Multiple mechanisms are involved in new imazamox-resistant varieties of durum and soft wheat

Weed control in wheat is one of the major goals of farmers in their efforts toward obtaining the highest crop yields for human foods. Several studies (dose-response, enzyme activity, absorption-translocation and metabolism) were conducted to characterize the resistance level of two new wheat cultivars called Rafalín (Triticum aestivum) and Antoñín (T. durum) that were obtained by conventional breeding based on Clearfield® technology; they are resistant (R) to imazamox compared to their sensitive (S) counterparts (Gazul and Simeto, respectively). The R-cultivars were 93.7-fold (Rafalín) and 43.7-fold (Antoñín) more resistant than their respective S-cultivars. The acetolactate synthase (ALS) enzyme activity revealed high resistance to imidazolinone (IMI) herbicides in R-cultivars, but no cross-resistance to other ALS herbicides was found. The Ser653Asn mutation that confers resistance to IMI herbicides was identified in the imi1 and imi2 genes of Rafalín and only in the imi1 gene of Antoñín. The 14C-imazamox absorption did not differ between the R- and S-cultivars. Imazamox was metabolized by Cyt-P450 into imazamox-hydroxyl and imazamox-glucoside in the R-cultivars, altering their translocation patterns. The differential sensitivity to imazamox between R-cultivars was due to the number of resistance genes that carry each genotype. The R-cultivars Rafalín and Antoñín could be excellent weed control tools.

. Dose-response curves of the fresh weight reduction with respect to untreated control plants of imazamox-susceptible and imazamox-resistant cultivars from the different soft (Triticum aestivum) and durum (Triticum durum) wheat varieties when evaluated at 30 DAA. Vertical bars ± standard error (n = 10).  Table 1. Parameters of the sigmoidal equation used to estimate the imazamox dose (g ai ha −1 ) needed to reduce the weight of a population by 50% (GR 50 ) in susceptible and resistant wheat cultivars of the soft (Triticum aestivum) and durum (Triticum durum) varieties. c = lower limit, d = upper limit, b = Hill's slope, R 2 aj = 1 − (sums of squares of the regression/corrected total sums of squares). RF = Resistance factor = GR 50 R/GR 50 S. CI95 values are the upper and lower limits (±) of the 95% confidence intervals (n = 10).
Scientific REPORTS | 7: 14839 | DOI: 10.1038/s41598-017-13874-3 The R-cultivars showed multiple resistance to herbicides in the IMI family; however, they were not resistant to the other families of ALS inhibitor herbicides (Fig. 2, Table 2). According to the 95% confidence intervals (CI), the S-cultivars showed no significant differences in either the GR 50 or I 50 parameters. Independent of the wheat variety, the Rafalín cultivar is the one that withstands a higher dose of imazamox. ALS sequencing. The predicted amino acid sequence of S cultivars was presented as the same consensus of accessions imi1-AY210407 and imi2-AY210408 in wheat, corresponding to the imi1 and imi2 genes, respectively. The R-cultivar Rafalín presented two mutations at the Ser653 position (also known as Ser627 15,20 ) in the imi1 and imi2 genes, whereas the Antoñín cultivar presented the same mutation, but in the imi2 gene. The codon change was AAC to AGC, resulting in an amino acid substitution from serine to asparagine. No mutation was found in the imi3 gene (Fig. 3). 14 C-Imazamox absorption and translocation. The four wheat cultivars presented a high 14 C-imazamox absorption rate. At 12 h after application (HAA), the S-cultivars presented an absorption level of over 87%, absorbing up to more than 94% at 96 HAA, while the R cultivars exhibited an average absorption of 73 and 88% at 12 and 96 HAA, respectively. Although the S-cultivars absorbed more 14 C-imazamox, the differences were not significant with respect to the R-cultivars (Fig. 4a).
The S-cultivars showed high rates of 14 C-imazamox translocation from the treated leaves to the rest of the plants and roots. The greatest differences in translocation were observed at 96 HAA. The rates of translocation to the roots of the R-cultivars were 16.1 and 16.6% for Rafalín and Antoñín, respectively, for the absorbed herbicide, translocating 8.5-10.3% less herbicide to the roots than the S-cultivars and retaining approximately 60% in the treated leaves of both imazamox-resistant cultivars. At that time, the S-cultivars presented a higher translocation of 14 C-imazamox from the treated leaves to the rest of the plants and roots (Table 3).    A Phosphor Imager was used to confirm the previous results. At 96 HAA, the R-cultivar plants translocated smaller amounts of 14 C-imazamox from the treated leaf to the roots than S-cultivar plants. This finding shows that the 14 C-imazamox translocation could have contributed to the resistance of the R-cultivars (Fig. 4b).
Imazamox metabolism. In this study, the R-cultivar plants (Rafalín and Antoñín) presented rapid imazamox metabolism to hydroxylated imazamox (imazamox-OH) and a glucose conjugate (imazamox-glucose), which were metabolites that were not found in the S-cultivars (Gazul and Simeto). In taking into account that these metabolites come from imazamox, we considered the sum of all of them to be 100%, and we calculated the imazamox percentage that was metabolized in the leaves and roots. The R-cultivars had a high metabolite content at 96 HAA compared to the S-cultivars. In the case of the Rafalín and Antoñín cultivars, over 90% corresponded to imazamox metabolites, while the Gazul and Simeto cultivars did not metabolize this herbicide. It should be noted that the Antoñín cultivar presented higher contents of the glycosylated metabolite (84.1 μg g −1 in leaves and 47.6 μg g −1 in roots) compared to the Rafalín cultivar (50.4 μg g −1 in the leaves and 29.1 μg g −1 in the roots). Imazamox metabolism inhibition was also observed in R-cultivar plants that were treated with malathion. This finding suggests that cytochrome P450 monooxygenase (Cyt-P450) is involved in this detoxification mechanism (Fig. 5).

Discussion
Clearfield® wheat that is resistant to imazamox is a new and highly effective tool for weed control that has been developed in North America and in most Latin American countries. This study reports the first European case of two imazamox-resistant wheat cultivars, based on Clearfield® technology, which were obtained by the crossbreeding of Pantera T. aestivum Clearfield® 20 and a local imazamox-susceptible cultivar (Gazul) and another Triticum durum Clearfield® obtained by the crossing of Simeto (local imazamox-susceptible T. durum cultivar) and a Tritordeum Clearfield® 15 .
Dose-response assays demonstrated the high susceptibility of the S-cultivars (Gazul and Simeto) at low doses, whereas the R-cultivars (Rafalín and Antoñín) presented a small reduction in fresh weight. The differences between the R-cultivars could be explained by the different biochemical, morphological, physiological and molecular traits [21][22][23] . However, to achieve total growth reduction in a resistant plant, the grower needs to apply at least double the rate of herbicide as that of its corresponding GR 50 24 . This rule implies that the Antoñín and Rafalín cultivars require imazamox doses that are 8 and 11 times, respectively, higher than the recommended field dose of 40 g ia ha −1 of imazamox to produce total damage in the crop. This high level of resistance to the imazamox of wheat R-cultivars is enough to be a useful weed control tool, with advantages for farmers. The adequate use of herbicide-tolerant crops and the adoption of the associated agronomic practices may enhance farmland biodiversity and reduce the risk of weeds evolving herbicide resistance 8 .
Resistance to IMI herbicides is usually the result of a point mutation in the ALS gene that causes an amino acid substitution in the ALS enzyme 10,25 . The high I 50 rate of the R-cultivars suggests that the resistance mechanism is related to the ALS enzyme in wheat cultivars 9,16,20 . The similar specific activity of ALS between them also suggests that ALS overexpression is not involved as a resistance mechanism. Similar results were described in Sinapis alba 26 and T. aestivum 27 , in which ALS overexpression was not involved as a resistance mechanism.
The higher GR 50 and I 50 values estimated for Rafalín were due to the fact that this cultivar presented a mutation at the Ser-653 position in the imi1 and imi2 genes, which are located on the long arms of chromosomes 6B and 6D, respectively. Consequently, this resistant cultivar has a higher imazamox resistance level than the Antoñin  cultivar, and it only carried a single resistance gene (imi2). The R-wheat cultivars have satisfactorily acquired these alleles from their respective resistant parents, i.e., the imi1 and imi2 genes from the Pantera 20 cultivar were transferred to Rafalín, and imi2 from Tritordeum 15 was transferred to the Antoñín cultivar. The resistant allele imi-2, common in Clearfield® crops, endows sufficient resistance level at recommended field rates to IMI herbicides 28 . The resistance conferred by the ALS-resistant imi1 gene resulted in an additive resistance level to that conferred by imi2 gene. These mutations only confer resistance to IMI herbicides, but not cross-resistance to other ALS inhibitors 6 . The fitness cost associated to ALS-resistant alleles is small and easily detectable 10 . However, this fact can not generalize and the impact of each specific ALS gene mutation needs to be individually evaluated 29 , taking into account the growing conditions and species 10 . In addition, the epistatic effects of multiple resistance alleles on plant fitness cost is yet unknown 10,30 . Therefore, further studies are needed to determine the epistatic effect of ALS-resistant allele in imi1 gene of cultivar Rafalín, on fitness cost and the possible yield drag. The ALS isoform from the D genome, which corresponds to the imi2 gene, presents more ALS activity compared to isoforms of the A and B genomes 31 . This finding explains why the imi2 mutation is sufficient for making the Antoñín cultivar resistant to imazamox. Our results are consistent with other studies that report a single and/or double mutation in the imi1-and imi2-ALS genes that confer resistance to IMI herbicides in wheat cultivars 16,20,28,32 , and other Clearfield® crops such as rice 33 , barley 34 , sunflowers 35 and chickpeas 12,25 . In addition, a mutation at the Ala122 position of the imi2 gene that was identified in wheat improved the resistance to IMI herbicides 12 .
The effectiveness of an herbicide depends on the retention of the product on the leaf 23 , the foliar absorption of the active ingredient and finally its translocation to the site of action 21,22 . Herbicide foliar retention is influenced by leaf morphological characteristics 36 , and it is not a major mechanism that confers herbicide resistance 23,37 . Our results suggest that the leaf morphology of wheat cultivars is not related to greater or lesser herbicide retention, as was demonstrated in other imazamox-resistant wheat cultivars 9,20,27 .
Imazamox is absorbed and translocated very quickly. However, given that resistance to ALS inhibitor herbicides is generally associated with mutations in the ALS gene 9-11 , the absorption and translocation are not usually studied, and information on these mechanisms in Clearfield® crops is scarce. We recorded high absorption rates between the S-and R-cultivars, confirming that this mechanism is not involved in the resistance of the latter. By contrast, the 14 C-imazamox translocation results suggested that this parameter could play an important role in the resistance of the Rafalín and Antoñín cultivars because they retained most of the herbicide in the treated leaf. However, the translocation differences were a physiological and metabolic response to the different sensitivities to imazamox between the S-and R-cultivars, because this differential translocation was neutralized with malathion. The lowest translocation observed in the R-cultivars can be explained by the fact that these cultivars metabolized imazamox, and the identified metabolites (imazamox-OH and imazamox-glucose) have limited mobility 38 . Therefore, the observations made during the absorption and translocation assays in the wheat R-cultivars not only showed the translocated 14 C-imazamox but also these 14 C-metabolites.
Clearfield® crops generally have an enhanced ability to metabolize IMI herbicides before they reach the target site 39 . Imazamox metabolism was documented for the cultivar Clearfield® Pantera 20 , and although there are no studies in Tritordeum, it is evident that the R-cultivars (Rafalín and Antoñín) acquired the ability to metabolize IMI herbicides from their R counterparts. The mechanism by which Clearfield® crops gain tolerance to IMI herbicides has not yet been fully characterized 40 . In our study, malathion applications confirmed that Cyt-P450 enzymes play an important role in the imazamox detoxification of the R-cultivars into compounds that are harmless (imazamox-OH and imazamox-glucose) to the plants 22 . These enzymes are mediators of herbicide degradation that are involved in multiple herbicide resistance 11,41,42 . Cyt-P450 enzymes are responsible for the hydroxylation of a methyl group on the imazamox molecule studies 9,20 , followed by a glucosyl transferase catalysis producing a glucose conjugation 42 . This reaction occurs rapidly and is not reversible 41 . In addition, IMI herbicide metabolism not only has very little effect on the ALS, but the herbicide is also poorly translocated 39 . However, it can not be attributed that the metabolism of imazamox was due solely to Cyt-P450 genes, because other secondary metabolism pathways can be involved in resistance to ALS-inhibiting herbicides, resulting in an accumulation of different non-target-site-resistance genes, each of them conferring a moderate level of resistance 43 .

Conclusions
The differential response to imazamox between R-and S-cultivars was primarily due to the number of resistance genes that carry each genotype (at target-site level), but it must also be attributed to the enhanced imazamox metabolism into non-toxic compounds (imazamox-OH and imazamox-glucose), which is mediated by the Cyt-P450 (non-target-site genes) and is responsible for altering the translocation patterns. These mechanisms confer high resistance to IMI herbicides in the cultivars Rafalín (T. aestivum) and Antoñín (T. durum), allowing them to survive at higher doses than the recommended field dose of imazamox (40 g ai ha −1 ), being a great advantage for farmers in terms of weed management.

Material and Methods
Plant material. Two wheat cultivars (T. aestivum and T. durum) were used. For T. aestivum the cultivar Rafalín resistant to imazamox (R), and the cultivar Gazul as the susceptible one (S) were used. For T. durum, the R-and S-cultivars were Antoñín and Simeto, respectively. The cultivar Rafalín comes from the crossbreeding of Pantera Clearfield® (R) x Gazul (S) cultivars. Pantera was previously characterized by this research group 20 , and it has two mutations (imi1 and imi2) that confer resistant to imazamox. The resistant biotype of T. durum (Antoñín) comes from the crossing of Simeto (S) x Tritordeum, which presents the mutation imi2 15  Growth conditions. Seeds were sown in Petri dishes with two layers of filter paper moistened with distilled water. They were kept at 4 °C in the dark during 48 h. After this period, the seeds were transferred to a growth chamber until germination, with a temperature regime of 27/14 °C day/night with a photoperiod of 14/10 h, respectively. The seedlings were placed in pots (1 L) containing a mixture of peat and sand (1:1) as substrate, and taken to the greenhouse, where the plants grew at 25-28/12-14 °C day/night with 16 h of photoperiod. The natural light was supplemented by 900 µmol m −2 s −1 photosynthetic photon flux density delivered by incandescent and fluorescent lights. Once the plants reached a growth state corresponding to 3-4 true leaves, the treatments were performed.
Foliar herbicide retention. Foliar herbicide retention assays were performed following the method adapted by Jiménez et al. 27 . Plants were treated with a solution containing 40 g ai ha −1 of imazamox + 1.25 L ha −1 of adjuvant Dash (34.5% w/v methyl oleate/methyl palmitate) + 100 mg L −1 Na-fluorescein in the same treatment chamber used in dose-response assays. Na-fluorescein was used as a labeling reagent to determine the amount of herbicide solution retained. Once the herbicide solution from the leaf (20-25 min) was dry, the plants were cut at ground level and washed individually in Erlenmeyer's containing 50 mL of NaOH 5 mM shaking them vigorously for 30 seconds. The washing solution was recovered in glass flasks and the fluorescein absorbance was immediately measured at 490 exc /510 em nm (Hitachi F-2500 spectrofluorimeter). The cut tissues were packed in cellulose envelopes and dried in an oven at 80 °C for 72 h. Ten plants of each cultivar were used in a completely random design. Retention was expressed as µL of imazamox solution per g of dry matter. ALS enzyme activity. ALS activity was determined following the methodology used by Hatami et al. 44 with slight modifications. Samples of three grams of leaf tissue were taken and immediately frozen in liquid N 2 . Then, the samples were macerated in a mortar using 5 mg of polyvinylpyrrolidone (PVPP). An extraction buffer composed of 1 M K-phosphate buffer solution (pH 7.5), 10 mM sodium pyruvate, 5 mM MgCl 2 , 50 mM thiamine pyrophosphate, 100 μM flavin adenine dinucleotide (FAD), 12 mM dithiothreitol and glycerol (1:9 v/v) was added. The solution was agitated for 10 min at 4 °C. The homogenate was filtered through four layers of cheesecloth and centrifuged (20,000 rpm for 20 min). The supernatant containing a crude ALS enzyme extract was immediately used for the enzyme assays. To assay the ALS activity, 90 µL of enzyme extract was added to 110 µL of freshly prepared assay buffer (0.08 M K-phosphate buffer solution (pH 7.5), 0.5 M sodium pyruvate, 0.1 M MgCl 2 , 0.5 mM thiamine pyrophosphate, and 1 µM FAD). Then, increasing concentrations (0, 0.1, 1, 5, 10, 50, 100 and 1000 µM) of ALS inhibiting herbicides were added. The herbicides of technical grade were imazamox, imazapyr, imazaquin, tribenuron methyl, bispyribac sodium, flucarbazone and florasulam. Standard compounds used with 96.5-98% purity were provided by Sigma-Aldrich, Spain. The mixture was incubated for 60 min at 37 °C. The reaction was stopped after the addition of 50 µL of H 2 SO 4 and incubated at 60 °C for 15 min to decarboxylate acetolactate to acetoin. Finally, 250 µL of a freshly prepared solution of creatine in water (5 g L −1 ) and 250 µL of a solution of naphthol in sodium hydroxide (50 g L −1 NaOH 5 M) was added. It was again incubated at 60 °C for 15 min to facilitate decarboxylation of acetolactate to acetoin. Absorbance of acetoin was measured with a spectrophotometer (Beckman DU-640, Fullerton, CA, USA) at A520 nm. The total content of ALS in the raw extract was measured using the colorimetric method using the commercial kit-protocol No. P5656 (Sigma-Aldrich, Madrid, Spain) following the manufacturer´s instructions at 595 nm. The background was subtracted using control tubes. Three replicates per cultivar were made, each with extract from the mixture of the three plants.  16 to amplify a 617 bp-length fragment was observed. A polymerase chain reaction (PCR) reaction was set up with Certamp complex enzyme mix (Biotools B&M Labs, Madrid, Spain) following the manufacturer's instructions. PCR products (5 µL) were digested with the restriction enzyme Msp I (Invitrogen, CA, USA) to try to identify the three expected ALS alleles (from genomes A, B or D) of the catalytic subunit present in wheat varieties 16 . Both PCR and digestion products were resolved on 1% agarose gels and viewed under UV light. Ten PCR products of each allele and each cultivar were sequenced by Sanger technology. The assembly of the sequences was carried out by SeqMan Pro (Version 11, DNASTAR; Madison, WI, USA) and Geneious (Version 8.1.8, Biomatters Ltd, and Auckland, New Zealand) software's. ALS sequences of the wheat accessions imi1-AY210407, imi2-AY210408 and imi3-AY273827 from GenBank, were included in the alignment. Simultaneously, whole plants treated with 14 C-imazamox and rinsed with water: acetone (9:1 v/v), were fixed on filter paper (25 × 12.5 cm) and dried at room temperature. Finally, they were placed for 4 h on a phosphor film to visualize the 14 C-imazamox by a phosphor imager (Cyclone, Perkin-Elmer, and Packard Bioscience BV). Three plants of each cultivar were used at each evaluation time.
Imazamox metabolism. The methodology described by Rojano-Delgado et al. 45 was followed. Ten plants of each wheat cultivar were treated with imazamox at 40 g ai ha −1 (field dose) as in the dose-response assays. Jointly, a group of 10 plants of each cultivar were treated with Malathion (1000 g ai ha −1 ) 1 h before imazamox application, to evaluate if the Cyt-P450 was involved in the imazamox metabolism. A group of plants was remained as control. Those plants treated with herbicide and the controls were cut at 96 HAA and were washed with 60 mL of distilled water to remove the imazamox and soil residues on the leaf surface and finally stored at −40 °C until use. The samples were macerated in a porcelain mortar to a fine powder using liquid nitrogen. Next, 500 mg of each sample was mixed with 10 mL of methanol:water (9:1 v/v), and the metabolites were extracted using ultrasound at 70-W ultrasonication power for 10 min (duty cycle 0.7 s s −1 ). Supernatant was separated by centrifugation (15 min at 15000 rpm), and evaporated to dryness under an airstream. The solid residue of this fraction was reconstituted in 500 µL of methanol:water (9:1 v/v), and filtered through a nylon filter syringe (45 μm pore size and 13 mm i.d.; Millipore, Ireland) before chromatographic analysis.
For the determination of imazamox and its metabolites in extracts from plants, a liquid chromatography-diode array detector was employed. A hydrophilic interaction liquid chromatography column C18 (20 cm × 4.6 cm, 3 μm particle size) was used for the separation of the target compounds. Fifty µL of the reconstituted phase was injected into the liquid chromatography with 1% acetic acid solution as mobile phase A, and 100% methanol as mobile phase B. The elution program started with 5% mobile phase B and followed the linear gradient: step 1: 5 to 20% methanol for 10 min; step 2: 20 to 80% methanol for 10 min; step 3: 80 to 100% methanol for 5 min; and step 4: 100 to 5% methanol for 10 min. The constant flow rate and column temperature were 1 mL min −1 at 40 °C. Chromatographic grade and liquid chromatography-mass spectrometer grade solvents were used for liquid chromatography-diode array detector and liquid chromatography-time-of-flight/mass spectrometer analysis, respectively.
The analyses were performed in an Agilent 1200 Series LC system interfaced to an Agilent 6540 UHD Accurate-Mass liquid chromatography-time-of-flight/mass spectrometer detector (Palo Alto, USA), equipped with an Agilent Jet Stream Technology electrospray ion source operating in the positive ionization mode. The separation conditions were identical to those for the liquid chromatography-diode array detector determination, except for the use of the respective liquid chromatography-time-of-flight/mass spectrometer grade solvents.
After analysis, imazamox and metabolites were determined by the liquid chromatography-diode array detector analysis (measurement wavelength, 240 nm). A 15 Gold HPLC System from Beckman Coulter (Fullerton, USA) equipped with a 26 System Gold Diode Array detector (wavelength range 190-600 nm) was used in this case. Chromatographic peaks were assigned according to retention times using as a reference the imazamox peak identified by spiking extracts with the commercial standard. Quantification of imazamox metabolites was based on the calibration model for imazamox, and the results were expressed as μg of analytic g −1 fresh weight. Three replications per sample were analyzed. Statistical analysis. The percentage data of fresh weight reduction and ALS enzyme activity were submitted to a non-linear regression analysis. The dose of imazamox needed to reduce the weight of a population (GR 50 ) and to inhibit ALS activity (I 50 ) by 50% was calculated. The drc statistical package in the program R version 3.2.5 was used to conduct the following log-logistic model of four parameters 46 where Y is the percentage of fresh weight reduction with respect to the control, c and d are coefficients corresponding to the upper and lower asymptotic limits, b is the Hill slope, g is the imazamox dose (GR 50 , or I 50 ) at the mean point of inflexion between the upper and lower asymptote and x (independent variable) corresponds to the glyphosate dose. The data were plotted using SigmaPlot 11.0 (Systat Software, Inc., USA). Resistance factor was calculated as: RF = GR 50 or I 50 (R)/GR 50 or I 50 (S).
The data obtained in the spray retention, 14 C-imazamox absorption and translocation, and imazamox metabolism were subjected to ANOVA. For each analysis, assumptions such as equality of variance and normal distribution were evaluated. When required, the Tukey HSD test at 5% probability was used to separate means. A statistical analysis was performed using Statistix software (version 9.0) from Analytical Software (USA).