Detection of Target-Site Herbicide Resistance in the Common Ragweed: Nucleotide Polymorphism Genotyping by Targeted Amplicon Sequencing

: Background: The spread of herbicide-resistance Ambrosia artemisiifolia threatens not only the production of agricultural crops, but also the composition of weed communities. The reduction of their spread would positively affect the biodiversity and beneﬁcial weed communities in the arable habitats. Detection of resistant populations would help to reduce herbicide exposure which may contribute to the development of sustainable agroecosystems. Methods: This study focuses on the application of target-site resistance (TSR) diagnostic of A. artemisiifolia caused by different herbicides. We used targeted amplicon sequencing (TAS) on Illumina Miseq platform to detect amino acid changes in herbicide target enzymes of resistant and wild-type plants. Results: 16 mutation points of four enzymes targeted by four herbicide groups, such as Photosystem II (PSII), Acetohydroxyacid synthase (AHAS), 5-enolpyruvylshikimate 3-phosphate synthase (EPSPS) and IX oxidase (PPO) inhibitors have been identiﬁed in common ragweed populations, so far. All the 16 mutation points were analyzed and identiﬁed. Out of these, two mutations were detected in resistant biotypes. Conclusions: The applied next-generation sequencing-targeted amplicon sequencing (NGS-TAS) method on A. artemisiifolia resistant and wild-type populations enable TSR detection of large sample numbers in a single reaction. The NGS-TAS provides information about the evolved herbicide resistance that supports the integrated weed control through the reduction of herbicide exposure which may preserve ecological properties in agroecosystems. for tryptophan substitution at amino acid position 574 in AHAS enzyme was responsible for the imazethapyr herbicide resistance. Amino acid substitution at the position 219 (valine/isoleucine) in psbA gene product, D1 protein, was responsible for the linuron resistance. Amino acid substitutions in EPSPS and PPX2 enzymes at known resistance-conferring mutations sites were not found.


Introduction
Extensive weed management in agricultural fields is a major threat to agroecosystem biodiversity as weed communities play a key role in maintaining heterogeneity. The total number of weed species in agricultural fields has increased since the 1950s. Although this has resulted in a more diverse composition of agricultural weeds, it has negatively affected landscape structure and land-use intensity. As a result, crop intensification methods, such as herbicide treatment, have become critical to control weed communities to improve agricultural crop production [1]. It is important not to endanger beneficial weed species while using weed control. To define the correct management of invasive species, local environmental drivers have to be identified due to their severe effects on individual performance [2].
Common ragweed (Ambrosia artemisiifolia L.: Asteraceae, A. artemisiifolia) is a highly adaptive, invasive weed and has a negative impact on biodiversity [3,4]. Its population can be found not only in agricultural habitats where herbicides are normally used but also in field margins and paths, abandoned fields, and forest edges that could be affected by herbicide or intensified mowing [5,6]. Common ragweed has spread from North America to habitats in temperate climates such as Eurasia, Asia and Australia [7,8], and has become distributed in all countries of Southern and Eastern Europe [9][10][11][12]. In North America large common ragweed populations have extensively arisen along roadsides and arable fields as well, which was recently observed mainly in Eastern Canada [13].
Many weed scientists deliberate on the knowledge and practice gap of the perspectives and approaches in weed ecology and management [14][15][16]. Due to the complexity of the problem, it requires transdisciplinary. Transdisciplinary approaches contribute to the most promising technological improvements managing weed control and agroecosystem diversity and reduce the potential unwanted environmental impacts of weed management. The negative correlation between weed diversity and crop yield loss has been reported [17]. As a consequence of the diverse weed management and crop production systems, the more adaptable weedy plants may replace the invasive species and the given ecosystems benefit [15]. Therefore, a well-thought herbicide use strategy should be effective and reasonable. The weed thresholds are required to promote the usage of optimal herbicides to protect crop yields and arable habitats [18].
To improve weed management and avoid ecological damages, monitoring the different herbicide-resistance in a population seems to be a great solution to find the optimal chemical crop protection. In this study, we focus on the subject of the common ragweed chemical control. Our primary goal is to achieve the controlled eradication of ragweed or at least to maintain its low moderate level while minimizing the ecological damage caused by the applied herbicide.
Common ragweed has evolved different herbicide resistance modes, making it particularly problematic to control [19]. Herbicide resistance can be categorized into two categories based on mode of action: target-site resistance (TSR) and non-target-site resistance (NTSR) [20][21][22]. Here, we focus on TSR resistance, which can be caused by mutation or polymorphism that results in either increased abundance of the target protein or a structural change in the target protein that reduces affinity to the herbicide [23,24]. Structural changes are generally due one or several amino acid substitutions [25][26][27][28][29]. The accumulation of multiple amino acid substitutions in a target protein has been known as multiple target-site resistance (MTSR) which triggers an increased resistance level of individual plants [22,30,31]. MTSR has been detected in different weeds conferring resistance to Photosystem II (PSII) and Acetohydroxyacid synthase (AHAS) inhibitors [32][33][34]. Resistant genotypes proliferate in areas where herbicides are used for weed control. In order to effectively combat herbicide-resistant common ragweed, effective methods to measure the prevalence of TSR and MTSR within target populations are needed.
PS II inhibitors bind to the D1 protein of the photosystem II (EC: 1.10.3.9) complex, blocking electron transport and stopping CO 2 fixation. The D1 protein is encoded by the chloroplast psbA gene. The first documented herbicide resistant population of A. artemisiifolia Diversity 2021, 13, 118 3 of 16 was discovered in 1976 and was resistant to the PSII-inhibitor atrazine (a triazines) [35]. Resistance was conferred by a point mutation in the psbA gene resulting in a serine-toglycine substitution, which alters D1 protein conformation, causing reduced herbicide target interaction [36,37]. To detect the causal mutation (Ser264Gly) in common ragweed, a bidirectional allele-specific PCR (Bi-PASA) method was developed [38]. Resistant biotypes and populations have also been reported against linuron (belonging to ureas) in eastern Canada [39,40]. TSR, conferred by the psbA mutation Val219Ile, was detected after linuron treatment using TaqMan single nucleotide polymorphism (SNP) genotyping and Sanger sequencing [41]. Moreover, three additional mutations at the site of action (Ala251Val, Phe255Ile, and Asn266Thr) have been confirmed in populations of other species, each with a unique cross resistance pattern [42].
AHAS inhibitors are among the most used common ragweed inhibitors worldwide. AHAS (EC 4.1.3.18 also known as acetolactate synthase, ALS) catalyses the first common step in the pathway for synthesis of the branched-chain amino acids leucine, isoleucine, and valine in plants [43]. AHAS inhibitors have low mammalian toxicity, are broad spectrum and highly effective in low doses. However, resistance to AHAS inhibitors tends to arise rapidly in weed populations. Multiple studies have reported that AHAS inhibitor-resistant weed species are more numerous than weed species resistant to other herbicide groups [44,45]. Many AHAS inhibitor-resistant common ragweed populations have been reported. Cross resistance in these populations is common, with resistant individuals displaying resistance to additional herbicides, including cloransulam-methyl (triazolopyrimidines), chlorimuron (sulfonylureas), and imazaquin (imidazolinones) [46]. Resistance in common ragweed against AHAS inhibitors is conferred by the substitution Trp574Leu [47]. Although eight different AHAS amino acid substitutions confer resistance to AHAS inhibitors in other species, only Trp574Leu has been reported in common ragweed [20,48].
Several polymorphisms in the epsps gene of many plant species confer resistance to glyphosate. Resistance is conferred by mutations at Pro106 and Thr102, and recently an example of MTSR was reported in Amaranthus hybridus that contained the triple-substitution Thr102Ile, Ala103Val and Pro106Ser [49][50][51][52]. Glyphosate resistance in A. artemisiifolia was first reported in Missouri and Arkansas in 2004 but the mechanism of the resistance has not yet been determined [53].
PPO inhibitors have been used widely to control weeds in different crop cultures since the 1960s. PPO (EC 1.3.3.4) is a key enzyme in the tetrapyrrole biosynthetic pathway that produces heme and chlorophyll in plastids and mitochondria [54]. Two different nuclear-encoded PPO isozymes are in plants: plastidal PPO1 encoded by the gene ppx1 and mitochondrial PPO2 encoded by the ppx2 gene [55]. The weed reported to evolve resistance to PPO inhibitors was Amaranthus tuberculatus, which contained a glycine deletion at position 210 (∆G210) in PPX2 (GenBank accession: DQ386114) [56]. In A. artemisiifolia, an Arg98Leu substitution in the ppx2 gene contributes to PPO resistance [57].
Multi-resistant A. artemisiifolia populations can be found worldwide, containing multiple resistance-conferring mutations in genes encoding PSII, AHAS, EPSPS, and PPO [46]. The investigation of specific mutations conferring herbicide resistance in populations is essential for the selection of effective herbicides for weed control. Targeted amplicon sequencing (TAS) is an effective approach that takes advantage of NGS (next generation sequencing) to detect specific mutations. To sequence large numbers of targeted gene regions, NGS is using indexed primers to label samples that are expansively used for population genetics and for genome-wide genotyping in plants with ultralow SNP densities [58].
This study presents an NGS-TAS method for TSR identification in A. artemisiifolia. In order to discover multiple polymorphisms in psbA, ahas, epsps, and ppx2 that confer resistance, we analysed sensitive biotypes, as well as imazethapyr-resistant (imidazolinones, AHAS inhibitor) and linuron-resistant (ureas, PSII inhibitor) biotypes. This method may contribute to the sustainable maintenance of biological integrity in agroecosystems by screening herbicide resistant common ragweed populations. Seedlots were sampled from sites where ragweed was tested and identified as imazethapyr and linuron resistant. Resistance seeds were germinated in petri dishes at 30 • C/20 • C 16 h photoperiod and were planted in pots and were growing in greenhouse. Leaves of 20-20 planted individuals were further collected. Samples were frozen in liquid nitrogen immediately after collection and stored at −80 • C until DNA and RNA analysis.
Total DNA from leaf tissues was isolated according to the protocol Doyle, J. J. and Doyle, J. L. [59]. The DNA quantity and purity were assessed on NanoDrop2000 (Thermo Scientific, Waltham, MA, USA). For the experiments, DNA of resistant and sensitive biotypes were used according to the following classification: (i) bulked sample of 20 sensitive genotypes deriving from Hungarian sampling site; (ii) bulked sample of 20 imazethapyr (imidazolinones) resistant genotypes collected from Mirabel sampling site (AMI) and (iii) bulked sample of 20 linuron (ureas) resistant genotypes collected from Ste-Clotilde Legault sampling site (AMU). All individuals, both susceptible and resistant genotypes, were analyzed by Sanger method.

Identification of A. artemisiifolia Herbicide Target Enzyme mRNAs
In order to identify in silico coding sequences (cds) of herbicide target enzymes of each herbicide group, we used the transcript dataset of A. artemisiifolia, deposited in the National Centre for Biotechnology Information's (NCBI) Transcriptome Shotgun Assembly (TSA) dataset under the accession GEZL01000001 [60]. Reference genes were selected based on taxonomic relationship from closely related species and were downloaded from the UniProt database: A. artemisiifolia, psbA (B5MF75); Xanthium sp., ahas (Q41716); Cichorium intybus, ppx1 (Q9SPL6); Amaranthus palmeri, ppx2 (A0A291B3V5); Erigeron canadensis, epsps (G4U4J5) [61]. The identity of reference genes and identified genes was 96-99% [62]. The investigated sequences were deposited in the NCBI GenBank database (Accession numbers see below).

Intron Analysis, Amplification and Cloning of A. artemisiifolia Herbicide Target Enzyme Genes
Multiple primer pairs were designed in order to cover all possible mutation points along the whole psbA and ahas gene sequences by using Primer3Plus [63] (Figures S1 and S2).
In the case of intron containing epsps gene, the exon-intron bounds were predicted based on complete gene sequences of closely related species by using ClustalW aligning the sequences of Conyza canadensis (AY545667.1) and Amaranthus palmeri (FJ861242.1 and FJ861243.1) [64] ( Figure S3).
In order to validate polymorphisms that characterize herbicide resistance, the imazethapyr, linuron resistant, and sensitive bulked samples were amplified. The PCR amplifications was performed by using Dream Taq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA) The PCR products were separated on 1.5% agarose gel (Promega, Madison, WI, USA) and were purified using NucleoSpin Gel and PCR Clean-up system (Macharey-Nagel GmbH & Co, Düren, Germany). The pGEM-T Easy Vector System kit (A1360 Promega, Madison, WI, USA) and JM109 Competent Cells were used to clone PCR products [65]. Sanger sequencing of the cloned fragments was performed with ABI 3130xl Genetic Analyzer (Applied Biosystems, Carlsbad, CA, USA). Sequenced fragments of herbicide target genes were aligned to predicted gene sequences certifying amino acid substitutions in the mutation points.

PCR Amplification for NGS-TAS Experiments
The KAPA HiFi HotStart PCR Kit (Kapa Biosystems, Wilmington, NC, USA) was used for amplification of target fragments. In NGS-TAS experiments of investigated regions of psbA, ahas, epsps, and ppx2 genes, primers were designed by using Primer3Plus ( Figure S7).
The PCR amplifications were performed and the PCR products were separated on 1.5% agarose gel (Promega, Madison, USA). PCR products were controlled with Agilent 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany) using Agilent DNA 1000 Kit and were purified using NucleoSpin Gel and PCR Clean-up system (Macharey-Nagel GmbH &Co, Düren, Germany).

NGS-TAS Experiments
For NGS-TAS experiments, resistant samples were used and grouped with the following nomenclature.
Sample AMI was named based on the imidazolinone (imazethapyr) resistant bulked DNA of A. artemisiifolia and sample AMU was named based on the urea (linuron) resistant bulked DNA of A. artemisiifolia. Numbers 1-3 were three biological repeats of the same bulked DNA (AMI1-3, AMU1-3). The concentrations of PCR products were measured by NanoDrop2000 (Thermo Scientific, Waltham, USA) and fragments for each sample were diluted relative to the PCR product with the lowest concentration. AMI1control and AMU1control contained the not diluted samples of the first biological repeats. In the AMI and AMU groups bulked DNA of 20 individuals were used.
In order to obtain long fragments that cover all the mutation points in each gene, fragments were sequenced on MiSeq 550 platform obtaining 300 bp paired-end reads detailed as follows: locus-specific PCR products were purified using 1.  Figure S7. PCR cycling parameters for index PCRs were as follows: initial denaturation at 95 • C for 3 min; 8 cycles at 95 • C for 30 s, 55 • C for 30 s, 72 • C for 30 s; final extension at 72 • C for 5 min. PCR products were purified using 1.0 volume KAPA PureBeads and eluted in 20 µL of 10 mM Tris-HCl pH 8. The product libraries were quantified and qualified by using High Sensitivity D1000 ScreenTape on TapeStation 2200 instrument (Agilent Technologies, Santa Clara, CA, USA) ( Figure S11). Equimolar concentrations of libraries were pooled, diluted to 4 nM, and combined with other sample pools to gain the desired sequencing depth. Sequencing was carried out using Illumina MiSeq platform and 600-cycle Reagent Kit v3 (Illumina, Inc. San Diego, CA, USA). Samples were demultiplexed and adapter-trimmed by using MiSeq Control Software.

Bioinformatics Analysis
Pear software [66] was used to merge paired-end reads with minimum overlap size 10 bp, without discard uncalled bases, and with 30 parallel threads in processing (arguments -v 10 -u 1 -j 30). All the samples contained the reads amplified by the 7 primer pairs covering all the mutation points. Accordingly, the reads belonging to each fragment were counted in each sample ( Figure S12). Usearch software [67] was applied to collect identical sequences for resistance polymorphism. Usearch was used in three steps: (i) Quality filter with fastq_filter command with fastq_maxee 1.0 and fastq_minlen 160 parameters; (ii) Singleton filter with fastq_uniques command; (iii) Grouping with cluster_otus parameter with minsize 10. Groups were shown the similarity threshold 99% representing the different amplicon products with nucleotide polymorphism for the investigated gene.

Results
In order to perform NGS-TAS analysis, it would be necessary to know cds and gene sequences of herbicide target enzymes. However, no whole genome sequence data of A. artemisiifolia yet exists. To design correct NGS-TAS primer pairs, it is essential to know at least the sequences at the intron-exon boundaries close to the mutation points. Accordingly, gene sequences and point mutations of the herbicide target enzymes were determined in sensitive genotypes by using in silico NGS transcriptome analysis (NGS-TA), SS Sanger sequencing, and NGS-TAS (Figure 1).
Diversity 2021, 13, x FOR PEER REVIEW 7 of 1 Figure 1. Flow chart of target-site resistance (TSR) diagnostic method using the following steps: (i Next generation sequencing (NGS) Transcriptome analysis (NGS-TA). Open reading frames of target enzymes were identified based on transcriptome database; (ii) Sanger sequencing method (SS). Specific primers were designed covering the mutation points, fragments were amplified by PCR, PCR products were cloned and sequenced; (iii) Next-generation sequencing-Targeted amplicon sequencing (NGS-TAS) method. PCR products were sequenced on Illumina MiSeq550 platform resulting 300bp paired-end reads that were sorted based on primer motifs, assembled, and grouped.

Identification of A. artemisiifolia Herbicide Target Enzyme cds and Genes
In silico analysis of investigated herbicide target enzyme cds was performed with BLAST+. The accession numbers of predicted GOIs are MT425203 (psbA), MK096760 (ahas), MK096765 (epsps), and MK096762 (ppx2). Open reading frame (ORF) sequences were predicted using NCBI ORF finder [68]. The predicted sequence of the psbA gene consisted of one exon with a total length of 1062bp, as previously described (Table 1) [37]. The psbA gene was amplified with 2 primer pairs covering the resistance mutation points as described before ( Figure S1, GenBank accession MT879746).
The structure of the ahas gene contained one exon with a total length of 1965bp. The whole ahas gene was amplified with five primer pairs covering the resistance mutation points as described before (Figures S2 and S4, GenBank accession MT415954) [20,47].
The first published weed species to evolve resistance to PPO-inhibiting herbicides was Amaranthus tuberculatus [56]. The ppx2 gene (GenBank accession DQ386118) from the resistant biotype of A. tuberculatus contained a glycine deletion at position 210 (∆G210) that was shown to confer resistance. This deletion in A. artemisiifolia has not been published, yet, although the substitution Arg98Leu was discovered to contribute to PPO-inhibitor resistance A. artemisiifolia [57]. So far, no complete gene sequence exists of the whole ppx2 gene in plants covering all of the introns and exons. Therefore, identifying the complete sequence of ppx2 gene is not possible only using Sanger sequencing. By using the NGS-TAS, a 324 bp fragment of ppx2 gene was amplified with one primer pair covering the resistance mutation point Arg98 [55]. Electropherogram of Sanger sequenced PCR product revealed two product lengths. The differences between the two fragments were 13 bp localised in one intron ( Figures S4 and S5, GenBank accession MT879747, MT879748).
In order to obtain accurate information about the investigated enzymes, in silico predicted cds of Illumina data and the Sanger sequenced data of the appropriate PCR products were compared. The two technologies showed slight variability (1-2% differences) in amino acid sequences across ahas and epsps, however these discrepancies were not at mutation points. There was no difference in sequencing results of the psbA gene between the two methods ( Figure S6).

NGS-TAS Experiments of AMI and AMU Biological Repeats
In NGS-TAS experiments specific regions of psbA, ahas, epsps, and ppx2 genes were amplified by different primers (Figure 2, Figures S9 and S10). Sample AMI was named based on the imidazolinone (imazethapyr) resistant bulked DNA of A. artemisiifolia and sample AMU was named based on the urea (linuron) resistant bulked DNA of A. artemisiifolia. Both AMI and AMU contained three replicates.  In NGS-TAS experiments specific regions of psbA, ahas, epsps, and ppx2 genes were amplified by different primers (Figure 2, Figures S9 and S10). Sample AMI was named based on the imidazolinone (imazethapyr) resistant bulked DNA of A. artemisiifolia and sample AMU was named based on the urea (linuron) resistant bulked DNA of A. artemisiifolia. Both AMI and AMU contained three replicates.  Each sample contained the PCR products of psbA, ahas, epsps, and ppx2 genes. The control of AMI1-3 group was labeled as AMI1control and contained PCR products of AMI1. Similarly, AMU1control served as a control for AMU1-3 group and contained PCR products of AMU1. Both in AMI and AMU samples, the concentrations of the PCR products of the four genes were measured in each replicate, respectively. Afterwards, the lowest concentration was used as a base for further dilution of the other PCR products to avoid error in sequencing analysis (Table 2, Figure S8). In NGS-TAS analyses the samples were sequenced in four lanes. Based on the ratio of concentration/contig number of the samples, we found that all concentrations of AMI and AMU samples were optimal for the diagnostic procedure.

Bioinformatics Analysis of AMI and AMU Groups
Samples with different concentrations showed different distributions in contig number ( Table 2). As a result of amplicon sequencing the number of Illumina MiSeq 300 bp reads were 276,462 (forward), 296,722 (reverse) in case of AMI and 316,465 (forward), 369,748 (reverse) in case of AMU samples. After quality filtering dropped reads were 0.7% on average of AMI samples and 4.7% on average of AMU samples. PEAR software was used to assemble cleaned reads of which unpaired ones were filtered out with the average ratio 5.45 and 9.05% of AMI and AMU samples, respectively. Numbers of filtered and assembled reads were 255,357-287,197. In order to separate the investigated gene fragments we clustered the contigs (assembled reads) into groups based on appropriate primer motifs by using Usearch software. During clustering singletons were dropped (unique groups with one sequence). Average ratios of the survived contigs were 69.8% and 69.7%. The group number was 18 and 12 in AMI and AMU samples ( Figure S12).  (Table 2). Undiluted AMI1control and AMU1 control samples showed an equable read distribution in each fragment, therefore dilution of samples in the same concentration is recommended. Confirmation of a fragment with such a large number of reads provides an opportunity for large-sample analysis.

Detection of Mutation Points in Resistant Biotypes
To be able to identify the mutations causing resistance in A. artemisiifolia, bulked and amplified DNA sequences of imazethapyr and linuron resistant samples were further analysed by NGS-TAS that covered 16 known resistance mutations sites in A. artemisiifolia. In the AMI group, sequence alignment analysis showed that a single nucleotide in the ahas gene was changed from TGG to TTG and resulted in a Trp574Leu substitution conferring herbicide resistance [47] (Figure 3). Other amino acid substitutions in the D1 protein, EPSPS and PPX2 enzymes at known resistance-conferring mutations sites were not found. Based on these data, the investigated population was susceptible to PSII, EPSP synthase and PPO inhibitors (Figure 3). In the AMU group, sequence alignment analysis showed that a single triplet was changed (GTA to ATA) in the psbA gene and resulted in a resistance-conferring Val219Ile substitution [41] (Figure 3). Other amino acid substitutions in AHAS, EPSPS, and PPX2 enzymes at known resistance-conferring mutations sites were not found. Based on this data the investigated biotype was susceptible to AHAS, EPSP synthase, and PPO inhibitors in terms of TSR ( Figure 3).  DNA sequencing revealed that an inferred leucine for tryptophan substitution at amino acid position 574 in AHAS enzyme was responsible for the imazethapyr herbicide resistance. Amino acid substitution at the position 219 (valine/isoleucine) in psbA gene product, D1 protein, was responsible for the linuron resistance. Amino acid substitutions in EPSPS and PPX2 enzymes at known resistance-conferring mutations sites were not found.

Discussion
Chemical weed control is one of the most common ways to fight against weeds in Figure 3. Amino acid mutations in investigated gene products. Imidazolinone and urea resistant bulked samples were analyzed by NGS-TAS that covered 16 known resistance-conferring mutations sites in A. artemisiifolia. Marking: green, possible mutation sites; red, mutant amino acids. DNA sequencing revealed that an inferred leucine for tryptophan substitution at amino acid position 574 in AHAS enzyme was responsible for the imazethapyr herbicide resistance. Amino acid substitution at the position 219 (valine/isoleucine) in psbA gene product, D1 protein, was responsible for the linuron resistance. Amino acid substitutions in EPSPS and PPX2 enzymes at known resistance-conferring mutations sites were not found.

Discussion
Chemical weed control is one of the most common ways to fight against weeds in agricultural areas, giving a high selection pressure for resistant biotypes. After herbicide treatment, the rapid diagnosis of resistance mutations present in surviving individuals using would inform and improve the effectiveness of the next herbicide treatment. In order to maintain sustainability in crop fields and beneficial weed communities that are required for natural ecosystems, the monitoring of resistant common ragweed populations between treatments is important. In this study, we present a molecular biology-based approach (NGS-TAS) to detect herbicide resistant common ragweed populations in arable habitats [69,70].
It is important that only a few herbicides with new mechanism of action (MOA) have been released since the 1980s. The increasing number of herbicide resistant weeds and the lack of discovery of new MOAs make it difficult or even impossible to use existing herbicides effectively. Some studies predict that these problems will endanger the sustainability of weed control [16,71]. Therefore, it is essential to inform farmers about integrated and resistant weed management strategies that are less harmful to the environment [16].
Determining TSR would largely improve the weed management strategy of arable lands. Field sampling of populations and testing them by NGS-TAS method would guide farmers to select the proper herbicides thereby decreasing the impact of agricultural chemical inputs on the populations of crop fields. Thereby, unnecessary and ineffective chemical usage could be avoided [16].
In this study, we describe an effective diagnostic process using NGS-TAS to get information about evolved TSR in A. artemisiifolia populations. In NGS-TAS, TSR is determined using SNP genotyping following genome-wide genotyping. We demonstrate that NGS-TAS analysis is a method that can monitor TSR against four different enzymes targeted by different herbicides in A. artemisiifolia simultaneously. In order to analyse SNPs as part of the resistance mechanism of A. artemisiifolia, coding sequences and complete genes of D1 protein, AHAS and EPSPS were identified. However, the sequence of PPX2 only partially was identified at the region of interest as its length and genomic data is unknown. Mutations and polymorphisms in these four proteins that had previously been reported to confer resistance were specifically investigated. The developed NGS-TAS markers can identify 16 amino acid substitutions of the investigated common ragweed genes among which 4 (psbA: Val219, Ser264, ahas: Trp 574 and ppx2:Arg98) were proven. As a result, the investigated samples characterized by imidazolinone (imazethapyr) and urea (linuron) TSR were proved to carry mutation points at Trp574Leu and Val219Ile in the ahas and psbA genes, respectively. Although no MTSR existed in the studied resistant samples (AMI, AMU), the NGS-TAS method can be used to detect multiple herbicide resistance, which needs to be confirmed by specific MTSR samples.
NGS technology can examine a large number of samples simultaneously using fragment or sample-specific indexed primers. Therefore, amplicon primers were provided that are suitable for discriminating mutation points. In this study, we demonstrated the trial of a pooled sample evolved mutations points of herbicide target genes supporting resistance of A. artemisiifolia populations and biotypes.

Conclusions
Invasive weed species appeared to be the bane of biodiversity, ecosystem services, and food security. Integrated weed management practices provide several comprehensive solutions for weed control to reduce weed coverage in agricultural fields. Unfortunately, many weed management approaches used nowadays are highly unsustainable. Moreover, many studies report on the rapid evolution of herbicide-resistant weed populations. This complex evolution of herbicide-resistant weeds makes it difficult to use sustainable herbicide technology [20]. Therefore, applying weed management tools for invasive plants are often ineffective at producing long-term benefits [72]. In order to map evolved resistance Diversity 2021, 13, 118 13 of 16 in populations, different herbicide resistance mechanisms have to be investigated, such as TSR.
To summarize, NGS-TAS method is a powerful approach that was used for the detection of amino acid polymorphism and mutations that induce TSR. Therefore, it can improve effective weed control taking into account biological integrity and sustainability in agroecosystems by selecting the optimal and effective herbicide usage in cropping systems.