Repeated origins, widespread gene flow, and allelic interactions of target-site herbicide resistance mutations
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
This paper studies the evolution of herbicide resistance in Amaranthus tuberculatus, a widespread agricultural weed. By illuminating how adaptive mutations arose and spread in this remarkable example of rapid human-induced adaptation, the study will be of interest to a broad audience, ranging from plant biologists interested in herbicide resistance to evolutionary biologists and population geneticists studying the fundamental factors and processes that govern rapid adaptation. The paper applies innovative population genetic methodology to support its primary finding that resistance mutations have evolved multiple times in parallel.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their names with the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
- Evolutionary Biology (eLife)
Abstract
Causal mutations and their frequency in agricultural fields are well-characterized for herbicide resistance. However, we still lack understanding of their evolutionary history: the extent of parallelism in the origins of target-site resistance (TSR), how long these mutations persist, how quickly they spread, and allelic interactions that mediate their selective advantage. We addressed these questions with genomic data from 19 agricultural populations of common waterhemp ( Amaranthus tuberculatus ), which we show to have undergone a massive expansion over the past century, with a contemporary effective population size estimate of 8 x 10 7 . We found variation at seven characterized TSR loci, two of which had multiple amino acid substitutions, and three of which were common. These three common resistance variants show extreme parallelism in their mutational origins, with gene flow having shaped their distribution across the landscape. Allele age estimates supported a strong role of adaptation from de novo mutations, with a median age of 30 suggesting that most resistance alleles arose soon after the onset of herbicide use. However, resistant lineages varied in both their age and evidence for selection over two different timescales, implying considerable heterogeneity in the forces that govern their persistence. Two such forces are intra- and inter-locus allelic interactions; we report a signal of extended haplotype competition between two common TSR alleles, and extreme linkage with genome-wide alleles with known functions in resistance adaptation. Together, this work reveals a remarkable example of spatial parallel evolution in a metapopulation, with important implications for the management of herbicide resistance.
Article activity feed
-
-
-
Evaluation Summary:
This paper studies the evolution of herbicide resistance in Amaranthus tuberculatus, a widespread agricultural weed. By illuminating how adaptive mutations arose and spread in this remarkable example of rapid human-induced adaptation, the study will be of interest to a broad audience, ranging from plant biologists interested in herbicide resistance to evolutionary biologists and population geneticists studying the fundamental factors and processes that govern rapid adaptation. The paper applies innovative population genetic methodology to support its primary finding that resistance mutations have evolved multiple times in parallel.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. …
Evaluation Summary:
This paper studies the evolution of herbicide resistance in Amaranthus tuberculatus, a widespread agricultural weed. By illuminating how adaptive mutations arose and spread in this remarkable example of rapid human-induced adaptation, the study will be of interest to a broad audience, ranging from plant biologists interested in herbicide resistance to evolutionary biologists and population geneticists studying the fundamental factors and processes that govern rapid adaptation. The paper applies innovative population genetic methodology to support its primary finding that resistance mutations have evolved multiple times in parallel.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their names with the authors.)
-
Reviewer #1 (Public Review):
The rapid evolution of target site resistance to pesticides in plant and insect populations poses a challenge to evolutionary models, as it requires a sufficient supply of adaptive genetic variation that is sometimes difficult to reconcile with classical population genetic theory. Kreiner et al address this conundrum by conducting an in-depth study of the evolutionary patterns associated with the evolution of target site resistance to acetolactate synthase (ALS) inhibiting and protoporphyrinogen oxidase (PPO) inhibiting herbicides in Amaranthus tuberculatus, an agricultural weed that is widespread in North America. Several of the mutations responsible for resistance to these herbicides have been previously characterized molecularly, so the adaptive alleles are known. Kreiner et al focus specifically on the …
Reviewer #1 (Public Review):
The rapid evolution of target site resistance to pesticides in plant and insect populations poses a challenge to evolutionary models, as it requires a sufficient supply of adaptive genetic variation that is sometimes difficult to reconcile with classical population genetic theory. Kreiner et al address this conundrum by conducting an in-depth study of the evolutionary patterns associated with the evolution of target site resistance to acetolactate synthase (ALS) inhibiting and protoporphyrinogen oxidase (PPO) inhibiting herbicides in Amaranthus tuberculatus, an agricultural weed that is widespread in North America. Several of the mutations responsible for resistance to these herbicides have been previously characterized molecularly, so the adaptive alleles are known. Kreiner et al focus specifically on the question of how often these individual resistance mutations have arisen within different populations, whether this occurred from de novo mutations or standing genetic variation, and how resistance mutations have subsequently spread across the landscape. They find remarkable parallelism in the evolution of resistance mutations, detecting instances of repeated mutational origin of individual mutations even within localized subpopulations, as well as some examples that appear more consistent with evolution from standing variation. Their interpretation is that such rapid adaptation was facilitated by a massive recent population size expansion. They also present results suggesting that the geographic distribution of resistance mutations has been shaped by both intra- and inter-locus allelic interactions due to haplotype competition.
This is a well-written study that presents intriguing results on the evolutionary dynamics of resistance evolution. The conducted analyses stand out through the application of cutting-edge population genetic methodology. In particular, the use of ancestral recombination graph (ARG)-based methods to infer tree-sequences along the genome is an innovative approach that allows the authors to disentangle independent mutational origins of resistance mutations, date individual resistance alleles, and infer selection coefficients over time at these loci. The implementation of these novel analyses appears sound. To the extent that we can trust them to provide correct results, the conclusions are well supported by the data.
While the use of sophisticated new methodology clearly represents a major appeal of the study, it also raises some concerns about the robustness of the results. At present, we simply do not yet have a very good understanding of how accurate the results from ARG-based inference methods are in the light that some of the assumptions underlying these methods are certainly violated in real-world populations. Spatial population structure, complex selection scenarios, variable mutation and recombination rates, or phasing errors are just some of the factors that could potentially mislead the resulting estimates. The error bars provided by these approaches, for instance for the allele age estimates shown in Figure 3, paint just part of the picture, given that they reflect only stochasticity in the MCMC analyses, but not any systematic errors due to violation of the underlying assumptions.
However, it is also clear that a thorough analysis of all potential factors that could limit the robustness of the conducted analyses and lead to biases in the results would be a serious undertaking, which I would consider as beyond the scope of the present study. To do this rigorously would presumably require comprehensive simulation analyses and, ideally, evaluation against positive controls where the true tree-sequences are known. Nevertheless, as I outline in my specific recommendations below, I believe that there are a number of simpler tests the authors could easily perform to at least test the robustness of their results to phasing errors, misspecification of the recombination rate, and the use of different demographic inference methods.
-
Reviewer #2 (Public Review):
In this manuscript, the authors investigate the origins of herbicide resistance alleles in an agricultural weed A. tuberculatus.
The authors collected genomic data from 18 populations across the Mid West of the US and Ontario in Canada. They focus on known mutations in two genes that are targeted by herbicides (PPO and ALS). One deletion in PPO and two non-synonymous mutations in ALS are common (8-34%) and lend themselves to more in depth analysis. A sweep analysis around the target genes that compares chromosomes with and without a resistance mutation reveal an increase in long range haplotypes (as measured by XPEHH) and a decrease in diversity on chromosomes with the PPO deletion, but not for the two common ALS alleles. This shows that the common resistance alleles are not due to a hard sweep.
Next the …
Reviewer #2 (Public Review):
In this manuscript, the authors investigate the origins of herbicide resistance alleles in an agricultural weed A. tuberculatus.
The authors collected genomic data from 18 populations across the Mid West of the US and Ontario in Canada. They focus on known mutations in two genes that are targeted by herbicides (PPO and ALS). One deletion in PPO and two non-synonymous mutations in ALS are common (8-34%) and lend themselves to more in depth analysis. A sweep analysis around the target genes that compares chromosomes with and without a resistance mutation reveal an increase in long range haplotypes (as measured by XPEHH) and a decrease in diversity on chromosomes with the PPO deletion, but not for the two common ALS alleles. This shows that the common resistance alleles are not due to a hard sweep.
Next the authors use a phylogenetic approach (taking into account recombination) to determine how often each of the three common alleles has originated. They find 6 and 2 origins for the two ALS alleles and 3 origins for the PPO deletion. They then map these origins onto the 18 sampled populations and see that some of the origins are found only in certain regions, but others are spread across Ontarios and the US. Next, the authors try to determine *when* the alleles arose and whether they had already been present as standing genetic variation before being picked up by selection. Finally, the authors try to determine interactions between the alleles.
In my opinion, the map with the origins of the alleles (fig 2B) is the key result of the paper. It shows multiple origins, co-occurrence of origins (local soft sweeps), and it shows that there is migration that allowed alleles to spread to different locations, but not enough to spread the alleles evenly across all locations. In a way, this result is a worst-case scenario for those who try to prevent weed resistance, because the results show that resistance evolved at least 9 times independently (which means it is hard to prevent) and also that the resistance alleles are able to spread across space - so even if a farmer or community can prevent evolution, resistant plants may arrive from elsewhere. The parts of the paper that deal with the ages of the alleles and genomic interactions of the alleles are not as convincing to me.
-