The dazzling complexity of multicellular organisms is generated from just a single genome, which must be interpreted differently in each cell type to produce a unique view of this common material. Epigenetic modifications, which influence the structure and functional properties of the genome but do not change its primary sequence, are essential to this act of interpretation during development. In particular, neurodevelopment is a time of dramatic reconfiguration of epigenetic patterning, including major changes to the DNA modification landscape. The nuclear base 5-hydroxymethylcytosine (5hmC), an oxidized derivative of the classic DNA modification 5-methylcytosine (5mC), is highly enriched within the genomes of neurons relative to other cell types and is predominantly localized to active and accessible genomic regions. However, it is unclear how 5hmC patterns are established during neurogenesis and how the modified base contributes to neuronal gene regulation.
Here, I provide insight into neuronal 5hmC biology by studying the patterning and functional roles of the modification during olfactory sensory neuron (OSN) development. Using a genome-wide approach to map 5mC and 5hmC distributions in defined developmental stages of the olfactory epithelium - from multipotent stem cells to neuronal progenitors to mature olfactory sensory neurons - I found that 5hmC levels increase during neuronal differentiation, particularly within the bodies of actively transcribed genes. Altering 5hmC patterning in mOSNs by overexpressing an enzyme that regulates its production demonstrated that 5hmC within the gene body modulates transcriptional output. Moreover, this manipulation disrupted the expression of olfactory receptors and axon guidance molecules, suggesting that 5hmC patterning is critical for the formation of neuronal identity.
In the final section, I examine connections between de novo DNA methylation and 5hmC patterning in olfactory sensory neurons. I show that Dnmt3a, a de novo DNA methyltransferase implicated in neuronal development and function, is necessary to generate elevated levels of 5hmC in OSNs. Moreover, Dnmt3a-dependent 5mC and 5hmC patterning is found within accessible regions of the genome, and its loss globally disrupts gene expression. Lastly, I provide evidence that neuronal DNA modifications are critical for neuronal function by demonstrating that odorant-induced gene activation is impaired in the absence of Dnmt3a.