Nxf1 encodes a nuclear export factor, but names do not limit molecular function. Initiation, elongation, processing, nuclear export, translation, and turnover of mRNA are generally coupled, but allow multiple for regulating gene expression at nominally distinct stages. We discovered that a natural variant of Nxf1 in wild mice genetically suppresses mutations caused by integration of an endogenous retrovirus–by increasing the amount and proportion of correctly processed mRNA expressed from the mutated gene, at the expense of mutant-specific, abnormal RNAs (Floyd 2003, Concepcion 2009). This suggests that Nxf1 influences alternative processing choices for at least some pre-mRNAs, in addition to its better-known recruitment to export complexes after processing. Genome editing in mice showed that a single amino acid substitution in the TAP_C domain is sufficient for this effect (Concepcion 2015). Current work aims to understand how Nxf1 variants alters gene expression and potentially remodel its protein interaction network. This work is funded by a grant from the National Institute for General Medical Sciences (NIGMS).

Zfp423 encodes a protein with 30 C2H2 zinc finger domains. Zfp423 protein interacts with both lineage and signal-dependent transcription factors to repress or activate target genes. Mice that lack Zfp423 have midline defects in brain development, similar in kind to several human disorders (Alcaraz, 2006). The brain abnormalities are most apparent in cerebellum. We showed that variation in the phenotype results from both modifier genes and non-genetic causes (Alcaraz, 2011). Consistent with the mouse phenotype, the human homolog ZNF423 is mutated in a some patients with Joubert Syndrome related disorders (Chaki, 2012). Zfp423 has many likely sites of action and is expressed in a dynamic pattern during development. One important site is cerebellar granule cell precursors, where Zfp423 is required for proliferative response to SHH through the primary cilium (Hong, 2016). We have made a large number of new mouse models to determine which Zfp423 phenotypes are cell autonomous, to model pathogenicity of genetic variants identified in patients, and to test for epistatic interactions between Zfp423 and Tulp3, one of its transcriptional targets. This work is funded by the National Institute for Neurolgical Disorders and Stroke (NINDS).

Zfp804a encodes a single zinc finger. Genome-wide studies by others identified human ZNF804A as one gene associated with schizophrenia and bipolar disorder. While predicted to be a transcription factor, little is known about its functional properties, interaction partners, or effects on brain development. We have created several custom mouse strains to allow visualization and purification of native Zfp804a and to test its requirements for normal development and behavior. This project was initiated with funds from the National Institute for Mental Health (NIMH).

Nmf9 encodes a protein of uncertain function. Mice that lack this gene have defects in vestibular function, fear learning, and precision of circadian behaviors (Zhang 2015). The gene is unusually conserved. The core protein sequence, give or take a domain, is identifiable in unicellular organisms that are sister groups to animals (choanozoa) but missing from nearly all fungi. Invertebrate species from every major phylum have a single copy except for tunicates, which have none. Jawed vertebrates have two paralogous copies, with the derived version having lost an RA domain that is constant among invertebrates. Mammals have only the derived, deleted copy. Mutations introduced by genome editing in mice, zebrafish, and fruit flies all produce defects in movement and balance, suggesting a level of functional conservation. However, while mice lacking Nmf9 have a comparatively mild phenotype, flies that lack the gene have severe defects and die as pupae or young adults. This project was developed by a graduate student funded by an institutional training grant from the National Institute for General Medical Sciences (NIGMS).