UC San Diego

The nonsense-mediated RNA decay (NMD) pathway serves as a quality control mechanism and regulator of normal gene expression. Here, we report our analysis of two gene paralogs, UPF3A and UPF3B, which we found to have opposing roles in NMD. Previous gain- of-function studies indicated that UPF3A encodes a protein exhibiting little or no NMD activity. We were therefore perplexed as to what property allowed UPF3A to survive since it arose at the dawn of vertebrates, over 400 million years ago. Using loss-of-function approaches, we found that UPF3A is a potent NMD repressor both in vitro and in vivo. We generated Upf3a-null mice and found that global UPF3A knockout causes mouse embryonic lethality and conditional knockout in male germ cells leads to spermatogenic defects, particularly at the meiotic stage where UPF3A is highly expressed. We propose that UPF3A serves as a molecular rheostat to upregulate critical NMD target mRNAs at specific developmental time points, a model supported by RNAseq analysis. In contrast to UPF3A, UPF3B does not function in the testis; instead it is critical for neural development and human cognition. Mutations in UPF3B cause a form of X- linked intellectual disability, with patients often also suffering from autism or schizophrenia. To understand UPF3B function in vivo, we generated Upf3b-null mice and found that they display defects in fear-conditioned learning and pre-pulse inhibition (PPI), a measure of sensorimotor inhibition often deficient in individuals with schizophrenia. Consistent with these defects, cortical pyramidal neurons from Upf3b-null mice exhibit decreased spine density and RNAseq analysis identified transcripts encoding key neural regulators as targets of UPF3B. Our findings demonstrated that UPF3A and UPF3B have opposing functions in the NMD pathway and act in distinct developmental processes.

A major focus of my thesis work was to further investigate the role of UPF3B in neurogenesis using the olfactory system as a model. We and found (at the transcript level) fewer early OE cell types in the Upf3b-null mouse but similar numbers of mOSNs, the cells responsible for odorant detection. Pooled RNA sequencing of Upf3b-null and wild type mOSNs revealed decreased class-II olfactory receptor (olfr) expression, which may (at least partially) underlie a previously observed partial olfaction defect. When comparing pooled data with our single cell RNA sequencing (scRNAseq) results, we also found a significantly reduced number of cells expressing one or more olfrs. Closer examination of the scRNAseq data also indicated a role for UPF3B in immune response and presented the possibility that multiple OE cell types are bifunctional, playing a role in both detecting odorant molecules and responding to infection. This work was also the first to define the “translome” of pooled mOSNs, building a framework for what transcripts are more often translated in a healthy model and how that can change in a disease (in our case UPF3B-depleted) state.