Abstract
Necrosis and apoptosis are two distinct types of mechanisms that mediate ischemic injury. But a signaling point of convergence between them has yet to be identified. Here, we show that activated death-associated protein kinase 1 (DAPK1), phosphorylates p53 at serine-23 (pS23) via a direct binding of DAPK1 death domain (DAPK1DD) to the DNA binding motif of p53 (p53DM). We uncover that the pS23 acts as a functional version of p53 and mediates necrotic and apoptotic neuronal death; in the nucleus, pS23 induces the expression of proapoptotic genes, such as Bax, whereas in the mitochondrial matrix, pS23 triggers necrosis via interaction with cyclophilin D (CypD) in cultured cortical neurons from mice. Deletion of DAPK1DD (DAPK1DDΔ) or application of Tat-p53DM that interrupts DAPK1–p53 interaction blocks these dual pathways of pS23 actions in mouse cortical neurons. Thus, the DAPK1–p53 interaction is a signaling point of convergence of necrotic and apoptotic pathways and is a desirable target for the treatment of ischemic insults.
Introduction
In ischemia, lack of oxygen and glucose depletes the cellular energy and hence causes a wide range of deleterious cellular reactions (Lipton, 1999; Moskowitz et al., 2010). These include loss of ionic homeostasis, production of reactive oxygen species (Lipton, 1999; Aarts et al., 2002; Brennan et al., 2009), production of reactive nitrogen species (Aarts et al., 2002), and mitochondrial dysfunctions (Schinzel et al., 2005; Cheung et al., 2007; Vaseva et al., 2012). As a consequence, cells die via necrosis (i.e., swelling, rupture and cytolysis), apoptosis (i.e., chromatin condensation and DNA breaks), or both, depending in part upon the cell types in the brain (Chan, 2001; Yuan, 2009). However, whether there is a “core signaling program” that controls these two distinct types of cell death pathways needs to be determined.
Necrosis and apoptosis are activated during ischemic injuries through multiple cellular pathways, such as Ca2+ overloading and reactive oxygen species/reactive nitrogen species production (Lipton, 1999; Lo et al., 2005; Yuan, 2009; Moskowitz et al., 2010), and are involved in an excessive stimulation of glutamate transmitter receptors (Lipton and Rosenberg, 1994; Hardingham and Bading, 2010; Moskowitz et al., 2010). However, these receptors also function physiologically in the regulation of gene expression, plasticity, and neuronal survival and hence are not desirable targets in ischemic neuronal death (Ginsberg, 2009). We recently reported that death-associated protein kinase 1 (DAPK1) constitutes a specific cell death signaling molecule that is directly linked to glutamate receptor channels (Tu et al., 2010). DAPK1 is a newly identified Ca2+/calmodulin-dependent protein kinase (Bialik and Kimchi, 2006) and plays a role in several modes of cell death, including apoptosis and autophagy (Raveh et al., 2000; Van Eldik, 2002). However, downstream signaling of DAPK1 that specifically mediates ischemic injuries remains unknown.
Here, we identified a direct binding of the DAPK1 death domain (DAPK1DD) to the DNA binding motif (DM) of the tumor suppressor protein p53 (p53DM). We synthesized a membrane-permeable p53DM peptide consisting of amino acid residues 270–281 (p53DM) that specifically blocks DAPK1DD-p53DM binding in mouse cortical neurons. We also devised multiple genetic deletion/rescue experiments in the cultured cortical neurons from mice and demonstrated that an interaction between DAPK1 and p53 is a signaling point of convergence of necrosis and apoptosis in ischemic insults.
Materials and Methods
Yeast two-hybrid screen.
Yeast two-hybrid analysis was performed using the L40 yeast strain, which carries reporter genes His3 and LacZ under the control of upstream LexA binding sites. A cDNA encoding DAPK1DD (residues 1318–1393) was cloned from full-length DAPK11–1430 (Accession: NP_598823) into pBHA (LexA-DD fusion vector) as bait and VP16 as an activation domain (AD) for screening a mouse brain cDNA library, constructed in pGAD10 (GAL4 activation domain vector, Clontech). A total of 1 × 107 transformants were screened. Colonies on plates lacking histidine (supplemented with 2.5 mm of 3-aminotriazole) with a blue color with X-gal were selected for plasmid recovery, as described previously (Tu et al., 2010).
Coimmunoprecipitation and GST affinity binding assays.
The cerebral cortex from adult male C57BL/6J mice at 4 months old of age was homogenized in ice-cold lysis buffer containing (mm) the following: 50 Tris-HCl, 150 NaCl, 1% NP-40, 2 EDTA, 1 Na-orthovanadate, pH 7.4, and proteinase inhibitor mixture (5 μl/100 mg tissue, Sigma). After clearing debris by centrifugation at 14,000 × g at 4°C, protein concentration in the extracts was determined using the Bradford assay (Bio-Rad), as described previously (Tu et al., 2010). The nuclear/mitochondrial fractions were prepared using cell fractionation kits (BioVision), according to the manufacturer's instructions. The extracts (500 μg proteins) were incubated with nonspecific IgG (IgG, 2 μg) or polyclonal rabbit anti-DAPK1 (2 μg; Thermo Fisher Scientific, catalog #PA5-17055) or anti-p53 (2 μg, Abcam, catalog #ab26) overnight at 4°C, followed by the addition of 40 μl of Protein G-Sepharose (Sigma) for 3 h at 4°C. The precipitates were washed four times with lysis buffer and denatured with SDS sample buffer and separated using 12% SDS-PAGE. The proteins were transferred onto nitrocellulose membranes using a Bio-Rad mini-protein-III wet transfer unit overnight at 4°C. Transfer membranes were then incubated with blocking solution (5% nonfat dried milk dissolved in TBST buffer containing 10 mm Tris-HCl, 150 mm NaCl, and 0.1% Tween-20 for 1 h at room temperature, washed three times, and incubated with anti-goat primary antibody against anti-DAPK1 (Epitomics, catalog #3798-1) or anti-p53 (Millipore, catalog #04-1083) for 1 h at room temperature. The membranes were washed three times with TBST buffer and incubated with the appropriate secondary antibodies (1:1000 dilution) for 1 h followed by washing four times. Signal detection was performed using an enhanced chemiluminescence kit (GE Healthcare). The lanes marked “input” were loaded with 10% of the starting material used for immunoprecipitation.
To determine a binding sequence of DAPK1 DD with p53 protein, the GST-p53 deletion mutants (GST-p531–61, GST-p5361–126, GST-p53126–241, GST-p53241–281, and GST-p53281–390) were generated from full-length cDNA mouse p531–390 (Addgene, pLCRcala (p53), plasmid 12136), as we described previously (Tu et al., 2010; Yang et al., 2012). Purified GST fusion proteins were separated using SDS-PAGE and transferred onto a nitrocellulose membrane, which was washed with distilled water and blocked with TBST for 1 h at room temperature. The membrane was then incubated with affinity binding buffer containing 50 mm Tris-HCl, pH 7.5, 200 mm NaCl, 12 mm α-mercaptoethanol, 1.0% polyethylene glycol, 10 μg/ml protease inhibitors, and 500 μg/ml purified Flag-tagged DAPK1 DD (Invitrogen) for 1 h at room temperature and washed four times for 5 min with affinity binding buffer. Bound DAPK1DD was detected with anti-Flag (1: 2000, Invitrogen).
Luciferase reporter assays.
The p53-response luciferase construct that encodes firefly luciferase reporter genes under the control of a minimal CMV promoter and tandem repeats of the p53 transcriptional response element was used to study the transcriptional activity of p53 (QIAGEN). cDAPK1, which lacks a calmodulin regulatory domain (amino acids 266–312), and cDAPK1DDΔ were generated as we described previously (Tu et al., 2010). Cells were transiently cotransfected with the indicated expression plasmids and reporter constructs in HEK293 cells. A pool of three specific 19–25 nt p53 siRNA and the scrambled control (s-siRNA) were purchased from Santa Cruz Biotechnology (sc-29435). At 48 h after transfection, whole-cell lysates were prepared, and luciferase activity was assessed using the Promega dual luciferase assay system. The firefly luminescence signal was normalized to Renilla luminescence signal. The cell lysates were also prepared and probed with antibodies against p53 and GAPDH.
Cortical cultures and cell death analysis.
The male heterozygous mutant mice with a deficiency in the expression of DAPK1 (DAPK1+/− mice with the C57BL/6J background) (Tu et al., 2010), p53 (p53+/− mice with the 129Sv background, The Jackson Laboratory), cyclophilin D (CypD+/− mice with the C57BL/6–129 mixed background, The Jackson Laboratory), or Bax gene (Bax+/− mice with the C57BL/6 background, The Jackson Laboratory) were bred with their respective heterozygous female mice. The cerebral cortex was isolated from the homozygous E20 embryos, including DAPK1−/−, p53−/−, CypD−/−, or Bax−/− and the respective wild-type controls (DAPK1+/+, p53+/+, CypD+/+, and Bax+/+ embryos, as described previously) (Tu et al., 2010; Yang et al., 2012). Cells were dissociated and purified using a papain dissociation kit (Worthington Biochemical) plated at a density of 100–150 cells/mm2 on 19 mm coverslips coated with 30 μg/ml poly-d-lysine and 2 μg/ml laminin. The cells were placed in fresh serum-free Neurobasal Medium (21103, Invitrogen) plus 2% B27, and the medium was replaced every 4 d. Some cultures were stained for β-tubulin III (Tuj1), a neuronal marker to confirm that these cultures were >85% neurons.
On day 10 (DIV10), wDAPK1, cDAPK1, cDAPK1DDΔ, wild-type p53, or mutant p53A23 was expressed using the respective rAAV1/2- virus particles by the addition of 1 μl of the virus particles (9 × 1012 genomic particle/ml) in the 200 culture media. At 72 h after infection, a transgene was expressed in >90% neurons. The cultures were stained with propidium iodide (PI, Sigma) 3 d (DIV13) or with TUNEL (Millipore) 5 d (DIV15) after virus particle infections. In some experiments, 5 μm Tat-p53DM or Tat-s-p53DM, or 50 μm BIP-V5 (Millipore), or 10 μm pifithrin-α (Sigma) was applied 2 d after virus infections. Tat-p53DM (Tat-RVCACPGRDRRT) or Tat-s-p53DM (Tat-CCPGECVRTRRR) peptides with 99% purity were synthesized by AnaSpec. The peptides were numbered and the experimenters were unaware of which one was used in all experiments.
Cloning and generation of rAAV1/2 virus particles.
cDNA encoding p53 in pLCRcala (Eddgene, plasmid 12136) or p53A23 in pBluescript II SK (Addgene, plasmid 12167) was used as the template for PCR amplification. The PCR products were digested using SalI and HindIII. The resulting fragment was ligated to a SalI-digested rAVE-GFP vector (GenDetect) to generate rAVE-p53-GFP, or rAVE-p53A23-GFP under the control of a CAG promoter and was terminated using the polyadenylation signal in the 3′ long terminal repeat. Downstream of GFP was a woodchuck hepatitis virus regulatory element that enhances the transgene expression. The rAVE-CAG-p53-GFP or rAVE-CAG-p53A23-GFP was cotransfected with the AAV helper 1/2 into HEK-293 cells to generate high titers (9 × 1012 genomic particles/ml) of rAAV1/2 infectious particles. Cloning and generation of lenti virus (pLenti) particles for the expression of wDAPK1, cDAPK1, or cDAPK1DDΔ were described previously (Tu et al., 2010).
Real-time PCR.
RNA was isolated using TRIzol (Invitrogen) from the cultured neurons 72 h after transfection of the respective genes (indicated in the figure legends) and reverse transcribed using MMLV reverse transcriptase (Invitrogen) and random primers. PCR was performed in triplicate using SYBR Green (SA-Biosciences) and a 7900HT Fast Real-Time PCR machine (Applied Biosystems). The following primers (5′ to 3′) were used: Bax forward, TGAAGACAGGGGCCTTTTTG and Bax reverse, AATTCGCCGGAGACACTCG; Puma forward, AGCAGCACTTAGAGTCGCC and Puma reverse, CCTGGGTAAGGGGAGGAGT; and p53 forward, CTCTCCCCCGCAAAAGAAAAA and p53 reverse, CGGAACATCTCGAAGCGTTTA. The results were computed using a standard curve made with cDNA pooled from all samples.
Oxygen and glucose deprivation (OGD).
The cultured neurons were transferred to an anaerobic chamber containing a 5% CO2, 10% H2, and 85% N2 atmosphere. The cells were then washed 3× with 500 μl of deoxygenated glucose-free bicarbonate solution and maintained anoxic at 37°C for 60 min and 90 min, respectively. OGD was terminated by washing the cultures with oxygenated glucose-containing (20 mm) bicarbonate solution. The cultures were maintained for 72 h at 37°C in a humidified 5% CO2/10% H2/85% O2 atmosphere. The cultures were used for Western blotting, real-time PCR, and PI and TUNEL staining. The numbers of PI- and TUNEL- labeled cells were expressed as a percentage of total numbers of DAPI-labeled cells per condition.
Statistical analysis.
All variance values in the text and figure legends are presented as mean ± SEM. Parametric tests, including t test and ANOVA, were used where assumptions of normality and equal variance (F test) were met. In some analyses, Student-Newman-Keuls post hoc tests after ANOVA were performed, as described in the respective figure legends.
Animal care.
Male adult (4 months old of age) C57BL/6J mice were used in this study. Care and experiments with mice were in accordance with institutional guidelines of the Animal Care and Use Committee (Huazhong University of Science and Technology, Wuhan, China).
Results
DAPK1DD binds to p53DM directly
To identify a specific cell death signaling molecule downstream of DAPK1, we performed yeast two-hybrid analysis using DAPK1DD as bait. The screen of a mouse whole-brain cDNA library detected p53, extracellular signal-regulated kinase (ERK) and Tau as interacting substrates of DAPK1 (Fig. 1A). Among them, p53 is a transcriptional regulator and controls the program of necrotic and apoptotic pathways of cell death (Copani et al., 2001; Green and Kroemer, 2009; Maiuri et al., 2009; Vaseva et al., 2012). To validate this interaction in brain cells in vivo, we precipitated the endogenous DAPK1 complex in the cellular extracts of the adult mouse cerebral cortex using antibodies against DAPK1 and p53. We probed the precipitates with antibodies against p53, ERK, or Tau and showed the existence of endogenous p53, ERK, and Tau proteins in the DAPK1 complex (Fig. 1B). These findings demonstrated that DAPK1 was physically associated with p53, ERK, and Tau proteins.
To map a binding region of DAPK1DD with p53, we generated a glutathione S-transferase (GST) protein fused to p53 (GST-p53) and GST-p53 deletion mutants, including GST-p531–61, GST-p5361–126, GST-p53126–241, GST-p53241–281, and GST-p53281–390 (Fig. 1C). We performed an affinity binding assay by incubating the GST fusion proteins with a Flag-tagged DAPK1DD in vitro. We found that DAPK1DD bound to a p53 DNA binding motif consisting of amino acid 241–281 (p53DM241–281; Fig. 1D).
We subsequently screened a series of p53DM241–281 deletions and identified a peptide consisting of amino acid 270–281 (270RVCACPGRDRRT281, p53DM) in the p53DM241–281 fragment as a minimal binding region of DAPK1DD (Fig. 1E). The direct binding of DAPK1DD to p53DM was verified using peptide blocking experiments (Fig. 1F,G). In these experiments, a synthesized p53DM peptide at 5 μm completely blocked the binding of Flag-tagged DAPK1DD to GST-p53DM in vitro but did not affect DAPK1 association with ERK or Tau protein (Fig. 1H). These data demonstrated that DAPK1 interacted with p53 in brain cells in vivo via a direct binding of DAPK1 DD to p53DM.
DAPK1 functionally interacts with p53
To determine whether DAPK1 functionally interacts with p53, we used a p53 transcriptional response element reporter system to analyze p53 targeting genes in HEK293 cells. We coexpressed constitutively active DAPK1 (cDAPK1) lacking an auto-inhibitory domain of DAPK1 (amino acids 266–312), or wild-type DAPK1 (wDAPK1) with the reporter in HEK293 cells. We found that cDAPK1, as an active form of DAPK1, induced a fivefold increase of the reporter activity (Fig. 2A). This increase was abolished by knocking down of endogenous p53 using a small interference RNA (p53siRNA) that specifically targets to the p53 gene. A scrambled p53siRNA (s-siRNA) was used as a negative control. Deletion of cDAPK1DD (cDAPK1DDΔ), which eliminates cDAPK1–p53DM binding, was ineffective to alter the p53 reporter genes. Because p53 protein was not changed when cDAPK1 was expressed (Fig. 2B), cDAPK1 activated the p53 reporter via a direct binding of DAPK1DD to p53DM.
DAPK1 phosphorylates p53 at serine-23 (pS23)
DAPK1 is considered a kinase of human p53 at serine-20 (pS20) (Craig et al., 2007), a residue corresponding to mouse p53 at serine-23 (p53S23, 19QETFSGL25). We thus tested whether activated DAPK1 phosphorylates p53S23 (pS23) in the cortical neurons from mice. We reconstituted the p53−/− cortical neurons with the expression of cDAPK1 and a wild-type p53 or its mutant p53A23. We also developed an antibody against pS23. Our results demonstrated that anti-pS23 specifically recognized pS23, but not mutant p53A23 (Fig. 2C). The activated cDAPK1, but not inactivated wDAPK1, induced pS23 (Fig. 2D). We also expressed cDAPK1DDΔ and added a membrane-permeable p53DM peptide (Tat-p53DM) generated by fusing p53DM to the cell-membrane transduction domain of the HIV-1 Tat protein. We found that the expression of cDAPK1DDΔ or application of Tat-p53DM effectively blocked the direct binding of DAPK1DD to p53DM and completely eliminated anti-pS23 signals in the cytosolic fraction of the proteins, whereas a scrambled Tat-s-p53DM did not (Fig. 2D,E). In addition to p53, DAPK1 interacted with ERK and Tau proteins. Thus, it is likely that pS23 was phosphorylated indirectly via DAPK1-ERK or DAPK1-Tau interactions. To address this possibility, we expressed small interference RNA (siRNA) that specifically silences the ERK (E-siRNA) or Tau (T-siRNA) gene in the cultured neurons when cDAPK1 was expressed. s-siRNA was expressed as a control. We found that neither E-siRNA nor T-siRNA altered pS23 level in the cytoplasmic fraction of the cultured neurons in the presence of cDAPK1 (Fig. 2E). Previous studies indicated that expression of exogenous p53 is sufficient to cause the death of cultured neurons (Slack et al., 1996; Xiang et al., 1996; Jordán et al., 1997). Accordingly, we hypothesized that the expression of exogenous p53 may induce pS23 production. To test this hypothesis, we introduced exogenous p53 gene in the p53+/+ and p53−/− neurons, respectively, and showed that it did not induce pS23 expression (Fig. 2F). Thus, cDAPK1 catalyzes p53 into the pS23 via the direct binding of cDAPK1DD to p53DM.
pS23 nuclear translocation induces proapoptotic gene expression
p53 is a transcriptional regulator and controls proapoptotic gene expression (Brady et al., 2011; Vaseva et al., 2012). We next determined whether pS23 represents a functional version of p53. First, we examined its subcellular distributions. We purified a nuclear fraction from the cortical cells. Western bloting revealed that the pS23, p53, and a nuclear matrix protein p84 were enriched in the nucleus of the p53+/+ neurons expressing cDAPK1 (Fig. 2G). In contrast to pS23, DAPK1 protein was largely restricted to the cytoplasm (Fig. 2D). Thus, activated DAPK1 catalyzed p53 into pS23 outside the nucleus. The presence of pS23 in the nucleus reflects pS23 nuclear translocation. We also stained the cultured neurons with antibodies against the pS23 and p53 (Fig. 2H) or mitochondrial protein (COX-IV, Fig. 2I). Our results verified that the pS23 was located in both the mitochondria and the nucleus in the cultured cortical neurons when cDAPK1 was expressed (Fig. 2H,I). Of note, p53 protein was detectable in the nucleus in some cultured cortical neurons under the basal condition, although it seemed to be absent in the nucleus in some cultured neurons (Fig. 2H). This result is consistent with the results in Figure 2G, showing the presence of p53 protein in the nucleus of a pool of the cultured neurons.
Then, we determined the transcriptional activity of pS23 in cortical neurons through analysis of p53 proapoptotic genes. Our data revealed that Bax and Puma genes (Fig. 3A–E), but not p53 mRNA (Fig. 3F), were expressed when cDAPK1, but not wDAPK1, was reconstituted in the p53+/+ neurons. Deletion of the p53 gene (p53−/− neurons) or reconstitution of p53A23 in the p53−/− neurons blocked the effects of cDAPK1. We found that application of pifithrin-α, but not with the vehicle control, completely suppressed the cDAPK1-induced expression of Bax and Puma genes (Fig. 3C,E), but not pS23 (Fig. 3G). Thus, cDAPK1 activated the p53 proapoptotic genes via activation of pS23 transcription. We also treated the cortical cultures with a DNA-damaging agent (camptothecin [CPT]). Consistent with previous studies (Uo et al., 2007; Brochier et al., 2013), we found that CPT at 10 μm induced the expression of Bax genes (Fig. 3H). However, when a mutant p53A23 was introduced into the p53−/− neurons, it did not compromise the CPT effects, showing that CPT induced the expression of p53 proapoptotic genes via the pS23-independent mechanism.
pS23 induces apoptosis via Bax expression
Bax controls the programs of apoptosis through interaction with Bcl2 family proteins at the outer mitochondrial membrane, resulting in Permeability Transition Pore (PTP) opening (Morris et al., 2001; Steckley et al., 2007; Meulmeester and Jochemsen, 2008). Consistent with this notion, we showed that cDAPK1 activated PTP (Fig. 4A). Application of Tat-p53DM that directly intercepts DAPK1DD-p53DM binding completely inhibited the effects of cDAPK1 (Fig. 4B), Deletion of the p53 gene (p53−/−) produced the same degree of inhibition as p53DM. This inhibition was completely reversed by the introduction of wild-type p53 gene, but not its mutant p53A23, into the p53−/− neurons, but this was partially affected by application of Bax inhibitor or deletion of Bax gene (Fig. 4C). Thus, the cDAPK1-induced activation of PTP was mediated by pS23 through the direct binding of cDAPK1DD to p53DM, but was partially dependent on Bax expression. The PTP opening was associated with cytochrome-c (Cyto-c) release (Fig. 4D), Caspase-3 (Casp3) cleavage (Fig. 4E,F), and TUNEL labeling (Fig. 4G,H). Expression of exogenous p53 in the p53−/− neurons alone in the absence of cDAPK1 caused no neuronal apoptosis (Fig. 4H). This observation is, however, inconsistent with the previous reports that the expression of an exogenous p53 gene is sufficient for the induction of neuronal apoptosis (Slack et al., 1996; Xiang et al., 1996; Jordán et al., 1997). This inconsistency could be due to the different p53 genes used in the different studies; the mouse p53 gene was used in the present study, whereas the human p53 gene was used in the previous reports. Thus, our results demonstrated that pS23 was an essential substrate of cDAPK1 and activated apoptosis by expression of the p53 proapoptotic genes. As mentioned above, application of pifithrin-α that blocks p53 transcription or of a Bax Inhibitory Peptide (BIP-V5) that directly inhibits Bax, or deletion of Bax gene (Bax−/−), produced only partial inhibition of PTP (Fig. 4C), although each of these treatments was able to block Cyto-c release (Fig. 4D), Casp3 activation (Fig. 4E,F), and neuronal apoptosis (Fig. 4G,H). Thus, PTP activation by pS23 involved both Bax-dependent and -independent mechanisms.
pS23 interacts with CypD
p53 interacts with CypD in the mitochondrial matrix and activates PTP, leading to cell necrosis (Vaseva et al., 2012). Accordingly, we focused our studies on an interaction of pS23 with CypD. We coexpressed cDAPK1 in the p53−/− cortical cells with wild-type p53 or its mutant p53A23. Mitochondrial fractions were then purified from these cortical cultures. To verify the purification of individual mitochondrial fractions, we used anti-VDAC (voltage-dependent anion channel; Fig. 5A) for mitochondrial (Mit) membrane and anti-uMTCK (ubiquitous mitochondrial creatine kinase, Fig. 5B) for the outer membrane (OM), and anti-CypD for the inner membrane (IM; Fig. 5C) loading controls. We found that pS23 was enriched in the mitochondrial IM when cDAPK1 and p53 were both expressed (Fig. 5A–D). Incubation of the cultured neurons with 5 μm Tat-p53DM, but not Tat-s-p53DM blocked the pS23 expression in the mitochondrial IM (Fig. 5C). Anti-CypD antibody precipitated endogenous pS23 in the IM fraction from the p53+/+, but not p53−/− neurons when cDAPK1 was expressed (Fig. 5E). We subsequently used antibodies against p53 (Fig. 5F) or pS23 (Fig. 5G) to precipitate endogenous CypD. Blots of the precipitates revealed that endogenous CypD was physically associated with both p53 and pS23 proteins in CypD+/+ neurons (Fig. 5F,G). This association was not affected by inhibition of p53 transcription through application of pifithrin-α (Fig. 5H). Furthermore, we found that deletion of CypD gene produced partial inhibition of PTP activation (Fig. 5I), but it did not affect Cyto-c release and Casp3 activation (Fig. 5J). Thus, these results showed that pS23 was physically associated with CypD in the mitochondrial inner membrane compartments. But this association was not required for Cyto-c release and neuronal apoptosis in response to DAPK1 activation.
pS23 induces cell necrosis via interaction with CypD
We next tested whether the pS23–CypD association induces neuronal necrosis. We expressed cDAPK1 in the cultured cortical neurons. Three days later, the cultures were stained with PI alone (Fig. 6A,B) or together with TUNEL (Fig. 6C). We found that both PI and TUNEL labeling was effectively protected by deletion of p53 gene (p53−/−; Fig. 6D). Significantly, this protection was completely reversed by reconstitution of a wild-type p53, but not a mutant p53A23 in p53−/− mice (Fig. 6D), showing the pS23-dependent process. Inhibition of Bax by application of BIP-V5 or inhibition of p53 transcriptional activity by application of pifithrin-α partially inhibited PI labeling (Fig. 6B), suggesting that pS23 activates neuronal necrosis via a Bax-independent process. To determine a role of CypD in Bax-independent necrosis, we cultured cortical neurons lacking the CypD gene (CypD−/− cortical neurons). We found that CypD−/− neurons were largely resistant to the pS23-induced PI labeling (Fig. 6D), although they remained to be vulnerable to TUNEL labeling (Fig. 6D). Thus, CypD is a substrate of pS23 for cDAPK1-induced necrosis. Previous studies showed that expression of exogenous p53 was necrotic to cortical neurons. It is likely that p53 shares the similar effects as cDAPK1 for the induction of neuronal necrosis. In line with this idea, we showed that deletion of CypD gene completely blocked p53-induced PI labeling (Fig. 6E). When Tat-p53DM that interrupts DAPK1–p53 interaction was applied, it protected against cDAPK1, but not p53-induced cell death (Fig. 6E). Together, these results revealed that p53 constituted a substrate of DAPK1 for the induction of CypD-dependent necrosis and Bax-dependent apoptosis.
DAPK1–p53 interaction mediates OGD injury
Having determined that the pS23 acts as a signaling point of convergence of both necrotic and apoptotic pathways of cell death, we next determined whether it is involved in ischemic injuries. We challenged the DAPK1+/+ and DAPK1−/− neurons (DIV14), in which wild-type p53 or its mutant p53A23 was reconstituted, with OGD for 60 or 90 min. OGD is widely considered an in vitro cellular model of ischemic stroke (Tu et al., 2010). At 24, 30, 36, or 48 h after 60 min OGD treatment, the cultures were used for the measurement of pS23 accumulation. We found that pS23 level increased 24 h after OGD challenge and peaked at 48 h later. pS23 accumulation was observed in DAPK1+/+, but not in DAPK1−/−, cells (Fig. 7A,B) and was associated with no change in p53 mRNA (Fig. 7C) and protein levels (data are not shown). Thus, OGD challenge activated DAPK1 and in turn causes pS23, but not p53, expression in the cultured cortical neurons. After pS23 expression, the cultures were stained with PI together with TUNEL (Fig. 7D,E). In case of 60 min OGD, 46% cells were necrotic, according to PI labeling and 21% cells were apoptotic, according to TUNEL labeling (Fig. 7F). When the cultured cells were challenged with 90 min OGD, 89% cells were labeled with PI and 18% cells were labeled with TUNEL (Fig. 7F), showing that the majority of cells were killed by OGD challenge. Pretreatment of the cultures with 10 μm Tat-p53DM effectively protected the cultures from PI and TUNEL labeling (Fig. 7F). The similar protection from OGD injury was observed in the cultured p53−/− neurons (Fig. 7D–F). Thus, the DAPK1–p53 interaction was indispensable for both cell necrosis and apoptosis in ischemic insults.
Discussion
In the present study, we showed that activated DAPK1 catalyzes p53 into pS23 via a direct binding of DAPK1DD to p53DM. We demonstrated that pS23 represents a functional mode of p53, essential for both transcription-dependent and -independent pathways of apoptosis and necrosis. We also revealed that a small membrane-permeable peptide inhibitor (Tat-p53DM) effectively intercepted the DAPK1–p53 interaction in the cortical neurons. Thus, our present study has not only uncovered a core cell death signaling complex that merges the dual pathways of brain cell death in ischemia but also identified a desirable cell death inhibitor against ischemic injuries.
The p53 protein plays a role in several modes of cell death depending in part upon its subcellular distributions (Meulmeester and Jochemsen, 2008; Brady et al., 2011; Vaseva et al., 2012). In the nucleus, p53 is a crucial transcriptional regulator and controls the program of transcription-dependent apoptosis in response to various signals, including DNA damage, oxidative stress, and ischemia (Meulmeester and Jochemsen, 2008). The main actions of p53 in cell apoptosis involve transcriptional induction of proapoptotic genes, such as Bax, Puma, and Noxa, that in turn trigger mitochondrial pathways, including Cyto-c release and Casp3 accumulation (Green and Kroemer, 2009). Recently, it was shown that, after stress, the cytoplasmic pool of p53 rapidly relocates onto the mitochondrial surface, where it physically interacts with proapoptotic Bcl2 family members, leading to PTP opening and apoptosis (Mihara et al., 2003). Our present study demonstrated that p53 can be phosphorylated at serine-23 through a direct binding of activated DAPK1 to p53DM. The experiments with reconstitution of wild-type p53 or its mutant p53A23 in p53−/− neurons revealed that pS23 represents a functional mode of p53 in the nucleus that activates proapoptotic genes, such as Bax and Puma in cortical cells. Bax is one of the earliest-studied p53 gene targets that interact with mitochondrial proapoptotic signals (Morris et al., 2001; Steckley et al., 2007; Walensky and Gavathiotis, 2011). Consistent with these previous studies, our results showed that the induction of neuronal apoptosis by pS23 is completely abolished by deletion of Bax gene or application of a Bax inhibitor, BIP-V5. On the other hand, our data exhibited that pS23 is absent in the mitochondrial outer membrane components. Thus, the induction of Bax expression by pS23 possibly reflects an essential event for neuronal apoptosis in response to DAPK1 activation.
In our present study, we showed that recruitment of pS23 by cDAPK1 into the nucleus activates the mitochondrial PTP-dependent programs of apoptosis. Our data also revealed that PTP activation by pS23 is partially affected by deletion of Bax gene or inhibition of Bax via application of BIP-V5. This finding indicates that pS23 regulates mitochondrial PTP through both Bax-dependent and -independent mechanisms. Consistent with this idea, our results demonstrated that pS23 is indeed transferred into the inner mitochondrial membrane and interacts with CypD, which is known as a key regulator of mitochondrial PTP (Vaseva et al., 2012).
Through activation of PTP at the inner membrane, CypD controls the programs of cell necrosis via activation of mitochondrial PTP (Baines et al., 2005; Basso et al., 2005; Schinzel et al., 2005). Necrosis is a rapid loss of the cellular membrane potential because of energy depletion and ion pump/channel failure and is a central cellular event in ischemia/reperfusion injuries. Earlier studies showed that CypD−/− neurons are resistant to ischemia-induced necrosis in a mouse model of stroke (Vaseva et al., 2012). Of note, these CypD-deficient neurons remain vulnerable to Bax-dependent apoptosis. In line with these earlier studies, our data showed that the DAPK1–p53 interaction and subsequent pS23 accumulation evoke two functionally distinct mitochondrial death pathways: transcription-dependent Bax expression and transcription-independent pS23–CypD interaction at the mitochondrial inner membrane, as illustrated in Figure 7G.
Our previous studies revealed that DAPK1 is linked to excitatory glutamate receptor channels and becomes activated in response to ischemic insults (Tu et al., 2010). Here, we demonstrated that the signaling events downstream of DAPK1 involve transcription-dependent and -independent pathways through the expression of Bax gene or interaction with CypD in cortical neurons. Accordingly, we synthesized the membrane-permeable peptide Tat-p53DM that showed protective effect against ischemic injury in the cultured cortical neurons in vitro. Necrosis and apoptosis are two distinct mechanisms that mediate ischemic injury in brain cells. Depending on the challenges of the cultured neurons to OGD (60 or 90 min OGD), the percentage of cells undergoing necrosis versus apoptosis varies. In some subsets of cells, we observed some overlap between apoptosis and necrosis (TUNEL+/PI+) in response to OGD treatment (Fig. 7D,E). As interception of DAPK1–p53 interaction effectively blocks these two distinct pathways of neuronal death in ischemia, it is likely to be a preferred therapeutic target for ischemic stroke.
Footnotes
This work was supported by National Natural Science Foundation of China Grants 81130079 and 91232302 to Y.L., 81200863 to L.P., 81271270 to Y.S., and 81361120245 and Ministry of Science and Technology of China Grant 2011DFG33250 to L.-Q.Z. We thank Ruojian Wen (Huazhong University of Science and Technology) for cell cultures, Kunpeng Zhao (Huazhong University of Science and Technology) for cell death assays and real-time PCR, and Qing Tian for comments on the manuscript.
The authors declare no competing financial interests.
- Correspondence should be addressed to either Dr. Ling-Qiang Zhu or Youming Lu, Department of Physiology, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, China, zhulq{at}hust.edu.cn or lym{at}hust.edu.cn