The regulation of survival and cell death is a key determinant of cell fate. Recent evidence shows that survival and death machineries are regulated along the cell cycle. In the present paper, we show that BimEL [a BH3 (Bcl-2 homology 3)-only member of the Bcl-2 family of proteins; Bim is Bcl-2-interacting mediator of cell death; EL is the extra-long form] is phosphorylated in mitosis. This post-translational modification is dependent on MEK (mitogen-activated protein kinase/extracellular-signal-regulated kinase kinase) and growth factor signalling. Interestingly, FGF (fibroblast growth factor) signalling seems to play an essential role in this process, since, in the presence of serum, inhibition of FGF receptors abrogated phosphorylation of Bim in mitosis. Moreover, we have shown bFGF (basic FGF) to be sufficient to induce phosphorylation of Bim in serum-free conditions in any phase of the cell cycle, and also to significantly rescue cells from serum-deprivation-induced apoptosis. Our results show that, in mitosis, Bim is phosphorylated downstream of growth factor signalling in a MEK-dependent manner, with FGF signalling playing an important role. We suggest that phosphorylation of Bim is a decisive step for the survival of proliferating cells.
The regulated co-ordination of cell division and cell death is a key determinant for normal physiological functions and cell number homoeostasis [1,2]. Apoptosis (programmed cell death) is a mechanism by which organisms eliminate unwanted or damaged cells [3–5]. Cells have developed surveillance mechanisms that control the progression through the cell cycle by ensuring that the initiation of one event occurs only after successful completion of the previous one . The evidence pointing to the cross-talk between the cell cycle and apoptosis is sparse, but not lacking, although the details are still largely unknown [7–9]. Caspases and the pro-apoptotic Bcl-2 family of proteins have been shown to be up-regulated in S phase after G0 arrest [10,11]. In addition, Bcl-2 and Mcl-1 have been reported to influence and be influenced by the cell cycle (for a review see ). Moreover, activation of several cell-cycle-related proteins has been implicated in apoptosis in post-mitotic neurons . This communication between cell-cycle regulators and the apoptotic machinery led us to the hypothesis that competence for self-destruction of cells might change along the cycle.
The mitochondria play a central role in apoptosis by releasing death-promoting molecules, under apoptotic stimuli such as cytochrome c, thus engaging downstream executioner pathways for subsequent cell death [4,5]. The mitochondrial ‘gates’ are crucial in initiating or restraining the downstream cascades that lead to apoptosis. The Bcl-2 family proteins, consisting of both pro- and anti-apoptotic members, share from one to four homology domains (BH1–BH4), and act as gatekeepers by regulating the translocation of death-promoting molecules from the mitochondrial intermembrane space [12,14]. The members of the Bcl-2 family of proteins are regulated at various levels, each of them being specifically downstream of different, but partially overlapping, apoptotic and survival signals, but all of them in unison sense a wide range of stimuli . Moreover, the ability of these proteins to interact among themselves through the BH domains provides cells with a mechanism for integration of information from within and around the cell (for a review see ). A subset of Bcl-2-related proteins share only the BH3 domain, and are referred to as ‘BH3-only proteins’ . All members of this subgroup are pro-apoptotic, and are thought to exert their function by activating or enhancing other pro-apoptotic Bcl-2-related proteins, or by preventing the action of anti-apoptotic proteins, hence tilting the balance between anti- and pro-apoptotic Bcl-2 molecules .
Bim (Bcl-2-interacting mediator of cell death) is a BH3-only protein that exists mostly in three isoforms, BimS (short), BimL (long) and BimEL (extra-long), generated by alternative splicing . BimL and BimEL have been reported to be modulated at the level of expression [18–21], post-translational modifications and localization. In healthy cells, BimL and BimEL are sequestered to the microtubules (MTs). In response to several apoptotic stimuli, BimL and BimEL are released and re-localized to the intracellular organelles, and trigger apoptosis [22,23]. Release of Bim from the MTs and subsequent apoptosis has also been described to occur as a consequence of phosphorylation downstream of the JNK (c-Jun N-terminal kinase) pathway . Interestingly, other reports have shown that Bim may be attenuated through phosphorylation downstream of survival pathways in a MEK [mitogen-activated protein kinase/ERK (extracellular-signal-regulated kinase) kinase]/ERK-dependent manner [25–28], resulting in degradation [29–31] and possibly loss of ability to interact with anti-apoptotic Bcl-2 and Bcl-XL.
In the present paper, we report that in NIH 3T3 cells, Bim is phosphorylated in mitosis downstream of growth factor signalling in a MEK-dependent manner, suggesting that this pro-apoptotic molecule is attenuated in this phase of the cell cycle. Moreover, we show that FGF (fibroblast growth factor) signalling, shown previously to be involved in survival and proliferation [32,33], plays a major role in phosphorylation of Bim in mitosis. This report provides evidence for the cross-talk between survival and apoptotic machineries in mitosis, indicating intricate quality-control mechanisms in surveillance of cell division.
DMEM (Dulbecco's modified Eagle's medium) with GlutaMAX™-I and 1000 mg/ml glucose, heat-inactivated FBS (foetal bovine serum), non-essential amino acids, PBS, Lipofectamine™, Plus™ reagent and human recombinant bFGF (basic FGF) were obtained from Invitrogen. Nocodazole, taxol, thymidine, poly(L-lysine) and anti-(active ERK p42/p44) antibodies were from Sigma. SU5402, roscovitine, U0126 and LY294002 were from Calbiochem. Purvalanol A was from Tocris (Bristol, U.K.). Anti-Bim antibody was from Stressgen. Anti-pSer65-Bim antibody was from Upstate Biotechnology. Anti-HA (haemagglutinin) monoclonal antibody was from Santa Cruz Biotechnology. Anti-(total ERK p42/p44), anti-Akt and anti-pAkt antibodies were from Cell Signaling Technology. Protein A–Sepharose beads were from Amersham Biosciences. NheI, XhoI and CIP (calf intestine phosphatase) enzymes were from New England Biolabs.
HA–Bim fusion was constructed by subcloning a HA tag in-frame with the N-terminus of BimEL into the NheI and XhoI sites of pCMS-EGFP+BimEL as described in . The frame was confirmed by local automatic sequencing facilities.
Cell culture and cell-cycle synchronization
NIH 3T3, Swiss 3T3 and HEK-293T (human embryonic kidney) cells were cultured in DMEM plus GlutaMAX™-I and 10% (v/v) FBS at 37 °C, under 5% CO2, except that the medium for HEK-293T cells was supplemented with 0.1 mM minimum essential medium non-essential amino acids. Mitotic NIH 3T3 cells obtained by shake-off were arrested previously by treating with nocodazole (100 ng/ml) or taxol (400 nM) for 12 h. The detached (enriched in mitosis) and adherent cells (non-mitotic) were collected separately. G1/S-phase NIH 3T3 cells were obtained by double-thymidine block: cells at 40–50% confluence were incubated with 2.5 mM thymidine for 12 h, washed with PBS, and cultured with fresh medium for 8 h, followed by a second treatment with 2.5 mM thymidine for 12 h. Mitotic arrest was obtained by thymidine–nocodazole block: NIH 3T3 cells at 70–80% confluence were arrested in G1/S phase using 2.5 mM thymidine for 12 h, washed with PBS, and cultured in the presence of 100 ng/ml nocodazole for 12 h. G2-phase cells were obtained as described for mitotic cells, except that Cdk (cyclin-dependent kinase) inhibitor roscovitine (35 μM) or purvalanol A (15 μM) was added 5 h post-G1/S release. G0 quiescent cells were obtained by culturing Swiss 3T3 cells in DMEM supplemented with 0.5% (v/v) FBS for 32 h. Transient serum-deprivation experiments were performed by washing NIH 3T3 cells three times with serum-free medium, followed by incubation with fresh serum-free DMEM for 2 h.
DNA content analysis by flow cytometry (FACS)
Cells were collected after scraping, washed with cold PBS and centrifuged at 3300 g for 5 min. Cells were resuspended in 300 μl of cold PBS and fixed by adding 700 μl of 100% ethanol (70% final), incubated for 20 min at 4 °C, centrifuged at 3300 g for 5 min and resuspended in PBS and 0.1% (v/v) Triton X-100 (Sigma) containing 5 μg/ml RNase (Roche) for 20 min at room temperature (21 °C). Cells were then stained with propidium iodide (Sigma) by incubating with 25 μg/ml propidium iodide in PBS. After 10 min, cells were washed twice with PBS, centrifuged at 3300 g for 5 min and resuspended in PBS. Cells were analysed on a FACScalibur flow cytometer, and results were analysed using CellQuest (Becton Dickinson).
Protein extraction, protein determination, SDS/PAGE and Western blotting
Cells were lysed using cold lysis buffer [25 mM Hepes, 5 mM MgCl2, 1 mM EGTA and 0.5% (v/v) Triton X-100, pH 7.5, supplemented with 2 mM NaF, 1 mM DTT (dithiothreitol), 2 mM PMSF, 20 μg/ml aprotinin, 1.5 μg/ml benzamidine, 10 μg/ml leupeptin and 1 μg/ml pepstatin A] and centrifuged at 20000 g for 15 min at 4 °C. Supernatants were collected, and 6×sample buffer [350 mM Tris/HCl, pH 6.8, 10.3% (w/v) SDS, 300 μl/ml glycerol, 93 μg/ml DTT, 0.12 mg/ml Bromophenol Blue] was added to a final concentration of 1× and heated at 99 °C for 5 min. Protein determination was achieved using Bio-Rad Protein Assay reagent before the addition of sample buffer. Samples were loaded and separated on SDS/discontinuous 4–12% (w/v) acrylamide–bisacrylamide (Bio-Rad) gels. Blotting was done using nitrocellulose membranes (Scheicher & Schuell). Membranes were blocked with PBS and 5% (w/v) non-fat dried milk for 1 h, incubated with antibodies diluted 1:1000 in blocking solution (1 h at room temperature), washed with PBS and 0.1% (v/v) Tween 20, incubated with the respective secondary antibodies diluted 1:1000 in blocking solution (1 h at room temperature) and washed with PBS. For anti-pSer65-Bim antibody staining, blocking was carried out overnight at 4 °C in TBST [Tris-buffered saline with 0.1% (v/v) Tween 20] and 5% (w/v) non-fat dried milk. Primary antibody dilution was 1:2000 in TBST with 5% (w/v) non-fat dried milk (1 h at room temperature), and washes were performed using TBST and 0.5% (w/v) BSA. An ECL® (enhanced chemiluminescence) kit (Amersham Biosciences) was used for detection, according to the manufacturer's instructions.
Protein extracts of mitotic cells were obtained as described above. Before the addition of sample buffer, extracts were quantified and incubated with CIP in 1× Buffer 3 (New England Biolabs) at 50 units of enzyme per 100 μg of total protein at 30 °C for 30 min.
Transient transfection and IP (immunoprecipitation)
Transient transfections were performed in 2.5×105 HEK-293T cells by using 0.5 μg of cDNA, pre-complexed with 0.3 mg/ml Plus™ Reagent and 0.12 mg/ml Lipofectamine™, according to the manufacturer's instructions. After 90 min of incubation at 37 °C under 5% CO2, the medium containing the DNA complexes was replaced by fresh medium. After a further 18 h, cells were treated or not with 20 ng/ml bFGF for 15 min and then collected. Cells were harvested, washed with cold PBS and homogenized in IP buffer [40 mM Tris/HCl, pH 8.0, 300 mM NaCl, 2% (v/v) Nonidet P40, 20% (v/v) glycerol, 50 mM NaF, 1 mM β-glycerophosphate, 1 mM PMSF, 20 μg/ml aprotinin, 1.5 μg/ml benzamidine, 10 μg/ml leupeptin and 1 μg/ml pepstatin A). Homogenates were centrifuged at 20000 g for 10 min at 4 °C, and the supernatants were incubated with solid Protein A–Sepharose beads pre-treated with 0.2 mg/ml anti-HA antibody or anti-pSer65-Bim antibody. After 3 h of incubation on a rocking platform at 4 °C, the samples were washed three times with IP buffer for 10 min, heat-denatured and analysed by SDS/PAGE.
TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling) assay
Cells were plated on poly(L-lysine)-coated coverslips. Cells were washed with PBS and fixed with 4% (w/v) paraformaldehyde for 25 min at room temperature, then washed again with PBS and permeabilized using PBS and 0.1% (v/v) Triton X-100 for 15 min at room temperature. After washing with PBS, FITC-conjugated TUNEL staining was carried out using the In Situ Cell Death Detection Kit (Roche), according to the manufacturer's instructions. Cells were washed three times with PBS and incubated with 100 ng/ml DAPI (4,6-diamidino-2-phenylindole) (Sigma) for 5 min at room temperature for nuclei staining. Cells were washed with PBS, and coverslips were mounted on glass slides using Vectashield mounting medium (Vector). Cells were visualized under a Leica DM LB2 microscope. Image processing and cell counting were carried out using Image J software.
BimL and BimEL, but not BimS, are phosphorylated in mitosis, but not in G1/S or G2 phases of the cell cycle
Using established cell-synchronization methods , NIH 3T3 cells were arrested in G1/S phase or in mitosis (metaphase arrest). The metaphase-arrested cells exhibited the typical round and loosely attached mitotic phenotype, in contrast with the untreated cycling cells or cells arrested in G1/S phase (Figure 1A). DNA-content analysis revealed that nocodazole- or taxol-treated cells detached by shake-off were predominantly in mitosis, displaying mostly 4n (tetraploid) DNA content, in contrast with untreated cycling mixed population of cells, which were highly enriched in G1 phase (Figure 1B). The remaining adherent population of nocodazole- or taxol-treated cells was mostly enriched in G2 phase, consisting of cells that had not yet entered mitosis, as assessed by DNA-content analysis (enrichment in 4n DNA content, Figure 1B, panel 3) and morphology (flat adherent cells, not shown). When re-plated in fresh medium, metaphase-arrested cells were alive and competent to resume the cell cycle. At 12 h after re-plating, morphological features (results not shown) and flow-cytometry analysis revealed that they were viable, since we could not detect a significant sub-G1 population of cells. Moreover, the cells were able to proliferate, exhibiting more cells with 2n (diploid) DNA, indicating exit from mitosis and entry into G1 phase (Figure 1B, panel 5). Western-blot analysis of whole-cell protein extracts using anti-Bim polyclonal antibody detected the three most abundant alternative splice isoforms of the protein (BimS, L and EL) (Figure 1C). Extracts obtained from shaken- off metaphase-arrested cells exhibited slow-migrating bands for BimL and BimEL (but not for BimS), in addition to the bands observed migrating in the same position as protein obtained from control or G1/G2-phase adherent cells (Figure 1C). All further experiments will focus only on the BimEL isoform.
Metaphase-arrested cells display phosphorylation of BimL and BimEL protein, but not BimS
In order to rule out that the slow-migrating band of Bim could be due to the MT-perturbing agents used to arrest cells in metaphase (see the Experimental section), we obtained an enriched population of mitotic cells without using nocodazole or taxol. Cells were arrested in G1/S phase (monitored by flow cytometry) and then released, by washing and adding fresh medium, resulting in cell-cycle progression. At 8 h post-release, a considerable amount of cells had already reached mitosis, endorsed by morphological features (results not shown) and FACS analysis (Figure 2A, upper panel). The cells were shaken off, and Western-blot analysis of protein extracts obtained from the detached cells revealed that the post-translational modification of BimEL took place in normal mitotic cells independently of pharmacological intervention (Figure 2A). To test whether this slow-migrating band was due to phosphorylation, a regulatory mechanism widely used by cells during mitosis , extracts prepared from the detached cells were treated with CIP or left untreated. The slow-migrating band was lost on phosphatase treatment, indicating it was indeed a phosphorylated form of Bim (Figure 2A, lower panel).
Bona fide phosphorylation of Bim in mitosis
To assess if nocodazole could induce phosphorylation of Bim, cycling Swiss 3T3 cells were arrested in G0 phase and were then maintained for 20 h in the presence of the drug, in order to allow the activation of any putative signalling cascades that might induce phosphorylation of Bim. As a positive control for phosphorylated Bim (P-Bim), G0-phase cells were cultured in the presence of serum and nocodazole, resulting in cell-cycle re-entry and mitotic arrest 28 h later (Figure 2B, upper panel). Western-blot analysis of extracts from nocodazole-treated G0-phase cells demonstrated clearly that it failed to induce phosphorylation of Bim, as observed by the protein migrating at a level similar to untreated cycling cells. In contrast, phosphorylation was clearly evident in mitotic cells that were allowed to re-enter the cycle after G0 arrest (Figure 2B, lower panel). These results confirmed that phosphorylation of Bim takes place in mitosis and is independent of nocodazole treatment. Hence, for all further experiments, nocodazole was used as the mitosis-arresting agent.
To investigate the phosphorylation status of BimEL in G2 phase, the Cdk inhibitor roscovitine was used to prevent Cdk1 activation and consequent entry into mitosis . G2-arrested cells were tetraploid to the same extent as mitotic cells, in contrast with G1/S-arrested cells (Figure 3A). Morphological features allowed us to ensure that these cells were indeed in G2 phase and not in mitosis, since they were flat in shape and strongly adherent, whereas the latter displayed a rounded-up shape and were loosely attached to the tissue culture dish (Figures 3B and 3C). Western-blot analysis confirmed that, similarly to G1/S, G2-arrested cells did not possess significant amounts of P-Bim, in contrast with mitotic cells (Figure 3D). These results show that if cells are held in G2 phase and restrained from entering mitosis by inhibiting Cdks (most likely due to inhibition of Cdk1, the primary engine of mitosis), Bim is not phosphorylated. These results support the conclusion that, under cell culture conditions, Bim is specifically phosphorylated in mitosis and not in G1/S or G2 phases of the cell cycle.
Bim is phosphorylated in mitosis, but not in G1/S or G2 phase, and inhibition of Cdks and MEK, but not PI3K, correlates with loss of phosphorylation of Bim in mitosis
Inhibition of Cdks or MEK, but not PI3K (phosphoinositide 3-kinase) induces loss of phosphorylation of Bim in mitosis
Cdk1 in association with cyclin B are the most significant and decisive players for entry and maintenance of mitosis . Taking into account that Bim is specifically phosphorylated in this phase of the cell cycle, we assessed the importance of mitosis-related signalling for the phosphorylation event to occur by perturbing mitosis through pharmacological inhibition of Cdk1 in metaphase-arrested cells. Roscovitine or purvalanol A treatment on mitotic cells greatly reduced phosphorylation of Bim, in contrast with cells held in mitosis for the same amount of time without inhibiting Cdks (Figures 3E and 4E).
Addition of exogenous bFGF causes phosphorylation of Bim in a MEK/ERK-dependent manner, independently of the cell cycle
It has been shown that the MEK/ERK pathway is critical for survival and required for G2/M progression . We addressed the role of the MEK/ERK pathway in phosphorylation of Bim in mitosis by inhibiting MEK1 and 2, the upstream kinases and activators of ERK1 and 2. By using its specific inhibitor U0126 , phosphorylation of Bim was fully reversed, suggesting that the MEK/ERK pathway plays an important role for phosphorylation of Bim in mitosis (Figure 3F). Results from other investigators using other cell types and stimuli also implicate the MEK/ERK pathway in phosphorylating Bim [25,29–31]. Inhibition of PI3K (another well known pro-survival kinase) by using its specific inhibitor LY294002  had no effect on phosphorylation of Bim, since the migration pattern of Bim was equal to that of untreated mitotic cells (Figure 3F). The specificity of MEK and PI3K inhibitors was assessed by using anti-phospho antibodies specific for their respective targets, ERK1/2 and Akt. Levels of phosphorylated ERK and Akt decreased dramatically in the presence of their respective kinase inhibitors (U0126 for MEK and LY294002 for PI3K), whereas no change was observed in presence of inhibitors for their reciprocal kinases. Protein levels of total ERK and total Akt remained unchanged in all conditions (results not shown). The involvement of the MEK/ERK pathway ushered in the ‘usual suspects’ for the survival cascades and growth factor signalling, suggesting a cross-talk between the cell cycle and growth factor signal reception by the cell.
bFGF causes phosphorylation of Bim independently of the cell cycle in a MEK/ERK-dependent manner
FGF signalling has been shown to regulate mitogenic and survival cascades via MEK and ERK . Hence, we set out to investigate whether bFGF, a growth factor known to activate the MEK/ERK pathway in NIH 3T3 fibroblasts , was competent to induce phosphorylation of Bim and override the requirement for cells to be in mitosis for Bim to be phosphorylated. As shown in Figure 4(A), incubating a heterogeneous population of cells with 1, 5 or 10 ng/ml bFGF resulted in increasingly higher levels of P-Bim. At 10 ng/ml, the phosphorylation was even higher than that of mitotic cells, in contrast with protein obtained from cells in G1/S (Figure 4A). To confirm that the slow-migrating band of Bim was indeed a phosphorylated form of the protein, HEK-293T cells were transfected with empty vector or HA–Bim, and extracts were immunoprecipitated using anti-HA or anti-pSer65-Bim antibodies. As shown in Figure 4(B), the anti-HA antibody was competent to immunoprecipitate both the normal (control) and the slow-migrating species of HA–Bim (upon bFGF treatment of the cells) equally well (left-hand panel). In contrast, the anti-pSer65-Bim antibody recognized the slow-migrating HA–Bim, but could only marginally immunoprecipitate the normal migrating band, indicating that P-Bim accounts for the slow-migrating band (Figure 4B, right-hand panel). To understand further if bFGF was acting through the MEK/ERK pathway, G1/S and mitotic cells were treated with or without the MEK inhibitor U0126, and then stimulated with or without bFGF. In Figure 4(C), we present Western blots for the conditions indicated for phosphorylated active ERK (P-ERK), total ERK (as control) and Bim. As expected, there was a robust activation of ERK and phosphorylation of Bim by treating cells with bFGF alone (Figure 4C, lanes 3 and 7), indicating that Bim is downstream of the MEK/ERK pathway, which in turn is downstream of bFGF signalling. Lanes 2, 4, 6 and 8 of Figure 4(C) show that U0126 greatly reduced basal phosphorylation and activation of ERK in all phases of the cell cycle, with no significant amounts of P-Bim detectable even in the presence of bFGF (lanes 4 and 8). Our results presented in Figure 4(C) regarding phosphorylation of Bim downstream of bFGF and in mitosis were confirmed further by Western blot, using the anti-pSer65-Bim antibody (Figure 4D, middle panel), showing that the intensity of signal obtained in each experimental condition using the anti-pSer65-Bim antibody was consistent with the slow-migrating pattern of Bim using the anti-(total Bim) antibody (Figure 4D, upper panel).
Experiments shown in Figures 4(C) and 4(D) showed that total Bim protein levels seem to be significantly higher in G1/S-phase conditions than in mitosis. Although detailed analysis of this observation was beyond the scope of the present work, it is in agreement with a report showing that several BH3-only proteins (including Bim) are up-regulated in G1/S-phase, downstream of E2F1 . P-Bim may also undergo degradation in mitosis, thus contributing to the relative lower amount of protein in this phase of the cell cycle. These issues are currently being addressed in our laboratory.
Cdk inhibitor purvalanol A  (or roscovitine, as shown earlier in Figure 3E) significantly reduced phosphorylation of Bim in mitosis. However, bFGF was able to override this effect (Figure 4E, lanes 4–6). Similar results were obtained when using a cycling population of cells (Figure 4E, lanes 1–3). These results not only showed that the Cdk inhibitors used did not interfere with MEK pathway, but also suggest that bFGF- and mitosis-induced phosphorylation of Bim are outcomes of independent pathways. Alternatively, it is also likely that phosphorylation of Bim in mitosis might occur due to increased sensitivity of cells to growth factors signalling through MEK during this phase of the cell cycle. This hypothesis would explain why Bim undergoes phosphorylation in mitosis without additional growth factors, and why this effect could be reverted by inhibiting Cdks, keeping in mind that Cdk1 in association with CyclinB is the major driving force for mitosis. This would also explain why we were able to block phosphorylation of Bim in mitosis by inhibiting MEK, and to induce phosphorylation of Bim in non-mitotic cells by treating them with exogenous bFGF. Furthermore, phosphorylation of Bim through activation of growth factor signalling pathways and consequent activation of MEK/ERK corroborates results from other researchers [25–29,31].
Phosphorylation of Bim in mitosis is dependent on growth factors
To test the significance of growth factors in phosphorylating Bim in mitosis, we transiently deprived mitotic cells of growth factors by serum starvation and assayed for phosphorylation of the protein. In Figure 5(A) we show, using Western blot, that phosphorylation of Bim and ERK in mitosis in the presence of serum could be reduced significantly by transiently serum-starving the cells, which could be reverted by adding exogenous bFGF. This result clearly illustrates that growth factors are indeed necessary for phosphorylation of Bim in mitosis, in agreement with earlier results regarding bFGF (Figure 4).
Growth-factor-dependent phosphorylation of Bim in mitosis
We next sought to clarify the contribution of FGF signalling by using SU5402  to pharmacologically inhibit FGF receptors and assess the status of Bim in mitosis. However, we first titrated the inhibitor by treating cells with or without SU5402 at different concentrations (Figure 5B). Western-blot analysis showed that bFGF was sufficient to induce phosphorylation of BimEL in the absence of serum. Phosphorylation was prevented in the presence of increasing concentrations of SU5402, being completely abrogated by inhibiting the receptors for 2 h at 100 μM (Figure 5B). Under the same conditions, we were also able to fully revert the phosphorylation phenotype of Bim and reduce phosphorylation of ERK in mitosis in the presence of serum (Figure 5C). This result suggests strongly that FGF signalling plays a critical role in inducing phosphorylation of Bim in mitosis in NIH 3T3 fibroblasts.
Phosphorylation of Bim correlates with bFGF rescuing of serum-deprivation-induced cell death
Based on our results showing that serum deprivation caused a loss of phosphorylation of Bim in mitosis, and that bFGF was sufficient to rescue this phenotype (Figure 5A), we investigated whether bFGF alone was competent to rescue cells from serum-deprivation-induced apoptosis. Mitotic cells were transiently deprived of serum to extinguish growth factor signalling and abrogate phosphorylation of Bim. Cells were then released from metaphase arrest, resulting in progression through cytokinesis to G1-phase (results not shown). Prolonged serum starvation resulted in high levels of cell death, in contrast with cells cultured in the presence of serum or bFGF (Figure 6A). As assessed by DNA-content analysis, 38% of the serum-starved cells were apoptotic, displaying sub-G1 DNA content. Treating cells with bFGF alone significantly reversed this effect (Figure 6B). We do not exclude a requirement for other growth factors, as reflected by the lack of complete rescue of cell death. Moreover, we observed significant levels of apoptosis to occur in conditions of serum-starvation in early G1 and even before cytokinesis. Serum-deprived cells displayed apoptotic features at 3 and 6 h post-mitotic release, as observed by light microscopy and TUNEL assay (results not shown, and Figures 6C and 6D), which could be reversed by treatment with bFGF. These data suggest that decision-making events for the initiation of apoptosis may occur even during mitosis. Taken together, our results strongly indicate that FGF signalling, leading to activation of survival pathways (MEK/ERK) and phosphorylation of Bim, attenuates the pro-apoptotic consequences of serum deprivation.
EFGF rescues cells from serum-deprivation-induced cell death
The harmonious orchestration of cell-fate decisions primarily requires regulation of cell division and cell death through intra- and inter-cellular signalling. These processes are intricately connected, and each one of them may modulate the other, influencing physiology and pathology [1,2]. The cross-talk between cell cycle and apoptosis is still poorly understood [10,11]. In the present study, we have focused on the regulation of cell survival/death signalling in mitosis, specifically the pathways that regulate the pro-apoptotic Bcl-2 family BH3-only protein Bim.
Our results showed that BimEL was phosphorylated specifically in mitosis. We have also shown that this post-translational modification occurred in metaphase-arrested and also in normal mitotic cells obtained from a cycling population, suggesting a biological role for phosphorylation of Bim. Our data indicate that certain signalling pathways activated during mitosis might trigger this post-translational modification. This idea was reinforced by the fact that perturbation of mitosis through pharmacological inhibition of Cdk1 (an essential kinase during mitosis) during metaphase arrest was sufficient to significantly reduce phosphorylation of Bim. Furthermore, inhibition of MEK/ERK pathway in this phase of the cell cycle resulted in complete loss of phosphorylation of Bim, unveiling a link between cell cycle and survival/apoptosis players MEK, ERK and Bim.
In agreement with other reports [25,28–31], we found that exogenous activation of MEK/ERK pathway was sufficient to phosphorylate Bim in interphase (G1/S-arrested or cycling cells). Namely, bFGF, a known activator of the MEK/ERK pathway through FGF receptors, was competent to induce phosphorylation of Bim in a MEK/ERK-dependent manner. We have also shown this pathway to be competent to induce phosphorylation of Bim during metaphase arrest, even in the presence of mitosis-perturbing agents, such as Cdk inhibitors (as discussed above). This result suggested that either growth-factor-dependent phosphorylation of Bim and phosphorylation of Bim in mitosis were independent events or growth factor signalling was being activated as a consequence of cells being in mitosis. The second hypothesis appeared to be more plausible, since phosphorylation of Bim in mitosis was dependent on growth factors, as shown in the serum-deprivation and inhibition of MEK experiments. Concomitantly, inhibition of FGF receptors using a specific inhibitor also prevented phosphorylation of Bim in mitosis in the presence of serum, indicating that FGF signalling plays a pivotal role in this biochemical event. Overall, our data were consistent with reports describing phosphorylation of Bim in cycling cells to occur downstream of growth factor signalling in a MEK/ERK-dependent manner, such as nerve growth factor, epidermal growth factor, serum, or interleukin-2 or -3, depending on the cell typesused [25–29,31].
Growth-factor-dependent MEK/ERK-mediated phosphorylation of Bim has been shown to restrict the apoptotic activity of the protein [25–28,31]. Although the mechanisms by which P-Bim is restrained from causing apoptosis are still not understood, P-Bim has been described to be targeted for proteasome degradation [29–31]. However, other mechanisms responsible for loss of proapoptotic properties of Bim, such as relocation or revised interaction with interacting pro- or anti-apoptotic partners cannot be excluded. Moreover, it has been reported that Bim is kept constrained from initiating apoptosis by sequestration to the MTs, being mobilized off them on stimulation by pro-apoptotic agents [22,24]. It is well documented that, in mitosis, MTs undergo profound rearrangements during the spindle formation. Innumerable MAPs (MT-associated proteins) are displaced from MTs, and MT-stabilizing proteins are loaded on to them [44,45]. We speculate that Bim may be displaced from MTs during the spindle formation, and pose a danger of turning on apoptosis due to release and relocation, hence the need to phosphorylate to incapacitate the molecule. These issues are currently being addressed in our laboratory.
We have observed that mitotic-arrested cells transiently deprived of serum were able to progress to G1 phase when released (by washing away nocodazole), but showing a significant number undergoing apoptosis, as compared with cells treated with 10% FBS or bFGF at the time of release. bFGF alone was able largely to rescue these dying cells, although not completely, suggesting that FGF signalling is relevant, but probably not the only player involved. Our results, together with a number of reports from different research groups showing that MEK/ERK-dependent phosphorylation of Bim attenuates the pro-apoptotic features of the protein, strongly suggest that phosphorylation of Bim is an important step during mitosis, establishing a link between cell cycle and survival signals. Based on our data, we suggest that loss of phosphorylation of Bim in mitosis may trigger apoptosis at the M to G1 transition. Regulation of apoptosis through post-translational modification mechanisms may be specially important in mitosis, since transcription is essentially shut down due to chromatin condensation, hence de novo RNA and protein synthesis become highly constrained.
The Bim-knockout mice show no striking generalized phenotype. The strongest effects observed mainly concern the immune system, consisting of a 2–4-fold increase in B- and T-cell numbers, and being prone to autoimmunity. Nevertheless, the mice are largely uncompromised . The possibility of compensation by other redundant BH3-only proteins in the Bim-knockout mouse during development cannot be excluded. Numerous evidence published recently has established that phosphorylation of Bim influences the function of the protein towards pro- [24,47] or anti-apoptosis [25,26,29–31]. Our results, taken together with these findings, encourage us to speculate whether a knock-in mouse carrying the phosphorylation mutants of Bim will exhibit a more drastic phenotype with respect to viability than Bim-knockout. This approach might uncover new functional implications for regulation of Bim in development.
We thank Dr E. Lam for help with cell-synchronization techniques and for providing cell lines; Dr L.A. Greene for the pCMS-EGFP+BimEL plasmid; Professor M. Raff and Dr M. Mallo for discussion; Dr J. Leon, D. Calado, S. Godinho and M. Rebelo for discussion and comments on the manuscript; Chatterjee laboratory members (Dr T. Pais, C. Figueiredo, A. Mena, A. Veloso, R. Peixoto and M. Matos) for help provided, discussion and comments on the manuscript; N. Moreno for assistance with imaging techniques. This work was funded by research grant POCTI/BCI/42249/2001 from Fundação para a Ciência e a Tecnologia (FCT), Portugal, and Instituto Gulbenkian de Ciência/Fundação Calouste Gulbenkian to S.C. M.G. and A.D.A. are recipients of FCT fellowships SFRH/BD/2729/2000 and fellowship associated to grant POCTI/CBO/47565/2002 respectively. M.G. is on the Programa Gulbenkian de Doutoramento em Biologia e Medicina (Gulbenkian Ph.D. Programme in Biology and Medicine).
basic fibroblast growth factor
Bcl-2-interacting mediator of cell death
calf intestine phosphatase
Dulbecco's modified Eagle's medium
foetal bovine serum
fibroblast growth factor
human embryonic kidney
mitogen-activated protein kinase/ERK kinase
Tris-buffered saline with Tween 20
terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling