Ubiquitin operates in two fundamentally different modes: it can act as a degradative signal, in which polyubiquitin chains target proteins for destruction by the proteasome organelle, or as a regulatory modification, in which a single ubiquitin alters a substrate’s activity or interactions without destroying it. Over the past several years our lab has shown that both modes shape the chromatin landscape during neuronal maturation, placing our work at the intersection of ubiquitin biology and neuronal epigenetics. Below I describe the work that has led us to the questions that currently occupy our lab.
As a postdoctoral fellow, I identified the molecular genetic basis of a new recessive intellectual-disability disorder, Ferguson–Bonni neurodevelopmental syndrome (OMIM 619699), as loss-of-function mutations in ANAPC7, which encodes APC7, a core subunit of the Anaphase-Promoting Complex (APC). The APC is a ubiquitin E3 ligase long studied for clearing mitotic proteins so that dividing cells can exit mitosis. None of the other 15 core APC subunits had ever been associated with an inherited human disease. Indeed, why should non-dividing neurons be susceptible to a defect in the clearance of mitotic proteins? In Ferguson et al. (Molecular Cell, 2022), we showed that APC7 has a crucial, previously unsuspected role in neuronal differentiation. Specifically, it is required for efficient ubiquitination by the APC and we identified Ki-67 as a key neuronal substrate. When insufficiently cleared from post-mitotic neurons, Ki-67 shifts from the chromosome periphery to concentrate in constitutive heterochromatin (densely compacted genomic regions that that must be silenced for neurons to function). We speculated that, because Ki-67 is rich in intrinsically disordered repeat domains, its accumulation could alter the condensation of constitutive chromatin that occurs through phase separation (see below).
After starting my lab at UCSD, we realized that we had found not a single, especially important substrate but a general principle: the canonical mitotic regulators—Ki-67, the chromosome-passenger complex (CPC) and its kinase Aurora B, and topoisomerase 2α—are all normally ubiquitinated to be cleared from neurons as they exit mitosis and differentiate (Ledvin et al., Developmental Cell, 2023). When the APC loses activity, these proteins persist in phosphorylated, mitotic-like forms, with Aurora B accumulating in heterochromatin and driving aberrant phosphorylation of its substrate histone H3. The CPC had been shown to affect chromatin state if present during interphase, but its roles outside mitosis were still unknown. Proteins long thought to be purely mitotic are thus repurposed to regulate heterochromatin in post-mitotic neurons.
This study established that ubiquitination in the context of neurodevelopmental disorders merited scientific interest. It also produced a promising translational result: we administered two brain-penetrant Aurora B inhibitors subcutaneously and found that they fully corrected abnormal histone H3 phosphorylation in the mutant brain. Lowering Ki-67 genetically likewise restored heterochromatin function. In collaboration with biochemist and ubiquitin expert Nick Brown (UNC–Chapel Hill), we are now defining how this machinery recognizes its neuronal substrates.
The APC work drew our attention to Ki-67 itself—a protein universally used as a marker of cell proliferation, but whose function in post-mitotic neurons was utterly unknown. We generated the first brain-specific Ki-67 knockout mouse and, comparing these with Anapc7 mutant mice that have too much Ki-67, discovered that Ki-67 determines the size and number of the silenced heterochromatic particles within the nucleus. It does so without changing chromatin marks or which genes neurons express, suggesting Ki-67 operates through a physical rather than a chemical mechanism. Live-cell imaging revealed that Ki-67-rich structures behave like liquid droplets that merge and split, one hallmark of phase separation. Collaborating with Galia Debelouchina (UCSD Dept. of Chemistry and Biochemistry), we used purified proteins and biophysical methods, as well as molecular dynamics simulations, to show how Ki-67 drives this behavior (Abasi et al., Nature Communications, 2026). These findings recast a familiar proliferation marker as a physical organizer of the neuronal nucleus, and they provide one of the clearest examples of phase separation setting heterochromatin architecture in neurons, with direct consequences for genome integrity in the developing brain.
Polycomb complexes at active loci in the developing brain
The specific modification we study, H2AK119ub (or, more simply, H2Aub), is placed onto histone H2A by Polycomb Repressive Complex 1 (PRC1) and removed by the deubiquitinase BAP1. Loss of PRC1 function de-represses genes and has been studied primarily in the context of various cancers; little was known about what it might do in neurons, except that mutations in PRC1 components cause inherited neurodevelopmental disease. We therefore mapped H2Aub in the mouse brain at an early postnatal and adult timepoints, while developing new bioinformatic methods to deepen our analyses of our CUT&RUN data. To our surprise, we found that as neurons mature, this “silencing” mark is depleted from genes that are silenced and redirected toward active loci (specifically, active enhancers and promoters) (Parmar et al., PLoS Genetics, 2025). Mapping other PRC-dependent histone modifications showed that Polycomb Repressive Complex 2 likewise controls actively expressed genes in maturing neurons through the modification H3K27me1, which we found to be much more dynamic than the widely known PRC2-dependent, heterochromatin-associated modification H3K27me3. These findings overturn the textbook view that Polycomb complexes only act at repressive loci.
Chromatin control of inter-organ small-molecule handling
One offshoot of the PRC chromatin mapping project led my lab in a new direction in collaboration with Sanjay Nigam (UCSD Dept. of Pediatrics), who studies small molecule transporters. We decided together to explore how the roughly 1,000 genes that absorb, distribute, metabolize, and clear small molecules—and so mediate communication between organs—are regulated as the body matures. This question had never been examined at the chromatin level, but the approaches we had developed for the PLoS Genetics paper made this feasible. Mapping several chromatin marks in mouse kidney, liver, and brain at embryonic and adult stages, we found that the liver remodels chromatin state at these loci to the greatest degree during development, with coordinated activation in kidney and liver that tracked with changes in gene expression (Parmar et al., Epigenetics & Chromatin, 2025). The findings bear on how the body achieves metabolic autonomy after birth, on handling of exogenous drugs, and on diseases such as chronic kidney disease and liver failure. We are now focusing on transporters for specific disease-relevant molecules, including bile acids in the liver and urate, uremic toxins, and creatinine in the kidney.
BAP1 and the formation of chromatin identity in neurons
If H2Aub is redistributed during neurodevelopment, what happens when its removal goes wrong? Children born with mutations in BAP1, the deubiquitinase that erases H2Aub, develop Küry–Isidor syndrome, which causes severe global delay, behavioral problems, and seizures. Like PRC1/2, BAP1 had been studied almost exclusively in the context of cancer; its role in the brain had never been examined. Because complete loss of BAP1 is embryonic-lethal, we recently generated conditional mouse models removing BAP1 from excitatory neurons of the forebrain or cerebellum, and paired behavioral analyses with CUT&RUN, ATAC-seq, and transcriptomics. We found that BAP1 orchestrates a multifaceted genetic program essential for neuronal gene expression—a program distinct from its roles in other cell types. Remarkably, when Bap1 is lost, the distinction between active (euchromatin) and silenced (heterochromatin) chromatin breaks down and these two functional compartments begin to resemble one another. The genes most affected by Bap1 loss are involved in the formation and function of synapses; in collaboration with Kim Dore’s lab (UCSD Dept. of Neurosciences), we performed brain-wave recordings and single-cell electrophysiology to show that the mutant neurons are abnormally hyperexcitable. This work recasts BAP1 as a central regulator of the neuronal epigenome and gene regulation.
Our map of the gene pathways disrupted by BAP1 loss also relates to its other life as a tumor suppressor gene in cancer. BAP1 mutations occur in many malignancies but are most famously associated with uveal melanoma, a diagnosis I encounter frequently as an ophthalmic pathologist. Loss of BAP1 in uveal melanoma is strongly associated with metastasis and poor prognosis. We found that the gene expression changes that occur in Bap1 deficient neurons resemble those that occur in tumor cells that have lost Bap1, for instance, dysregulated expression of cell adhesion pathways and lineage-defining transcription factors. These shared signatures hint at a single unifying function for BAP1 that underlies both cancer’s invasive properties and the neurite outgrowth required for neuron-to-neuron connectivity in the brain.
We have established iPSC-derived neurons in the lab bearing patient BAP1 mutations. We are also examining mice lacking the PRC1 ubiquitin-ligase subunits Ring1A and Ring1B: where BAP1 erases H2Aub, these enzymes deposit it, and comparing the two will let us test how writing versus erasing the mark shapes gene expression, DNA methylation, chromatin folding, behavior, and neurophysiology.
My interest in inherited neurological disease began in graduate school with Miriam Meisler at the University of Michigan, where I studied FIG4, a phosphatase that, with PIKfyve and VAC14, controls the endolysosomal signaling lipid PI(3,5)P₂. FIG4 mutations underlie several neurodegenerative diseases, ranging from a severe syndrome in children to a fatal form of Charcot-Marie-Tooth disease (CMTJ4) and several forms of amyotrophic lateral sclerosis (ALS). We had shown that insufficient PI(3,5)P₂impairs autophagy in neurons and astrocytes and produces severe neurodegeneration (Ferguson et al., Human Molecular Genetics, 2009), and that neuronal Fig4 expression is both necessary and sufficient to prevent it (Ferguson et al., Human Molecular Genetics, 2012). We also found that Fig4 loss causes congenital CNS hypomyelination through a non-cell-autonomous effect on oligodendrocytes (Winters & Ferguson et al., Journal of Neuroscience, 2011), defined the destabilizing mechanism of the common CMT4J allele (Lenk et al., PLoS Genetics, 2011), and helped delineate the defining features of the human disease (Nicholson et al., Brain, 2011).
Yet despite decades of study, it has remained unclear how PI(3,5)P₂ deficiency actually harms the brain, because the lysosome is involved in so many cellular processes. Borrowing the proteomic strategy that had worked for the APC, I applied unbiased proteomics to the lysosome with Steve Gygi (Harvard) and Brandon Gassaway (BYU), quantifying more than 10,000 brain proteins to define the full set of PI(3,5)P₂-dependent lysosomal substrates—and found that lysosomal failure drives hyperactivation of innate immunity. With Andrew Mendiola (UCSD, Department of Pharmacology) we showed that microglia contribute to oxidative damage, and that these abnormalities begin in the first days of life, supporting the idea that disturbances during development can set the stage for later neurodegeneration. This first comprehensive account of how disrupted lysosomal lipids injure the brain appeared as Wong et al. (Neurobiology of Disease, 2026). Our data implicate hyperactivation of the cGAS–STING pathway as the central hub in astrocytes that drives neuroinflammation. Because cGAS–STING inhibitors are already in trials for common neurodegenerative diseases, they may warrant evaluation for FIG4-related disease.
The most recent line of research in my lab arose directly from my work as a neuropathologist: I am part of a multi-institutional study with our oculoplastics and ophthalmology colleagues on the pathology of orbital inflammatory disease. We recently used periostin immunohistochemistry to show that this matrix protein is strongly and specifically expressed in zones of pathologic fibrosis—but not in the inflammatory infiltrate or epithelium—in both nonspecific orbital inflammation and IgG4-related orbital disease (Zuraski et al., Orbit, 2025), suggesting that periostin is a tissue biomarker and a possible contributor to fibrosis in these conditions.
We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.