Pro197Thr Substitution in Ahas Gene Causing Resistance to Pyroxsulam Herbicide in Rigid Ryegrass ( Lolium Rigidum Gaud.)

: Lolium rigidum Gaud. is a cross-pollinated species characterized by high genetic diversity and it was detected as one of the most herbicide resistance-prone weeds, globally. Acetohydroxyacid synthase (AHAS) resistant populations cause signiﬁcant problems in cereal production; therefore, monitoring the development of AHAS resistance is widely recommended. Using next-generation sequencing (NGS), a de novo transcriptome sequencing dataset was presented to identify the complete open reading frame (ORF) of AHAS enzyme in L. rigidum and design markers to amplify fragments consisting of all of the eight resistance-conferring amino acid mutation sites. Pro197Thr, Pro197Ala, Pro197Ser, Pro197Gln, and Trp574Leu amino acid substitutions have been observed in samples. Although the Pro197Thr amino acid substitution was already described in SU and IMI resistant populations, this is the ﬁrst report to reveal that the Pro197Thr in AHAS enzyme confers a high level of resistance (ED 50 3.569) to pyroxsulam herbicide (Triazolopyrimidine).


Introduction
Rigid ryegrass (Lolium rigidum Gaud.) is a diploid annual grass species distributed worldwide [1]. L. rigidum is one of the most widespread and harmful weeds in the Mediterranean area and it was introduced to North and South America, South Africa, and Australia [2]. Globally, L. rigidum occurs as weeds of cereal crops, vineyards, and orchards [3]. Though it was intentionally introduced to south-western Australia as a pasture plant species, it is considered as one of the most problematic weeds [4,5].
L. rigidum is one of the most important herbicide-resistant weeds causing economic problems worldwide [6]. Since the 1980s, the development of broad and complex herbicide resistance patterns, such as multiple and cross resistance, has been detected across L. rigidum populations [7][8][9].
However, in weed control the dominant strategy is to use different herbicides that are highly effective on weeds, within these populations, several plants were documented carrying multiple resistance to herbicides including ten different chemical classes [10,11]. Multiple resistance to PSII, acetyl coenzyme A carboxylase (ACCase) and acetolactate synthase (ALS) inhibitors has been demonstrated in Australian and Spanish populations [5,12,13]. This observation has significant consequences on the treatment management for arable lands, as it indicates that the selection with a single herbicide can cause a complex and highly unpredictable pattern of cross-resistance [14].
An effective herbicide is able to penetrate the plants to their site of action at a lethal dose. Herbicides can inhibit plant specific enzymes on a specific target site, thereby altering plant metabolic mechanisms. If the structure of the target site is changing, then the effect of the herbicide is limited, therefore, plants can develop target-site resistance (TSR) against several herbicides. The mechanism of TSR includes point mutations in a target gene causing amino acid changes and lower herbicide binding capacity to the target enzyme. TSR often involves mutations in genes encoding protein targets of herbicides that affect herbicide binding either at or near the catalytic domains or in regions that affect access to them [15]. The mutant target protein shows increased expression in response to herbicide exposure [16]. Because of the rapid evolution of TSR mutations triggered by genetic and other factors, an increased number of herbicide resistant weed populations can be observed on arable fields [17]. This problem will endanger the sustainability of weed control [18], therefore, it is necessary to detect herbicide resistance in agricultural fields. Determining herbicide resistance in surviving weeds after chemical treatment facilitates the planning and effectiveness of subsequent herbicide treatment. In the next-generation sequencingtargeted amplicon sequencing (NGS-TAS) method, TSR is determined by single nucleotide polymorphism (SNP) genotyping and allows identifying the formed multiple TSR (MTSR) in weed population [19].
A remarkable phenomenon of herbicide resistance in L. rigidum cross-pollinated weed species, is the high rate of occurrence of cross-resistance that has been confirmed globally. For instance, resistance to AHAS (acetohydroxyacid synthase, EC 4.1.3.18, also known as ALS) inhibitors has been found in Australia, Chile, France, Greece, Israel, and South Africa [20]. Moreover, triasulfuron-resistant populations were found to be resistant also to imidazolinone herbicides [21]. As a consequence of amino acid substitutions, resistant plants were detected in the genetically diverse and resistance-prone L. rigidum genus. Examining a single population, 1-6 different mutations were found in the ahas gene sequence causing herbicide resistance [22]. So far, 13 Sites of Action were detected in resistant populations in the case of 35 herbicides, therefore, L. rigidum was ranked as one of the most herbicide resistance-prone weeds [20].
Resistance to AHAS targeting herbicides is the most common problem among weed species. Most weed species are more resistant to AHAS-inhibiting herbicides than to any other herbicides [8,23]. AHAS catalyzes the first common step in the pathway for synthesis of the branched-chain amino acids such as leucine, isoleucine, and valine in plants. This herbicide group consists of five different chemical families: sulfonylureas (SU), imidazolinones (IMI), sulfonylaminocarbonyl-triazolinones (SCT), pyrimidinylthiobenzoates (PTB), and triazolopyrimidines (TP). AHAS inhibitor herbicides have been widely used as a result of their effectiveness against several weeds at low doses with low mammalian toxicity and their selectivity for major crops [24,25]. Pyroxsulam (molecular formula: C14H13F3N6O5S) (IUPAC name: N-(5,7-dimethoxy [1,2,4] triazolo[1,5-a]pyrimidin-2-yl)-2-methoxy-4-(trifluoromethyl) pyridine-3-sulfonamide) is a member of the TP chemical family. Pyroxsulam herbicide inhibits the production of the AHAS enzyme in plants. Therefore, it can control annual grasses and broadleaf weeds in cereal crops, corn, soybean, sunflower, cotton, tree, and vine crops.
A few point mutations may occur in the sequence of a target enzyme gene, evolving resistance to herbicides. In plants, point mutations causing amino acid substitutions at eight different codons of the ahas gene, have been confirmed to cause resistance in fieldevolved resistant weed species [17,26]. So far in AHAS enzyme, two amino acid substitutions at the position of Pro197 and Trp574 (according to Arabidopsis thaliana numbering) have been reported in L. rigidum samples as a consequence of SU and IMI. Considering AHAS enzyme, six resistance-endowing amino acid substitutions were identified in one L. rigidum population: Pro197Ala, Pro197Arg, Pro197Gln, Pro197Leu, Pro197Ser, and Trp574Leu [20,22,27]. Furthermore, Ala122Gly, Pro197Thr, Ala205, Asp376Asn, Asp376Glu amino acid changes were observed in Greek and Italian populations after SU and IMI treatment [28,29]. Although AHAS has a crucial role in the development of herbicide resistance, interestingly, its whole gene sequence is still unknown. It would be essential to identify it to be able to discover new amino acid substitutions in order to fully understand its role in herbicide resistance.
L. rigidum, as a highly susceptible weed, poses a serious economic threat to intensive cropping systems. The use of herbicides is a dominant part of integrated weed management practices, so the knowledge of developed herbicide resistance is important (Beckie 2006). Our aim was to determine the degree of AHAS resistance and the type of resistance (TSR or NTSR) in the suspected Tunisian population of L. rigidum. The site resistance can be confirmed by examining mutations that cause known resistance.
In this study, the complete AHAS enzyme mRNA sequence in sensitive L. rigidum was identified using Illumina de novo sequencing technology, which sequence information was gap-filling information in this area. After pyroxsulam treatment resistant populations were produced in order to analyze the mutations in the ahas gene coding sequences. Based on Illumina data, point mutations were identified using Sanger sequencing ( Figure 1). After RNA extraction, library preparation was prepared for Illumina NextSeq500 paired-end sequencing. The quality of the raw reads was checked, and de novo assembly of transcripts was performed using Trinity software and the ORF of the ahas gene was identified. (c) DNA analysis. After DNA extraction, gene fragments of the ahas gene were amplified by PCR and their nucleotide sequence was identified by Sanger sequencing. As a result, amino acid substitutions were identified.

Plant Materials and Growth Conditions
Seeds of 50 putative pyroxsulam-resistant L. rigidum mature plants were collected from a wheat field beside Ben Arous city, in north-eastern Tunisia where pyroxsulam (Pallas TM 45 OD, Dow AgroSciences) was used at a normal field dose of 18.4 g ai/ha. Immediately, after collection of all seeds, they were air-dried and stored at 4 • C until used in consecutive experiments. Sensitive seeds of L. rigidum that have never been treated with herbicides were obtained from Seed Bank of Hungarian University of Agriculture and Life Sciences Georgikon Campus, Institute of Plant Protection. Seeds were germinated in Petri dishes at 30 • C/20 • C 16 h photoperiod and were planted in pots and were stored in a greenhouse.

De Novo Transcriptome Sequencing and Assembly
Leaves of five herbicide-susceptible samples were collected and frozen in liquid nitrogen immediately after collection and were stored at −80 • C. The total RNA from plant tissues were isolated using TaKaRa Plant RNA Extraction Kit following the manufacturer's instruction (Takara Bio Inc.; Shiga, Japan). The purity and concentration of the RNA sample were assessed on the Agilent TapeStation system (Agilent Technologies Inc.; Santa Clara, CA, USA) and Qubit 4 Fluorometer (Thermo Fisher Scientific; Waltham, MA, USA) (Table S1, Figure S1). As starting material, 2.0 µg of total RNA was used from which cDNA synthesis and library construction were performed. TruSeq™ RNA sample preparation kit (Low-Throughput protocol) was used to enrich the mRNA, synthesize cDNA, and prepare the library for Illumina NextSeq paired-end sequencing ( Figure S2). The quality of the raw (2 × 80 base long) reads was checked by using FastQC quality control software [30]. Quality trimming at a Phred score ≥30 was performed by using self-developed software [31]. De novo assembly of transcripts was performed using Trinity [32] with 25 kmer size. Assembled contigs were used as L. rigidum reference transcriptome. The transcript abundances were estimated using the RSEMprogram (http://deweylab.github.io/RSEM/ accessed on 10 June 2021) with the scripts provided in the Trinity package. Raw reads and assembled transcript dataset were deposited in the NCBI Sequence Read Archive (SRA: SRR8469510) and Transcriptome Shotgun Assembly (TSA: GHHB00000000.1; BioProject PRJNA512627).

Identification of Ahas mRNA in L. Rigidum
Ahas mRNA of Lolium multiflorum (GenBank: AF310684.2) was used as a reference. To dynamically translate all reading frames, Blastn algorithm was used [33]. The mRNA sequences were identified from BLAST and aligned into a single gene by using ClustalW [34]. Open reading frames (ORF) were identified using NCBI ORF finder (https://www.ncbi.nlm.nih.gov/orffinder/ accessed on 10 June 2021). The investigated sequence of L. rigidum was deposited in the NCBI GenBank MK492446.

Dose-Response Analysis
Twenty putative pyroxsulam-resistant plants at the three-leaf stage in each pot were treated with pyroxsulam herbicide as commercial herbicide formulation PallasTM 45 OD, using a plot sprayer delivering 250 L ha −1 , at a pressure of 300 kPa and a speed of 1.2 ms −1 , with a boom equipped multirange flat fan hydraulic spray nozzle (TeeJetR© XR11002). Relative biomass reduction bioassay was performed using six application rates: 0, 4.6, 9.2, 18.4, 36.8, 73.6, and 147.2 (18.4 g ai/ha as the recommended field dose). The treatment was applied on sensitive samples, as control and on resistant samples. Leaves of sensitive samples were collected at 0 day after treatment (DAT) for DNA extraction. The pots were placed in a greenhouse and survivors were assessed at 21 DAT. The herbicide treatment was applied in three technical replicates with a complete randomized block. The dose-response experiment was repeated once more. The nonlinear regression analysis was done using the "drc" package in R v4.0.4 (https://cran.r-project.org/web/packages/drc/drc.pdf accessed on 10 June 2021) ( Figure S3). According to the biomass weight of the resistant population, effective dose (ED 50 ) was estimated using the four-parameter logistic Equation (1) [35,36]: ED 50 , the effective dose causing 50% reduction in biomass, indicates the level of resistance (R/S ratios).

Ahas Gene of L. Rigidum Amplification and Sequencing
Total DNA was extracted from leaf tissues of 50-50 putative pyroxsulam-resistant and sensitive samples [37]. Three primer pairs were designed using the software Primer3Plus to amplify the earlier described resistance-conferring mutation sites based on the ahas mRNA in L. rigidum (Table S1). The PCR was performed in a final volume of 15 µL (Thermo Fisher Scientific; Waltham, MA, USA). Amplification was carried out in an Eppendorf Mastercycler ep384 (Eppendorf AG; Hamburg, Germany). PCR products were controlled with Agilent 2100 Bioanalyzer (Agilent Technologies Inc.; Santa Clara, CA, USA) using Agilent DNA 1000 Kit and were purified using NucleoSpin Gel and PCR Clean-up System (Macharey-Nagel GmbH &Co; Düren, Germany). The pGEM-T Vector System II (Promega; Madison, WI, USA) was used for cloning according to the manufacturer's instructions and fragments were sequenced ABI 3130x1 Genetic Analyzer (Applied Biosystems; Carlsbad, CA, USA). This 16-capillary array genetic analyzer was used with 50 cm capillary and 3130 POP-7 separation polymers.

Results
In this study, an RNA-seq approach is presented to identify complete coding sequences (cds) of the ahas gene in L. rigidum and mutation points involved resistance to TP herbicide, pyroxsulam. A novel amino acid substitution was identified among the earlier described resistance-conferring mutation sites using Sanger sequencing.

Transcriptome Reconstruction and Identification of Ahas Cds
The L. rigidum mRNA library was sequenced using NextSeq500 platform. After quality filtering, 18,636,868 cleaned reads were obtained from the total raw reads of 18,642,689 (2 × 80 bp paired-end) showing Phred score 34 and GC% 55 on average. Based on FastQC report, sequences represented more than 0.1% of the total and low-quality bases (corresponding to a 0.1% sequencing error rate) were removed. After transcriptome assembly by Trinity, de novo transcript dataset resulted in 74,279 contigs with lengths from 201 to 15,531bp (GC% 52.5). This transcript dataset was used to identify mRNA of ahas gene in L. rigidum. The DN19950_c0_g2_i3 contig was used to determine the ORF of ahas gene. The similarity between L. multiflorum as the reference and the identified cds in L. rigidum was 98%. The determined ahas mRNA sequence was 1929bp long.

Dose-Response Experiment
The population of L. rigidum contains different amino acid substitutions. Each sample contains one of the amino acid substitutions of Pro197Thr, Pro197Ala, Pro197Ser, Pro197Gln, or Trp574Leu. The resistance patterns of the combined substitution population were characterized. The estimated ED 50 is 3.569 that indicates the 50% survival of the resistance population to pyroxsulam that showed Pro197 or Trp574 substitution (Figure 2). Following a pyroxsulam field application rate of 18.4 g ha −1 , L. rigidum putative pyroxsulam-resistant population survived with 98.9% growth reduction.

Determination of Mutation Points of Ahas Gene in L. Rigidum
Three parts of the ahas gene were amplified and sequenced by Sanger sequencing. The primer pairs covered all of eight resistance-conferring amino acid mutation sites at three different regions: 418, 259, and 399 bp ( Figure S4, Table S2). The Ahas gene fragments were amplified by PCR from 50-50 resistant and sensitive samples and sequenced by the Sanger method ( Figure S5). Results of sequence alignment analysis showed single nucleotide substitutions in Pro197 and Trp574 positions in resistant samples (Figure 3). Wild type of L. rigidum samples showed no mutations at the eight known nucleotide sequences of the ahas gene that correspond to the resistance.

Determination of Mutation Points of Ahas Gene in L. Rigidum
Three parts of the ahas gene were amplified and sequenced by Sanger sequencing. The primer pairs covered all of eight resistance-conferring amino acid mutation sites at three different regions: 418, 259, and 399 bp ( Figure S4, Table S2). The Ahas gene fragments were amplified by PCR from 50-50 resistant and sensitive samples and sequenced by the Sanger method ( Figure S5). Results of sequence alignment analysis showed single nucleotide substitutions in Pro197 and Trp574 positions in resistant samples (Figure 3). Wild type of L. rigidum samples showed no mutations at the eight known nucleotide sequences of the ahas gene that correspond to the resistance.  Sequence alignment analysis showed a single nucleotide change of CCG to ACG at the 197 positions that resulted in a Pro/Thr substitution in the resistant samples ( Figures S6 and S7). This is the first Pro197Thr substitution in an ahas resistant L. rigidum population that was reported after triazolopyrimidines herbicide treatment. Moreover, amino acid mutations were detected at Pro197Ala, Pro197Ser, Pro197Gln, and Trp574Leu ( Figure S8) in resistant samples as previously reported [22]. A nucleotide substitution (GAT/GAC) was observed at the Asp376 amino acid position that did not evolve amino acid substitution ( Figure S9).

Discussion
In arable lands, chemical weed control is one of the most common and unavoidable methods of controlling weeds, which means high selection pressure for resistant weeds. The presence of high levels of herbicide resistance risks the sustainability of current weed control strategies, the integrated weed control [38,39]. In this study, Illumina RNA-sequencing and the "gold standard" Sanger sequencing were used to investigate the sequence of the ORF ahas gene in L. rigidum. Rigid ryegrass was noticed as a highly resistance-prone weed species and numerous populations were described as resistant to AHAS inhibitors, including SU and IMI. Populations containing Pro197 Ala122, Ala205, Asp376, and Leu 574 amino acid mutations showed high resistance (RI from 8 to 70) after SU treatment. Types of mutant ALS alleles, such as Pro197Ala, Pro197Gln, Pro197Leu, Pro197Ser, and Pro197Thr were observed in these samples [28]. The same biotypes were detected in our samples after pyroxsulam treatment. This is the first report that reveals that the Pro197Thr amino acid change in the AHAS enzyme can confer a high level of resistance (ED 50 is 3.569) to pyroxsulam (TP) herbicide in L. rigidum. The different AHAS-mutations confer different resistance against the AHAS inhibitor herbicide chemical families [20]. It has been already shown that several weed species carrying the Pro197 and Thr574 amino acid substitutions developed high levels of resistance after TP herbicide treatment. In this study, the investigated L. rigidum population that contained amino acid substitutions at the 197 and 574 positions, also showed high resistance against the pyroxsulam herbicide. So far, the Pro197Thr resistance-endowing mutation position has been reported in four weed species after TP treatment [20]. Substitutions at Pro197Thr are known to cause a high level of resistance to TP in Raphanus raphanistrum [40], in Kochia scoparia [41] and moderated resistance to TP in Apera spica-venti [26] and Alopecurus aequalis [42]. In this study, fifty L. rigidum resistant plants were examined and every individual carried only one mutation point out of the eight in the ahas gene at the known positions of resistance-endowing amino acid substitutions. Single Pro197Thr substitution was observed in 16 samples (Table S3). As the pyroxsulam resistant biotypes contained different single amino acid substitution in the ahas gene, the statistical analysis describes the mixed population. To obtain more information about the Pro197Thr amino acid substitution in the ahas gene, larger populations need to be examined. Moreover, the Thr197Pro mutation was described, after SU and IMI treatment, as the same mutation that occurs after pyroxsulam treatment; therefore, presence of cross-resistance is presumed. The existence of cross-resistance in this allelic variant should be demonstrated by treatments with other herbicides belonging to other AHAS inhibitor families (e.g., SU, IMI). Additional herbicides that demonstrate the presence or absence of cross-resistance will be included in our further experiments. Indeed, the detected mutation would provide important information about possible cross-resistance and can help to plan treatment with AHAS inhibitors.

Conclusions
Due to the rapid evolution of herbicide resistance, it is necessary to constantly monitor resistant populations on the arable fields in order to maintain weed control considering environmental protection and sustainable agriculture. The pyroxsulam-resistant L. rigidum population showed a TSR mutation detected at positions Pro197 and Leu574. The knowledge of resistance-inducing effects of compounds containing AHAS, including pyroxsulam-triggered TSR, would help to develop diagnostic methods, such as NGS-TAS. By diagnosing the presence of TSR and MTSR, chemical weed control would become more predictable, and it will guide farmers on how to plan integrated and resistant weed management strategies that are less harmful to the environment.