Enhanced Herbicide Metabolism and Target-Site Mutations Confer Multiple Resistance to Fomesafen and Nicosulfuron in Amaranthus retroflexus L.

Simple Summary Amaranthus retroflexus L. is the most prevalent broadleaf weed in autumn crops such as soybean and corn in northeastern China. In this study, we investigated the target-site resistance mechanism of HW-01 and ST-1 populations of A. retroflexus, which are resistant to fomesafen and nicosulfuron, and the non-target-site resistance of cytochrome P450s- and GSTs-based herbicide metabolic was also confirmed in the HW-01 population. This study helps to provide scientific and theoretical guidance for the control of resistant populations of A. retroflexus. Abstract Amaranthus retroflexus L. is a highly competitive broadleaf weed of corn–soybean rotation in northeastern China. In recent years, the herbicide(s) resistance evolution has been threatening its effective management in crop fields. One resistant A. retroflexus (HW-01) population that survived the protoporphyrinogen oxidase (PPO) inhibitor fomesafen and acetolactate synthase (ALS) inhibitor nicosulfuron applied at their field-recommended rate was collected from a soybean field in Wudalianchi City, Heilongjiang Province. This study aimed to investigate the resistance mechanisms of fomesafen and nicosulfuron and determine the resistance profile of HW-01 to other herbicides. Whole plant dose–response bioassays revealed that HW-01 had evolved resistance to fomesafen (50.7-fold) and nicosulfuron (5.2-fold). Gene sequencing showed that the HW-01 population has a mutation in PPX2 (Arg-128-Gly) and a rare mutation in ALS (Ala-205-Val, eight/twenty mutations/total plants). In vitro enzyme activity assays showed that ALS extracted from the HW-01 plants was less sensitive to nicosulfuron (3.2-fold) than ST-1 plants. Pre-treatment with the cytochrome P450 inhibitors malathion, piperonyl butoxide (PBO), 3-amino-1,2,4-triazole (amitrole), and the GSTs inhibitor 4-chloro-7-nitrobenzofurazan (NBD-Cl) significantly increased fomesafen and nicosulfuron sensitivity in the HW-01 population compared with that of the sensitive (S) population ST-1. Moreover, the rapid fomesafen and nicosulfuron metabolism in the HW-01 plants was also confirmed via HPLC-MS/MS analysis. Furthermore, the HW-01 population showed multiple resistance (MR) to PPO, ALS, and PSII inhibitors, with resistance index (RI) values ranging from 3.8 to 9.6. This study confirmed MR to PPO-, ALS-, and PSII-inhibiting herbicides in the A. retroflexus population HW-01, as well as confirming that the cytochrome P450- and GST-based herbicide metabolic along with TSR mechanisms contribute to their multiple resistance to fomesafen and nicosulfuron.


Introduction
Amaranthus retroflexus L. is a widely distributed and troublesome annual broadleaf weed in northern China [1,2]. It is a monoecious species that is highly self-fertilized with occasional outcrossing, and due to large-scale seed production (about 1.5 million seeds),

Plant Materials
In September 2017, seeds of a suspected resistant A. retroflexus population (HW-01) were collected from a soybean field with a continuous application history of fomesafen or nicosulfuron (>10 years) in Wudalianchi City (48.51 • N, 126.13 • E), Heilongjiang Province. In these regions, the growers found that fomesafen had poor control of A. retroflexus under the field-recommended rate. An herbicide-susceptible population of A. retroflexus (ST-1) was collected from Mountain Tai (36.05 • N, 117.03 • E), Tai'an City, Shandong Province, where there is no history of herbicide application. Mature seeds of suspected resistant (R, HW-01) and susceptible (S, ST-1) A. retroflexus populations were randomly collected from at least 50 individual plants, and the sample size was approximately 0.2 ha. After drying, the seeds were stored at 4 • C in paper bags until use.

Whole-Plant Dose-Response Experiments
For accelerating germination, A. retroflexus seeds were placed in Petri dishes containing two layers of Whatman No. 1 filter paper and 6 mL of deionized water. Then, the Petri dish was placed in the growth chamber (33/23 • C, 12/12 h day/night) for cultivation. After the radicle germinated, 10 seedlings (after thinned) were planted in a plastic pot measuring 15 cm in diameter and 12 cm in height and containing loam. The soil organic matter content was 1.7% and filtered using a 3-mm sieve. Plants were grown in the greenhouse with a 14-h photoperiod and a temperature of 25/15 • C day/night. The pots were watered every two days to maintain moisture.
Post-emergence application was conducted when seedlings reached the 3-4-leaves stage using a mobile nozzle cabinet sprayer equipped with flat fan nozzles (TeeJet 9503EVS, Greenman Machinery, Beijing, China) delivering a spraying volume of 450 L ha −1 at 280 KPa. Based on the results of preliminary experiments, dose-response tests were conducted with applications of nicosulfuron and fomesafen and five other herbicides to determine the level of resistance of HW-01 and ST-1 to these herbicides, with the higher dose used for HW-01; the herbicides as well as the doses are shown in Table 1. At 21 days after treatment (DAT), the aboveground materials from each pot were harvested and oven-dried at 80 • C for 72 h in an air-blowing box (Model DHG-9140A, Changzhou Noki Instrument Co., Ltd., Changzhou, China). Then, the dry weights were recorded. The aboveground dry weight of the plants obtained here was divided by the dry weight of the untreated control, and the result was expressed as a percentage. The experiment was designed in a completely randomized design, and the whole experiment was repeated twice. Table 1. Herbicide rates applied in the whole-plant experiment.

Herbicide a Group b
Application Rate (g ai ha −1 )

Gene Sequencing and In Vitro Assay of ALS Activity
The fresh leaf tissues from each population (R, HW-01; S, ST-1; 20 individuals per population) at the 3-4-leaves stage were harvested and stored at −80 • C. For ALS, PPX1, and PPX2 gene sequencing, RNA extraction, primers, and methods were identical to our previous report [2], and the primer pairs used to amplify ALS, PPX1, and PPX2 of A. retroflexus are shown in Table S1. Then, the in vitro activity of ALS was extracted and determined according to the method described by Yu et al. and Han et al. [29,30]. The nicosulfuron (97%, Shandong Rainbow chemical Co., Ltd., Weifang, China, provided) concentrations used for the in vitro activity assays were 0.02, 0.2, 2, 20, 200, 2000 µM for the HW-01 population while 0.001, 0.01, 0.1, 1, 10, 100 µM were used for the ST-1 population. Two ALS extractions were performed for each population. The activity of each ALS extract was measured in three technical replicates and averaged.

Effect of P450 and GST Inhibitors on Fomesafen and Nicosulfuron Resistance
To test whether metabolic resistance was involved in the fomesafen and nicosulfuron resistance found in the HW-01 population, R and S plants (3-4-leaves stage) were treated with fomesafen and nicosulfuron, respectively, with or without the following cytochrome P450s inhibitors: malathion (Binnong, Shandong, China) at 1500 g ai ha −1 ; PBO (TCI, Shanghai, China) at 1500 g ai ha −1 ; amitrole (TCI, Shanghai, China) at 13.1 g ai ha −1 ; and GST-inhibitor NBD-Cl (TCI, Shanghai, China) at 270 g ai ha −1 . Malathion, PBO, and amitrole are indicators of P450-mediated metabolic resistance and NBD-Cl is an indicator of GST-mediated metabolic resistance in weeds [31]. All cytochrome P450 inhibitors were applied 2 h before the treatment with fomesafen or nicosulfuron. In contrast, NBD-Cl was applied to plants 2 d before herbicide treatment according to the procedure outlined by Varanasi et al. and Ma et al. [28,32,33]. The application rates of fomesafen and nicosulfuron are described in Table 1. These P450-and GST-inhibitory bioassays were performed simultaneously in whole-plant dose-response experiments. The methods described above were equally applicable to the plant-growing and herbicide applications. At 21 DAT, the aboveground dry weights of the plants were recorded and expressed as percentages of the control group. The whole experiment was double-repeated and had a completely random design.

Analysis of Fomesafen and Nicosulfuron Metabolism in A. retroflexus
Technical-grade fomesafen (98%) and nicosulfuron (97%) were provided by Shandong Binnong Technology Co., Ltd., Binzhou, China and Shandong Rainbow chemical Co., Ltd., respectively. Micropipettes were used to apply 4.0 µg of fomesafen and 1.0 µg of nicosulfuron on S and HW-01 plants at the 3-4-leaves stage (4.0 or 1.0 µg per individual), respectively. After 1, 3, 5, 7, and 9 days of herbicide treatment, plants of S and R were selected for extraction, and the method was based on the one described by Bai et al. [34], with five time samples and three replicates. The experiment was created in a completely randomized design, and the whole experiment was carried out twice. The HPLC-MS/MS parameters are listed in Table 2, and validation of the analysis methods is in Supplementary Table S2. Instrumentation (Thermo Fisher, Vanquish UHPLC-TSQ Quantis, Waltham, MA, USA) and HPLC-MS/MS analytical conditions of fomesafen were for separation using a C 18 column (Agilent Eclipse Plus C 18 ) at a flow rate of 0.3 mL min −1 at 30 • C. Mobile phase A was 0.1% formic acid in the water, while mobile phase B was methanol (LC grade). The injected sample was subjected to gradient elution at 90% component A (10%B), and the B component reached 92% (8%A) in 2.5 min, followed by adjustment of the mobile phase ratio to 90%A and 10%B in 3.5 min, which was maintained in this state for 1.5 min, after which it returned to the initial state. Detection conditions were conducted with positive ions (3500V), negative ions (2800V), sheath gas (30 Arb), aux gas (5 Arb), sweep gas (0 Arb), ion transfer tube temp (325 • C), vaporizer (350 • C). The HPLC-MS/MS parameters include min dwell time (ms) 124.00, RF lens (V) 204.
Nicosulfuron was conducted using an Acquity UPLC™ system (Waters, Milford, CT, USA) and separated using BEH C 18 column (Waters). The gradient elution (0.4 mL min −1 ) started with 10% component A (mobile phase A: water; mobile phase B: methanol; LC grade) at the time of sample injection and was maintained for 6 min; then, it linearly increased to 90%A (10%B) at 8 min, returning to the initial state at 10 min. The MS/MS analysis conditions were performed with a desolvation temperature of 400 • C, source temperature of 110 • C, capillary voltage of 3.20 kV, desolvation gas (N 2 ) flow of 520 L h −1 , and cone gas (N 2 ) flow of 82 L h −1 . In addition, multiple reaction monitoring (MRM) and a cone voltage of 45 V were utilized. Standard sample mass spectrograms of fomesafen and nicosulfuron were listed in Figure S1.

Statistical Analyses
Except for gene sequencing experiments, all data from twice-repeated experiments were analyzed via ANOVA (SPSS v19.0, IBM, Armonk, NY, USA). The data generated from two runs were pooled, as the test for homogeneity of variance showed that the variance across runs was similar.
SigmaPlot (Version12.5; SigmaPlot Software Inc., San Jose, CA, USA) was used for further analysis through the following equation: where b is the relative slope around the herbicide dose resulting in 50% growth inhibition or 50% ALS activity inhibition, c is the lower limit, and d is the upper limit. In the regression equation, the independent variable (x) was the herbicide rate, the dependent variable (y) was the growth response (percentage of the untreated control) or ALS activity (percentage of untreated control), and the ED 50 represented the GR 50 (the dose causing a 50% dry weight growth reduction in the aboveground) or I 50 (the dose causing 50% ALS activity inhibition). The resistance index (RI) was calculated by dividing the GR 50 or I 50 value of the resistant population by that of the susceptible population.
GR 50 reduction was calculated by calculating the GR 50 value for fomesafen or nicosulfuron alone minus the GR 50 value for the application of fomesafen or nicosulfuron plus inhibitors divided by the GR 50 value for the application of fomesafen or nicosulfuron alone, and the result was expressed as a percentage.
In metabolic assay experiments, the residual dose was residual herbicide in A. retroflexus measured via HPLC-MS/MS, uptake dose was calculated using the total herbicide applied in leaves minus herbicide residue in acetonitrile washed off, and metabolism dose was calculated using the uptake dose minus the residual dose.
All the data are presented as means of replicates ± standard error (SE), and the means were separated using Fisher's protected least significant difference (LSD) test at the p < 0.05 significance level.

Multiple Resistance to Fomesafen and Nicosulfuron and Target-Site Mutation(s) Identification in A. retroflexus
Dose-response experiments confirmed that the suspected resistant population HW-01 (R) was highly resistant to fomesafen, while the ST-1 (S) population was sensitive ( Table 3). The GR 50 values of fomesafen to the ST-1 and HW-01 populations were 4.6 g ha -1 and 235.7 g ha -1 , respectively, which results in RI values (HW-01/ST-1) of 51-fold resistance.
In addition, the GR 50 values of nicosulfuron were 3.7 g ha −1 for ST-1 and 19.2 g ha −1 for HW-01, indicating the RI value was 5.2-fold and showed moderate resistance. In addition, gene sequencing revealed that all 20 detected plants of the HW-01hold Arg-128-Gly (AGG to GGG) mutation in the PPX2 gene compared with the susceptible population ( Figure 1A). However, no single nucleotide polymorphism was observed in the amplified fragments of the HW-01 population compared with the PPX1 gene sequence of the ST-01 population. However, there was a remarkable phenomenon that a considerable percentage (12/20) of the HW-01 population had no mutation (Ala-205-Val, GCT to GTT) in the ALS gene and developed resistance to nicosulfuron ( Figure 1B). To further explore the TSR of the HW-01 population to nicosulfuron, it is necessary to determine the activity of ALS in vitro.
where b is the relative slope around the herbicide dose resulting in 50% growth inhibition, c is the lower limit, d is the upper limit, x is the herbicide dose, and y is the growth response (percentage of the untreated control); GR 50 , dose required to reduce plant dry weight by 50%. d RI = GR 50 (HW-01)/GR 50 (ST-1).
Biology 2023, 12, x 6 of 16 All the data are presented as means of replicates ± standard error (SE), and the means were separated using Fisher's protected least significant difference (LSD) test at the p < 0.05 significance level.

Multiple Resistance to Fomesafen and Nicosulfuron and Target-Site Mutation(s) Identification in A. retroflexus
Dose-response experiments confirmed that the suspected resistant population HW-01 (R) was highly resistant to fomesafen, while the ST-1 (S) population was sensitive ( Table  3). The GR50 values of fomesafen to the ST-1 and HW-01 populations were 4.6 g ha -1 and 235.7 g ha -1 , respectively, which results in RI values (HW-01/ST-1) of 51-fold resistance. In addition, the GR50 values of nicosulfuron were 3.7 g ha −1 for ST-1 and 19.2 g ha −1 for HW-01, indicating the RI value was 5.2-fold and showed moderate resistance. In addition, gene sequencing revealed that all 20 detected plants of the HW-01hold Arg-128-Gly (AGG to GGG) mutation in the PPX2 gene compared with the susceptible population ( Figure 1A). However, no single nucleotide polymorphism was observed in the amplified fragments of the HW-01 population compared with the PPX1 gene sequence of the ST-01 population. However, there was a remarkable phenomenon that a considerable percentage (12/20) of the HW-01 population had no mutation (Ala-205-Val, GCT to GTT) in the ALS gene and developed resistance to nicosulfuron ( Figure 1B). To further explore the TSR of the HW-01 population to nicosulfuron, it is necessary to determine the activity of ALS in vitro.

In Vitro Assay of ALS for Nicosulfuron Activity
The results of the in vitro ALS assay showed that the total ALS activity between the HW-01 and ST-1 populations was approximately similar in the absence of nicosulfuron (15.11 ± 0.48 nmol of acetoin mg −1 protein min −1 and 16.78 ± 0.53 nmol of acetoin mg −1 protein min −1 , respectively). However, the addition of nicosulfuron in the reaction almost completely inhibited ALS activity (when concentration ≥ 2 µM; Figure 2 − logED50)]}, where b is the relative slope around the herbicide dose resulting in 50% growth inhibition, c is the lower limit, d is the upper limit, x is the herbicide dose, and y is the growth response (percentage of the untreated control); GR50, dose required to reduce plant dry weight by 50%. d RI = GR50 (HW-01)/GR50 (ST-1).

In Vitro Assay of ALS for Nicosulfuron Activity
The results of the in vitro ALS assay showed that the total ALS activity between the HW-01 and ST-1 populations was approximately similar in the absence of nicosulfuron (15.11 ± 0.48 nmol of acetoin mg −1 protein min −1 and 16.78 ± 0.53 nmol of acetoin mg −1 protein min −1 , respectively). However, the addition of nicosulfuron in the reaction almost completely inhibited ALS activity (when concentration ≥ 2 µM; Figure 2) in both S and HW-01 plants. The I50 values of HW-01 and ST-1 plants were 0.24 µM and 0.076 µM, respectively, which results in a RI value that is 3.2-fold lower (Figure 2).

Impact of P450 and GST Inhibitors on Fomesafen and Nicosulfuron Resistance
The P450 and GST inhibitors used alone had no significant effects on the plant growth of either the HW-01 or ST-1 populations. However, pre-treatment with malathion, PBO, and amitrole significantly increased the toxicity of fomesafen to HW-01 plants, with GR 50 values being significantly reduced by 83%, 80%, and 68%, respectively (Table 4, Figure 3). By comparison, the pre-treatment of malathion, PBO, and amitrole caused GR 50 values of fomesafen to reduce by 1.5%, 5.9%, and 5.1%, respectively, in the ST-1 population (Table 4, Figure 3). For the effect of P450 and GST inhibitors on nicosulfuron sensitivity in both populations, results showed that the GR 50 reductions seen in HW-01 plants caused by malathion, PBO, and amitrole plus nicosulfuron were 66%, 53%, and 60% (Table 4, Figure 4), respectively, compared to nicosulfuron alone. In contrast, no significant reductions in the GR 50 values of nicosulfuron were observed in the ST-1 population with or without the pre-treatment of malathion, PBO, and amitrole (GR 50 reductions were 8.6%, 3.2%, and 3.3%, respectively, Table 4, Figure 4). These results indicated that the cytochrome P450 inhibitors PBO, amitrole, and malathion had effects on the fomesafen and nicosulfuron resistance in A. retroflexus. Moreover, GST-inhibiting NBD-Cl pre-treatment led to fomesafen GR 50 values being reduced by 75% and 3.4% (Table 4, Figure 3) for the HW-01 population and ST-1 population, respectively. Additionally, the pre-treatment of NBD-Cl led to nicosulfuron GR 50 values being reduced by 64% and 5.3% (Table 4, Figure 4) for the HW-01 population and ST-1 population, respectively. Therefore, these findings suggest that the enhanced  of either the HW-01 or ST-1 populations. However, pre-treatment with malathion, PBO, and amitrole significantly increased the toxicity of fomesafen to HW-01 plants, with GR50 values being significantly reduced by 83%, 80%, and 68%, respectively (Table 4, Figure 3). By comparison, the pre-treatment of malathion, PBO, and amitrole caused GR50 values of fomesafen to reduce by 1.5%, 5.9%, and 5.1%, respectively, in the ST-1 population (Table  4, Figure 3). For the effect of P450 and GST inhibitors on nicosulfuron sensitivity in both populations, results showed that the GR50 reductions seen in HW-01 plants caused by malathion, PBO, and amitrole plus nicosulfuron were 66%, 53%, and 60% (Table 4, Figure  4), respectively, compared to nicosulfuron alone. In contrast, no significant reductions in the GR50 values of nicosulfuron were observed in the ST-1 population with or without the pre-treatment of malathion, PBO, and amitrole (GR50 reductions were 8.6%, 3.2%, and 3.3%, respectively, Table 4, Figure 4). These results indicated that the cytochrome P450 inhibitors PBO, amitrole, and malathion had effects on the fomesafen and nicosulfuron resistance in A. retroflexus. Moreover, GST-inhibiting NBD-Cl pre-treatment led to fomesafen GR50 values being reduced by 75% and 3.4% (Table 4, Figure 3) for the HW-01 population and ST-1 population, respectively. Additionally, the pre-treatment of NBD-Cl led to nicosulfuron GR50 values being reduced by 64% and 5.3% (Table 4, Figure 4) for the HW-01 population and ST-1 population, respectively. Therefore, these findings suggest that the enhanced herbicide metabolism mediated by P450s and/or GSTs contribute to the fomesafen and nicosulfuron resistance in A. retroflexus.

Fomesafen and Nicosulfuron Metabolism in A. retroflexus
The HPLC/MS-MS results showed that fomesafen and nicosulfuron absorption were not significantly differed in the R and S plants during the experiment, and over time, fomesafen and nicosulfuron absorption increased in all the tested plants (Tables 5 and 6).

Dose Response to Other Herbicides
This study also determined the sensitivity of A. retroflexus to PPO inhibitors (fluoroglycofen-ethyl, acifluorfen, lactofen, flumioxazin, and cloransulam-methyl) and a PSII inhibitor (bentazone). The ST-1 population was susceptible to all the herbicides used in the present study. Based on the RI values (Table 3), the HW-01 population was 9.0-fold more resistant to fluoroglycofen-ethyl relative to the ST-1 population, 7.4-fold more resistant to acifluorfen, 9.6-fold more resistant to cloransulam-methyl, and 8.4-fold more resistant to bentazone compared with ST-1. In addition, RI values of 3.8 and 4.0 were observed for the HW-01 population for flumioxazin and lactofen, respectively, and compared to the RI values of other herbicides, the HW-01 population did not develop strong resistance to these two herbicides. As a result, HW-01 plants developed resistance to all the above-tested herbicides, but the response was various.

Discussion
PPO inhibitors, especially in China, have been used for approximately 50 years since their introduction into soybean to control broadleaf weed species [35]. For the resistance evolution of PPO inhibitors in weeds worldwide, the first PPO inhibitor-resistant A. tuberculatus population was reported from a field with a long history of continuous soybean production and repeated selection via acifluorfen in Kansas, USA [36]. The HW-01 population tested in this study originated from soybean fields with a history of continuous use of fomesafen or nicosulfuron for over 10 years, and the 50.1-fold increase in the resistance of the resistant A. retroflexus population HW-01 to fomesafen is most likely due to selection pressure caused by the continuous application of this herbicide. In addition, Lamego et al. demonstrated that the continuous use of ALS inhibitors resulted in the development of herbicide resistance in Bidens Subalternans [37]. Additionally, Ma et al. also pointed out that the concentrated use of herbicides often leads to the selection of genes that confer herbicide resistance in weed populations [38]. Therefore, after long-term screening with fomesafen and nicosulfuron, plants with herbicide-resistant genes may gradually become the main body of the HW-01 population through reproduction, leading to the development of resistance to these two herbicides in this population.
The in vitro ALS activity assay showed that the ALS was less susceptible to nicosulfuron in the HW-01 population than in the ST-1 population. However, the altered ALS sensitivity to nicosulfuron in HW-01 was not fully consistent with the level of nicosulfuronresistance tested in the whole-plant dose-response experiment. This phenomenon may be attributed to several factors: firstly, the samples of the in vitro ALS activity assay were randomly harvested from the population level, and the preceding target gene sequences analysis observed that HW-01 plants had a relatively low frequency of mutations in the ALS gene (eight/twenty, mutations/total plants); secondly, NTSR in the HW-01 population cannot be ignored and most likely played a vital role in the resistance of A. retroflexus to nicosulfuron and fomesafen. Therefore, here, we further investigated the NTSR mechanism in A. retroflexus.
The cytochrome P450 inhibitors used in this study included the organophosphate insecticide malathion, the synergistic chemical PBO, and the herbicide amitrole, which all inhibit plant P450s and are known to target different P450s enzymes [31,39]. Ma et al. and Oliveira et al. reported that malathion combined with mesotrione, tembotrione, or topramezone increased biomass reduction and herbicide efficacy in HPPDresistant A. tuberculatus, which confirmed the enhanced metabolism (NTSR) mechanism in the HPPD-resistant A. tuberculatus [31,33]. Varanasi et al. also documented that the fomesafen resistance could be partially reversed by P450-inhibitor malathion and GST-inhibitor NBD-Cl in A. palmeri [28]. In this study, the P450 inhibitors (malathion, PBO, and amitrole) and GST inhibitor (NBD-Cl) drastically reversed the resistance of the HW-01 plants to fomesafen and nicosulfuron, suggesting the presence of NTSR mediated by P450s and GSTs in HW-01 which contributes to the herbicide-resistant phenotype by enhancing herbicide metabolism. To the best of our knowledge, this is the first report on NTSR to fomesafen and nicosulfuron resistance in A. retroflexus, and with the continued use of multiple sites of action (SOA) herbicides, likely, populations of A. retroflexus in this dual resistance state (TSR coexisting with NTSR) will become more numerous. Generally, NTSR confers unpredictable cross-resistance patterns, which may threaten the possibility of herbicide mixing in delaying resistance evolution [40,41]. Worse, a new herbicide MOA is not expected to occur in the near future [42]. Therefore, NTSR increases the difficulty of herbicide control, based on the emergence of TSR in weeds. From this point of view, the multiple-herbicide-resistant A. retroflexus may be becoming troublesome in weed management in the soybean and/or corn production regions of China.
The use of P450 or GST inhibitors (malathion, PBO, amitrole, NBD-Cl) may be part of a future solution for NTSR, and it opens an avenue of research that warrants further exploration. Studies have demonstrated that synergists can reverse herbicide resistance [43,44]. However, the main problem with P450 or GST inhibitors is that these molecules may also reduce crop selectivity and have unintended environmental effects. In this study, the use of herbicides with P450 and GST inhibitors significantly reduced GR 50 values compared to herbicide application alone (p < 0.05), and it confirmed the existence of a metabolic resistance phenotype in A. retroflexus. However, how P450s and GSTs mediate herbicide metabolism and which enzyme family members are involved in metabolic resistance merits need further investigation.
There have been many cases of mutations in ALS and PPX2 in Amaranthus leading to their resistance to ALS and PPO inhibitors and more and more target-site mutations have been reported [17,18,21]. However, with the discovery of the first case of amaranth with non-target resistance [45], more and more studies have confirmed that non-target resistance plays an important role in the resistance of amaranth to herbicides, among which Varanasi et al. demonstrated that the population of palmer amaranth that developed non-target resistance delivered cross-resistance to PPO inhibitors, and this population developed some degree of resistance to herbicides of multiple-herbicide mechanisms of action (including ALS inhibitors) [28]. Rangani et al. revealed that palmer amaranth developed metabolic resistance to s-metolachlor in which the GST gene family played an important role [46]. Küpper et al. highlighted the involvement of the P450 enzyme family in the palmer amaranth detoxification of tembotrione [47]. Obenland et al. experimentally confirmed that resistance to carfentrazone-ethyl in A. tuberculatus arises due to the production of NTSR within this population [48]. The resistant population in this study, HW-01, experimentally confirmed the involvement of P450s and GSTs in the detoxification metabolism of nicosulfuron and fomesafen. The presence of NTSR in HW-01 may be one of the reasons for the formation of multiple resistance patterns. It has demonstrated that using low or sublethal herbicide rates was one of the main reasons attributed to the evolution of metabolic resistance, resulting in the accumulation of metabolic genes over several generations and resistance to different types of herbicides [25,40].
The HW-01 population already had a high index of resistance to PPO, ALS, and PSII inhibitors (Table 3), and the population developed multiple resistance to herbicides belonging to these chemical groups, probably because the widespread use of these three types of herbicides in soybean-corn rotation fields contributed to the formation of NTSR in HW-01. The herbicide resistance caused by multiple target-site mutations has also been reported in A. retroflexus and A. palmeri [17,49]. Multiple resistance due to more than one target gene resistance mutation has also been reported in lots of weed species, such as Alopecurus Linn and Echinochloa Beauv [50][51][52]. However, there was no history of a largescale use of flumioxazin in China's soybean/corn fields [53], and the HW-01population showed low resistance to flumioxazin, so growers should be vigilant about the use of flumioxazin to prevent overuse that could lead to resistance to the herbicide in A. retroflexus. In previous studies, it was also reported that wild radish evolved low-level resistance to diflufenican after only four applications [54,55].
The reasons for the results of this investigation were most likely diverse. In terms of TSR, a mutation in the PPX2 gene (Arg-128-Gly) was confirmed in the HW-01 pop-ulation, which could be a reason for conferring cross-resistance to flumioxazin; PPX1 was compared and no mutation was found; however, whether there is an overexpression of PPX1 in the HW-01 population causing TSR still needs further experimental verification. Additionally, we have demonstrated the presence of NTSR in the HW-01 resistant population; therefore, enhanced cytochrome P450-and GST-mediated herbicide metabolism may be another reason for HW-01 resistance to flumioxazin. However, NTSR has different metabolic effects on different herbicides, so whether NTSR is involved in the metabolic detoxification of HW-01 to flumioxazin and to what extent it plays a role still needs further experimental confirmation.

Conclusions
In summary, in this study, we identified one A. retroflexus population, which has evolved multiple resistance to PPO, ALS, and PSII inhibitors. The multiple resistance to fomesafen (PPO inhibitor) and nicosulfuron (ALS inhibitor) was endowed by P450-and GST-mediated enhanced herbicide metabolism and target-site mutations. For resistance development, the specific herbicide application history in the soybean and corn rotation of agricultural practice likely facilitated the rapid evolution of ALS and PPO inhibitor resistance in A. retroflexus [56]. This phenomenon poses a threat to the chemical control of weeds in crop fields, and the "co-evolution" of the target gene and metabolism for herbicide resistance should be paid attention to and further understood.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biology12040592/s1. Figure S1: Typical chromatograms of fomesafen and nicosulfuro-standard and extracted Amaranthus retroflexus L samples [HW-01 population (R) and ST-1 population (S)]. Table S1: Primers used to amplify the PPX1, PPX2, and ALS genes of A. retroflexus. Table S2: The analysis method of fomesafen and nicosulfuron in A. retroflexus using HPLC-MS/MS was validated in terms of linearity, limit of quantification (LOQ), accuracy, and precision.