Coordinated regulation of p53 apoptotic targets BAX and PUMA by SMAR1 through an identical MAR element
2010; Springer Nature; Volume: 29; Issue: 4 Linguagem: Inglês
10.1038/emboj.2009.395
ISSN1460-2075
AutoresSurajit Sinha, Sunil K. Malonia, Smriti Mittal, Kamini Singh, Kadreppa Sreenath, Rohan Kamat, Robin Mukhopadhyaya, Jayanta K. Pal, Samit Chattopadhyay,
Tópico(s)MicroRNA in disease regulation
ResumoArticle14 January 2010free access Coordinated regulation of p53 apoptotic targets BAX and PUMA by SMAR1 through an identical MAR element Surajit Sinha Surajit Sinha National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Sunil Kumar Malonia Sunil Kumar Malonia National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Smriti P K Mittal Smriti P K Mittal National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Department of Biotechnology, University of Pune, Pune, India Search for more papers by this author Kamini Singh Kamini Singh National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Sreenath Kadreppa Sreenath Kadreppa National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Rohan Kamat Rohan Kamat Virology Lab, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Navi Mumbai, India Search for more papers by this author Robin Mukhopadhyaya Robin Mukhopadhyaya Virology Lab, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Navi Mumbai, India Search for more papers by this author Jayanta K Pal Jayanta K Pal Department of Biotechnology, University of Pune, Pune, India Search for more papers by this author Samit Chattopadhyay Corresponding Author Samit Chattopadhyay National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Surajit Sinha Surajit Sinha National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Sunil Kumar Malonia Sunil Kumar Malonia National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Smriti P K Mittal Smriti P K Mittal National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Department of Biotechnology, University of Pune, Pune, India Search for more papers by this author Kamini Singh Kamini Singh National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Sreenath Kadreppa Sreenath Kadreppa National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Rohan Kamat Rohan Kamat Virology Lab, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Navi Mumbai, India Search for more papers by this author Robin Mukhopadhyaya Robin Mukhopadhyaya Virology Lab, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Navi Mumbai, India Search for more papers by this author Jayanta K Pal Jayanta K Pal Department of Biotechnology, University of Pune, Pune, India Search for more papers by this author Samit Chattopadhyay Corresponding Author Samit Chattopadhyay National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India Search for more papers by this author Author Information Surajit Sinha1, Sunil Kumar Malonia1, Smriti P K Mittal1,2, Kamini Singh1, Sreenath Kadreppa1, Rohan Kamat3, Robin Mukhopadhyaya3, Jayanta K Pal2 and Samit Chattopadhyay 1 1National Centre for Cell Science (NCCS), Pune University Campus, Ganeshkhind, Pune, India 2Department of Biotechnology, University of Pune, Pune, India 3Virology Lab, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Navi Mumbai, India *Corresponding author. National Center for Cell Science, Pune University Campus, Ganeshkhind, Pune 411007, India. Tel.: +91 20 2570 8152; Fax: +91 20 2569 2259; E-mail: [email protected] The EMBO Journal (2010)29:830-842https://doi.org/10.1038/emboj.2009.395 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info How tumour suppressor p53 bifurcates cell cycle arrest and apoptosis and executes these distinct pathways is not clearly understood. We show that BAX and PUMA promoters harbour an identical MAR element and are transcriptional targets of SMAR1. On mild DNA damage, SMAR1 selectively represses BAX and PUMA through binding to the MAR independently of inducing p53 deacetylation through HDAC1. This generates an anti-apoptotic response leading to cell cycle arrest. Importantly, knockdown of SMAR1 induces apoptosis, which is abrogated in the absence of p53. Conversely, apoptotic DNA damage results in increased size and number of promyelocytic leukaemia (PML) nuclear bodies with consequent sequestration of SMAR1. This facilitates p53 acetylation and restricts SMAR1 binding to BAX and PUMA MAR leading to apoptosis. Thus, our study establishes MAR as a damage responsive cis element and SMAR1–PML crosstalk as a switch that modulates the decision between cell cycle arrest and apoptosis in response to DNA damage. Introduction The tumour suppressor p53 is the cellular sentinel of the mammalian cell cycle and an indispensable component of the DNA damage response pathway. Activation of p53 in response to DNA damage results in either cell cycle arrest or apoptosis. Although genes that regulate cellular processes such as arrest and apoptosis are essentially p53 targets, activation of p53 always results in specific and selective transcription of p53-regulated genes (Riley et al, 2008). Thus, it is likely that unique sets of p53-regulated genes operate in tandem to bring about a desired outcome in response to specific stimuli. How p53 executes these two distinct functions in a promoter-specific manner remains largely unclear. Recent reports suggest that activation of specific promoters by p53 is achieved through its interaction with heterologous transcription factors such as Hzf, human cellular apoptosis susceptibility (hCAS)/CSE1L and ASPP family proteins (Samuels-Lev et al, 2001; Das et al, 2007; Tanaka et al, 2007). In addition, under conditions of stress, different phosphorylation and acetylation modules stabilize p53 enhancing its sequence-specific DNA binding and transcriptional activity (Sakaguchi et al, 1998). Although phosphorylation at amino-terminus is required for p53 stability, acetylation at carboxyl-terminus has been shown to be indispensable for p53 transcriptional activation (Tang et al, 2008). The tumour suppressor promyelocytic leukaemia (PML) protein is involved in the regulation of p53-dependent and -independent apoptosis. The PML protein localizes to multi-protein sub-nuclear structures termed as PML nuclear bodies (PML-NBs), which act as sensors of DNA damage and cellular stress. On genotoxic stress, the PML protein functions as a transcriptional co-activator of p53 by recruiting it to the PML-NBs, wherein PML facilitates p53 phosphorylation and acetylation through recruitment of factors such as HIPK2 and CBP (Bernardi and Pandolfi, 2007). Further, the PML gene itself is a transcriptional target of p53 (de Stanchina et al, 2004). Scaffold/matrix attachment regions (S/MARs) are regulatory DNA sequences mostly present upstream of various promoters. Matrix attachment region-binding proteins (MARBPs), which bind to such regulatory sequences, interact with numerous chromatin modifying factors and facilitate transcription in response to diverse stimuli (Zaidi et al, 2005). SMAR1 is an MARBP identified in mouse double positive thymocytes, wherein it binds to MARβ sequence at TCRβ locus and affects V(D)J recombination (Chattopadhyay et al, 2000; Kaul-Ghanekar et al, 2005). Subsequently, SMAR1 has been characterized as a tumour suppressor by virtue of its ability to interact with p53 and delay tumour growth in mouse melanoma model (Kaul et al, 2003; Jalota et al, 2005). The tumour suppressor p53 is also an MARBP that associates with the nuclear matrix and this association increases after DNA damage (Jiang et al, 2001). The PML-NBs are also associated with the nuclear matrix indirectly. However, the biological and functional significances of nuclear matrix-bound p53 and nuclear matrix-associated PML-NBs in context to p53 regulation are not known. In this study, we show that p53 target gene SMAR1 modulates the cellular response to genotoxic stress by a dual mechanism. First, SMAR1 interacts with p53 and facilitates p53 deacetylation through recruitment of HDAC1. Second, SMAR1 represses the transcription of BAX and PUMA by binding to an identical 25 bp MAR element in their promoters. We show that on mild DNA damage, SMAR1 generates an anti-apoptotic response by promoting p53 deacetylation and specifically repressing Bax and Puma expression. Reducing the expression of SMAR1 by shRNA leads to significant increase in p53-dependent apoptosis. However, severe DNA damage results in sequestration of SMAR1 into the PML-NBs. This facilitates p53 acetylation and transactivation of BAX and PUMA leading to apoptosis in cancer cells. Silencing of PML by specific siRNA abrogates DNA damage-induced apoptosis through transrepression of BAX and PUMA by SMAR1. Thus, sequestration of SMAR1 into the PML-NBs acts as a molecular switch to dictate p53-dependent cell arrest and apoptosis on DNA damage and we suggest that SMAR1 may be a new molecular target for cancer therapy. Results SMAR1 deacetylates p53 through association with HDAC1 Earlier, we have shown that SMAR1 interacts with p53 and stabilizes it in the nucleus (Jalota et al, 2005). We showed that SMAR1 inhibits Cyclin D1 expression by recruitment of HDAC1–mSin3A repressor complex on the promoter MAR (Rampalli et al, 2005). As SMAR1 and p53 are both matrix-associated transcription factors and both interact with HDAC1 independently, we investigated whether they are associated together in a complex. Double co-immunoprecipitation assay in HCT116 p53+/+ cells revealed the presence of detectable SMAR1–HDAC1–p53 complexes in vivo (Figure 1A). However, in the same eluate, SMAR1 did not show binding to another MAR protein PARP, thus revealing the specificity of the interaction. As p53 exists in a complex with HDAC1 along with SMAR1, it is possible that SMAR1 might deacetylate p53 endogenously by recruiting HDAC1 and thus keep its transactivation potential under check. To validate this, we investigated the levels of p53 acetylation on SMAR1 knockdown. SMAR1-specific shRNAs (sh745 and sh1077) were generated targeting two different regions of SMAR1 mRNA. The knockdown and specificity of the shRNAs were checked by western blotting (Supplementary Figure S1). Notably, knockdown of SMAR1 significantly induced p53 acetylation at Lys 373/382 (Figure 1B). However, total p53 level and p53 acetylation at Lys 320 were not affected. To understand the physiological relevance of this regulation in response to genotoxic stress, HCT116 p53+/+ cells were transduced with GFP expressing control and SMAR1 adenoviruses followed by UV (100 J/m2) treatment. Interestingly, SMAR1 overexpression strongly inhibited p53 acetylation at Lys 373/382, which was otherwise induced significantly on UV treatment (Figure 1C, panel 3, lane 2 versus lanes 3 and 4). Further, we performed an in vivo deacetylation assay. HCT116 p53−/− cells were transfected with p53 alone or in combination with p300 and SMAR1 in the presence or absence of Trichostatin A (TSA). Although p300 induced p53 acetylation at Lys-373/382 (Figure 1D, lane 3), presence of SMAR1 efficiently reduced p53 acetylation (lane 6), which was reversed on treatment with HDAC-inhibitor TSA (lane 5). Notably, although total p53 levels are high in SMAR1-transfected cells as SMAR1 is involved in p53 stabilization (Jalota et al, 2005), p53 acetylation levels are reduced in the presence of SMAR1 and the absence of TSA (lane 6). Finally, to confirm whether SMAR1 imparts deacetylase activity on p53 through recruitment of HDAC1, we carried out reversible co-immunoprecipitation to evaluate the association of p53 with HDAC1 in the presence and absence of SMAR1. Although knockdown of SMAR1 reduced the amount of p53 pulled with HDAC1, overexpression resulted in strong increase in their association (Figure 1E). This suggests that SMAR1-mediated p53 deacetylation is through recruitment of HDAC1. Figure 1.SMAR1 interacts with and regulates p53 acetylation endogenously. (A) Double co-immunoprecipitation assay to check the in vivo association of SMAR1–HDAC1 complex with p53. One milligram of cell lysate from HCT116 p53+/+ cells was immunoprecipitated sequentially with SMAR1 and HDAC1 antibodies. The eluted fraction was probed with p53 antibody. (B) HCT116 p53+/+ cells were transfected with SMAR1 shRNA (sh1077). Western blotting shows endogenous acetylation status of p53 at lys 373/382 and total p53 level 30 and 60 h after transfection. (C) HCT116 p53+/+ cells transduced with GFP expressing control and SMAR1 adenoviruses and treated with UV (100 J/m2) for 24 h. The levels of p53 and p53 acetylation status in comparison with SMAR1 expression are shown. (D) In vitro deacetylation assay of p53 by SMAR1. HCT116 p53−/− cells were transfected with p53, p300 expression plasmids in different combinations and treated with TSA (200 nM, 16 h) as given in the figure. GFP expression plasmid was transfected to monitor transfection efficiency. (E) Reversible co-immunoprecipitation assay in HCT116 p53+/+ cells showing differential association of p53 with HDAC1 in SMAR1 knockdown and overexpressed cells (left panel). Input controls are shown in right panel. Download figure Download PowerPoint BAX and PUMA are transcriptional targets of SMAR1 As SMAR1 modulates p53 acetylation through HDAC1, we evaluated the expression of p53 target genes. Overexpression of SMAR1 resulted in significant downregulation of Bax and Puma, whereas the levels of p53AIP1 and Apaf1, which are also p53 targets remained unchanged (Figure 2A). However, the levels of total p53 and p21 increased significantly corroborating our earlier data that overexpression of SMAR1 induces cell cycle arrest through p53 Ser 15 phosphorylation (Jalota et al, 2005). Thus, on one hand, SMAR1 enhances the expression of p21 and, on the other, it inhibits the expression of apoptotic genes BAX and PUMA. Interestingly, SMAR1 overexpression in HCT116 p53−/− cells also repressed Bax and Puma, but failed to induce p21 (Figure 2B). As SMAR1 interacts with HDAC1 and exists in a complex with p53, it is possible that SMAR1 deacetylates p53 through HDAC1. We, therefore, investigated whether knockdown of HDAC1 alters p53 acetylation status and alleviates the repression of Bax and Puma by SMAR1. Silencing of HDAC1 not only abrogated SMAR1-mediated repression of Bax and Puma, but also induced their basal expression. This suggests that SMAR1 negatively regulates the expression of Bax and Puma through HDAC1. In addition, HDAC1 knockdown induced p53 acetylation corroborating with increased Bax and Puma levels (Figure 2C) indicating that deacetylation of p53 by SMAR1 is dependent on HDAC1. Apaf1 expression, however, remained unaltered on HDAC1 knockdown. Reporter assays using full-length BAX and PUMA promoters in HCT116 p53−/− cells exhibited strong inverse correlation on SMAR1 ectopic expression and knockdown (Figure 2D and E), whereas p53AIP1 promoter did not show any significant change (Supplementary Figure S2). Again, p21 reporter was induced by SMAR1 in HCT116 p53+/+ cells, but failed to show read out in p53−/− cells (Figure 2F). Thus, induction of p21 by SMAR1 is p53 dependent, whereas transrepression of BAX and PUMA is p53 independent. Finally, to evaluate the biological significance of SMAR1-mediated repression of Bax and Puma, HCT116 p53+/+ cells were transduced with GFP expressing control adenovirus (Ad-V) and SMAR1 adenovirus (Ad-SM) followed by UV (100 J/m2) and Camptothecin (10 μM) treatment 48 h post-transduction. Surface staining using annexin-V-conjugated Cy3 revealed significant decrease in apoptotic population in SMAR1 overexpressed cells (Figure 2G). The statistical representation of percentage apoptosis is given in Supplementary Figure S2B. Propidium iodide staining in HCT116 p53+/+ cells (Figure 2H) and mouse embryonic fibroblast (Figure 2I) transduced with SMAR1 adenovirus and treated with UV and Camptothecin showed that SMAR1 can significantly protect these cells from genotoxic stress-induced apoptosis. Thus, the protective function of SMAR1 on genotoxic stress is most likely attributed to its function as a negative regulator of Bax and Puma. Figure 2.SMAR1 regulates the expression of key apoptotic molecules Bax and Puma and inhibits apoptosis. (A) HCT116 p53+/+ and (B) HCT116 p53−/− cells were transduced with GFP expressing control adenovirus (Ad-V) and SMAR1 Adenovirus (Ad-SM). Forty eight hours post-transduction, the levels of p53, p21, Bax, Puma, Apaf1 and p53AIP1 were determined. (C) HCT116 p53+/+ cells were transfected with HDAC1 siRNA (Santacruz) and levels of Bax, Puma and p53 acetylation were determined in the presence and absence of SMAR1 (C3). (D, E) Luciferase activity of full-length promoters of BAX and PUMA on SMAR1 overexpression and knockdown using two different shRNAs (sh745 and sh1077) in HCT116 p53−/− cells. A deletion mutant of SMAR1 lacking the DNA binding and the protein interacting domain (ΔSMAR1) was used as a control. Bars indicate s.d. from three independent experiments. (F) p21 luciferase assay on SMAR1 overexpression in HCT116 p53+/+ and p53−/− cells. (G) Annexin staining of HCT116 p53+/+ cells treated with Camptothecin (10 μM, 12 h) after transduction of GFP expressing control and SMAR1 adenoviruses. Cell cycle analysis in (H) HCT116 p53+/+ cells and (I) MEFs transduced with GFP expressing control (Ad-V) and SMAR1 (Ad-SM) adenoviruses. Forty eight hours post-transduction, cells were treated with UV (100 J/m2, 24 h) and Campthothecin (10 μM, 12 h) and thereafter stained with propidium iodide. Download figure Download PowerPoint SMAR1 induces an anti-apoptotic signal in response to mild DNA damage The tumour suppressor p53 induces cell cycle arrest after mild DNA damage and apoptosis after severe irreparable damage. Mild DNA damage results in transient acetylation of p53 and causes cell cycle arrest, whereas severe damage promotes prolonged accumulation and sustained p53 acetylation levels leading to apoptosis (Aylon and Oren, 2007). To further decipher the protective function of SMAR1, we investigated its responsiveness to mild DNA damage. Exposure of HCT116 p53+/+ cells to low-dose UV (5 J/m2) led to a steady increase in SMAR1 expression 24 h onwards (Figure 3A, panel 1). However, p53 acetylation levels increased from 4 h and peaked at 16 h, which corroborated with increased Bax and Puma levels at these time points. Conversely, p53 acetylation, Bax and Puma expression were reduced at 36–48 h, the time during which SMAR1 is induced (Figure 3A, panels 2, 3 and 4). This is in agreement with our earlier data that SMAR1 facilitates p53 deacetylation and represses Bax and Puma expression. Both p53AIP1 and PML, which are p53 targets were not induced at low UV dose suggesting that other post-translational modifications of p53 are required for their transactivation (Oda et al, 2000). Although Bax and Puma were inhibited by 48 h with concomitant deacetylation of p53, p21 levels increased and remain unaltered suggesting the induction of cell cycle arrest. Knockdown of SMAR1 resulted in robust increase in p53 acetylation, Bax and Puma expression with subsequent PARP cleavage (Figure 3B). Cell cycle analysis by propidium iodide staining showed that low-dose UV treatment did not induce significant cell death (∼4%), but caused distinct G1 arrest (Figure 3C, II and III). Notably, SMAR1 knockdown alone induced significant cell death (∼13%), which further increased (∼20%) on treatment with low-dose UV (Figure 3C, IV and V). Statistical representation of percentage sub-G1 population from three independent experiments is shown in Figure 3D. Similar results were observed in HEK 293 and MCF-7 cells (Supplementary Figure S3A and S3B). SMAR1 knockdown in MEFs also induced significant cell death through induction of Bax and Puma (Figure 3E; Supplementary Figure S3C). However, although shRNA-mediated knockdown of SMAR1 in HCT116 p53−/− cells results in a modest increase in Bax and Puma expression (Supplementary Figure S3D), no significant cell death was observed (Figure 3F). Thus, p53 is required for driving the cells towards apoptosis. Recently, it has been shown that acetylation is indispensable for p53-dependent transactivation and, therefore, expression of p53 targets Bax and Puma (Tang et al, 2008). Strikingly, in Figure 3A, we find that the levels of Bax and Puma increase well before increase in acetylated p53. It is possible that under this condition, activation of factors such as hCAS and ASPP family of proteins (Samuels-Lev et al, 2001; Tanaka et al, 2007), which can modulate p53 activity, may also contribute significantly to the expression of BAX and PUMA. Taken together, our results suggest that knockdown of SMAR1 affects the cellular apoptotic response by enhancing p53 acetylation at Lys 373/382 and increasing the expression of Bax and Puma. Figure 3.SMAR1 induces an anti-apoptotic signal in response to low-dose DNA damage. (A) HCT116 p53+/+ cells were UV irradiated with 5 J/m2 UV and cells collected at different time points as indicated. Immunoblot analysis was carried out as shown. (B) Expression levels of Bax, Puma, p53, acetylated p53 and PARP cleavage on shRNA (sh1077)-mediated knockdown of SMAR1 in HCT116 p53+/+ cells. (C) Propidium iodide staining in control HCT116 p53+/+ cells (48 h), cells treated with UV (5 J/m2, 48 h) and SMAR1 knockdown UV-treated cells (5 J/m2, 48 h) showing percentage apoptosis represented by sub-G1 population. (D) Statistical representation of percentage sub-G1 population on low-dose UV treatment and SMAR1 knockdown from three independent experiments. Error bars represent standard deviation. (E) Propidium iodide staining in MEFs transduced with control and SMAR1 lentivirus (F) Knockdown of SMAR1 by transient transfection in HCT116 p53+/+ and HCT116 p53−/− cells showing differential sub-G1 population. Download figure Download PowerPoint SMAR1 drives p53 apoptotic target gene specificity through MARs Earlier, we have shown that only BAX and PUMA and not other p53 targets were specifically repressed by SMAR1 overexpression. As SMAR1 is an MARBP and these proteins have propensity to bind AT-rich sequences, which often flank various promoters, we analysed the promoter sequences of p53 inducible genes. Computational analysis using MARWIZ software (Singh et al, 1997) predicted potential MARs in BAX and PUMA promoter within ∼700 bp upstream of p53 response element (p53RE), but not in p53AIP1 (Supplementary Figure S4A–S4C). Sequence alignment of the promoters up to ∼700 bp upstream of the p53RE revealed two stretches of sequences P1 (30 mer) and P2 (25 mer), which are identical in BAX and PUMA promoter (Figure 4A, blue and green boxes). Although the sequence encompassing the region P1 is located outside the MAR, the region P2 lies within the MAR region of both promoters (Supplementary Figure S4A and S4B). We then evaluated whether SMAR1 binds to these sequences. For this, two probes of 40 mer each were designed corresponding to the two segments P1 and P2. For binding reactions, bacterially expressed recombinant protein R5 (GST 350–548 aa) corresponding to the DNA-binding region of SMAR1 and recombinant protein R6 representing the protein interaction domain (GST 160–350 aa) were used. Interestingly, although P1 failed to form any complex, P2 formed nucleoprotein complex with R5 (Figure 4B, lane 4). The binding specificity was further showed by competition experiments showing loss of binding with the addition of 10-fold molar excess cold P2 (Figure 4C, lane 6). Under similar experimental conditions, GST and R6 failed to form any complex with P2 indicating the specific DNA-binding activity of SMAR1 (Figure 4C, lanes 2 and 3). To further check the specificity of SMAR1 binding to P2, we also designed two more similar sized probes P3 and P4 that lie proximal to P2 (Figure 4A, black boxes). Electrophoretic mobility shift assay (EMSA) studies with probes P3 and P4 did not show any complex formation (data not shown). Thus, SMAR1 binds to a specific and identical MAR such as sequence P2 present in both BAX and PUMA promoters. As the repressor activity of SMAR1 is attributed to its association with HDAC1, we further confirmed the binding of SMAR1 and recruitment of HDAC1 by chromatin immunoprecipitation (ChIP) assay. Cross-linked chromatin from HCT116 p53+/+ cells were pulled with SMAR1 and HDAC1 antibodies, respectively, and PCR amplified for BAX and PUMA promoters using primers spanning their respective MAR regions. Both BAX and PUMA promoters gave strong signal in PCR, whereas under similar experimental conditions, p21 and GAPDH promoters were not amplified (Figure 4D). Thus, SMAR1 recruits HDAC1 to the MAR regions of both BAX and PUMA, but not to p21 promoter. Notably, the MAR sites and the p53RE are juxtaposed on BAX and PUMA promoters. As SMAR1–HDAC1 complex interacts with p53, we reasoned that this complex might be associated with the nuclear matrix and more specifically to these MARs. To verify the co-occupancy of SMAR1/HDAC1 and SMAR1/p53 complexes on these MARs, we isolated the nuclear matrix from HCT116 p53+/+cells and performed sequential ChIP using SMAR1/HDAC1 and SMAR1/p53 antibodies. The purity of the isolated nuclear matrix was verified using Lamin B1 and histone H1 antibodies, respectively (Supplementary Figure S4D). Although BAX and PUMA showed detectable amplification, p53AIP1 and GAPDH promoters were not amplified (Figure 4E, lanes 3 and 4). Further, to study the occupancy and recruitment of SMAR1 on these MARs, cross-linked chromatin from UV-treated (5 J/m2) HCT116 p53+/+ cells at different time intervals was pulled with SMAR1, HDAC1 and acetylated p53 lys373/382 antibodies. Although BAX (Figure 4F, left panel) and PUMA (right panel) promoters were PCR detected in anti-SMAR1 pulled fraction, no amplification was observed in p53AIP1 (Supplementary Figure S4E). Furthermore, the kinetics of SMAR1 occupancy showed an oscillatory pattern with initial endogenous-bound SMAR1 slowly getting disappeared 4 h after UV treatment and reappeared by around 24 h. Strikingly, the occupancy of SMAR1 and HDAC1 inversely correlated with p53 acetylation status at these loci (Figure 4F, panels 1, 2 and 3). Interestingly, immunostaining showed strong translocation of SMAR1 into the nucleolus at 12 and 24 h after UV irradiation (Supplementary Figure S4F, white arrows). To verify this observation, we isolated nucleoplasmic and nucleolar fractions from HCT116 p53+/+ cells after irradiation with 5 J/m2 at different time points and probed for SMAR1 in these compartments. Our data shows that SMAR1 is present in the nucleolar fraction of un-irradiated cells albeit at very low levels, but expression increases significantly 8 h onwards after UV stress. Interestingly, although SMAR1 level increases in the nucleolar fraction, it decreases in the nucleoplasmic fraction (Figure 4G). This suggests that SMAR1 translocates into the nucleolus on UV stress. This can possibly explain the disappearance of SMAR1 from BAX and PUMA promoter MAR (Figure 4F) after 8 h of low-dose UV treatment leading to increased expression of Bax and Puma at 8 h (Figure 3A). However, 24 h after UV stress when SMAR1 is induced, it reappears in the nucleoplasmic fraction as well as in the nucleolus. This is again consistent with our ChIP data, wherein SMAR1 binding to the MAR element is restored after 24 h (Figure 4F), the time points after which Bax and Puma expression is maximally repressed as evident by immunoblot (Figure 3A). In the nucleolus, SMAR1 inhibits ribosomal gene transcription to cause cell cycle arrest (unpublished data). The nucleolar translocation of SMAR1 facilitates p53 acetylation, which then unleashes its transactivation potential on BAX and PUMA promoter resulting in pronounced expression of Bax and Puma. What post-translational modifications cause SMAR1 translocation into the nucleolus and binding to the MAR is currently under investigation. Together, these data suggest that nucleolar sequestration of SMAR1 may partly facilitate p53 acetylation and allow acetylated p53 to activate BAX and PUMA. Nonetheless, once SMAR1 is induced, it facilitates p53 deacetylation through HDAC1 and inhibits the transcription of BAX and PUMA. Thus, occupancy of SMAR1 on BAX and PUMA MAR selectively regulates their expression in response to DNA damage. Figure 4.SMAR1 regulates BAX and PUMA through MAR. (A) Schematic representation of BAX and PUMA promoter aligned with complementary strands showing the locations of two identical sequences P1 (blue) and P2 (pink in BAX and green in PUMA) in proximity to p53 response element (red). The sequence of P2 is boxed. (B) EMSA showing specific binding of SMAR1 (R5) to probe P2 (lane 4), but not to probe P1. (C) Purified GST (lane 2), R6 (lane 3) and R5 in increasing doses (lanes 4 and 5) were incubated with probe P2. Formation of complex (I) was visualized by autoradiography. Binding specificity of the complex in the presence of 10-fold molar excess of cold competitor (cold probe P2) is shown in lane 6. Free probe is denoted as FP. (D) C
Referência(s)