Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Epigenetics & Chromatin https://doi.org/10.1186/s13072-019-0298-1 RESEARCH Open Access MeCP2‑E1 isoform is a dynamically expressed, weakly DNA‑bound protein with different protein and DNA interactions compared to MeCP2‑E2 Alexia Martínez de Paz1,25, Leila Khajavi2,3, Hélène Martin3, Rafael Claveria‑Gimeno4,5,6, Susanne Tom Dieck7, Manjinder S. Cheema1, Jose V. Sanchez‑Mut8, Malgorzata M. Moksa9,10, Annaick Carles9,10, Nick I. Brodie11, Taimoor I. Sheikh12,13, Melissa E. Freeman1, Evgeniy V. Petrotchenko11, Christoph H. Borchers11,14,15,16, Erin M. Schuman7, Matthias Zytnicki2, Adrian Velazquez‑Campoy4,5,17,18,19, Olga Abian4,5,6,17,19, Martin Hirst8,9,20, Manel Esteller21,22,23, John B. Vincent12,13,24, Cécile E. Malnou3 and Juan Ausió1* Abstract Background: MeCP2—a chromatin-binding protein associated with Rett syndrome—has two main isoforms, MeCP2-E1 and MeCP2-E2, differing in a few N-terminal amino acid residues. Previous studies have shown brain region-specific expression of these isoforms which, in addition to their different cellular localization and differential expression during brain development, suggest that they may also have non-overlapping molecular mechanisms. However, differential functions of MeCP2-E1 and E2 remain largely unexplored. Results: Here, we show that the N-terminal domains (NTD) of MeCP2-E1 and E2 modulate the ability of the methyl- binding domain (MBD) to interact with DNA as well as influencing the turn-over rates, binding dynamics, response to neuronal depolarization, and circadian oscillations of the two isoforms. Our proteomics data indicate that both iso‑ forms exhibit unique interacting protein partners. Moreover, genome-wide analysis using ChIP-seq provide evidence for a shared as well as a specific regulation of different sets of genes. Conclusions: Our study supports the idea that Rett syndrome might arise from simultaneous impairment of cellular processes involving non-overlapping functions of MECP2 isoforms. For instance, MeCP2-E1 mutations might impact stimuli-dependent chromatin regulation, while MeCP2-E2 mutations could result in aberrant ribosomal expression. Overall, our findings provide insight into the functional complexity of MeCP2 by dissecting differential aspects of its two isoforms. Keywords: MeCP2, Isoforms, Chromatin, Rett syndrome *Correspondence: jausio@uvic.ca 1 Department of Biochemistry and Microbiology, University of Victoria, Petch Building 260, Victoria, BC V8W 3P6, Canada Full list of author information is available at the end of the article © The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat​iveco​mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 2 of 16 Background sequence is identical for both isoforms (Figs. 1a and Methyl CpG-binding protein 2 (MeCP2) was first iden- 2a). tified through its ability to bind methylated DNA [1]. MeCP2-E1 is likely the ancestral form of the protein, Mutations in the MECP2 gene were later associated as orthologues are present across vertebrate evolution, with Rett syndrome (RTT; OMIM 312750), a severe whereas orthologous sequences of the exon 2 coding neurological disorder that is among the most common region have only been found in mammalian genomes causes of intellectual disability in females [2]. [5]. Although splicing variants often encode proteins MeCP2 gene has four exons than can be alternatively with different functions, in the case of MeCP2-E1 and spliced to produce two transcripts. The transcript skip- E2 isoforms, this remains still controversial [6, 7]. The ping exon 2 has translation initiation in exon 1 and presence of a polyalanine tract followed by a polygly- encodes MeCP2-E1. This isoform is slightly longer (498 cine tract in E1 N-terminal domain (NTD) could be an amino acids in humans) and has 21 unique N-terminal indication of a potential functional difference [8]. In amino acids. When exon 2 is included in the transcript, this regard, polyalanine domains within various protein translation initiates in exon 2 to give rise to MeCP2-E2, families are thought to have a convergent origin, sug- a shorter variant (486 amino acids in humans) with 9 gesting that a specific function for these tracts has been unique N-terminal amino acids [3, 4]. The remaining selected by evolutionary pressure [9]. The existence of Fig. 1 Biophysical characterization of the MeCP2-E1 and E2 NTD-MBD domain interaction with DNA. a Schematic representation of the MeCP2-E1 and E2 isoforms depicting the unique NTD amino acid sequences and shared domains. b Fluorescence thermal denaturation curves for E1 and E2 NTD-MBD protein fragments in the presence of unmethylated and mCpG-dsDNA. Unfolding traces were fitted considering a two-state unfolding model. c Unfolding stability parameters obtained from thermal denaturations followed by intrinsic tryptophan fluorescence. d Calorimetric titrations of E1 and E2 NTD-MBD proteins interacting with dsDNA plots show the thermograms (thermal power as a function of time) and the binding isotherms (normalized heats as a function of the dsDNA/protein molar ratio). e Buffer-independent dsDNA binding parameters ­(Kd, dissociation constant; ΔG: Gibbs free energy of interaction; ΔH: enthalpy of interaction; −TΔS: entropic contribution of interaction; ΔCP: heat capacity of interaction; nH: number of protons exchanged upon complex formation) obtained from calorimetric titrations at pH 7 Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 3 of 16 Fig. 2 Isoforms N-terminal processing, turn-over rates, and dynamics. a Mass spectrometry sequencing of the N-terminal end of the MeCP2 protein (in vitro). N-terminal peptide coverage alignment chart and high-resolution mass spectra showing N-methionine excision (NME) and N-acetylation (NA) of the N-termini of MeCP2-E1 and MeCP2-E2. NA (+42 Da) of N-terminus amino acid is shown highlighted in yellow. b–d Cycloheximide-chase assays of the E1 and E2 MeCP2 isoforms. Densitometries and representative western blots performed after cycloheximide treatments of b SH-SY5Y cells overexpressing (OE) E1 and E2 isoforms fused to GFP, c differentiated SH-SY5Y, and d rat cultured cortical neurons with detection of endogenous E1 and E2 isoforms. e Densitometric analysis and representative western blots showing endogenous E1 and E2 levels in frontal cortices of mice euthanized at 12 a.m. and 12 p.m. f KCL depolarization and representative Western blots analysis of total endogenous MeCP2 of cultured cortical neurons and E1 and E2 of transfected cultured cortical neurons overexpressing Flag-MeCP2-E1 or E2. Represented data are mean ± S.E.M. (n = 7–8). * P < 0.05 two-tailed Mann–Whitney test. MeCP2 levels were normalized using β actin and/or histone H3 non-overlapping functions of the E1 and E2 isoforms E2 does not functionally compensate for the lack of E1 is supported by a difference in their relative abundance [12, 13]. However, the high degree of structural simi- during development and in diverse regions of the brain larity between MeCP2 isoforms points towards a high [10, 11]. Moreover, Rett syndrome-causing mutations extent of functional overlapping, and some findings described so far involve solely the E1 isoform, and iso- reinforce this idea. For instance, E2 expressed at levels form-specific mouse knockouts show Rett-related phe- comparable to those of E1 was reported to prevent key notypes for E1 knockout but not for E2, suggesting that Rett-like phenotypes in mice models of Rett syndrome, indicating that part of the difference between isoforms Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 4 of 16 could simply be related to the disparity in temporospa- for unmethylated dsDNA, thus resulting in E1 isoform tial expression and protein levels [7]. having a slightly lower discrimination capability for meth- Given the poorly understood nature of the structural ylated/unmethylated dsDNA (Fig. 1e). Strikingly, the and functional differences between E1 and E2 isoforms, main intermolecular DNA-binding driving forces for the we decided to investigate this further. Our study compre- two isoforms are of different nature, displaying opposed hensively describes for the first time differences between thermodynamic binding profiles: dsDNA interaction MeCP2 isoforms, using various complementary biophys- with E1 is enthalpically driven and with E2 is entropically ical, biochemical, and genomic approaches. This work driven; thus, while E1 interacts with favorable binding provides a detailed framework for the further under- enthalpy (ΔH) and unfavorable binding entropy (−TΔS), standing of the many fold functional aspects of MeCP2, E2 interacts with negligible binding enthalpy and favora- thus shedding light onto the pathophysiology of Rett syn- ble binding entropy (Fig. 1e). Therefore, the interaction drome and other neurological disorders. of E1 isoform with dsDNA is mainly driven by specific interactions between the protein and the dsDNA (i.e., Results hydrogen bonds and electrostatic interactions), and the Biophysical characterization of MeCP2 isoforms N‑terminal interaction of E2 isoform with dsDNA is mainly driven domains by unspecific interactions (i.e., hydrophobic desolvation As mentioned in the introduction, the different function- and steric arrangements). In addition, E1 isoform exhib- ality of the MeCP2 isoforms has long remained contro- its a larger binding heat capacity (ΔCP) and the formation versial. However, there are many indirect hints to suggest of its complex with dsDNA releases a larger number of otherwise, including a different pattern of expression protons (nH). Overall, these observations indicate that during mouse brain development (Additional file 1: Fig. the amino acid residues at the N-terminal regions of E1 S1 A) [11] and various evidences previously mentioned. and E2 NTDs have significant influence not only on pro- Interestingly, the two MeCP2 isoforms differ only in their tein stability, but also on the interaction with the dsDNA: N-terminal domain (NTD) (Fig. 1a), which has been E1 is slightly less stable and exhibits lower affinity for previously described to lack any DNA-specific binding dsDNA than E2 isoform. Fluorescence recovery after structure, but has the ability to stabilize the neighboring photobleaching (FRAP) data for the two isoforms sup- methyl-binding domain (MBD) and its binding to meth- ports this, with E1 having a more rapid recovery trajec- ylated DNA [14, 15]. A partial folding of this unstruc- tory than E2, suggesting looser binding, although t-half tured region might contribute to the interaction with and mobile fractions were not significantly different double-stranded DNA (dsDNA), thus having a differen- (Additional file 1: Fig. S1C). These properties could also tial impact on E1 and E2 binding properties. Therefore, be reflecting a differential ability of MeCP2 isoforms to we decided to compare the different biophysical proper- interact with other molecules, different turn-over rate, ties of E1 and E2 NTDs. Constructs consisting of the E1 intracellular trafficking, or susceptibility to undergo post- or E2 specific NTD followed by the MBD were analyzed translational modifications. as previously described [15]. Thermal unfolding studies of E1/NTD-MBD and E2/NTD-MBD (Fig. 1b) indicate Higher MeCP2‑E1 protein turn‑over in neuronal systems that E1 isoform shows a slightly lower mid-transition reflects its involvement in dynamic processes temperature (temperature at which 50% of the protein MeCP2 is an intrinsically disordered protein (IDP) [16], is unfolded, Tm) in all situations considered according to thus highly susceptible to proteolytic degradation. The the two-state unfolding model (Fig. 1c; Additional file 1: lower affinity of the E1 NTD-MBD region for DNA or its Fig. S1B), showing a slightly lower structural stability. lower folding stability might reflect a higher presence in Following the same trend, E1 isoform also shows a solution or the occurrence of a larger exposed surface to diminished unfolding enthalpy (ΔH) (Fig. 1c), indicating be targeted for proteasomal degradation [17]. Therefore, a lower cooperativity in the thermal unfolding, suggest- we decided to compare the half-lives of the two MeCP2 ing that amino acid residues located at the NTD might be isoforms in different neuronal systems. The end-terminal important for the stability of the folded regions located amino acid has great impact in protein degradation [18], in the MBD. The nature of protein–DNA interactions and thus, we first assessed the N-terminal processing of was further assessed by determining their thermody- both proteins. The NTDs were expressed in HEK293T namic profile with isothermal titration calorimetry (ITC), cells, and purified and analyzed by mass spectrometry considering a single binding site model [15] (Fig. 1d, e). (MS). Our previous MS sequencing of the N-terminal Results show that compared to E2, E1 exhibits ninefold tael of MeCP2-E1 [5] showed no peptides with N-ter- lower binding affinity (higher dissociation constant, Kd) minal methionine (NM), indicating complete NM exci- for methylated dsDNA and fivefold lower binding affinity sion (NME) at the first residue (P1) position (Fig. 2a Top Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 5 of 16 panel). Acetylation of the initial alanine residue (P′1) (Fig. 2f ). By contrast, E2 shows a stable pattern, exhibit- after NME was observed. ing levels similar to those of non-treated cells throughout In addition, we observed some peptide reads with ala- the whole duration of the experiment (Fig. 2f ). Hence, nine 1, or alanine 1 and 2, or alanine 1–4, or 1–5 excised our data confirm the existence of different dynamics of and acetylation of the subsequent alanine. For MeCP2- MeCP2 isoforms that are consistent with a different role E2, on the other hand, we found reads in which N-ter- of the proteins within the neuronal context. minal methionine (P1 position) is retained and acetylated and few peptide reads with NME and acetylation of the Genome‑wide distribution of MeCP2‑E1 and MeCP2‑E2 penultimate valine (P’1) (Fig. 2a Bottom panel). All post- isoforms translational modifications (PTMs) reported received The differences described between E1 and E2 in terms of Ascores of 1000. The complete methionine excision their affinity for methylated DNA and dynamics might and the presence of alanine as first residues in E1 sup- have an influence on and/or reflect a differential genomic port a faster turn-over rate of this isoform, compared distribution. Chromatin immunoprecipitation and to the E2 bearing a methionine or valine as N-terminal sequencing (ChIP-seq) of E1 and E2 in frontal cortices amino acids [19]. We then decided to test this possibil- of mice euthanized at 12 a.m. and 12 p.m. showed broad ity by performing cycloheximide (CHX) chase assays and distributions for both isoforms, as previously noted for Western blot (WB) in three different contexts: undiffer- total MeCP2 (Additional file 3: Fig. S3A and B, [21–23]. entiated SH-SY5Y neuroblastoma cells transfected with The overall distribution of the isoforms along different E1- and E2-EGFP (enhanced green fluorescent protein) genomic regions was similar (FDR ≤ 0.001, SICER algo- fusion proteins (Fig. 2b), differentiated SH-SY5Y (Fig. 2c), rithm: window 600 bp; gap 200 bp; Fig. 3a). Despite the Additional file 2: Fig. S2A), and DIV7 rat cortical neu- similar isoform’s general distribution, we were able to rons (Fig. 2d). These experiments consistently showed identify different significantly enriched binding motifs a tendency of E1 to be degraded faster than E2, both at specific for E1 (using E2 bound peaks as background), the endogenous level and when the isoforms were over- GCT​ GAG​ C (e-value: 1.2 e−145), and GCC​ACA​GCA expressed in cell cultures (detected using specific in (e-value: 2 e−99) indicating a differential binding site house-made antibodies, Additional file 2: Fig. S2B). Our preference (Fig. 3b). As MeCP2 has been described to results prompted us to investigate if MeCP2 isoforms be a transcriptional regulator, we analyzed the distribu- would show differential behavior in two highly dynamic tion of MeCP2 isoforms around transcribed regions. The neuronal settings, the circadian cycle and neuronal acti- average binding profiles of E1 and E2 to regions span- vation. First, we took advantage of the system which we ning 1.5 kb upstream of transcription start sites (TSS) to reported previously displaying total MeCP2 24 h oscil- 1.5 kb downstream transcription end sites (TES) demon- lations [20]. Analysis of frontal cortices obtained at 12 strate a consistently similar binding pattern for both iso- a.m. and 12 p.m. shows a noticeable 30% reduction of E1 forms, with a marked depletion at TSS and a peak at TES protein level at 12 p.m., while E2 levels remain similar (Fig. 3c). at these two times (Fig. 2e). The second scenario involv- Interestingly, a closer inspection around TSS regions ing MeCP2 dynamics was neuronal activation after KCl (Fig. 3d) revealed a marked depletion of E2 precisely at exposure. Protein levels were measured at different time the TSS. In contrast, the E1 isoform is depleted before points after depolarization (55 mM KCl). The inability the TSS and corresponding to the −1 nucleosome region, of the MeCP2-E2 antibody to detect this isoform in rat with a slight increase on the TSS at 12 a.m. that decreases neurons prompted us to first determine the endogenous at 12 p.m. These results suggest a differentiated role of levels of total MeCP2 after treatment. Due to the high the two isoforms in shaping the chromatin structure E1 abundance compared to E2, this mostly corresponds around the TSS. to the E1 isoform. Next, we assessed the two isoforms’ We then clustered the genes based on their MeCP2 dynamics by transfecting cultured rat neurons with flag- occupancy. Heatmap clusters and profiles for the log2 tagged E1 and E2. The results show a fast increase of total ratio plots failed to reveal any differential binding of the MeCP2 levels immediately after KCl treatment followed isoforms to specific gene clusters (data not shown); how- by a decrease to basal levels at around 4 h after treat- ever, we detected daily differences for each isoform occu- ment (Fig. 2f ). Interestingly, we observe a completely dif- pancy throughout gene bodies (Fig. 4a). For instance, E1 ferent pattern between MeCP2 isoforms. As expected, cluster 1 showed a flat profile at 12 a.m. and an increased E1 shows a trend similar to that of total MeCP2, rapid binding at 12 p.m. (Figure 4b). In the case of E2, cluster 4 upregulation upon depolarization that is maintained exhibited an increased binding at 12 p.m. compared to 12 during 3–4 h, and then, protein levels decrease to reach, a.m., while cluster 5 displayed lower binding at 12 p.m. in this case, approximately 50% of the initial E1 levels (Figure 4c). Functional pathways associated with genes Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 6 of 16 Fig. 3 Genome-wide distribution and dynamics of MeCP2 isoforms. a Bar plot depicting the distributions of regions enriched in the MeCP2 isoforms across eight defined genomic features. b DNA-binding motifs enriched for MeCP2-E1 (and excluding E2) shown as motif logos based on aligned, over-represented patterns found in the binding sites. The overall height of each stack indicates the sequence conservation at that position (measured in bits), whereas the height of symbols within the stack reflects the relative frequency of the corresponding amino or nucleic acid at that position. c ChIP-seq average profiles across 3 Kb upstream the TSS and 3 Kb downstream the TES of genic regions occupied by E1 at 12 a.m. and 12 p.m. (top) and E2 at 12 a.m. and 12 p.m. (bottom). d Details of the previous representations focusing in a 6 Kb region surrounding the TSS of E1 and E2 bound genes (top and bottom panels, respectively) present in each cluster were analyzed using the Kyoto general trends observed in our ChIP-seq-analysis, despite encyclopedia of genes and genomes (KEGG) (Fig. 4d left the very slight variations of the isoforms’ occupancies graphs). All three clusters were enriched in genes related during the day (Fig. 4d right graphs). Overall, our results to sensory transduction like olfaction or taste, and with suggest that beyond the common functions in which histone proteins (E1 was associated with genes encoding both isoforms are involved, they regulate different sets H2A family members (p value: 1.16 e­ −06) and E2 mainly of genes and display distinct dynamics on their genomic with members of the histone cluster 1 (cluster 4 p value: occupancy, reinforcing the existence of non-overlapping 1.65 ­e−13 and cluster 5 p value: 2.23 e­ −27). MeCP2 iso- roles. form-specific enrichments were related to neuroactive ligand–receptor interaction in E1 and ribosomal proteins MeCP2‑E1 and E2 protein partners in E2. Interestingly, cluster 5 contains several genes asso- IDP proteins are characterized by their inability to ciated with the neurodegenerative diseases Huntington acquire a stable secondary structure when free in solu- (p value: 4.17 ­e−08), Parkinson (p value: 9.87 ­e−06) and tion. This confers the structural flexibility that enables Alzheimer (p value: 9.15 e­−06). ChIP-qPCR validations them to serve as scaffolds for the recruitment of partners of randomly selected genes of each cluster confirmed the and thus function as interacting hubs [24]. Interestingly, Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 7 of 16 Fig. 4 MeCP2-E1 and E2 isoforms display diurnal dynamic genomic binding. a Heatmaps representing the log2 ratios obtained for E1 and E2 ChIP experiments; each column is divided into five clusters using the k-means algorithm. Protein occupancy is represented by color intensity, where the darker the color, the higher the protein enrichment. b Comparison of E1 enrichment at 12 a.m. vs. 12 p.m. showing occupancy differences in different clusters of interest. c Heatmap depicting the E2 12 a.m. vs. 12 p.m. shows a dynamic binding in clusters 4 and 5 (yellow and orange, respectively). d Left graphs: top-enriched functional pathways (−log10 (P)) of genes included in dynamic E1 and E2 gene clusters [Kyoto encyclopedia of genes and genomes (KEGG)]. Right graphs: ChIP-qPCR results validating MeCP2-E1 and E2 variations in occupancy of genes included in each of the gene clusters obtained Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 8 of 16 IDPs, including MeCP2, usually acquire ordered struc- decided to perform a comprehensive proteomic analysis tures upon binding to their interacting partners, allowing to look for MeCP2-E1 and E2 protein partners. Endog- the exposure of molecular recognition features (MoRFs) enous E1 and E2 from whole brain lysates were immuno- to further make contacts with other molecules [25, 26]. precipitated, using normal rabbit IgG and blocking of E1 Thus, the possibility exists that the aforementioned E1 and E2 antibodies with blocking peptides were used as and E2 differences in unfolding temperature and affinity negative controls. Co-immunoprecipitated proteins were for DNA could expose differential interacting surfaces. separated by SDS-PAGE and different gel fractions sec- These attributes together with their previously discussed tioned for protein identification by mass spectrometric expression patterns [3, 4, 27] raise the possibility that E1 analysis (Fig. 5a). and E2 might be involved in non-overlapping molecu- We chose proteins identified by at least 2 significantly lar functions that perhaps could be defined through the matching peptides which were absent from the negative identification of their protein interactors. Therefore, we controls. This filter rendered 40 interacting proteins for Fig. 5 MeCP2-E1 and E2 interacting proteins. a Schematic workflow of the proteomic analysis. b, c Proteins partners identified for each isoform. d Selected pathways enriched in E1 and E2 interacting proteins as identified by functional clustering (DAVID Gene Ontology Bioinformatics Resources). Newly identified interactors are shown in black and previously identified interactors are highlighted in orange Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 9 of 16 the E1 isoform and 7 for E2 (Fig. 5b, c). As a good vali- proteins [47, 48]. Of note, one of the best characterized dation for our approach, we detected several previously components of the nuclear matrix is the attached region- described MeCP2 interactors (Fig. 5b, c, interactors high- binding protein (ARBP), a chicken MeCP2 orthologue lighted in orange [28–34]). [49] that binds methylated DNA within matrix attach- Functional clustering of co-eluted proteins (DAVID ment region (MAR) elements [48, 50]. [35]) uncovered functional enrichments, especially for Overall, the lack of shared protein partners by the E1 (Fig. 5d). E1 co-eluted proteins are highly enriched MeCP2-E1 and E2 isoforms suggests their involvement in for β-Tubulins, the building blocks of microtubules, and similar general mechanisms like RNA processing, chro- microtubule-associated proteins such Adducin 1 (Add1) matin control of transcription, or microtubule regulation, or microtubule-associated protein 6 (Map6). Importantly, but performing non-redundant functions through the microtubule assembly initiates from the centrosome, interaction with different partners. organelle associated with MeCP2 function in microtu- bule stability, and mitotic spindle organization [36–38]. Discussion Proteins related to mRNA splicing and mRNA process- The existence of mutations affecting only the MeCP2-E1 ing were also highly represented among E1 partners (for isoform in Rett patients [e.g., p.Ala2Val; [13, 51]] sug- example, 116 kDa U5 small nuclear ribonucleoprotein gest that endogenous E2 expression cannot compensate component [Eftud2], Heterogeneous nuclear ribonucleo- for the lack of functional E1. An important question thus proteins L [Hnrnpl] or DEAD (Asp-Glu-Ala-Asp) box arises as to whether this is simply related to the lower lev- polypeptides 5 and 17 [Ddx5 and Ddx17]). MeCP2 func- els of E2 found in neurons [27] or it is due to the exist- tions on RNA splicing or mRNA processing have been ence of E1-specific functions that cannot be provided by previously described [33, 39, 40], but still lack deep inves- the E2 isoform. The different cellular distribution of the tigation. As we expected, functions related to chromatin two isoforms and their distribution during brain develop- regulation are also enriched among MeCP2-E1 partners, ment (Additional file 1: Fig. S1) also suggest a different as we found the nucleosome-core histone H2A and the functionality. variant H3.3, the chromatin regulators Brg1 associated The NTD is the only structural feature that differs factor 170 (BAF170), member of the switch/sucrose non- between the two MeCP2 isoforms, and currently, there is fermenting (SWI/SNF) complex, and MTA2, subunit of a lack of information regarding any potential functional the nucleosome remodeling deacetylase (NuRD) com- difference between E1 and E2 NTDs. The NTD of these plex [41]. Functional network analysis (STRINGv10 [42]) isoforms has generically been described as a highly disor- revealed a higher than expected number of connections dered region able to acquire secondary structure, as dem- between all E1 and E2 interactors (Additional file 4: Fig. onstrated by the coil-to-helix transitions exhibited in the S4; p value < 0.001), and suggests the participation of E2 presence of hydrogen-bond stabilizers [26]. Such confor- in processes similar to those involving E1, but through mational transitions contribute to enhancing the MBD the interaction with a different set of protein partners. In affinity for DNA [14]. To the best of our knowledge, the this regard, among E2 interactors, we found the microtu- biophysical characterization of the interaction of NTD- bule-associated protein RP/EB family members 1 and 3 MBD fragments of the two MeCP2 isoforms with DNA (Mapre1 and Mapre3), important for microtubule organ- is the first of its kind (Fig. 1). Most of MeCP2 functions ization [43]. The E2 interactor fused in sarcoma (FUS) is rely on its ability to bind nucleic acids, and in this regard, involved in mRNA processing, with Mecp2 being one of these results uncover a fundamental ninefold difference its known target genes [44]. In the chromatin regulation in affinity for DNA of E2 over E1 isoform. This could be group, we found that E2 specifically interacts with two one of the basic structural features responsible for shap- recently described MeCP2 protein partners: Transducin- ing the functional discrepancies between the isoforms. β-like 1 (Tbl1) and Tbl1-related 1 (Tbl1r1), components The main differences observed here between the of the nuclear receptor co-repressor (N-CoR) complex MeCP2 isoforms can be summarized as follows: E1 (the [28, 29]. Interestingly, E2 also interacts with the polymer- major isoform of MeCP2 in neurons) shows a lower ase I transcription and release factor (Ptrf ), and protein DNA-binding affinity and a lower structural stability involved in ribosomal DNA (rDNA) transcription [45]. (Fig. 1). E1 also exhibits a higher basal degradation rate E1 co-eluted proteins include spectrin β1, lamin B2, the in various neuronal settings and enhanced dynamic band 4.1 proteins B and N, and matrin 3 (the latter was fluctuations of protein levels via diurnal rhythm oscil- previously reported to interact with E1 in neuronal nuclei lations and neuronal depolarization (Fig. 2). Within [31]), components of the nuclear matrix [46], classically a neuronal context, these attributes are especially defined as a fibrogranular structure which consists of interesting given the peculiar chromatin relation- nucleoskeleton/nuclear lamina networks and associated ship between MeCP2 and the linker histone H1 in Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 10 of 16 this tissue. Under normal conditions, neurons express [16]. Our ChIP-seq results reinforce this MeCP2 function MeCP2 to near histone-octamer levels and contain half and add a dynamic component to it. the amount of H1. By contrast, when MeCP2 is not Quite unexpectedly, all gene clusters displaying expressed, H1 returns to the levels which are observed dynamic diurnal binding of the two MeCP2 isoforms in other somatic tissues [52]. Therefore, it is likely were also enriched in gene-encoding replication-depend- that MeCP2-E1 could function within this setting as a ent (RD) histones. Expression of RD histones has been DNA-methylation-dependent highly dynamic linker recently detected in terminally differentiated cells and tis- histone, needed to allow for rapid chromatin structural sues, including neurons and brain [56, 57]. Some of these changes in response to external stimuli. This is particu- genes encode histone isotypes with a certain sequence larly important in neurons, given their versatile ability divergence [56], and could possibly affect the histone to readily modify gene expression as a result of their interactions within the nucleosome. Further analyses unique methylome [53, 54]. will be required to assess if MeCP2 has any regulatory In this regard, signal transduction from the cell surface effect on the expression of such genes, but it is tempting to the genome often relies on the cytoskeleton–nucle- to speculate a possible role in the generation of variant oskeleton–chromatin interconnection [46]. Importantly, nucleosomes present in adult brain [56]. we have identified MeCP2-E1 and E2 protein partners An additional distinctive function for MeCP2-E2 which associated with every part of this system: the cytoskel- could be inferred from our results has to do with ribo- eton (i.e., tubulins, Map6, and Mapre1 and 3), nuclear somal gene expression regulation. Our co-immuno-pre- envelope/matrix-associated proteins (Lamin B2, Band cipitation experiment demonstrated the interaction of E2 4.1 proteins, Spectrin, or Matrin 3), and chromatin (his- with Ptrf (Fig. 5c). Transcription of ribosomal genes has tone proteins, Hp1γ, Mta2, or Baf170) (Fig. 5). Our find- been reported to be dependent on the DNA methylation ings open up the possibility of MeCP2 functioning as an status [58] and occurs through the formation of nucleo- important player in signal transduction. In particular, E1 tide loops linking initiation and termination gene regions, could play a prominent role in the neuron-specific nucle- process in which Ptrf participation is essential [45]. Our oskeleton–chromatin connection, due to its remarkable ChIP-seq results provide evidence for a dynamic MeCP2- abundance and the dynamism observed upon the applica- E2 genomic binding to ribosomal genes (Fig. 4). In sup- tion of external stimuli such as neuronal depolarization. port to our results, MeCP2 has been previously linked to Because of its higher abundance, E1 appears to exhibit nucleolar changes during neuronal maturation [59]. a more dynamic behavior; however, E2 also exhibits an Another observation made through our ChIP-seq oscillating genomic-binding nature. Surprisingly, most of analysis is that the two isoforms possess significant dif- the daily MeCP2 isoform-binding differences observed ferences in DNA-binding site preference; E1 targets are overlap with genes encoding for sensorial receptors, such significantly enriched for the DNA motifs GCT​GAG​C as olfactory (ORs) and taste (TASRs) receptors (Fig. 4d). and GCC​ACA​GCA. Interestingly, the latter motif con- This seemingly counterintuitive result is very interesting tains the trinucleotide CAC, described by Lagger and as preliminary observations have shown expression of colleagues to be, when methylated, a high affinity-bind- ORs and TASRs in brain regions not related to the direct ing site for MeCP2 in brain [22]. detection of odors and flavors [55]. The study of the so- called “ectopic ORs” (outside olfactory epithelium) is in Conclusion its infancy, but apparently, they act as chemoreceptors Overall, the present work provides support to the which are important to maintain cellular homeostasis, notion that Rett syndrome arises from the simultane- and some of them are able to activate complex cellular ous impairment of different cellular functions involving responses mediated by neurotransmitters or hormones both MeCP2 isoforms. For instance, mutations of E1 may [55]. The expression of olfactory receptors has been have a larger impact in neuronal chromatin structure and described to be upregulated in MeCP2 KO mice and stimuli-dependent gene expression dynamics. We have downregulated in mice overexpressing the protein. These previously described in MeCP2 KO mice a decrease in data support a potential role for MeCP2 in transcrip- the circadian gene oscillations of the brain-derived neu- tional regulation of these genes in different regions of the rotrophic factor (Bdnf) and somatostatin (Sst) genes [20]. brain structures such as cerebellum, amygdala, and hypo- By contrast, similar mutations in E2 could be involved in thalamus [21]. More importantly, the expression of these the deregulation of ribosomal expression or microtubule receptors in brain is altered in neurodegenerative dis- control. eases such as Parkinson’s, Alzheimer’s and in prefrontal The seemingly contradictory literature available to cortex in schizophrenia [55], disorders in which MeCP2 date regarding the degree of functional overlapping of expression has been also observed to be dysregulated MeCP2 isoforms is likely the result of the lack of studies Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 11 of 16 carried out on the endogenous native proteins. Over- nitrocellulose membranes (GE Healthcare), and block- all, our results provide strong support for the existence ing by incubating in 5% skimmed milk in PBS with 0.1% of both different and overlapping functions between the Tween 20. Membranes were incubated with specific anti- two. Importantly, our study uncovers some functional bodies, either overnight at 4 °C or for 1 h at room tem- aspects of MeCP2 that were previously unknown. It perature. Antibodies and dilutions used were as follows: opens the door to further investigation that will be help- MeCP2 1:5000, β-actin 1:20,000, Histone H4 1:20,000 ful to understand the role of this complex protein in the Histone H3 1:20,000 (Abcam), Flag 1:5000 (Sigma- healthy state and the consequences of its deregulation in Aldrich), EGFP 1:1000 (Thermo Fisher), in house rabbit Rett syndrome and other neurological disorders. polyclonal antibodies MeCP2 E1 1:1000 and E2 1:500. Fluorescent secondary antibodies were (Li-Cor) 1:10,000. Methods Densitometries were performed with Li-Cor Image Stu- Animals and cell lines dio Light software. The Mecp2-null mouse strain [60] was purchased from Jackson Laboratories (003890B6.129P2(C)- Protein expression Mecp2 < tm1.1Bird >/J), maintained on a CD1 back- Plasmids were transformed into BL21 (DE3) Star E. coli ground, and genotyped as described in [61]. C57BL/6 strain. Protein expression was induced with 1 mM iso- mice and Sprague–Dawley rats were maintained under propyl 1-thio-β-d-galactopyranoside (IPTG) at 18 °C standard animal house conditions (12 h dark–light overnight. Cells were sonicated and treated with ben- cycles on ad libitum food and water intake). Cortical pri- zonase (Merck-Millipore). Proteins were purified using mary neurons were prepared from Sprague–Dawley rat immobilized metal ion affinity chromatography (IMAC) embryos as described previously [62]. For depolarization in a HiTrap TALON column (GE-Healthcare Life Sci- assay, neuronal activity was blocked by pre-treatment ences) with two washing steps before a 10–150 mM imi- of the cells with 1 mM tetrodotoxin (TTX), 100 mM dazole elution gradient. Removal of the histidine-tag was DL-2-amino-5-phosphonovalerate (APV), and 20 mM performed with GST-tagged PreScission Protease (GE- 6-cyano-7-nitroquinoxaline-2, 3-dione (CNQX). Depo- Healthcare). Further purification was performed using a larization was achieved by 30 min exposure to 55 mM HiTrap TALON column and a GST TALON column (GE- KCl in Tyrode buffer. Transfections with plasmids Healthcare). The identity of all proteins was checked by expressing Flag-MeCP2 isoforms (kind gift from Angus mass spectrometry (4800plus MALDI-TOF/MS, Thermo Wilson [63]) were performed using the Amaxa rat neu- Fisher). Stability and binding assays were performed at ron nucleofector kit (Lonza) following the manufacturer’s different pH and buffer conditions [50 mM Tris (pH 7–9), instructions. 0–150 mM NaCl; 50 mM Pipes (pH 7); 50 mM Phos- Human cell lines SH-SY5Y and HEK 294 were obtained phate (pH 7)]. Buffer exchanges were performed using a from the American Type Culture Collection (ATCC) and 10 kDa-pore size ultrafiltration device (Amicon centrifu- were maintained in DMEM supplemented with 10% FBS, gal filter, Merck-Millipore) at 4000 rpm and 4 °C. penicillin–streptomycin and l-glutamine (Gibco) in a 37 °C, 5% ­CO2 humidified incubator. Differentiation was Double‑stranded DNA performed as previously described [64]. Cells were trans- HPLC-purified single-stranded DNA fragments were fected with MeCP2-E1 and MeCP2-E2 C-terminal GFP purchased from Integrated DNA Technologies and [5] using Lipofectamine 3000 (Invitrogen). Cyclohex- annealed. Sequences correspond to the promoter IV of imide (CHX) treatments were performed 48 h after the mouse brain-derived neurotrophic factor (BDNF) transfection. [15]. Western blot Fluorescence spectroscopy Mouse brains (brain development) for each development Thermal unfolding studies were performed in a Cary time period were used to prepare western blot samples Eclipse fluorescence spectrophotometer (Varian—Agi- as per [65]. Nuclear extracts of HEK293 cells express- lent) in three steps. Fluorescence emission spectra were ing the 3 × Flag versions of E1 and E2 were used to pre- recorded from 300 to 400 nm (excitation 290 nm and pare serial dilutions and calculate MeCP2-E1/MeCP2-E2 bandwidth 5 nm). Protein concentration was set at 5 µM. ratios. Frozen frontal cortices (circadian) and cell pel- Thermal stability assays were performed at a heating rate lets (CHX) were homogenized on Laemmli buffer 3% of 1 °C/min and at the wavelength for maximal spec- B-Mercaptoethanol. In all experiments, proteins were tral change (330 nm). Thermal unfolding experiments separated by sodium dodecyl sulfate-polyacrylamide gel were analyzed considering a two-state unfolding model electrophoresis (SDS-PAGE) [66] and transferred onto [15]. The stabilizing effect upon dsDNA interaction was Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 12 of 16 assessed performing thermal denaturations in the pres- was used for the MS/MS scans. XCalibur 2.2 (Thermo- ence of methylated and unmethylated DNA (at 10 µM) Fisher Scientific) was used to acquire the raw data files under the same conditions. and further processed with the PEAKS 7 search engine (Bioinformatics Solutions) using a database consisting of Isothermal titration calorimetry (ITC) the wild-type MeCP2 constructs. Ascores were assigned Proteins–dsDNA interactions were characterized using for the peptides and PTMs using the PEAKS 7 software. an Auto-iTC200 microcalorimeter (MicroCal—Malvern Instruments). 3–5 µM protein was titrated with 50 µM Chromatin immunoprecipitation dsDNA (2 µL titrant every 150 s). Interaction parameters Mice were euthanized 6 and 18 h after first light stimu- were obtained as previously described [15]. Buffer-inde- lus (12 a.m. and 12 p.m., respectively). One frontal cor- pendent binding parameters and protons released were tex was crosslinked in 1x PBS 0.5% formaldehyde at room determined by calorimetric titrations using buffers with temperature for 5 min and quenched by adding 0.125 M different ionization enthalpies [15], and binding heat Glycine for 5 min. The tissue was centrifuge and washed capacity was determined by calorimetric titrations at dif- twice with ice-cold PBS. The pellet was resuspended in ferent temperatures. 1 mL ChIP lysis buffer [10 mM HEPES (pH7.9), 1.5 M ­MgCl2, 10 mM KCl, 0.5 mM DTT, and 0.1% NP-40], Fluorescence recovery after photobleaching (FRAP) dounce-homogenized (10 strokes), and incubated on MeCP2-E1 and MeCP2-E2 isoforms were expressed in ice for 20 min. Nuclei were centrifuged at 2500×g for (Nunc™; NalgeNunc). Experiments were performed at HEK293T cells in chambered cover glass culture plates 5 min and resuspended in 300 µL of RIPA buffer [50 mM Tris–HCl (pH 8), 150 mM NaCl, 0.5% SDS, 0.5% Sodium 37 °C and 5% C­ O2. Confocal time-lapse images of frames deoxycholate, 1% Triton-100]. Suspension was sonicated (512 × 512 pixels) were captured at 488 nm excitation in a Bioruptor (Diagenode) at high power for 15 min with 0.05 transmissions for GFP-tagged protein post- with 30 s on/off intervals. Chromatin was centrifuged bleach recovery. Images were recorded with a minimum at 16,000×g for 10 min. Supernatant was diluted ten of 10 pre-bleach frames, 250 µs bleach time with 405 nm times with RIPA buffer without SDS and pre-cleared for laser line at 100% transmission, and 150 post-bleach 2 h with 20 µL of Dynabeads Protein G magnetic beads frames were recorded at equal time intervals. (Thermo Fisher). Eight microns of MeCP2 E1 and E2 antibodies (produced in house, Additional file 2: Fig. Cycloheximide‑chase assay S2B) and Normal Rabbit IgG (Cell Signaling Technolo- Cells were treated with 10 µg/mL of cycloheximide gies) were bound to 25 µL of Dynabeads Protein G fol- (Sigma-Aldrich) and harvested at the indicated times. lowing the manufacturer’s instructions and incubated 1 h Samples collected at each time point were then analyzed in PBS/5% BSA. Inmunoprecipitations were performed by western blot. at 4 °C overnight while tumbling. Supernatants were discarded and beads washed twice with 1 mL of low salt Mass spectrometry to determine MeCP2 PTMs buffer [50 mM Tris–HCl (pH 8.0), 150 mM NaCl, 0.1% All protein samples were digested overnight at 37 °C SDS, 1% NP-40, 1 mM EDTA, 0.5% Sodium deoxycolate], with trypsin, using 50:1 protein:enzyme ratio. Digested twice with 1 mL of high salt buffer [50 mM Tris–HCl (pH peptide mixtures were desalted using ­C18 reverse phase 8.0), 500 mM NaCl, 0.1% SDS, 1% NP-40, 1 mM EDTA, columns, and then loaded onto a 50 cm × 75 μm ID 0.5% Sodium deoxycolate], twice with 1 mL of LiCl buffer column with RSLC 2 μm ­C18 packing material (EASY- [50 mM Tris–HCl (pH 8.0), 250 mM LiCl, 0.1% SDS, 1% then eluted into a Q-Exactive™ Hybrid Quadrupole- Spray, Thermo-Fisher) with an integrated emitter, and NP-40, 1 mM EDTA, 0.5% sodium deoxycholate], and Orbitrap™ mass spectrometer (Thermo-Fisher) using an twice with 1 mL of TE. Chromatin was eluted in elution buffer (100 mM ­NaHCO3, 1% SDS) and reversal of the Easy-Spray nLC 1000 chromatography system (Thermo- crosslinking was carried out for 5 h at 65 °C. The samples Fisher). The mass spectrometer was operated with 1 mass were then incubated with proteinase K for 1 h at 65 °C spectrometry (MS) spectrum followed by 10 MS/MS and RNase A for 30 min at 42 °C. DNA was purified with spectra in a data-dependent mode. The MS was acquired a PCR purification kit (Qiagen). with a resolution of 70,000 FWHM (full width at half maximum), a target of 1 × 106 ions, and a maximum scan Bioinformatics analysis of ChIP‑Seq data time of 120 ms. Using a relative collision energy of 27%, ChIP-seq libraries were pooled and sequenced paired- the MS/MS scans were acquired with a resolution of end 75 on a HiSeq 2500 (Illumina). We detected binding 17,500 FWHM, a target of 1 × 106 ions, and a maximum sites using SICER (window 600 bp, gap 200 bp) [67], and scan time of 120 ms. A dynamic exclusion time of 15 s detected differential binding sites (e.g., E1 vs E2) with Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 13 of 16 the same tool. Peaks were called in each sample inde- Coomassie blue and different sections were excised pendently. Increased and decreased peaks in the two for subsequent analysis. Mass spectrometry was then replicates were then combined to provide reliable peaks, performed using a nano-HPLC system (Easy-nLC II, using MSPC ­(10−4 and ­10−8 thresholds on p values defin- Thermo Fisher), coupled to the ESI source of an LTQ ing stringent and weak peaks, respectively [68]. Location Orbitrap Velos (Thermo Fisher), using conditions of the peaks was performed with ChipSeeker [69] on described previously [74]. MS data were acquired using mm10 annotation. For visualization purposes, ChIP sig- a data-dependent method. The data acquisition also nal has been normalized to input signal using the signal utilized dynamic exclusion, with an exclusion window extraction scaling (SES) method [70]. RSAT peak-motif of 10 ppm and exclusion duration of 10 s. MS events tool [71] was used detect over-represented motifs in the used 60,000 resolution FTMS scans, and MS/MS events detected peaks. For cluster generation, we merged rep- used ITMS scans, with a scan range of m/z 400–2000 in licates and computed the fold-change distribution along the MS scan. MS/MS data were analyzed using Mascot. the gene body, the TSS, and the TES. We then com- The data were compared to the Uniprot Mouse data- puted the fold-change distribution along each gene and base, using trypsin digestion with up to three missed clustered the genes into five groups using the k-means cleavages, a peptide tolerance of 5 ppm, and MS/MS method. Pathway enrichment analysis was performed tolerance of 0.3 Da. Acetylated N-termini and oxidation with WebGestalt [72]. of methionine were included as variable modifications. qPCR Supplementary information Each PCR was carried out in triplicate using SYBR Green Supplementary information accompanies this paper at https​://doi. org/10.1186/s1307​2-019-0298-1. PCR Master Mix (Applied Biosystems) and following the manufacturer’s instructions. Fluorescent signals were Additional file 1: Fig. S1. (A) Western blot analysis of the changes in the acquired by the Stratagene Mx3005P qPCR System (Agi- level of expression of MeCP2-E1 and E2 isoforms during brain develop‑ lent Technologies), and primer sequences upon request. ment (n = 6). (B) Summary of thermal unfolding results for NTD-MBD E1 and NTD-MBD E2 performed in different conditions (pH 7, pH8 and pH9 and at pH7 in presence of methylated dsDNA or unmethylated dsDNA). Co‑immunoprecipitation and mass spectrometry (C) Fluorescence recovery after photobleaching (FRAP) to determine One whole brain/immunoprecipitation was processed diffusion and binding kinetics of human wild type MeCP2-E1 and E2 as per [73] with slight modifications. MNase digestion in HEK293T. (1) Real time post-bleach recovery of GFP tagged isoforms (bleaching 1000 ms/frame for 2 frames). Red lightening indicates bleach was performed during 30 min at 37 °C with 150 units spots. (2) and (3) Comparative illustration of average half maximal recov‑ MNase (Worthington). After centrifugation, super- ery time and mobile fraction of wild type MeCP2-W1 and E2, respectively. natant was kept on ice and reaction was stopped by Mean and S.E. are shown. (4) FRAP recovery curves normalized to 1, show‑ ing chromocenter recovery in 151 frames. addition of EDTA (10 mM final concentration). The Additional file 2: Fig. S2. A (1) Western blot showing the levels of total remaining pellet was resuspended in digestion buffer MeCP2 and MeCP2 isoforms upon SH-SY5Y differentiation. (2) Expression with 2 mM ­CaCl2, and 50 ud MNase were added and changes of general (Sox2) and neuronal (Syt1, Syn1, Tubb3) differentia‑ incubated for 20 min at 37 °C. The reaction was stopped tion markers as detected by Reverse Transcriptase quantitative Poly‑ merase Chain Reaction (RT qPCR). Data represent mean ± S.E.M (n = 3). by addition of EDTA (10 mM final concentration), the (B) (1) Western blot showing untransfected HEK 293 cells and HEK 293 sample was centrifuged and supernatant kept. Superna- cells expressing 3xFlag-MeCP2 E1 and 3xFlag-MeCP2 E2. Immunoblots tant was pre-cleared for 1 h with 50 µL of Dynabeads were performed using the following antibodies: panMeCP2, MeCP2-E1 and MeCP2-E2 (Three upper panel). Equal loadings were assessed by Protein G magnetic beads (Thermo Fisher). Negative Coomasie blue gel staining (lower panel). (2) Western blot performed controls for the IP were obtained by blocking E1 and after immunoprecipitation of endogenous MeCP2-E1 and MeCP2-E2 E2 paratopes with E1 and E2 specific peptides [E1: from whole brain lysates. Normal rabbit IgG was used as negative control. Immunoprecipitated proteins were run in 10% gels to differentiate CAAAAPSGGG​G GG​EEER; E2: MVAGMLGLREEKC isoforms sizes and panMeCP2 antibody was used for staining. (3) Ponceau (New England Peptides)]. Antibodies were blocked staining showing equal loadings of wild type (WT) and Adrian Bird’s knock during 45 min by tumbling at room temperature with out (KO) mouse (60) brain samples. Samples were run on 12% SDS PAGE (66). (4) Western blots of Total MeCP2, MeCP2-E1 and MeCP2-E2 antibod‑ a fivefold excess of the specific peptide. 10 µg of each ies. H4 was used as a normalizer. antibody [E1, E1 blocked, E2, E2 blocked and normal Additional file 3: Fig. S3. (A) Summary table showing the amount of rabbit IgG (Cell Signaling Technologies)] were bound ChIP seq mapped reads per sample and number of reads post filtering. to 50 µl of magnetic beads according to the manu- (B) Integrative genomics viewer (IGV) (23) profiles of E1 and E2 ChIP-seq based on sequence reads of each sample. Data represent a 20 Mb region facturer’s instructions. Immunoprecipitations were of chromosome 17. Numbers in the right of each track represent track carried out overnight at 4 °C while tumbling. Anti- heights. body–protein complexes were washed and proteins Additional file 4: Fig. S4. MeCP2 isoforms functional networks as deter‑ eluted with 2x SDS buffer, boiled for 10 min and pro- mined by using STRINGv10 software. teins were run on SDS-PAGE. The gel was stained with Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 14 of 16 Abbreviations Author details 1 MeCP2: methyl CpG-binding protein 2; NTD: N-terminal domain; MBD: methyl- Department of Biochemistry and Microbiology, University of Victoria, Petch binding domain; ChIP: chromatin immunoprecipitation; dsDNA: double- Building 260, Victoria, BC V8W 3P6, Canada. 2 Unité de Mathématiques et stranded DNA; ITC: isothermal titration calorimetry; FRAP: fluorescence recov‑ Informatique Appliquées, Toulouse INRA, Auzeville, BP 52627, 31326 Casta‑ ery after photobleaching; IDP: intrinsically disordered protein; NM: N-terminal net‑Tolosan Cedex, France. 3 Centre de Physiopathologie de Toulouse Purpan, methionine; NME: N-terminal methionine excision; PTM: post-translational INSERM, UMR 1043, CNRS, UMR 5282, Université Toulouse III Paul Sabatier, Tou‑ modification; CHX: cycloheximide; WB: western blot; EGFP: enhanced green louse, France. 4 Institute of Biocomputation and Physics of Complex Systems fluorescent protein; DIV: days in vitro; KEGG: Kyoto encyclopedia of genes and (BIFI), Joint Units IQFR‑CSIC‑BIFI and GBsC‑CSC‑BIFI, Universidad de Zaragoza, genomes; TSS: transcriptional start site; TES: transcriptional end site; qPCR: 50018 Saragossa, Spain. 5 Instituto Aragonés de Ciencias de la Salud (IACS), quantitative polymerase chain reaction; MoRF: molecular recognition features; 50009 Saragossa, Spain. 6 Aragon Institute for Health Research (IIS Aragon), SDS-PAGE: sodium dodecyl sulfate–polyacrylamide gel electrophoresis; MAR: 50009 Saragossa, Spain. 7 Synaptic Plasticity Department, Max-Planck-Institute matrix attachment region; OR: olfactory receptor; TASR: taste receptor. for Brain Research, Frankfurt/Main, Germany. 8 School of Life Sciences, Brain Mind Institute, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Acknowledgements Switzerland. 9 Michael Smith Laboratories, University of British Columbia, Van‑ Not applicable. couver, BC V6T 1Z4, Canada. 10 Department of Microbiology and Immunology, University of British Columbia, Vancouver, BC V6T 1Z4, Canada. 11 University Authors’ contributions of Victoria-Genome British Columbia Proteomics Centre, Vancouver Island AMdP and JA designed and wrote the paper. AMdP, MSC, HM, CM, MEF, and Technology Park, #3101‑4464 Markham Street, Victoria, BC V8Z7X8, Canada. 12 TIS performed the biochemical/molecular biology experiments; RC-G, OA, and Molecular Neuropsychiatry & Development (MiND) Lab, Campbell Family AV-C designed and performed the calorimetry and spectroscopy experiments; Mental Health Research Institute, Centre for Addiction and Mental Health, MMM provided assistance with construction of the ChIP libraries; NIB and EVP Toronto, ON M5T 1R8, Canada. 13 Institute of Medical Science, University carried out the proteomic analyses; LK, MZ, and AC conducted the bioinfor‑ of Toronto, Toronto, ON M5S 1A8, Canada. 14 Department of Biochemistry matic analyses of the CHIP-seq data; STD and EMS performed the immuno‑ and Microbiology, University of Victoria, Room 270d, Petch Building, 3800 fluorescence; JVS-M, MZ, AV-C, JBV, MH, ME, and CM contributed to the writing Finnerty Road, Victoria, BC V8P 5C2, Canada. 15 Gerald Bronfman Department and discussion of specific sections of the manuscript. All authors read and of Oncology, Jewish General Hospital, Suite 720, 5100 de Maisonneuve Boule‑ approved the final manuscript. vard West, Montreal, QC H4A 3T2, Canada. 16 Proteomics Centre, Segal Cancer Centre, Lady Davis Institute, Jewish General Hospital, McGill University, 3755 Funding Côte‑Sainte‑Catherine Road, Montreal, QC H3T 1E2, Canada. 17 Department This work was supported by the Canadian Institutes of Health Research of Biochemistry and Molecular and Cell Biology, Universidad de Zaragoza, [MOP-130417 to JA and MOP-102758 to JBV]. Genome Canada and Genome 50009 Saragossa, Spain. 18 Fundación ARAID, Government of Aragon, British Columbia for financial support for the University of Victoria-Genome 50018 Saragossa, Spain. 19 Biomedical Reseach Networking Centre for Liver BC Proteomics Centre [204PRO and 214PRO], Leading Edge Endowment Fund and Digestive Diseases (CIBERehd), Madrid, Spain. 20 Michael Smith Genome (University of Victoria), the Segal McGill Chair in Molecular Oncology at McGill Sciences Centre, BC Cancer Agency, Vancouver, BC V5Z 1L3, Canada. 21 Cancer University, the Warren Y. Soper Charitable Trust and the Alvin Segal Family Epigenetics and Biology Program (PEBC), Bellvitge Biomedical Research Foundation to the Jewish General Hospital to CHB. Max Planck Society and Institute (IDIBELL), Avinguda Gran Vía de L’Hospitalet 199‑203. L’Hospitalet DFG CRC 1080; DFG CRC 902 to EMS. European Research Council under the de Llobregat, Barcelona, Catalonia, Spain. 22 Physiological Sciences Depart‑ European Union’s Horizon 2020 research and innovation program [743216 to ment, School of Medicine and Health Sciences, University of Barcelona (UB), EMS and 727264 (EPIPHARM) to ME]. Spanish Ministerio de Economia y Com‑ Catalonia, Spain. 23 Institució Catalana de Recerca I Estudis Avançats (ICREA), petitividad [BFU2013-47064-P and BFU2016-78232-P to AVC and SAF2014- Barcelona, Catalonia, Spain. 24 Department of Psychiatry, University of Toronto, 55000-R to ME]. Instituto de Salud Carlos III and co-funded by European Union Toronto, ON M5T 1R8, Canada. 25 Present Address: Department of Pathology (ERDF/ESF, “Investing in your future”) [PI15/00663, and Miguel Servet contract and Laboratory Medicine, Weill Cornell Medical College, Cornell University, CPII13/0017] to OA. Spanish Ministerio de Educacion Cultura y Deporte [FPU New York, NY 10065, USA. 13/3870] to RCG. Terry Fox Research Institute Program Project [TFRI #1039] and Canadian Cancer Society Research Institute [CCSRI #703489] to MH. European Received: 18 May 2019 Accepted: 22 August 2019 Community’s Seventh Framework Programme (FP7/2007–2013)/ERC [268626/ EPINORC project]; Integrated Project of Excellence [PIE13/00022] (ONCOPRO‑ FILE); CIBER 2016 CB16/12/00312 (CIBERONC); the Spanish Cancer Research Network [RD12/0036/0039]; the Cellex Foundation; Obra Social “La Caixa”; and the Health and Science Departments of the Catalan Government (Generalitat References de Catalunya) AGAUR [2014SGR633] to ME. ME is an ICREA Research Professor. 1. Lewis JD, Meehan RR, Henzel WJ, Maurer-Fogy I, Jeppesen P, Klein F, et al. French Association of Rett Syndrome funded research conducted by CEM, Purification, sequence, and cellular localization of a novel chromosomal HM and LK, genotoul bioinformatics platform Toulouse Midi-Pyrenees (Bioinfo protein that binds to methylated DNA. Cell. 1992;69(6):905–14. Genotoul) provided computing and storage resources. 2. Amir RE, Van den Veyver IB, Wan M, Tran CQ, Francke U, Zoghbi HY. Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl- Availability of data and materials CpG-binding protein 2. Nat Genet. 1999;23(2):185–8. Data sets generated and/or analyzed during the current study have been 3. Mnatzakanian GN, Lohi H, Munteanu I, Alfred SE, Yamada T, MacLeod PJ, deposited to the Gene Expression Omnibus Database (GEO) under the acces‑ et al. A previously unidentified MECP2 open reading frame defines a new sion number GSE130277. protein isoform relevant to Rett syndrome. Nat Genet. 2004;36(4):339–41. 4. Kriaucionis S, Bird A. The major form of MeCP2 has a novel N-terminus Ethics approval and consent to participate generated by alternative splicing. Nucleic Acids Res. 2004;32(5):1818–23. The experimental procedures were in accordance with all legislations defined 5. Sheikh TI, de Paz AM, Akhtar S, Ausio J, Vincent JB. MeCP2_E1 N-terminal by the European Union and approved by the local ethics committees from modifications affect its degradation rate and are disrupted by the Ala2Val Barcelona (UB-IDIBELL) and Toulouse (permit number: 04-U1043-DG-06). Rett mutation. Hum Mol Genet. 2017;26(21):4132–41. 6. Itoh M, Tahimic CG, Ide S, Otsuki A, Sasaoka T, Noguchi S, et al. Methyl Consent for publication CpG-binding protein isoform MeCP2_e2 is dispensable for Rett syndrome Not applicable. phenotypes but essential for embryo viability and placenta develop‑ ment. J Biol Chem. 2012;287(17):13859–67. Competing interests 7. Kerr B, Soto CJ, Saez M, Abrams A, Walz K, Young JI. Transgenic comple‑ The authors declare that they have no competing interests. mentation of MeCP2 deficiency: phenotypic rescue of Mecp2-null mice by isoform-specific transgenes. Eur J Hum Genet. 2012;20(1):69–76. Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 15 of 16 8. Kashi Y, King D, Soller M. Simple sequence repeats as a source of quantita‑ 32. Sephton CF, Cenik C, Kucukural A, Dammer EB, Cenik B, Han Y, et al. Identi‑ tive genetic variation. Trends Genet. 1997;13(2):74–8. fication of neuronal RNA targets of TDP-43-containing ribonucleoprotein 9. Lavoie H, Debeane F, Trinh QD, Turcotte JF, Corbeil-Girard LP, Dicaire complexes. J Biol Chem. 2011;286(2):1204–15. MJ, et al. Polymorphism, shared functions and convergent evolution of 33. Tsujimura K, Irie K, Nakashima H, Egashira Y, Fukao Y, Fujiwara M, et al. genes with sequences coding for polyalanine domains. Hum Mol Genet. miR-199a links MeCP2 with mTOR signaling and its dysregulation leads to 2003;12(22):2967–79. Rett syndrome phenotypes. Cell Rep. 2015;12(11):1887–901. 10. Dragich JM, Kim YH, Arnold AP, Schanen NC. Differential distribution of 34. Baker SA, Lombardi LM, Zoghbi HY. Karyopherin alpha 3 and karyopherin the MeCP2 splice variants in the postnatal mouse brain. J Comp Neurol. alpha 4 proteins mediate the nuclear import of methyl-CpG binding 2007;501(4):526–42. protein 2. J Biol Chem. 2015;290(37):22485–93. 11. Olson CO, Zachariah RM, Ezeonwuka CD, Liyanage VR, Rastegar M. Brain 35. da Huang W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: region-specific expression of MeCP2 isoforms correlates with DNA meth‑ paths toward the comprehensive functional analysis of large gene lists. ylation within Mecp2 regulatory elements. PLoS ONE. 2014;9(3):e90645. Nucleic Acids Res. 2009;37(1):1–13. 12. Quenard A, Yilmaz S, Fontaine H, Bienvenu T, Moncla A, des Portes V, et al. 36. Kalnins VI, Subrahmanyan L, Fedoroff S. Assembly of glial intermedi‑ Deleterious mutations in exon 1 of MECP2 in Rett syndrome. Eur J Med ate filament protein is initiated in the centriolar region. Brain Res. Genet. 2006;49(4):313–22. 1985;345(2):322–6. 13. Fichou Y, Nectoux J, Bahi-Buisson N, Rosas-Vargas H, Girard B, Chelly J, 37. Bergo A, Strollo M, Gai M, Barbiero I, Stefanelli G, Sertic S, et al. et al. The first missense mutation causing Rett syndrome specifically Methyl-CpG binding protein 2 (MeCP2) localizes at the centrosome affecting the MeCP2_e1 isoform. Neurogenetics. 2009;10(2):127–33. and is required for proper mitotic spindle organization. J Biol Chem. 14. Ghosh RP, Horowitz-Scherer RA, Nikitina T, Shlyakhtenko LS, Woodcock 2015;290(6):3223–37. CL. MeCP2 binds cooperatively to its substrate and competes with 38. Delepine C, Nectoux J, Bahi-Buisson N, Chelly J, Bienvenu T. MeCP2 histone H1 for chromatin binding sites. Mol Cell Biol. 2010;30:4656–70. deficiency is associated with impaired microtubule stability. FEBS Lett. 15. Claveria-Gimeno R, Lanuza PM, Morales-Chueca I, Jorge-Torres OC, Vega 2013;587(2):245–53. S, Abian O, et al. The intervening domain from MeCP2 enhances the DNA 39. Young JI, Hong EP, Castle JC, Crespo-Barreto J, Bowman AB, Rose MF, et al. affinity of the methyl binding domain and provides an independent DNA Regulation of RNA splicing by the methylation-dependent transcrip‑ interaction site. Sci Rep. 2017;7:41635. tional repressor methyl-CpG binding protein 2. Proc Natl Acad Sci USA. 16. Ausió J, de Paz A, Esteller M. MeCP2: the long trip from a chromatin 2005;102(49):17551–8. protein to neurological disorders. Trends Mol Med. 2014;20(9):487–98. 40. Maunakea AK, Chepelev I, Cui K, Zhao K. Intragenic DNA methylation 17. Suskiewicz MJ, Sussman JL, Silman I, Shaul Y. Context-dependent modulates alternative splicing by recruiting MeCP2 to promote exon resistance to proteolysis of intrinsically disordered proteins. Protein Sci. recognition. Cell Res. 2013;23(11):1256–69. 2011;20(8):1285–97. 41. Zhang Y, Ng HH, Erdjument-Bromage H, Tempst P, Bird A, Reinberg 18. Varshavsky A. The N-end rule pathway of protein degradation. Genes D. Analysis of the NuRD subunits reveals a histone deacetylase core Cells. 1997;2(1):13–28. complex and a connection with DNA methylation. Genes Dev. 19. Gonda DK, Bachmair A, Wunning I, Tobias JW, Lane WS, Varshavsky 1999;13(15):1924–35. A. Universality and structure of the N-end rule. J Biol Chem. 42. Szklarczyk D, Franceschini A, Wyder S, Forslund K, Heller D, Huerta-Cepas 1989;264(28):16700–12. J, et al. STRING v10: protein-protein interaction networks, integrated over 20. Martinez de Paz A, Vicente Sanchez-Mut J, Samitier-Marti M, Petazzi P, the tree of life. Nucleic Acids Res. 2015;43(Database issue):D447–52. Saez M, Szczesna K, et al. Circadian cycle-dependent MeCP2 and brain 43. Yang C, Wu J, de Heus C, Grigoriev I, Liv N, Yao Y, et al. EB1 and EB3 regu‑ chromatin changes. PLoS ONE. 2015;10(4):e0123693. late microtubule minus end organization and Golgi morphology. J Cell 21. Gabel HW, Kinde B, Stroud H, Gilbert CS, Harmin DA, Kastan NR, et al. Biol. 2017;216(10):3179–98. Disruption of DNA-methylation-dependent long gene repression in Rett 44. Tan AY, Riley TR, Coady T, Bussemaker HJ, Manley JL. TLS/FUS (translocated syndrome. Nature. 2015;522(7554):89–93. in liposarcoma/fused in sarcoma) regulates target gene transcription 22. Lagger S, Connelly JC, Schweikert G, Webb S, Selfridge J, Ramsahoye via single-stranded DNA response elements. Proc Natl Acad Sci USA. BH, et al. MeCP2 recognizes cytosine methylated tri-nucleotide and 2012;109(16):6030–5. di-nucleotide sequences to tune transcription in the mammalian brain. 45. Liu L, Pilch PF. PTRF/Cavin-1 promotes efficient ribosomal RNA transcrip‑ PLoS Genet. 2017;13(5):e1006793. tion in response to metabolic challenges. eLife. 2016;5:e17508. 23. Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz 46. Simon DN, Wilson KL. The nucleoskeleton as a genome-asso‑ G, et al. Integrative genomics viewer. Nat Biotechnol. 2011;29(1):24–6. ciated dynamic ‘network of networks’. Nat Rev Mol Cell Biol. 24. Bah A, Forman-Kay JD. Modulation of intrinsically disordered pro‑ 2011;12(11):695–708. tein function by post-translational modifications. J Biol Chem. 47. Davie JR. Histone modifications, chromatin structure, and the nuclear 2016;291(13):6696–705. matrix. J Cell Biochem. 1996;62(2):149–57. 25. Babu MM, van der Lee R, de Groot NS, Gsponer J. Intrinsically disordered 48. Jackson Da. In Encyclopedia of molecular cell biology and molecular proteins: regulation and disease. Curr Opin Struct Biol. 2011;21(3):432–40. medicine; 2006. 26. Hite KC, Kalashnikova AA, Hansen JC. Coil-to-helix transitions in intrinsi‑ 49. Weitzel JM, Buhrmester H, Stratling WH. Chicken MAR-binding protein cally disordered methyl CpG binding protein 2 and its isolated domains. ARBP is homologous to rat methyl-CpG-binding protein MeCP2. Mol Cell Protein Sci. 2009;21(4):531–8. Biol. 1997;17(9):5656–66. 27. Kaddoum L, Panayotis N, Mazarguil H, Giglia-Mari G, Roux JC, Joly E. 50. Cockerill PN, Garrard WT. Chromosomal loop anchorage of the kappa Isoform-specific anti-MeCP2 antibodies confirm that expression of the e1 immunoglobulin gene occurs next to the enhancer in a region contain‑ isoform strongly predominates in the brain. F1000Res. 2013;2:204. ing topoisomerase II sites. Cell. 1986;44(2):273–82. 28. Bourgeron T. From the genetic architecture to synaptic plasticity in 51. Saunders CJ, Minassian BE, Chow EW, Zhao W, Vincent JB. Novel exon autism spectrum disorder. Nat Rev Neurosci. 2015;16(9):551–63. 1 mutations in MECP2 implicate isoform MeCP2_e1 in classical Rett 29. Kruusvee V, Lyst MJ, Taylor C, Tarnauskaite Z, Bird AP, Cook AG. Structure syndrome. Am J Med Genet A. 2009;149A(5):1019–23. of the MeCP2-TBLR1 complex reveals a molecular basis for Rett syndrome 52. Skene PJ, Illingworth RS, Webb S, Kerr AR, James KD, Turner DJ, et al. and related disorders. Proc Natl Acad Sci USA. 2017;114(16):E3243–50. Neuronal MeCP2 is expressed at near histone-octamer levels and globally 30. Agarwal N, Hardt T, Brero A, Nowak D, Rothbauer U, Becker A, et al. MeCP2 alters the chromatin state. Mol Cell. 2010;37(4):457–68. interacts with HP1 and modulates its heterochromatin association during 53. Guo JU, Ma DK, Mo H, Ball MP, Jang MH, Bonaguidi MA, et al. Neuronal myogenic differentiation. Nucleic Acids Res. 2007;35(16):5402–8. activity modifies the DNA methylation landscape in the adult brain. Nat 31. Yasui DH, Gonzales ML, Aflatooni JO, Crary FK, Hu DJ, Gavino BJ, et al. Neurosci. 2011;14(10):1345–51. Mice with an isoform-ablating Mecp2 exon 1 mutation recapitulate the 54. Fasolino M, Zhou Z. The crucial role of DNA methylation and MeCP2 in neurologic deficits of Rett syndrome. Hum Mol Genet. 2014;23:2447–58. neuronal function. Genes (Basel). 2017;8(5):141. Martínez de Paz et al. Epigenetics & Chromatin (2019) 12:63 Page 16 of 16 55. Ferrer I, Garcia-Esparcia P, Carmona M, Carro E, Aronica E, Kovacs GG, et al. and phosphorylation and increases its chromatin binding affinity. Epige‑ Olfactory receptors in non-chemosensory organs: the nervous system in netics. 2017;12(11):934–44. health and disease. Front Aging Neurosci. 2016;8:163. 66. Laemmli UK. Cleavage of structural proteins during the assembly of the 56. Banday AR, Baumgartner M, Al Seesi S, Karunakaran DK, Venkatesh A, head of bacteriophage T4. Nature. 1970;227(259):680–5. Congdon S, et al. Replication-dependent histone genes are actively 67. Zang C, Schones DE, Zeng C, Cui K, Zhao K, Peng W. A clustering transcribed in differentiating and aging retinal neurons. Cell Cycle. approach for identification of enriched domains from histone modifica‑ 2014;13(16):2526–41. tion ChIP-Seq data. Bioinformatics (Oxford, England). 2009;25(15):1952–8. 57. Lyons SM, Cunningham CH, Welch JD, Groh B, Guo AY, Wei B, et al. A 68. Jalili V, Matteucci M, Masseroli M, Morelli MJ. Using combined evidence subset of replication-dependent histone mRNAs are expressed as poly‑ from replicates to evaluate ChIP-seq peaks. Bioinformatics (Oxford, Eng‑ adenylated RNAs in terminally differentiated tissues. Nucleic Acids Res. land). 2015;31(17):2761–9. 2016;44(19):9190–205. 69. Yu G, Wang LG, He QY. ChIPseeker: an R/Bioconductor package for ChIP 58. Nemeth A, Guibert S, Tiwari VK, Ohlsson R, Langst G. Epigenetic regula‑ peak annotation, comparison and visualization. Bioinformatics (Oxford, tion of TTF-I-mediated promoter-terminator interactions of rRNA genes. England). 2015;31(14):2382–3. EMBO J. 2008;27(8):1255–65. 70. Ramirez F, Dundar F, Diehl S, Gruning BA, Manke T. deepTools: a flex‑ 59. Singleton MK, Gonzales ML, Leung KN, Yasui DH, Schroeder DI, Dunaway ible platform for exploring deep-sequencing data. Nucleic Acids Res. K, et al. MeCP2 is required for global heterochromatic and nucleolar 2014;42(Web Server issue):W187–91. changes during activity-dependent neuronal maturation. Neurobiol Dis. 71. Thomas-Chollier M, Herrmann C, Defrance M, Sand O, Thieffry D, van 2011;43(1):190–200. Helden J. RSAT peak-motifs: motif analysis in full-size ChIP-seq datasets. 60. Guy J, Hendrich B, Holmes M, Martin JE, Bird A. A mouse Mecp2-null Nucleic Acids Res. 2012;40(4):e31. mutation causes neurological symptoms that mimic Rett syndrome. Nat 72. Wang J, Duncan D, Shi Z, Zhang B. WEB-based GEne SeT AnaLysis Toolkit Genet. 2001;27(3):322–6. (WebGestalt): update 2013. Nucleic Acids Res. 2013;41(Web Server 61. Bedogni F, Cobolli Gigli C, Pozzi D, Rossi RL, Scaramuzza L, Rossetti G, et al. issue):W77–83. Defects during Mecp2 null embryonic cortex development precede the 73. Draker R, Ng MK, Sarcinella E, Ignatchenko V, Kislinger T, Cheung P. A com‑ onset of overt neurological symptoms. Cereb Cortex. 2016;26(6):2517–29. bination of H2A.Z and H4 acetylation recruits Brd2 to chromatin during 62. Bonnaud EM, Szelechowski M, Betourne A, Foret C, Thouard A, transcriptional activation. PLoS Genet. 2012;8(11):e1003047. Gonzalez-Dunia D, et al. Borna disease virus phosphoprotein modu‑ 74. Brodie NI, Popov KI, Petrotchenko EV, Dokholyan NV, Borchers CH. lates epigenetic signaling in neurons to control viral replication. J Virol. Solving protein structures using short-distance cross-linking con‑ 2015;89(11):5996–6008. straints as a guide for discrete molecular dynamics simulations. Sci Adv. 63. Matsumura S, Persson LM, Wong L, Wilson AC. The latency-associated 2017;3(7):e1700479. nuclear antigen interacts with MeCP2 and nucleosomes through sepa‑ rate domains. J Virol. 2010;84(5):2318–30. 64. Forster JI, Koglsberger S, Trefois C, Boyd O, Baumuratov AS, Buck L, Publisher’s Note et al. Characterization of differentiated SH-SY5Y as neuronal screen‑ Springer Nature remains neutral with regard to jurisdictional claims in pub‑ ing model reveals increased oxidative vulnerability. J Biomol Screen. lished maps and institutional affiliations. 2016;21(5):496–509. 65. Good KV, Martinez de Paz A, Tyagi M, Cheema MS, Thambirajah AA, Gretz‑ inger TL, et al. Trichostatin A decreases the levels of MeCP2 expression Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions