SEARCH

SEARCH BY CITATION

Keywords:

  • Bcl-2-associated death protein;
  • Ca2+ signalling;
  • mitogen- and stress-activated kinase;
  • mitogen-activated protein kinase;
  • ribosomal S6 kinase

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. References

Activated extracellular-signal-regulated kinase (Erk) phosphorylates and activates downstream kinases including ribosomal S6 kinase 2 (Rsk2/RPS6KA3) and mitogen- and stress-activated kinase 1 (Msk1, RPS6KA5). Rsk2 plays an important role in neuronal plasticity, as patients with Coffin–Lowry syndrome, where Rsk2 is dysfunctional, have impaired cognitive function. However, the relative role of neuronal Rsk2 and Msk1 in activating proteins downstream of Erk is unclear. In PC12 cells and in cortical neurones, the calcium ionophore A23187-induced phosphorylation of Erk, Msk1, Rsk2 and also the Bcl-2-associated death protein (Bad), which protects against neurotoxicity. Specific knockdown of Msk1 with small interfering RNA reduced the ability of A23187 to induce Bad phosphorylation in both PC12 cells and cortical neurones. Conversely, specific knockdown of Rsk2 potentiated Bad phosphorylation following A23187 treatment, and also elevated Erk phosphorylation in both cell types. This indicates that Msk1 rather than Rsk2 mediates neuronal Bad phosphorylation following Ca2+ influx and implicates Rsk2 in a negative-feedback regulation of Erk activity.

Abbreviations used
Bad

Bcl-2-associated death protein

CREB

cAMP response element binding protein

csiRNA

control small interfering RNA

Erk

extracellular-signal-regulated kinase

JNK

c-jun N-terminal kinase

MAPK

mitogen-activated protein kinase

MKP

map k phosphatase

Msk1

mitogen- and stress-activated kinase 1

NMDA

N-methyl-d-aspartate

OD

optical density

PBS

phosphate-buffered saline

Rsk2

ribosomal S6 kinase 2

siRNA

small interfering RNA

Stimulation of the N-methyl-d-aspartate (NMDA) class of ionotropic glutamate receptor allows Ca2+ influx into neurones, which triggers a cascade of second messenger responses leading to various diverse effects on neuronal function. A major consequence of such activity is the stimulation of kinase cascades, in particular the extracellular-signal regulated kinase (Erk), c-jun N-terminal kinase (JNK) and p38 mitogen-activated protein kinase (p38 MAPK) pathways. Activation of Erks, JNKs and p38 MAPKs has been detected following NMDA-R stimulation in a number of different brain areas (Xia et al. 1996; Vincent et al. 1998; Fuller et al. 2001; Burr and Morris 2002). While there is conflicting evidence regarding p38 MAPK and JNK activation, it is widely accepted that Erk activation occurs following NMDA-R-induced Ca2+ influx. The activated Erk is strongly implicated in synaptic plasticity (English and Sweatt 1996, 1997), and also in both excitotoxic and neuroprotective effects (Ikonomovic et al. 1997; Kawasaki et al. 1997; Mukherjee et al. 1999; Gonzalez-Zulueta et al. 2000; Irving et al. 2000; Giardina and Beart 2002; Sutton and Chandler 2002; Takeda and Ichijo 2002; Gozdz et al. 2003; Hughes et al. 2003; Hetman 2004; Hetman and Gozdz 2004). This probably reflects Erk activation of various downstream kinases with distinct long-term effects.

Ribosomal S6 kinase (Rsks) have been known for some time to be mediators of Erk signalling (Frodin and Gammeltoft 1999; Hauge and Frodin 2006). Of four Rsk genes, only Rsk2 is substantially expressed in brain regions such as the cerebral cortex and hippocampus (Zeniou et al. 2002). Hence it was thought that Rsk2 mediated many of the effects of NMDA-induced Erk activation in neurones, including phosphorylation of the cAMP response element binding protein (CREB) (West et al. 2002). However, the closely related mitogen- and stress-activated kinase (Msk) kinases – Msk1 and Msk2 (Deak et al. 1998; Pierrat et al. 1998; New et al. 1999) are also activated by Erk (Deak et al. 1998; Pierrat et al. 1998; New et al. 1999; Wiggin et al. 2002), and there is now considerable evidence that CREB phosphorylation in neurones is in fact mediated by primarily Msk1 rather than Rsk2 (Deak et al. 1998; Wiggin et al. 2002; Hauge and Frodin 2006).

Conversely, members of the Bcl-2 proteins play a major role in neurodegeneration/survival responses. Activation of Bcl-2-associated death protein (Bad) by dephosphorylation promotes neuronal death (Wang et al. 1999; Henshall et al. 2002). Bad phosphorylation is believed therefore to be neuroprotective. While Msk1 can phosphorylate Bad in vitro (She et al. 2002), current in vivo evidence suggests Rsk2 is primarily involved in Bad phosphorylation (Tan et al. 1999; She et al. 2002; Eisenmann et al. 2003). This would implicate Rsk2 rather than Msk1 in neuroprotection, and in fact, recent evidence suggests that Msk1 activation promotes excitotoxicity (Hughes et al. 2003).

We recently demonstrated that neuronal Msk1 is potently activated by physiological levels of NMDA receptor stimulation, without any corresponding activation of Rsk2 (Rakhit et al. 2005). We therefore tested the hypothesis that neuronal Rsk2 might be activated by direct influx of Ca2+ ions into neurones (a stronger stimulus), and that, if activated, Rsk2 would mediate potentially neuroprotective effects of Erk activation, such as Bad phosphorylation. This would be particularly relevant to our understanding of the CNS dysfunction in Coffin–Lowry syndrome, which is caused by a deletion in the Rsk2 gene.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. References

Materials

Antibodies towards phospho-ERK (pErk), phospho-T581-MSK1 (pMSK1), phospho-S380-Rsk and phospho-S112-Bad (pBad) were purchased from Cell Signalling Technology (Hitchin, UK); the MSK1-specific antibody was provided by GlaxoSmithKline (Harlow, UK), while the RSK2-specific antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

The small interfering RNA (siRNA) oligonucleotides (double-stranded with two nucleotide overhangs) were designed according to the ‘STEALTHTM’ RNAi designer (Invitrogen, Paisley, UK) at http://www.rnaidesigner.invitrogen.com/sirna/, along with scrambled sequence [control siRNAs (csiRNAs)]. The formulation is optimised to increase siRNA stability and minimise stress/immune responses to siRNAs (Samarsky and Datta 2004). The antisense sequences were as follows: RSK2 siRNA, 5′-GGACCTGGTGTCAAAGATGCTTCAT-3′; RSK2 csiRNA, 5′-GGATGGTACTGAGAATCGTTCCCAT-3′; MSK1 siRNA, 5′-GCTAAAGACCTCCTTCAGCGTCTTT-3′; MSK2 csiRNA, 5′GCTAGATCCTCCACTGCGCTAATTT-3′. Sequences were checked for lack of homology to other transcripts by Basic Local Alignment Search Tool analysis against the RefSeq, Est and rat whole genome databases.

Cell culture

The PC12 cell culture conditions were as previously described (James et al. 2006). Primary neuronal cultures were prepared as described (Morris 1995; Simpson and Morris 2000). The cortex was dissected out from between 10 and 15 embryos at the 17 day of pregnancy and pooled. The tissue was collected in Minimum Essential Medium (Invitrogen) and minced finely, then dispersed in a solution of trypsin/EDTA for an hour at 37°C. After trituration through a sterile pestles, the cells were suspended in Dulbecco’s Minimum Essential Medium (Invitrogen) containing penicillin/streptomycin and 20% foetal calf serum. Cortical cells were plated on laminin/poly-d-Lysine coated sterile culture plates and after 24 h, the medium was changed to Neurobasal medium (Invitrogen) with B27 supplement to promote selective growth of neuronal cells. Cells were routinely stimulated after 7 days in culture (Morris 1995; Simpson and Morris 2000). The siRNA oligos were transfected using Lipofectamine 2000TM (Invitrogen) according to our standard procedures (Conway et al. 2004; James et al. 2006). Tracer amounts of fluorescently tagged inactive siRNAs were added alongside active or control siRNAs, during liposome formation, to allow identification of transfected neurones (Brown et al. 2004; Bektas et al. 2005; Goparaju et al. 2005).

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis and western analysis of samples

After stimulation, cells were washed in phosphate-buffered saline (PBS) and scraped into 1 mL PBS. Cells were centrifuged at 500 g and the supernatant discarded, the cell pellet being dissolved in extraction buffer (5 mmol/L HEPES, 1.5 mmol/L MgCl2, 0.2 mmol/L EDTA, 26% glycerol, 300 mmol/L NaCl, 0.1% dithiothreitol and 0.025% Nonidet P40). Pestles were used to aid grinding of cellular material. This was then frozen and thawed three times in quick succession and then centrifuged (2100 g, 20 min). The supernatant was added to 2X gel loading sample buffer (0.25 mol/L Tris, 2 mmol/L Na4P2O7, 5 mmol/L EDTA, 5% Glycerol and 2% sodium dodecyl sulphate, pH 6.7) boiled, and then loaded on a gel. Proteins were separated using a 10% polyacrylamide gel and western blotted onto nitrocellulose paper. The blots were incubated in TBST (10 mmol/L Tris, 0.1 mol/L NaCl and 0.1% TWEEN 20, pH 7.5) plus 5% non-fat milk for 1 h to block non-specific binding. This was followed by incubation in primary antibody made in TBST in 1% bovine serum albumin for 18 h at 4°C. Antibody dilutions used were as follows: pERK 1/2000; RSK2 1/100; pRSK 1/250; MSK1 20 μg/10 mL; pMSK1 1/250; pBad 1/1000. The blots were then washed 3 × 5 min in TBST and incubated with secondary antibody conjugated to horseradish peroxidase for 1 h, washed again and bands were detected by a chemiluminescence kit (Amersham, Little Chalfont, UK). Integrative optical density (OD) readings for bands were obtained using NIH Image (W. Rasband; NIH, Bethesda, MD, USA). To normalise data between experiments, OD values after drug treatment were expressed as the change in OD (ΔOD) after subtraction of the corresponding value after vehicle treatment. Data for kinase inhibition studies were log-transformed prior to analysis due to the large range of absolute values obtained. Results were analysed using paired t-test (data within a single culture paired) or anova (Minitab software, Coventry, UK).

Immunocytochemistry

Cultured cells were fixed using ice-cold 4%p-formaldehyde in PBS, for 10 min. After 1 h in blocking serum (PBS with 15% normal goat serum), primary antibody solution (PBS with primary antibody, 3% normal goat serum and 0.005% Triton-X100) was added overnight at 4°C. Antibodies were used at the following dilutions: pERK 1/50; RSK2 1/50; pMSK1 1/100; pBad 1/50. The cells were then incubated with the secondary antibody solution (PBS with specific Alexa 594-conjugated anti-rabbit or anti-mouse antibody (Molecular Probes, Eugene, OR, USA) and 15% normal goat serum) for 60 min at 17–19°C. The cells were then coverslipped in Vectashield (Vector Laboratories, Peterborough, UK). Immunocytochemical staining intensity was measured (in arbitrary OD units) using NIH Image. Intensity of staining was monitored over individual cells as previously described (Morris 1995; Simpson and Morris 2000) and expressed as a percentage of staining on control cells on the same slide. Statistical differences between control and treatment groups were assessed using a Mann–Whitney test.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. References

In both PC12 cells and cortical neurones, phosphorylated Erk1 and 2 appeared as two bands at 42 and 44 kDa respectively, with the 44 kDa band predominating.

Initially, we tested the ability of the Ca2+ ionophore A23187 to induce phosphorylation of Erk, Msk1 and Rsk2. Elevated levels of pMsk1 and pRsk immunoreactivity (-ir) were observed in both PC12 cells and cortical neurones 10 min after exposure to A23187 (Fig. 1) Similarly, we detected a clear increase in phospho-Bad (pBad) levels following exposure of both PC12 cells and cortical neurones to A23187 (Fig. 1).

image

Figure 1.  Effect of A23187 on pMsk1, pRsk, pErk and pBad levels. (a and b) Results are shown following treatment with vehicle (veh) or A23187 (50 μmol/L) for 10 min in (a) PC12 neurones or (b) cortical neurones. The multiple bands detected with the Rsk2 antibody in cortical neurones probably represent different Rsk2 isoforms (Sassone-Corsi et al. 1999). Results shown are from a single experiment, which was repeated at least three times with equivalent results.

Download figure to PowerPoint

Both Msk1 and Rsk2 have been suggested to be involved in Bad phosphorylation (Tan et al. 1999; She et al. 2002). Oligonucleotide siRNAs, along with corresponding control oligonucleotides with scrambled sequence but with identical G/C content (csiRNAs), were designed to target rat Rsk2 and Msk1. A number of different concentrations of the siRNA were then tested in PC12 cells. After 48 h treatment, it was found that 2 or 4 nmol/L Rsk2 siRNA produced a selective suppression of the levels of Rsk2 protein, with no effect on the levels of Msk1 protein (Figs 2a and c). Similarly, 2 or 4 nmol/L Msk1 siRNA produced a selective suppression of the levels of Msk1 protein, with no effect on the levels of Rsk2 protein (Figs 2b and c). Higher concentrations showed similar efficacy, but with some loss of selectivity (data not shown). The csiRNAs did not produce any significant effect on either protein. The 4 nmol/L dose was then selected for use in further studies in PC12 cells. Over the series of experiments, a reproducible suppression of the target protein of around 60% was achieved in both cases.

image

Figure 2.  Specificity and efficacy of Msk1 siRNA and Rsk2 siRNA. (a) Levels of total Rsk2-ir and total Msk1-ir in PC12 neurones pre-treated for 48 h with 4 nmol/L siRNA or control siRNA (csiRNAs) targetting Msk1 (left) or Rsk2 (right). (b) Semi-quantitative analysis of the effect of Rsk2 siRNA or Msk1 siRNA on Rsk2 levels. Results shown are mean optical density (OD) +/− SEM, for = 9/group. anova: F(3,32) = 33.2, < 0.001; ***< 0.001, Fisher’s post hoc test. (c) Semi-quantitative analysis of the effect of Rsk2 siRNA or Msk1 siRNA on Msk1 levels. Results shown are mean OD +/− SEM, for = 9/group. anova: F(3,32) = 2.98, < 0.001; **< 0.01, Fisher’s post hoc test.

Download figure to PowerPoint

Following pre-treatment with either Rsk2 siRNA or control csiRNA (4 nmol/L), PC12 cells were then treated with either vehicle or A23187. The Rsk2 siRNA did not show any ability to attenuate the phosphorylation of Bad following A23187 treatment (Figs 3a and c). In fact, there was an augmentation of the increase in pBad levels (Figs 3a and c). This surprising result raised the possibility that Rsk2 might in fact be acting to inhibit Bad phosphorylation. We tested the hypothesis that Rsk2 knockdown increased the activity of the Erk-Bad pathway. Treatment with Rsk2 siRNA increased the levels of pErk detected after A23187 treatment (compare lane 4 with lane 3, Fig. 3b), while suppressing the levels of pRsk detected. These effects were confirmed by semi-quantitative analysis of the western blot band intensities (Figs 3c and d).

image

Figure 3.  Effect of Rsk2 siRNA on Bad and Erk phosphorylation. (a) Levels of total Msk1-ir, total Rsk2-ir and pBad-ir in PC12 neurones pre-treated for 48 h with 4 nmol/L siRNA or control siRNA (csiRNA) targetting Rsk2 mRNA and then treated with vehicle or A23187 (50 μmol/L) for 10 min. (b) Levels of pRsk-ir and pErk-ir in PC12 neurones pre-treated for 48 h with 4 nmol/well siRNA or csiRNA targetting Rsk2 mRNA and then treated with vehicle or A23187 (50 μmol/L) for 10 min. (c) Semi-quantitative analysis of the effect of Rsk2 siRNA and A23187 on Bad and Erk phosphorylation. Western blot band intensities were measured as described in Materials and methods. Results are shown as the change in integrated optical density (ΔOD) relative to the signal in vehicle-treated wells pre-treated with csiRNA. *< 0.05; **< 0.01 versus veh-csiRNA; #< 0.05 versus A23187-csiRNA (t-test); anova: Effect of treatment –F(2,12) = 37.3, < 0.001; †< 0.05, Fisher’s post hoc test. (d) Effect of siRNA as compared with the corresponding csiRNA on pBad-ir and pErk-ir, after A23187 treatment. Results from A23187-treated wells pre-treated with siRNA are shown as the change in integrated optical density (ΔOD) relative to the signal in A23187-treated wells pre-treated with csiRNA. #< 0.05 versus A23187-csiRNA (paired t-test).

Download figure to PowerPoint

In contrast to the results with Rsk2 siRNA, Msk1 siRNA treatment suppressed the phosphorylation of Bad following A23187 treatment (Figs 4a–c).

image

Figure 4.  Effect of Msk1 siRNA on Bad phosphorylation. (a) Levels of total Msk1-ir, total Rsk2-ir and pBad-ir in PC12 neurones pre-treated for 48 h with 4 nmol/L siRNA or control siRNA targetting Msk1 MRNA, and then treated with vehicle or A23187 (50 μmol/L) for 10 min. (b and c) Semi-quantitative analysis of the effect of Msk1 siRNA on Bad phosphorylation. Western blot band intensities were measured as described in Materials and methods. (b) Effect of siRNA and A23187 on pBad-ir. Results are shown as the change in integrated optical density (ΔOD) relative to the signal in vehicle-treated wells pre-treated with csiRNA. *< 0.05 versus veh-csiRNA; #< 0.05 versus A23187-csiRNA (t-test); anova: Effect of treatment –F(2,5) = 19.1, < 0.05, †< 0.05; Fisher’s post hoc test. (c) Effect of siRNA as compared with the corresponding csiRNA on pBad-ir, after A23187 treatment. Results from A23187-treated wells pre-treated with siRNA are shown as the change in integrated optical density (ΔOD) relative to the signal in A23187-treated wells pre-treated with csiRNA. #< 0.05 versus A23187-csiRNA (paired t-test).

Download figure to PowerPoint

These results indicated contrasting roles for Msk1 and Rsk2 in signalling responses to Ca2+ in PC12 cells. Next, we assessed whether these contrasting roles of these two closely related kinases might be characteristic of neurones in general, by conducting equivalent studies in cultured rat cortical neurones. As transfection efficiencies in primary neurones are much lower than in PC12 cells, siRNAs were transfected together with trace amounts of fluorescently labelled inactive marker siRNA, and transfected cells were identified via fluorescence microscopy. Protein levels were monitored in parallel by immunocytochemistry.

The Msk1-specific antibody did not give any specific signal above background when used for immunocytochemistry, therefore, rather than monitoring the degree of Msk1 protein knockdown, the degree of suppression of Msk1 activation was assessed using the phospho-Msk1 antibody. No detectable pMsk1-ir, pErk-ir or pBad-ir was observed after vehicle treatment of cultured neurones. However, after a 10 min treatment with A23187, strong nuclear pMsk1-ir was observed (Figs 5a and b) confirming the results from the PC12 cells that exposure of neurones to the calcium ionophore leads to Msk1 phosphorylation. Staining for pErk and pBad was also induced by A23187 treatment and was also predominatly nuclear (Figs 6 and 7). Initial experiments indicated that 2 nmol/L Rsk2 siRNA produced a selective suppression of the levels of Rsk2-ir (Figs 6a and c). There was no significant effect of Rsk2 siRNA on the levels of pMsk1-ir (Fig. 6c), although there was a trend for the pMsk1-ir staining to be increased. Similarly, 2 nmol/L Msk1 siRNA produced a selective suppression of the levels of pMsk1-ir following treatment with A23187, with no effect on the levels of Rsk2 protein (Figs 7a and c).

image

Figure 5.  Ca2+ influx leads to phosphorylation of Msk1 in cultured cortical neurones. (a and b) Cortical neurones were treated for 10 min with (a) vehicle or (b) 50 μmol/L A23187, and the levels of pMsk-ir were detected by immunocytochemistry. (c) Cortical neurones were treated for 24 h with 2 nmol/L siRNA, and observed by phase microscopy. Scale bars represent 20 μm.

Download figure to PowerPoint

image

Figure 6.  Effects of Rsk2 siRNA in cultured cortical neurones. (a) Levels of total Rsk2-ir (left panels) and fluorescently labelled transfected cells (right panels), incubated with Rsk2 csiRNA (upper panels) or with Rsk2 siRNA (lower panels). (b) Levels of pErk-ir (left panels) and fluorescently labelled transfected cells (right panels), incubated with Rsk2 csiRNA (upper panels) or with Rsk2 siRNA (lower panels), and then treated with A23187 (50 μmol/L) for 10 min. Scale bar represents 20 μm. (c) Semi-quantitative analysis of the effect of Rsk2 siRNA on total Rsk2-ir and pMsk1-ir. ***< 0.001 versus Rsk2 csiRNA; Mann–Whitney test (= 53–85 cells/group). (d) Semi-quantitative analysis of the effect of Rsk2 siRNA on pBad-ir and pErk-ir. **< 0.01; ***< 0.001 versus Rsk2 csiRNA; Mann–Whitney test (= 117–215 cells/group).

Download figure to PowerPoint

image

Figure 7.  Effects of Msk1 siRNA in cultured cortical neurones. (a) Levels of pMsk1-ir (left panels) and fluorescently labelled transfected cells (right panels), incubated with Msk1 csiRNA (upper panels) or with Msk1 siRNA (lower panels). (b) Levels of pBad-ir (left panels) and fluorescently labelled transfected cells (right panels), incubated with Msk1 csiRNA (upper panels) or with Msk1 siRNA (lower panels), and then treated with A23187 (50 μmol/L) for 10 min. Scale bar represents 20 μm. (c) Semi-quantitative analysis of the effect of Msk1 siRNA on total Rsk2-ir and pMsk1-ir. ***< 0.001 versus Msk1 csiRNA; Mann–Whitney test (= 47–166 cells/group). (d) Semi-quantitative analysis of the effect of Msk1 siRNA on pBad-ir. ***< 0.001 versus Msk1 csiRNA; Mann–Whitney test (= 133–134 cells/group).

Download figure to PowerPoint

In the cortical neurones, as with the PC12 cells, the Rsk2 siRNA did not show any ability to attenuate the phosphorylation of Bad following A23187 treatment. In fact, there was an augmentation of the increase in pBad levels (Fig. 6d). Furthermore, the Rsk2 knockdown increased the levels of pErk detected after A23187 treatment (Figs 6b and d).

In contrast to the results with Rsk2 siRNA, Msk1 siRNA treatment suppressed the phosphorylation of Bad following A23187 treatment (Figs 7b and d).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. References

The importance of Erk signalling in many of the long-term responses to glutamate receptor stimulation and Ca2+ influx in neuronal cells is now well established. However, there is little information available on the pathways that are activated downstream of Erk. Studies in transfected cells have suggested that there is substantial overlap between the protein targets phosphorylated by Rsks and Msks, and it is not entirely clear at present how the consequences of neuronal Msk1 activation differ from those of Rsk activation. While there is evidence for an important role for Rsk2 in neuronal signalling, reinforced by the deficits in neuronal function associated with Coffin–Lowry syndrome, recent reports have tended to suggest that Msk1 fulfils many of the functions previously ascribed to Rsk2.

Accumulating evidence implicates Msk1 in Ca2+ and Erk-dependent signalling. Ca2+ influx is reportedly a powerful stimulus for Msk1 phosphorylation and activation in HeLa cells (Tomas-Zuber et al. 2000). We observed a clear increase in the levels of pErk-ir, pMsk1-ir and pRsk2-ir following treatment with A23187, in both PC12 cells and cortical neurones. This implies activation of both Msk1 and Rsk2 in neurones by Ca2+ influx. We have previously shown that activation of neuronal Msk1 following NMDA receptor stimulation is mediated by Erk (Rakhit et al. 2005), and other workers have linked Rsk2 activation to Ca2+ influx and Erk activity (Xing et al. 1998; Thomas et al. 2005).

The phosphorylation of Bad is thought to be one of the major mechanisms through which growth factors promote cell survival (Shimamura et al. 2000; Zhu et al. 2002). Phosphorylation of Bad, at key residues, including S112, promotes binding to 14-3-3 proteins, and sequesters Bad away from the mitochondrial membrane. This prevents dimerisation with Bcl-2, which would otherwise facilitate apoptosis (Bonni et al. 1999; Wang et al. 1999; Henshall et al. 2002). The neuroprotective and pro-survival actions of brain-derived neurotrophic factor are thought to be mediated at least in part via phosphorylation of Bad (Bonni et al. 1999), and we have observed increased Bad phosphorylation following brain-derived neurotrophic factor treatment in both cortical neurones and PC12 cells (Clark and Morris, unpublished). We also observed clear phosphorylation of Bad following A23187 treatment in both cell types. While substantial Ca2+ influx would be predicted to promote neurodegeneration, it is likely that the Ca2+ influx will also trigger an attempted neuroprotective response, which may be represented by the increased Bad phosphorylation. Indeed, there is some evidence that A23187 can exert a protective effect via Bad phosphorylation (Cheng et al. 2002). As A23187 induces phosphorylation of Erk, Rsk and Msk1 in both cell types, either Rsk2 or Msk1 could be mediating Bad phosphorylation via the Erk pathway.

The RNA interference approach for selectively manipulating gene function has become one of the most valuable techniques in cell biology in recent years. The use of siRNAs is particularly effective in cultured cells (Huppi et al. 2005), provided that appropriate steps are taken to maximise efficacy and specificity. The Rsk2 and Msk1 antibodies used here have been characterised previously (Merienne et al. 2001; Hughes et al. 2003). The formulation of siRNAs used here is optimised to increase stability and eliminate stress/immune responses to siRNAs (Samarsky and Datta 2004). Our preliminary experiments identified a concentration of siRNA producing effective knockdown of the target protein, without affecting a closely related protein. Thus Msk1 siRNA reduced the levels of Msk1 but not Rsk2, while Rsk2 siRNA reduced the levels of Rsk2 but not Msk1.

While Rsk2 knockdown did not reduce the levels of pBad-ir in either PC12 cells or cortical neurones, we found that selective knockdown of Msk1 substantially suppressed Bad phosphorylation in response to A23187. The fact that the suppression of Bad phosphorylation was not total is not surprising, considering that the knockdown of Msk1 protein levels by the siRNA is itself incomplete, and the residual Msk1 in the cells would be expected to enable partial function of Msk1-dependent processes.

While most attention has focussed on the role of Rsk2 in Bad phosphorylation, phosphorylation of Bad (at S112) by Msk1 has been reported in peripheral cells in vitro and in vivo (Lizcano et al. 2000; She et al. 2002). In fact, She et al. suggested that Rsk2 and Msk1 provided parallel routes to Bad phosphorylation downstream of Erk. Our data support the primary involvement of Msk1 in neuronal Bad phosphorylation in response to Ca2+ influx. The predominant immunostaining for both Msk1 and pBad appeared to reside in the nuclear region of the cortical neurones. Msk1 is regarded as a mainly nuclear kinase, although some Msk1 is also present in the cytoplasm, i.e. (Funding et al. 2006). In most cell types, Bad and pBad show a cytoplasmic/mitochondrial localisation. However, an apparently nuclear location in neurones has been observed (Stein and Johnson 2002), and may represent compartmentalisation to the nuclear membrane. The role of Msk1 in phosphorylating Bad implied by our results, and those of other groups (Lizcano et al. 2000; She et al. 2002), is therefore most likely mediated by cytoplasmic Msk1 interacting with Bad at the nuclear membrane.

In contrast, no suppression of Bad phosphorylation was observed following selective knockdown of Rsk2. This immediately implies that neuronal Rsk2 does not play an important role in the direct phosphorylation of Bad. While this is in conflict with a considerable amount of evidence in the literature, there may be parallels with the Rsk2-mediated phosphorylation of CREB, which was widely accepted for many years, but has recently been appreciated as of minor importance compared with the role of Msk1. The ability of Rsk2 to phosphorylate Bad in vitro may not indicate an in vivo role. Similarly, much of the strongest evidence for Rsk2 involvement in Bad phosphorylation comes from the effects of transfected dominant-negative constructs, but of course these may also have effects on related proteins such as Msk1 by sequestering activating kinases or scaffolding molecules. Enhanced Bad phosphorylation in the absence of Rsk activation has been noted in the CNS in vivo (Jin et al. 2002). Our data clearly suggest that Msk1 rather than Rsk2 mediates neuronal Bad phosphorylation in response to Ca2+.

Surprisingly, in PC12 cells and in cortical neurones, when Rsk2 levels are suppressed, the levels of pBAD were in fact increased relative to controls following Ca2+ influx. This intriguing finding would be explicable if Rsk2 is mediating a feedback inhibition on some part of the pathway linking Ca2+ influx to Bad phosphorylation. We therefore tested the hypothesis that Rsk2 knockdown might result in elevated Erk phosphorylation in response to Ca2+ influx and found that pErk levels were increased following knockdown of Rsk2. Hence we conclude that Rsk2 is involved in a negative feedback inhibition upstream of Erk.

Interestingly, markedly increased Erk phosphorylation has been noted in skeletal muscle, in response to various stimuli, in mice lacking a functional Rsk2 gene (Dufresne et al. 2001). Thus feedback inhibition of Erk signalling may be a widespread role of Rsk2. While the potential site of Rsk2-mediated inhibition of the Erk signalling cascade is currently unclear, several possibilities exist. Rsk2 can phosphorylate Sos and inhibit Ras activation (Douville and Downward 1997), but Ras and Sos are probably not involved in the neuronal Erk response to Ca2+ (Szeberenyi et al. 1992; Miranti et al. 1995; Anborgh et al. 1999). Rsks may phosphorylate 14-3-3 proteins and hence regulate Raf activity (Kinuya et al. 2000). Alternatively, the level of activity of Erk is profoundly regulated by the MAPK phosphatase Map K1/Dual specificity Phosphatase 1 (MKP1/DUSP1). MKP1 is transcriptionally induced following Erk-activation, and it is conceivable that this induction involves Rsk2. The siRNAs were applied for 48 h in our studies to allow time for Rsk2 depletion, so there is time for MKP1 gene expression to subside in the absence of Rsk2. However, MKP1 transcriptional induction occurs after physiological levels of NMDA receptor activation (Davis et al. 2000), where there is probably no Rsk2 activation (Rakhit et al. 2005), so this mechanism maybe unlikely to explain the feedback action of Rsk2. Equally, we note that MKP1 contains a consensus Rsk phosphorylation site (R-X-X-S) at S334, conserved from rodents to humans, within the MKP1 C-terminal domain – a region that regulates its phosphatase activity (Hutter et al. 2002). It could be speculated that Rsk2-mediated phosphorylation at this site might enhance MKP1 function and thereby suppress subsequent Erk activity.

Coffin–Lowry Syndrome is an X chromosome-linked disorder caused by disruption of the gene encoding Rsk2, characterised by severe mental retardation. As would be predicted from our data, there are some indications in the literature that cells from patients with Coffin–Lowry Syndrome show elevated Erk activity (Soloaga et al. 2003). While it has been assumed that the mental retardation results from dysfunction of Rsk2 downstream targets, our results raise the intriguing possibility that the deficits may rather result from inappropriately high Erk activity. Germline mutations that lead directly to elevated Erk activity are associated with mental impairment (Rodriguez-Viciana et al. 2006), and we have recently shown that abnormally high Erk activity is the direct cause of hippocampal plasticity deficits in a mouse model of type I neurofbromatosis (Guilding et al. 2007).

Our data implicate Msk1 in the phosphorylation of Bad in response to Erk activation. This potentially neuroprotective action contrasts with evidence that Msk1 activation contributes to neurotoxicity (Hughes et al. 2003). It may be that under excitotoxic conditions, excessive Msk1 activation leads to phosphorylation of an additional set of pro-degenerative substrates – i.e. the serine–threonine kinase II (STKII) (Sapkota et al. 2001). Nonetheless, our data clearly suggest that a component of Msk1 activity is neuroprotective. The dichotomous data, implicating Msk1 in both protective and degenerative actions, mirror the evidence available for Erk (Hetman 2004).

While Msk1 clearly participates in the phosphorylation of Bad in response to Ca2+ influx and Erk activation, a potentially neuroprotective action, Rsk2 appears to mediate a feedback inhibition of Erk. This may also contribute to neuroprotection by limiting Erk activation in response to stimuli such as Ca2+ influx. While our data imply that this function may not be invoked by physiological levels of NMDA receptor stimulation, this may fulfil an important neuroprotective role by limiting extreme levels of Erk activation under more extreme stimuli.

Overall, despite the evidence in the literature that there is substantial overlap between the proteins phosphorylated by Rsks and Msk1, our results demonstrate that neuronal Rsk2 and Msk1 play distinct but potentially coordinated roles in signalling downstream of Erk in neurones.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. References
  • Anborgh P. H., Qian X., Papageorge A. G., Vass W. C., DeClue J. E. and Lowy D. R. (1999) Ras-specific exchange factor GRF: oligomerization through its Dbl homology domain and calcium-dependent activation of Raf. Mol. Cell Biol. 19, 46114622.
  • Bektas M., Payne S. G., Liu H., Goparaju S., Milstien S. and Spiegel S. (2005) A novel acylglycerol kinase that produces lysophosphatidic acid modulates cross talk with EGFR in prostate cancer cells. J. Cell Biol. 169, 801811.
  • Bonni A., Brunet A., West A. E., Datta S. R., Takasu M. A. and Greenberg M. E. (1999) Cell survival promoted by the Ras-MAPK signaling pathway by transcription-dependent and -independent mechanisms. Science 286, 13581362.
  • Brown J. M., Boysen M. S., Chung S., Fabiyi O., Morrison R. F., Mandrup S. and McIntosh M. K. (2004) Conjugated linoleic acid induces human adipocyte delipidation: autocrine/paracrine regulation of MAP/ERK signalling by adipocytokines. J. Biol. Chem. 279, 2673526747.
  • Burr P. and Morris B. J. (2002) Involvement of NMDA receptors and a p21Ras-like guanosine triphosphatase in the constitutive activation of nuclear factor-kappa-B in cortical neurons. Exp. Brain Res. 147, 273279.
  • Cheng C.-H., Hsieh C.-L., Shu K.-H., Chen Y.-L. and Chen H.-C. (2002) Effect of calcium channel antagonist diltiazem and calcium ionophore A23187 on cyclosporine A-induced apoptosis of renal tubular cells. FEBS Lett. 516, 191196.
  • Conway A.-M., James A. B., O’Kane E. M., Rakhit S. and Morris B. J. (2004) Regulation of myosin light chain phosphorylation by RhoB in neuronal cells. Exp. Cell Res. 300, 3542.
  • Davis S., Vanhoutte P., Pages C., Caboche J. and Laroche S. (2000) The MAPK/ERK cascade targets both Elk-1 and cAMP response element-binding protein to control long-term potentiation-dependent gene expression in the dentate gyrus in vivo. J. Neurosci. 20, 45634572.
  • Deak M., Clifton A. D., Lucocq J. M. and Alessi D. R. (1998) Mitogen- and stress-activated protein kinase-1 (MSK1) is directly activated by MAPK and SAPK2/p38, and may mediate activation of CREB. EMBO J. 17, 44264441.
  • Douville E. and Downward J. (1997) EGF induced SOS phosphorylation in PC12 cells involves P90 RSK-2. Oncogene 15, 373383.
  • Dufresne S. D., Bjorbak C., El-Haschimi K., Zhao Y., Aschenbach W. G., Moller D. E. and Goodyear L. J. (2001) Altered extracellular signal-regulated kinase signaling and glycogen metabolism in skeletal muscle from p90 ribosomal S6 kinase 2 knockout mice. Mol. Cell Biol. 21, 8187.
  • Eisenmann K. M., VanBrocklin M. W., Staffend N. A., Kitchen S. M. and Koo H.-M. (2003) Mitogen-activated protein kinase pathway-dependent tumor-specific survival signaling in melanoma cells through inactivation of the proapoptotic protein Bad. Cancer Res. 63, 83308337.
  • English J. D. and Sweatt J. D. (1996) Activation of p42 mitogen-activated protein kinase in hippocampal long term potentiation. J. Biol. Chem. 271, 2432924332.
  • English J. D. and Sweatt J. D. (1997) A requirement for the MAP kinase cascade in hippocampal LTP. J. Biol. Chem. 272, 1910319106.
  • Frodin M. and Gammeltoft S. (1999) Role and regulation of 90 kDa ribosomal S6 kinase (RSK) in signal transduction. Mol. Cell Endocrinol. 151, 6577.
  • Fuller G., Veitch K., Lai Kwan H., Cruise L. and Morris B. J. (2001) Activation of p44/p42 MAP kinase in striatal neurons via kainate receptors and PI3 kinase. Mol. Brain Res. 89, 126132.
  • Funding A. T., Johansen C., Kragballe K. et al. (2006) Mitogen- and stress-activated protein kinase 1 is activated in lesional psoriatic epidermis and regulates the expression of pro-inflammatory cytokines. J. Invest. Dermatol. 126, 17841791.
  • Giardina S. F. and Beart P. M. (2002) Kainate receptor-mediated apoptosis in primary cultures of cerebellar granule cells is attenuated by mitogen-activated protein and cyclin-dependent kinase inhibitors. Br. J. Pharmacol. 135, 17331742.
  • Gonzalez-Zulueta M., Feldman A. B., Klesse L. J., Kalb R. G., Dillman J. F., Parada L. F., Dawson T. M. and Dawson V. L. (2000) Requirement for nitric oxide activation of p21Ras/ERK in neuronal ischaemic preconditioning. Proc. Natl Acad. Sci. USA 97, 436441.
  • Goparaju S. K., Jolly P. S., Watterson K. R. et al. (2005) The S1P2 receptor negatively regulates platelet-derived growth factor-induced motility and proliferation. Mol. Cell Biol. 25, 42374249.
  • Gozdz A., Habas A., Jaworski J., Zielinska M., Albrecht J., Chlystun M., Jalili A. and Hetman M. (2003) Role of N-methyl-d-aspartate receptors in the neuroprotective activation of extracellular signal-regulated kinase 1/2 by cisplatin. J. Biol. Chem. 278, 4366343671.
  • Guilding C., Stone T. W. and Morris B. J. (2007) Restored plasticity in a mouse model of neurofibromatosis type 1 via inhibition of hyperactive ERK and CREB. Eur. J. Neurosci. 25, 99105.
  • Hauge C. and Frodin M. (2006) RSK and MSK in MAP kinase signalling. J. Cell Sci. 119, 30213023.
  • Henshall D. C., Araki T., Schindler C. K., Lan J.-Q., Tiekoter K. L., Taki W. and Simon R. P. (2002) Activation of Bcl-2-associated death protein and counter-response of Akt within cell populations during seizure-induced neuronal death. J. Neurosci. 22, 84588465.
  • Hetman M. (2004) Role of extracellular signal regulated kinases in neuronal death and survival. Eur. J. Biochem. 271, 2049.
  • Hetman M. and Gozdz A. (2004) Role of extracellular signal regulated kinases 1 and 2 in neuronal survival. Eur. J. Biochem. 271, 20502055.
  • Hughes J. P., Staton P. C., Wilkinson M. G., Strijbos P. J. L. M., Skaper S. D., Arthur J. S. C. and Reith A. D. (2003) Mitogen and stress response kinase-1 (MSK1) mediates excitotoxic induced death of hippocampal neurones. J. Neurochem. 86, 2532.
  • Huppi K., Martin S. E. and Caplen N. J. (2005) Defining and assaying RNAi in mammalian cells. Mol. Cell 17, 110.
  • Hutter D., Chen P., Li J., Barnes J. and Liu Y. (2002) The carboxyl-terminal domains of MKP-1 and MKP-2 have inhibitory effects on their phosphatase activity. Mol. Cell Biochem. 233, 107117.
  • Ikonomovic S., Kharlamov E., Manev H., Ikonomovic M. D. and Grayson D. R. (1997) GABA and NMDA in the prevention of apoptotic-like cell death in vitro. Neurochem. Int. 31, 283290.
  • Irving E. A., Barone F. C., Reith A. D., Hadingham S. J. and Parsons A. A. (2000) Differential activation of MAPK/ERK and p38/SAPK in neurones and glia following focal cerebral ischaemia in the rat. Mol. Brain Res. 77, 6575.
  • James A. B., Conway A.-M. and Morris B. J. (2006) Regulation of the neuronal proteasome by Zif268 (Egr1). J. Neurosci. 26, 16241634.
  • Jin K., Mao X. O., Zhu Y. and Greenberg D. A. (2002) MEK and ERK protect hypoxic cortical neurons via phosphorylation of Bad. J. Neurochem. 80, 119125.
  • Kawasaki H., Morooka T., Shimohama S., Kimura J., Hirano T., Gotoh Y. and Nishida E. (1997) Activation and involvement of p38 mitogen-activated protein kinase in glutamate-induced apoptosis in rat cerebellar granule cells. J. Biol. Chem. 272, 1851818521.
  • Kinuya M., Takishima K. and Mamiya G. (2000) Detection of kinases that phosphorylate 14-3-3 binding sites of Raf-1 using in situ gel kinase assay. Bio. Pharm. Bull. 23, 11581162.
  • Lizcano J. M., Morrice N. and Cohen P. (2000) Regulation of BAD by cAMP-dependent protein kinase is mediated via phosphorylation of a novel site, Ser155. Biochem. J. 349, 547557.
  • Merienne K., Pannetier S., Harel-Bellan A. and Sassone-Corsi P. (2001) Mitogen-regulated RSK2-CBP interaction controls their kinase and acetylase activities. Mol. Cell Biol. 21, 70897096.
  • Miranti C., Ginty D., Huang G., Chatila T. and Greenberg M. (1995) Calcium activates serum response factor-dependent transcription by a Ras- and Elk-1-independent mechanism that involves a Ca2+/calmodulin- dependent kinase. Mol. Cell Biol. 15, 36723684.
  • Morris B. J. (1995) Stimulation of immediate-early gene expression in striatal neurones by nitric oxide. J. Biol. Chem. 270, 2474024744.
  • Mukherjee P. K., DeCoster M. A., Campbell F. Z., Davis R. J. and Bazan N. G. (1999) Glutamate receptor signaling interplay modulates stress-sensitive mitogen-activated protein kinases and neuronal cell death. J. Biol. Chem. 274, 64936498.
  • New L., Zhao M., Li Y., Basset W. W., Feng Y., Ludwig S., Padova F. D., Gram H. and Han J. (1999) Cloning and characterization of RLPK, a novel RSK-related protein kinase. J. Biol. Chem. 274, 10261032.
  • Pierrat B., Correia J. S., Mary J. L., Tomas-Zuber M. and Lesslauer W. (1998) RSK-B, a novel ribosomal S6 kinase family member, is a CREB kinase under dominant control of p38 mitogen-activated protein kinase (p38MAPK). J. Biol. Chem. 273, 2966129671.
  • Rakhit S., Clark C. J., O’Shaughnessy C. T. and Morris B. J. (2005) N-Methyl-d-aspartate and brain-derived neurotrophic factor induce distinct profiles of extracellular signal-regulated kinase, mitogen- and stress-activated kinase, and ribosomal S6 kinase phosphorylation in cortical neurons. Mol. Pharmacol. 67, 11581165.
  • Rodriguez-Viciana P., Tetsu O., Tidyman W. E., Estep A. L., Conger B. A., Cruz M. S., McCormick F. and Rauen K. A. (2006) Germline mutations in genes within the MAPK pathway cause cardio-facio-cutaneous syndrome. Science 311, 12871290.
  • Samarsky D. and Datta M. (2004) Expediting target identification and validation through RNAi. Exp. Rev. Mol. Diagn. 4, 1114.
  • Sapkota G. P., Kieloch A., Lizcano J. M., Lain S., Arthur J. S. C., Williams M. R., Morrice N., Deak M. and Alessi D. R. (2001) Phosphorylation of the protein kinase mutated in Peutz–Jeghers cancer syndrome, LKB1/STK11, at Ser431 by p90RSK and cAMP-dependent Protein Kinase, but not its farnesylation at Cys433, is essential for LKB1 to suppress cell growth. J. Biol. Chem. 276, 1946919482.
  • Sassone-Corsi P., Mizzen C. A., Cheung P., Crosio C., Monaco L., Jacquot S., Hanauer A. e. and Allis C. D. (1999) Requirement of Rsk-2 for epidermal growth factor-activated phosphorylation of histone H3. Science 285, 886891.
  • She Q.-B., Ma W.-Y., Zhong S. and Dong Z. (2002) Activation of JNK1, RSK2, and MSK1 is involved in serine 112 phosphorylation of Bad by ultraviolet B radiation. J. Biol. Chem. 277, 2403924048.
  • Shimamura A., Ballif B. A., Richards S. A. and Blenis J. J. (2000) Rsk1 mediates a MEK-MAP kinase cell survival signal. Curr. Biol. 10, 127135.
  • Simpson C. S. and Morris B. J. (2000) Regulation of neuronal cell adhesion molecule (NCAM) expression by NF-kB. J. Biol. Chem. 275, 1687916884.
  • Soloaga A., Thomson S., Wiggin G. R., Rampersaud N., Dyson M. H., Hazzalin C. A., Mahadevan L. C. and Arthur J. S. C. (2003) MSK2 and MSK1 mediate the mitogen- and stress-induced phosphorylation of histone H3 and HMG-14. EMBO J. 22, 27882797.
  • Stein T. D. and Johnson J. A. (2002) Lack of neurodegeneration in transgenic mice overexpressing mutant amyloid precursor protein is associated with increased levels of transthyretin and the activation of cell survival pathways. J. Neurosci. 22, 73807388.
  • Sutton G. and Chandler L. J. (2002) Activity-dependent NMDA receptor-mediated activation of protein kinase B/Akt in cortical neuronal cultures. J. Neurochem. 82, 10971105.
  • Szeberenyi J., Erhardt P., Cai H. and Cooper G. M. (1992) Role of Ras in signal transduction from the nerve growth factor receptor: relationship to protein kinase C, calcium and cyclic AMP. Oncogene 7, 21052113.
  • Takeda K. and Ichijo H. (2002) Neuronal p38 MAPK signalling: an emerging regulator of cell fate and function in the nervous system. Genes cells 7, 10991111.
  • Tan Y., Ruan H., Demeter M. R. and Comb M. J. (1999) p90RSK blocks Bad-mediated cell death via a protein kinase C-dependent pathway. J. Biol. Chem. 274, 3485934867.
  • Thomas G. M., Rumbaugh G. R., Harrar D. B. and Huganir R. L. (2005) Ribosomal S6 kinase 2 interacts with and phosphorylates PDZ domain-containing proteins and regulates AMPA receptor transmission. Proc. Natl Acad. Sci. USA 102, 1500615011.
  • Tomas-Zuber M., Mary J. L. and Lesslauer W. (2000) Control sites of ribosomal S6 kinase B and persistent activation through tumor necrosis factor. J. Biol. Chem. 275, 2354923558.
  • Vincent S. R., Sebben M., Dumuis A. and Bockaert J. (1998) Neurotransmitter regulation of MAP kinase signaling in striatal neurons in primary culture. Synapse 29, 2936.
  • Wang H.-G., Pathan N., Ethell I. M., Krajewski S., Yamaguchi Y., Shibasaki F., McKeon F., Bobo T., Franke T. F. and Reed J. C. (1999) Ca2+-Induced apoptosis through calcineurin dephosphorylation of BAD. Science 284, 339343.
  • West A. E., Griffith E. C. and Greenberg M. E. (2002) Regulation of transcription factors by neuronal activity. Nat. Rev. Neurosci. 3, 921931.
  • Wiggin G. R., Soloaga A., Foster J. M., Murray-Tait V., Cohen P. and Arthur J. S. C. (2002) MSK1 and MSK2 are required for the mitogen- and stress-induced phosphorylation of CREB and ATF1 in fibroblasts. Mol. Cell Biol. 22, 28712881.
  • Xia Z. G., Dudek H., Miranti C. K. and Greenberg M. E. (1996) Calcium influx via the NMDA receptor induces immediate early gene transcription by a MAP kinase/ERK-dependent mechanism. J. Neurosci. 16, 54255436.
  • Xing J., Kornhauser J. M., Xia Z., Thiele E. A. and Greenberg M. E. (1998) NGF activates ERK and p38 MAK kinase pathways to stimulate CREB serine133 phosphorylation. Mol. Cell Biol. 18, 19461955.
  • Zeniou M., Ding T., Trivier E. and Hanauer A. (2002) Expression analysis of RSK gene family members: the RSK2 gene, mutated in Coffin–Lowry syndrome, is prominently expressed in brain structures essential for cognitive function and learning. Hum. Mol. Genet. 11, 29292940.
  • Zhu Y., Yang G.-Y., Ahlemeyer B., Pang L., Che X.-M., Culmsee C., Klumpp S. and Krieglstein J. (2002) Transforming growth factor-beta 1 increases Bad phosphorylation and protects neurons against damage. J. Neurosci. 22, 38983909.