Secreted protein toxins are widely used weapons in conflicts between organisms. Elucidating how organisms genetically adapt to defend themselves against these toxins is fundamental to understanding the coevolutionary dynamics of competing organisms. Within yeast communities, “killer” toxins are secreted to kill nearby sensitive yeast, providing a fitness advantage in competitive growth environments. Natural yeast isolates vary in their sensitivity to these toxins, but to date, no polymorphic genetic factors contributing to defense have been identified. We investigated the variation in resistance to the killer toxin K28 across diverse natural isolates of the Saccharomyces cerevisiae population. Using large-scale linkage mapping, we discovered a novel defense factor, which we named KTD1 . We identified many KTD1 alleles, which provided different levels of K28 resistance. KTD1 is a member of the DUP240 gene family of unknown function, which is rapidly evolving in a region spanning its two encoded transmembrane helices. We found that this domain is critical to KTD1 ’s protective ability. Our findings implicate KTD1 as a key polymorphic factor in the defense against K28 toxin.

Secreted protein toxins are widely used weapons in conflicts between organisms. Killer yeast produce killer toxins that inhibit the growth of nearby sensitive yeast. We investigated variation in resistance to the killer toxin K28 across diverse natural isolates of the Saccharomyces cerevisiae population and discovered a novel defense factor, which we named KTD1, that is an important determinant of K28 toxin resistance. KTD1 is a member of the DUP240 gene family of unknown function. We uncovered a putative role of DUP240 proteins in killer toxin defense and identified a region that is undergoing rapid evolution and is critical to KTD1's protective ability. Our findings implicate KTD1 as a key factor in the defense against killer toxin K28.

Understanding the genetic causes of trait variation is a primary goal of genetic research. One way that individuals can vary genetically is through the existence of variable pangenomic genes - genes that are only present in some individuals in a population. The presence or absence of entire genes could have large effects on trait variation. However, variable pangenomic genes can be missed in standard genotyping workflows, due to reliance on aligning short-read sequencing to reference genomes. A popular method for studying the genetic basis of trait variation is linkage mapping, which identifies quantitative trait loci (QTLs), regions of the genome that harbor causative genetic variants. Large-scale linkage mapping in the budding yeast Saccharomyces cerevisiae has found thousands of QTLs affecting myriad yeast phenotypes. To enable the resolution of QTLs caused by variable pangenomic genes, we used long-read sequencing to generate highly complete de novo assemblies of 16 diverse yeast isolates. With these assemblies we resolved growth QTLs to specific genes that are absent from the reference genome but present in the broader yeast population at appreciable frequency. Copies of genes also recombine onto chromosomes where they are absent in the reference genome, and we found that these copies generate additional QTLs whose resolution requires pangenome characterization. Our findings demonstrate the power of long-read sequencing to identify the genetic basis of trait variation.