Identification of 15 new bypassable essential genes of fission yeast

Every organism has a different set of genes essential for its viability. This indicates that an organism can become tolerant to the loss of an essential gene under certain circumstances during evolution, via the manifestation of ‘masked’ alternative mechanisms. In our quest to systematically uncover masked mechanisms in eukaryotic cells, we developed an extragenic suppressor screening method using haploid spores deleted of an essential gene in the fission yeast Schizosaccharomyces pombe. We screened for the ‘bypass’ suppressors of lethality of 92 randomly selected genes that are essential for viability in standard laboratory culture conditions. Remarkably, extragenic mutations bypassed the essentiality of as many as 20 genes (22%), 15 of which have not been previously reported. Half of the bypass-suppressible genes were involved in mitochondria function; we also identified multiple genes regulating RNA processing. 18 suppressible genes were conserved in the budding yeast Saccharomyces cerevisiae, but 13 of them were non-essential in that species. These trends are consistent with a recent independent bypass-of-essentiality (BOE) screening of 142 fission yeast genes conducted with more elaborate methodology (Li et al., 2019). Thus, our study reinforces the emerging view that BOE is not a rare event and that each organism may be endowed with secondary or backup mechanisms that can substitute for primary mechanisms in various biological processes. Furthermore, the robustness of our simple spore-based methodology paves the way for genome-scale BOE screening.


34
A recent genome-wide study using S. cerevisiae gave an insight into the 'evolvability' of 35 essential cellular processes (Liu et al., 2015), which can be also termed 'bypass-of-essentiality' 36 (BOE) (Li et al., 2019). The study surveyed the viability of every essential gene disruptant in S. 37 cerevisiae (1,106 genes), and found that 9% of the gene disruptants proliferate and form colonies 38 spontaneously (i.e. without artificial mutagenesis). Genome analysis showed that most of the 39 proliferating strains had gained an extra chromosome (i.e. aneuploidy), which is typically an 40 outcome of chromosome missegregation. This is a reasonable path to BOE in S. cerevisiae, 41 because its haploid is tolerant to a chromosome gain for 13 of 16 chromosomes (Torres et al., 42 2007). However, we speculated that there might be many more bypassable essential genes in yeast, 43 as some non-bypassable essential gene disruptants might recover their viability by acquiring 44 extragenic mutations, which are rarely introduced without mutagenesis. 45 2 Comprehensive identification of suppressor mutations would help to elucidate secondary or 1 backup mechanisms that can substitute for primary mechanisms. Hitherto 'masked', these 2 alternative mechanisms may act as the dominant pathways in specific cell types and/or diseased 3 cells. To this end, we designed a BOE screening using the fission yeast S. pombe, which has a 4 similar number (1,260) of essential genes to S. cerevisiae (Kim et al., 2010). A notable difference 5 from S. cerevisiae is that S. pombe has only 3 chromosomes, and the haploid yeast is inviable 6 when an extra copy of either chromosome I or II (the two larger chromosomes) is inherited (Niwa 7 and Yanagida, 1985). Thus, BOE via extra chromosome gain is likely an infrequent event in S. 8 pombe. In the present study, we carried out BOE screening for randomly selected 92 essential 9 genes in S. pombe, based on UV mutagenesis of spores in which essential genes were deleted. containing 1 g/L sodium glutamate instead of 3.75 g/L).

21
Gene disruption 22 Essential genes were selected based on information found in the Pombase database (Wood et al., 23 2012). 92 genes on chromosome II were randomly selected. Conventional one-step replacement 24 was conducted using ~500-bp homologous sequences (5'UTR and 3'UTR of the gene to be 25 deleted) (Krawchuk and Wahls, 1999). A tandem G418-resistance (kanMX) /ura4+ cassette was 26 used as the selection marker (however, ura4+ marker was not actually used for selection). For 27 most genes, we directly generated a linear construct (5'UTR-G418-ura4+-3'UTR) by two rounds 28 of PCR using two sets of primers (i.e. nested PCR). In some cases, the PCR fragment was cloned 29 into a vector using an Infusion kit (Takara), and the linear construct was amplified with T7/T3 30 primer set from the plasmid template. The linear DNA was transformed into the G29 diploid strain 31 using the standard lithium acetate/PEG-mediated method, and disruption of the target gene was 32 confirmed by colony PCR using KOD-Fx-Neo or KOF-ONE kit (Toyobo). When the endogenous 33 gene and G418-ura4+ cassette had a similar length, we used a longer version of G418-ura4+ 34 cassette to distinguish disrupted and endogenous alleles by length. PCR primers for gene 35 disruptions and their confirmation are listed in Table S1.

41
The solution was shaken at 80 rpm at room temperature overnight to digest non-sporulated cells. 42 The spores were harvested and further treated with 30% ethanol for 30 min (80 rpm, room  Prior to UV mutagenesis, the viability and purity of spores were determined by plating onto 2 normal YE5S plate and YE5S supplemented with G418 (100 µg/ml) and cycloheximide (100 3 µg/ml), respectively. No haploid spores were expected to grow on the G418/cycloheximide plate, 4 since an essential gene had been replaced with G418. However, colonies were always formed 5 typically at ~1 × 10 -6 frequencies. Cells in these colonies were diploids, which we interpreted to be 6 derived from diploid spores generated at low frequency during meiosis; diploid spores would also 7 be resistant to glusulase or ethanol. In cases in which the putative diploid contamination frequency 8 was < 5 × 10 -5 , we moved on to UV mutagenesis and screening. In cases in which the 9 contamination frequency was ≥ 5 × 10 -5 , we discarded the sample and repeated the spore isolation 10 process. The reason for differences in the prevalence of putative diploid spores is unknown.  non-spore-forming colonies as candidate BOE haploids; colonies that did not match this criterion 23 were likely diploids and disregarded. The candidate colonies were subjected to colony PCR, with 24 which the disruption of the target gene was reconfirmed (see Fig. 2B). For top3Δ, we performed 25 mutagenesis in a rad13Δ (DNA repair-deficient) background, in order to decrease UV power 26 (1,500 µJ/cm 2 , 5% viability) and avoid cytotoxicity. However, since we obtained expected BOE 27 results for cut7Δ in the presence of rad13+, we did not introduce rad13Δ for any other genes.

29
Whole-genome sequencing and sequence analysis 30 To identify suppressor mutations, bulk segregant analysis was performed. Survivor strains were 31 crossed with a wild-type strain, and spores were plated on G418-containing YE5S plates. After 7 d, 32 ~1,000 colonies were collected and DNA was extracted with Dr. GenTLE (Takara). Genomic  spores were plated on G418-containing plates and simultaneously mutagenized by UV irradiation. 42 If a haploid colony is obtained on this plate, it has likely acquired a suppressor mutation(s), 43 indicating that the essentiality of the gene has been bypassed. 44 We first applied this method to two gene disruptants, cut7Δ (kinesin-5) and top3Δ (type I . For cut7Δ, we obtained a total of 8 haploid colonies in the first 2 experiment and 30 more in a later experiment, in which 5-fold more spores were mutagenized 3 ( Fig. 2A, B). We analysed 26 colonies by target sequencing of pkl1 and msd1 genes (Msd1 is a 4 positive regulator of Pkl1 (Yukawa et al., 2015)), whole-genome sequencing, and/or genetic 5 linkage test (pkl1 locus is close to the rpl42 locus, at which a mutation was introduced to confer 6 cycloheximide resistance in our strain). The combined results suggested that suppressor mutations 7 reside in pkl1 for 19 strains and in msd1 for the remaining 7 strains (Fig. 2C, E). Mutagenesis of 8 top3Δ yielded 3 haploid strains (Fig. 2D), and direct sequencing of the rqh1 gene identified a 9 mutation in all cases (Fig. 2E). Thus, our screening successfully elucidated known BOE 10 relationships. 11 We then expanded the screening to 92 essential genes located on chromosome II. For 20 of 12 these, we obtained 1~17 haploid colonies, which corresponds to 22% (Fig. 3). This frequency is 13 much higher than that obtained in the previous mutagenesis-free screening in S. cerevisiae (Liu et

15
The 20 suppressible genes possess divergent known biological functions. 10 genes (50%) 16 were related to mitochondrial function (Fig. 4A). This may be partly explained by the fact that, in 17 the regular medium containing >2% glucose, cell proliferation does not depend much on 18 mitochondrial respiration (Takeda et al., 2015). 6 genes were associated with RNA processing and 19 ribosome functions; the basis of these trends are unclear. Overall, 90% of the genes had clear 20 orthologues in S. cerevisiae and H. sapiens, indicating that BOE is not limited to unconserved 21 genes (Fig. 4B, C). However, the orthologues of 70% genes were reported to be non-essential in S. 22 cerevisiae (Fig. 4D). An obvious next step would be to identify suppressor mutations of each 23 survivor to understand how an essential mechanism can be bypassed. 27% of the genes in one or more BOE assays, which is a similar frequency to ours.

32
Coincidentally, in the two studies, 29 common essential genes were screened. Upon 33 comparison, 21 genes were not bypassable in both studies, whereas 5 genes were common BOE 34 hits. 2 and 1 hits were uniquely found in their and our studies, respectively (Fig. 3). Thus, the 35 comparison indicates that both screens had a good agreement in bypassability, but also suggests 36 that a single screen cannot identify all possible BOE.  (Murakami et al., 2007). However, we could 4 not find dis3 mutations in any of the 3 BOE strains we obtained, indicating that other genes had 5 acquired suppressive mutations. 6 In summary, we have established an alternative sensitive-and perhaps less 7 labour-intensive-methodology for mutagenesis-based BOE screening in fission yeast, and 8 expanded the list of genes whose essentiality is bypassable. Our methodology allows for a 9 straightforward scale-up of the screen, from which we expect to reveal masked cellular 10 mechanisms.