SUMOylation of MeCP2 is essential for ... - Wiley Online Library

JOURNAL OF NEUROCHEMISTRY

| 2014 | 128 | 798–806

doi: 10.1111/jnc.12523

*Institute of Neuroscience, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China †Department of Biochemistry and Molecular Cell Biology & Shanghai Key Laboratory of Tumor Microenvironment and Inflammation, Shanghai Jiao Tong University School of Medicine (SJTU-SM), Shanghai, China ‡Department of Neonatology, Children’s Hospital of Fudan University, Shanghai, China

Abstract Methyl CpG binding protein 2 (MeCP2) binds to methylated DNA and acts as a transcriptional repressor. Mutations of human MECP2 gene lead to Rett syndrome, a severe neural developmental disorder. Here, we report that the MeCP2 protein can be modified by covalent linkage to small ubiquitinlike modifier (SUMO) and SUMOylation at lysine 223 is necessary for its transcriptional repression function. SUMOylation of MeCP2 is required for the recruitment of histone deacetylase complexes 1/2 complex. Mutation of MeCP2 lysine 223 to arginine abolishes its suppression of gene

expression in mouse primary cortical neurons. Significantly, mutation of MeCP2 K223 site leads to developmental deficiency of rat hippocampal synapses in vitro and in vivo. Thus, the SUMOylation of MeCP2 at K223 is a critical switch for transcriptional repression and plays a crucial function in regulating synaptic development in the central nervous system. Keywords: gene regulation, histone deacetylase, posttranslational modification, SUMOylation, transcription repressor. J. Neurochem. (2014) 128, 798–806.

Methyl CpG binding protein 2 (MeCP2) primarily binds to methylated CpG islands and plays a critical role in mediating transcriptional repression by recruiting histone deacetylase complexes (HDACs) (Nan et al. 1998). Mutations of mecp2 gene in human lead to a severe neurodevelopmental disorder known as Rett syndrome (Amir et al. 1999). The transcriptional repression activity of MeCP2 protein is regulated by post-translational modifications, such as phosphorylation and acetylation (Chen et al. 2003; Zhou et al. 2006; Tao et al. 2009; Zocchi and Sassone-Corsi 2012). Recent study has shown that covalent linkage with small ubiquitin-like modifier (SUMO) protein, a processing termed SUMOylation, is also an important form of post-translational modification function in regulating activity of many transcription factors. In most of the cases, SUMOylation correlates with transcription repression (Gill 2005). Thus, we set out to examine whether MeCP2 could also be SUMOylated and whether SUMOylation of MeCP2 could contribute to neural development.

Materials and methods

798

Experimental animals For biochemistry experiments, primary cortical neurons from either sex of embryonic 15 days of C57BL/6 mouse were used. Primary hippocampal neurons from either sex of post-natal 1 day of Sprague–Dawley rat were used for immunocytochemistry experiments. Either sex of SD rat new born pups after in utero electroporation was used for immunohistochemistry experiments. The use and care of animals complied with the guideline of the Received June 28, 2013; revised manuscript received October 27, 2013; accepted October 31, 2013. Address correspondence and reprint requests to Zilong Qiu, Institute of Neuroscience, Shanghai Institute of Biological Sciences, Chinese Academy of Sciences, 320 Yueyang Road, Shanghai 200031, China. E-mail: [email protected]; Wen-Hao Zhou, Department of Neonatology, Children’s Hospital of Fudan University, 399 Wanyuan Road, Shanghai 201102, China. E-mail: [email protected] Abbreviations used: HDAC, histone deacetylase complex; MeCP2, methyl CpG binding protein 2; SUMO, small ubiquitin-like modifier.

© 2013 International Society for Neurochemistry, J. Neurochem. (2014) 128, 798--806

SUMOylation of MeCP2 is required for transcriptional repression

Animal Advisory Committee at the Shanghai Institutes for Biological Science, Chinese Academy of Sciences. The ARRIVE guidelines were also complied with. Expression and purification of SUMOylated protein HA-MeCP2, myc-Ubc9, and His-SUMO1 were co-transfected into 293T Cells. Cells were lysated in 6M Guanidine Hydrocholoride (GuHCl) lysis buffer. After sonication, covalently modified His-SUMOHA-MeCP2 was purified by Ni-NTA His Bind Resin (70666; Merck Millipore, Bedford, MA, USA) under denature condition according to the protocol. After rotating 2–3 h in 25°C, beads were washed extensively with 8 M Urea with pH 8.0 and pH 6.3 for four times each. Beads were eluted with elution buffer containing 250 mM imidazole. Short hairpin RNA and replacement plasmids The MeCP2 RNAi sequence was AAGTCAGAAGACCAGGATC. The pSuper-MeCP2 RNAi construct was made by inserting the short hairpin RNA sequence into the pSuper.Retro.neo.Green fluorescent protein (GFP) vector according to the protocol. The MeCP2 expressing construct was made by cloning the MeCP2 cDNA into FUGW by BamHI and EcoRI site. For the replacement construct, the H1 promoter driving MeCP2 short hairpin cassette was inserted into the PacI site of the expressing construct. RNA extraction, reverse transcription, and Quantitative-PCR Total RNA from in vitro cultured neuron was extracted by TRIzol Reagent (155926-026; Invitrogen, Carlsbad, CA, USA). Reverse transcription was carried out with Reverse Transcriptase M-MLV (RNase H free) (D2639A; Takara Bio, Shiga, Japan). QuantitativePCR was carried out on the Rotor-Gene Q (Qiagen, Valencia, CA, USA). Primer sequences are as follows: Slco1b2-For, ctgtaaccagctgtggagca; Slco1b2-Rev, caggttctgggtttccttca; Hrasls5-For, atcttcagcaaccgagctgt; Hrasls5-Rev, ctcctgggcacaccatatct; Npc1 l1-For, aattgctggcattcttgtcc; Npc1 l1-Rev, cacgttcccaagtgtccttt; Il25-For, accacaaccagacggtcttc; Il25-Rev, ctgcttcaggtagggctttg; Phlda3-For, ctgttcttcatggcgtgcta; Phlda3-Rev, aggggtgagggaagaagtgt; Sgca-For, cgtcacctgaatggaaacct; Sgca-Rev, tgttccaaaggatgcacaaa; Tnnt2-For, ctgagacagaggaggccaac; Tnnt2-Rev, ttctcgaagtgagcctcgat; Tcea3-For, agcagctctgaaggcagaag; Tcea3-Rev, ctcatcactggccatttcct; Rgs5-For, tcatttcaatcctgcccttc; Rgs5-Rev, tgacaggaggcatctgagtg; Tcea3-For, agcagctctgaaggcagaag; Tcea3-Rev, ctcatcactggccatttcct, Dbc1-For, caggctgtctgcaccagtaa; Dbc1-Rev, atggatttgtgcaggagacc; Kcnh3-For, catggaatgtacccccagac; Kcnh3-Rev, cacctgttcagacagctcca; glyceraldehyde-3-phosphate dehydrogenase (GAPDH)-For, atcccagagctgaacgggaagc; GAPDH-Rev, ttgggggtaggaacacggaagg. Crh-For: aggc tctggtgtggagaaact; Crh-Rev: gttaggggcgctctcttctcc; neurexophilin4For: cacgccgcagtattaatggc; neurexophilin4-Rev: gagaggattggat gggctgg; Neurotensin-For: acatccaagatcagcaaagcaa; NeurotensinRev: catgtctcctgcttcctcgg Cell culture and transfection 293T cell was cultured in Dulbecco’s modified Eagle’s medium (11995-073; Invitrogen) containing 10% fetal bovine serum. Transfection was carried out with Fugene HD according to the protocol. Cultures of primary cortical neurons were prepared from E15.5 mouse and maintained in Neurobasal medium (21103-049; Invitrogen) supplemented with Glutmax (35050-061; Invitrogen)

799

and B27 (12587-010; Invitrogen). Neurons were transfected with lentivirus packaged in Shanghai Sunbio Biomedical Technology. Cultures of primary cortical neurons were prepared from P1 SD rats and maintained in Neurobasal-A medium (10888-022). Neurons were transfected by Calcium phosphate/DNA co-precipitation at 57DIV. Immunoprecipitation 293T cells expressing Myc-MeCP2 alone or together with HA-HDAC were lysed in radio immunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.1%Triton X-100) 48 h after transfection. Anti-HA/Myc-Tag mouse monoclonal antibody conjugated to agarose (M20012L/ 20013L; Abmart, Shanghai, China) was used for precipitation. For endogenous SUMO1 immunoprecipition, neurons were lysed in the RIPA containing 20 mM N-Ethylmaleimide (E3876; Sigma, St Louis, MO, USA). Protein A/G plus agarose (sc-2003; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and rabbit anti-SUMO-1 antibody (sc-9060; Santa Cruz Biotechnology) were used for precipitation and normal rabbit IgG (sc-2027; Santa Cruz Biotechnology) was used as a negative control. Western blot Proteins were directly lysed with 1 9 sodium dodecyl sulfate loading buffer (62.5 mM Tris-HCl, pH = 6.8, 1% w/v sodium dodecyl sulfate, 10% glycerol, 50 mM dithiothreitol, 0.01% w/v bromophenol blue) and boiled for 10 min. Immunoblot was performed using appropriate dilution of primary antibodies (antiHA, 1 : 2000; Abmart. anti-Myc, 1 : 2000; Abmart. anti-MeCP2, #3456, 1 : 1000; Cell Signaling Technology, Beverly, MA, USA. anti-HDAC1, ab7028, 1 : 2000; Abcam, Cambridge, UK. antiHDAC2, ab7029, 1 : 2000; Abcam) and secondary antibodies (horseradish peroxidase–conjugated antibody to mouse, 1 : 5000, GE Healthcare Life Sciences, Pittsburgh, PA, USA). The immunoreactive bands were detected by Pierce enhanced chemiluminescence western blotting substrate (34080; Pierce, Rockford, IL, USA). Densitometry of the western blot protein bands was analyzed using ImageJ, NIH, rsbweb.nih.gov/ij/. Microarray analysis Mouse cortical neurons were transfected with GFP and MeCP2 RNAi lentivirus, respectively, for 5 days, before RNA collection. Total RNAs were collected, each sample with triple biological duplicates, and sent to be analyzed by Agilent mouse whole-genome 4X44K chip (design ID: 014868). Microarray experiment and data analysis were performed in Shanghai Biotechnology Corporation (www.shbiochip.com). Statistic analysis is performed in Shanghai Novel Biotech Corporation (www.novelbio.com). Briefly, the Bayes-based LIMMA (Linear Models for Microarray Data) model was used for initial screening. Those genes with fold change (FC) either log2 FC ≥ 0.58 or log2 FC ≤ 0.585, and further with pvalue ≤ 0.05, FDR ≤ 0.05, were left for further analysis.

In utero electroporation Electroporation was carried out as described previously (NavarroQuiroga et al. 2007) with some modifications. E16.5 SD rats were used. Electroporation was performed by Electro Squireportator T830 (BTX Instrument Division, Harvard Apparatus Inc.,


800

J. Cheng et al.

Holliston, MA, USA) at five repeats of 60 V for 50 ms, with an interval of 100 ms. Statistical analysis The results were expressed as Mean SEM. Statistical significance (p < 0.05) was assessed using Student’s t-test and one way ANOVA, as indicated in Figure legends. Immunocytochemistry and Immunohistochemistry For immunocytochemistry, rat hippocampal neurons 11-13DIV was fixed with 4% (wt/vol) paraformaldehyde in 0.1 M phosphate buffer. After blocked in 10% (vol/vol) goat serum, the cells were incubated in the antibodies (anti-GFP, A11122, 1 : 2000; Invitrogen. anti-vesicular glutamate transporter 1, AB5905, 1 : 5000; Millipore Corporation, Bedford, MA, USA) at 4°C overnight. The cells were incubated in Alexa Fluor 488 goat anti-rabbit and Alexa Fluor 555 goat anti-guinea pig secondary antibodies (Molecular Probes, Eugene, OR, USA) diluted at 1 : 2000 at 25°C on the second day. Experimental animals were anesthetized and then perfused with 4% (wt/vol) paraformaldehyde in 0.1 M phosphate buffer. Whole brains were removed from the skulls, post-fixed, cryopreserved in 30% (wt/vol) sucrose, embedded in OCT and sectioned coronally (50 lm). Sections were first permeabilized with 0.1% Triton X-100 in 0.1 M phosphate-buffered saline and then blocked and incubated in primary antibodies overnight at 4°C. The sections were incubated in Alexa Fluor 488 goat anti-rabbit secondary antibody on the second day. Stained sections were visualized using confocal microscopy (A1R/A1 confocal laser scanning microscope, Nikon Micro-Imaging, Tokyo, Japan). The spine density was measured using Image J. ANOVA was performed to assess statistical differences between control and experimental conditions.

Results MeCP2 could be SUMOylated in 293T cells and in mouse primary cortical neurons To determine whether MeCP2 could be SUMOylated, we first co-expressed HA-tagged MeCP2 (HA-MeCP2) along with SUMO1 and UBC9, an E2-ligase in the SUMOylation pathway, in 293T cells. We found that the HA-MeCP2 protein showed significant shifting bands (around 95 kd and 130 kd) on sodium dodecyl sulfate–polyacrylamide gel electrophoresis when SUMO1 and UBC9 were present, suggesting that MeCP2 can be SUMOylated (Fig. 1a). Next, we used a denature method to further confirm the covalent SUMOylation modification in MeCP2. We co-transfected with HA-MeCP2 with or without UBC9 and tandem Histidines tagged SUMO1 into 293T cells. After obtaining the cell lysate with denature reagent, we purified HisSUMO1-HA-MeCP2 protein with Ni-NTA beads. After purification, we examined the His-SUMO1-HA-MeCP2 fusion protein with western blot using anti-HA antibody. Consistently, we found that HA-MeCP2 could be covalently modified by His-SUMO1 and showed modified bands after co-transfected with UBC9 and SUMO1 (Fig. 1b). Note that

the un-modified MeCP2 protein itself also contains tandem Histidines residues and therefore could also be purified by Ni-NTA beads and therefore shows the around 75 kd band in the absence of UBC9 and His-SUMO1. To further investigate whether MeCP2 could be SUMOylated in neurons, we performed immunoprecipitation experiments against the endogenous proteins in mouse cortical neurons in primary cultures. We found that MeCP2 could be precipitated by anti-SUMO1 antibody, confirming that the endogenous neuronal MeCP2 can be SUMOylated (Fig. 1c). We further confirmed the SUMOylation modification in MeCP2 by performing immunoprecipitation using antiMeCP2 antibody in 293T cells. We found that SUMOylated MeCP2 bands are present after immunoblot with antiSUMO1 antibody (Fig. 1d). Next, we wanted to know whether endogenous MeCP2 could be SUMOylated in mouse cortical neurons. We performed immunoprecipitation experiments with antiSUMO1 antibody in neurons upon depolarization stimulus and found that SUMOylation of MeCP2 may show slightly increased after depolarization stimulus for 30 min (Fig. 1e). Interestingly, the SUMOylated MeCP2 in mouse cortical neurons (around 110 kd) showed a different size of band shifting comparing to previous data in 293 cells (95 kd and 130 kd), suggesting that there are different sites of SUMOylation in mouse and human cells. SUMOylation of MeCP2 may occur at multiple lysine sites, which exhibit the conserved tetrapeptide motif Ψ-K-XE/, whereas Ψ stands for hydrophobic amino acids. We found that there is one potential lysine site, K223 of MeCP2, was notably conserved across vertebrates (Fig. 1f). Therefore, we set out to examine whether K223 of MeCP2 could be SUMOylated. SUMOylation of MeCP2 on K223 is crucial for recruiting HDAC complex We generated MeCP2 with point mutation by replacing lysine to arginine at K223 site (MeCP2K223R). We found that K223R mutation of MeCP2 resulted in a significant reduction in the level of SUMOylation (Fig. 2a), suggesting that K223 is the primary SUMOylated residue in MeCP2 protein. Given the importance of SUMO in recruiting HDAC complex, we propose that mutation at lysine K223 may disrupt the interaction between MeCP2 and HDAC complex. This hypothesis was tested by further immunoprecipitation experiments. We co-expressed HA-HDAC1 with myc tagged wild type and K223R MeCP2 in 293T cells, respectively. We found that immunoprecipitation with HA antibodies showed strong interaction between myc-MeCP2 wild type and HA-HDAC1, and this interaction was greatly reduced when Myc-MeCP2K223R was co-expressed with HA-HDAC1 (Fig. 2b, c). Furthermore, we transfected 293T cells with either myc-MeCP2 WT or MeCP2K223R, and examine the interaction of these proteins with endogenous HDAC1 and



(a)

801

(d)

(b) (e)

(f) (c)

Fig. 1 Methyl CpG binding protein 2 (MeCP2) is a substrate of SUMOylation. (a) Western blot showed that when expressed together with UBC9 and small ubiquitin-like modifier (SUMO)1, covalent linkage to SUMO1 enlarged the molecular weight of MeCP2, resulting in the band shifts on the sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). Protein lysates from 293T cells were collected 48 h after transfection and immunobloted with antibodies indicated. (b) Covalent linkage of SUMO1 to MeCP2. Protein lysates from 293T cells were collected 48 h after transfection, denature and purified by Ni-NTA beads, and immunobloted with antibodies indicated. (c) Immunoprecipitation of endogenously expressed MeCP2 with SUMO1 antibody in

cultured cortical neuron. Neuronal lysates from mouse primary cortical cultures were collected 5 DIV and immunoprecipitated with antibodies indicated and immunobloted by anti-MeCP2 antibody. (d) Immunoprecipitation of endogenous SUMOylated MeCP2 in 293T cells. Cell lysates were collected 48 h after transfection and immunoprecipitated with antibodies indicated and immunobloted by anti-SUMO1 antibody. (e) Activity-dependent SUMOylation of MeCP2. Neuronal protein samples were immunoprecipited with anti-SUMO1 antibody with or without KCl (50 mM) stimulus and immunobloted with anti-MeCP2 antibody. (f) The K223 site of MeCP2 is conserved across majority species of vertebrates.

HDAC2, using immunoprecipitation with myc antibodies. We found that K223R mutation fully abolished the interaction between MeCP2 and HDAC1/2 (Fig. 2d, e, f). These evidences indicate that SUMOylation of MeCP2 K223 residue is critical for its interaction with the HDAC1/2 complex. Next, we used a human ubiquitin c promoter driving GFP reporter system to examine whether transcriptional repression activity of MeCP2 requires SUMOylation of MeCP2 at

K223 (Fig. 2g). We found that over-expression of wild type MeCP2 significantly represses reporter gene expression, but K223R mutation of MeCP2 lost its repressive activity, further confirming that SUMOylation of K223 of MeCP2 is essential for its transcriptional repression activity (Fig. 2h). It is reported that MeCP2 could also function as a transcriptional activator and numerous genes were found to be up-regulated in MeCP2 transgenic mouse (Chahrour et al.


802

J. Cheng et al.

(a)

(d)

(b)

(e)

(f)

(c)

(g)

(h)

(i)

(j)

Fig. 2 SUMOylation of Methyl CpG binding protein 2 (MeCP2) at K223 is necessary for its interaction with histone deacetylase complex (HDAC) and the transcriptional repression function. (a) The SUMOylation level of K223R was reduced when compared to wild type MeCP2 in 293T cells. HA-tagged wild type and K223R mutation of MeCP2 were co-transfected with small ubiquitin-like modifier (SUMO)1, respectively. Anti-HA western blots were used to examine SUMOylated MeCP2. (b) Mutation from lysine to arginine at K223 disrupts the interaction between MeCP2 and HDAC complex. HA-HDAC1 was cotransfected with wild type and K223R myc-MeCP2 into 293T cells. Immunoprecipitation using anti-HA antibody was performed, followed by western blot using anti-myc antibody. (c) Quantization of the MycMeCP2 protein level precipitated by HA-HDAC1 in (b). Student’s t-test, *p = 0.022384. (d) K223R disrupts interaction between myc-MeCP2 and endogenous HDAC1,2. Wild type and K223R mutation of myc

tagged MeCP2 were transfected into 293T cells. Immunoprecipitations using anti-myc antibody were performed, followed by western blot using anti-HDAC1,2 antibodies as indicated. (e, f) Quantization of the HDAC1/2 protein level precipitated by Myc-MeCP2. Student’s t-test, *p = 0.000228219 (e). *p = 0.002103538 (f). (g) Schematic illustration of Ubi-GFP reporter. (h) SUMOylation of MeCP2 K223 site is critical for repressing reporter gene expression. 293T cells were transfected with Ubi-GFP reporter along with control, MeCP2-WT and MeCP2K223R, respectively. GFP protein levels were examined by western blot 48 h after transfection. (i) Schematic illustration of lentiviral-based MeCP2 wild type and K223R rescue constructs. (j) Quantitative-PCR showed the mRNA level of the genes in primary cultured neuron infected with lentivirus harboring GFP along as control, and RNAi along with RNAi-resistant wild type MeCP2 (wt-rescue) or K223R (K223R-rescue). One way ANOVA *p < 0.05.



2008). We would like to address whether SUMOylation of MeCP2 at K223 is required for its transcriptional activation activity. To avoid the potential inference of endogenous MeCP2, we carried out a replacement experiment by simultaneously down-regulating endogenous MeCP2 with RNAi and expressing either wild-type or mutated MeCP2 in cultured mouse cortical neurons. The replacement construct contains two cassettes, a short hairpin RNA expressing cassette against MeCP2 and an RNAi-resistant wild-type or mutant MeCP2 cDNA expression element (Fig. 2i). We then asked whether the expression of three candidate genes, Corticotropin-releasing hormone (Crh), Neurexophilin4, and Neurotensin, which are found to be up-regulated in previous work, could be regulated by wild-type and K223R mutant of MeCP2. We examined mRNA of target genes by quantitative PCR in three groups of neurons – the control construct (containing only GFP), and replacement constructs containing RNAi-resistant wild-type or mutant MeCP2 cDNA. Surprisingly, we found that only wild type, but not K223R mutant of MeCP2 could up-regulate expression levels of target genes (Fig. 2j). These evidences strongly suggest that

(a)

SUMOylation on K223 MeCP2 plays a critical role in regulating gene expression. SUMOylation of MeCP2 on K223 is required for transcriptional repression in neurons To further address the transcription repression function of MeCP2 in neurons, we carried out whole-genome cDNA microarray to screen for neuronal target genes that are repressed by MeCP2. We used lentivirus harboring MeCP2 RNAi to down-regulate MeCP2 expression in cultured mouse cortical neurons (Fig. 3a, b) and collected RNAs for microarray analysis. We found that the expression level of numerous genes were up-regulated when MeCP2 was knocked down, suggesting that they are potential target genes for transcriptional repression by MeCP2 (Fig. 3c, Table S1). Twelve candidate genes were chosen for further validation by quantitative PCR (Fig. 3d). We would like to address whether SUMOylation of MeCP2 at K223 is required for its transcriptional repression activity. We used the replacement strategy described previously (Fig. 2i). In this experiment, we used quantitative PCR to

(c)

(b)

Fig. 3 SUMOylation of Methyl CpG binding protein 2 (MeCP2) K223R is critical for regulating gene expression in neurons. (a) Lentivirus harboring short hairpin RNA sequence targeting MeCP2 down-regulates the protein level of MeCP2. Western blot showed that the protein level of MeCP2 was down-regulated in cultured neurons infected by MeCP2 RNAi lentivirus compared with control GFP expressing virus. (b) Quantization of the protein level of MeCP2 in cultured neuron infected by lentivirus expressing GFP or MeCP2 RNAi. Student’s t-test, *p < 0.05. (c) Heat map for microarray data with MeCP2 RNAi. (d) Validation of the microarray result by quantitative PCR. mRNA level quantized by realtime PCR using cDNA of cultured neuron infected by lentivirus expressing GFP or MeCP2 RNAi. Consistent with the microarray result, the mRNA levels of these twelve genes were up-regulated in MeCP2 knockdown neurons. Student’s t-test, *p < 0.05. (e) Quantitative-PCR showed the mRNA level of the genes in primary cultured neuron infected with lentivirus harboring GFP along as control, RNAi targeting MeCP2, and RNAi along with RNAi-resistant wild type MeCP2 (wt-rescue) or K223R (K223R-rescue). One way ANOVA p < 0.05.

803

(d)

(e)


804

J. Cheng et al.

(a)

(b)

(c)

(d) (e)

Fig. 4 SUMOylation of Methyl CpG binding protein 2 (MeCP2) at K223 is essential for synapse development in vitro and in vivo. (a) Representative images of cultured hippocampal neurons transfected with control vector (GFP), RNAi targeting MeCP2, RNAi together with wild type MeCP2 or K223R. Scale bar = 2lm. (b) Quantification of the excitatory synapse density showed in (a), Data represent mean SEM, n = 11– 24, one way ANOVA: p = 4.29E-05. (c) Representative rat brain section

image showing that GFP encoding plasmids were delivered into developing hippocampal neurons. Scale bar = 1 mm. (d) Representative images of P14 hippocampal CA1 neuron transfected with control vector (GFP), RNAi targeting MeCP2, RNAi together with wild type MeCP2 or K223R. Bar = 2 lm. (e) Quantification of the spine density showed in (d). Data represent mean SEM, n = 18–25 for each condition, one way ANOVA: p = 0.000627.

examine the expression of potential candidate genes in four groups of neurons – those expressing control construct (containing only GFP), MeCP2 RNAi construct, and replacement constructs (containing either RNAi-resistant wild-type or

K223R mutant MeCP2 cDNA). We found that the expression of MeCP2 RNAi resulted in up-regulated level of expression of five genes that are known to be important for neuronal functions, as compared with that found in neurons expressing



control construct. Expression of the replacement construct encoding RNAi-resistant wild-type MeCP2 fully restored the expression of these genes to the control level, whereas the replacement with MeCP2K223R had no effects on RNAiinduced up-regulatoin of these genes (Fig. 3e). These data indicate that K223R mutation results in a loss-of-function of MeCP2, further confirming the critical importance of MeCP2 SUMOylation for its transcriptional repression activity.

805

The role of MeCP2 SUMOylation in synapse development Loss-of-function of MeCP2 leads to developmental defects of both excitatory (Chao et al.2007) and inhibitory (Chao et al. 2010) synapses. As in the rescue experiment MeCP2K223R could not prevent the up-regulation of gene expression caused by MeCP2 RNAi, we asked whether it may affect synaptic defects caused by MeCP2 deletion. We found that in cultured rat hippocampal neurons, knockdown of MeCP2 with short hairpin RNA transfection significantly reduced the density of excitatory synapses, as measured by the density of immunostained puncta of vesicular glutamate transporter 1. Co-expression of wild-type MeCP2 with short hairpin RNA fully restored the normal synaptic density, whereas cotransfection with MeCP2K223R had no effect (Fig. 4a, b). Finally, using in utero electroporation assay, we manipulate the expression of MeCP2 in vivo by introducing MeCP2 RNAi construct together with rescue constructs encoding either wild-type MeCP2 or MeCP2K223R into in developing rat hippocampus (Fig. 4c). Consistent with the effect on synapse density in cell cultures, knockdown of MeCP2 in vivo notably reduced the development of excitatory synapses, represented by density of spines on the dendrites and this spine loss could be fully rescued by coexpression with wild-type MeCP2 but not MeCP2K223R (Fig. 4d, e). These data strongly support the notion that SUMOylation of K223 site of MeCP2 plays a critical role in regulating excitatory synapse formation in vitro and in vivo.

provides new insight into the molecular mechanism by which MeCP2 exerts its suppressive function to gene expression. SUMOylation modifications are often associated with changes in subcellular localization of modified protein. We carried out immunocytochemistry for wild type and K223R mutant form of MeCP2 in 293T cell and in cortical neurons and found both forms of MeCP2 remain strictly in the nucleus, suggesting SUMOylation of K223 doesn’t affect the subcellular localization of MeCP2 (Data not shown). Besides regulating transcription repression, SUMO is thought to be a modifier of chromatin structure and function (Cube~ nas-Potts and Matunis 2013). More and more recent studies showed that MeCP2 is involved in normal chromatin structure maintenance (Skene et al. 2010; Baker et al. 2013). Our SUMO assay data indicate that MeCP2 has multiple SUMOylated lysines and SUMOylation at other sites may contribute to alteration of chromatin state. MeCP2’s binding to 5mC and 5hmC lead to transcription repression and activation separately (Mellen et al. 2012) and the balance may also be regulated by MeCP2 SUMOylation. Previous reports showed that MeCP2 is critical for synaptogenesis (Chao et al. 2007; Cohen et al. 2011; Li et al. 2011). Dysregulation of synaptogenesis may contribute to the pathological basis of Autism Spectrum Disorder. Our finding that SUMOylation of MeCP2 is necessary for synapse formation suggests potential developmental regulation of MeCP2 SUMOylation, and defective regulation of this SUMOylation may contribute to the pathogenesis of Rett Syndrome. Taking together with previous evidence that SUMOylation of MEF2, another important activity-dependent transcriptional factor, also play critical roles in synapse formation, the finding of MeCP2 as a new target of SUMOylation further highlights the importance of protein SUMOylation in synaptic development (Shalizi et al. 2006; Craig and Henley 2012).

Discussion

Acknowledgements

As a nuclear protein, MeCP2 is highly expressed in postmitotic neurons and in mature brain. Both phosphorylation (Zhou et al. 2006; Tao et al. 2009) and acetylation (Zocchi and Sassone-Corsi 2012) are critical for the transcriptional repression function of MeCP2 by regulating its chromatin association. Here, we identified a novel form of posttranslational modification in MeCP2 that regulates its interaction with transcriptional repression complex HDAC1/2. SUMOylation of MeCP2 contributed to its transcriptional repression and failure of SUMOylation at K223 disrupted the normal gene expression, leading to abnormal synaptic development. SUMO is a critical regulator for transcriptional repression (Yang and Sharrocks 2004). Our finding that SUMOylation of MeCP2 at K223 is necessary for its interaction with HDAC complex

We thank Dr Mu-ming Poo for critical reading of this manuscript. This work was supported by the 973 Program of China (2011CBA00400), CAS Hundreds of Talents Program, Strategic Priority Research Program of Chinese Academy of Sciences, Grant No.XDB02050400 to ZQ, and SKLN-2010A08 from State Key Laboratory of Neuroscience, Chinese Academy of Sciences to JY. National Natural Science Foundation of China (91232712) to WHZ. Authors declare that they do not have conflict of interests for this work.

Supporting information Additional supporting information may be found in the online version of this article at the publisher's web-site: Table S1. The list of differential expressing genes identified from MeCP2 RNAi microarray.


806

J. Cheng et al.

References Amir R. E., Van den Veyver I. B., Wan M., Tran C. Q., Francke U. and Zoghbi H. Y. (1999) Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nat. Genet. 23, 185–188. Baker S. A., Chen L., Wilkins A. D., Yu P., Lichtarge O. and Zoghbi H. Y. (2013). An AT-Hook domain in MeCP2 determines the clinical course of Rett syndrome and related disorders. Cell 152, 984–996. doi: http://dx.doi.org/10.1016/j.cell.2013.01.038 Chahrour M., Jung S. Y., Shaw C., Zhou X., Wong S. T. C., Qin J. and Zoghbi H. Y. (2008) MeCP2, a key contributor to neurological disease, activates and represses transcription. Science 320, 1224–1229. doi:10.1126/science.1153252. Chao H.-T., Zoghbi H. Y. and Rosenmund C. (2007). MeCP2 controls excitatory synaptic strength by regulating glutamatergic synapse number. Neuron, 56, 58–65. doi: http://dx.doi.org/10.1016/j. neuron.2007.08.018 Chao H.-T., Chen H., Samaco R. C., Xue M., Chahrour M., Yoo J. and … Zoghbi, H. Y. (2010) Dysfunction in GABA signalling mediates autism-like stereotypies and Rett syndrome phenotypes. Nature 468, 263–269. Chen W. G., Chang Q., Lin Y., Meissner A., West A. E., Griffith E. C. and … Greenberg, M. E. (2003) Derepression of BDNF transcription involves calcium-dependent phosphorylation of MeCP2. Science 302, 885–889. doi:10.1126/science.1086446. Cohen S., Gabel H. W., Hemberg M et al. (2011). Genome-wide activity-dependent MeCP2 phosphorylation regulates nervous system development and function. Neuron 72, 72–85. doi:http:// dx.doi.org/10.1016/j.neuron.2011.08.022 Craig T. J. and Henley J. M. (2012). Protein SUMOylation in spine structure and function. Curr. Opin. Neurobiol. 22, 480–487. doi: http://dx.doi.org/10.1016/j.conb.2011.10.017 Cube~ nas-Potts C. and Matunis M. J. (2013). SUMO: a multifaceted modifier of chromatin structure and function. Dev. Cell 24, 1–12. doi: http://dx.doi.org/10.1016/j.devcel.2012.11.020 Gill G. (2005). Something about SUMO inhibits transcription. Curr. Opin. Genet. Dev. 15, 536–541. doi: http://dx.doi.org/10.1016/j. gde.2005.07.004

Li H., Zhong X., Chau K. F., Williams E. C. and Chang Q. (2011) Loss of activity-induced phosphorylation of MeCP2 enhances synaptogenesis, LTP and spatial memory. Nat. Neurosci. 14, 1001–1008. Mellen M., Ayata P., Dewell S., Kriaucionis S. and Heintz N. (2012). MeCP2 binds to 5hmC enriched within active genes and accessible chromatin in the nervous system. Cell 151, 1417–1430. doi: http:// dx.doi.org/10.1016/j.cell.2012.11.022 Nan X., Ng H.-H., Johnson C. A., Laherty C. D., Turner B. M., Eisenman R. N. and Bird A. (1998) Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393, 386–389. Navarro-Quiroga I., Chittajallu R., Gallo V. and Haydar T. F. (2007) Long-term, selective gene expression in developing and adult hippocampal pyramidal neurons using focal in utero electroporation. J. Neurosci. 27, 5007–5011. doi:10.1523/ JNEUROSCI.0867-07.2007. Shalizi A., Gaudilliere B., Yuan Z., Stegm€uller J., Shirogane T., Ge Q. and … Bonni, A. (2006) A calcium-regulated MEF2 sumoylation switch controls postsynaptic differentiation. Science 311, 1012–1017. doi:10.1126/science.1122513. Skene P. J., Illingworth R. S., Webb S., Kerr A. R. W., James K. D., Turner D. J., Andrews R. and Bird A. P. (2010). Neuronal MeCP2 is expressed at near histone-octamer levels and globally alters the chromatin state. Mol. Cell 37, 457–468. doi: http://dx.doi.org/10. 1016/j.molcel.2010.01.030 Tao J., Hu K., Chang Q., Wu H., Sherman N. E., Martinowich K. and … Sun, Y. E. (2009) Phosphorylation of MeCP2 at Serine 80 regulates its chromatin association and neurological function. Proc. Natl Acad. Sci. 106, 4882–4887. doi:10.1073/pnas.0811648106. Yang S.-H. and Sharrocks A. D. (2004). SUMO promotes HDACmediated transcriptional repression. Mol. Cell 13, 611–617. doi: http://dx.doi.org/10.1016/S1097-2765(04)00060-7 Zhou Z., Hong E. J., Cohen S. et al. (2006). Brain-specific phosphorylation of MeCP2 regulates activity-dependent bdnf transcription, dendritic growth, and spine maturation. Neuron 52, 255–269. doi: http://dx.doi.org/10.1016/j.neuron.2006.09.037 Zocchi L. and Sassone-Corsi P. (2012). SIRT1-mediated deacetylation of MeCP2 contributes to BDNF expression. Epigenetics 7, 695– 700.
