Abstract
Viral infection leads to a robust cellular response whereby the infected cell produces hundreds of molecular regulators to combat infection. Currently, non-canonical components, e.g., long noncoding RNAs (lncRNAs) have been added to the repertoire of immune regulators involved in the antiviral program. Interestingly, studies utilizing next-generation sequencing technologies show that a subset of the >10,000 lncRNAs in the mammalian genome contain small open reading frames (smORFs) associated with active translation, i.e., many lncRNAs are not noncoding. Here, we use genome-wide high-throughput methods to identify potential micropeptides in smORF-containing lncRNAs involved in the immune response. Using influenza as a viral infection model, we performed RNA-seq and ribosome profiling to track expression and translation of putative lncRNAs that may encode for peptides and identify tens of potential candidates. Interestingly, many of these peptides are highly conserved at the protein level, strongly suggesting biological relevance and activity. By perusing publicly available data sets, four potential peptides of interest seem common to stress induction and/or are highly conserved; potential peptides from the MMP24-AS1, ZFAS1, RP11-622K12.1, and MIR22HG genes. Interestingly, using an antibody against the potential peptide encoded by MIR22HG RNA, we show that the peptide is stably expressed in the absence of infection, and upregulated in response to infection, corroborating the prediction of the ribosome profiling results. These data show the utility of perturbation approaches in identifying potentially relevant novel molecules encoded in the genome.