Introduction

Progression through the cell cycle is mediated by the sequential activation of cyclin-dependent kinases (Cdks), which phosphorylate substrates critical for advancing the cell cycle. Cdk activity is regulated by cyclin binding, Cdk inhibitors, proteolysis, localization and phosphorylation (reviewed in Morgan, 1997). Several cell cycle Cdks require phosphorylation of a conserved threonine residue within the T-loop for full activity (reviewed in Kaldis, 1999). The enzymes that catalyze this activation are known as Cdk-activating kinases or Caks. Biochemical purification of an activity from mammalian and Xenopus lysates capable of activating cdc2 in vitro (Solomon et al., 1992) led to the identification of a trimeric complex of the MO15/Cdk7 kinase together with cyclin H and a third subunit termed Mat1 (ménage-à-trois; reviewed in Nigg, 1996). In vitro, the mammalian Cdk7–cyclin H–Mat1 trimer can phosphorylate and activate Cdc2/Cdk1, Cdk2, Cdk3, Cdk4 and Cdk6 in complex with their cognate cyclin partners (reviewed in Kaldis, 1999).

Shortly after their identification in Cak fractions, Cdk7–cyclin H–Mat1, along with the homologous complex Kin28–Ccl1–Tfb3 in budding yeast Saccharomyces cerevisiae, were found to be components of the nine-subunit basal transcription factor TFIIH (Feaver et al., 1994; Mäkelä et al., 1995; Serizawa et al., 1995; Shiekhattar et al., 1995). TFIIH provides several catalytic functions in mediating both basal transcription by RNA polymerase II (pol II) and transcription-coupled nucleotide excision repair (see Tirode et al., 1999, and references therein). These include the helicase activity afforded by the XPB and XPD subunits, which are involved in promoter melting prior to transcription initiation, in addition to the kinase activity provided by Cdk7–cyclin H–Mat1. TFIIH kinase activity is believed to be directed primarily at Ser5 of the repeated heptapeptide consensus sequence YSPTSPS (Gebara et al., 1997; Hengartner et al., 1998; Sun et al., 1998; Trigon et al., 1998), which comprises the C-terminal domain (CTD) of the large subunit of pol II. Phosphorylation of the CTD during pol II-mediated transcription has been shown to be critical in the recruitment of proteins necessary for proper processing of the nascent transcript (McCracken et al., 1997; Hirose and Manley, 1998; Hirose et al., 1999; Misteli and Spector, 1999; Komarnitsky et al., 2000; Rodriguez et al., 2000).

The functions of Mat1 in the trimeric Cdk7–cyclin H–Mat1 kinase and as a subunit of TFIIH are not well understood. Mat1 enhances complex formation of cyclin H and Cdk7 in the absence of Cdk7 T-loop phosphorylation (Devault et al., 1995; Fisher et al., 1995; Tassan et al., 1995; Martinez et al., 1997). Several studies have suggested a role for Mat1 in the control of substrate specificity whereby Mat1 shifts the kinase activity of dimeric Cdk7–cyclin H from Cdk substrates towards CTD substrates in vitro (Inamoto et al., 1997; Rossignol et al., 1997; Yankulov and Bentley, 1997). Cdk7–cyclin H–Mat1 has been shown to have a wide range of additional in vitro substrates, and phosphorylation of several of these appears to be Mat1 dependent and is likely to occur within the context of TFIIH. These include p53 (Ko et al., 1997), pRb (Wu et al., 2001), the POU domains of the Oct factors (Inamoto et al., 1997), retinoic acid receptor- (Rochette-Egly et al., 1997) and estrogen receptor- (Chen et al., 2000).

Primary sequence analysis indicates that Mat1 contains two conserved structural domains: a canonical RING finger domain and a coiled-coil domain. Structure–function mapping of Mat1 has suggested that the N-terminally located RING finger domain is associated with TFIIH-mediated transcriptional activation (Busso et al., 2000). The central coiled-coil domain is involved in establishing contacts with other TFIIH subunits, primarily with the XPD helicase (Busso et al., 2000). C-terminal sequences appear to be sufficient to mediate the assembly of Cdk7–cyclin H–Mat1 trimers (Tassan et al., 1995; Busso et al., 2000).

The two principal pathways in which Cdk7–cyclin H–Mat1 are suggested to act, namely Cdk activation and pol II-mediated transcription, are both essential for the viability of the cell. Studies have shown that homologous molecules in non-mammalian species are essential genes, and thus a clear dissection of the functions of Cdk7–cyclin H–Mat1 in a physiological setting has proven difficult (Valay et al., 1995; Faye et al., 1997; Larochelle et al., 1998). Compounding this problem is the fact that it has become apparent that differences in the actual in vivo capacities of homologous kinases from different species exist. Notably, the Kin28–Ccl1–Tfb3 complex of S.cerevisiae has been shown to function only in TFIIH-mediated transcription (Cismowski et al., 1995; Valay et al., 1995; Holstege et al., 1998) while Cdk activation is provided by a separate monomeric kinase named Cak1/Civ1 (Espinoza et al., 1996; Kaldis et al., 1996; Thuret et al., 1996; Kimmelman et al., 1999). In Schizosaccharomyces pombe, on the other hand, both the Cdk7-related Mcs6 kinase complex and a single-subunit kinase Csk1 activate Cdks in vivo (Hermand et al., 1998; Lee et al., 1999; Hermand et al., 2001).

To date, the best in vivo evidence that Cdk7–cyclin H–Mat1 functions as a Cak in metazoan species comes from experiments in Drosophila that utilized a temperature-sensitive allele of Cdk7 (DmCdk7) to show that activation of mitotic Cdk–cyclins was impeded at the restrictive temperature (Larochelle et al., 1998). However, a separate study utilizing a dominant-negative allele of DmCdk7 was unable to detect defects in Cak activity and instead described transcriptional defects (Leclerc et al., 2000). Although genetic studies in mammalian systems have not yet been reported, the recent biochemical characterization of a potential budding yeast Cak1/Civ1 homolog from mammalian cells (Nagahara et al., 1999; Kaldis and Solomon, 2000) has suggested that mammalian Cdk activation in vivo could be mediated by a single-subunit Cak in addition to, or perhaps in place of, Cdk7–cyclin H–Mat1. These putative kinases are unlikely to represent close sequence homologs of Cak1 based on the lack of evident Cak1 homologs in currently available mammalian databases.

Thus the actual in vivo functions of the mammalian Cdk7–cyclin H–Mat1 kinase remain elusive. In this report, we have attempted to add some solvency to this issue by generating a loss-of-function allele of Mat1 in mouse to address whether the Cdk7–cyclin H–Mat1 kinase functions in pol II-mediated transcription, cell cycle progression or both.

Results

Targeted disruption of the murine Mat1 gene

Mat1 genomic sequences were isolated from a 129-Sv library using a full-length human MAT1 cDNA as a probe. Restriction analysis and sequencing of several overlapping clones revealed that a single exon of Mat1 spanning 153 nucleotides of the cDNA (nucleotides 242–394 in DDBJ/EMBL/GenBank accession No. U35249) had been isolated. This exon encodes 51 N-terminally located amino acids encompassing most of the RING finger domain. A replacement-type targeting vector (Figure 1A) was constructed and electroporated into embryonic stem (ES) cells. Three clones were isolated (of 800 screened) that were confirmed to be targeted correctly by Southern blotting with both 5' and 3' external probes (Figure 1B). Targeted cells were then injected into BL/6 blastocysts and several of the resulting chimeras were found to transmit the targeted allele through the germline (Figure 1C).

Splicing around targeted exons has been demonstrated with other loci and in this case could lead potentially to the production of an in-frame Mat1 transcript. To investigate whether a truncated protein might be synthesized by the targeted allele, we performed western blot analysis on total lysates obtained from ES cells heterozygous for the targeted allele using an antibody against Mat1 (Figure 1D). No truncated Mat1 proteins were observed in the ES cell lysates, suggesting that the targeted allele represents a null allele of Mat1.

Disruption of Mat1 leads to early embryonic lethality

Mat1 heterozygous (Mat1^+/-) intercrosses were set to ascertain the viability of Mat1 null homozygotes (Mat1^-/-). PCR genotyping analysis of 104 adult offspring indicated that while both Mat1^+/+ and Mat1^+/- animals were observed at the expected frequencies, no homozygote null animals were obtained (Table I). This demonstrated that disruption of Mat1 leads to embryonic lethality. Dissection and genotyping of embryos at embryonic day E10.5, E9.0 and E7.5 of development also failed to identify viable homozygotes. In contrast, blastocysts at E4.0 showed close to Mendelian ratios for all genotypes, indicating that lethality probably occurred after implantation but before gastrulation (Table I). Indeed, the observation of resorbing embryos and empty decidua in roughly Mendelian numbers at later times of gestation suggests that Mat1^-/- embryos die shortly after implantation.

As Mat1 is an essential gene in S.cerevisiae, deletion of mammalian Mat1 might also be expected to compromise cell viability (Faye et al., 1997). Thus it was surprising that early embryogenesis would proceed in a genetically Mat1 null background. We reasoned that survival of Mat1^-/- embryos to the implantation stage of development might be due to maternally provided Mat1. In order to test this hypothesis, embryos obtained from Mat1^+/- intercrosses were isolated at eight-cell, 16-cell and blastocyst stages and immunostained with an anti-Mat1 antibody. Immunofluorescence analysis demonstrated decreasing Mat1 signal in approximately one-quarter of the embryos as development proceeded such that by the blastocyst stage, Mat1 immunoreactivity was barely observable (Figure 2). Subsequent PCR analysis revealed that embryos with diminishing signal genotyped as Mat1^-/- (Figure 2 and data not shown). The timing of the lethality in Mat1^-/- embryos therefore probably reflects the depletion of maternal Mat1 protein below threshold levels required to sustain an essential function.

Interestingly, Mat1^-/- blastocysts were phenotypically indistinguishable from wild-type controls, demonstrating that Mat1^-/- cells have the developmental capacity to differentiate into both inner cell mass (ICM) and trophectodermal lineages. TUNEL analysis of blastocyst stage embryos derived from Mat1^+/- intercrosses showed no differences in either the ICM or trophectoderm cells of Mat1^-/- and wild-type controls (data not shown). This indicated that pre-implantation embryos severely depleted of Mat1 protein stores were not committed rapidly to an apoptotic cell fate at this stage.

Mat1 is required for the survival of mitotic but not post-mitotic lineages

To examine more directly the role of Mat1 in cell proliferation and differentiation, we followed the development of pre-implantation embryos derived from Mat1^+/- intercrosses in culture. Regardless of genotype, most embryos hatched from the zona pelucida and formed blastocyst outgrowths, which were maintained in culture for 7 days. Both Mat1^+/- and Mat1^+/+ outgrowths were characterized by the establishment of a rapidly proliferating cluster of ICM-derived cells growing on a base of trophectoderm-derived trophoblast giant cells that had attached to the culture dish (Figure 3A and B). After 4–5 days in culture, a subpopulation of cells, presumably representing endodermal lineages by morphology, was often observed to differentiate from the ICM and migrate away from the wild-type outgrowths.

In contrast, the Mat1^-/- embryos gave rise to outgrowths with trophectodermal cells yet failed to develop proliferative ICM cells and, as a consequence, endodermal derivatives (Figure 3A and B). Cells derived from the ICM of the blastocyst would often be observed associated with the remnants of the zona pelucida in Mat1^-/- outgrowths, but these cells were never noted to proliferate. Hoechst staining of these cells revealed condensed, fragmented nuclei, suggesting that the Mat1^-/- ICM cells underwent apoptosis during the first few days in culture (data not shown). The trophoblast cells established in the Mat1^-/- outgrowths, on the other hand, were indistinguishable from wild-type counterparts by light microscopy over the first few days in culture (Figure 3A). However, as the cells were maintained further in culture, the Mat1^-/- trophoblast cells were noted to be smaller than the controls. We determined that this was not a secondary phenotype resulting from the absence of ICM lineages in the Mat1^-/- outgrowths, as control outgrowths dissected of their ICM early in culture developed giant cells comparable to those with an intact ICM.

In order to address the function of Mat1 beyond embryonic lethality, we took advantage of the blastocyst outgrowth system as a means of generating a population of trophoblast giant cells on coverslips (20–40 cells per outgrowth). Genotyping of outgrown cells was accomplished by staining with Mat1 antibodies (Figure 3C), which in all cases analyzed was in agreement with PCR genotyping of ICM cells dissected away from wild-type outgrowths. It should be noted that genotyping by immunofluorescence does not distinguish between Mat1^+/+ and Mat1^+/-, and so outgrowths genotyped in this manner are referred to collectively as wild-type (Mat1^wt).

Mat1 specifically regulates the steady-state levels of Cdk7 and cyclin H

Mat1 has been suggested to act as an assembly factor for the Cdk7–cyclin H–Mat1 trimer (Devault et al., 1995; Fisher et al., 1995; Tassan et al., 1995). We therefore wanted to address whether or not the expression of Cdk7 and cyclin H was affected in Mat1^-/- cells. In order to do this, outgrown trophoblast cells were subjected to immunofluorescence with antibodies specific for Cdk7 and cyclin H (Figure 4A). Analysis of wild-type cells revealed strong nuclear staining of both Cdk7 and cyclin H (Figure 4A). Mat1^-/- cells, however, consistently revealed that the signals for both Cdk7 and cyclin H were severely diminished (Figure 4A). In order to be sure that this was not due to epitope masking in the Mat1^-/- cells, immunofluorescence with additional antibodies directed at Cdk7 and cyclin H was performed and found to confirm our initial observations (data not shown).

In S.cerevisiae, strains harboring temperature-sensitive alleles of the Mat1 homolog (TFB3) exhibit severely compromised transcription upon shifting to the non-permissive temperature (Faye et al., 1997). We therefore wanted to determine whether the diminished levels of Cdk7 and cyclin H were specific to Mat1 loss or instead a manifestation of a more general transcriptional deregulation. Steady-state transcription can be assayed indirectly by monitoring protein expression, particularly when assaying the levels of proteins with a rapid turnover. We therefore analyzed cells derived from blastocyst outgrowths for expression of several proteins with variable turnover rates by immunofluorescence. These included proliferating cell nuclear antigen (PCNA), p53, Cdk2, cyclin E, cyclin D1, Cdk6 and the large subunit of pol II (Figures 4B and 5A). We found that the expression of all of these proteins was comparable in both Mat1^-/- and Mat1^wt cells. These observations suggest that the loss of mammalian Mat1 does not deregulate transcription or translation globally and that Mat1 specifically regulates the steady-state levels of Cdk7 and cyclin H in vivo.

Mat1 modulates pol II CTD phosphorylation

Many studies have implicated the CTD of the large subunit of RNA polymerase II as a substrate of TFIIH kinase activity. We therefore wanted to determine whether the phosphorylation status of the CTD was affected in Mat1^-/- cells. This was accomplished by immunofluorescence analysis utilizing monoclonal antibodies recognizing specific phospho-epitopes of pol II (Figure 5).

Using an antibody (8WG16) that is specific for non-phosphorylated Ser2 of the CTD but that also recognizes partially phosphorylated pol II (Thompson et al., 1989), we were able to visualize the majority of cellular pol II. We found that both Mat1^wt and Mat1^-/- nuclei immunostained comparably throughout the nucleus, excluding nucleolar compartments (Figure 5A). This observation indicated that expression of pol II was unaffected by the loss of Mat1.

Phosphorylation of Ser5 of the heptapeptide repeat was visualized using an antibody (H14) specific for this phospho-epitope (Patturajan et al., 1998). Immuno staining of Mat1 wild-type cells with H14 revealed fine punctate immunoreactivity throughout the nucleus (Figure 5B and C). Individual cells of an outgrowth often showed varying degrees of intensity of H14 signal, with some cells exhibiting very intense staining (Figure 5B and C). In contrast, Mat1^-/- cells revealed a considerable decrease in H14 immunoreactivity (Figure 5B and C). Although some degree of variability in H14 staining was also noted in Mat1^-/- cells, we never observed any cells exhibiting the high levels of Ser5 immunoreactivity that had been noted in wild-type cells.

We then looked at the phosphorylation status of Ser2 using an antibody (H5) specific for this phospho-epitope (Patturajan et al., 1998). As had been observed with H14 staining, Mat1^wt cells exhibited a finely particulate staining throughout the nucleoplasm that, unlike the H14 staining, did not vary significantly from one cell to another (Figure 5D and E). Immunostaining of Mat1^-/- cells with H5 revealed a dramatic decrease in signal when compared with Mat1^wt control cells (Figure 5D and E). Additionally, in 30% of the Mat1^-/- cells (Figure 5D and E; 67 of 231 cells scored), we observed intense H5 immunoreactivity localizing to 20–30 discreet nuclear domains. A similar staining pattern in Mat1^wt control cells was only noted in a single nucleus of several hundred scored.

Functional de novo transcription and translation in Mat1^-/- cells

The apparent discrepancy between our observation that multiple proteins were detected at wild-type levels and the observed reduction in pol II CTD phosphorylation prompted us to assay more directly the ability of Mat1^-/- cells to engage in de novo transcription and translation. To this end, a cytomegalovirus (CMV) promoter-driven green fluorescent protein (GFP) expression plasmid was microinjected into outgrown cells at day 6.0 in culture (equivalent to 10.5 days post-coitum). Injected cells were returned to an incubator and assayed for GFP expression by fluorescence microscopy the following day. Regardless of genotype, all microinjected cells exhibited strong expression of the GFP protein (Figure 6). This result showed that Mat1^-/- cells were comparable with control cells in their ability to engage in de novo transcription and translation of the CMV-GFP expression plasmid.

Mat1 is required for S phase entry in trophoblast giant cells

DNA synthesis in trophoblast giant cells occurs by endoreduplication whereby successive rounds of G and S phases proceed without intervening mitosis (Barlow et al., 1972; Gardner, 1983). As a result, endocycling giant cells acquire vast quantities of DNA in their nuclei, thereby becoming polyploid (Varmuza et al., 1988). Throughout our analysis of outgrown cells, it was noted that Mat1^-/- nuclei were consistently smaller in size and stained less intensely with the DNA stain Hoechst 33342 than control nuclei (compare Mat1^-/- and Mat1^wt nuclei in Figures 3,4,5,6 and inset in Figure 7). These observations suggested that the Mat1^-/- cells had reduced amounts of DNA in their nuclei.

In order to examine this possibility in detail, the DNA content of outgrown cells and control mouse embryo fibroblasts (MEFs) was quantified by measuring the fluorescence intensity of Hoechst 33342-stained nuclei from normalized digital images [Figure 7; n(MEF) = 3146, n(Mat1^-/-) = 171, n(Mat1^wt) = 362]. This analysis revealed the DNA content of the Mat1^-/- cells to be heterogeneous and was on average greater than the 2C and 4C amounts measured for the MEFs (Figure 7). This indicated that the majority of Mat1^-/- cells had engaged in endoreduplication and become polyploid. However, comparison with Mat1^wt cells uncovered a dramatic reduction of DNA content in the Mat1^-/- cells (Figure 7). This observation indicated that although the Mat1^-/- trophoblasts had entered cycles of endoreduplication, they were nonetheless compromised in some aspect of the endocycle.

We reasoned that the lower DNA content in Mat1^-/- trophoblasts could reflect either that the endocycle was progressing more slowly than in Mat1^wt counterparts or, alternatively, it could reflect an arrest of the endocycle. To differentiate between these possibilities, E3.5 blastocysts and blastocyst outgrowths (day 4.5 in culture equivalent to 9 days post-coitum) were cultured in the presence of bromodeoxyuridine (BrdU; for 15 and 46 h, respectively) and then immunostained with an antibody specific for BrdU to identify cells that had undergone DNA synthesis during the labeling period (Figure 8). Incorporation of BrdU into blastocyst stage nuclei revealed no difference between the Mat1^-/- and Mat1^wt embryos, both of which showed extensive labeling of ICM and trophectoderm cells (Figure 8A). In contrast, while the vast majority of giant cells readily incorporated BrdU in control outgrowths, no BrdU-positive cells were detected in the Mat1^-/- outgrowths (Figure 8B). These results demonstrate that the Mat1^-/- cells were unable to enter S phase and were arrested in the endocycle.

Discussion

Mat1 is an essential gene

Studies in yeast and Drosophila have indicated that Cdk7–cyclin H–Mat1 homologs are essential genes (Valay et al., 1995; Faye et al., 1997; Larochelle et al., 1998). It was therefore surprising that although disruption of the murine Mat1 gene leads to embryonic lethality, the timing of the lethality was relatively advanced in terms of murine development. We observed that Mat1^-/- embryos inherit a reservoir of maternal Mat1 protein (and possibly mRNA) that diminishes with time as development proceeds. Depletion of maternal Mat1 is likely to reflect the half-life of Mat1 protein as well as the fact that the protein is diluted to ever increasing numbers of nuclei as embryonic cells divide. The coincidence of the peri-implantation lethality of Mat1^-/- embryos with the depletion of observable Mat1 therefore suggests that protein levels are reduced to a point no longer capable of sustaining a vital Mat1 function, and lethality ensues.

Mat1 is required for mitotic proliferation and for endoreduplication: implications for Cak activity

Our blastocyst outgrowth experiments indicated that mitotic ICM cells required Mat1 for survival while post-mitotic trophoblast giant cells did not. This suggests that Mat1 is essential for some aspect of the mitotic cell cycle. As Drosophila Cdk7 has been demonstrated to be required for activation of mitotic Cdk–cyclins (Larochelle et al., 1998), it is tempting to speculate that the inability of murine Mat1^-/- ICM cells to proliferate might also reflect a defect in the activating phosphorylation of mammalian mitotic Cdks.

The switch from the mitotic cell cycle to the endocycle consisting of alternating S and G phases is coupled to several realignments of the cell cycle machinery. Most notably, oscillating cyclin E activity, presumably in complex with Cdk2, has emerged as the critical regulator of endocycles. In Drosophila, this is achieved by oscillation of cyclin E protein levels (Knoblich et al., 1994; Sauer et al., 1995; Lilly and Spradling, 1996). Cyclin E activity also appears to be critical for murine trophoblast giant cell endocycles as constitutive levels of cyclin E lead to arrested endoreduplication (Singer et al., 1999). Our data demonstrate that Mat1 is required for progression of murine trophoblast cells through the endocycle. As Cdk7–cyclin H–Mat1 activates Cdk2–cyclin E complexes in vitro, the endocycle arrest we have described could reflect a defect in the activating phosphorylation of Cdk2–cyclin E. Formal proof of this hypothesis, however, requires further biochemical examination that the paucity of cells in our experimental system unfortunately does not allow.

Mat1 loss leads to destabilization of Cdk7 and cyclin H and disruption of TFIIH kinase

Mat1 has been described variably as an assembly or targeting factor for the Cdk7–cyclin H–Mat1 kinase. Our data provide in vivo evidence that the steady-state levels of both Cdk7 and cyclin H are Mat1 dependent. The decrease in Cdk7 and cyclin H levels we have observed is probably due to the disruption of the trimeric complex and subsequent destabilization of Cdk7 and cyclin H proteins. This observation, together with the evidence indicating that Mat1 mediates the interaction of Cdk7 and cyclin H with core TFIIH (Rossignol et al., 1997; Busso et al., 2000; Schultz et al., 2000), suggests that loss of Mat1 would effectively disrupt TFIIH kinase activity. Destabilization of core TFIIH does not appear to accompany loss of Mat1, as suggested by the viability of Mat1^-/- cells as well as the apparent transcriptional integrity of these cells.

Mat1 modulates RNA polymerase II CTD phosphorylation in vivo

The two residues within the heptapeptide repeat sequence of the CTD believed to be the most highly modified in vivo are Ser2 and Ser5 (reviewed in Dahmus, 1996). In vitro, amino acid substitution of synthetic heptapeptide repeats has been used principally to link TFIIH kinase activity to phosphorylation of Ser5 (Gebara et al., 1997; Hengartner et al., 1998; Sun et al., 1998; Trigon et al., 1998). Our data provide evidence that mammalian Mat1 mediates phosphorylation of pol II on Ser5 to a great extent in vivo. The fact that most but not all cellular Ser5 phosphorylation is depleted suggests the existence of additional CTD kinases with in vivo affinity for this residue.

We have also observed a decrease in total cellular Ser2 phosphorylation in Mat1^-/- cells, suggesting that TFIIH kinase might also catalyze the phosphorylation of Ser2 in vivo. Interestingly, a recent report utilizing chromatin immunoprecipitation in yeast showed that Ser2 phosphorylation is associated predominantly with the 3' ends of actively transcribed loci while TFIIH kinase is believed to dissociate from the nascent transcript shortly after promoter clearance (Komarnitsky et al., 2000). Thus the decrease in Ser2 phosphorylation that we observed in the Mat1^-/- cells could be an indirect consequence of Mat1 loss rather than a direct defect in TFIIH-catalyzed Ser2 phosphorylation. If so, our data are consistent with a model in which TFIIH-mediated kinase activity during promoter clearance would in some way regulate the activity of Ser2 kinase(s) that then act during elongation. Such regulation could involve the direct activation of Ser2 kinases by TFIIH kinase. Alternatively, TFIIH may serve to enhance the substrate affinity of Ser2 kinases by phosphorylating Ser5 during promoter clearance.

Is Mat1-mediated TFIIH kinase function required for pol II transcription?

Surprisingly, disruption of murine Mat1 was not accompanied by a detectable decrease in transcription or translation in Mat1^-/- cells despite apparent defects in CTD phosphorylation. This could be inferred indirectly by observing that the steady-state levels of several cellular proteins (including short-lived cyclin D1) were found to be comparable in both mutant and control cells. More directly, Mat1^-/- cells were capable of de novo transcription and translation of a microinjected reporter plasmid. This status of apparent transcriptional integrity was maintained several days after the cells arrested in the endocycle, which appeared to occur concomitantly with loss of maternally provided Mat1 protein.

Our results differ from those obtained in budding yeast, where loss of TFIIH kinase activity leads to an extremely rapid and near total loss in the transcription of the vast majority of S.cerevisiae genes (Holstege et al., 1998). Accordingly, the transcriptional deficit accompanying the disruption of Mat1 or Cdk7 homologs in S.cerevisiae rapidly leads to loss of viability (Valay et al., 1995; Faye et al., 1997). Our results suggest that the strict dependence of pol II transcription on TFIIH kinase activity in yeast is not maintained in murine trophoblast giant cells, perhaps reflecting unexplored differences in the role of TFIIH kinase in yeast and mammals.

It will be interesting to determine whether the phenotypes associated with Mat1 loss that we describe are specific to early embryonic cell lineages. In this regard, preliminary data utilizing a Cre recombinase-mediated loxP conditional disruption of Mat1 suggest that the Mat1 functions described herein may represent a more general feature of mammalian cells as adult post-mitotic lineages are viable for many months in vivo in the absence of Mat1 (N.Korsisaari, D.J.Rossi, M.Henkemeyer and T.P.Mäkelä, unpublished data).

Materials and methods

Targeting of murine Mat1 and PCR genotyping

A 129-Sv mouse genomic library (Stratagene) was screened with a ³²P-labeled probe from a human MAT1 cDNA corresponding to the open reading frame. Two identified positive clones were subjected to restriction mapping and sequencing, leading to the identification of a single exon corresponding to nucleotides 242–394 of the murine Mat1 cDNA (DDBJ/EMBL/GenBank accession No. U35249) with large introns on both sides. A target vector was constructed by inserting 3.0 kb HindIII, and 3.2 kb KpnI–EcoRI genomic fragments flanking the targeted exon into a conditional replacement vector (pDELBOY-3X; http://www.hi.helsinki.fi/tm/pub/Del/Delboy3x.jpg) derived from pK-11 (kind gift of Drs E.Meyers and G.Martin). XhoI-linearized vector was electroporated into the R1 ES cell line (Nagy et al., 1993), and gancyclovir and G418 double-resistant clones were isolated (Joyner, 1993). Homologous recombination was screened for by Southern blotting with 5' and 3' external probes, both of which were diagnostic upon SacI digestion of genomic DNA. Three out of 800 clones confirmed with both probes and cells from one of these lines (clone 6.85) were injected into C57/BL6 blastocysts. Three chimeric males were identified to be transmitting the targeted allele by Southern blotting and by PCR genotyping. Mice were maintained on inbred (129-Sv), mixed (129-Sv X CD1) and outbred (CD1 >90%) backgrounds with no observable difference in phenotype. PCR genotyping was achieved with the following primers: M7, CAACTAAAGATACCTCCAGCTCC; M10, GCCCTATTTCAGGAGCCAGTCC; and N4, GTCAGTTTCATA GCCTGAAGAACG. M7 and M10 amplify a 190 bp wild-type band, while N4 and M10 amplify a 310 bp mutant band.

Blastocyst outgrowth

Mat1^+/- heterozygous animals were naturally mated and fertilized embryos were flushed from the oviducts at E2.5 with M2 medium (Sigma). Embryos were transferred into M16 medium (Sigma) at 37°C, 5% CO₂ under mineral oil (Sigma) for 48 h. Expanded blastocysts were then transferred to 24-well plates seeded with glass coverslips and maintained in ES medium (Joyner, 1993) for up to 7 days.

BrdU labeling

E3.5 blastocysts and outgrowths at day 5 in culture were cultured in ES cell medium supplemented with 10 M BrdU (Sigma) for 16 and 46 h, respectively. Cells were then fixed in paraformaldehyde (PFA) for 20 min. Following washing in phosphate-buffered saline (PBS), the cells were treated with 0.5 M HCl for 30 min, after which immunofluorescence was performed as described below.

Immunofluorescence and antibodies

For immunofluorescence experiments, we used rabbit polyclonal antibodies against Mat1 (FL-309), Cdk2 (M-2), cyclin E (M-20) and mouse monoclonal antibodies against PCNA (PC10), cyclin D1 (72–13G) and Cdk7 (C-4), all from Santa Cruz. In addition, mouse monoclonal antibodies were used against pol II epitopes H5, 8WG16 (Research Diagnostics) and H14 (Covance), against p53 (Pab 122; Pharmingen), against Cdk6 (Neomarkers), against BrdU (Roche Molecular Biochemicals) and against Mat1 (Transduction Labs). The Cdk7 and cyclin H antibodies were rabbit polyclonal antibodies (Mäkelä et al., 1995). Secondary antibodies included fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgM (Chemicon), FITC-conjugated anti-mouse IgG (Roche Molecular Biochemicals) and rhodamine-conjugated anti-rabbit IgG (Roche Molecular Biochemicals) all used at 1:200 dilution in 5% fetal calf serum (FCS) and 0.5% saponin in PBS. Immunofluoresence was performed variably depending upon the antibodies used. In general, outgrowths grown on glass coverslips were fixed for 20 min in 1.75–3.5% PFA followed by a permeabilization with either 0.5% Triton X-100 in PBS for 20 min or with cold methanol for 2 min and blocking for 30 min in 10% FCS in PBS. Primary and secondary antibodies were left on cells for 1 h, after which DNA was stained with Hoechst 33342 (0.5 g/ml) for 5 min. Each step was followed by three washes of 5 min with PBS. Coverslips were mounted onto glass slides for epifluorescence microscopy with a Zeiss Axiophot 2 microscope and images were photographed with a Sensicam 12 bit digital camera (PCO CCD Imaging). DNA content was quantified from the digital images with NIH Image software.

Microinjection and detection of GFP

Cells cultured in ES cell medium buffered with 10 mM HEPES pH 7.4 were injected with 25 ng/l of pEGFP-N2 (Clontech), 0.1 mg/ml Texas Red dextran (mol. wt 70 000 Da; Molecular Probes) in water. Injection duration was 0.5 s with 120 hPa pressure using an Eppendorf microinjector and transjector (Eppendorf 5246) and a Zeiss Axiovert 135 microscope. After injection, cells were washed twice with ES medium, returned to a 37°C incubator with 5% CO₂ and analyzed 24 h later by fluorescence microscopy.

Acknowledgements

We would like to thank Drs Gerald Faye, Gail Martin and E.Meyers for reagents, Susanna Räsänen for assistance with the animal colony, and members of Makelab for comments on the manuscript. T.P.M. is supported by grants from Academy of Finland, University of Helsinki, Finnish Cancer Organization, Finnish Cancer Institute and Sigrid Juselius Foundation. D.J.R. is a graduate student of Helsinki Biomedical Graduate School; N.K. and A.P. are graduate students of the Helsinki Graduate School in Biotechnology and Molecular Biology.

References

Barlow P, Owen DA and Graham C (1972) DNA synthesis in the preimplantation mouse embryo. J Embryol Exp Morphol, 27, 431–445. | PubMed | ChemPort |

Busso D, Keriel A, Sandrock B, Poterszman A, Gileadi O and Egly JM (2000) Distinct regions of MAT1 regulate cdk7 kinase and TFIIH transcription activities. J Biol Chem, 275, 22815–22823. | Article | PubMed | ChemPort |

Chen D, Riedl T, Washbrook E, Pace PE, Coombes RC, Egly JM and Ali S (2000) Activation of estrogen receptor by S118 phosphorylation involves a ligand-dependent interaction with TFIIH and participation of CDK7. Mol Cell, 6, 127–137. | PubMed | ISI | ChemPort |

Cismowski MJ, Laff GM, Solomon MJ and Reed SI (1995) KIN28 encodes a C-terminal domain kinase that controls mRNA transcription in Saccharomyces cerevisiae but lacks cyclin-dependent kinase-activating kinase (CAK) activity. Mol Cell Biol, 15, 2983–2992. | PubMed | ISI | ChemPort |

Dahmus ME (1996) Reversible phosphorylation of the C-terminal domain of RNA polymerase II. J Biol Chem, 271, 19009–19012. | PubMed | ISI | ChemPort |

Devault A, Martinez AM, Fesquet D, Labbe JC, Morin N, Tassan JP, Nigg EA, Cavadore JC and Doree M (1995) MAT1 ('menage a trois') a new RING finger protein subunit stabilizing cyclin H–cdk7 complexes in starfish and Xenopus CAK. EMBO J, 14, 5027–5036. | PubMed | ISI | ChemPort |

Espinoza FH, Farrell A, Erdjumentbromage H, Tempst P and Morgan DO (1996) A cyclin-dependent kinase-activating kinase (Cak) in budding yeast unrelated to vertebrate Cak. Science, 273, 1714–1717. | PubMed | ISI | ChemPort |

Faye G, Simon M, Valay JG, Fesquet D and Facca C (1997) Rig2, a RING finger protein that interacts with the Kin28/Ccl1 CTD kinase in yeast. Mol Gen Genet, 255, 460–466. | Article | PubMed | ChemPort |

Feaver WJ, Svejstrup JQ, Henry NL and Kornberg RD (1994) Relationship of CDK-activating kinase and RNA polymerase II CTD kinase TFIIH/TFIIK. Cell, 79, 1103–1109. | PubMed | ISI | ChemPort |

Fisher RP, Jin P, Chamberlin HM and Morgan DO (1995) Alternative mechanisms of CAK assembly require an assembly factor or an activating kinase. Cell, 83, 47–57. | PubMed | ChemPort |

Gardner RL (1983) Origin and differentiation of extraembryonic tissues in the mouse. Int Rev Exp Pathol, 24, 63–133. | PubMed | ChemPort |

Gebara MM, Sayre MH and Corden JL (1997) Phosphorylation of the carboxy-terminal repeat domain in RNA polymerase II by cyclin-dependent kinases is sufficient to inhibit transcription. J Cell Biochem, 64, 390–402. | Article | PubMed | ChemPort |

Hengartner CJ, Myer VE, Liao SM, Wilson CJ, Koh SS and Young RA (1998) Temporal regulation of RNA polymerase II by Srb10 and Kin28 cyclin dependent kinases. Mol Cell, 2, 43–53. | PubMed | ISI | ChemPort |

Hermand D, Pihlak A, Westerling T, Damagnez V, Vandenhaute J, Cottarel G and Mäkelä TP (1998) Fission yeast Csk1 is a CAK-activating kinase (CAKAK). EMBO J, 17, 7230–7238. | Article | PubMed | ChemPort |

Hermand D et al. (2001) Specificity of Cdk activation in vivo by the two Caks Mcs6 and Csk1 in fission yeast. EMBO J, 20, 82–90. | Article | PubMed | ChemPort |

Hirose Y and Manley JL (1998) RNA polymerase II is an essential mRNA polyadenylation factor. Nature, 395, 93–96. | Article | PubMed | ISI | ChemPort |

Hirose Y, Tacke R and Manley JL (1999) Phosphorylated RNA polymerase II stimulates pre-mRNA splicing. Genes Dev, 13, 1234–1239. | PubMed | ISI | ChemPort |

Holstege FC, Jennings EG, Wyrick JJ, Lee TI, Hengartner CJ, Green MR, Golub TR, Lander ES and Young RA (1998) Dissecting the regulatory circuitry of a eukaryotic genome. Cell, 95, 717–728. | PubMed | ISI | ChemPort |

Inamoto S, Segil N, Pan ZQ, Kimura M and Roeder RG (1997) The cyclin-dependent kinase-activating kinase (CAK) assembly factor, MAT1, targets and enhances CAK activity on the POU domains of octamer transcription factors. J Biol Chem, 272, 29852–29858. | Article | PubMed | ISI | ChemPort |

Joyner AL (1993) Gene Targeting. IRL Press, Oxford, UK.

Kaldis P (1999) The cdk-activating kinase (CAK): from yeast to mammals. Cell Mol Life Sci, 55, 284–296. | Article | PubMed | ChemPort |

Kaldis P and Solomon MJ (2000) Analysis of CAK activities from human cells. Eur J Biochem, 267, 4213–4221. | Article | PubMed | ChemPort |

Kaldis P, Sutton A and Solomon MJ (1996) The Cdk-activating kinase (Cak) from budding yeast. Cell, 86, 553–564. | PubMed | ISI | ChemPort |

Kimmelman J, Kaldis P, Hengartner CJ, Laff GM, Koh SS, Young RA and Solomon MJ (1999) Activating phosphorylation of the Kin28p subunit of yeast TFIIH by Cak1p. Mol Cell Biol, 19, 4774–4787. | PubMed | ChemPort |

Knoblich JA, Sauer K, Jones L, Richardson H, Saint R and Lehner CF (1994) Cyclin E controls S phase progression and its down-regulation during Drosophila embryogenesis is required for the arrest of cell proliferation. Cell, 77, 107–120. | PubMed | ISI | ChemPort |

Ko LJ, Shieh SY, Chen X, Jayaraman L, Tamai K, Taya Y, Prives C and Pan ZQ (1997) p53 is phosphorylated by CDK7–cyclin H in a p36MAT1-dependent manner. Mol Cell Biol, 17, 7220–7229. | PubMed | ISI | ChemPort |

Komarnitsky P, Cho EJ and Buratowski S (2000) Different phosphorylated forms of RNA polymerase II and associated mRNA processing factors during transcription. Genes Dev, 14, 2452–2460. | Article | PubMed | ISI | ChemPort |

Larochelle S, Pandur J, Fisher RP, Salz HK and Suter B (1998) Cdk7 is essential for mitosis and for in vivo Cdk-activating kinase activity. Genes Dev, 12, 370–381. | PubMed | ISI | ChemPort |

Leclerc V, Raisin S and Leopold P (2000) Dominant-negative mutants reveal a role for the cdk7 kinase at the mid-blastula transition in Drosophila embryos. EMBO J, 19, 1567–1575. | Article | PubMed | ChemPort |

Lee KM, Saiz JE, Barton WA and Fisher RP (1999) Cdc2 activation in fission yeast depends on Mcs6 and Csk1, two partially redundant Cdk-activating kinases (CAKs). Curr Biol, 9, 441–444. | Article | PubMed | ISI | ChemPort |

Lilly MA and Spradling AC (1996) The Drosophila endocycle is controlled by cyclin E and lacks a checkpoint ensuring S-phase completion. Genes Dev, 10, 2514–2526. | PubMed | ISI | ChemPort |

Mäkelä TP, Parvin JD, Kim J, Huber LJ, Sharp PA and Weinberg RA (1995) A kinase-deficient transcription factor IIH is functional in basal and activated transcription. Proc Natl Acad Sci USA, 92, 5174–5178. | PubMed | ChemPort |

Martinez AM, Afshar M, Martin F, Cavadore JC, Labbe JC and Doree M (1997) Dual phosphorylation of the T-loop in cdk7: its role in controlling cyclin H binding and CAK activity. EMBO J, 16, 343–354. | Article | PubMed | ChemPort |

McCracken S et al. (1997) 5'-capping enzymes are targeted to pre-mRNA by binding to the phosphorylated carboxy-terminal domain of RNA polymerase II. Genes Dev, 11, 3306–3318. | PubMed | ISI | ChemPort |

Misteli T and Spector DL (1999) RNA polymerase II targets pre-mRNA splicing factors to transcription sites in vivo. Mol Cell, 3, 697–705. | Article | PubMed | ISI | ChemPort |

Morgan DO (1997) Cyclin-dependent kinases: engines, clocks and microprocessors. Annu Rev Cell Dev Biol, 13, 261–291. | Article | PubMed | ISI | ChemPort |

Nagahara H, Ezhevsky SA, Vocero-Akbani AM, Kaldis P, Solomon MJ and Dowdy SF (1999) Transforming growth factor targeted inactivation of cyclin E:cyclin-dependent kinase 2 (Cdk2) complexes by inhibition of Cdk2 activating kinase activity. Proc Natl Acad Sci USA, 96, 14961–14966. | Article | PubMed | ChemPort |

Nagy A, Rossant J, Nagy R, Abramow-Newerly W and Roder JC (1993) Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Natl Acad Sci USA, 90, 8424–8428. | PubMed | ChemPort |

Nigg EA (1996) Cyclin-dependent kinase 7: at the cross-roads of transcription, DNA repair and cell cycle control? Curr Opin Cell Biol, 8, 312–317. | Article | PubMed | ISI | ChemPort |

Patturajan M, Schulte RJ, Sefton BM, Berezney R, Vincent M, Bensaude O, Warren SL and Corden JL (1998) Growth-related changes in phosphorylation of yeast RNA polymerase II. J Biol Chem, 273, 4689–4694. | Article | PubMed | ISI | ChemPort |

Rochette-Egly C, Adam S, Rossignol M, Egly JM and Chambon P (1997) Stimulation of RAR activation function AF-1 through binding to the general transcription factor TFIIH and phosphorylation by CDK7. Cell, 90, 97–107. | PubMed | ChemPort |

Rodriguez CR, Cho EJ, Keogh MC, Moore CL, Greenleaf AL and Buratowski S (2000) Kin28, the TFIIH-associated carboxy-terminal domain kinase, facilitates the recruitment of mRNA processing machinery to RNA polymerase II. Mol Cell Biol, 20, 104–112. | Article | PubMed | ISI | ChemPort |

Rossignol M, Kolbcheynel I and Egly JM (1997) Substrate specificity of the cdk-activating kinase (Cak) is altered upon association with TFIIH. EMBO J, 16, 1628–1637. | Article | PubMed | ChemPort |

Sauer K, Knoblich JA, Richardson H and Lehner CF (1995) Distinct modes of cyclin E/cdc2c kinase regulation and S-phase control in mitotic and endoreduplication cycles of Drosophila embryogenesis. Genes Dev, 9, 1327–1339. | PubMed | ISI | ChemPort |

Schultz P, Fribourg S, Poterszman A, Mallouh V, Moras D and Egly JM (2000) Molecular structure of human TFIIH. Cell, 102, 599–607. | PubMed | ISI | ChemPort |

Serizawa H, Mäkelä TP, Conaway JW, Conaway RC, Weinberg RA and Young RA (1995) Association of Cdk-activating kinase subunits with transcription factor TFIIH. Nature, 374, 280–282. | Article | PubMed | ISI | ChemPort |

Shiekhattar R, Mermelstein F, Fisher RP, Drapkin R, Dynlacht B, Wessling HC, Morgan DO and Reinberg D (1995) Cdk-activating kinase complex is a component of human transcription factor TFIIH. Nature, 374, 283–287. | Article | PubMed | ISI | ChemPort |

Singer JD, Gurian-West M, Clurman B and Roberts JM (1999) Cullin-3 targets cyclin E for ubiquitination and controls S phase in mammalian cells. Genes Dev, 13, 2375–2387. | Article | PubMed | ISI | ChemPort |

Solomon MJ, Lee T and Kirschner MW (1992) Role of phosphorylation in p34cdc2 activation: identification of an activating kinase. Mol Biol Cell, 3, 13–27. | PubMed | ISI | ChemPort |

Sun X, Zhang Y, Cho H, Rickert P, Lees E, Lane W and Reinberg D (1998) NAT, a human complex containing Srb polypeptides that functions as a negative regulator of activated transcription. Mol Cell, 2, 213–222. | PubMed | ISI | ChemPort |

Tassan JP, Jaquenoud M, Fry AM, Frutiger S, Hughes GJ and Nigg EA (1995) In vitro assembly of a functional human CDK7–cyclin H complex requires MAT1, a novel 36 kDa RING finger protein. EMBO J, 14, 5608–5617. | PubMed | ChemPort |

Thompson NE, Steinberg TH, Aronson DB and Burgess RR (1989) Inhibition of in vivo and in vitro transcription by monoclonal antibodies prepared against wheat germ RNA polymerase II that react with the heptapeptide repeat of eukaryotic RNA polymerase II. J Biol Chem, 264, 11511–11520. | PubMed | ISI | ChemPort |

Thuret JY, Valay JG, Faye G and Mann C (1996) Civ1 (Cak in vivo), a novel Cdk-activating kinase. Cell, 86, 565–576. | PubMed | ISI | ChemPort |

Tirode F, Busso D, Coin F and Egly JM (1999) Reconstitution of the transcription factor TFIIH: assignment of functions for the three enzymatic subunits, XPB, XPD and cdk7. Mol Cell, 3, 87–95. | PubMed | ISI | ChemPort |

Trigon S, Serizawa H, Conaway JW, Conaway RC, Jackson SP and Morange M (1998) Characterization of the residues phosphorylated in vitro by different C-terminal domain kinases. J Biol Chem, 273, 6769–6775. | Article | PubMed | ISI | ChemPort |

Valay JG, Simon M, Dubois MF, Bensaude O, Facca C and Faye G (1995) The KIN28 gene is required both for RNA polymerase II mediated transcription and phosphorylation of the Rpb1p CTD. J Mol Biol, 249, 535–544. | Article | PubMed | ISI | ChemPort |

Varmuza S, Prideaux V, Kothary R and Rossant J (1988) Polytene chromosomes in mouse trophoblast giant cells. Development, 102, 127–134. | PubMed | ISI | ChemPort |

Wu L, Chen P, Shum CH, Chen C, Barsky LW, Weinberg KI, Jong A and Triche TJ (2001) MAT1-modulated CAK activity regulates cell cycle G₁ exit. Mol Cell Biol, 21, 260–270. | Article | PubMed | ChemPort |

Yankulov KY and Bentley DL (1997) Regulation of Cdk7 substrate specificity by Mat1 and TFIIH. EMBO J, 16, 1638–1646. | Article | PubMed | ChemPort |