Structural Basis of MeCP2 Distribution on Non-CpG Methylated and Hydroxymethylated DNA

Abstract

The Rett syndrome-associated methyl-CpG binding protein 2 (MeCP2) selectively binds methylated DNA to regulate transcription during the development of mature neurons. Like other members of the methyl-CpG binding domain (MBD) family, MeCP2 functions through the recognition of symmetrical 5-methylcytosines in CpG (mCG) dinucleotides. Advances in base level resolution epigenetic mapping techniques have revealed, however, that MeCP2 can bind asymmetrically methylated and hydroxymethylated CpA (h/mCA) dinucleotides and this alternative binding selectivity modifies gene expression in the developing mammalian brain. The structural determinants of binding to mCA and hydroxymethylated DNA have not been previously investigated. Here, we employ ITC and NMR spectroscopy to characterize MeCP2 binding to methylated and hydroxymethylated mCG and mCA DNA; examine the effects of Rett syndrome-associated missense mutations; and make comparisons to the related and evolutionarily most ancient protein, MBD2. These analyses reveal that MeCP2 binds mCA with high affinity in a strand-specific and orientation-dependent manner. In contrast, MBD2 does not show high affinity or methyl-specific binding to mCA. The Rett-associated missense mutations (T158M, R106W, and P101S) destabilize the MeCP2 MBD and disrupt recognition of mCG and mCA equally. Finally, hydroxymethylation of a high-affinity mCA site does not alter the binding properties; whereas, hemi-hydroxylation of the equivalent cytosine in an mCG site decreases affinity and specificity. Based on these findings, we suggest MeCP2 recognition of methylated/hydroxymethylated CpA dinucleotides functions as an epigenetic switch redistributing MeCP2 among mCG and mCA loci.

Quantitative modelling predicts the impact of DNA methylation on RNA polymerase II traffic

“…We assume that MeCP2 occupies methylated cytosines with probability times the probability of binding to a particular motif (SI Section 5). Binding probabilities for different motifs are based on known binding affinities (35) and relative binding strengths (15). To create simulated ChIP fragments, we assume that if a DNA fragment contains at least one MeCP2 bound to it, it will be present in the simulated ChIP-seq.…”

Neuronal non-CG methylation is an essential target for MeCP2 function

“…MeCP2 binding to mCG and mCAC sites has been characterised in vitro using multiple techniques and the co-crystal structures of both these interactions have been solved (Ho et al, 2008;Lagger et al, 2017;Lei et al, 2019;Lewis et al, 1992;Mellén et al, 2012;Sperlazza et al, 2017). A third potential MeCP2 binding motif, mCAT, has been proposed, though this interaction appears to be weaker and is barely detectable in some assays (Lagger et al, 2017;Liu et al, 2018).…”

“…This modification arises postnatally due to high levels of DNMT3A, which persist throughout adulthood (Feng et al, 2005;Stroud et al, 2017). The discovery that MeCP2 targets mCAC sites preferentially over all other forms of mCH (Gabel et al, 2015;Lagger et al, 2017;Liu et al, 2018;Sperlazza et al, 2017) raised the possibility that this DNA binding specificity contributes to neuronal function. Non-CG methylation almost doubles the number of available MeCP2 binding sites in neurons (Lagger et al, 2017;Lister et al, 2013;Mo et al, 2015) and the timing of increased CAC methylation coincides with the increase in neuronal MeCP2 protein levels during the first few weeks of life (Guo et al, 2014;Lister et al, 2013;Skene et al, 2010;Stroud et al, 2017).…”

Integrated analysis of human transcriptome data for Rett syndrome finds a network of involved genes

“…It influences alternative splicing, and miRNA expression (Chahrour et al, 2008;Ehrhart et al, 2016;Gonzales et al, 2012;Jordan et al, 2007;Liyanage & Rastegar, 2014). RTT is also a disorder which is well known for epigenetic implications because MECP2 not only binds on methylated DNA sites but is also involved in maintenance and modification of the methylation pattern of DNA (Bienvenu & Chelly, 2006;Sperlazza et al, 2017). For a previous study we created a comprehensive visualization of known MECP2 actions which can be found on WikiPathways (http://www.wikipathways.org/instance/WP3584) (Ehrhart et al, 2016).…”