Portal de Periódicos da CAPES

A link between cell cycle and cell death: Bax and Bcl-2 modulate Cdk2 activation during thymocyte apoptosis

1998; Springer Nature; Volume: 17; Issue: 24 Linguagem: Inglês

10.1093/emboj/17.24.7209

1460-2075

Gabriel Gil‐Gómez,

DNA Repair Mechanisms

Article15 December 1998free access A link between cell cycle and cell death: Bax and Bcl-2 modulate Cdk2 activation during thymocyte apoptosis Gabriel Gil-Gómez Gabriel Gil-Gómez Division of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands Search for more papers by this author Anton Berns Anton Berns Division of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands Search for more papers by this author Hugh J.M. Brady Corresponding Author Hugh J.M. Brady Division of Molecular Immunology, The National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA UK Cancer Biology and Molecular Haematology Units, Camelia Botnar Laboratories, Institute of Child Health, 30 Guilford Street, London, WC1N 1EH UK Search for more papers by this author Gabriel Gil-Gómez Gabriel Gil-Gómez Division of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands Search for more papers by this author Anton Berns Anton Berns Division of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands Search for more papers by this author Hugh J.M. Brady Corresponding Author Hugh J.M. Brady Division of Molecular Immunology, The National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA UK Cancer Biology and Molecular Haematology Units, Camelia Botnar Laboratories, Institute of Child Health, 30 Guilford Street, London, WC1N 1EH UK Search for more papers by this author Author Information Gabriel Gil-Gómez1, Anton Berns1 and Hugh J.M. Brady 2,3 1Division of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands 2Division of Molecular Immunology, The National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA UK 3Cancer Biology and Molecular Haematology Units, Camelia Botnar Laboratories, Institute of Child Health, 30 Guilford Street, London, WC1N 1EH UK *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:7209-7218https://doi.org/10.1093/emboj/17.24.7209 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Resting thymocytes undergoing apoptosis in response to specific stimuli degrade the cdk inhibitor p27Kip1 and upregulate Cdk2 kinase activity. Inhibition of Cdk2 kinase activity efficiently blocks cell death via certain apoptosis pathways whereas overexpression of Cdk2 accelerates such cell death, suggesting its involvement in the signal transduction pathways activated by certain apoptotic stimuli. We found that Cdk2 activation during thymocyte apoptosis can be regulated by p53, Bax and Bcl-2. The highly elevated Cdk2 kinase activity in the apoptosing thymocytes is not associated with its canonical cyclins, cyclin E and cyclin A, and requires de novo synthesis of proteins for activation to take place. We therefore propose Cdk2 activation to be a crucial event in distinct pathways of apoptosis and the point at which the cell cycle and cell death pathways interact. Introduction Apoptosis is a ubiquitous physiological process used to eliminate damaged or unwanted cells in multicellular organisms. For example, apoptosis serves to remove autoreactive or non-functional T cells during their development in the thymus (Suhr and Sprent, 1994). Deregulation of apoptosis is now widely believed to be involved in the pathogenesis of many human diseases (reviewed in Thompson, 1995). The failure of apoptotic cell death has been implicated in various cancers (Lowe et al., 1994; Symonds et al., 1994) as well as autoimmune disorders (Watanabe-Fukunaga et al., 1992; Suda et al., 1993). Conversely, increased apoptosis is associated with a variety of diseases involving cell loss, such as neurodegenerative disorders (Loo et al., 1993; Takashima et al., 1993; Thompson, 1995) and AIDS (Gougeon and Montagnier, 1993). Apoptosis is an active process, regulated by gene products which are conserved from nematodes to mammals and either block or accelerate programmed cell death (reviewed in Penninger and Kroemer, 1998). An understanding of the action of these gene products will provide insights into the process of cell death, and by implication may have profound consequences for the treatment of many human diseases. Bax and Bcl-2 are two of the gene products that can regulate apoptosis. Bax overexpression has been shown to accelerate cell death in response to certain apoptotic stimuli such as growth factor withdrawal, γ-irradiation or glucocorticoid treatment (Oltvai et al., 1993; Brady et al., 1996a), whereas Bcl-2 can inhibit cell death in such circumstances (Sentman et al., 1991; Strasser et al., 1991). Bax and Bcl-2 can form heterodimers and overexpression of one antagonizes the other's effect (Oltvai et al., 1993). They are part of two classes of Bcl-2-related proteins. One class includes Bcl-2 and Bcl-XL which can inhibit apoptosis, the other class includes proteins such as Bax, Bak and Bcl-XS which promote apoptosis (White, 1996). Like apoptosis, cell division is a fundamental and ubiquitous process in multicellular organisms. The molecules which regulate cell-cycle progression and mitosis, such as the cyclins and cyclin-dependent kinases (cdks) (Norbury and Nurse, 1992; Morgan, 1995), are well characterized. Evidence exists to suggest that the cell cycle and apoptosis may be interconnected (Meikrantz and Schlegel, 1995). For example, expression of the proto-oncogene c-myc stimulates cell proliferation and can also predispose cells to apoptosis when growth factors are limiting (Evan et al., 1992). Recent work now indicates that the apoptosis regulatory proteins themselves can directly impinge on the cell-cycle machinery (Brady et al., 1996b; Linette et al., 1996; Mazel et al., 1996; O'Reilly et al., 1996). We and others (Brady et al., 1996b; Linette et al., 1996; Mazel et al., 1996; O‘Reilly et al., 1996) have demonstrated that overexpression of the apoptosis regulatory proteins Bax and Bcl-2 in T cells of transgenic mice leads to perturbations in the dividing population of thymocytes. The effect of the bax transgene is to increase the number of cycling thymocytes (Brady et al., 1996b) whereas bcl-2 has the opposite effect (Brady et al., 1996b; Linette et al., 1996; Mazel et al., 1996; O’Reilly et al., 1996). Furthermore, by studying the interleukin-2 (IL-2)-mediated transition from G1 to S phase in activated T cells, it was shown that bcl-2 overexpression delays entry into S phase (Brady et al., 1996b; Linette et al., 1996) whereas bax speeds it up (Brady et al., 1996b). This correlates with the levels of p27Kip1 in the activated T cells since bcl-2 overexpression is seen to delay the degradation of p27Kip1 whereas bax accelerates it. p27Kip1 is a negative regulator of cdk activity (Polyak et al., 1994; Totoshima and Hunter, 1994). The cdks in turn are positively regulated by association with cyclins and are required for cell-cycle progression (Norbury and Nurse, 1992; Morgan, 1995; Elledge et al., 1996). Several reports have shown the activation of various Cdks to correlate with apoptosis (Meikrantz et al., 1994; Shi et al., 1994; Li et al., 1995; Wei et al., 1997; Zhang et al., 1997; Levkau et al., 1998). It has been shown that dominant negative alleles of Cdk2 can alleviate tumour necrosis factor-α (TNFα)- and staurosporine-induced apoptosis in HeLa cells (Meikrantz and Schlegel, 1996) and more recently, apoptosis induced by growth factor withdrawal from human umbilical vein endothelial cells (Levkau et al., 1998). It has been suggested that it is the association with cyclin A in these situations which upregulates Cdk2 activity. Given the data suggesting a connection between Bax/Bcl-2 and the Cdk2 inhibitor p27Kip1, we examined whether Bax and Bcl-2 could modulate Cdk2 activity during thymocyte apoptosis. We demonstrate a linear relationship between p27Kip1 degradation, Cdk2 activation and thymocyte apoptosis which is modulated by p53, Bax and Bcl-2. Furthermore, we show that inhibitors of Cdk2 activity can block various apoptotic pathways although not the pathway regulated by CD95 and, conversely, that Cdk2 overexpression directly accelerates apoptosis in response to specific stimuli. Finally, we also show that the increased Cdk2 kinase activity in dying thymocytes integrates two different signal transduction pathways, does not seem to be mediated by either cyclin E or cyclin A and requires de novo protein synthesis. We propose that Cdk2 activation is a crucial event in distinct pathways of apoptosis and the point at which the cell cycle and cell death pathways interact. Results Cdk2 activation is necessary for thymocyte apoptosis Recent reports have implicated the cell-cycle regulatory protein Cdk2 in apoptosis. In these studies the study of the significance of Cdk2 activation for apoptosis is complicated by the fact that the model cells are either proliferating (Meikrantz and Schlegel, 1996) or undergoing apoptosis as a consequence of growth factor withdrawal (Levkau et al., 1998). In both cases, the cells examined contain large amounts of Cdk2 activity related to normal cell-cycle progression which cannot be distinguished from Cdk2 activity related to apoptosis. Therefore, we have focused on the involvement of Cdk2 in thymocyte apoptosis. Thymocytes from normal mice are a naturally occurring synchronized population since ∼90% of them are quiescent, whereas only ∼10% are cycling (Huesmann et al., 1991), which makes it possible to separate Cdk2 activity as a regulator of cell-cycle progression from Cdk2 activity involved in apoptosis regulation. Thymocytes were removed from wild-type mice and cultured in vitro in medium alone, exposed to γ-irradiation or treated with the glucocorticoid dexamethasone. To determine the level of its activity, Cdk2 was immunoprecipitated from equivalent numbers of thymocytes and then co-incubated with histone H1 and [γ-32P]ATP (Kranenburg et al., 1995). Cdk2 phosphorylates histone H1 and the amount of 32P incorporated into the histone serves as a measure of Cdk2 kinase activity. The level of histone H1 phosphorylation following immunoprecipitation of Cdk2 from the dying thymocytes increased with time when the cells were kept in culture alone. The level of Cdk2 kinase activity increased even more dramatically following an apoptotic stimulus. The percentage of apoptotic cells present during the process, as determined by a flow cytometry-based technique (Nicoletti et al., 1991), correlated with the elevation in Cdk2 kinase activity for all the stimuli tested (Figure 1A). The amount of antibody used in the immunoprecipitations was sufficient to allow the recovery of Cdk2 kinase activities in proliferating cells several fold higher than those measured in apoptotic thymocytes (e.g. see Figure 6) thus showing that the amounts of antibody used in the immunoprecipitations were saturating. Figure 1.Cdk2 activation during thymocyte apoptosis. (A) Thymocytes from wild-type mice were either treated with 5 Gy of γ-radiation, 2 μM dexamethasone or untreated then put in culture. Aliquots (5×105 cells) were taken at the times indicated and analysed by Western blotting using antibodies against p27Kip1. As a loading control, blots were also probed with a pan-actin antibody. The percentage apoptosis was determined for duplicate aliquots of the thymocytes used in the Western blots. Replicas from the Western blot samples were assayed for Cdk2 kinase activity. (B) Thymocytes from wild-type and bax transgenic mice were plated out either without any treatment, after γ-irradiation (5 Gy) or treated with 2 μM dexamethasone in the presence or absence of 25 μM roscovitine dissolved in DMSO. Equivalent amounts of DMSO were included in the samples not treated with roscovitine as a control. Aliquots (5×105 cells) were taken in duplicate at the times indicated and the percentage of cells in apoptosis determined. The mean values and mean deviation are shown. The experiments were repeated on at least three separate occasions and a representative experiment is shown. (C) Wild-type thymocytes were treated with anti-CD95 antibody (1 μg/ml) ± cycloheximide (30 μg/ml) and plated out in the presence or absence of roscovitine (25 μM) for an incubation period of 18 h. The experiments were repeated on at least three separate occasions and a representative experiment is shown. (D) Exponentially growing neo clone B1 and Cdk2 clone 6 were rendered quiescent as described in Materials and methods and plated out either with no treatment (Control), after complete IL-3 withdrawal, after treatment with 5 Gy of γ-radiation or in the presence of 50 μM etoposide. Except in the case of IL-3 withdrawal-induced apoptosis, the BaF3 clones were incubated in starvation medium during the apoptosis time courses. Aliquots (5×105 cells) were withdrawn in duplicate from the cultures at the times indicated and apoptosis levels measured as indicated in Materials and methods. Similar results were obtained on at least four separate occasions and a representative experiment is shown. Independent single-cell derived BaF3 clones with similar expression levels of the transfected Cdk2 protein behaved similarly to the clone shown (data not shown). Download figure Download PowerPoint Roscovitine was developed as a competitive analogue of ATP which selectively inhibits cdk kinase activity with specificity for Cdk1 (Cdc2), Cdk2 and Cdk5 (Meijer, 1996; Meijer et al., 1997). We analysed the consequences of incubation with roscovitine for thymocyte apoptosis. Thymocyte suspensions from wild-type and bax transgenic mice were treated as above with γ-radiation as well as dexamethasone, in the presence of 25 μM roscovitine or an equivalent volume of dimethyl sulfoxide (DMSO). As shown in Figure 1B, roscovitine was able to block both p53-independent (dexamethasone) and p53-dependent(γ-radiation and etoposide; data not shown) apoptosis in thymocytes from wild-type mice. Roscovitine was also able to inhibit bax-accelerated apoptosis in transgenic thymocytes as well as the residual apoptosis measured in cultures of bcl-2 transgenic thymocytes triggered to die with γ-radiation or dexamethasone (data not shown). However, not all apoptotic stimuli were blocked by roscovitine. CD95 (Fas/APO-1)-induced apoptosis in thymocytes was insensitive to the presence of roscovitine, with or without cycloheximide (Figure 1C). Cdk2 kinase activity was unchanged following CD95-induced apoptosis in thymocytes (data not shown). CD95-induced apoptosis does not require protein synthesis and is accentuated by the presence of cycloheximide (Ogasawara et al., 1995). The lack of inhibition by roscovitine on CD95-induced apoptosis is consistent with other data, including our own, that show Bax/Bcl-2-mediated and CD95-induced apoptosis of primary T cells to be separate and distinct pathways (Strasser et al., 1995; Brady et al., 1996b). Roscovitine can also inhibit Cdk5 and Cdk1 activity (Meijer et al., 1997). However, Cdk5 kinase activity is not present in thymocytes, being restricted to neuronal cells (Tsai et al., 1993). We also assayed Cdk1 activity and found that, as described previously by others (Norbury et al., 1994), its activity declines during apoptosis of thymocytes as the proliferative fraction of thymocytes initially present disappears (data not shown). Therefore, it appears that it is the ability of roscovitine to inhibit Cdk2 activity which permits its use to inhibit certain apoptotic stimuli. We have reproduced these effects on thymocyte apoptosis with other Cdk2 inhibitors (data not shown), such as flavopiridol (de Azevedo et al., 1996), butyrolactone I (Kitagawa et al., 1993) and olomucine (Meijer, 1996), which in the case of flavopiridol and butyrolactone do not share structural homology with roscovitine (Meijer, 1996). Cdk2 inhibitors have been described as apoptosis-inducing drugs in proliferating cell lines (van Engeland et al., 1997; Schutte et al., 1997). Under these conditions, Cdk2 inhibition leads to a cell-cycle block followed by apoptosis. These observations clearly show that the effect of Cdk2 inhibitors on cell death is dependent on the cell-cycle status of the target cell. We have been able to reproduce the apoptosis-inducing properties of Cdk2 inhibitors in the proliferating pro-B cell line BaF3 and recapitulate their apoptosis protecting effects, as seen in thymocytes, after rendering BaF3 cells quiescent by IL-3 deprivation (described below for Figure 1D; data not shown). In order to prove that Cdk2 activation can lead to apoptosis, we generated stable transfectants expressing Cdk2 wild-type protein in IL-3-dependent BaF3 cells. Since all our observations have been made using thymocytes which are essentially a resting cell population, we looked for conditions suitable to render the BaF3 cells quiescent prior to inducing apoptosis. We titrated down the amount of IL-3 necessary to arrest the BaF3 cells without allowing them to initiate the apoptotic programme. We then found the concentration of IL-3-conditioned medium necessary to have at least 90% of a BaF3 culture in G0/G1 but <10% undergoing apoptosis after 15 h in low IL-3. Exponentially growing cultures of control cells and transfected BaF3 clones, as described below, were exhaustively washed and plated out in low IL-3 (starving medium) before triggering them with the apoptotic stimuli. We generated stably transfected BaF3 clones constitutively expressing wild-type Cdk2 and studied the kinetics of apoptosis induction after IL-3 withdrawal, γ-irradiation and etoposide treatment. The levels of expression of the wild-type human Cdk2 were comparable to the levels of endogenous mouse Cdk2 present in BaF3 cells (data not shown). As shown in Figure 1D, overexpression of the wild-type Cdk2 leads to a dramatic acceleration in the kinetics of apoptosis appearance in the cultures for both p53-independent (spontaneous apoptosis and IL-3 deprivation) and p53-dependent stimuli (γ-radiation and etoposide treatment). Similar results were obtained with several independent clones expressing comparable levels of Cdk2 (data not shown). The data from the BaF3 transfectants clearly illustrate that elevated Cdk2 activity alone is sufficient to accelerate apoptosis in resting cells. Bax and Bcl-2 levels influence p27Kip1 degradation in thymocytes triggered to die In activated T cells bcl-2 overexpression is seen to delay the degradation of p27Kip1 whereas bax accelerates it (Brady et al., 1996b). Given the apoptosis regulatory role of Bax and Bcl-2, we examined whether they could also regulate p27Kip1 levels in a population of cells triggered to die. As thymocytes are virtually all resting, the level of p27Kip1 in the total thymocyte population is very high and allowed us to follow the changes in its levels upon triggering of apoptosis. Western blot analysis of wild-type and bax transgenic thymocytes following γ-irradiation shows that there is a time-dependent degradation of p27Kip1 (Figure 2A). As previously seen with p27Kip1 levels during the IL-2-driven G1→S progression in activated T cells (Brady et al., 1996b), bax overexpression in thymocytes accelerates p27Kip1 degradation (Figure 2A) while the opposite is seen for bcl-2 (Figure 2B). This result is surprising since p27Kip1 acts as a cell-cycle brake which needs to be degraded to bring cells into S phase, and given the fact that ∼90% of the initial population of thymocytes are in the G0/G1 state and do not enter S phase, during apoptosis (Pellicciari et al., 1996; data not shown). This suggests that p27Kip1 not only functions as a cell-cycle regulator but also controls apoptotic processes. Figure 2.p27Kip1 levels during apoptosis. bax (A) and bcl-2 (B) transgenic thymocytes and thymocytes from wild-type littermate controls were treated with 5 Gy of γ-radiation then put in culture. Aliquots (5×105 cells) were taken at the times indicated and analysed by Western blotting using antibodies against p27Kip1. As a loading control, blots were also probed with a pan-actin antibody. The percentage apoptosis was determined for duplicate aliquots of the thymocytes used in the Western blots. Download figure Download PowerPoint Cdk2 is activated in apoptotic thymocytes in a Bax- and Bcl-2-regulated manner p27Kip1 is a negative regulator of the activity of cdks (Polyak et al., 1994; Totoshima and Hunter, 1994), notably Cdk2. We therefore examined whether Bax and Bcl-2 could modulate Cdk2 activity during thymocyte apoptosis. As we have reported previously (Brady et al., 1996a), thymocytes from bax transgenic mice are more susceptible to die after triggering with p53-dependent (such asγ-irradiation) or p53-independent (such as dexamethasone) apoptotic stimuli. Measurement of the Cdk2 kinase activity in the conditions described above showed that Bax-accelerated apoptosis correlated with the earlier appearance of Cdk2 kinase activity (Figure 3A). Conversely, overexpression of the survival gene bcl-2 in transgenic thymocytes led to resistance to apoptotic stimuli and to a delayed rise in Cdk2 kinase activity (Figure 3B). As a positive control (BaF), exponentially growing BaF3 cells were processed in the same way as the thymocyte samples for Cdk2 kinase assay. As a negative control (BaF + Block), a similar amount of cell extract was immunoprecipitated in the presence of an excess of a blocking peptide for the Cdk2 antibody. These results show that Cdk2 activation during apoptosis is a highly regulated process under the control of known apoptosis regulators such as Bax and Bcl-2. Figure 3.Cdk2 kinase activity during apoptosis. Thymocytes from wild-type mice (5×105 cells/time point), bax (A) and bcl-2 (B) transgenic mice were plated out either without any treatment, after γ-irradiation (5 Gy) or in the presence of 2 μM dexamethasone. The numbers at the bottom of each line indicate the percentage of apoptotic cells at the corresponding time point. As a positive control (BaF), 2×105 exponentially growing BaF3 cells were processed in the same way as the thymocyte samples for Cdk2 kinase assay. As a negative control (BaF + Block), a similar amount of cell extract was immuno-precipitated in the presence of an excess of the M2 peptide. All the time courses were repeated on three separate occasions and a representative experiment is shown. The relative intensity of the radioactive bands between two separate treatments can vary slightly due to different autoradiographic exposure times. Download figure Download PowerPoint p53 is upstream of Cdk2 activation during apoptosis In order to establish firmly the correlation between Cdk2 activation and apoptosis, we carried out similar experiments as described above using thymocytes from p53+/− and p53−/− mice. Thymocytes from p53−/− mice are highly resistant to genotoxic agents such as γ-radiation and chemotherapeutic drugs, while remaining fully sensitive to dexamethasone-induced apoptosis (Clarke et al., 1993; Lowe et al., 1993). As in wild-type thymocytes (Figures 1A and 2), there is a time-dependent degradation of p27Kip1 and an increase in Cdk2 kinase activity in thymocytes from p53+/− mice triggered to die by γ-radiation or dexamethasone treatment, which correlates with the number of apoptotic cells (Figure 4). As expected, there is no difference in the susceptibility of either p53−/− or p53+/− thymocytes to dexamethasone-induced cell death in terms of cells undergoing apoptosis and this correlates with the kinetics of p27Kip1 degradation and Cdk2 induction. Whereas p53+/− thymocytes behave comparably to wild-type thymocytes in γ-radiation-induced apoptosis, p53−/− thymocytes do not show increases in cell death, p27Kip1 is not degraded and Cdk2 kinase activity does not rise during the same time course (Figure 4). Cdk2 activity does begin to appear by the 10 h time point of p53−/−γ-irradiated thymocytes as by then even p53−/− thymocytes are beginning to apoptose. From these results we conclude that there is a correlation between the levels of both p53-dependent and p53-independent apoptosis in thymocytes and the activation of Cdk2 kinase activity. These results also prove that Cdk2 activation during apoptosis is downstream of the action of p53, consistent with the current view that p53 is involved in the sensing of DNA damage induced by γ-radiation and in triggering DNA repair mechanisms or apoptosis depending on the severity of the damage (reviewed in Cox and Lane, 1995). Figure 4.p53 control of Cdk2 kinase activation during apoptosis. Thymocytes from p53+/− and p53−/− mice (5×105 cells/time point) were plated out after γ-irradiation (5 Gy) or in the presence of 2 μM dexamethasone. Samples were taken at the time points indicated and processed for Western blotting with anti-p27Kip1 antibodies or Cdk2 kinase assay as described in Materials and methods. The numbers at the bottom of each line indicate the percentage of apoptotic cells at the corresponding time point. All the time courses were repeated three times and a representative experiment is shown. Download figure Download PowerPoint Cdk2 activation during apoptosis results from changes in its specific activity and is not associated with cyclins A or E Biochemically, the changes in Cdk2 activity measured during apoptosis might be due to an increase in the levels of Cdk2 or an increased specific activity of the protein. In order to distinguish between these two possibilities, we measured the abundance of Cdk2 during the apoptosis time course by Western blot analysis. The amount of Cdk2 protein present did not change appreciably during the apoptosis of γ-irradiated bax transgenic thymocytes which give rise to high levels of Cdk2 kinase activity (Figure 5A). In contrast, Cdk2 kinase activity increased dramatically during the same time course (Figure 3). These results demonstrate that it is the specific activity of Cdk2 rather than its abundance which changes during apoptosis. The relative abundance of the two Cdk2 bands detected by the anti-Cdk2 antibody does not change during the apoptosis time course suggesting that the phosphorylation state of Cdk2 does not contribute to its effect on apoptosis. This is in agreement with the results of Levkau et al. (1998) which also detect no change in Cdk phosphorylation status during apoptosis. Figure 5.Cdk2 specific activity increases during apoptosis. (A) Cdk2 levels during apoptosis. bax transgenic thymocytes and thymocytes from wild-type littermate controls were treated with 5 Gy of γ-radiation then put in culture. Aliquots (5×105 cells) were taken at the times indicated and analysed by Western blotting using antibodies against Cdk2. As a loading control blots were also probed with a pan-actin antibody. Exponentially growing BaF3 cells (5×105) were used as a positive control. The percentage apoptosis was determined for duplicate aliquots of the thymocytes used in the Western blots. (B) Cell-cycle regulatory proteins during apoptosis. Thymocytes from wild-type mice were either treated with 5 Gy of γ-radiation, 2 μM dexamethasone or untreated then put in culture. Aliquots (2×106 cells) were taken at the times indicated and analysed by Western blotting using antibodies against the indicated proteins. The percentage apoptosis was determined for duplicate aliquots of the thymocytes used in the Western blots. As proliferating cell controls, 106 day 2 concanavalin A-stimulated mouse splenocytes were processed (ConA). Download figure Download PowerPoint Figure 6.Cyclin E- and A-associated kinase activity during apoptosis. (A) Wild-type and bax transgenic thymocytes were treated with 5 Gy of γ-radiation. Aliquots (5×105 cells) were withdrawn at the times indicated and processed for histone H1 kinase activity after immunoprecipitation of cyclin A or cyclin E. Radioactive bands were excised from the gel and counted by liquid scintillation. These experiments were repeated on several separate occasions and a representative experiment is shown. As controls, 5×105 exponentially growing BaF3 cells were processed similarly for cyclin A- and E-associated kinase activity and compared with histone H1 kinase activity measured after Cdk2 immunoprecipitation (B). For comparison, the 32P counts of cyclin A- and E-associated kinase activities in wild-type (WT) and bax transgenic (bax) thymocytes at t0 are shown. Download figure Download PowerPoint During the cell cycle, Cdk2 activity is regulated by association with a cyclin subunit, namely cyclin E, then cyclin A (reviewed in Norbury and Nurse, 1992; Sherr, 1993). We investigated whether the Cdk2 kinase activity present in apoptosing thymocytes was associated with either cyclin. We first measured the levels of different cell-cycle regulators during thymocyte apoptosis. The levels of both cyclin E and cyclin A were low, as expected in a mostly resting cell population, and declined in time (Figure 5B), correlating with the disappearance of the proliferative fraction of thymocytes (data not shown). As shown before, p27Kip1 levels decreased as thymocytes underwent apoptosis. Actin levels were measured as loading controls and shown to be constant throughout the time course. D-type cyclins can also bind Cdk2, although only D2–Cdk2 complexes appeared to be active (Ewen et al., 1993). Cyclin D3 levels turned out to be dramatically downregulated quickly after inducing apoptosis in thymocytes (Figure 5B). Similar results were obtained for cyclin D2 (data not shown). These res