Analysis of multiple ethyl methanesulfonate-mutagenized Caenorhabditis elegans strains by whole-genome sequencing

Abstract

Whole-genome sequencing (WGS) of organisms displaying a specific mutant phenotype is a powerful approach to identify the genetic determinants of a plethora of biological processes. We have previously validated the feasibility of this approach by identifying a point-mutated locus responsible for a specific phenotype, observed in an ethyl methanesulfonate (EMS)-mutagenized Caenorhabditis elegans strain. Here we describe the genome-wide mutational profile of 17 EMS-mutagenized genomes as assessed with a bioinformatic pipeline, called MAQGene. Surprisingly, we find that while outcrossing mutagenized strains does reduce the total number of mutations, a striking mutational load is still observed even in outcrossed strains. Such genetic complexity has to be taken into account when establishing a causative relationship between genotype and phenotype. Even though unintentional, the 17 sequenced strains described here provide a resource of allelic variants in almost 1000 genes, including 62 premature stop codons, which represent candidate knockout alleles that will be of further use for the C. elegans community to study gene function.

Figure 1.—

WGS identifies a cis-regulatory deletion in ot358 animals. (A) Expression of a GFP-based AIY cell fate marker (otIs173) in wild-type or ot358 adults (red: AIY interneuron). (B) Percentage of adults showing expression of the AIY GFP reporter otIs173 (array[fosmid WRM064cD04] is a chromosomally integrated array containing a ttx-3 rescuing genomic fragment in the context of the fosmid shown in C, kindly provided by V. Reinke). (C) ttx-3 genomic locus. The ot358 deletion removes a cis-regulatory element regulating ttx-3 expression (enhancer in blue). Our screen and previous screens have also identified strains bearing mutations in splice sites or coding regions of the ttx-3 locus (in green are mutations reported here for the first time). (D) Number of noncovered regions in ot358 in the 0.6-cM mapping region. Left, noncovered in ot358; right, noncovered in ot358 but covered in ot354. Tested and confirmed deletions are denoted in parentheses. (E) Number of noncovered regions in ot358 or ot177 in the whole genome. ot358 left, noncovered in ot358; ot358 right, noncovered in ot358 but covered in ot354. ot177 left, noncovered in ot177; ot177 right, noncovered in ot177 but covered in ot358. Tested and confirmed deletions are denoted in parentheses.

Figure 2.—

Lack of correlation between fold outcross and variant number. (A) y-axis represents unique variant numbers divided by total number of sequenced bases (those with at least 3× and at most 60× sequencing depth) for each data set. This accounts for differences in coverage between different data sets. Unique variants exclude those shared between two or more data sets. (B) This graph presents unique variants per analyzed chromosome, excluding those linked to the mutation or transgene reporter. To deal with linkage, we did not include variants on chromosomes that contained (a) the phenotype-causing mutation isolated for, (b) the integrated fluorescent transgene used to score the mutant phenotype, and (c) the X chromosome, which is subject to fewer recombination events due to method of outcrossing: F₁ males from the first outcross (which are XO for this chromosome and therefore do not undergo meiotic recombination) are mated with the wild-type strain. F₂ progeny from this second outcross containing the mutant phenotype are then isolated. This analysis accounts for mutation hitchhiking effects.

Figure 3.—

lin-59 lethality is suppressed by lsy-12. (A) lin-59 gene structure. The table at the bottom displays viability of each lin-59(−) strain. (B) Outcrossing of OH7677 reveals a modifier locus on chromosome V that suppresses the observed lethality likely due to lin-59(ot104). (C) Variants discovered on chromosome V in OH7677. Variants were processed and filtered with MAQGene (Bigelow et al. 2009). One of the missense variants resides within the lsy-12 locus. (D) lsy-12 gene structure. This schematic shows an incomplete version of the lsy-12 locus per the last release of WormBase. Transcript mapping shows that the lsy-12 locus extends in the more upstream located mys-3 gene, a histone acetyltransferase (M. M. O'Meara and O. Hobert, unpublished results). (E) A second independently isolated allele of lsy-12 (Sarin et al. 2007) suppresses two other independently isolated lin-59 alleles (Potts et al. 2009) that display lethality at 25°.

Figure 4.—

Outcrossing reduces total variant number but introduces new variants. Variant numbers that are unique to the ot354 data set, i.e., not found in any other of the 16 sequenced mutant genomes, are considered. ot354; otIs173 was outcrossed against the laboratory wild-type N2 Bristol C. elegans strain and F₂ homozygous mutant; reporter animals were singled out after each outcross. x/x are numbers that indicate the state of each unique variant position (wild type, heterozygous, etc.) in the other strain's genome, revealing both the loss and the gain of variants after six additional outcrossing events. Homozygous mutations were defined by those positions covered by more than ten reads holding a variant:total read ratio x > 0.86. Heterozygous mutations were defined by those positions covered by more than two reads holding a variant:total read ratio 0.4 > x ≥ 0.86. See materials and methods for the value of the other parameters.