Targeted high throughput mutagenesis of the human spliceosome reveals its in vivo operating principles [screen]

A major impediment to studying the human spliceosome in vivo has been the inability to program point mutations in endogenous genes on a large scale. CRISPR-Cas9 technology now provides such opportunities. Due to its scalability and ability to introduce point mutations, we chose CRISPR-Cas9 base editing for a forward genetic screen of the spliceosome in an eHAP haploid human cell line background. We maintained eHAP FNLS cells as haploid so that we could subsequently assign genotype-phenotype relationships. We designed a single guide RNA (sgRNA) library targeting a hand-curated list of 153 human spliceosomal proteins which engage at various steps of the splicing cycle. After mutagenesis, we interrogated the spliceosome using the potent inhibitor pladienolide B (PB), which targets U2 snRNP by binding to a pocket between the SF3B1 and PHF5A subunits of the SF3b complex, preventing stabilization of the U2-branchpoint RNA duplex. Validation and genomic sequencing revealed resistance mutations in SF3B1 and PHF5A in residues adjacent to the compound binding pocket. We mapped hypersensitive mutants to U2 snRNP components, but also to factors that act as late as the second chemical step, after SF3b has dissociated. Strikingly, we obtained resistance mutants in SUGP1, a spliceosomal G-patch protein of unknown function that lacks orthologs in yeast and is also a newly proposed tumor suppressor whose loss underpins the splicing changes induced by cancer-associated SF3B1 mutations. Overall design: We generated a monoclonal stable cell line expressing FNLS, a cytosine base editor, in an eHAP haploid cell background. A CRISPR-Cas9 base editing screen was performed using a library of 43618 sgRNAs (including 100o non-targeting sgRNAs) targeting 153 proteins. After six days to allow for editing and selection of transduced cells, we then split the selected pool into treatment arms (control/DMSO vs. 2 nM pladienolid B (PB), which approximates its EC50).We cultured cells for two weeks while maintaining a representation of 500 cells per sgRNA. On days 0, 8 and 14 we isolated genomic DNA and amplified and sequenced the sgRNA inserts; we also did so for the input plasmid library. Libraries were sequenced with paired-end to identify both sgRNA sequence as well as a unique barcode linked to each sgRNA.