharmingen). The resulting viruses were graphy. Cyclin H/cdk7/Mat1 complexes were produced in Sf9 used to infect Sf9 cells and to express the protein. Insect cells cells and purified by the GST tag of Mat1 or the His-tag of were triple infected with recombinant viruses expressing cdk7, cyclin H as described above. As substrate served 5 mmofa Mat1, and cyclin H either in its wild-type or mutant form. synthetic peptide comprising four times the hepta repeats Forty-eight hours after infection cells were harvested by (YSPTSPS) of the C- terminal domain of the large subunit of centrifugation, washed twice with ice-cold PBS and then lysed human RNA polymerase II or GST-cdk2K33R. Kinase in either a hypotonic buffer (20 mM NaH2PO4, pH 8.0, reactions were performed in 20 mM Tris/HCl, pH 7.8, TM 20 mM NaCl, 0.5 mM DTT, protease inhibitors (complete , 100 mM KCl, 10 mM MgCl2, 0.1% Tween20, 1 mM ATP Roche Mannheim) for purification by the His tag) or in a (unlabelled), 1 mCi [g-32P]ATP using 1 pmole purified trimeric HEPES – buffered solution (10 mM HEPES pH 7.4, 10 mM cyclin H/cdk7/Mat1. The kinase reaction was started by the NaCl, 1 mM EDTA, 1 mM DTT for purification by the GST addition of ATP (labelled and unlabelled) and incubated for tag). For purification by the GST tag cells were lysed at 48C, 30 min at 378C. After stopping the reaction with SDS loading sonicated and cleared by centrifugation. The supernatant was buffer the reaction mixture was analysed by SDS poly- mixed with the glutathion Sepharose (Amersham – Pharma- acrylamide gel electrophoresis. Phosphorylation of the CTD cia) and binding of the GST fusion protein was allowed to peptide or cdk2K33R was visualized by autoradiography. take place for 2 h at 48C. The resin was washed with phosphate buffered saline (PBS). Elution of trimeric cyclin H/ SDS polyacrylamide gel electrophoresis and Western blot cdk7/Mat1 complex or monomeric subunits was achieved by analysis elution with 50 mM Tris/HCl pH 8.0, 5 mM MgCl2, 150 mM NaCl, 1 mM DTT, 5 mM glutathion. For the purification of Proteins were analysed by SDS gel electrophoresis according His tagged proteins expressing cells were extracted for 15 min to the procedure of Laemmli (1970). For Western blot on ice. The extracts were sonicated and cleared by analysis proteins were transferred to a PVDF membrane by centrifugation. The supernatant was diluted with 1/30 volume tank blotting with 20 mM Tris/HCl, pH 8.7, 150 mM glycine 1 M phosphate buffer, pH 8.0 and 2/5 volume 1 M KCl and as transfer buffer. Membranes were blocked in PBS with subsequently incubated with Ni2+ – NTA agarose (Qiagen) 0.1% Tween20 and 5% dry milk for 1 h at room for 1 h on ice. The resin was washed with 20 mM Tris/HCl, temperature. The membrane was incubated with the primary pH 8.0, 300 mM KCl, 10 mM MgCl2, 0.1% Tween 20 and antibody in PBS-Tween 20 with 1% dry milk for another 20 mM imidazole. Proteins were eluted from the resin by hour. The membrane was then washed with PBS-Tween20 increasing the imidazole concentration stepwise between 50 three times before incubating with a peroxidase-coupled and 100 mM. secondary antibody in a dilution of 1 : 30 000 in PBS Tween 20 with 1% dry milk. Signals were developed and visualized by the Lumilight system of Roche Diagnostics (Mannheim, In vitro phosphorylation assays Germany). Recombinant protein kinase CK2 was expressed and purified as described earlier (Appel et al., 1995). Recombinant cyclin H proteins (0.5 mg) from baculovirus infected insect cells were Acknowledgments mixed with recombinant protein kinase CK2 in 20 ml kinase We thank Wolfgang Nastainczyk for peptide synthesis, buffer (50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 5 mM Ju¨ rgen Gu¨ nther for expert technical assistance and Dr MgCl2,1mM DTT) essentially as described (Go¨ tz et al., Claudia Go¨ tz and Dr Michael Faust for many helpful 1996). The phosphorylation reaction was started by the discussions. M Montenarh is supported by grant Mo309/ addition of 1 mCi [g-32P]ATP. After 30 min at 378C the 11-3 from Deutsche Forschungsgemeinschaft and from reaction was stopped by addition of 10 ml SDS loading buffer Fonds der Chemischen Industrie.

References

Adamczewski JP, Rossignol M, Tassan J-P, Nigg EA, El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons Moncollin V and Egly J-M. (1996). EMBO J., 15, 1877 – R, Trent JM, Lin D, Mercer WE, Kinzler KW and 1884. Vogelstein B. (1993). Cell, 75, 817 – 825. Akoulitchev S, Chuikov S and Reinberg D. (2000). Nature, Feaver WJ, Svejstrup JQ, Henry NL and Kornberg RD. 407, 102 – 106. (1994). Cell, 79, 1103 – 1109. Andersen G, Busso D, Poterszman A, Hwang JR, Wurtz JM, Fisher RP, Jin P, Chamberlin HM and Morgan DO. (1995). Ripp R, Thierry JC, Egly JM and Moras D. (1997). EMBO Cell, 83, 47 – 57. J., 16, 958 – 967. Fisher RP and Morgan DO. (1994). Cell, 78, 713 – 724. Angiolillo A, Bramucci M, Marsili V, Panara F, Miano A, Go¨ tz C, Wagner P, Issinger O-G and Montenarh M. (1996). Amici D and Gianfranceschi GL. (1993). Mol. Cell Oncogene, 13, 391 – 398. Biochem., 125, 65 – 76. Guerra B, Boldyreff B, Sarno S, Cesaro L, Issinger OG and Appel K, Wagner P, Boldyreff B, Issinger O-G and Pinna LA. (1999a). Pharmacol. Ther., 82, 303 – 313. Montenarh M. (1995). Oncogene, 11, 1971 – 1978. Guerra B and Issinger OG. (1999b). Electrophoresis, 20, Devault A, Martinez AM, Fesquet D, Labbe JC, Morin N, 391 – 408. Tassan JP, Nigg EA, Cavadore JC and Doree M. (1995). Hathaway GM, Lubben TH and Traugh JA. (1980). J. Biol. EMBO J., 14, 5027 – 5036. Chem., 255, 8038 – 8041.

Oncogene CK2 phosphorylation of cyclin H E Schneider et al 5037 Johnson DG and Walker CL. (1999). Annu. Rev. Pharmacol. Morgan DO. (1995). Nature, 374, 131 – 134. Toxicol., 39, 295 – 312. Pinna LA and Meggio F. (1997). Prog. Cell Cycle Res., 3, Kaldis P. (1999). Cell. Mol. Life Sci., 55, 284 – 296. 77 – 97. KoLJ,ShiehSY,ChenXB,JayaramanL,TamaiK,TayaY, Schaefer L, Roy R, Humbert S, Moncollin V, Vermeulen W, Prives C and Pan ZQ. (1997). Mol. Cell. Biol., 17, 7220 – Hoeijmakers JH, Chambon P and Egly JM. (1993). 7229. Science, 260, 58 – 63. Labbe JC, Martinez AM, Fesquet D, Capony JP, Darbon Schneider E, Montenarh M and Wagner P. (1998). Oncogene, JM, Derancourt J, Devault A, Morin N, Cavadore JC and 17, 2733 – 2742. Doree M. (1994). EMBO J., 13, 5155 – 5164. Schuster N, Go¨ tz C, Faust M, Schneider E, Prowald A, Laemmli UK. (1970). Nature, 227, 680 – 682. Jungbluth A and Montenarh M. (2001). J. Cell. Biochem., Lu H, Fisher RP, Bailey P and Levine AJ. (1997). Mol. Cell. 81, 172 – 181. Biol., 17, 5923 – 5934. Svejstrup JQ, Vichi P and Egly J-M. (1996). Trends Biochem. Makela TP, Tassan JP, Nigg EA, Frutiger S, Hughes GJ and Sci., 21, 346 – 350. Weinberg RA. (1994). Nature, 371, 254 – 257. Tassan JP, Jaquenoud M, Fry AM, Frutiger S, Hughes GJ Marin O, Meggio F and Pinna LA. (1994). Biochem. Biophys. and Nigg EA. (1995). EMBO J., 14, 5608 – 5617. Res. Commun., 198, 898 – 905. Meggio F, Marin O and Pinna LA. (1994). Cell Mol. Biol. Res., 40, 401 – 409.

Oncogene