MeCP2 is a microsatellite binding protein that protects CA repeats from nucleosome invasion

MeCP2 binds hydroxymethylated CA repeats Despites of decades of research on the Rett syndrome protein MeCP2, its function remains unclear. Ibrahim et al. show that MeCP2 is a hydroxymethylated cytosine-adenosine (CA) repeat-binding protein that modulates chromatin architecture at a distance from the transcription start site (see the Perspective by Zhou and Zoghbi). MeCP2 accumulates and spreads around modified CA repeats and competes for nucleosome occupancy. Loss of MeCP2 results in a widespread increase in nucleosome density inside lamina-associated domains and transcriptional dysregulation of genes enriched in CA repeats. These results shed light on the underlying molecular mechanism of Rett syndrome, a severe disease associated with mutations in MeCP2. Science, abd5581, this issue p. eabd5581; see also abj5027, p. 1390 The Rett syndrome protein MeCP2 is a DNA microsatellite CA repeat–binding protein that regulates chromatin architecture. INTRODUCTION Rett syndrome is a severe neurodevelopmental disorder that is mainly caused by mutations in the methyl-CpG-binding protein 2 gene (MeCP2). Initially, MeCP2 was identified as an essential brain protein that binds to methylated CpG (mCG) via its methyl-binding domain (MBD) and acts as transcriptional repressor. However, during early brain development, the postnatal accumulation of MeCP2 parallels the genome-wide high-level accumulation of hydroxymethylcytosine (hmC) and methylated CpA (mCA), suggesting that MeCP2 may also recognize and bind to DNA sequences that contain these modified nucleotides. The ability of MeCP2 to recognize both mCA and hmC as well as mCG has led to conflicting conclusions regarding its function in transcriptional regulation, because these cytosine modifications are associated with either repression (mCA and mCG) or activation (hmC) of transcription. The unambiguous identification of the MeCP2 target sequence(s) would help to clarify this issue. RATIONALE CA repeats (CAn) represent ~1% of the mouse genome and belong to the microsatellite family. They are widely distributed throughout the genome and have been shown to affect transcription of nearby genes. Our recent data reveal that CAn are methylated (mCAn) or hydroxymethylated (hmCAn) in various cell types. In a search for proteins that could specifically recognize and bind these CA repeats, we identified MeCP2 as a specific reader of CA repeats. We hypothesized that the methylation status of CAn is essential for the recognition and binding of MeCP2, possibly through recognition of the modified nucleotides in CA repeats with distinct affinities, relevant for its neuronal function in transcriptional regulation. RESULTS Here we show that within the MBD family, MeCP2 is the only protein that specifically recognizes and binds to CA repeats, with much stronger affinity than mCG and mCA. MeCP2 selectively recognizes CA repeat DNA in a strand-specific manner and requires at least five consecutive CA dinucleotides to optimally bind DNA. While MeCP2 can bind in vitro to modified and nonmodified CA repeats, it exhibits impressive selectivity toward hydroxymethylated CA repeats, which are modified by DNA (cytosine-5)-methyltransferase 3A. The modified cytosine, only when located within a CA repeat, serves as a nucleation point for both MeCP2 accumulation and spreading around the repeat, which, in turn, correlates with nucleosome exclusion. In addition, loss of MeCP2 results in widespread increase in nucleosome density within lamina-associated domains (LADs) and transcriptional dysregulation of CA repeat–enriched genes located outside LADs. We have also dissected the molecular basis of the MeCP2 hydroxymethylated CA repeat recognition by solving the crystal structure of MeCP2 in complex with hmCAn. The CA repeat creates a well-defined DNA shape, with a considerably modified geometry, including a widened major groove and negative roll parameters, located precisely at the modification site. We show that the molecular recognition of the hydroxymethylated CA repeat specifically occurs through Arg133, a key MeCP2 residue whose mutation causes Rett syndrome. CONCLUSION Our work identifies MeCP2 as a hydroxymethylated CA repeat DNA binding protein that targets the 5hmC-CA-rich sequence, which are specifically located on one strand. Our data provide insights into the origin of Rett syndrome at the molecular level and suggest that this neurodevelopmental disorder could be viewed as a chromatin disease, originating from the inability of mutant MeCP2 to bind and protect the CA repeats from nucleosome invasion. Our results open a previously unexplored area of research focused on understanding the role of specific protein binding to microsatellites and other repeats in neurological diseases of unknown etiology. The Rett syndrome protein MeCP2 is a microsatellite CA repeat binding protein regulating chromatin architecture. MeCP2 is a microsatellite binding protein that specifically recognizes hydroxymethylated CA repeats. Depletion of MeCP2 alters the chromatin organization of CA repeats and LADs and results in nucleosome accumulation on CA repeats and genome-wide transcriptional dysregulation. WT, wild-type; KO, knockout. The Rett syndrome protein MeCP2 was described as a methyl-CpG-binding protein, but its exact function remains unknown. Here we show that mouse MeCP2 is a microsatellite binding protein that specifically recognizes hydroxymethylated CA repeats. Depletion of MeCP2 alters chromatin organization of CA repeats and lamina-associated domains and results in nucleosome accumulation on CA repeats and genome-wide transcriptional dysregulation. The structure of MeCP2 in complex with a hydroxymethylated CA repeat reveals a characteristic DNA shape, with considerably modified geometry at the 5-hydroxymethylcytosine, which is recognized specifically by Arg133, a key residue whose mutation causes Rett syndrome. Our work identifies MeCP2 as a microsatellite DNA binding protein that targets the 5hmC-modified CA-rich strand and maintains genome regions nucleosome-free, suggesting a role for MeCP2 dysfunction in Rett syndrome.