Introduction

Infection with oncogenic human papillomavirus (HPV) types, especially HPV 16 and HPV 18, has been recognized as the major cause of cervical dysplasia and carcinoma.1,2,3 Indeed, the tumors as well as tumor-derived cell lines, CaSki and HeLa cells that harbor respectively HPV 16 and 18 strains, continue to express E6 and E7 viral oncoproteins.4 In vitro studies have shown that E6 protein is able to form stable complexes with the wild-type p53 protein and this binding promotes its degradation via the ubiquitin pathway.5,6 The E7 protein interacts with the retinoblastoma protein, the p105Rb, and turns off its suppressor function.7 In the HPV-negative cervical carcinoma cell lines, such as C33A, the two proteins p53 and p105Rb are also involved in the transformation process through mutations.8 As a consequence, inactivation and subsequent loss of cell cycle regulation as well as apoptosis-inducing activities are suppressed. It was especially confirmed that transformed fibroblasts expressing HPV 16 or HPV 18 E6 loose the G1 checkpoint activity very early.9 The G2 checkpoint function is also attenuated and this may be related to the ablation of a p53-regulated G2/M checkpoint (reviewed in10). Moreover, HPV 18 E6 abolishes p53-mediated apoptosis11 which normally modulates the cytotoxicity of anticancer agents.12

Staurosporine (ST), a potent inhibitor of multiple protein kinases, has been identified as an anti-proliferative drug through its cell growth inhibitory capacities. Indeed, ST enhances differentiation of human promyelocytic leukemia cells,13 inhibits tumor cell invasion,14 and arrests cell cycle progression.15 Low concentrations of ST (∼10 nM) lead to G1 cell cycle accumulation.16 In contrast, higher concentrations exert a preferential G2/M arrest and/or DNA polyploïdy (∼50 nM).17,18 Tumor cell lines seem to require higher doses of ST in order to stop the cells either in G1 or G2/M, in comparison with non transformed cells.15,19

Many reports have also confirmed that staurosporine has the potential to kill various human cell lines.20,21,22,23 Therefore, synthetic analogs of ST have been developed like UCN-01 (7-hydroxystaurosporine) and CGP 41251 (4′-N-benzoyl staurosporine), that are now used as anti-cancer agents in phase I clinical trials.24,25 However, the molecular pathways to achieve tumor suppression remain unclear.

Recent cell-free systems permit to explore apoptotic cell death processes.26,27,28,29 With these models, it was confirmed that cytoplasmic extracts can induce apoptotic changes in nuclei from a wide variety of human and animal cells.26,27,29,30,31,32 Such in vitro systems give several advantages for morphological and biochemical analysis of apoptotic death pathways. Indeed, apoptosis is presented as a three stage process: an initiation phase followed by an integration phase and finally an irreversible execution phase (reviewed in33). The latter one begins with condensation of chromatin at the nuclear periphery, followed by blebbing of the nuclear and cytoplasmic membranes. It culminates within 1 h or so, with a fragmentation of residual nuclear structures into discrete membrane-bounded apoptotic bodies.26,34,35,36 However, whether cell death triggered by specific cellular molecules depends on the cell's position in the ongoing cell cycle is unclear. Growing cells can be synchronized and the use of nuclei from cell populations in specific phases of the cycle in presence of cytoplasmic extracts, from G1, S or G2/M cells, can be a valuable approach to study cell death.

We show that the three cancer cells may either be reversibly blocked in the G1 or G2/M phase of the cell cycle, or dye, depending on dose and/or time ST exposure. We observed that cytoplasmic extracts from apoptotic cells are sufficient to initiate events of programmed cell death on healthy nuclei in a cell-free system. Interestingly, we find that nuclei of HPV positive CaSki and HeLa cells enter apoptosis only in the presence of G2/M phase cytoplasmic extracts, while nuclei of HPV negative C33A cells need G1 extracts in order to promote the cell death.

Results

ST induces transient and reversible cell-cycle arrest either in G1 or G2/M

The cell cycle was investigated by flow cytometry comparing the DNA content of CaSki, HeLa and C33A cells treated with two different ranges of ST for 24 h (Figure 1). Low concentrations of ST (2–40 nM) led to cell-cycle arrest in G1 (Figure 1A,B). Higher doses of ST (80–140 nM) led to a G2/M arrest (Figure 1C,D). These results were obtained whatever cells were or not previously synchronized by a double thymidine block (Figure 1A,C versus B,D). Both blocks, G1 and G2/M, started to be observed 12 h post-treatment (data not shown). After 24 h of ST exposure, the cells were completely arrested. At this latter time, the two blocks were still reversible when we changed the media and withdrew the staurosporine. As shown in Figure 2 for CaSki cell line, the cells started progressively to recycle reaching step by step the different phases of the cell cycle, as typically synchronized cells (Figure 2A,B). The trypan blue exclusion assay demonstrated that more than 95% of the cells were viable along all these experiments (data not shown). Moreover, we observed a doubling of cell populations 24 h after ST removal (Figure 2C). Identical results were obtained with HeLa and C33A cell lines.

Figure 1
figure 1

G1 or G2/M arrest of cervical cancer-derived cell lines after treatment by ST for 24 h. Cells were either S-presynchronized (A,C) or not (B,D), and treated with 20 nM ST (A,B) or 100 nM ST (C,D) for up to 24 h. The cells were stained with propidium iodide and DNA content was determined using flow cytometry (A–D); the numbers indicate the percentage of cells in G1 and G2/M. The right panels show a representative DNA content obtained from HeLa cells. On left, each histogram and bar represent the mean and standard deviation of triplicate measurements

Full size image
Figure 2
figure 2

ST removal allowed the cells to restart in the cell cycle. The results of DNA content in CaSki cells, arrested either in G1 (A) or G2/M (B) after 20 nM or 100 nM of ST treatment respectively, were obtained at 0, 8, 16 and 24 h after cells were washed and released into fresh medium. At indicated times, cells were harvested for propidium iodide staining, and analyzed by flow cytometry. In (C) the histograms present the mean of viable cells after ST removal, and bars indicate the standard deviation of triplicate measurements. Cell viability was determined by trypan blue exclusion assay

Full size image

Long term exposure or high concentrations of ST induce apoptosis

When the three cervical cancer cells were maintained in the present of ST (2–140 nM) beyond 24 h, they consistently entered in apoptosis. In this context, the results summarized in Table 1 show that a short (4 h) or longer delay (after the 24 h necessary to get cell synchronization) was required for the three cell lines to enter in programmed cell death, as detected by double labeling Annexin-V/IP. Indeed, the delay was closely dependent on cell lines and cell phase position (Table 1).

Table 1 Determination of the delay before apoptosis entry after ST treatment
Full size table

In order to further explore the programmed cell death process in our cervical carcinoma cell lines, we treated them with ST concentrations exceeding 140 nM. In none synchronized cells after such ST regimens, apoptosis was consistently induced without clear cell-cycle arrests. Nevertheless, when the cells were first blocked in G1 or G2/M, we have been able to show that the entry into cell death started within 4 h from G2/M for CaSki and HeLa cells, and from G1 for C33A cells (Figure 3B). These results were determined by double labeling Annexin-V/IP and confirmed later by cell cycle analysis, TUNEL assays and time-dependent production of internucleosomal DNA ladders (Figure 3C–E). The fluorescence microscopy analysis after DAPI staining revealed characteristic apoptotic nuclear events, such as hypercondensation of the chromatin, reduction of nuclear volume and DNA extrusion through the nuclear envelope (Figure 3F).

Figure 3
figure 3

Staurosporine (150 nM) induced apoptosis in G2/M arrested CaSki, HeLa, and G1 arrested C33A cells. Annexin-V labeling, IP staining and TUNEL assay were performed for the three cell lines, at each time point. Most representative data demonstrating induction of apoptosis are shown: Annexin-V labeling 0 and 4 h post-treatment (A,B), IP staining 8 h (C) and TUNEL assay 12 h (negative control in blue, positive control in red, DNA sample in green) (D). The DNA cleavage was analyzed as oligonucleosome-sized fragments in ethidium bromide-stained agarose gel (ΦX: DNA marker ΦX174) (E). Apoptosis was confirmed by morphology analysis after DAPI staining under fluorescence microscope following 8 h exposure to ST. The white arrowhead indicates the condensation of the chromatin, and the white arrows point dense nuclear bodies (F, ×400)

Full size image

These results show that ST-induced apoptosis takes place from different cell cycle phases, according to the HPV status, either from G2/M for HPV positive CaSki and HeLa cells, or from G1 for HPV negative C33A cervical tumoral cells. We decided next to use cell-free approaches in order to confirm this difference.

Cytoplasmic extracts from apoptotic cells promote typical nuclear apoptosis changes independently of the nuclei origins

We developed a cell-free system to explore cell death in our three cell lines. Purified isolated healthy nuclei have been prepared from G1 and G2/M arrested cells initially treated with ST (20 and 100 nM, respectively for 24 h). Double thymidine-blocked cells allowed us to prepare S-phase nuclei. Cytoplasmic extracts were obtained from ST-induced apoptotic cells (200 nM for 36 h) (Figure 4A). After incubation of the cytoplasmic extracts with various types of isolated healthy nuclei, we observed typical apoptotic morphological changes, accompanied by a DNA fragmentation as revealed by TUNEL analysis and agarose gel electrophoresis (Figure 4B,C). This occurred within the first 60 min at 37°C with a chromatin condensation around the nuclear periphery (Figure 4B: 20, 40 min, white arrows) accompanied by nuclear size reduction (Figure 4B: 40 min). The DNA was highly condensed against the margins of the nuclei, which retained only an amorphous fibrillar network throughout their interior.30 Later there appeared the process of DNA extrusion without nuclear membrane dislocation (Figure 4B: 60 min, white arrows). These changes were identical to those described during programmed cell death occurring in whole cells (Figure 3F). A slight difference may however be noticed between the nuclear morphology in whole cells and in the cell-free system. Nuclei typically appeared more spherical in isolated nuclei incubated in the presence of apoptotic extracts than in nuclei from whole cells (Figure 4B, 40 min versusFigure 3F, 2 min). This could be linked to the presence of a supporting cytoplasmic matrix surrounding the nucleus in intact cells.

Figure 4
figure 4

Representative time-course of morphological and biochemical changes obtained in HeLa isolated nuclei upon addition of cytoplasmic extracts from apoptotic cells. Nuclei in G1, S and G2/M phases were incubated with cytoplasmic extracts from apoptotic cells (200 nM ST, 36 h) (A). They were removed at the indicated times (0, 20, 40 and 60 min), and examined under fluorescence microscope at ×400 magnification after DAPI staining. The white arrows indicates the nuclear chromatin condensation (40 min) and point later the DNA extrusion (60 min) (B). Time-dependent DNA cleavage was concomitantly observed by TUNEL analysis (negative control in black, positive control in red, DNA sample in green), and gel electrophoresis (ΦX: DNA marker ΦX174) (C)

Full size image

The apoptotic process was extremely efficient since 95% of the healthy nuclei presented structural modifications in a synchronous manner in our cell-free assay. This process was blocked at 4°C even after 3 h incubation and did not appear when nuclei were incubated either in presence of the dilution buffer alone or with cytoplasmic extracts obtained from non apoptotic cells (data not shown).

Apoptosis induction is dependent on the origin of the cytoplasmic extracts with a clear difference between HPV positive and negative cell lines

Since we demonstrated a difference for entrance into apoptosis between CaSki, HeLa and C33A cells without nuclear involvement, we decided to look for a likely cytoplasmic implication. G1, S and G2/M cytoplasms and nuclei were extracted from synchronized viable cells as described above. Each cytoplasmic extract was incubated with different isolated nuclei for 60 min at 37°C. No typical changes related to apoptosis have been noticed in the different combinations that we explored (Table 2). The addition of pro-apoptotic dose of ST (200 nM) in our cell-free system allowed us to find a difference between cells according to the cell cycle origin of the cytoplasmic extracts. Thus, we demonstrated that the ST-induced death of CaSki and HeLa nuclei was dictated by cytoplasm which has to be prepared from G2/M cells. In opposite, the nuclei from HPV negative C33A cells were sensitive to ST only in the presence of G1 cytoplasmic extracts (Table 2). Moreover, we have been able to promote apoptosis features in CaSki and HeLa nuclei with G1 C33A cytoplasm and vice versa (data not shown). As illustrated in Figure 5 for CaSki and C33A isolated nuclei, the apoptosis process was typical. Using the different extracts (Figure 5A) incubated in the presence of 200 nM of ST, we showed characteristic morphological changes (Figure 5C) and internucleosomal DNA fragmentation as well as positive TUNEL analysis (Figure 5D).

Table 2 Comparison of different in vitro cell-free systems
Full size table
Figure 5
figure 5

Morphological changes and DNA fragmentation observed in nuclei of either HPV positive or negative cells upon addition of G2/M or G1 phase cytoplasmic extracts, supplemented with 200 nM of ST. G1, S or G2/M phase CaSki and C33A nuclei were incubated respectively with G2/M CaSki and G1 C33A cytoplasmic extracts (A,B). The apoptotic reaction was performed at 37°C for 60 min in the presence of 200 nM ST. The morphological nucleus observations (×400), after DAPI staining, was in agreement with results obtained in Figure 4 (C). The increase of DNA cleavage was confirmed by flow cytometry after TUNEL labeling (both left panels), and agarose gel electrophoresis (right panel) (D)

Full size image

In our cell-free system, the S-phase cytoplasm could never provoke apoptotic features in different tested nuclei. Identically, 200 nM of ST without cytoplasmic extracts did not induce apoptosis (Table 2).

Discussion

In the present study, we investigated the susceptibility of HPV positive and negative human cervical carcinoma-derived cell lines to react to staurosporine (ST), a member of a new class of drugs active on the cell cycle, and potentially interesting in further anticancer therapeutic strategies. Our results showed that the three tumor cell lines tested were sensitive to ST that induced either cell-cycle arrest or apoptosis. Interestingly, in this latter process we observed a difference in the exit from the cell cycle between HPV positive and negative cell lines, strictly dependent on the cytoplasmic fraction as based on our cell-free system.

Cancer of the uterine cervix remains a public health problem worldwide. The treatment of this disease is based on surgery with or without radiotherapy, whereas chemotherapy is largely recognized as poorly effective.37 Human papillomaviruses (HPVs) have been strongly associated with cervical cancer and its precursor lesions.1,2,3,38,39,40 HPV 16 and 18 genotypes account for greater than 65% of all HPV DNA-positive invasive cervical carcinomas in multi-institutional studies1,41 and are found in the majority of human cervical carcinoma-derived cell lines.42 In this context, the intervention of HPV infection in the carcinogenesis is mainly supported by E6 and E7 oncoproteins which abrogate p53 and pRb functions; the genes of those two factors retain usually a wild-type form.43 The recent demonstration, that p53 and pRb disruptions are effective in the way to create human tumor cells, contributes to the knowledge of HPV-induced cell transformation.44

In order to explore new drugs to treat uterine cervix cancer, we decided to study the responses of three different human cervical carcinoma cell lines under staurosporine exposure. CaSki and HeLa cells were selected for their HPV positive status (types 16 and 18, respectively) associated to a wild-type p53 and pRb expression.45 In contrast, our third tumor cell line (C33A cells) present mutations on both genes without HPV infection.46 Two different ranges of ST were tested showing that the three cell lines might be reversibly arrested either in G1 or G2/M cell-cycle phase, depending on the dose and duration of the treatment.

The first range of tested ST drug, 10–40 nM (Figure 1), had the ability to block the cells in G1, as described in a wide variety of normal or transformed mammalian cells, through both CDK2 activity inhibition and CDK inhibitor protein induction such as p21CIP1/WAF1/SDI1 and p27KIP1.17,19,47,48,49 Conversely, growth arrest capacity seems to disappear in foreskin keratinocytes transfected and immortalized by the E6/E7 HPV oncoproteins.50 Our data suggest that the three long established carcinoma-derived cell lines used in the present study retained their original parental sensitivity to ST, as it was recently demonstrated for human cancer cells.51 Concerning the G2/M arrest, it was observed with ST doses ranging from 40 to 140 nM (Figure 1). This result is in agreement with previous data obtained in normal and transformed cells15,17,52 including the E6/E7 immortalized foreskin keratinocytes.50 The mechanism involved in such ST-induced G2/M block is related to a reduction of the CDK1/cyclin B kinase activity secondary to a cdc25C phosphatase inactivation.53,54

Interestingly both G1 and G2/M cell-cycle arrests were reversible in the first 24 h allowing cells to be synchronized after staurosporine removal (Figure 2). These data suggest that the ST drug as well as analogs, i.e. UCN-01, might be useful in future anticancer therapeutic strategies. The antiproliferative effects of ST were observed in our study for drug concentrations ranging from 10 to 140 nM and up to a 24 h treatment duration. Over 140 nM dose or 24 h of exposure, the ST irreversibly induced apoptosis. The exact mechanism of such apoptosis induction remains elusive which made us decide to use this model as well as a cell free system to approach this question.

Programmed cell death was explored in our study by both morphological and biochemical analysis. By using 150 nM of ST on synchronized cells, we observed in each cell line characteristic morphological signs of apoptosis. These are mainly represented by cell shrinkage and nuclear condensation associated to a blebbing process involving both nuclear and cytoplasmic membranes, that lead to the final fragmentation step producing apoptotic bodies.26,34,35,36 TUNEL assay and visualization of a sub-G1 peak by cell-cycle analysis confirmed this final apoptotic event. Interestingly, our cycle study revealed a possible relation between the sub-G1 peak generation and the depletion of CaSki and HeLa cells in the G2/M fraction, and a reduction of C33A cells in the G1 fraction (Figure 3). These results suggested to us that ST-induced apoptosis took place from different cell-cycle positions. The entry in apoptosis occurred from G2/M for CaSki and HeLa cell lines, while it specifically appeared from G1 for C33A cells. In contrast, the three cervical carcinoma-derived cells were resistant to cell death entry from the S-phase after ST treatment.

In order to demonstrate a relationship between cell-cycle exit to apoptosis and the type of cervical cell lines, we decided to develop a cell-free system analysis. This model recapitulates the selective effects observed in intact cells permitting further to analyze the cellular and/or molecular targets involved in specific transduction pathways. The cell-free system turned out to be powerful in the cell cycle research for several years in Xenopus laevis field,55,56 and more recently in mammalian cells as shown by Krude et al.57 in a HeLa cell-free system. Likewise, we demonstrated that the programmed cell death as described above is under the dependence of the cytoplasm. First, only cytoplasmic extracts from any apoptotic cells were able to induce, whatever the origin of isolated nuclei, specific morphological modifications related to apoptosis (Figure 4). Secondly, we showed that only some specific cytoplasmic fractions originating from viable cells and supplemented with 200 nM pro-apoptotic staurosporine, had the ability to provoke nuclear changes (Figure 5). Thus, apoptosis entry of CaSki and HeLa cells was dependent on cytoplasmic components and specifically related to the G2/M cell-cycle checkpoint (Table 2). In opposite, C33A apoptosis was strictly related to the G1 phase. Remarkably, this difference between cells could be associated with the HPV status. However, we can not conclude that the effects observed are due solely to HPV, since the p53 and p105Rb status is also different between the HPV positive and HPV negative cell lines.

Concerning the S-phase exploration, our cell-free model failed to activate apoptosis with any type of S-phase cytoplasmic extracts supplemented with ST. Identical S-phase protective activity has previously been reported in two different apoptosis induction models. Fas-induced apoptosis in T cells as well as oncogene promotion of cell death in fibroblasts were inefficient in S-phase arrested cells.58,59

Our overall results indicate that apoptosis induction takes place in the cytoplasm and is independent of the nucleus issue. This apparently passive role of nuclei in the execution of apoptosis is consistent with previous experiments which found apoptotic changes proceed exclusively in cytoplasm.60,61,62 This latter one presented an ubiquitous apoptosis transduction pathway which might likely be modulated by staurosporine.60,63 Such modulation could take place at ST-sensitive components which have to be multiple and different since our results demonstrated that HPV positive cells were exclusively sensitive to cell death in G2/M phase and HPV negative cells in G1 phase. Additionally to the HPV infection, the three cervical cancer cell lines differ on their p53 gene status besides a supposedly identical p53 function abrogation. Wild-type p53 causes cell-cycle arrest at late G1 phase and induction of apoptosis.64 Moreover, p53 will also be required to arrest human cells in G2 after DNA damage.65,66 Usually, it is believed that cells with a loss of p53 function present a p53-independent apoptotic sensitivity.12,67 Nevertheless, former studies based on HPV positive CaSki and Hela cells after cisplatin treatment showed a clear p53 and p21WAF1 protein expression leading to the hypothesis that this drug may use a p53-dependent pathway even in the presence of E6 and E7 oncoproteins.68,69 Taken together, our data did not rule out the involvement of a p53-dependent pathway in the ST-induced apoptosis, independent of the HPV and p53 status. Indeed, it has been recently shown that under some circumstances, the wild-type p53 function of mutant p53 genes may be restored.70

In summary, our results showed that staurosporine may reversibly arrest cervical cancer cell lines in G1 or G2/M depending on the concentration and the duration of exposure. Beyond certain ST concentrations or/and over 24 h exposure, an irreversible apoptotic entry was initiated. We demonstrated that this death process was strictly dependent on the cytoplasm, related to the cell-cycle phase and to the cell lines.

To conclude, the unspecific protein kinase inhibitor staurosporine has an antiproliferative effect by inducing G1 or G2/M arrest, and apoptosis. Further explorations focused on HPV and cell cycle regulatory proteins are important to understand the complicated mechanisms involved in apoptosis. Therefore, staurosporine, or its derivatives, might be useful in the treatment of cervical neoplastic diseases.

Materials and Methods

Cell cultures and S-phase synchronization

HPV18-positive HeLa (ATCC CCL 2, Rockville, MD, USA) and HPV-negative C33A (ATCC HTB 31) human cervical cancer-derived cell lines were cultured in MEM-Earle medium (Sigma, St Louis, MO, USA), supplemented with 10% (v/v) fetal calf serum (Gibco BRL, Cergy Pontoise, France), 60 mg/l gentamycine, 1% (v/v) non essential amino acid (Biochrom KG, Berlin, Germany). Five per cent (v/v) sodium pyruvate (Sigma) were added for C33A cells. HPV16-positive CaSki (ATCC CRL 1550) human cervical cancer cell line was grown in RPMI 1640 medium (Sigma) containing 10% (v/v) fetal calf serum, 60 mg/ml gentamycine and 1% (v/v) non essential amino acid. Carcinoma cell lines were incubated at 37°C in a humidified atmosphere of a 5% (v/v) CO2 in air.

For synchronization, 3×105 cells were seeded in 100×20 mm Optilux® petri dish (Becton Dickinson, Plymouth, UK). When they were about 80% confluent, thymidine was added to a final concentration of 2.5 mM. After 25 h of incubation at 37°C, cells were released into fresh culture medium and incubated in the usual growth medium for 8 h to allow cells exiting from S-phase.37 After, a second incubation with 1 mM thymidine for 14 h, cells were washed twice with PBS, once with fresh culture medium and used for further experiments.

Staurosporine treatments

Staurosporine (ST) (Sigma) was dissolved in dimethyl sulfoxide. ST was used on unsynchronized or S-synchronized CaSki, HeLa and C33A cell lines. Dose response and kinetic studies were performed using 2, 10, 20, 30, 40, 50, 80, 100, 120 and 140 nM ST for periods of 4, 8, 12, 16, 18 and 24 h. At this last time point, part of the cells were either maintained in the presence of the same concentration of ST, or washed and released in fresh culture medium for additional periods of 4, 8, 12, 16, 24 and 36 h. Moreover, in other experiments, cell lines were treated with 150, 160, 180 and 200 nM for periods of 4, 8, 12, 16, 18 and 24 h. Cells were analyzed by flow cytometry, and fluorescence microscopy. All the experiments reported here were undertaken at least three times.

Isolated nuclei

The isolation of interphase nuclei was carried out as previously described.38,39 Briefly, 2–5.108 cells were harvested by trypsinization (0.15% trypsin, 0.015%. EDTA), washed once with cold PBS, and centrifuged for 5 min at 1600×g. A second wash was performed in 20 vol of ice-cold hypotonic nuclei buffer (NB) (10 mM PIPES pH 7.4, 10 mM KCl, 100 μM PMSF, 2 mM MgCl2, 1 mM DTT, 10 μM cytochalasin B). One tablet of COMPLETE® protease inhibitor (Roche, Meylan, France) was added per 20 ml of NB. The cells were then resuspended in 10 vol of NB, and swelled for 20 min on ice. After swelling, cells were disrupted with 20–30 strokes of a tight-fitting dounce homogenizer. Up to 2 ml of homogenate were layered over 5 ml NB plus 30% sucrose, and centrifuged for 5 min at 1000×g. The nuclei pellet was further resuspended and washed three times in cold PBS and finally pelleted at 1000×g for 5 min. Then nuclei were stored in 5 ml of freezing buffer (FB) (10 mM PIPES pH 7.4, 80 mM KCl, 20 mM NaCl, 5 mM EGTA, 250 mM sucrose, 500 μM spermidine, 200 μM spermine, 50% glycerol, 100 μM PMSF, 1 mM DTT, COMPLETE®) at −20°C.

Cytoplasmic extracts

The cytoplasmic extracts were prepared as described previously.30,38 Briefly, cells were collected after treatment with trypsin-EDTA (0.15%, 0.015% respectively), washed once with cold PBS, once with extraction buffer (EB) (50 mM PIPES pH 7.4, 50 mM KCl, 5 mM EGTA, 100 μM PMSF, 2 mM MgCl2, 1 mM DTT, 20 μM cytochalasin B, COMPLETE®), resuspended in 1 vol of EB, and pelleted at 2000×g for 5 min directly into a dounce homogenizer. The cell pellet was resuspended in 1 vol of 94% (v/v) ethanol and quickly frozen at −80°C. The complete lysis was obtained after three cycles of freezing-thawing. After the lysate was centrifuged at 15 000×g for 45 min, the supernatant was clarified by high-speed centrifugation in a Beckman L8-70 Ultracentrifuge at 150 000×g for 3 h. The cytoplasmic extracts were stored under liquid nitrogen.

Reconstituted in vitro cell-free system

Nuclei were washed once with cold dilution buffer (DB) (10 mM HEPES, 40 mM β-glycerophosphate, 50 mM NaCl, 2 mM MgCl2, 5 mM EGTA, 1 mM DTT, 2 mM ATP, 10 mM phosphocreatine, 50 μg/ml creatine phosphokinase), and resuspended to a final concentration of 5×107 nuclei per ml of DB. The cytoplasmic fractions were diluted to a final concentration of 8 mg of protein per ml of DB. Cytoplasm from non apoptotic cells supplemented or not with 200 nM ST, or cytoplasm from apoptotic cells were incubated with isolated nuclei from viable cells at 4°C, and 37°C for 10, 20, 40, 50, 60 and 70 min. Nuclei were then analyzed by flow cytometry and fluorescence microscopy. All experiments were performed at least three times.

Flow cytometry analysis

Cell cycle distribution was analyzed by flow cytometry as follows: cells were harvested by trypsinization, washed three times with cold PBS, and fixed with 70% (v/v) cold ethanol for 1 h at 4°C. After washing twice with cold PBS, 106 cells were resuspended in 1 ml of staining solution (PBS 1×, 1 mg/ml RNase A Dnase-free, 1 mg/ml propidium iodide) for 60 min at 4°C. DNA content was determinated by flow cytometry using a FACS Calibur with the CellQUEST (Becton Dickinson, Heidelberg, Germany) and ModFit (Verity Software House, USA) softwares for profile analysis. For each experiment, a statistical analysis was realized with GraphPad Prism software, and results were scored to achieve statistical significance.

Assessment of apoptosis

Cell death was assessed by measuring the binding of FITC-labeled Annexin V protein to the phospholipid phosphatidylserine which is present on the external surface of the membrane of apoptotic cells.40 106 cells were stained according to the manufacturer's instructions (Annexin-V-FLUOS staining Kit, Roche). Apoptosis induction was also quantified by determining the percentage of cells with a sub-diploïd DNA content by FACS Calibur analysis. Apoptotic cell death was also studied by in situ labeling of cell preparations using terminal deoxynucleotidyl transferase and fluoresceinated dUTP (TUNEL assay). The In Situ Cell Death Detection Kit (Roche) was used according to the manufacturer's instructions.

Nuclear morphology from whole cells or isolated nuclei were also observed by fluorescence microscopy After DAPI (Sigma) nuclear staining. The analysis was performed using a Zeiss Axiophot microscope equipped with the following filter set (Excitation filter UV-G365, Chromatic division filter FT-395, Arrest filter LP-420). Images were acquired using a COHU camera controlled by the Cytovision Applied Imaging analysis program.

Cellular DNA from apoptotic whole cells, or from isolated nuclei of the cell-free system was extracted with chloroform/isoamyl alcohol (24/1). DNA corresponding roughly to 106 cells was separated through a 2.5% agarose gel (100 V for 60 min). After electrophoresis, gels were stained with an ethidium bromide solution (0.5 μg/ml), and the DNA fragmentation was visualized by UV light. All images were acquired using a Gel Doc 1000, and controlled by the Multi-Analyst Image Analysis Systems from Bio-Rad.