UC Irvine

The auditory brainstem is specialized to localize sounds with extreme temporal precision, which requires highly accurate development of the circuit that facilitates this function. In the embryonic chicken, this circuit is comprised of the axonal projection from the cochlear nucleus (nucleus magnocellularis; NM) to the monolayer coincidence detector nucleus laminaris (NL). Our lab has previously shown that formation of this circuit requires non-apoptotic activity of the apoptotic protease caspase-3. Pharmacological inhibition of caspase-3 activity in NM axons results in dual aberrant phenotypes: NM axons that overshoot NL, and NL cells that do not form a monolayer. How does caspase activity mediate auditory brainstem development? We characterized the proteome of auditory brainstems that had been treated with caspase-3 inhibitor or vehicle, screening for peptides that bore the biochemical signature of caspase proteolysis and that were present in control brainstems but absent in caspase-inhibited brainstems. The 288 proteins with at least one such peptide (i.e. likely caspase-3 substrates) were disproportionately proteins associated with extracellular vesicles (EVs; membrane-bound nanoparticles that carry cargo and information between cells). We purified EVs from chick auditory brainstems and found that their contents were enriched for caspase substrates. We next sought to determine how caspase activity occurs in the auditory brainstem without inducing apoptosis. One possibility was that caspases cleave proteins at different sites than during apoptosis. We analyzed the cleavage sites of auditory brainstem caspase substrates and found that they were cleaved at a novel cleavage site: IX(G/R)D instead of DEVD. The proteins with (G/R)D cleavage sites were enriched for proteins associated with the cytoskeleton. The human apoptotic proteomic database Degrabase also showed evidence that specific motifs are associated with cytoskeletal proteins, suggesting that non-apoptotic caspase activity may cleave substrates at sites enriched in cytoskeletal proteins to change cell morphology without killing the cell. Finally, based on the preponderance of EV proteins and RNA-binding proteins (RBPs) in caspase substrates, as well as the established enrichment of RBPs in EVs, we hypothesized that caspase activity controls the loading of RBPs (and therefore RNAs) into EVs. We sequenced the RNA of EVs derived from brainstems treated with caspase inhibitor or vehicle solution. We found that a highly abundant, long intergenic non-coding RNA (CREVASSE; Caspase-Regulated, Extracellular-Vesicle-Associated, Single-Stranded Effector) was more abundant in EVs from caspase-inhibited brainstems than EVs from control brainstems. We corrected some inconsistencies between the Ensembl annotation for this lncRNA and the annotation indicated by our sequencing data, and we predicted RNA binding partners for the resulting transcripts. Gene Set Enrichment Analysis of the ranked RNA binding partners showed that CREVASSE binds RNAs involved in development and differentiation, especially in the nervous system, auditory system, and brainstem. Together, these data suggest that auditory brainstem caspase activity regulates axon guidance by targeting cytoskeletal proteins, and it facilitates timely intercellular communication leading to neuronal differentiation (likely of NL cells) once caspase activity subsides.