Title: Substrate specificity of the cdk-activating kinase (CAK) is altered upon association with TFIIH
Abstract: Article1 April 1997free access Substrate specificity of the cdk-activating kinase (CAK) is altered upon association with TFIIH Mireille Rossignol Mireille Rossignol Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France Search for more papers by this author Isabelle Kolb-Cheynel Isabelle Kolb-Cheynel Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France Search for more papers by this author Jean-Marc Egly Corresponding Author Jean-Marc Egly Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France Search for more papers by this author Mireille Rossignol Mireille Rossignol Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France Search for more papers by this author Isabelle Kolb-Cheynel Isabelle Kolb-Cheynel Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France Search for more papers by this author Jean-Marc Egly Corresponding Author Jean-Marc Egly Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France Search for more papers by this author Author Information Mireille Rossignol1, Isabelle Kolb-Cheynel1 and Jean-Marc Egly 1 1Institut de Génétique et de Biologie Moléculaire et Cellulaire, UPR 6520 (CNRS), Unité 184 (INSERM), 1 rue Laurent Fries, BP 163, 67404 Illkirch Cédex, CU de Strasbourg, France The EMBO Journal (1997)16:1628-1637https://doi.org/10.1093/emboj/16.7.1628 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The transcription/DNA repair factor TFIIH consists of nine subunits, several exhibiting known functions: helicase/ATPase, kinase activity and DNA binding. Three subunits of TFIIH, cdk7, cyclin H and MAT1, form a ternary complex, cdk-activating kinase (CAK), found either on its own or as part of TFIIH. In the present work, we demonstrate that purified human CAK complex (free CAK) and recombinant CAK (rCAK) produced in insect cells exhibit a strong preference for the cyclin-dependent kinase 2 (cdk2) over a ctd oligopeptide substrate (which mimics the carboxy-terminal domain of the RNA polymerase II). In contrast, TFIIH preferentially phosphorylates the ctd as well as TFIIEα, but not cdk2. TFIIH was resolved into four subcomplexes: the kinase complex composed of cdk7, cyclin H and MAT1; the core TFIIH which contains XPB, p62, p52, p44 and p34; and two other subcomplexes in which XPD is found associated with either the kinase complex or with the core TFIIH. Using these fractions, we demonstrate that TFIIH lacking the CAK subcomplex completely recovers its transcriptional activity in the presence of free CAK. Furthermore, studies examining the interactions between TFIIH subunits provide evidence that CAK is integrated within TFIIH via XPB and XPD. Introduction TFIIH was the first of the basal transcription factors shown to play a role in cellular activities other than expression of protein-coding genes (reviewed in Hoeijmakers et al., 1996; Moncollin et al., 1997). This may be attributed in part to its multisubunit composition of at least nine peptides, XPB, XPD, p62, p52, p44, cdk7, cyclin H, p34 and MAT1. The two helicase subunits (XPB and XPD) of TFIIH are known to play a role in DNA nucleotide excision repair, thereby providing evidence that transcription is intimately coupled to DNA repair (Schaeffer et al., 1993, 1994; Sung et al., 1993). The importance of this link is evident in three transcription/repair syndromes; xeroderma pigmentosum, Cockayne syndrome and trichothiodystrophy, which are due to defects in genes encoding TFIIH subunits (Vermeulen et al., 1994; Broughton et al., 1995; Hoeijmakers et al., 1996). More recently, cdk7, a kinase known to have a role in the cell cycle, was identified as a component of TFIIH (Roy et al., 1994; Feaver et al., 1994; Shiekhattar et al., 1995; Serizawa et al., 1995). This kinase belongs to a family of cyclin-dependent kinases (cdks) which are key regulatory components that coordinate numerous events such as cell cycle progression, DNA replication and transcription. Cdk activity is controlled by transient association with a specific cyclin, a regulatory subunit or inhibitory protein and activated by phosphorylation in its T-loop domain by a cdk-activating kinase (CAK) (reviewed in Morgan, 1995). Cdk7 forms a ternary complex with cyclin H (Fisher and Morgan, 1994; Mäkelä et al., 1994; Tassan et al., 1994) and MAT1 (Devault et al., 1995; Fisher et al., 1995; Tassan et al., 1995; Adamczewski et al., 1996), and together they form the CAK responsible for activating cdk1, cdk2 and cdk4 (reviewed in Nigg, 1996). The presence of these kinase subunits in TFIIH suggests a role in transcription (Roy et al., 1994; Fisher et al., 1995; Serizawa et al., 1995; Shiekhattar et al., 1995; Adamczewski et al., 1996). The kinase activity of TFIIH is directed towards the carboxy-terminal domain (CTD) of the largest subunit of the RNA polymerase II (RNA pol II) (Feaver et al., 1991; Lu et al., 1992; Serizawa et al., 1992). This phylogenetically highly conserved domain consists of a heptapeptide sequence (YSPTSPS) tandemly repeated up to 52 times. Phosphorylation of the CTD initiates promoter clearance and transcription elongation (Goodrich and Tjian, 1994; Dahmus, 1995). This includes the dissociation of pre-initiation complex proteins. The fact that both the TATA-binding protein (TBP) and TFIIE (Usheva et al., 1992; Maxon et al., 1994) interact with the non-phosphorylated but not with the phosphorylated form of the RNA pol II supports this idea. The cdk7 kinase can phosphorylate a synthetic ctd oligopeptide (Roy et al., 1994) and is therefore thought to be the kinase responsible for the phosphorylation of the RNA pol II CTD. This is supported by in vivo studies in yeast with thermosensitive mutants of Kin28 and Ccl1, the yeast homologue of cdk7 and cyclin H respectively (Feaver et al., 1991; Cismowski et al., 1995; Svejstrup et al., 1996), which demonstrate that these peptides are required for both RNA pol II phosphorylation and transcription (Valay et al., 1995). However, TFIIH containing inactivated CAK complex can support transcription in vitro (Mäkelä et al., 1995). Moreover, depending on the promoter, CTD and its phosphorylation are dispensable for transcription in vitro, but indispensable in vivo (Dahmus, 1995; Gerber et al., 1995). We have demonstrated that the TFIIH-associated kinase complex (cdk7, cyclin H and MAT1) is UV irradiation sensitive but that the non-associated CAK complex (hereafter referred to as free CAK, Adamczewski et al., 1996) is not, suggesting for the first time that cdk7 behaves differently when it is associated with TFIIH. This prompted us to investigate further the possible differences in roles of free CAK versus TFIIH-associated CAK. In this study, we have purified from HeLa cell whole cell extract free CAK composed of cdk7, cyclin H and MAT1, and from baculovirus-infected Sf9 insect cells, a recombinant kinase complex (rCAK). By characterizing their activities, we demonstrate that free CAK, rCAK and TFIIH exhibit different substrate specificities towards a synthetic oligopeptide mimicking the CTD (referred to as ctd) and cdk2, suggesting differential functions dependent on the composition of these complexes. This is supported further in a TFIIH-dependent transcription system, where only the entire TFIIH is capable of phosphorylating the CTD of the RNA pol II and supporting RNA synthesis. As the organization of CAK in TFIIH seems to influence its kinase activity, we characterized several TFIIH subcomplexes and defined the interactions of the kinase complex subunits with the remaining subunits of TFIIH. Results Purification of a CAK complex free of TFIIH To purify the kinase complex (free CAK) composed of cdk7, cyclin H and MAT1, HeLa cell whole cell extract was fractionated on heparin Ultrogel, Sulfopropyl and DEAE columns (see Figure 1A and Materials and methods). The purification of a kinase complex free of TFIIH was followed by Western blot analysis as well as by a ctd kinase assay in which a synthetic peptide (ctd) was used as a substrate (data not shown and Figure 1C). The DEAE-eluted fractions were purified further with antibody against cdk7 (Ab-cdk7) cross-linked to protein A–Sepharose and the elution was performed with an excess of the corresponding epitope peptide (Figure 1A). The eluted fraction contains the three polypeptides, cdk7, cyclin H and MAT1 (Figure 1B and C, WB, lanes 2). The contaminant bands observed at the top of the gel (Figure 1B) are either bovine serum albumin (BSA; 70 kDa), since protein A–Sepharose cross-linked to Ab-cdk7 was saturated with BSA before use, or are artefacts which originate from the silver staining procedure (50–60 kDa). Neither p62 nor XPD was detected in this fraction (Figure 1C, WB, compare lanes 1 and 2) indicating the absence of TFIIH. Furthermore, this complex possesses a ctd kinase activity (Figure 1C, Kinase, lane 2). Together, our results demonstrate that cdk7, cyclin H and MAT1 are associated to form a ternary complex distinct from TFIIH, that possesses a ctd kinase activity. Moreover, the stoichiometry of the three kinase subunits is conserved between TFIIH and the free CAK complex. Figure 1.(A) Scheme of purification of human free CAK and TFIIH. WCE, whole cell extract; AS, ammonium sulfate; PO4, phosphate buffer. The free CAK last step of purification consists of an immunopurification followed by an elution with an excess of cdk7 carboxy-terminal peptide [amino acids 312–329; Pep(312–329)], mimicking the Ab-cdk7 epitope. (B) The eluted fraction was resolved by SDS–PAGE and the gel was silver stained. The position of the molecular weight markers (MW) in kiloDaltons is shown. (C) Free CAK fraction was also analysed by Western blot (WB) and tested for its ctd kinase activity. Cdk7, cyclin H and MAT1 polypeptides are indicated. A purified TFIIH fraction (lane 1) was used as a positive control. Ctd4 indicates the synthetic peptide mimicking four repeats of the CTD of the largest subunit of the RNA polymerase II. Download figure Download PowerPoint Production of a recombinant CAK complex To study further the properties of the CAK complex and its interaction with the other TFIIH subunits, we reconstituted a recombinant CAK (hereafter referred to as rCAK) composed of cdk7, cyclin H and MAT1, in a baculovirus expression vector system. Sf9 insect cells were co-infected with three recombinant baculoviruses encoding cdk7, MAT1 and a histidine-tagged cyclin H (His-cyclin H). The rCAK was purified from the crude Sf9 cell extract via the histidine-tagged cyclin H, using nickel chelate affinity chromatography. The eluted fractions were immunoprecipitated with Ab-cdk7 cross-linked to protein A–Sepharose and the elution was performed with an excess of the corresponding epitope peptide (Figure 2A). The eluted fraction was resolved by SDS–PAGE and analysed by Coomassie staining and Western blot (Figure 2B). As shown in Figure 2B, the eluted fraction contains a highly purified complex composed of the three subunits, cdk7, His-cyclin H and MAT1, which is resolved into two major bands; the first band corresponds to the co-migration of cdk7 (40 kDa) (Figure 2B, lanes 2 and 4) and His-cyclin H (39.8 kDa) (Figure 2B, lanes 2 and 3), and the second band corresponds to MAT1 (32 kDa) (lanes 2 and 4). Figure 2.(A) Scheme of production and purification of the recombinant CAK (rCAK) from Sf9 cells co-infected with baculovirus encoding cdk7, His-cyclin H and MAT1 respectively. In the last step of the purification, rCAK was eluted with an excess of cdk7 carboxy-terminal peptide [Pep(312–329)]. (B) The purified rCAK was analysed by SDS–PAGE followed by either Coomassie blue staining or Western blot (WB). Since cdk7 and His-cyclin H have the same apparent molecular weight, these two polypeptides were revealed separately. Cdk7, His-cyclin H and MAT1 polypeptides are indicated. The position of the molecular weight markers (MW) in kiloDaltons is shown. Download figure Download PowerPoint Free CAK and rCAK phosphorylate TBP To characterize the kinase activity further, we first compared the behaviour of TFIIH, free CAK and rCAK. The nucleotide specificities of the three kinases are identical: ATP, dATP and GTP all act as cofactors for the kinase activity towards ctd or cdk2 in competition assays, while CTP or cAMP had no effect on the phosphorylation of ctd or cdk2 (Roy et al., 1994; data not shown). All three kinases are similarly inhibited, with IC50 values between 10 and 50 μM, by 5,6 dichloro-1-β-D-ribofuranosylbenzimidazole (DRB), a known inhibitor of the phosphorylation of RNA pol II CTD in vivo (Dubois et al., 1994) and TFIIH kinase activity (Yankulov et al., 1995). These results indicate that TFIIH, free CAK and rCAK have the same nucleotide specificity and respond similarly to DRB. To determine whether free and rCAK were able to phosphorylate known substrates of TFIIH, kinase assays were carried out using the recombinant transcription factors TFIIEα and TBP (Ohkuma and Roeder, 1994; Roy et al., 1994; Yankulov et al., 1995). TFIIH, free CAK and rCAK phosphorylate TBP (Figure 3, Kinase, TBP), whereas TFIIEα is only phosphorylated by TFIIH (Figure 3, Kinase, TFIIEα, lane 1) and not by free CAK and rCAK (lanes 2 and 3). To ensure that the three kinase complexes (TFIIH, free CAK and rCAK) were functional, we tested their ability to activate cdk2 in a histone H1 kinase assay. Free CAK (lane 2) and rCAK (lane 3) as well as TFIIH (lane 1) phosphorylate cdk2 and subsequently activate its ability to phosphorylate histone H1 (Figure 3, CAK). Together, these results suggest that CAK gains TFIIEα substrate specificity when associated with TFIIH, without losing its cdk-activating kinase activity. However, we could not exclude the possibility that an unidentified kinase tightly associated with TFIIH could also be responsible for TFIIEα phosphorylation. Figure 3.Comparison of kinase activities between TFIIH, free CAK and rCAK. The cdk7 concentration of each purified fraction (TFIIH: HAP fraction, free CAK) was estimated by Western blot analysis, and adjusted to equivalent values for the kinase assays. Recombinant substrates TFIIEα (100 ng) and TBP (100 ng), overexpressed and purified from Escherichia coli, were tested in the kinase assay (Kinase). The phosphorylation of histone H1 (1 μg) was tested in a CAK assay (CAK). Download figure Download PowerPoint CTD versus cdk2 substrate specificity of free or TFIIH-associated CAK We demonstrated that UV irradiation of living cells partially inhibits TFIIH kinase activity whereas the free CAK is not affected, indicating a potential specificity in the respective enzyme function (Adamczewski et al., 1996). This prompted us to test if, as a function of its state (either free or TFIIH-associated), the kinase complex exhibits a substrate specificity. This study was carried out using an equimolar mixture of cdk2, a substrate implicated in the regulation of the cell cycle, and ctd, a substrate important in transcription. The cdk7 concentration of each purified fraction was estimated by Western blot analysis. The cdk7 concentration of TFIIH and free CAK was adjusted to equivalent values for the following kinase assays. To ensure that the phosphorylated band observed corresponds to cdk2, the gel was Coomassie stained and the radioactive band superimposed with the cdk2 band on the protein gel. Both free CAK and rCAK exhibited a preference for cdk2 over the ctd substrate (Figure 4A, lanes 2 and 3, compare cdk2 with ctd). Quantification of the radioactive signals and calculation of cdk2/ctd ratios illustrate an ∼3-fold higher efficiency in cdk2 phosphorylation by both free and rCAK, in our in vitro experimental conditions (Figure 4A, see cdk2/ctd4 values at the bottom of the panel). In these conditions, we noticed that TFIIH has a preference for the ctd peptide (Figure 4A, lane 1). Figure 4.Substrate preferences of CAK versus TFIIH. (A) Differential phosphorylation of cdk2 and ctd4 by TFIIH, free CAK and rCAK. The kinase activity of TFIIH, free CAK and rCAK was assayed using a mixture containing both oligopeptide ctd4 (1 μg) and GST–cdk2 (50 ng). Quantification of the radioactive signals was performed using a PhosphorImager, and calculations of the cdk2/ctd ratios are indicated at the bottom of the figure. (B) The phosphorylation of RNA pol II (Kinase) was tested in an in vitro transcription assay containing all the basal transcription factors and, when indicated (at the top of the figure), TFIIH (HAP fraction), free CAK and rCAK. IIA and IIO are designated as the non-phosphorylated and the hyperphosphorylated forms of RNA pol II respectively. The production of the 309 nucleotide transcript (309 nt) was visualized in a transcription assay (Transcription). Download figure Download PowerPoint Given the preference of the free CAK towards cdk2, we investigated the ability of the three kinase complexes to phosphorylate the CTD of the RNA pol II, a natural substrate of TFIIH (Lu et al., 1992; Serizawa et al., 1992). Knowing that the CTD phosphorylation was greatly enhanced when RNA pol II was integrated in the transcription complex, we set up the following assay. The phosphorylation of RNA pol II was investigated in an in vitro reconstituted transcription assay containing, in addition to RNA pol II, the general transcription factors TFIIA, TFIIB, TBP, TFIIE and TFIIF, a linear DNA template containing the adenovirus major late promoter (AdMLP) and (as indicated at the top of Figure 4B) either TFIIH, free CAK or rCAK. The phosphorylation of RNA pol II was visualized by Western blot analysis using an antibody raised against its largest subunit (Dubois et al., 1994, see also Materials and methods). Only the TFIIH fraction resulted in the shift from the lower molecular weight RNA pol IIA (IIA, the non-phosphorylated form), to the higher molecular weight RNA pol IIO (IIO, the hyperphosphorylated form, Figure 4B, lane 1). This shift is due exclusively to TFIIH and not to some contamination present in the other transcription factors (lane 4). Although free CAK and rCAK use ctd synthetic peptide as a substrate (Figure 4A), neither of them lead to this shift (Figure 4B, lanes 2 and 3). When the phosphorylation of the RNA pol II was analysed under the same conditions but in the presence of radiolabelled ATP, incorporation of 32P was observed only in the presence of TFIIH. The labelled band corresponded to the IIO form of the RNA pol II (data not shown). To investigate whether the presence of transcription factors and DNA template could inhibit the free CAK and rCAK CTD kinase activity, the RNA pol II phosphorylation was tested in a simple kinase assay devoid of any basal transcription factors and DNA template. Under these conditions, only incubation with TFIIH leads to the IIO form of the RNA pol II, although at a very low level, whereas incubation with free CAK and rCAK do not (data not shown). Production of the AdMLP-specific transcript was detected only in the presence of TFIIH (lane 1) but not with the free CAK (lane 2) or rCAK fractions (lane 3) (Figure 4B, Transcription). These results demonstrate that CAK does not support transcription in the absence of the other TFIIH subunits. Together, our data show that CAK requires an association with TFIIH to phosphorylate the CTD of the RNA pol II and to play a role in transcription. TFIIH can be resolved into different subcomplexes Glycerol gradient analysis performed under high salt conditions indicated that the ternary kinase complex containing cdk7, cyclin H and MAT1 could be resolved from TFIIH (Adamczewski et al., 1996). We have also observed that XPD could be partially dissociated from the other TFIIH subunits (Roy et al., 1994; Schaeffer et al., 1994). To characterize the organization of TFIIH further, we set up the following assay according to the scheme outlined in Figure 5A. A purified TFIIH fraction was treated with high salt (1.2 M KCl) and then immunoprecipitated in a buffer containing 1.2 M KCl, with an excess of either Ab-cdk7, Ab-p62 or Ab-XPD (see Materials and methods). The corresponding supernatant (Sn) and the proteins adsorbed on the three immunoadsorbants (Bd) were analysed by Western blot. The supernatant from a fraction immunoprecipitated with Ab-cdk7 (Sn/Ab-cdk7) contained all the TFIIH subunits, XPB, some XPD, p62, p52, p44 and p34, but not cdk7, cyclin H or MAT1 (Figure 5B, compare lanes 1 and 2). Analysis of the proteins bound to Ab-cdk7 cross-linked to protein A–Sepharose (Bd/Ab-cdk7) demonstrated that Ab-cdk7 immunoprecipitates not only cdk7, cyclin H and MAT1 but also some XPD (Figure 5C, lane 2). This indicates that Ab-cdk7 is able to immunoprecipitate two subcomplexes: one containing, in addition to cdk7, the two subunits of the CAK complex, cyclin H and MAT1, and a second complex, CAK–XPD, which contains the three CAK subunits in addition to a fourth subunit XPD. Analysis of the Sn/Ab-p62 fraction revealed the absence of not only p62 but also p52, p44, p34 and the majority of XPB (Figure 5B, lane 3), thus demonstrating that these five subunits strongly interact with each other. It should be noted that part of XPD also immunoprecipitates with the five other subunits (XPB, p62, p52, p44 and p34; see Figure 5C, lane 3). These results indicate a tight association between XPB, p62, p52, p44 and p34, making up a large complex also called core TFIIH. In the Sn/Ab-XPD fraction, only XPD was missing (Figure 5B, lane 4). Furthermore, Ab-XPD immunoprecipitates cdk7, a subunit belonging to the CAK complex, as well as p62, a subunit of the core TFIIH (Figure 5C, lane 4). Together, these results indicate that TFIIH can be resolved into two major complexes: the core TFIIH, which contains XPB, p62, p52, p44 and p34, and the CAK complex, which contains cdk7, cyclin H and MAT1. XPD is found in two subcomplexes, associated with the core of TFIIH (core TFIIH–XPD) and with the CAK complex (CAK–XPD). Figure 5.Analysis of TFIIH subcomplexes. (A) Scheme of TFIIH dissociation by 1.2 M KCl treatment and immunoprecipitation with Ab-cdk7, Ab–p62 or Ab-XPD. (B) The supernatants (Sn) were collected and analysed by Western blot for the presence of all TFIIH subunits. (C) The proteins remaining on the beads (Bd) after extensive washing were analysed by Western blot using Ab-XPD, Ab-p62 (subunit of the core TFIIH) and Ab–cdk7 (subunit of the kinase complex). Download figure Download PowerPoint Protein–protein interactions between the kinase complex and other subunits of TFIIH To understand how the CAK complex is integrated into TFIIH, we identified the interactions between the kinase complex subunits and the other subunits of TFIIH. The cDNAs encoding XPB, XPD, p62, p44, p34, cdk7, His-cyclin H and MAT1 were inserted into baculovirus expression vectors (see Materials and methods). Each subunit of TFIIH, alone or in combination with each of the kinase complex subunits, was co-expressed in Sf9 cells. Sf9 cell crude extracts were made and protein expression tested by Western blot analysis using the appropriate antibodies raised against TFIIH polypeptides (data not shown). Immunoprecipitation with Ab-cdk7 bound to protein A–Sepharose was performed on crude extracts containing either cdk7 with the TFIIH subunit being tested (XPB, XPD, p62, p44, p34, cyclin H and MAT1; Figure 6, lanes 2, 4, 6, 8, 10, 12, 14 and 15) or, as a negative control, the subunit alone (lanes 1, 3, 5, 7, 9, 11 and 13). The interactions between cyclin H or MAT1 and the TFIIH subunits were also carried out according to the same scheme presented for cdk7 in Figure 6, and the results obtained are summarized in Table I. Data obtained from at least four sets of independent experiments allow us to distinguish three categories of interaction: +, − and +/−. A (+) indicates an interaction reflecting similar levels of each protein tested on a Western blot (for example see Figure 6, lane 2), a (−) indicates no detectable interaction (lane 10) and a (+/−) reflects a weak interaction, indicated by the presence of a strong signal for one partner and a weak signal for the second protein (lane 6). These experiments demonstrate that cdk7 interacts not only with cyclin H as previously observed (Fisher and Morgan, 1994; Tassan et al., 1995; Adamczewski et al., 1996) but also with MAT1 (Figure 6, lanes 13 and 15), whereas no interaction was detected between cyclin H and MAT1. Interactions were also detected between cdk7 and XPB, MAT1 and XPB, and MAT1 and XPD (Figure 6, lanes 2 and Table I). We also detected a weak interaction between cdk7 and p62, cdk7 and p44, and cyclin H and XPB (Figure 6, lanes 6 and 8 and Table I). Since the level of human recombinant TFIIH subunits overexpressed in Sf9 cells greatly exceeds the concentration of endogenously expressed TFIIH subunits (data not shown), the interactions observed are most likely not mediated by insect cell proteins and thus could be considered as direct interactions. Note that the same interactions were observed when individually expressed TFIIH subunits were incubated together before immunoprecipitation (data not shown). Figure 6.Interaction of cdk7 with the other subunits of TFIIH. Sf9 cells were (co)infected with recombinant baculoviruses expressing TFIIH subunits individually or in combination with cdk7 as indicated at the top of the figure. Sf9 crude extracts were made and immunoprecipitated using Ab-cdk7 cross-linked to protein A–Sepharose. After extensive washing, the proteins absorbed were then analysed by Western blot with antibodies generated towards the various TFIIH subunits. TFIIH fractions (M) were run in parallel in order to determine the expected migration of the various TFIIH subunits. Since His-cyclin H (lane 14) and cdk7 (lane 15) have the same apparent molecular weight, these two polypeptides were immunoblotted separately. Download figure Download PowerPoint Table 1. Qualitative summary of the interactions between the kinase complex subunits themselves and the subunits of TFIIH Cdk7 Cyclin H MAT1 rCAK XPB + +/− +a + XPD − − + + p62 +/− − − +/− p44 +/− − − +/− Cdk7 nt +a + nt Cyclin H +a nt − nt p34 − − − − MAT1 + − nt nt The averages of at least four independent experiments, similar to the one presented in Figure 6, are shown. +, +/− and − correspond to detected, very weak (but detectable) and no interaction between the kinase subunits and TFIIH subunits. nt, indicates that the interaction was not tested. a The interaction was also positive in the two-hybrid system. Knowing that cdk7, cyclin H and MAT1 form a ternary complex which is part of TFIIH, we also investigated the interactions between this ternary complex and the other subunits of TFIIH in vitro. The kinase complex (cdk7, cyclin H and MAT1) was produced in Sf9 cells as was either XPB, XPD, p62, p44 or p34. Crude cell extracts containing approximately equal amounts of the kinase complex and of the additional subunit being tested were mixed, pre-incubated and immunoprecipitated with Ab-cdk7 (Table I). Similar results were also obtained when the immunoprecipitations were performed with the antibodies raised against the other subunits being tested (Table I). The interaction between the kinase complex and the other TFIIH subunits takes place via the two helicases, XPB and XPD, and also, but at a very low level, via two subunits of the core TFIIH, p62 and p44. Free CAK and rCAK restore the transcription activity of TFIIH devoid of CAK We subsequently investigated the effect of the CAK complex on TFIIH transcription activity. A highly purified reconstituted transcription assay lacking TFIIH, in which a DNA containing the AdMLP serves as template, was used. Transcription was performed with a complete TFIIH or with a TFIIH lacking CAK (Sn/Ab-cdk7; Figure 5B, lane 2). The transcription activity of TFIIH lacking the CAK complex was decreased compared with the activity of the complete TFIIH (Figure 7A, compare lanes 1 and 2). Interestingly, when either free CAK or rCAK is added to the TFIIH lacking the CAK, the transcription activity is almost completely restored (Figure 7A, compare lane 1 with lanes 5 and 6). Quantification of the radioactive signals demonstrates that TFIIH devoid of CAK possesses 30% of the normal TFIIH transcription activity, and addition of either free CAK or rCAK allows the recovery of 80% of this activity. The residual transcription activity observed with TFIIH lacking CAK is not due to the presence of residual CAK since the Sn/Ab-cdk7 containing the TFIIH devoid of CAK does not exhibit any kinase activity towards the ctd oligopeptide (Figure 7B, compare lanes 1 and 2). Together, these results demonstrate: first, that a TFIIH lacking CAK can support in vitro transcription from the AdMLP; second, that although not absolutely necessary for the transcri