Plant Material.

Spotted spurge plants were collected from a ‘Sea Isle 1’ seashore paspalum field in Cook County, GA. The location coordinates will not be disclosed to protect the privacy of the landowner. A 27-m² area was sprayed on July 1, 2014, with metsulfuron (Manor 60WG, Nufarm Americas, Burr Ridge, IL) at 84 g ai ha⁻¹, which is eight times the standard use rate. Plants that were not controlled from this treatment were removed from the field by hand on August 12, 2014. Spotted spurge was also collected in Griffin, GA, from a susceptible (S) population. The two biotypes were then transplanted to separate plastic pots of 79-cm² surface areas and 10-cm depths and filled 80 : 20 (v/v) sand : peat moss. Pots were placed in a greenhouse set at 32/25 C day/night temperatures at the University of Georgia Griffin Campus. Approximately 20 plants of each biotype were irrigated as needed to prevent moisture deficiencies. Seeds were planted immediately after collection by hand and scattered over pots of 3.8-cm diam and 20-cm depth with the aforementioned potting medium. Pots were fertigated (MacroN 28-7-14 sprayable fertilizer, LESCO Inc., Cleveland, OH) weekly and allowed to produce three to five branches before treatments. Pots were thinned to single plants before treatments were applied.

Metsulfuron Dose–Response Experiments.

The resistance level of the R biotype was compared with the S biotype in a rate titration of metsulfuron. Treatments were applied in a spray chamber calibrated to deliver 187 L ha⁻¹ with a flat fan nozzle (8002E, TeeJet Spraying Systems Co., Roswell, GA). Metsulfuron-methyl (60%, Alligare LLC, Opelika, AL) was applied at 1.3, 2.6, 5.3, 10.5, 21, 42, 84, 168, or 336 g ha⁻¹. This range was chosen based on standard use rates for spurge control in turfgrass (10.5 to 21 g ha⁻¹). Nontreated checks of the two biotypes were included. A nonionic surfactant (Chem Nut 80-20, mixture of alkyl and alkylaryl polyoxyethylene glycol, 80%, Chem Nut Inc., Albany, GA) was added to the spray solution at 0.25% (v/v). Plants were returned to the greenhouse at ∼ 1 h after treatment and did not receive irrigation until 24 h after treatment. Injury was visually evaluated at 4 wk after treatment (WAT) on a scale of 0 (no injury) to 100% (complete desiccation). Shoot biomass was harvested 4 WAT, oven-dried for 72 h at 60 C, and then weighed.

Multiple and Cross-Resistance Experiment.

In another experiment, the two biotypes were treated with dicamba, glyphosate, metsulfuron, triclopyr, or trifloxysulfuron. A nontreated check was included. Application rates and product information are presented in Table 1. Metsulfuron and trifloxysulfuron were applied with the aforementioned surfactant. Injury was visually evaluated 4 WAT as previously described.

Table 1.

Injury of two spotted spurge biotypes at 4 wk after treatments with six herbicides in two greenhouse experiments at Griffin, GA. Results were pooled over experimental runs.

The design for greenhouse experiments was a randomized complete block with four replications. A block design was chosen to minimize potential vari-ability in greenhouse location on plant responses to treatments. Both experiments were repeated once. Data were subjected to ANOVA with PROC GLM in SAS (SAS v. 9.3, Cary, NC). For plant injury, data were plotted and regressed with an exponential growth function and linear equation to describe the relationship of herbicide rate with plant responses,

Figure 1.

Injury of two spotted spurge biotypes after treatment with metsulfuron in two greenhouse experiments in Griffin, GA. Results were pooled over two experimental runs. Vertical bars represent standard errors of the mean (n = 8). Data for the susceptible (S) biotype was regressed with the following equation: y = β₀{1 − [exp(−β₁x)]}, where y is injury, β₀ is the asymptote, and β₁ is the slope estimate. Data for the resistant (R) biotype was regressed with the following equation: y = β₀ + β₁x, where β₀ and β₁ are the intercept and slope, respectively. Regression for the S biotype: y = 87.68{1 − [exp(−0.23x)]}. Standard errors for β₀ and β₁ measured 3.3 and 0.04, respectively. The 95% confidence intervals for β₀ and β₁ measured 81–94 and 0.14–0.31, respectively. Regression for the R biotype is 4.56 + 0.002x. The standard error for the intercept and slope measured 0.9 and 0.007, respectively. The I₅₀ for the R and S biotypes measured >336 g ha⁻¹ and 3.7 g ha⁻¹, respectively. The 95% confidence interval for the I₅₀ estimate for the S biotype is 2.6–4.9.

Figure 2.

Dry shoot biomass of two spotted spurge biotypes at 4 wk after treatment with metsulfuron in two greenhouse experiments in Griffin, GA. Results were pooled over experimental runs. Vertical bars represent standard errors of the mean (n = 8). Biomass for the nontreated averaged 0.23 (±0.02 SEM) and 0.15 (±0.02 SEM) g plant⁻¹ for the R and S biotypes, respectively. Data were regressed with the following equation: y = β₀ + β₁[exp(−β₂x)], where y is shoot biomass, β₀ is the lower asymptote, β₁ is the maximum predicted response, β₂ is the slope, and x is metsulfuron rate. Regression for the R biotype: 73.63 + 25.29[exp(−0.011x)]. Standard errors for β₀, β₁, and β₂ measured 22.3, 21.8, and 0.03, respectively. The 95% confidence intervals for β₀, β₁, and β₂ measured 29–118, −18 to 69, and −0.04 to 0.06, respectively. Regression for the susceptible biotype: 24.91 + 42.21[exp(−0.21x)]. Standard errors for β₀, β₁, and β₂ measured 4.2, 16.9, and 0.16, respectively. The 95% confidence intervals for β₀, β₁, and β₂ measured 16–33, 9–76, and −0.1 to 0.5, respectively. The GR₅₀ for the R and S biotypes are >336 and 2.5 g ha⁻¹, respectively. The 95% confidence interval for the GR₅₀ of the S biotype is <1.3–7.

ALS Gene Assembly, Mapping, and SNP Detection.

Plants used for gene sequencing were collected from the populations that were seeded for the aforementioned experiments. To identify potential target site mutations, massively parallel sequencing using the Illumina HiSeq (Illumina, San Diego, CA;  http://www.illumina.com/) platform was utilized in lieu of traditional short-read capillary sequencing using polymerase chain reaction. Methodologies for the assembly and polymorphism detection in a nonmodel organism with no reference genome or transcriptome were based on suggestions by Brautigam and Gowik (2010). RNA was extracted using a standard RNA extraction kit (RNeasy Plant Mini Kit, Qiagen, Venlo, The Netherlands;  http://www.qiagen.com). Illumina sequencing, including all RNA preparation steps before sequencing, was conducted at the Hudson Alpha Institute for Biotechnology (Huntsville, AL, USA;  http://gsl.hudsonalpha.org/).

Sequencing reads were processed using the Trin-ity de novo assembly pipeline ( http://trinityrnaseq.sourceforge.net/) (Grabherr et al. 2011; Haas et al. 2013). Reads of the two biotypes were separately paired, trimmed, and de novo assembled using Trin-ity RNA-Seq de novo assembler. Contiguous assembled sequences (contigs) identified as ALS-expressed genes were identified using a local BLAST search. To facilitate the BLAST search, full-length ALS protein sequences were downloaded from the National Center for Biotechnology Information (NCBI;  http://www.ncbi.nlm.nih.gov/). Local BLAST was conducted using tblastn within CLC Genomics Workbench (CLC Bio, Primset, Denmark;  http://www.clcbio.com/). The assemblies of the two biotypes were converted to BLAST databases, and NCBI protein sequences were searched against the assembly databases. Contigs were identified and extracted from each assembly as similar to ALS proteins. Extracted putative ALS contigs were aligned and compared using alignment within CLC Genomics Workbench. The ALS sequences of two biotypes were submitted to NCBI (GenBank KT382543 and KT382544 for R and S biotypes, respectively).

Metsulfuron Dose–Response Experiments.

A biotype-by-rate interaction was detected for injury; thus, results are presented by biotype. The metsulfuron rate that injured the S and R biotypes 50% measured 3.7 and >336 g ha⁻¹, respectively (Figure 1). The R biotype was injured <8% from all application rates, but ≥42 g ha⁻¹ injured the S biotype >87%. Similarly, a metsulfuron rate that reduced biomass of the S and R biotypes by 50% from the nontreated measured 2.5 and >336 g ha⁻¹, respectively (Figure 2).

The resistance factor for this spotted spurge biotype is >90-fold greater than the S biotype. Trezzi et al. (2005) reported that wild poinsettia (Euphorbia heterophylla L.) from Brazil had >24-fold resistance to imazethapyr than an S biotype. It was also determined that the biotype was resistant to metsulfuron, nicosulfuron, and a protoporphyrinogen oxidase inhibitor, fomesafen. Researchers have identified comparable resistant levels to ALS inhibitors in annual bluegrass populations from turfgrass systems (Cross et al. 2013; McElroy et al. 2013). Similar resistance to ALS inhibitors has been confirmed in horseweed [Conyza canadensis (L.) Cronq.], pigweeds (Amaranthus spp.), and rice barnyardgrass [Echinochloa phyllopogon (Stapf) Koso-Pol.] (Osuna et al. 2002; Whaley et al. 2006; Zheng et al. 2011).

Multiple and Cross-Resistance Experiment.

Metsulfuron at 21 g ha⁻¹ injured the S biotype 75% at 4 WAT but did not injure the R biotype (Table 1). Metsulfuron was more injurious to the S biotype than trifloxysulfuron at 29 g ai ha⁻¹ (75% vs. 45%), but trifloxysulfuron did not injure the R biotype. Injury to the R biotype was equivalent to the S biotype from dicamba, glyphosate, and triclopyr. These herbicides averaged 39, 44, and 85% injury, respectively. Results suggest this biotype is resistant to another sulfonylurea used for spotted spurge control in turf, trifloxysulfuron.

Turfgrass managers have limitations with alternative mechanisms of action to ALS inhibitors in warm-season grasses. For example, triclopyr is only labeled for cool-season grasses and zoysiagrass (Zoysia japonica Steud.) because of excessive injury potential from labeled use rates (0.28 to 1.12 kg ha⁻¹) on most warm-season species during active growth (Cudney et al. 1997; McElroy and Breeden 2006). Glyphosate was less efficacious than triclopyr at rates evaluated and would be limited to spot applications for POST control of spotted spurge. Dicamba is safe on most major turfgrass species and provided equivalent control on both biotypes after 4 wk. However, dicamba was not as efficacious as triclopyr and may require tank mixtures with other herbicides for best results.

Another limitation to POST control of spotted spurge is herbicide efficacy on mature plants. In field experiments, dicamba, triclopyr, and other herbicides required greater use rates to control the R biotype when plants were ∼10 cm tall with multiple (>4) branches, compared with seedlings treated before branching (McCullough, personal observation). Reduced control from herbicides applied to mature weeds has been reported for common cocklebur (Xanthium strumarium L.), pitted morningglory (Ipomoea lacunosa L.), smooth pigweed (Amaranthus hybridus L.), and other annual weeds (Barrentine 1989; DeFelice et al. 1989; Klingman et al. 1992). Furthermore, mature weeds may have produced viable seed that could further spread resistant populations after control. Practitioners may need to modify application rates and regimens of alternative herbicides to metsulfuron for controlling mature populations of ALS-resistant spurge. Further research is needed to evaluate the efficacy of other ALS inhib-itors, Photosystem II inhibitors, and organic arsenicals for controlling this biotype.

Gene Sequencing of the ALS Enzyme.

The R biotype contained two amino acid substitutions: Asp-341 to Ser and Trp-574 to Leu (Figure 3; Table 2). A mutation was not detected in the Pro-197 codon. Considering Asp-341 is not one of the 18 amino acid in the ALS-inhibiting herbicide-binding region and Trp-574 to Leu is historically correlated with target site resistance, it can be concluded that an amino acid substitution to Leu-574 is the molecular mechanism of resistance for this spotted spurge biotype (McCourt et al. 2006).

Table 2.

Missense mutations in the susceptible and resistant spotted spurge biotypes as revealed by nucleotide read mapping and translation to amino acid sequence.

Figure 3.

Alignment of acetolactate synthase (ALS) sequences flanking codon Trp₅₇₄ from Arabidopsis thaliana and the spurge populations. The yellow highlight indicates the Trp₅₇₄ codon in A4. thaliana. The resistant population shows a mutation (TGG to TTG) conferring Trp₅₇₄ to Leu substitution in ALS enzymes. (Color for this figure is available in the online version of this article.)

The Trp-574 to Leu substitution has conferred resistance to sulfonylureas, imidazolinones, and pyrimidinyl-benzoic acids (McCourt et al. 2006; McElroy et al. 2013; Yu and Powles 2013). This specific mutation has been identified as the molecular basis for ALS resistance in annual bluegrass, Powell amaranth [Amaranthus powellii (S.) Wats.], kochia [Kochia scoparia (L.) Schrad.], rigid ryegrass (Lolium rigidum Gaudin), wild mustard (Sinapis arvensis L.), and other weed species (Christoffers et al. 2006; McElroy et al. 2013; Warwick et al. 2008; Yu et al. 2008; Yu and Powles 2013).

Implications from These Findings.

Spotted spurge with resistance to ALS inhibitors will require alternative cultural and chemical control methods for acceptable control. Selecting alternatives to ALS inhibitors with comparable efficacy will be critical in turfgrasses susceptible to injury from other mechanisms of action, such as bermudagrass [Cynodon dactylon (L.) Pers.] or seashore paspalum. Mechanical suppression or hand weeding may be an effective approach to controlling many resistant weeds in turfgrass. However, these techniques may not be practical in large areas with severe infestations of spotted spurge.

The prolific seed production of spotted spurge will contribute to the spread of R biotypes. Isoxaben or other PRE herbicides should be applied before soil temperatures reach 25 C for effective control (Asgarpour et al. 2015). The continued use of PRE herbicides throughout the summer may be necessary to control later flushes of seed germination in turfgrass. Further research is needed to determine the distribution of ALS resistance in spotted spurge and related Chamaesyce species in the United States.

This is the first report of ALS resistance in spotted spurge. This is also the first report of a herbicide-resistant broadleaf weed from a turfgrass system. The confirmation of ALS resistance in a weed with substantial seed production has serious implications for herbicide resistance management. Rotating POST herbicide mechanisms of action may provide acceptable control in certain turfgrass species. However, alternative herbicides to ALS inhibitors may have potential for drift, turfgrass injury, or limited efficacy. The identification of new weed species with resistance to ALS inhibitors should emphasize the importance of alternative management programs, including herbicide rotation, in turfgrass and other cropping systems.