Genomic-based epidemiology using EPSPS copy number and associated duplication markers
Glyphosate resistance in kochia was first detected in Kansas in 2007, followed by Colorado and Alberta in 2012, Oklahoma and Montana in 2013, and Texas, Wyoming, Idaho, and Oregon in 2014 (Heap, 2020). We surveyed a set of populations from across western North America for glyphosate resistance (Table 1). Most populations returned the expected phenotype; however, a few populations were designated as suspected GR based on field observations while all nine individuals tested had GS phenotype in our screening assay, resulting in a classification for the population as GS but with potential for it to be heterogeneous for phenotype (containing GR at a low frequency). For example, both KS5S and KS6S were suspected to be GR when sampled from the field but were classified as S by phenotyping (Table 1) and had GR individuals based on EPSPSgene copy number data (Table S1).
We identified three categories of EPSPS -duplication haplotypes, defined as follows: 1) increased EPSPS , Type I and II, and MGE copy numbers that correspond to increased EPSPS copy number (≥ 10) (Central Great Plains); 2) increased EPSPS , no Type I or II, MGE ≥ 10 (Northern Plains); 3) increased EPSPS , no Type I or II, MGE < 10 (Pacific Northwest, north-central Wyoming) (Figure 1). Although all GR individuals from across the continent had increasedEPSPS copy number, kochia in the Pacific Northwest and Northern Plains did not contain the type I and type II repeats associated with the tandem EPSPS duplication in the previously characterized population from Colorado (Figure 1). In contrast, the type I and II repeats were present in the Central Great Plains, and in ratios consistent with those reported by Patterson et al. (2019) (Table S1). In the Central Great Plains, Type I and II always amplified together, and Type I (long repeat) was nearly always present at higher copy number than Type II (short repeat) (Table S1), as in the originally characterized EPSPS repeat structure (Patterson et al., 2019). Some individuals had much higher EPSPS copy number than previously reported; for example, individuals collected in MT had between 20-30 copies with no presence of Type I or II and very high (>60 copies) MGE (Table S1). These individuals may represent independent origins via a different molecular genetic mechanism, which will require additional sequencing to assemble this specific EPSPS duplication haplotype.
All GS kochia samples in the survey had 1 copy of EPSPS , no amplification of Type I or II markers, and 4-6 copies of the MGE (Table S1). The absence of amplification of type I and II markers in GS kochia samples, which requires insertion of the MGE next to duplicated copies of EPSPS , confirms that the MGE is not inserted at the start of the EPSPS locus in any of the diverse susceptible populations sampled and further supports the hypothesis that MGE insertion occurred prior to the EPSPS duplication event, rather than the MGE having been present next to EPSPS ancestrally and co-duplicated withEPSPS (Patterson et al., 2019). MGE copy number was increased to >10 in one GS individual from each of four populations (located in CO, NE, and MT) without corresponding increase inEPSPS gene duplication (Table S1), suggesting copy number of this MGE varies across populations and it may represent an active element, or that the EPSPS duplication locus can segregate away from other MGE duplication sites in the genome.