The genomic basis of adaptation: target site and non-target site resistance mechanisms
Pesticides are designed to target particular genes in functional pathways of pest organisms. For example, quinone-outside inhibitor fungicides (QoIs) target the products of the fungal gene cytochrome b, a critical component of aerobic respiration. In plants, the herbicide glyphosate targets the protein product of EPSPS (5-enolpyruvylshikimate-3-phosphate synthase), a central gene in the shikimate acid pathway. One type of pesticide resistance— termed target site resistance—occurs when a mutation in the gene targeted by the pesticide alters the conformation of the protein, reducing or completely eliminating the ability of the pesticide to bind to the protein’s active site (Délye et al. 2013). A number of studies have examined target site resistance by sequencing the target locus and assessing the various mutations associated with resistance. Givena priori information about how pesticides work, thus providing obvious genetic candidates to explore, along with the relative simplicity of sequencing a single gene, we currently have a somewhat comprehensive understanding of the type and number of genetic variants associated with target site resistance, especially in herbicide resistant weeds (Tranel, Wright, & Heap 2021; Baucom 2016).
We still have much to learn about target-site resistance evolution, however, especially in the context of broader, genome-scale dynamics. In this special feature, Clarkson et al. (2021) move beyond investigating the dynamics of target-site resistance in isolation by explicitly examining the role of intragenic variation on resistance to pyrethroid, a class of insecticide that is used to control mosquito populations associated with malaria. Their investigation of whole-genome sequence data reveals a ‘a tale of two alleles’: two widespread large-effect target-site resistance alleles within the voltage-gated sodium channel (VGSC) gene appear to be on different evolutionary trajectories. One allele, likely an early ancestral mutation, is associated (i.e.,in strong positive linkage disequilibrium) with a subsequent explosion of 13 secondary non-synonymous mutations, whereas the second allele is associated with fewer mutations. Further, most of these mutations are background-dependent, occurring nearly exclusively on distinct haplotypes—haplotypes that are associated with different signatures of selection despite harbouring the same focal resistance allele, implying important compensatory or enhancing allelic interactions for resistance evolution.
A major contribution to our understanding of the predictability of evolution stems from work examining the repeatability of target site changes that confer pesticide resistance across insect and weed species, respectively (Martin & Orgogozo 2013). However, whether or not parallel genetic changes lead to resistance among fungal plant pathogens has yet to be succinctly summarized. In this issue, Hawkins and Fraaije (2021) investigate the extent of parallel evolution of individual mutations in target genes among species of fungal pathogens. Focusing on mutations associated with four classes of fungicide, they show that the target-genes vary substantially in the diversity of mutations detected. For two fungicide classes (Qols and MBCs) the same mutations are observed repeatedly across species. In contrast, a greater diversity of resistance mutations was uncovered within genes targeted by azole and SDHI fungicides, providing less evidence for extreme parallelism across species compared to QoIs and MBCs.
Another form of target-site resistance is from gene amplification, where increased copy number of the target locus leads to more functional protein and subsequent resistance. In a handful of weeds, an increase in the copy number of the EPSPS locus leads to high glyphosate resistance (reviewed in Gaines et al. 2019); while the underlying mechanism of this copy number increase has been described (Koo et al. 2018), we understand relatively little about the long term maintenance of copy number variation (CNV) and how gene amplification may influence interactions with other loci. Yakimowski, Teitel and Caruso (2021) quantified patterns of variation of target gene copy number and resistance phenotypes within and among populations—the ‘natural history’ of a resistance CNV—to provide insight into the evolution of glyphosate resistance in the agricultural weedAmaranthus palmeri in the eastern United States. They detected a steep increase in phenotypic glyphosate resistance at a threshold value of ~15 gene copies, but also found that populations with the highest mean resistance contained some low copy number individuals (albeit at low-frequency). From 15 to 160 gene copies the level of resistance changed very little; however, the proportion of low-resistance phenotypes gradually decreased in populations with increased copy number, suggesting that dosage of the target gene with increasing copy number might compensate for negative interactions with other loci. Potential positive interactions with other genes were also observed in populations from Georgia. Overall, target gene copy number variation explained a high proportion (~57%) of variation in phenotypic resistance among populations.
In another contribution to this special issue, Gaines et al. (2020) show that copy number variation of this target gene is also present in populations of A. palmeri from Brazil and Uruguay, indicating that copy number variation related to resistance is found broadly across the landscape. Interestingly, however, resistance in Argentinian lineages of A. palmeri was due in large part to non-target site resistance mechanisms—i.e. resistance mechanisms that do not involve the target site, such as altered translocation or detoxification of the pesticide, among others—rather than elevated copy number of the EPSPS locus. These results show both genomic flexibility in solving the problem of herbicide exposure and the independent, novel evolution of resistance across geography in this species.
Thus, in addition to target site resistance mechanisms, organisms can also evolve resistance through non-target site mechanisms. Non-target site resistance mechanisms, which are often thought to be due to polygenic variation, can both confer resistance as well as potentially supplement target-site effects. While both target site and non-target site mechanisms have previously been uncovered within the same herbicide resistant weed species (as in Gaines et al. 2020), the relative contribution of either type of mechanism has yet to be clearly delineated in any weed species. Using another glyphosate resistantAmaranthus species, A. tuberculatus (common waterhemp), Kreiner et al. (2021) uncovered the cryptic contribution of genome-wide alleles to glyphosate resistance. On the genomic background of agricultural populations harbouring high frequencies of target-site resistance mechanisms (Kreiner et al. 2019), the authors illustrated a near-equal importance of non-target and target-site mechanisms. Further, they uncovered hundreds of alleles associated with non-target site resistance that show not only evidence of recent strong selection from herbicides but a classic trade-off between effect size and allele frequency that implicates pleiotropy as a key constraint to the evolution of herbicide resistance.
As our understanding of the genetic architecture of pesticide resistance and governing selective processes deepens, a key question will be how consistently such alleles are involved across geographic scales. This question is addressed by Hartmann et al. (2020), who investigated the architecture of azole fungicide in a key wheat pathogen,Zymoseptoria tritici across three continents. They uncovered a suite of azole resistance-related loci across the genome including a novel large-effect gene, DHHC palmitoyl transferase. Along with key alleles conferring resistance to three other chemical classes of fungicides, the authors find evidence that the genomic architecture of fungicide resistance is largely distinct across continents, with the exception of large-effect genes that act as hotspots for convergence. Overall, this collection of work characterizing the genetic architecture of pesticide resistance uncovers remarkable complexity in monogenic and polygenic contributions and the processes that govern their assemblage across genomes, from background- and population-specific constraints to the potential for pleiotropic tradeoffs.