agronomy Occurrence and Management of PSII-Inhibitor-Resistant Chenopodium album L. in Atlantic Canadian Potato Production

: Potato producers in the Atlantic Canadian provinces of New Brunswick (NB) and Prince Edward Island (PE) rely on the photosystem II-inhibiting herbicide metribuzin for weed management. Recently, potato producers in the region have reported unacceptable common lambsquarters control following an application of metribuzin. Tissue and seed samples were collected from escaped common lambsquarters populations from across the potato producing regions of NB and PE and screened for the Ser264Gly mutation in psbA . Overall, 46% of sampled populations possessed the Ser264Gly mutation across the region. Cross-resistance testing to atrazine, metribuzin and linuron conﬁrmed populations with the Ser264Gly were resistant to triazines and triazinones but remained susceptible to linuron. Dose response analysis determined a moderate level of resistance to metribuzin in common lambsquarters which would not be controlled in producers ﬁelds. A ﬁeld experiment was conducted in Fredericton, NB and Harrington, PE, to determine if currently registered and unregistered products and tank-mixes would control PSII-inhibitor-resistant common lambsquarters in potato. All evaluated products, with the exception of S-metolachlor, provided control equivalent to the weed-free check without compromising potato yield or quality. This study demonstrates that PSII-inhibitor-resistant common lambsquarters are found in Atlantic Canadian potato production systems, but can be controlled with currently registered herbicides and rates with alternative modes of action. in PE16 to PSII-inhibiting herbicides is not consistent with previously published and suggest a possible alternative resistance mechanism. Together, this demonstrates that in addition the well characterized Ser264Gly mutation, common lambsquarters in may have additional mechanisms of resistance. Further studies sequencing psbA investigating additional mechanisms are warranted.


Introduction
The Atlantic Canadian provinces of New Brunswick (NB), and Prince Edward Island (PE), are the largest potato (Solanum tuberosum L.) producing regions in Canada with over 40% of total Canadian potato acreage seeded in these two provinces [1]. Shifts in management from two-pass to one-pass systems in the region have reduced reliance on hilling and cultivation for weed management. In a identify unknown incidence of herbicide resistance [11,12]. All samples were collected from potato fields with visible common lambsquarters escapes. Field history was unknown at time of collection. Leaf samples from 5 plants field −1 were collected from August until September prior to potato topkill. At each field location, a young leaf sample was collected and placed in a paper envelope. Leaf samples were stored in a plastic bag with 10 g of silica gel orange (Sigma-Aldrich, Mississauga, ON, Canada) at 4 °C prior to analysis. Samples were analyzed individually. Leaf samples from 39, 44, and 108 populations were collected in 2017, 2018 and 2019, respectively. Sample locations were GPSreferenced for seed collection (Figure 1). Where possible, seeds were collected from mature common lambsquarters plants in fields which had been previously sampled for tissue analysis prior to potato harvest in September and October. Inflorescences from 20 random plants field −1 were clipped, placed in a bag and hand-threshed. Seeds collected from a single field were bulked. Seed samples from 8, 37, and 3 populations were collected in 2017, 2018 and 2019, respectively.

Identifying Target-Site Mutation
Genomic DNA of common lambsquarters was extracted from fresh leaf tissue (n = 191) using Qiagen DNeasy Plant Mini Kit (Qiagen, Inc., Mississauga, Canada) according to the manufacturer's instructions. A pair of primers and a pair of probes were constructed based on the psbA gene sequence of common lambsquarters to test for the well-known Ser264Gly mutation [13]. PCR primers were 5 -TCATGGTTATTTCGGCCGATTGAT-3 and 5 -GCCAAGCAGCTAAGAAGAAGTGTAA-3 . The probes were 5 -CTTTCAATATGCTAGTTTCAA-3 (VIC; serine(wild type)) and 5 -CAATATGCT GGTTTCAA-3 (FAM; glycine(resistant)). Quantitative PCR was performed on an AriaMx (Agilent, CA, USA).

Single-Dose Assays
Seeds of collected populations (n = 48) were surface-sterilized in a 5% hypochlorite solution (Thermo Fisher Scientific, Mississauga, ON, Canada) with 0.1% Tween (Thermo Fisher Scientific, Mississauga, ON, Canada) for 30 min followed by triple rinsing with dH 2 O prior to planting. Ten seeds population −1 were seeded into square pots (8.5 × 8.5 × 10 cm, L × W × H) containing a 50:50 mix of sterilized Harrington, PE, field soil (Orthic Humo-Ferric Podzol, 3% organic matter, pH 6.6) and sterilized play sand (Quikrete Premium Play Sand, Home Depot, North York, ON, Canada). Greenhouse conditions were 22/16 • C day/night temperatures with a 14/10 h day/night photoperiod. Supplemental lighting was provided by high pressure sodium lights (550 µmol m −2 s −1 ). A single irrigation line to each pot provided 150 mL of water day −1 . One week after emergence, pots were thinned to 3 plants pot −1 and fertilized with 100 mL of commercial fertilizer (N-P-K, 24%-8%-16% w/w; Miracle Grow, Scotts, Marysville, OH, USA). At the two-to three-leaf stage, herbicide screening was conducted as a randomized complete block design using discriminating doses of the following herbicides: metribuzin (Sencor 75DF, Bayer CropScience Inc., Phoenix, AZ, USA) at 750 g a.i. ha −1 , atrazine (Aatrex 480, Syngenta Canada, Guelph, ON, Canada) at 1 kg a.i. ha −1 and linuron (Lorox L, Tessenderlo Kerley Inc., Phoenix, AZ, USA) at 1 kg a.i. ha −1 in addition to an untreated control [14]. Experimental units consisted of 3 pots of 3 plants pot −1 and were replicated thrice. All treatments were applied in a single-track research spray chamber (DeVries Manufacturing, Hollandale, MS, USA) equipped with a single TP8001 EVS (TeeJet Technologies, Springfield, MA, USA) nozzle set to deliver 200 L ha −1 at 210 kPa. Following treatment, plants were returned to the greenhouse and monitored. Three weeks after application (WAA), plants were rated as dead (susceptible) or alive (resistant). Remaining above-ground biomass was harvested, dried to constant moisture at 60 • C and weighed.

Herbicide Dose-Response
Seeds of select populations were surface sterilized as previously described. Soil and greenhouse conditions were as described above for single-dose assays. Number of seeds was adjusted for each population based on previous germination tests to achieve a target density of 10 plants pot −1 . One week after emergence, pots were thinned to 5 plants pot −1 and fertilized with commercial fertilizer (N-P-K, 24%-8%-16% w/w; Miracle-Gro, Scotts, Marysville, OH, USA) as required. When seedlings reached the 2 to 3 leaf stage, two independent dose-response experiments were established as randomized complete blocks. Experimental units consisted of 3 pots of 5 plants pot −1 replicated four times for each experiment. The doses for metribuzin and linuron were as follows: metribuzin 0, 110, 275, 550, 1100, 2200, 4400 and 8800 g a.i. ha −1 and linuron 0, 220.8, 552, 1104, 2208, 4416, 8832 and 17,664 g a.i. ha −1 . Treatments were applied as described above in a single-track research spray chamber. Following treatment, seedlings were returned to the greenhouse and monitored for 6 weeks after application (WAA). Plants were rated weekly on a 0 to 100% scale where 0 is no injury and 100 is plant death. Six WAA, remaining above-ground biomass was harvested and dried to constant weight at 60 • C. and Fredericton (OM 3%, pH 6.2) is classified as a Orthic Humo-Ferric Podzol. The experiment was established as a randomized complete block design with 3 replications and 14 treatments. Prior to planting, fields were disked, cultivated, harrowed twice and then potato rows formed. Fresh-cut and untreated potato (cv. Russet Burbank) sets were hand-planted at 38 cm plant spacing with 0.91 m spaced rows. Fertilizer, according to provincial recommendations, was banded at planting prior to row closure. Rows were closed mechanically following planting. Plots measured 5 m in length and consisted of 2 potato rows and a guard row on either side for a total of 4 potato rows per treatment.

Evaluation of PRE-Emergence Herbicide Options
In 2018, potatoes were planted on 28th May and 5th June and in 2019 on 29th May and 5th June in Harrington and Fredericton, respectively. Plots were hilled approximately 4 weeks after planting when potatoes reached the ground crack stage. Hilling completely covered emerged potato shoots. Treatments were applied immediately after hilling on 27th June in 2018 and 25th June in 2019 in both Harrington and Fredericton. Treatments were applied with a CO 2 -pressurized backpack sprayer equipped with a 1.5 m boom and 4, TP8002 vs. (TeeJet Technologies, Springfield, MA, USA) nozzles spaced 50 cm apart set to deliver 200 L ha −1 at 210 kPa. Treatments and rates are listed in Table 1, a weed-free and weedy check were also included. Plots were maintained weed-free by hand-weeding. Visual estimates of potato phytotoxicity were collected 4 and 8 WAA on a 0 to 100% scale where a rating of 0 was defined as no injury and 100 was defined as plant death. Common lambsquarters control was rated 4 and 8 WAA on a 0 to 100% scale where a rating of 0 was defined as no injury and 100 was defined as total common lambsquarters death. At 4 and 8 WAA, common lambsquarters density and biomass was determined in each plot through counting all common lambsquarters and other weed species within a 0.5 m 2 quadrat and cutting at the soil surface. Common lambsquarters were dried to constant moisture at 60 • C and weighed.
Potatoes were top-killed with diquat (300 g a.i. ha −1 ; Reglone, diquat ion 240 g a.i. L −1 , (Syngenta Canada, Guelph, ON, Canada) on September 24th and September 20th in 2018 and October 1st and September 17th in 2019 at Harrington and Fredericton, respectively. At physiological maturity, one row per treatment was mechanically harvested. Potatoes were graded with an optical grader (Celox-P-UHD, New Tec, Odense, Denmark) into small (3.8-5 cm) and Canada #1 (5-8.9 cm) size classes. Smaller (<3.8 cm), larger (>8.9 cm), misshapen and sunburnt potatoes were removed and counted as culls. Marketable yield was calculated as the sum of small and Canada #1 yield. Specific gravity was calculated using the weight-in-water and weight-in-air method with a 5 kg sample of marketable tubers.

Statistical Analysis
All statistical analysis was carried out in SAS v9.4 (SAS Institute, Cary, NC, USA). Data from dose-response experiments was analyzed with Proc NLIN in SAS using untransformed biomass data. Metribuzin and linuron data were fit with a 3-parameter log-logistic function (1) where d is the above-ground biomass when dose = 0, LD 50 is the herbicide dose providing 50% reduction in biomass and b is the slope of the curve at dose = LD 50 . Goodness-of-fit was evaluated using mean square error. (1) Data collected from the discriminating dose and the PRE-herbicide evaluation experiment were analyzed with generalized linear mixed effects models in SAS using Proc Glimmix and a Gaussian distribution. Fixed effects were population, herbicide treatment (products or doses) and their interactions and random effects were replication, year, location and their interactions. In-season data (potato phytotoxicity, common lambsquarters and weed visual ratings, density and biomass), were analyzed with repeated measures through time. Weed-free and weedy control were removed from analyses of visual ratings. Residual analysis was conducted using proc univariate to confirm ANOVA assumptions. Significance of covariance parameters was investigated with a Wald's test. No significant effect of year, location or their interaction was found, therefore, data was pooled for analysis. In-season data least square means were compared with Tukey's HSD (α = 0.05). Yield was compared to weedy check using linear contrast statements with a Type I error rate of α = 0.05.

PSII-Inhibitor-Resistant Common Lambsquarters Is Widespread across Atlantic Canada
Common lambsquarters populations possessing the Ser264Gly mutation were found across all sampled potato-producing regions of NB and PE demonstrating widespread incidence of PSII-inhibitor-resistance (Figure 1). Overall, 46% of sampled populations possessed the Ser264Gly mutation with a higher prevalence in sampled populations found in PE (53%) than NB (42%) ( Table 2). Where seed was available, single-dose herbicide assays confirmed all populations possessing the Ser264Gly mutation were resistant to atrazine and metribuzin, but not linuron ( Table 2). In addition, while fewer populations were sampled using single-dose assays compared to those with molecular markers, the relative percentage of resistant to wild-type populations across the region appeared to be similar between methods. These results are in confirmation with previous findings demonstrating the Ser264Gly mutation confers a high and moderate level of resistance to triazine and triazinone herbicides, respectively, but not other PSII-inhibitors [6]. Table 2. Results of single-dose herbicide assays, presence of serine264-glycine (Ser264Gly) mutation and specific cross resistance patterns for screened New Brunswick (NB) and Prince Edward Island (PE) common lambsquarters populations.

Location
Populations Screened The resistance index of collected common lambsquarters populations to metribuzin was between 127 and 263 ( Table 3). The metribuzin LD 50 of the susceptible population was 0.003 kg ha −1 whereas the LD 50 of resistant populations ranged from 0.38 to 0.79 kg ha −1 . All populations were susceptible to linuron and controlled with the lowest dose tested (0.22 kg ha −1 ) ( Table 3). The estimated LD 50 of populations to linuron ranged from 0.005 to 0.04 kg ha −1 . Estimated metribuzin LD 90 values, which give an indication of acceptable control in a producer field, for PSII-inhibitor-resistant common lambsquarters populations ranged from 1.02 to 4.08 kg ha −1 . Given that typical use rates of metribuzin in NB and PE potato production are 0.42 to 0.51 kg ha −1 and 0.56 to 0.62 kg ha −1 for one-and two-pass systems, respectively; no tested common lambsquarters populations would be controlled in producer fields. All resistant populations identified via dose-response possessed the Ser264Gly mutation, whereas, it was lacking in the susceptible population. The Ser264Gly mutation maintains susceptibility to non-triazine PSII-inhibiting herbicides such as linuron [6]. Indeed, linuron controlled all populations at all tested doses. Together, these results provide further support that a Ser264Gly mutation is a mechanism conferring resistance to PSII-inhibitors in Atlantic Canada. Table 3. Nonlinear regression parameter estimates ± SE for dose-response to metribuzin and linuron for susceptible and resistant common lambsquarters populations from New Brunswick (NB) and Prince Edward Island (PE) potato fields. Estimated LD 90 (kg ha −1 ) is also included. The use of molecular markers can greatly accelerate the speed and ease at which resistance-testing is conducted [15], yet several caveats remain. Indeed, while all susceptible populations did not possess the Ser264Gly mutation, several populations which lacked Ser264Gly displayed resistance to one or more of atrazine, metribuzin or linuron in the single-dose assay (Table 4). For example, two populations from NB (NB6, NB7) were found to lack the Ser264Gly mutation and were not controlled by 1 kg ha −1 of atrazine. These populations were controlled with 750 g ha −1 of metribuzin and 1 kg ha −1 of linuron. One population from PE (PE16) was not controlled with 1 kg ha −1 of atrazine nor 750 g ha −1 of metribuzin but was controlled by 1 kg ha −1 of linuron. Further, one population from NB (NB5) was not controlled by 1 kg ha −1 of atrazine or 1 kg ha −1 of linuron. Several psbA target-site mutations were identified which confer cross-resistance patterns consistent with our results. Populations NB6 and NB7 displayed low, high and low levels of resistance for metribuzin, atrazine and linuron, respectively (Table 4). This pattern is similar to the response of PSII-inhibitor-resistant Amaranthus retroflexus L., which possess a Phe274Val substitution [16]. The high level of resistance to metribuzin, atrazine and linuron found in NB5 is similar to A. retroflexus in Ontario with a Ala251Val substitution [16]. Other reports, however, found Ala251Val to confer low level resistance to the triazinone metamitron and remain susceptible to atrazine and linuron in Swedish C. album [17]. The moderate and low levels of resistance found in PE16 to PSII-inhibiting herbicides is not consistent with previously published mutations and suggest a possible alternative resistance mechanism. Together, this demonstrates that in addition to the well characterized Ser264Gly mutation, common lambsquarters populations in Atlantic Canada may have additional mechanisms of resistance. Further studies sequencing psbA investigating additional mechanisms are warranted. Table 4. Cross resistance patterns for populations lacking the Ser264Gly mutation. Classification in cross resistance pattern based off % biomass of treated compared to control: S (susceptible)-<25%; L (low)-25-50%; M (M)-51-75%; H (high)->76%.

Potato Yield Was Not Impacted by Preemergent Herbicides
Marketable potato yields of all treatments were greater than the weedy control and yield parameters of several treatments significantly differed from weedy check ( Table 6). Total tuber yield in the dimethenamid-P treatments (36.24 T ha −1 ) was significantly higher than weedy check (p = 0.025) and was the highest average total yield. Similarly, marketable tuber yield (33.98 T ha −1 ) and Canada #1 tuber yield (22.73 T ha −1 ) in dimethenamid-P treatments was significantly higher than weedy check (p = 0.012 and p = 0.013, respectively). This is consistent with previous reports of potato tolerance to dimethenamid-P up to rates of 2.9 kg ha −1 on low organic matter soils [26]. Marketable yield of metribuzin plus sulfentrazone treatments was significantly higher than the weedy check (p = 0.0156). Potato has a high tolerance to sulfentrazone with no negative impact on yield seen with rates up to 2.8 kg ha −1 [19]. In addition to control of PSII-inhibitor-resistant species, sulfentrazone in tank mixture with metribuzin effectively controls several other problematic weeds in Atlantic Canadian potato production including A. retroflexus [20,27,28], Solanum sarrachoides Sendtn. [20,27,28], Echinochloa crus-galli (L.) P. Beauv. [27] and Digitaria sanguinalis (L.) Scop [19]. Metribuzin plus sulfentrazone was recently registered for use in Atlantic Canadian potato production [29] and as such may be an effective alternative mode of action for potato producers.

Conclusions
Our results demonstrate PSII-inhibitor-resistant common lambsquarters is widespread across the potato-producing regions of New Brunswick and Prince Edward Island. In the majority of populations surveyed, PSII-inhibitor-resistance is conferred by the well characterized Ser264Gly mutation granting resistance to metribuzin and maintaining susceptibility to linuron. Results of screening with atrazine, metribuzin and linuron indicate additional mechanisms of resistance may be present in the region. Currently registered and unregistered preemergent herbicide active ingredients and tank mixtures effectively controlled PSII-inhibitor-resistant common lambsquarters without compromising potato yield or quality. To delay resistance, producers should focus on rotating herbicide modes of action and implementation of additional integrated weed management strategies. Weeds escaping herbicide treatment must be controlled to prevent proliferation of resistant populations [30], however, limited postemergent management options exist for potato. Future studies should focus on development of alternative management strategies such as herbicide tank-mixtures, crop rotations and cultivation for pre-and postemergent weed control in potato to delay the onset of further herbicide resistance and ensure long-term sustainability of Atlantic Canadian potato production.