A large-scale sORF screen identifies putative microproteins and provides insights into their interaction partners, localisation and function

The human genome contains thousands of potentially coding short open reading frames (sORFs). A growing set of microproteins translated from these sORFs are known to have important cellular functions. However, the majority remains uncharacterised. Thus, larger screens to find functional microproteins have become more vital. Here, we performed a high-throughput CRISPR/Cas9 knock-out screen with a customised library of 11,776 sORFs, curated from literature and databases to identify microproteins essential for cancer cell line growth. 16/17 tested candidates displayed a reproducible knockout phenotype. We selected our top six hits, consisting of 11 to 63 amino acids. Various of these candidates localised to distinct subcellular compartments and the majority showed specific interaction partners. Endogenous tagging demonstrated translation of an sORF in the CENPBD2P pseudogene that bears no resemblance to the CENPBD2P name-giving CENPB DNA binding domains. For two candidates, uORFs in the DSE and NUTF2 genes, the microprotein supplied in trans ameliorated the growth defect of the respective knock-out. RNA-seq analysis revealed however that gene expression changes in the knock-out could only partially be rescued. Overall, we identified various putative microproteins and a microprotein-producing pseudogene that might be involved in cancer cell growth, but also illustrate the limitations and caveats of sORF functional screening and characterisation.