Abstract
Herbicide resistance in weeds is one of the greatest challenges in modern food production. The grass species Lolium multiflorum is an excellent model species to investigate evolution under similar selection pressure because populations have repeatedly evolved resistance to many herbicides, utilizing a multitude of mechanisms to neutralize herbicide damage. In this work, we investigated the gene that encodes acetyl-CoA carboxylase (ACCase), the target site of the most successful herbicide group available for grass weed control. We sampled L. multiflorum populations from agricultural fields with history of intense herbicide use, and studied their response to three ACCase-inhibiting herbicides. To elucidate the mechanisms of herbicide resistance and the genetic relationship among populations, we resolved the haplotypes of 97 resistant and susceptible individuals by sequencing ACCase amplicons using long-read DNA sequencing technologies. Our dose-response data indicated the existence of many, often unpredictable, resistance patterns to ACCase-inhibiting herbicides, where populations exhibited as much as 37-fold reduction in herbicide response. The majority of the populations exhibited resistance to all three herbicides studied. Phylogenetic and molecular genetic analyses revealed multiple evolutionary origins of resistance-endowing ACCase haplotypes, as well as widespread admixture in the region regardless of cropping system. The amplicons generated were diverse, with haplotypes exhibiting 26-110 polymorphisms. Polymorphisms included insertions and deletions 1-31 bp in length, none of which were associated with the resistance phenotype based on an association analysis. We also found evidence that some populations have multiple mechanisms of resistance. Our results highlight the astounding genetic diversity in L. multiflorum populations, and the potential for repeated evolution of herbicide resistance across the landscape that challenges weed management approaches and jeopardizes sustainable weed control practices. We provide an in-depth discussion of the evolutionary and practical implications of our results.