Histone 4 Lysine 20 Methylation: A Case for Neurodevelopmental Disease

Abstract

Neurogenesis is an elegantly coordinated developmental process that must maintain a careful balance of proliferation and differentiation programs to be compatible with life. Due to the fine-tuning required for these processes, epigenetic mechanisms (e.g., DNA methylation and histone modifications) are employed, in addition to changes in mRNA transcription, to regulate gene expression. The purpose of this review is to highlight what we currently know about histone 4 lysine 20 (H4K20) methylation and its role in the developing brain. Utilizing publicly-available RNA-Sequencing data and published literature, we highlight the versatility of H4K20 methyl modifications in mediating diverse cellular events from gene silencing/chromatin compaction to DNA double-stranded break repair. From large-scale human DNA sequencing studies, we further propose that the lysine methyltransferase gene, KMT5B (OMIM: 610881), may fit into a category of epigenetic modifier genes that are critical for typical neurodevelopment, such as EHMT1 and ARID1B, which are associated with Kleefstra syndrome (OMIM: 610253) and Coffin-Siris syndrome (OMIM: 135900), respectively. Based on our current knowledge of the H4K20 methyl modification, we discuss emerging themes and interesting questions on how this histone modification, and particularly KMT5B expression, might impact neurodevelopment along with current challenges and potential avenues for future research.

Keywords: H4K20; KMT5A; KMT5B; KMT5C; SUV420H; epigenetic; histone methylation; lysine-methylation; mutations; neurodevelopment.

Figure 1

Schematic representation of chromatin organization and H4K20 methylation. (A) Models for accessible (euchromatin) vs. condensed (heterochromatin) nucleosomal structures. (B) Two each of the histone proteins, H2A, H2B, H3, and H4, come together to form a histone octamer, which is bound by ~1.65 turns of DNA to form the nucleosome. The nucleosome bound by the H1 linker histone forms the chromatosome, which is further compacted to form chromosomes. (C) Histone protein 4 has a globular head domain and an N-terminal tail, of which the lysine (K) 20 residue can be mono, di, or tri-methylated.

Figure 2

Diagram of major H4K20-related proteins and their proposed role in DNA double stranded break (DSB) repair. (A) Lysine methyl transferase (KMT) 5A, KMT5B, and KMT5C are the three main enzymes involved in the sequential mono, di, and tri-methylation of H4K20. Other potential H4K20 methyl transferases include the NSD gene family. Lysine demethylase 4A (KDM4A/JMJD2A), lysine-specific histone demethylase 1 neuronal isoform (LSD1n), plant homeodomain finger 8 (PHF8), and PHF2 are demethylases known to act on H4K20. Malignant brain tumor domain-containing proteins (MBTD1, L3MBTL1), Fanconi Anemia Complementation Group D2 (FANCD2), and p53-bindng protein (TP53BP1, 53BP1) are examples of readers of H4K20 methyl marks. Protein TP53BP1 is a checkpoint protein that binds H4K20me1/me2 and other histone modifications that occur in response to DNA damage to facilitate DSB repair via non-homologous end joining (NHEJ). (B) The NHEJ pathway is the preferred method of repair during the G1-phase of the cell cycle [11,12]. DNA DSBs (red bolt) are recognized by the MRN complex (MRE11/RAD50/NBS1) binding, which leads to autophosphorylation of ATM kinase and phosphorylation of histone H2AX on S 139 (γ-H2AX). This creates a biding site for mediator of DNA damage checkpoint protein 1 (MDC1), which recruits the E3 ubiquitin ligases, RNF8 and RNF168, to establish polyubiquitination marks at the break sites [4]. These ubiquitination events recruit TP53BP1 and BRCA1 to the damage site. TP53BP1 binds H4K20me1, H4K20me2, and H2AK15 (ubiquitinated by RNF168) and mediates a cascade of events that prevents end resection, signaling NHEJ factors, including p53, to either repair the DNA or execute apoptosis/autophagy [13].

Figure 3

Schematic of the full-length KMT5B protein with its main annotated functional domain, the SET domain. Data points represent single nucleotide variants (SNVs) identified within NDD [73,77,79,80], cancer [81], or control (missense or in-frame deletions) [78] populations. Only SNVs falling within the coding sequence are depicted (i.e., intronic, splice-blocking, and synonymous mutations are excluded). While NDD and cancer variants span the full protein and cluster near the SET domain (likely disrupting gene function), there is a clear lack of mutations near the SET domain among control individuals, indicating that protein altering variants that affect the expression of the SET domain are more-likely associated with disease phenotypes.

Figure 4

Developmental transcriptome data for the human KMT genes from the Allen BrainSpan Atlas [82]. (A) Scatterplot shows average expression across all brains regions by age from conception (−40 weeks) to 18 years old (936 weeks); and (B) bar graph shows average expression by brain region for data points from all individuals before birth (i.e., week 0). As defined by the Allen BrainSpan Atlas: A1C: primary auditory cortex (core); AMY: amygdaloid complex; CB: cerebellum; CBC: cerebellar cortex; CGE: caudal ganglionic eminence; DFC: dorsolateral prefrontal cortex; DTH: dorsal thalamus; HIP: hippocampus (hippocampal formation); IPC: posteroventral (inferior) parietal cortex; ITC: inferolateral temporal cortex (area TEv, area 20); LGE: lateral ganglionic eminence; M1C: primary motor cortex (area M1, area 4); M1C-S1C: primary motor-sensory cortex (samples); MD: mediodorsal nucleus of thalamus; MFC: anterior (rostral) cingulate (medial prefrontal) cortex; MGE: medial ganglionic eminence; Ocx: occipital neocortex; OFC: orbital frontal cortex; PCx: parietal neocortex; S1C: primary somatosensory cortex (area S1, areas 3,1,2); STC: posterior (caudal) superior temporal cortex (area 22c); STR: striatum; TCx: temporal neocortex (n = 1); URL: upper (rostral) rhombic lip; V1C: primary visual cortex (striate cortex, area V1/17); and VFC: ventrolateral prefrontal cortex. Error bars represent the standard deviation of the mean.

Figure 5

Proposed dual model for KMT5B haploinsufficiency. We propose that KMT5B loss-of-function could result in (1) aberrant gene expression changes and/or (2) defective DSB repair, leading to defects in proliferation and differentiation of neural cells responsible for governing NDD phenotypes.