Whole-Genome Sequencing and iPLEX MassARRAY Genotyping Map an EMS-Induced Mutation Affecting Cell Competition in Drosophila melanogaster

Abstract

Cell competition, the conditional loss of viable genotypes only when surrounded by other cells, is a phenomenon observed in certain genetic mosaic conditions. We conducted a chemical mutagenesis and screen to recover new mutations that affect cell competition between wild-type and RpS3 heterozygous cells. Mutations were identified by whole-genome sequencing, making use of software tools that greatly facilitate the distinction between newly induced mutations and other sources of apparent sequence polymorphism, thereby reducing false-positive and false-negative identification rates. In addition, we utilized iPLEX MassARRAY for genotyping recombinant chromosomes. These approaches permitted the mapping of a new mutation affecting cell competition when only a single allele existed, with a phenotype assessed only in genetic mosaics, without the benefit of complementation with existing mutations, deletions, or duplications. These techniques expand the utility of chemical mutagenesis and whole-genome sequencing for mutant identification. We discuss mutations in the Atm and Xrp1 genes identified in this screen.

Keywords: Drosophila melanogaster; Flybook; Xrp1; cell competition; iPLEX MassARRAY; whole-genome sequencing.

Figure 1

Flowchart of steps leading to identification and mapping of a single mutant allele without recourse to genetic complementation assays or collections of known mutants.

Figure 2

Scheme of a forward genetic screen to find genes required during cell competition. (A) The F2 crossing scheme for screening mutants on chromosomal arm 3R. In the F1 generation, single F1 male progeny with equal contributions of red and white eye tissues were selected (illustrated in C, D). After breeding with RpS3^Plac92 females, the F2 progeny with appropriate genotype were screened for candidate mutant strains (illustrated in E, F). Mutants were recovered and bred from balanced siblings. (B) Cartoon of the screen concept. FLP-FRT mediated mitotic recombination in RpS3 heterozygous cells produces unpigmented, non-Minute FRT82B homozygous cells and reciprocally recombinant cells homozygous for the mutation in the RpS3 gene. The latter die, leaving the FRT82B homozygous clones to compete with unrecombined genotype RpS3 heterozygous genotype. (C) Adult eye image of an unmutagenized control fly in the F1 generation. Genotype: y w eyFLP; FRT82B w+ arm-LacZ/FRT82B (unmutagenized). (D) Schematic of the F1 generation. Mosaic flies that are indistinguishable from controls in proportions of red and white eye cells were selected. These should lack novel mutants that cell-autonomously alter growth, in which the unpigmented clones should either be relatively smaller (in case of a growth deficit) or larger (in case of a mutation causing overgrowth). (E) Adult eye image of an unmutagenized control fly in the F2 generation. Few pigmented cells remain in the eye, which is largely homozygous for the unmutagenized FRT82B progenitor chromosome. (F) Schematic of the F2 screen. Mutants that might affect cell competition are selected from F2 genotypes where more pigmented cells survive than in controls (see Figure 3, B and D for examples). Genotype: y w eyFLP; FRT82B/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3.

Figure 3

Atm and Xrp1 mosaic phenotypes in Minute and non-Minute backgrounds. (A–D) Adult eye phenotypes from F2 screen (see Figure 2A for method and genotypes) (A) FRT82B control; (B) FRT82B M9-25; (C) FRT82B Atm³; (D) FRT82B M2-73. (B–D) Homozygous M9-25, Atm³, and M2-73 cells were less able to eliminate pigmented RpS3 heterozygous cells. (E, F) Late third instar wing imaginal disc. The RpS3/+ background is labeled for β-galactosidase expression in green; non-Minute clones are unlabeled. Cell death is detected with α-activated-Dcp-1 labeling in red. (E) Competitive cell death of RpS3/+ cells is largely limited to boundaries between the genotypes. (F) High cell death levels were observed throughout Atm³ clones. (G–I) Atm³ homozygous clones induced between 48 and 72 hr in a non-Minute background and dissected at late third instar larval stage from 48 to 72 hr after clone induction (ACI) as indicated. The FRT82B Atm³ heterozygous background was labeled for β-galactosidase expression, the FRT82B homozygous twin-spots were labeled more brightly, the FRT82B Atm³ homozygous clones were unlabeled. Atm³ homozygous clones were under-represented compared to twin-spots after 72 hr but not after shorter times. Scalebars represent 250 microns for panels A–D and 50 microns for panels E–I. Genotypes: (A) y w eyFLP;FRT82B/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3, (B) y w eyFLP;FRT82B M9-25/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3, (C) y w eyFLP;FRT82B Atm³/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3, (D) y w eyFLP;FRT82B M2-73/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3, (E) hsFLP;FRT82B/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3, (F) hsFLP;FRT82B Atm³/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3, (G–I) y w hsFLP;FRT82B Atm³/FRT82B P{w+, arm-LacZ}.

Figure 4

iPLEX MassARRAY genotyping maps the M2-73 mutant. (A, B) Recombinant mapping principles. ‘X’ illustrates that meiotic recombination occurs between homologs. (A) A cartoon representation of recombinant chromosome arms 3R expected to retain the M2-73 phenotype. These should retain M2-73 SNPs proximal to and including the locus responsible (asterisk), and exhibit various crossover points distal to the locus. When DNA from 50 such recombinants is pooled, the frequency of M2-73 SNPs is expected to decrease distal to the locus as the frequency of Tb SNPs increases. (B) A cartoon representation of recombinant chromosome arms 3R expected to lack the M2-73 phenotype. These should retain Tb SNPs distal to and including locus responsible for the M2-73 phenotype, and exhibit crossover points proximal to the locus. When DNA from such recombinants is pooled, the frequency of M2-73 SNPs should increase proximally to the locus as the frequency of Tb SNPs decreases. Comparing observed SNP frequencies for the two pools should limit the mutant locus to a chromosome segment that retains 100% M2-73 SNPs in phenotypically mutant recombinants and 100% Tb SNPs in phenotypically wild-type recombinants. (C) Distribution of the 17 SNPs that were assessed by iPLEX (crosses) and 22 exonic variants on the M2-73 chromosome (circles). An open circle shows the position of the coding mutation in Xrp1, closed circles show the other mutations. Some of the coding alleles were used as SNPs, in addition to noncoding SNPs. (D) Allele frequencies determined by iPLEX for 17 SNPs in the FRT82B M2-73/FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3 genotype. Deviation from 50% reflects amplification and detection bias in the iPLEX method (see section on Mapping a single allele using recombination and iPLEX MassARRAY). This heterozygote genotype is used because the recombinant DNA is also heterozygous with the FRT82B P{w+, arm-LacZ} P{w+, tub-Gal80} RpS3 chromosome, in which background the M2-73 phenotype is assessed. (E) Normalized frequencies of reference alleles (derived from the Tb¹ chromosome). Solid line, pooled recombinants exhibiting the M2-73 phenotype; dashed line, pooled recombinants lacking the M2-73 phenotype. Normalized allele frequency is calculated as (measured % Tb allele in recombinants)/(measured % Tb allele in (D) control) × 100, to account for any amplification or detection bias in the MassARRAY method. The arrow indicates SNP 19,029,906 (+BDGP Release 6) where Tb SNPs are recovered in ∼0% of phenotypically M2-73 recombinants and ∼100% of phenotypically wild-type recombinants. This SNP is ∼100 kb from a coding mutation in the Xrp1 gene (open circle in A).