New approaches to determine the function of neuronal epigenetic marks

Abstract Single cell profiling has revealed a striking cellular diversity and complexity in the mammalian nervous system. Neurons develop their unique features by intrinsic factors, but also must adapt to local environmental features to produce functional circuits. This raises the question: how do neurons maintain or return to stable cellular states in the context of variable inputs? Evidence suggests that this process requires transcriptional regulation by DNA methylation and the methyl-binding protein MeCP2. The levels of 5-methylcytosine (mC) and its oxidized form, 5-hydroxymethylcytosine (hmC), dramatically change in the developing mammalian brain in response to intrinsic cues and environmental input, and these marks show unique genomic patterns in different cell-types. MeCP2 modulates the effects of mC and hmC in neurons and is commonly mutated in Rett syndrome. Further, mutations in the TET enzymes responsible for converting mC to hmC have been recently associated with neurological disorders. Meanwhile, developmental studies indicate that neuron-specific non-CG methylation (mCH), as well as hmC and MeCP2 build up during the postnatal period, when neurons integrate into circuits and complete their final maturation into specific functional subtypes. This leads to the intriguing hypothesis that mC, hmC, and MeCP2 are critical for the establishment and maintenance of diverse, specialized neuronal subtypes within microcircuits. Recent results from us and others indicate that mC and hmC play distinct roles in neuron-specific gene regulation and function depending on sequence context (CG vs CH) and the genomic feature (promoter, enhancer, gene body). However, a critical barrier to testing our hypothesis is the lack of available methods to simultaneously analyze mC and hmC across different genomic contexts on an individual allele. The primary limitation is that the bisulfite chemistry that underlies all high-resolution mC profiling methods shears DNA to under 300 bp. Here we propose to take advantage of the long-read capabilities of nanopore sequencers to develop a method to accurately simultaneously profile mC and hmC in both CG and non-CG contexts across gene promoters and bodies (Aim 1). This will allow us to address how mC and hmC coordinate in different contexts to affect MeCP2 binding and transcriptional programs in Granule and Pukinje neuronal subtypes in the cerebellum (Aim 2). More specifically we will address how differences in mC and hmC profiles across cell types are read out by MeCP2 to maintain differential subtype-specific transcriptomes. We will further dissect how allele specific patterns of this methylation, which can only be decoded using long read sequencing, contribute to neuronal gene regulation. In the future, our approach, which can be easily implemented into standard nanopore workflows, will enable complete sequencing and methylation analysis of clinical or research samples in a single assay. This could lead to a profound increase in the number of available methylomes and our ability to interpret and exploit methylation information for clinical utility.