Molecular confirmation of resistance to PPO inhibitors in Amaranthus tuberculatus and Amaranthus palmeri, and isolation of the G399A PPO2 substitution in A. palmeri

Abstract

During the 2017 to 2019 growing seasons, samples of waterhemp and Palmer amaranth that had reportedly survived field-rate applications of protoporphyrinogen oxidase (PPO)–inhibiting herbicides were collected from the American Midwest and tested for target-site mutations known at the time to confer resistance. Target-site resistance was identified in nearly all (135 of 145) tested common waterhemp populations but in only 8 of 13 Palmer amaranth populations. Follow-up research on one population of Palmer amaranth (W-8), which tested negative for all such mutations, confirmed it was resistant to lactofen, with a magnitude of resistance comparable to that conferred by the ΔG210 PPO2 mutation. Gene sequences from both isoforms of PPO (PPO1 and PPO2) were compared between W-8 and known PPO inhibitor–sensitive sequence. A glycine-to-alanine substitution at the 399th amino acid position (G399A) of PPO2, recently identified to reduce target-site herbicide sensitivity, was observed in a subset of resistant W-8 plants. Because no missense mutation completely delimited resistant and sensitive sequences, we initially suspected the presence of a secondary, non-target-site resistance mechanism in this population. To isolate G399A, a segregating F₂ population was produced and screened with a delimiting rate of lactofen. χ² goodness-of-fit analysis of dead/alive ratings indicated single-locus inheritance of resistance in the F₂ population, and molecular markers for the W-8 parental PPO2 coding region co-segregated tightly, but not perfectly, with resistance. More research is needed to fully characterize Palmer amaranth PPO inhibitor–resistance mechanisms, which appear to be more diverse than those found in common waterhemp.

Introduction

Protoporphyrinogen oxidase (PPO)–inhibiting herbicides have been used for over 50 yr, primarily to control broadleaf weed species. These herbicides target and inhibit the two isoforms of PPO (PPO1 and PPO2) and ultimately cause accumulation of harmful reactive oxygen species, resulting in their herbicidal activity (Becerril and Duke 1989; Matringe et al. 1989a, 1989b). Characteristics responsible for the popularity of these herbicides include low mammalian toxicity, low use rates, broad herbicidal spectrum, and the residual activity for some (Hao et al. 2011). Usage of these herbicides peaked in the mid-1990s and saw a sharp decline with the introduction of glyphosate-resistant cropping systems (Dayan et al. 2017). As weed species have evolved resistance to glyphosate and other herbicides, PPO-inhibiting herbicides have seen a revitalization in usage (Legleiter et al. 2009; USDA-NASS, 2017).

To date, resistance to PPO inhibitors has been documented in 13 weed species globally (Heap 2020). The study of the first of these species, common waterhemp, revealed that resistance was caused by a three-nucleotide deletion in the coding sequence of PPO2, which resulted in the deletion of a glycine residue (ΔG210) at the 210th amino acid position of the protein sequence (Patzoldt et al. 2006). Since this discovery, the same mutation, substitutions of arginine (R128G and R128M) and of glycine (G399A) within the PPO2 coding sequence, and non-target-site mechanisms have been found to reduce sensitivity to PPO inhibitors in Palmer amaranth (Giacomini et al. 2017; Rangani et al. 2019; Salas et al. 2016; Varanasi et al. 2018). Additionally, a mutation of PPO1 was reported to cause resistance in goosegrass [Eleusine indica (L.) Gaertn.], highlighting the need to look beyond PPO2 to identify target-site resistance mechanisms (Bi et al. 2019).

The role of the G399A substitution of PPO2 in herbicide resistance is still not fully understood; it is generally less common than the other target-site mutations that have been reported (Wu et al. 2020). At first, it was thought that the G399A substitution was unlikely to be responsible for decreased affinity of the PPO2 enzyme for PPO-inhibiting herbicides, because glycine and alanine residues are highly similar. Further study led to the hypothesis that an additional methyl group on the substituted alanine residue introduced steric hindrance, which changed the shape of the binding pocket of PPO2 (Rangani et al. 2019). In vitro assays designed to test the activity of variant PPO2 enzymes in the presence of PPO-inhibiting herbicides showed that the G399A substitution conferred comparable resistance to that from the ΔG210 mutation for diphenyl ether, pyrimidinedione, triazolinone, N-phenylphthalimide, phenylpyrazole, and thiadiazole herbicides, though the presence of the G399A substitution was associated with very low enzymatic activity even in the absence of herbicide (Rangani et al. 2019). Although these in vitro results indicate a role for the G399A substitution in PPO-inhibitor resistance, this substitution was reported in a population that was hypothesized to possess an additional non-target-site resistance mechanism. Thus, the magnitude of whole-plant resistance conferred by the G399A substitution alone is not clear.

During the summer growing seasons of 2017 to 2019, leaf samples of common waterhemp and Palmer amaranth plants that had survived field applications of PPO inhibitors were collected from across the American Midwest and tested for target-site mutations known at the time to confer resistance. The primary objective of this testing was to provide rapid confirmation of resistance for site-specific weed management recommendations. Although our results are not meant as a random survey, they nevertheless provide insights into the distribution of target-site resistance in these species. Follow-up research on a Palmer amaranth population from Douglas County, Kansas (W-8) resulted in the identification of the G399A PPO2 substitution, which was not known at the time this research was initiated. Herein we summarize the results of our molecular testing to confirm resistance and provide evidence for our hypothesis that the G399A substitution may coexist with some secondary resistance mechanism(s) in W-8.

Materials and Methods

Resistance Survey

Leaf samples from Palmer amaranth and common waterhemp plants that had survived field applications of PPO-inhibiting herbicides were collected from 13 and 145 fields, spanning three (Illinois, Nebraska, and Kansas) and nine (Illinois, Nebraska, Kansas, South Dakota, Wisconsin, Michigan, Iowa, Missouri, and Ohio) states, respectively. DNA was extracted from leaf samples following a previously reported protocol (Xin and Chen 2012). To confirm resistance, each sample was used along with a previously reported allele-specific quantitative polymerase chain reaction (qPCR) (TaqMan; Bio-Rad, Hercules, CA) protocol to detect the presence of the ΔG210 mutation (Wuerffel et al. 2015). Additionally, Palmer amaranth samples were tested for R128G and R128M substitutions using analogous allele-specific qPCR protocols (Giacomini et al. 2017). Data were combined over years to document the prevalence and distribution of these three mutations.

Growth Conditions and Herbicide Application

All other plants used in this study were grown in 7.8- by 5.6- by 5.9-cm-deep inserts (one plant per insert) containing a 3:1:1:1 mixture of Sunshine LC1 (Sun Gro Horticulture, Agawam, MA)/soil/peat/torpedo sand plus 13-13-13 Osmocote (The Scotts Company, Marysville, OH). The plants were grown in a greenhouse maintained at 30 C/25 C day/night with artificial lighting from metal halide lamps programmed for a 16-h photoperiod. All herbicide applications included 1% v/v crop oil concentrate and were made on plants at the six- to seven-leaf stage using a compressed-air laboratory spray chamber (DeVries Manufacturing, Hollandale, MN) equipped with an 80° even flat-fan spray nozzle (TeeJet Technologies, Wheaton, IL) calibrated to deliver 187 L ha^–1.

Population Development

Of the populations described in this study, W-8 refers to a PPO inhibitor–resistant population of Palmer amaranth sourced from a soybean field in Douglas County, KS. Seed was obtained from this population during the fall of 2017 after plants that had survived a field-rate application of lactofen tested negative for G210 and R128 PPO2 mutations. Population 7 refers to a PPO inhibitor–sensitive population of Palmer amaranth collected from a soybean field in Chase County, KS in the fall of 2017. dG210 refers to a synthetic population of Palmer amaranth developed by crossing two plants known to be homozygous for the ΔG210 mutation from the KLPA population previously described by Lillie et al. (2019).

To identify PPO inhibitor–resistant individuals within the W-8 population, plants were grown from field-collected seed and treated with 219 g ai ha^–1 of lactofen (Valent USA, Walnut Creek, CA), a rate previously reported to delimit PPO inhibitor–sensitive and -resistant Palmer amaranth at the six- to seven-leaf stage (Lillie et al. 2020). A surviving W-8 plant was selected and used as the female parent in a cross to an individual P-7 plant in a pollen-exclusion tent to produce an F₁ population (RxS). The remaining surviving W-8 plants were allowed to intermate within a pollen-exclusion tent. Seed was subsequently harvested and combined to produce a bulked resistant population (RxR). Plants from the F₁ RxS population were screened with 219 g ai ha^–1 of lactofen, and the least injured male and female plants were used in a pairwise cross in a pollen-exclusion tent to produce an F₂ population (RxS_F₂). Other resistant RxS plants were individually crossed to P-7 individuals in separate pairwise crosses to produce several backcross populations (BC₁).

POST Dose Response

To quantify the level of resistance within the W-8 population, plants from RxR, P-7, and dG210 populations were grown under greenhouse conditions and treated with varying rates of lactofen. Plants from the P-7 population were treated with seven rates of lactofen, evenly spaced along a log_3.162 scale ranging from 0.219 to 219 g ai ha^–1. To better capture the entirety of the dose–response curve and achieve plant injury, these rates were shifted to range from 6.92 to 6,920 g ai ha^–1 in the remaining populations. An adjuvant-only control treatment was also included for each population. Two weeks after lactofen treatment, injury ratings were noted on a scale from 1 to 10, with 1 representing plant mortality and 10 representing no injury, before aboveground biomass was harvested, dried for at least 5 d at 37 C, and weighed for each plant. To account for the plant-to-plant variability in growth rate within and between Palmer amaranth populations, the dry weight of each plant was multiplied by its injury rating to obtain adjusted dry-weight measurements for each plant, which were then expressed relative to adjuvant-only treated plants of the respective population (Guo et al. 2015). Each treatment included five or six biological replicates, and the experiment was conducted twice, although the dG210 population was included in only the second run. The two experimental runs were pooled for the P-7 and RxR populations after no significant run effect was detected with ANOVA.

Adjusted dry-weight measures were used to generate dose–response curves in the drc (Ritz et al. 2015) package in R (version 3.5.2; R Core Team). The four-parameter log-logistic function used to generate dose–response curves is described in Equation 1:

([1]) $$y = c + {{d - c} \over {1 + {\rm{exp}}\left( {b\left( {\log \left( x \right) - \log \left( {E{D_{50}}} \right)} \right)} \right)}}$$

where y is the adjusted dry-weight measure expressed as a percentage of the average such measure of each population’s adjuvant-only control treatment, b is the slope of the curve, c is the lower asymptote, d is the upper asymptote, and ED ₅₀ is the dose required to reduce the response halfway between d and c. Estimates of the ED ₅₀ parameter from each population were compared to test for significant differences using the compParm function within the drc package in R (Ritz et al. 2015).

Target-Site Gene Sequence Comparison

To identify any target-site sequence polymorphisms that delimited resistant plants from W-8 and sensitive plants from P-7, leaf tissue was collected from 12 plants of each population prior to lactofen application at a rate of 219 g ai ha^–1. RNA was extracted from the tissue using a previously described TRI-zol protocol (Rio et al. 2010) of six W-8 plants showing resistance and five P-7 plants showing sensitivity 14 d after lactofen treatment. Following extraction, a 5-μl aliquot (approximate RNA concentration of 50 ng μl^–1) of each sample was used to create a complementary DNA (cDNA) library using the ProtoScript^® II First Strand cDNA Synthesis Kit and the supplied Oligo-dT primers (New England Biolabs, Ipswich, MA). Nucleic acid concentrations of each library were quantified with Nanodrop (Thermo Fisher Scientific, Waltham, MA) and diluted with water to 100 ng μl^–1. To amplify and isolate PPO1 and PPO2 coding sequence, each cDNA library was used as the template in two PCRs, one each to amplify the sequence of isoforms PPO1 and PPO2. Each reaction consisted of 12.3 μl water, 5 μl of 5X Green GoTaq Flexi Buffer (Promega, Madison, WI), 2.5 μl of 25 mM MgCl₂ (Promega), 2 μl of 10 mM equal mixture of deoxyribonucleotide triphosphates (i.e., each dNTP at 2.5 mM; New England Biolabs), 1 μl each of forward and reverse primers diluted to a concentration of 10 μM (Integrated DNA Technologies, Skokie, IL), 1 μl of cDNA template, and 0.2 μl Taq polymerase (Promega). Primer sequences used in these reactions were as follow: PPO1 (forward, 5′-ATTCTACAATGTCCGCCGCA-3′; and reverse, 5′-ACTTGTCTTTGTACTGTGAGAGG-3′); PPO2 (forward, 5′-TCCATTACCCACCTTTCACC-3′; and reverse, 5′-TTACGCGGTCTTCTCATCCAT-3′). Thermocycler settings were as follows: 95 C for 1 min; 35 cycles of 95 C for 15 s, 60 C for 15 s, and 72 C for 1.5 min; and 72 C for 5 min. The products of each reaction were separated using gel electrophoresis on a 1.5% agarose gel at 130 volts for 45 min, and bands of approximately 1,650 and 1,600 base pairs were extracted for PPO1 and PPO2, respectively, using the QIAquick Gel Extraction Kit (QIAGEN, Germantown, MD). The extracted amplicons were used in sequencing reactions with the BigDye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific). Each template used three reactions per target site to span the entire coding sequence; two with either the forward or reverse primer listed above and the third with an internal primer (PPO1, 5′-CCTGGTAAGATTAGGGCTGGT-3′; PPO2, 5′-CGTAAGTCACCTCTGGAATAA-3′). Products of these reactions were sent to the W.M. Keck Center at the Roy J. Carver Biotechnology Center at the University of Illinois, Urbana-Champaign for sequence analysis. The resulting chromatograms were visualized using Geneious (version 11.1.5; Kearse et al. 2012) and inspected to identify sequence variants that delimited W-8 and P-7 plants.

Comparison of Relative Gene Expression

To test for target-site expression differences between resistant and sensitive plants, leaf tissue was collected before lactofen treatment (219 g ai ha^–1) from 12 each of W-8 and P-7 plants separate from the previously mentioned target-site sequencing experiment. This tissue was used to generate cDNA libraries as described above for six plants of each population. The resulting libraries were used in qPCR assays to quantify gene expression levels of PPO1 and PPO2. Primers for PPO1 (forward, 5′-TGCTTACTATGGCGGTTGAC-3′; and reverse, 5′-CCGTAGGTTTGGAAGGAACA-3′) and PPO2 (forward, 5′-TGTGGTTGTCACTGCTCCA-3′; and reverse, 5′-GGGGATAAGAACTCCGAAGC-3′) were designed to span intronic regions, thereby excluding amplification of any genomic DNA contaminants, based on the sequences generated above and other reference PPO sequences of Palmer amaranth and common waterhemp. A previously identified qPCR control gene, elongation factor-1 alpha (EF1-alpha), was used to normalize gene expression values (Nicot et al. 2005). Primer sequences for EF1-alpha (forward, 5′-CACGCTTTGCTTGCTTTCACTCTTG-3′; and reverse, 5′-CGATTTCATCGTACCTAGCCTTGGAGTAC -3′) were designed for Amaranthus spp. and confirmed to have efficiency near 100% (±5%) in Palmer amaranth. Each qPCR mix consisted of 5 μl iTaq Universal SYBR Green Supermix (Bio-Rad, Hercules, CA), 1 μl forward primer and 1 μl reverse primer, each diluted to 10 μM (Integrated DNA Technologies), 1 μl template cDNA, and 2 μl water. A QuantStudio 7 Flex Real-Time PCR System was used with thermoprofile as follows: 95 C for 10 min and 35 cycles of 60 C for 15 s and 95 C for 15 s, immediately followed by a melt curve analysis for quality control of positive samples (Taylor et al. 2010). Three technical replicates were used for each biological replicate to reduce error, and samples showing multiple products in melt curve analysis were discarded from further analysis. Samples were pooled by population, and mean relative expression levels were compared to identify differences in target-site gene expression levels. Additionally, the efficiency of each developed primer set was quantified by testing on a log₁₀ dilution series of one cDNA library to ensure accurate fold-change calculations.

F₂ Segregation

To understand the genetic complexity of resistance present within the F₂ population, 8, 42, 31, and 178 plants of the P-7, RxR, RxS, and F₂ populations, respectively, were treated with lactofen (70 g ai ha^–1). This rate was chosen to delimit the P-7 and RxR populations based on results of dose–response experiments. Plants were recorded as dead or alive 14 d after treatment, and data from the F₂ population were used in a χ² goodness-of-fit test to a 3:1 alive/dead ratio.

Target-site Marker Segregation

Allele-specific PCR primer sets (forward, 5′-CCAGGTTGACAGTGGATTGA-3′, and reverse, 5′-CCCATAGAACACTCATGTAC-3′; forward, 5′-CAGTAGCTGTAATGGGCAAC-3′, and reverse, 5′-CACTAATATTCCACAGAAAAAAGAAT-3′) were designed to detect single-nucleotide polymorphisms that were homozygous in the resistant-parent PPO1 and PPO2 coding sequence, respectively, and not present in the sensitive parent. DNA was extracted from leaf tissue of 20 F₂ and 45 BC₁ plants collected prior to lactofen application (70 and 21.9 g ai ha^–1, respectively) using a previously described protocol (Xin and Chen 2012). The rate used in F₂ screening was chosen based on the results of our dose–response experiment, and a lower rate was chosen for the BC₁ populations based on observations while screening the F₂ population. Each DNA sample was used as the template in two PCR assays, with reagent concentrations the same as those used in target-site sequence amplification. Thermocycler settings were as follows: 95 C for 1 min; 35 cycles of 95 C for 15 s, 61 C for 15 s, and 72 C for 30 s; and 72 C for 5 min. Products were separated on a 1.5% agarose gel as described above, and the presence of an amplicon of ~400 base pairs represented the presence of sequence that had been inherited from the resistant parent. Full-gene sequencing of PPO2 from these F₂ and BC₁ plants as described above was conducted to confirm marker results.

Results and Discussion

Resistance Survey

Of the 145 common waterhemp populations tested, 135 (93%) contained the ΔG210 mutation at some frequency (Table 1). Additionally, the ΔG210 mutation was found in all nine states tested, indicating that this resistance mechanism is widespread and primarily responsible for PPO-inhibitor resistance in common waterhemp in the American Midwest. Conversely, of the 13 Palmer amaranth populations tested, only 8 (62%) contained the ΔG210 mutation (Table 2). The resistance within one additional population could be explained by the presence of the R128G substitution, leaving four populations (31%) without these mutations. Because the G399A substitution had not been reported at the time of this survey, we did not test for it in these populations. Though testing for the G399A substitution may help further explain resistance in some of these populations, the finding that target-site mutations did not explain all PPO inhibitor–resistant Palmer amaranth populations is consistent with previous studies (Copeland et al. 2018; Wu et al. 2020). A heightened proportion of PPO inhibitor–resistant Palmer amaranth populations that are not explained by these target-site mutations may be due to additional undocumented resistance mechanisms or to Palmer amaranth having more innate PPO-inhibitor tolerance than common waterhemp (Lillie et al. 2020). This innate tolerance may have led to applications not being completely effective for Palmer amaranth control and, consequently, sampling of populations that were not truly resistant.

Table 1. Presence of ΔG210 within common waterhemp populations sourced from different states. ^a

^a Abbreviations: KS, Kansas; NE, Nebraska; SD, South Dakota; IA, Iowa; MO, Missouri; IL, Illinois; WI, Wisconsin; MI, Michigan; OH, Ohio.

Table 2. Presence of target-site mutations within Palmer amaranth populations sourced from different states.

^a This population also tested positive for the ΔG210 mutation.

Dose Response

Our dose–response experiment clearly separated the resistant (W-8 and dG210) populations from P-7 (Figure 1). The P-7 population performed similarly to other PPO inhibitor–sensitive Palmer amaranth populations tested previously by Lillie et al. (2020). When ED ₅₀ estimates for each population tested in the dose–response study were compared, there was a significant difference between P-7 (ED ₅₀ = 0.32 g ai ha^–1; standard error = 0.148 g ai ha^–1), and both resistant populations, RxR (ED ₅₀ = 1,100 g ai ha^–1; standard error = 615 g ai ha^–1; P < 0.0001) and dG210 (ED ₅₀ = 330 g ai ha^–1; standard error = 252 g ai ha^–1; P < 0.0001). Additionally, no significant difference was found between the ED ₅₀ estimates of the RxR and dG210 populations (P = 0.50). These dose–response data confirmed the presence of PPO-inhibitor resistance within the W-8 population and indicated that the resistance is of a similar magnitude to that conferred by the ΔG210 mutation.

Figure 1. Dose–response curves generated from populations of Palmer amaranth with (dG210) and without (Sensitive) the ΔG210 mutation, as well as with the G399A substitution and possibly some other resistance mechanism (RxR). Points represent the average measure of adjusted dry weight of all biological replicates at that dose expressed as a percentage of the adjuvant-only control treatment within each population. Lactofen rates were shifted for the herbicide-resistant populations (6.92–6,920 g ai ha^–1) compared to the representative herbicide-sensitive population (0.0219–219 g ai ha^–1) so as to capture more of the response curve. Dose–response curves were fitted using a four-parameter, log-logistic function in R.

Target-Site Gene Sequence Comparison

Full gene sequences of PPO1 and PPO2 generated from resistant W-8 and sensitive P-7 plants are available in Supplementary Appendices S1 and S2, respectively. These sequence data confirm the absence of the ΔG210 and R128 mutations in these two populations. One synonymous substitution within PPO2 and no variants of PPO1 were found to delimit the two populations completely (although several variants were identified between individual plants), leading us to believe that resistance within W-8 is not entirely explained by one target-site variant. We did, however, detect the presence of the G399A substitution in four of the six W-8 plants tested. In both of the two plants lacking the G399A substitution, we saw heterozygosity at nucleotides near this site and within the same sequencing reaction, indicating that both alleles were captured in the sequencing reaction (data not shown). Therefore, these two plants are probably homozygous for the wild type at the G399 locus and have some other resistance mechanism allowing them to survive lactofen application. Additionally, comparison of PPO2 sequence from these two plants and sequence from P-7 individuals failed to identify any variants that delimit the two biotypes, suggesting that the additional resistance is coming from some non-target-site mechanism. Our findings that the G399A substitution coexists with some other resistance mechanism within a population of Palmer amaranth is consistent with a previous report (Rangani et al. 2019). Additionally, we discovered that the W-8 plant used as the resistant parent to generate the RxS populations was heterozygous for G399A (data not shown).

Relative Gene Expression Comparison

Each qPCR primer set was between 100% and 105% efficient, allowing for accurate fold-change calculations when using EF1-alpha as a normalizing transcript (data not shown). Results of qPCR indicated no significant difference in the constitutive expression levels of PPO1 or PPO2 between the resistant W-8 population and sensitive P-7 population (Figure 2). Though it is possible that expression is induced following herbicide application, this seems unlikely to be an effective resistance mechanism given the fast-acting nature of these herbicides. Furthermore, although induced expression of PPO1 was not investigated, expression analysis of PPO2 4 h after treatment revealed no significant difference in expression between R and S plants (data not shown). These results suggest that PPO resistance within W-8 is probably not conferred by increased target-site expression. This is not unexpected, as PPO resistance has not been attributed to increased target-site expression in any species to date. However, this protocol serves as the first reported expression assay for PPO1 and can be used in future studies to test for differential expression of PPO1 within related species that do not have divergent coding sequence.

Figure 2. Mean gene expression levels of protoporphyrinogen oxidase 1 (PPO1, A) and 2 (PPO2, B) in PPO-resistant and PPO-sensitive sample pools prior to lactofen treatment as determined by quantitative PCR assays. Gene expression values for each target site are normalized to that of elongation factor-1 alpha and are expressed as percentages of the resistant pool. Bars represent the range of values used to calculate the mean.

F₂ Segregation

The rate used to delimit sensitive and resistant plants within the F₂ population worked well to delimit the RxR and P-7 populations, killing all P-7 plants tested and only 3 of 42 RxR plants 14 d after treatment. We expected some sensitive plants within the RxR population as a result of including multiple individuals in the seed bulking step, some of which were probably heterozygous for the allele(s) essential for resistance. This results in the infrequent combining of sensitive alleles within individuals of the RxR population. As stated previously, target-site sequencing identified the W-8 plant used as the resistant parent to generate the RxS populations to be heterozygous for G399A. This finding helps explain the 56% survival rate of RxS plants screened with 70 g ai ha^–1 of lactofen. Although utilizing a plant heterozygous for a trait of interest as a parent is not ideal, the selection of resistant RxS plants to use in an F₁ cross ensured the F₂ population was generated from two heterozygous, resistant individuals. Survival rates within the F₂ population are listed in Table 3. Because the P value of the resulting test-statistic of χ² goodness-of-fit analyses was above 0.05, we failed to reject the hypothesis that the population fits a 3:1 alive/dead ratio, consistent with single-locus inheritance of resistance. This finding, along with the survival rate of the F₁ population, constitute strong evidence that there is one major locus contributing to resistance within the F₁ and F₂ populations, that being the G399A substitution within PPO2.

Table 3. Results of χ² goodness-of-fit analysis on alive/dead ratings taken 14 d after lactofen treatment on Palmer amaranth plants of a segregating F₂ population derived from a parental resistant-by-sensitive cross.

Target-Site Marker Segregation

PCR primer sets were designed to detect the presence of PPO1 and PPO2 alleles inherited from the resistant parent used in the RxS cross. These primer sets target synonymous alleles that were homozygous in the resistant parent and absent in the sensitive parent. These primers amplified all samples from the RxS population and not the sensitive parent, confirming the homozygosity of the resistant parent for both markers (although, as stated above, the resistant parent was not homozygous for G399A). Additionally, we saw approximately 75% of F₂ individuals and 50% of BC₁ individuals testing positive for these markers, confirming that the F₁ plants used in the crosses to produce these populations were each heterozygous for these markers (Table 4). We saw no apparent linkage of resistance with PPO1 coding sequence, as evidenced by the majority of dead F₂ plants displaying the PPO1 coding sequence from the resistant parent (Table 4). Conversely, we saw tight (but not perfect) linkage of resistance and PPO2 coding sequence, with nearly all surviving F₂ and BC₁ plants testing positive for the resistant-parent allele of PPO2 (Table 4). Full gene sequencing of PPO2 from these F₂ plants confirmed that marker results correlated with the presence of the G399A substitution (data not shown). A decreased rate of lactofen was used in screening the BC₁ population to ensure that we did not kill any plants that were heterozygous. This precaution seemingly resulted in some sensitive plants surviving lactofen application (Table 4). Observed linkage along with segregation data that support single-locus inheritance of resistance led us to conclude that we have isolated the G399A resistance mechanism from any putative secondary, non-target-site resistance mechanism(s) present within W-8. This isolation allows for the production of a fixed, homozygous line of G399A plants and the study of this resistance mechanism alone in comparison to other identified mechanisms.

Table 4. Segregation of survival following lactofen application, and parental target-site gene sequence originating from the resistant parent of a resistant-by-sensitive cross within F₂ and BC₁ populations of Palmer amaranth. ^a

^a R PPO2+ and R PPO2– represent individuals with or without, respectively, a PPO2 coding sequence from the resistant parent, and PPO1 inheritance is indicated analogously.

The lack of perfect linkage of resistance with PPO2 coding sequence may be attributed to imperfect phenotype assignment, in which an imperfect delimiting rate or plant-to-plant variability may cause a plant’s phenotype to be misidentified. An improper assignment of phenotype may also be responsible for the subset of W-8 plants classified as resistant that did not display the G399A substitution, which would nullify our hypothesis of a secondary resistance mechanism in this population. Additionally, error in genotype assignment may explain the lack of perfect linkage documented in the F₂ population and the resistant W-8 plants that were determined not to possess the G399A substitution. Further study is needed to understand the possibility of a secondary resistance mechanism within W-8. In such a study, a resistant plant from W-8 that does not possess the G399A substitution would be used in a parental cross to a PPO-sensitive plant to generate populations analogous to the ones described in this study. Repetition of the experiments described above with these new populations would confirm that plants without the G399A substitution consistently and heritably survive lactofen applications. If linkage of neither PPO locus with resistance is present, the populations generated by this future study would be sufficient to conduct a genome-wide association study to identify genomic regions associated with this secondary resistance mechanism (Korte and Farlow 2013).

Here we present the first report of a population possessing the G399A substitution directly compared to the ΔG210 mutation in dose–response studies, showing a resistance of approximately equal magnitude. Our results further confirm that mechanism(s) other than the ΔG210 mutation confer high levels of PPO-inhibitor resistance. With the isolation of G399A within the F₂ population reported in this study, it will be possible to examine the effect of this mutation in planta without the unknown effect of possible secondary resistance mechanisms that may be present within W-8. Future studies could characterize aspects of G399A such as the magnitude of resistance, degree of dominance, and possible fitness penalties associated with this mutation. Additionally, whole-plant studies could be used to elucidate cross-resistance conferred by this mutation, supplementing enzyme data provided by Rangani et al. (2019).