ARTICLE https://doi.org/10.1038/s41467-019-08928-1 OPEN Systematic analysis reveals the prevalence and principles of bypassable gene essentiality Jun Li1, Hai-Tao Wang1,2, Wei-Tao Wang1,3, Xiao-Ran Zhang1, Fang Suo1, Jing-Yi Ren1, Ying Bi1, Ying-Xi Xue1, Wen Hu1, Meng-Qiu Dong 1 & Li-Lin Du 1,4 Gene essentiality is a variable phenotypic trait, but to what extent and how essential genes can become dispensable for viability remain unclear. Here, we investigate ‘bypass of 1234567890():,; essentiality (BOE)’—an underexplored type of digenic genetic interaction that renders essential genes dispensable. Through analyzing essential genes on one of the six chromo- some arms of the fission yeast Schizosaccharomyces pombe,wefind that, remarkably, as many as 27% of them can be converted to non-essential genes by BOE interactions. Using this dataset we identify three principles of essentiality bypass: bypassable essential genes tend to have lower importance, tend to exhibit differential essentiality between species, and tend to act with other bypassable genes. In addition, we delineate mechanisms under- lying bypassable essentiality, including the previously unappreciated mechanism of dormant redundancy between paralogs. The new insights gained on bypassable essentiality deepen our understanding of genotype-phenotype relationships and will facilitate drug development related to essential genes. 1 National Institute of Biological Sciences, 102206 Beijing, China. 2 Chinese Academy of Medical Sciences and Peking Union Medical College, 100730 Beijing, China. 3 College of Biological Sciences, China Agricultural University, 100193 Beijing, China. 4 Tsinghua Institute of Multidisciplinary Biomedical Research, Tsinghua University, 100084 Beijing, China. These authors contributed equally: Jun Li, Hai-Tao Wang, Wei-Tao Wang. Correspondence and requests for materials should be addressed to L.-L.D. (email: [email protected]) NATURE COMMUNICATIONS | (2019) 10:1002 | https://doi.org/10.1038/s41467-019-08928-1 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08928-1 ccording to whether gene deletion causes inviability or Bypassable essential genes have not been analyzed in an Anot, genes can be classified as either essential genes or unbiased and systematic manner—two previous attempts with non-essential genes. Essential genes are considered the limited breadth and depth only identified 4 bypassable essential foundation for life and gene essentiality is regarded as a key genes in E. coli14 and 5 bypassable essential genes in S. cerevi- criterion when selecting drug targets for combating pathogens siae15—and thus the true extent of bypassable essentiality remains and cancer cells1,2. However, essentiality is not a static gene unknown. Here, we perform a large-scale and unbiased BOE property. In recent years, it has been shown in yeast3 and human analysis in S. pombe, and find that bypassable essential genes are cell lines4–6 that gene essentiality can vary between genetic much more prevalent than previously realized. The results of this backgrounds. Thus, to fully grasp the underpinnings of life and to systematic BOE analysis enable us to identify the general prin- improve drug target selection, it is imperative to understand ciples as well as functional and evolutionary implications of which genes can undergo essentiality change and how essentiality essentiality bypass. Furthermore, we demonstrate that BOE ana- change can happen. In particular, it is of both fundamental and lysis is especially conducive to inferring functional relationships practical values to investigate which and how essential genes in between genes. a well-defined genetic background can lose essentiality and be converted to non-essential genes. Essentiality loss can happen through high-frequency sponta- Results neous chromosome copy number changes7. Taking advantage of Systematic analysis of BOE interactions. We developed an this phenomenon, it has been shown in a systematic study that efficient BOE analysis procedure that uses “query strains” lacking in Saccharomyces cerevisiae, 9% of the essential genes can be the chromosomal copy of an essential gene (“query gene”) but rendered dispensable by spontaneous genetic events occurring harboring a counter-selectable episomal plasmid containing at ≥ 1% frequencies, which are most probably events altering that gene (Supplementary Figs. 1a-c). To enable the identification chromosome copy numbers8. These essential genes are termed of a broad range of suppressor types (Fig. 1a), we induced “evolvable essential genes”. For most of the evolvable essential genetic changes using the chemical mutagen methylnitroni- genes, the underlying genetic basis of the essentiality loss, namely, trosoguanidine (MNNG), the transposon piggyBac (PB), and an how many and which genes on the chromosome with an altered overexpression plasmid library, and termed the BOE suppressors copy number mediate the essentiality loss, remains unknown. thus obtained C-BOE, T-BOE, and OP-BOE suppressors, Furthermore, null mutations and missense mutations, which respectively (Supplementary Figs. 1b, d). T-BOE and OP-BOE happen spontaneously at frequencies much lower than 1%, can- suppressors were exhaustively identified; for query genes having not be surveyed by this approach. But such mutations are able to only C-BOE suppressors, we ensured that at least one C-BOE suppress the inviability phenotype of essential gene deletions and suppressor was identified. We experimentally verified all of the thus drive essentiality loss. Prominent examples from the litera- candidate suppressors by independently generating genetic ture include suppression of SEC14 gene deletion via “bypass alterations identical or similar to the ones found in the screen hits Sec14” mutations in S. cerevisiae9, suppression of the loss of (Supplementary Fig. 1e). MEC1 or RAD53 by deletion of SML1 in S. cerevisiae10, and Aiming to unbiasedly survey the bypassability of essential suppression of cdc25 deletion by either a wee1 deletion or a cdc2- genes, we targeted the essential genes located on the left arm of 3w mutation in the fission yeast Schizosaccharomyces pombe11,12. chromosome II (chrII-L) and were able to obtain BOE analysis To our knowledge, this type of digenic suppression interaction results on 142 (89%) of them (Fig. 1b and Supplementary Data 1); lacks a name. Because “bypass suppressor” is a term widely used these are a representative set of essential genes (Supplementary in the genetics literature to refer to the extragenic suppressor of a Fig. 1 f). The 17 (11%) remaining genes include the telomere deletion mutant13, we name this type of digenic interaction protection gene stn1, whose efficient bypass by chromosome “bypass of essentiality” (BOE). For simplicity, we refer to essential circularization hindered the search for BOE suppressors16.We genes that can be rendered non-essential by monogenic sup- failed to construct query strains for the other 16 genes, possibly pressors as “bypassable essential genes”. Compared to the evol- because cells cannot tolerate their altered expression levels when vable gene approach, using experimental means to identify these genes are expressed from plasmids. bypassable essential genes and their BOE suppressors can provide In total, we identified and verified 263 BOE interactions that a more comprehensive coverage of genes with variable essentiality render 38 (27%) of the 142 essential genes dispensable, including and can more directly reveal the exact genetic causes of essenti- all three previously known bypassable essential genes on chrII-L: ality loss. res1, slx8, and rhb117–19 (Fig. 2 and Supplementary Data 2). These abBypass of essentiality Mbp WT background Null mutant of gene B 0 A A 38 (27%) B 0.5 Bypassable Inviable Viable 104 (73%) Non-bypassable 1.0 Missense mutant of gene C Overexpression of gene D A A 142 analyzed essential C* D 1.5 genes on chrII-L Viable Viable I II III Analyzed (142) Chr. Not analyzed (17) Fig. 1 Systematic BOE analysis. a Three types of monogenic changes that can cause bypass of essentiality (BOE). In the wild-type (WT) background, gene A is essential. BOE may occur when another gene is deleted, mutated, or overexpressed. b Systematic BOE analysis for 89% (142/159) of the essential genes on chrII-L in fission yeast led to the identification of BOE suppressors for 27% (38/142) of them; these 38 genes were deemed bypassable. The other 104 genes were deemed non-bypassable 2 NATURE COMMUNICATIONS | (2019) 10:1002 | https://doi.org/10.1038/s41467-019-08928-1 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08928-1 ARTICLE ptl1 SPAC1A6.03c cyt1 dcr1 cyr1 SPCC18.17c vps1302 SPBC215.01 rpl4102 nte1 cox10 cbp6 hmt2 atp3 txc1 SPAC26H5.07c coq7 bgs4 ppr1 mdm12 mga2 sur2 pgr1 cbp3 mrp21mrps28 mrps17 slm5 mef1 SPBC83.06c mrpl8 pex11 mmm1 sip1 ppr8 sls1 SPBC1703.09 gpi3 vgl1 ppr4 bcs1 sac11 lac1 fzo1 pma2 SPAP27G11.14c mts4 ubp15 SPAC57A10.08c rpt3 ubr11 rpn12 rpt6 amk2 uba2 ufd1 stn1 alp13 meu31 cbs2 ssp2 sgf29 SPBC17D1.05 swd1 ssr2 luc7 taf73 usp101 pst1 swd2 ulp2 cbh1 nts1 yth1 nrp1 nup189 clr6 spf1 nup132 slx8 ulp1 cwf15 ubp8 cph2 pst3 ngg1 jmj3 vac8 zas1 cip2 dam1 cip1 png1 set1 btb2pmt3 snR101 tup12 ada2 rhp16 msi2 pst2 not3 sgf73 ure6 swd3 rrp2 not2 gyp10 pli1 SPAC18G6.13 meu23 hrp3 klf1 gcn5 cdc10 rps1701 clr5 bdf1 rps602 pab2 rpb2 rnf10 apl2 apl4 yox1 rep2 nab2 nrm1 sty1 res2 rep1 iws1 asp1 fap7 hit1 cdk9 fkh2 iec1 wdr74 red5 SPBC27B12.12c cnp20 res1 whi5 SPBC25B2.08 rmn1 rnc1 pop2 rsa1 red1 imp1 cds1 srp40 apm3 mnr2 spt5 tps3 spc7 ppn1 cdc18 mak5 SPBC13E7.03c nup124 T-BOE (verified using deletion) Query gene T-BOE (verified using insertion) sfc1 mdm31 itr2 Suppressor gene C-BOE OP-BOE lip5 cog8 erg27 pdf1 RNA metabolism sfc9 mdm35 mug89 Ribosome biogenesis atd1 sly1 erg26 sec24 sfc4 cbf11 ivn1 Mitochondrion-related Transcription / Chromatin organization tsr2 nup107 rpl2501 ubp10 pir2 its8 rhb1 Protein catabolism / Protein modification Lipid / Membrane organization / Transmembrane transport Unknown / Others spi1 mex67 rpl2502 spp42 dis3 omh5 tor2 Fig. 2 BOE interactions that bypass the 38 chrII-L essential genes.