Multiplex enCas12a screens show functional buffering by paralogs is systematically absent from genome-wide CRISPR/Cas9 knockout screens

Abstract

Major efforts on pooled library CRISPR knockout screening across hundreds of cell lines have identified genes whose disruption leads to fitness defects, a critical step in identifying candidate cancer targets. However, the number of essential genes detected from these monogenic knockout screens are very low compared to the number of constitutively expressed genes in a cell, raising 5 the question of why there are so few essential genes. Through a systematic analysis of screen data in cancer cell lines generated by the Cancer Dependency Map, we observed that half of all constitutively-expressed genes are never hits in any CRISPR screen, and that these neveressentials are highly enriched for paralogs. We investigated paralog buffering through systematic dual-gene CRISPR knockout screening by testing algorithmically defined ~400 candidate paralog 10 pairs with the enCas12a multiplex knockout system in three cell lines. We observed 24 synthetic lethal paralog pairs which have escaped detection by monogenic knockout screens at stringent thresholds. Nineteen of 24 (79%) synthetic lethal interactions were present in at least two out of three cell lines and 14 of 24 (58%) were present in all three cell lines tested, including alternate subunits of stable protein complexes as well as functionally redundant enzymes. Together these 15 observations strongly suggest that paralogs represent a targetable set of genetic dependencies that are systematically under-represented among cell-essential genes due to genetic buffering in monogenic CRISPR-based mammalian functional genomics approaches. 20 through immense monogenic screening efforts, multiple groups revealed lists of ~2000 highly concordant human essential genes, and comparison of CRISPR technology to orthogonal techniques such as random insertion of gene traps also showed consistent results (Blomen et al., 2015; Hart et al., 2015; Wang et al., 2015).

Discovery of novel putative tumor suppressors from CRISPR screens reveals rewired lipid metabolism in AML cells

“…Importantly, by comparing the mean log fold change of query gene knockouts in the “A” position vs . the same genes in the “B” position of the dual knockout vector, we find no positional bias in the multiplex knockout constructs ( Figure 4d ), consistent with our previous findings 23,34 . Single knockout fitness measurements effectively segregated known essential genes from nonessentials, confirming the efficacy of the primary screens ( Supplementary Figure 6 ).…”

“…Digenic perturbations in human cells, a more faithful replication of the yeast approach, are possible with Cas9 and its variants, but library construction, sequencing, and positional biases can be problematic 16,28–34 . Recently, we showed that an engineered variant of the Cas12a endonuclease, enCas12a 35 , could efficiently perform multiplex gene knockouts 34 , and we demonstrated its effectiveness in assaying synthetic lethality between targeted paralogs 23 . These developments in principle enable researchers to measure how biological networks vary across backgrounds, a powerful approach for deciphering complex biology 24,36,37 .…”

“…Data-driven analysis of correlation networks reveals clusters of functionally related genes whose emergent essentiality in specific cell backgrounds is often unexplained by the underlying lineage or mutational landscape 21 . Interestingly, in a recent study of paralogs whose functional buffering renders them systematically invisible to monogenic CRISPR knockout screens 22,23 , it was shown that the majority of context-dependent essential genes are constitutively expressed in cell lines 23 . Collectively these observations suggest that there is much unexplained variation in the genetic architecture, and emergent vulnerability, of tumor cells.…”

Combinatorial CRISPR screening reveals functional buffering in autophagy

“…Functional buffering by paralogs has recently been shown to be largely absent from single gRNA CRISPR screens 63 , suggesting paralogs to contribute to network redundancy. Indeed, we identify paralogs as functional buffers in autophagy acting in autophagosome assembly (ATG2A/B), phosphatidylinositol phosphorylation (PIK3CA/PI4K2A), and phagosome-lysosome fusion (RAB7A/5A and RAB7B/5A).…”

Chromosomal arrangement of synthetic lethal gene pairs: repulsion or attraction?

“…Therefore, we hypothesized that repulsion of synthetic lethal genes would not be observed on eukaryotic chromosomes. Furthermore, eukaryotic genomes are strongly influenced by gene duplication events (Friedman & Hughes 2001; Qian & Zhang 2014), which generate closely located paralog gene pairs accounting potentially for a significant proportion of the synthetic lethal genetic interactions (De Kegel & Ryan 2019; Dede et al 2020). Moreover, studies on budding yeast as a eukaryotic cell have indicated that gene pairs showing negative genetic interactions tend to participate in closely related biological processes (Kelley & Ideker 2005; Costanzo et al 2010; Bellay et al 2011; Costanzo et al 2016) and hence might be co-expressed.…”