Introduction
The first case of herbicide resistance was reported in 1957, and with the heavy reliance on herbicides since then, the number of unique herbicide-resistant weed biotypes is now well over 500 (Heap Reference Heap2025; Switzer Reference Switzer1957). Weed management is becoming more complex as weeds continue to develop resistance to one or more herbicides. The detection of herbicide-resistant weeds is a key step toward adopting nonchemical weed management practices aimed at mitigating the adverse effects of these weeds (Burgos et al. Reference Burgos, Tranel, Streibig, Davis, Shaner, Norsworthy and Ritz2013). The traditional method to confirm herbicide resistance in a weed is laborious and time-consuming and requires the generation of a dose response to the herbicide in question (Burgos et al. Reference Burgos, Tranel, Streibig, Davis, Shaner, Norsworthy and Ritz2013). Molecular methods are much quicker, although each assay generally targets only a single mechanism and at times only a single mutation conferring resistance (Burgos et al. Reference Burgos, Tranel, Streibig, Davis, Shaner, Norsworthy and Ritz2013). Nevertheless, there remains a paucity in the infrastructure and technical expertise needed to routinely conduct these molecular tests.
A large proportion of herbicide-resistance cases belong to the acetolactate synthase (ALS)-inhibiting herbicides (Heap Reference Heap2025). Out of the 500-plus unique cases reported by the International Survey of Herbicide-Resistant Weeds, more than 160 are from this group. The protein coded by the ALS gene is the target of molecules from this group of herbicides (Gerwick et al., Reference Gerwick, Subramanian and Loney-Gallant1990; LaRossa and Schloss Reference LaRossa and Schloss1984; Shaner et al. Reference Shaner, Anderson and Stidham1984). ALS inhibitors include five structurally diverse chemical classes: sulfonylureas, imidazolinones, triazolopyrimidines, pyrimidinylthio (oroxy)-benzoates, and sulfonylamino-carbonyltriazolinones (Zhou et al. Reference Zhou, Liu, Zhang and Liu2007). Several mutations in the ALS gene have been associated with resistance to one or several of these chemical classes (Heap Reference Heap2025).
Molecular methods to identify ALS inhibitor target-site resistance (TSR) are quicker and easier to perform than the classical spray test where the appropriate genetic test is available. Even if a rapid diagnostic tool is not readily available, Sanger sequencing of PCR amplicons can rapidly reveal the presence of resistance-conferring mutations, as several residues (eight in the case of ALS) have been reported to confer resistance when mutated (reviewed in Beckie and Tardif Reference Beckie and Tardif2012). Known mutations associated with resistance to these herbicides abound, and six have been previously reported in redroot pigweed (Amaranthus retroflexus L.): Ala-122-Thr (McNaughton et al. Reference McNaughton, Letarte, Lee and Tardif2005), Pro-197-Leu (Sibony et al. Reference Sibony, Michel, Haas, Rubin and Hurle2004), Ala-205-Val (McNaughton et al. Reference McNaughton, Letarte, Lee and Tardif2005), Asp-376-Glu (Huang et al. Reference Huang, Chen, Zhang, Huang, Wei, Zhou, Chen and Wang2016), Trp-574-Leu (McNaughton et al. Reference McNaughton, Letarte, Lee and Tardif2005; Scarabel et al. Reference Scarabel, Varotto and Sattin2007), and Ser-653-Thr (Chen et al. Reference Chen, Huang, Zhang, Huang, Jingchao and Xu Wang2015; McNaughton Reference McNaughton2001). The amino acid changes Pro-197-Leu and Pro-197-His consistently (in all 22 cases) confer resistance to sulfonylureas (Tranel et al. Reference Tranel, Wright and Heap2019). Trp-574-Leu is widely thought to provide broad-spectrum resistance to ALS-inhibiting herbicides. Indeed, no susceptibility to any ALS-inhibiting herbicides was identified in the 37 Trp-574-Leu cases reported (Tranel et al. Reference Tranel, Wright and Heap2019). Mutations Ala-122-Thr and Ser-653-Thr generally confer resistance to imidazolinone but not sulfonylureas (Tranel and Wright Reference Tranel and Wright2002). Huang et al. (Reference Huang, Huang, Chen, Chen, Wei and Zhang2019) recently characterized an ALS inhibitor–resistant A. retroflexus biotype harboring a Ser-653-Asn substitution, and this biotype survived the application of nicosulfuron and showed a high resistance index (RI) of 14.50. The RI is the ratio of the herbicide doses that provide a 50% reduction in growth between the resistant and the susceptible (control) populations.
In Canada, the first A. retroflexus biotypes resistant to ALS inhibitors were found in a soybean [Glycine max (L.) Merr.] and a corn (Zea mays L.) field in 1997 in Ontario. These biotypes were resistant to both imazethapyr and thifensulfuron-methyl (Ferguson et al. Reference Ferguson, Hamil and Tardif2001). In 2005, additional A. retroflexus populations were found in Ontario with the same resistance pattern. Three specific amino acid substitutions in the ALS gene—Ala-122-Thr, Ala-205-Val, and Trp-574-Leu—were found to confer resistance in these A. retroflexus populations (McNaughton et al. Reference McNaughton, Letarte, Lee and Tardif2005). In Manitoba, the first population of A. retroflexus resistant to the ALS inhibitor florasulam was documented in a wheat (Triticum aestivum L.) field in 2002 (Beckie et al. Reference Beckie, Leeson, Thomas, Brenzil, Hall, Holzgang, Lozinski and Shirriff2008). In Quebec, the first A. retroflexus population resistant to ALS inhibitors (imazethapyr) was found in 2009 in a soybean field (Heap Reference Heap2025). In Saskatchewan, the first ALS-resistant A. retroflexus showed resistance to thifensulfuron-methyl and tribenuron-methyl in a wheat field in 2010 (Heap Reference Heap2025). The mechanisms conferring ALS resistance in the Manitoba, Quebec, and Saskatchewan biotypes were not identified at the time.
Resistance to ALS-inhibiting herbicides can also be caused by non–target site resistance (NTSR) mechanisms such as increased herbicide metabolism, reduced translocation, or sequestration (Jugulam and Shyam Reference Jugulam and Shyam2019). NTSR to ALS inhibitors can co-occur with TSR (Bai et al. Reference Bai, Zhang, Li, Wang, Wang, Wang, Liu and Bai2019) or as a unique resistance mechanism (Yu and Powles Reference Yu and Powles2014). Enhanced metabolism of ALS inhibitors is predominantly caused by cytochrome P450–dependent monooxygenase (P450) enzymes and has been documented in common waterhemp [Amaranthus tuberculatus (Moq.) Sauer] (Shegrill et al. Reference Shergill, Bish, Juglam and Bradley2018) and Palmer amaranth (Amaranthus palmeri S. Watson) (Nakka et al. Reference Nakka, Thompson, Peterson and Juglam2017), which are both closely related to A. retroflexus, and numerous unrelated species. Other NTSR mechanisms to ALS inhibitors include glutathione S-transferases (GSTs), glycosyl-transferases, and ATP-binding cassette transporters (Jugulam and Shyam Reference Jugulam and Shyam2019). Malathion is a known P450 enzyme inhibitor of some, but not all, P450 enzymes and, when effective, may revert a resistant weed biotype using this NTSR mechanism to a more susceptible phenotype (Jugulam and Shyam Reference Jugulam and Shyam2019).
Following grower concerns about a lack of response in A. retroflexus to ALS-inhibiting herbicides, seeds from suspected resistant biotypes were collected from three independent fields in Manitoba (MB), Canada, in 2012, and one biotype was collected in the summer of 2016 in an identity-preserved soybean field located in Saint-Louis-de-Gonzague (QC, Canada). The suspect Manitoba A. retroflexus populations survived in-crop applications of imidazolinone herbicides, while the Quebec population suspected of resistance to ALS-inhibiting herbicides survived preemergence applications of two ALS-inhibiting herbicides from different families (imazethapyr and chlorimuron-ethyl). We have characterized these biotypes with the goal of identifying the level of resistance using dose–response experiments and the mechanism of resistance with sequencing of the ALS gene and NTSR testing. We report similar results from two independent studies conducted in different regions of Cananda.
Materials and Methods
Plant Material
The populations, one from Quebec and three from Manitoba, were investigated independently in separate studies within their respective regions. In Quebec, seeds from A. retroflexus plants that escaped control (ArQc-R) were harvested in 2016 in a field located in Saint-Louis-de-Gonzague (QC, Canada). This field had been treated with a preemergence application of imazethapyr and chlorimuron-ethyl, before identity-preserved soybean were sown. Seeds for the A. retroflexus susceptible control against which the Quebec population was tested were collected in a soybean field in 2015 (ArQc-S) and confirmed susceptible in a discriminatory-dose experiment. Seeds were stored at 4 to 6 C in the dark until they were sown. The seeds were sown in Pro-Mix® soil (Premier Tech, Rivière-du-Loup, QC, Canada) and placed in a growth chamber (Controlled Environments, Winnipeg, MB, Canada) with a thermoperiod of 35/30 C, a photoperiod of 16 h, and 70% relative humidity. Once germinated, seedlings were transferred in the greenhouse. The photoperiod was also set to 16 h, and day/night temperatures were 25/20 C.
Seeds of the suspect Manitoba populations (ArMB1, ArMB2, and ArMB3) were collected in 2012 in three independent fields. The specific herbicide use patterns of the three fields are not known; however, imidazolinone and sulfonylurea herbicides are common in-crop herbicides used in Manitoba to manage acetyl-CoA carboxylase inhibitor–resistant wild oat (Avena fatua L.) populations and many other weeds in several crops. Seeds of the susceptible A. retroflexus populations (ArSC1 and ArSC2) to which the suspected resistant populations were compared were collected at the Ian Morrison Research farm at Carman, MB, and confirmed susceptible with a discriminatory-dose (field rate) experiment. All seeds were stored in a refrigerator until experimentation once they arrived in the lab. Storage conditions of the suspected resistant populations between sampling and arrival in the lab are not known. Seeds of ArMB1, ArMB2, ArMB3, ArSC1, and ArSC2 were planted in small pots (10.2 by 7.6 by 10.2 cm) containing a mixture of 40% topsoil, 40% sand, 20% peat, and Sunshine-Mix (Sun Gro® Horticulture, Canada, Vancouver, BC, Canada) fertilized with 11 kg N ha−1 and 53 kg P ha−1 and thinned to 1 seedling per pot shortly after seedling recruitment. The pots were kept in a growth room at 25/20 C and 16/8-h day/night at a relative humidity of 75%.
DNA Extraction and ALS Gene Amplification and Sequencing
For the Québec populations, 10 plants were grown for the purpose of DNA extractions and ALS gene sequencing as described later. For the populations from Manitoba, plant genomic DNA from 8 to 12 individuals obtained from the low doses of the dose–response experiment from all A. retroflexus populations was extracted using Qiagen DNeasy Plant Mini Kit (Qiagen, Mississauga, ON, Canada) according to the manufacturer’s instructions. DNA quality and quantity were determined using electrophoresis and a spectrophotometer, respectively. Primers (Table 1) were designed using the A. retroflexus ALS gene sequence (AF363369.1) available in GenBank (Benson et al. Reference Benson, Karsch-Mizrachi, Lipman, Ostell and Wheeler2005). PCR conditions were as follows: 95 C for 5 min; 35 cycles of denaturation for 1 min at 95 C, annealing for 30 s at a melting temperature specific for each primer, elongation at 68 C for 30 s; 10 min at 68 C; and finally 4 C until sequenced. Amplicon sequencing was performed at the Genome Quebec Innovation Centre (Montreal, QC, Canada). Sequences were aligned and mutations visually inspected using the Staden package (Bonfield et al. Reference Bonfield, Smith and Staden1995).
Primers used for PCR amplification and sequence of ALS gene for the Quebec (QC) and the Manitoba (MB) Amaranthus retroflexus populations.
a Single-letter abbreviations for mixed base positions: R = A, G; S = G, C; W = A, T; Y = C, T.
The ALS gene of the five Manitoba populations was sequenced using a tiling approach on an iSeq 100 Sequencing System (Illumina, San Diego, CA, USA). Fourteen primer pairs (Table 1) were used to ensure adequate coverage of the entire region of interest (McNaughton et al. Reference McNaughton, Letarte, Lee and Tardif2005), including all eight known sites of mutations that confer resistance to ALS-inhibiting herbicides. PCR conditions for amplicon preparation were as follows: 5 min incubation at 95 C; 35 cycles of 0.30 min at 95 C, 0.30 min at the specific annealing temperature for each primer pair (available upon request), and 1 min elongation at 72 C; then 5 min at 72 C and 4 C until sequencing. The PCR products were purified using the AMPure beads XP (Beckman Coulter, Indianapolis, IN, USA) after PCR amplification and quantified using a Qubit Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. The concentration of the PCR products was then standardized for each sample to 50 ng in 60 µl. Then 14 amplicons of each sample were pooled for the labeling PCR reaction using index primers (IDT for Illumina DNA/RNA UD Indexes–Set A, Illumina). Labeling and sequencing were performed as per the manufacturer’s protocol. The resulting gene sequences were analyzed using the Illumina Local Run Manager (LRM) software (Resequencing module) (Illumina). Quality metrics were determined using Sequencing Analysis Viewer (SAV) software (Illumina), and Integrative Genomics Viewer (IGV) (Thorvaldsdóttir et al. Reference Thorvaldsdóttir, Robinson and Mesirov2012) software was used to visualize and compare the gene sequences of the five populations. Position numbering was based on the alignment with the Arabidopsis thaliana L. gene AY124092.
Resistance Evaluation and Dose Response
Following the identification of the mutation by sequencing, dose–response experiments for the Quebec population were performed in a greenhouse at Agriculture and Agri-Food Canada’s Saint-Jean-sur-Richelieu Research and Development Centre. For each biotype, the susceptible (population ArQc-S) and the putative resistant (ArQc-R), 18 individuals (= 2 repetitions of 9 individuals per treatment) were sprayed with each dose of the herbicides in a random complete block design. This entire experiment was repeated 3 times for a total of 54 plants per treatment. Plants were grown in the conditions described earlier. Two weeks after seeding, at the 4- to 6-leaf stage, plants from the susceptible and suspected resistant biotypes were sprayed using a DeVries Manufacturing (Hollandale, MN, USA) moving-nozzle cabinet sprayer equipped with a 8001E-VS even-banding nozzle calibrated to deliver 164 L ha−1 of spray solution at 207 kPa. Doses of 0, 0.25, 0.5, 1, 2, and 4 times the recommended doses of each herbicide were sprayed. The recommended doses of imazethapyr (imidazolinone family, Pursuit® 240, BASF Canada, Mississauga, ON, Canada), chlorimuron-ethyl (sulfonylurea family, Classic 25 DF, DuPont Canada, Mississauga, ON, Canada) flumetsulam (triazolopyrimidine family, Broadstrike, Dow Agrosciences Canada, Calgary, AB, Canada), and nicosulfuron (sulfonylurea family, Accent 75 DF, DuPont Canada) were 100.8, 36, 87.5 and 60 g ai ha−1, respectively. Shoot biomass was collected 32 d after treatment (DAT) by clipping plants at the soil surface and weighing them after they were dried for 4 d at 70 C.
Dose–response curves for the Manitoba populations were generated for the two most commonly used ALS-inhibiting families in that region. Imazethapyr herbicide (Pursuit® 240, BASF Canada) from the imidazolinone family or thifensulfuron-methyl herbicide (Pinnacle SG, DuPont Canada) from the sulfonylurea family were applied at doses of 0, 0.01, 0.1, 1, 10, 100, 1,000, 2,500, 5,000, and 10,000 g ai ha−1 to 5 replicates per treatment using a spray cabinet (Pesticide Spray Chamber, model 2, Agassiz Scientific, Saskatoon, SK, Canada) equipped with a TeeJet® 80015VS nozzle (Spraying Systems, Wheaton, IL, USA) calibrated to deliver 175 L ha−1 at 241.3 kPa. The recommended field doses for postemergence application of imazethapyr and thifensulfuron-methyl in Manitoba are 50 g ai ha−1 and 9.8 g ai ha−1, respectively. Shoot dry matter was determined 21 DAT. The experiment was repeated 4 times for each active ingredient.
NTSR Experiment
To determine the presence of NTSR in the three suspected resistant Manitoba A. retroflexus populations, seedlings from all Manitoba populations were treated with malathion before treatment with the ALS-inhibiting active ingredients. Seedlings were treated with a dose of 2,000 g ai ha−1 malathion (Hi-Yield Chemical, Bonham, TX, USA) 1 h before treatment (Ma et al. 2013) with the herbicides imazethapyr or thifensulfuron-methyl using the same spray conditions as noted earlier. Following this, 14 treatments were established that were composed of three doses (10, 100, and 1,000 g ai ha−1) of imazethapyr or thifensulfuron-methyl with malathion or without malathion, a control treatment of malathion alone (2,000 g ai ha−1), and a control treatment of water alone. This experiment was conducted in conjunction with the dose–response assay, and all treatments were applied on the same day. Shoot dry matter was determined at 21 DAT. The occurrence of NTSR was not investigated in the Quebec population.
Statistical Analysis
Statistical analysis was performed using SAS v. 15.3 (SAS Institute, Cary, NC, USA). Nonlinear mixed-model regression analysis was performed on percent of shoot dry weight (measured as percentage of the untreated control) in response to herbicide doses by using PROC NLMIXED in SAS (Seefeldt Reference Seefeldt, Jensen and Fuerst1995). Dry matter data were modeled to a four-parameter log-logistic function (Equation 1) (Bowley Reference Bowley2008; Seefeldt et al. Reference Seefeldt, Jensen and Fuerst1995).
$$Y = C + {{{D - C}}\over{{1 + {\rm{exp}}\left[ {b(\log \left( x \right) - {\rm{log}}\left( {{\rm{G}}{{\rm{R}}_{50}}} \right)} \right]}}}$$
where Y is the percent of shoot dry weight, x is the herbicide dose, C is the lower limit, D is the upper limit, GR50 is the dose required to reduce plant dry weight by 50% between the upper and lower limits, and b is the slope of the curve at the inflection point. In this experiment, the herbicide treatments were considered as fixed effects, and treatment replication within experimental replication (run) and experimental replication were considered random effects. The negative log likelihood of initial parameters was optimized following Coffey (Reference Coffey2016). To achieve convergence of the procedure, two strategies were followed as necessary: (1) a bounds statement was invoked to keep variance estimates greater than or equal to zero, and (2) the criterion of the relative gradient convergence was set to zero (Kiernan et al. Reference Kiernan, Tao and Gibbs2012).
Single degree-of-freedom estimates were used to compare the model parameters. The results of this were converted to letter separation at a significance level of 5% (α = 0.05). To determine the level of resistance (resistance factors), the GR50 value of the resistant population was divided by the GR50 value of the relevant susceptible reference population (no. 130, ArSC1/ArSC2).
To determine whether a P450-mediated NTSR mechanism was contributing to ALS resistance among the Manitoba A. retroflexus populations, relative biomass was expressed as the ratio of treatment biomass compared with the water control treatment. Data from this experiment were subjected to ANOVA using PROC MIXED. Assumptions of ANOVA were tested, including that residuals followed the gaussian distribution, that influential values were present, and that heterogeneity of variance occurred among treatments (Kozak and Piepho Reference Kozak and Piepho2018). Means were separated using Fisher’s protected LSD at a significance level of 5% (α = 0.05) to compare the mean biomass reduction of the herbicide treatment (imazethapyr or thifensulfuron-methyl) with or without malathion.
Results and Discussion
The molecular-first approach to examine potential herbicide-resistant biotypes was selected for these new A. retroflexus biotypes suspected of resistance to ALS-inhibiting herbicides collected in southern Quebec and southern Manitoba. At both locations, the results showed a G to A substitution of the second nucleotide of the serine codon at position 653, coding for asparagine (Ser-653-Asn) in its mutated form (Figure 1). The same mutation has been reported to confer resistance to ALS-inhibiting herbicides in other species, including A. palmeri (Molin et al. Reference Molin, Nandula, Wright and Bond2017), A. tuberculatus (Patzoldt and Tranel Reference Patzoldt and Tranel2007), green foxtail [Setaria viridis (L.) P. Beauv.] (Laplante et al. Reference Laplante, Rajcan and Tardif2009), false cleavers (Galium spurium L.) (Beckie et al. Reference Beckie, Warwick, Sauder, Kelln and Lozinski2012b), A. fatua (Beckie et al. Reference Beckie, Warwick and Sauder2012a), and downy bromegrass (Bromus tectorum L.) (Kumar and Jha Reference Kumar and Jha2017).
Identification of the mutation Ser-653-Asn (G to A) conferring resistance to imidazolinone herbicides in Amaranthus retroflexus. (A) Quebec population: Upper chromatogram: wild-type sequence showing, highlighted in yellow, the serine codon. Lower chromatogram: mutated sequence showing, highlighted in yellow, the asparagine codon. (B) Manitoba populations: Critical portion of the ALS gene sequence alignment from some samples of the five populations with mutations (in resistant populations ArMB1, ArMB2, and ArMB3) and without mutations (in susceptible populations ArSC1 and ArSC2) generated using Integrative Genomics Viewer (IGV) software. Mutated alleles are A (green) and T (red).
In addition to the Ser-653-Asn mutation, 4 of 12 individuals of the ArMB3 population showed another substitution (Ser-653-Ile) at the same location (G to T) (Figure 1). In S. viridis, this substitution causes low levels of resistance to imidazolinones and pyrimidinyl benzoates and intermediate resistance to sulfonylureas and triazolinones (Laplante et al. Reference Laplante, Rajcan and Tardif2009). This resistance pattern matches our observations in ArMB3 (Figure 2; Table 2). This mutation is much less common than the Ser-653-Asn mutation (Heap Reference Heap2025), and to our knowledge, this is the first report of it in A. retroflexus.
Dose–response curves of ArMB1, ArMB2, ArMB3, ArSC1, and ArSC2 Amaranthus retroflexus populations to imazethapyr (top) and thifensulfuron-methyl (bottom). Data points represent the mean of four experimental runs with five replicates each, and each line represents the best-fitting log-logistic model. The vertical bars represent ±1 standard error of the mean. Different letters below the legend represent significantly different lines. Vertical arrows along the abscissa indicate the field dose (imazethapyr = 50 g ai ha−1; thifensulfuron-methyl = 9.8 g ai ha−1).
Parameter estimates and resistance factors (RF) of dose–response curves of three suspected resistant and two susceptible Amaranthus retroflexus populations to imazethapyr.
a SE is the standard error of the estimate. Fisher’s LSD at a significance level of 5% (α = 0.05) was used to separate the means. When similar letters are shared between means within the column, these means are not significantly different from each other.
b RF is calculated as a ratio of GR50 of a resistant population to GR50 of the susceptible population.
Dose–response experiments confirmed sequencing results and indicated a resistant phenotype for the suspected biotypes when treated with imazethapyr, a herbicide of the imidazolinone family (Quebec populations: RI of 8.55 [Figure 3; Table 3]; Manitoba populations: mean RI = 32.9 to 182.6 [Figure 2; Table 2]). The Quebec biotype with the Ser-653-Asn mutation proved to be susceptible to chlorimuron-ethyl (sulfonylurea) and flumetsulam (triazolopyrimidine) with RIs of 0.6 and 1.1, respectively (Figure 3). An RI of 1.6 was obtained when treated with nicosulfuron (Figure 3). The same conclusions were drawn with respect to the response of the resistant versus susceptible populations at the field dose of the various active ingredients (Figures 2 and 3; Tables 2 and 3). Our results clearly show susceptibility to chlorimuron-ethyl in the Quebec population, and this herbicide, along with imazethapyr, was used in the field where this population was collected. Amaranthus retroflexus requires high soil temperature to germinate and will germinate throughout the summer if soil moisture is adequate (Weaver and McWilliams Reference Weaver and McWilliams1980). Chlorimuron-ethyl has a shorter half-life and soil-residual activity than imazethapyr (Dan et al. Reference Dan, Dan, Barroso, Procópio, Oliveira, Silva, Lima and Feldkircher2010; Van Acker Reference Van Acker2005). In the Quebec field where the resistant population was detected, chlorimuron-ethyl was soil applied as a preemergence herbicide, while imazethapyr was applied postemergence. The dose–response curves for all suspected resistant A retroflexus populations in Manitoba (ArMB1, ArMB2, and ArMB3) also confirmed resistance to imazethapyr in the suspected resistant A. retroflexus populations (Figures 2 and 4; Table 2). All resistant populations showed GR50 values 30 to 194 times greater than those for the two susceptible populations, and differences among the dose–response curves were found only in the GR50 parameters (Table 2). Although tested with a different sulfonylurea (thifensulfuron-methyl), the Manitoba populations with the Ser-653-Asn mutation also showed no resistance to a member of the sulfonylurea ALS inhibitors (Figures 2 and 5). Population ArMB3, on the other hand, showed a low level of resistance to thifensulfuron-methyl (RI = 7.35 relative to ArSC1). Even though both Manitoba control populations were susceptible to thifensulfuron-methyl at the field dose, the ArSC2 control population was much more sensitive to thifensulfuron-methyl (30 times lower GR50 parameter) than ArSC1. Consequently, the low GR50 parameter of ArSC2 resulted in RI values of the suspected resistant populations that would classify these populations as medium to highly resistant to thifensulfuron-methyl. A differential dose response to imazethapyr was not observed between the two susceptible populations (Figures 2 and 4; Table 2). A potential explanation of the different behavior of our susceptible controls to thifensulfuron-methyl could be that this is an example of “creeping resistance” as presented by Gressel (Reference Gressel2009), wherein one of the two susceptible biotypes has started to accumulate minor effect mutations toward eventually developing NTSR.
Dose–response experiment of the Amaranthus retroflexus Quebec biotype suspected of resistance (dark points) and a susceptible control (light points) treated with an imidazolinone (imazethapyr), two sulfonylureas (chlorimuron-ethyl and nicosulfuron), and a triazolopyrimidine (flumetsulam). Data points represent the mean of two experimental runs with nine replicates each, and each line represents the best-fitting log-logistic model. The vertical bars represent ±1 standard error of the mean. Different letters below the legend represent significantly different lines determined via single degree-of-freedom estimates for that panel. Vertical arrows along the abscissa indicate the field dose.
Parameter estimates and resistance factors (RF) of dose–response curves of the Quebec resistant and susceptible Amaranthus retroflexus populations to imazethapyr.
a Fisher’s LSD at a significance level of 5% (α = 0.05) was used to separate the means. Means followed by different letters are significantly different. SE is the standard error of the estimate.
b RF is calculated as a ratio of GR50 of a resistant population to GR50 of the susceptible population.
Response of resistant Amaranthus retroflexus populations ArMB1, ArMB2, and ArMB3 and susceptible control populations ArSC1 and ArSC2 to different doses of imazethapyr at 3 wk after treatment (WAT).
Response of resistant Amaranthus retroflexus populations ArMB1, ArMB2, and ArMB3 and susceptible control populations ArSC1 and ArSC2 to different doses of thifensulfuron-methyl at 3 wk after treatment (WAT).
Our results are different from those presented by Huang et al. (Reference Huang, Huang, Chen, Chen, Wei and Zhang2019). The biotypes from China described by these authors, with an amino acid change at position 653 (Ser-653-Asn), showed resistance to nicosulfuron with an estimated lethal dose of 600 and >1,000 g ai ha−1 for the resistant populations compared with a lethal dose less than 10 g ai ha−1 in the susceptible population. Both the susceptible and the resistant Quebec populations showed lethal doses of less than 10 g ai ha−1 to nicosulfuron (Figure 3), and the dose–response curves were not different from one another. The Canadian biotypes presented in our study with the same mutation were effectively susceptible to all sulfonylurea herbicides tested with lethal doses <10 g ai ha−1 among all populations for all three active ingredients—chlorimuron-ethyl, nicosulfuron, and thifensulfuron-methyl (Figures 2 and 3)—when excluding the hypersensitive susceptible control ArSC2. The differential response to thifensulfuron-methyl, but not imazethapyr, between the Manitoba control populations can lead to very different conclusions with respect to the level of resistance to sulfonylurea herbicides caused by the Ser-653-Asn mutation. Excluding the ArSC2 susceptible control population, our results are in accordance with many studies in which the same mutation in several other species also does not confer resistance to sulfonylureas, including A. thaliana (Chang and Duggleby Reference Chang and Duggleby1998; Sathasivan et al. Reference Sathasivan, Haughn and Murai1991), tobacco (Nicotiana tabacum L.) (Chong and Choi Reference Chong and Choi2000), corn (Zhu et al. Reference Zhu, Paterson, Tagliani, St.Clair, Baszczynski and Bowen1999), and A. tuberculatus (Patzoldt and Tranel Reference Patzoldt and Tranel2007).
Cytochrome P450 metabolism-based NTSR to imazethapyr or thifensulfuron-methyl was not observed in any of our populations where NTSR was tested (all Manitoba biotypes) (Table 4). The dose–response experiments (Figure 2), however, showed that ArSC2 was more sensitive to thifensulfuron-methyl compared with ArSC1. The mechanism of this differential response between the two control populations could not be explained by our sequence (Figure 1) or NTSR data (Table 4), where no evidence of NTSR via P450 monooxygenases that were inhibited by malathion was observed in either susceptible population. The difference in response between the two susceptible populations could be further investigated using other P450 inhibitors such as phorate or piperonyl butoxide (PBO) or the GST inhibitor 4-chloro-7-nitro-1,2,3-benzoxadiazole (NBD-Cl) (Jugulam and Shyam Reference Jugulam and Shyam2019). Alternatively, the differential response to thifensulfuron between the two control populations may be caused by other TSR mechanisms such as altered ALS enzyme activity (Jugulam and Shyam Reference Jugulam and Shyam2019; Tranel and Wright Reference Tranel and Wright2002) or target-site overexpression (Sen et al. Reference Sen, Hamouzova, Mikula, Bharati, Kosnarova, Hamouz, Roy and Soukup2021). The presence of different mechanisms explaining resistance to herbicides of the same mode of action or even the same herbicide is not uncommon among ALS inhibitors (Tranel and Wright Reference Tranel and Wright2002) or other herbicides (Simard et al. Reference Simard, Laforest, Soufiane, Benoit and Tardif2017). As described by Yu and Powles (Reference Yu and Powles2014), selection for resistance sometimes hides unsuspected complexities; herbicides can select for all possible resistance mechanisms that endow survival.
Biomass reduction in Amaranthus retroflexus plants treated with herbicide (imazethapyr or thifensulfuron-methyl) plus malathion compared with plants treated with no malathion in ArMB1, ArMB2, and ArMB3 populations. Within each population, means followed by different letters are significantly different based on Fisher’s protected LSD.
In conclusion, we have identified A. retroflexus biotypes resistant to ALS-inhibiting herbicides and characterized a Ser-653-Asn and a Ser-653-Ile mutation. The Ser-653-Asn biotypes were resistant to imazethapyr (imidazolinone), but susceptible to chlorimuron-ethyl and thifensulfuron-methyl (sulfonylureas) and flumetsulam (triazolopyrimidine). Low levels of herbicide resistance to thifensulfuron-methyl were observed in the three Manitoba biotypes when the RIs were determined using the most sensitive susceptible control biotype only. The biotype with the Ser-653-Ile substitution showed a low level of resistance to thifensulfuron-methyl relative to both control populations. Our results showed differential response among susceptible control populations to thifensulfuron-methyl, which may influence the RI and classification of resistance to sulfonylurea herbicides. The underlying mechanism for the differential sensitivity of the susceptible control populations remains unknown and calls for further studies.
Acknowledgments
The authors gratefully acknowledge the technical assistance of David Girardville for collecting the A. retroflexus biotype; Marie-Josée Simard for providing the susceptible biotype; and Sylvain Fortin, Mélanie Cadieux, Marie Ciotola, and Rebecca Dueck for assistance with the dose–response experiments.
Funding statement
This work was funded by Agriculture and Agri-Food Canada (Projects #J-001324 and #J-001751) and support from Plant Surveillance Initiative labs and the Western Grain Research Foundation to RHG and a University of Manitoba Graduate Fellowship award to SS.
Competing interests
The authors declare no conflicts of interest.
Open access
