Abstract
It is widely accepted that cell-type-specific gene expression is primarily achieved by cell-type-specific
presence of transcription factors (TFs), which bind to cognate DNA sequences. TFs then initiate changes in
higher-order chromatin structures by recruiting chromatin modifiers, including histone-modifying enzymes.
Unlike TFs, chromatin modifiers tend to be ubiquitously expressed. Among the plethora of chromatin
modifications, regulators of histone methylation are more frequently mutated in neurodevelopmental disorders
(NDDs) such as intellectual disabilities (IDs) and autism. Why is the brain so sensitive to dysregulation of
histone methylation? Is methyl-histone regulation in neurons unique? Investigation of a limited number of cell
types, cancer-cell lines, and embryonic stem cells has hampered our ability to address these questions.
The overarching goal of my research group is to contribute to the understanding of how methyl-histone
regulations underlie normal and pathological brain functions. Our focus is on the LSD1-PHF21A histone-
demethylation complex, which involves neuron-specific alternative splicing. LSD1 is a histone demethylase for
histone H3 lysine 4 (H3K4me). PHF21A was the first-discovered “zero reader,” which recognizes unmethylated
H3K4 (H3K4me0), the reaction product of canonical LSD1 (LSD1-c). Both LSD1 and PHF21A
haploinsufficiencies lead to NDDs, suggesting their importance in brain development. The neuronal LSD1
isoform (LSD1-n), which carries an alternative exon in its catalytic domain, was reported to have distinct
substrate specificity. However, the specific lysine(s) targeted by LSD1-n remains controversial. The goal of this
proposal is to determine the roles of the neuronal LSD1-PHF21A complex. Our preliminary study showed that
PHF21A also carries an alternative exon right upstream of the H3K4me0-recognizing PHD finger. This region
of PHF21A contains an AT-hook motif, which directly binds to DNA; we found that the alternative exon disrupts
the AT-hook, hence the DNA binding, but not H3K4me0 binding. These observations raise an exciting
possibility that the neuronal PHF21A isoform (PHF21A-n) recognizes nucleosomes in a distinct manner
compared to canonical PHF21A (PHF21A-c), thereby cooperating with LSD1-n to generate the neuronal
transcriptome for normal brain development.
We propose testing the hypothesis using multidisciplinary approaches encompassing cell biology,
biochemistry, and structural biology. The research plan was developed to provide both mechanistic insights
into the regulation of histone modifications and a better understanding of the pathogenesis of
neurodevelopment disorders, which could lead to novel approaches for brain-specific therapeutic targets.