Yue Guo, Changle Liu, Weihua Long, Jianqin Gao, Jiefu Zhang, Song Chen, Huiming Pu,*,Maolong Hu,d,*

a Institute of Industrial Crops, Jiangsu Academy of Agricultural Sciences Key Laboratory of Cotton and Rapeseed (Nanjing), Ministry of Agriculture, Nanjing 210014, Jiangsu, China

b Institute of Life Sciences, Jiangsu University, Zhenjiang 212013, Jiangsu, China

c School of Food and Biological Engineering, Jiangsu University, Zhenjiang 212013, Jiangsu, China

d Provincial Key Laboratory of Agrobiology, Jiangsu Academy of Agricultural Sciences, Nanjing 210014, Jiangsu, China

Keywords:Acetohydroxyacid synthase High-resistance rapeseed Additional round of mutagenesis Sulfonylurea herbicide Synergistic mutations

ABSTRACT With the increasing promotion of simplified rapeseed cultivation in recent years,the development of cultivars with high resistance to herbicides is urgently needed.We previously developed M342, which shows sulfonylurea herbicide resistance, by targeting acetohydroxyacid synthase (AHAS), a key enzyme in branched-chain amino acid synthesis.In the present study,we used a progeny line derived from M342 for an additional round of ethyl methane sulfonate mutagenesis, yielding the novel mutant DS3, which harbored two mutations in AHAS genes and showed high sulfonylurea resistance.One mutation was the substitution Trp574Leu, as in M342, according to Arabidopsis protein sequencing.The other site was a newly recognized substitution, Pro197Leu.A KASP marker targeting Pro197Leu was developed and reliably predicted the response to sulfonylurea herbicides in the F2 population.The combination of Trp574Leu and Pro197Leu in DS3 produced a synergistic effect that greatly increased herbicide resistance.Analysis of the protein structures of AHAS1 and AHAS3 in wild-type and single-gene mutant plants revealed three-dimensional protein conformational changes that could account for differences in herbicide resistance characteristics including toxicity tolerance, AHAS enzyme activity, and AHAS gene expression.

1.Introduction

Acetohydroxyacid synthase (AHAS, EC 2.2.1.6; also known as acetolactate synthase, ALS) is the first enzyme in the biosynthesis pathway of branched-chain amino acids [1,2].AHAS inhibition starves plants of valine,leucine,and isoleucine,eventually leading to death.AHAS is the common target of five herbicide groups:sulfonylurea (SU), imidazolinone (IMI), triazolopyrimidine,pyrimidinyl-thio(oxy)benzoate, and sulfonyl-aminocarbonyl-tria zolinone [3].Among these herbicides, SUs have been widely used due to their broad-spectrum weed control,very low residual rates,low mammalian toxicity, and wide crop selectivity [3].The availability of herbicide-resistant hybrids would allow the application of SU herbicides in crop production systems.The development and commercialization of several herbicide-resistant traits is achievable, given that SU-resistant plants have been discovered in many crops.

Rapeseed(Brassica napusL.)is an oilseed crop worldwide.There is a need for commercial herbicide-resistant rapeseed cultivars,both open-pollinated and F1hybrid.Several SU-resistant rapeseed lines with known mutation sites have been created by mutagenesis in recent years [4-8].TheB.napusgenome contains three functionalAHASgenes (AHAS1-3) [9].Line M45, with a Pro-197-Leu substitution inAHAS3according to the protein sequence ofArabidopsis thaliana(APSAT), confers a low level (0.6 g ai ha-1) of SU tolerance and was developed by ethyl methane sulfonate(EMS) mutagenesis [5].We previously [4] developed from an EMS-treated M2population the herbicide-resistant line M342 by Trp-574-Leu(APSAT)substitution inAHAS3,which conferred resistance to approximately 90 g ai ha-1(fourfold the recommended production concentration;4×RC)to SUs.Another rapeseed mutant developed by EMS mutagenesis,K5,carries a Pro-197-Ser mutation APSAT inAHAS1.It exhibited complete male sterility induced at a low concentration of 20 ppm tribenuron-methyl (TBM) [6] and resistance to TBM at a concentration of 15 g ai ha-1[8].The Pro-197-His mutation, which confers resistance to SU, was created in rapeseed by transforming thecsr-1gene fromArabidopsis[10].The amino acid substitution P197F (APSAT) inAHAS1,AHAS3, and bothAHAS1andAHAS3, which produces an SU-resistant phenotype,was generated in rapeseed using the CRISPR/Cas-base editing system [11].A P197S substitution (APSAT) generated by a CRISPR/Cas9-mediated cytosine base-editing system was also reported[12] to improve herbicide resistance in rapeseed.In other crops,Pro-197-His mutation resistance to SU was generated in flax by transgenic technology [13], and the amino acid substitution P197F (APSAT), which produces an SU-resistant phenotype, has been generated in watermelon [14] and wheat [15] by gene editing.

Generally,SU-resistant germplasm is insufficiently abundant to meet the requirements of commercial cultivar breeding worldwide,given that many countries do not currently allow genetically modified or gene-edited crops to enter the market.Resistance in rapeseed germplasm produced by EMS mutagenesis by singlepoint mutation ofAHASgenes is too low to be suitable for practical production.These herbicide-resistant resources have limited uses,including only a few conventional rapeseed breeding procedures.Thus far,line M342 is the Chinese rapeseed material with the highest resistance, conferred by a single-point mutation produced using EMS.When M342 is crossbred, the resistance concentration of the resulting hybrid is reduced to approximately half of the recommended herbicide concentration(1.5-2.0×RC).If this rapeseed cultivar were released into the market, farmers could easily incur herbicide damage from improper weed control practices [16].

Theoretically, some crops could achieve higher resistance with two or more simultaneousAHASgene mutations.Two mutant gene combinations have been reported to confer herbicide resistance in soybean [17] and sunflower [18,19].In soybean, genesAls1andAls2are unlinked and reside on different chromosomes:Als1(P178S indicates P197S APSAT) on chromosome 4 andAls2(W560L indicates W574L APSAT)on chromosome 6.The combination of these two mutations synergistically increases tolerance to ALS-inhibiting herbicides[17].TheAhasl1-3allele,which has been shown to confer high levels of IMI resistance, showed dominance overAhasl1-1in sunflower according to resistance identification and an AHAS enzyme activity assay[19],suggesting the possibility of producing commercial sunflower hybrids by stacking the two resistance alleles.TheAhasl1-1(Ala205Val APSAT) andAhasl1-4(Trp574Leu APSAT) alleles harbor distinct point mutations that lead to different amino acid substitutions.AlleleAhasl1-4conferred higher levels of herbicide resistance thanAhasl1-1when tested against imazapyr and metsulfuron-methyl[18].Resistance showed semi-dominant behavior of theAhasl1-1andAhasl1-4alleles for both herbicides, at either whole-plant or AHAS activity levels.These discoveries allowed us to develop the first double-mutant rapeseed (5N) with high SU resistance using gene pyramiding[16].To generate abundant high-resistance rapeseed materials,we performed additional rounds of seed mutagenesis on a progeny line of M342, selecting a novel high-resistance mutant (DS3) that conferred much higher SU resistance than M342.

In this study, we performed molecular analysis to identify the roles of mutations in DS3 and developed a molecular marker to assist in the isolation of SU-resistant rapeseed lines.Resistance identification, enzyme activity analysis,and expression level analysis ofAHASgenes were conducted among several resistant lines,and the resistance mechanism was investigated using protein structure analysis.The findings of this study will enrich the germplasm available for commercial herbicide-resistant hybrid breeding of rapeseed.

2.Materials and methods

2.1.Seed mutagenesis and selection of resistant rapeseed

The medium-resistance rapeseed line M342 was derived from line N131[4].To develop higher levels of SU resistance in rapeseed,we treated approximately 4 kg of dry seeds of EM28,a progeny line derived from M342, with EMS for second-round mutagenesis, followed by high-concentration (8× RC) TBM screening, as described in our previous report[4].We obtained line DS3, which presented much higher resistance than M342.By gene cloning,we found two mutations in DS3 that contained a new substitution of P182L(P197L APSAT) inAHAS1.We then crossed DS3 and N131 to identify the new site and isolated from the F2segregating population a resistant line designated M196, with a P197L mutation site distinct from the mutation of M342.

Crosses between M196 and susceptible rapeseed lines N340 and N341 were made to develop two F2populations(N340 × M196 and N341 × M196) for further genetic analysis.The F2progeny were grown in the field and sprayed with TBM at 50 mg L-1(1× RC).Plant resistance was tested at 3 weeks after spraying and divided into three phenotype categories:resistant(R, no symptoms), moderate resistance (M, chlorosis or necrosis on some leaves, but no death), and susceptible (S, dead).

2.2.Evaluation of herbicide tolerance

2.2.1.Toxicology experiments

Lines N131,M342,M196,and DS3 were sown and grown under natural light conditions in plastic pots (30 cm diameter × 28 cm height; six plants per pot) containing a 1:1:1 mixture of 3.275 kg peat moss/perlite/vermiculite.At the 3-5 leaf stage, plants from each line were sprayed with one of nine different doses of TBM and mesosulfuron-methyl (MES):0, 0.25, 0.5, 1, 2, 4, 16, 32, or 64× RC (Fig.1).Each treatment had three biological replicates.Toxicological effects in response to TBM and MES were evaluated and recorded in terms of growth rate inhibition (%) and the concentration corresponding to 50% inhibition of growth rate(EC50, mg L-1) of each rapeseed plant at 3 weeks after herbicide application.The relationship between growth rate and plant fresh weight was plotted and used to determine biological endpoints to estimate the effect of SU toxicity on rapeseed plants.The GraphPad Prism 5 software was used to fit the results to a sigmoidal dose-response curve as follows [20]:

Fig 1.Toxicological responses of resistant rapeseed mutants (M196, M342, and DS3) and the susceptible WT (N131) following the application of SU herbicides.(A)Toxicological responses of rapeseed to tribenuron methyl(TBM).(B)Toxicological responses of rapeseed to mesosulfuron-methyl(MES)applied at 0.25,0.5,1,2,4,16,32,and 64× RC (recommended concentration).Values are means±standard errors (SEs) of at least three independent measurements from different experiments (n = 18), with at least three replicated measurements.Bars with asterisks indicate significant differences compared with corresponding control samples.*, P < 0.05; **, P <0.01 (Student’s ttest).

where BR is the biological response and BRminand BRmaxare the minimal and maximal biological responses, respectively.Hillslope describes the steepness of the curve, and all fits had a hillslope value of 1.0.

2.2.2.Isobole analysis method

Possible interactions betweenBnAHAS1-3RandBnAHAS3Rin DS3 in the phenotypic response to SU herbicide were investigated by plotting an isobole,or line of equal effect (i.e., zero interaction)[21].A combination effect was predicted based on the concept of dose equivalence, which assumes that an equally effective dose of treatment A will add to the dose of treatment B if the two are combined [22].The plant response of the three-dimensional (3D)phenotype was then reduced to a 2D graph by isobole analysis,selecting a quantitatively defined effect as the criterion for the graphical representation of joint action of both components.In this study,theX-,Y-,andZ-axes in 3D space represented the contributions ofBnAHAS1-3R,BnAHAS3R,and the plant response to SU herbicides, respectively.TheX- andY-axes in 2D space represented respectively the herbicide dose estimated to produce 50% phytotoxicity whenBnAHAS3Rwas present and that whenBnAHAS1-3Rwas present.The isobole method is always effective, regardless of the mechanism of action.

2.3.Polymerase chain reaction (PCR) amplification and sequence comparison

At three weeks after herbicide spraying treatments, genomic DNA was extracted from young rapeseed leaves using cetyltrimethylammonium bromide buffer, as previously described[23].TheBnAHAS1-BnAHAS3primer pairs(Table S2)were designed by Hu et al.[24] and synthesized by Genscript Biotech (Nanjing,Jiangsu, China).Open reading frames(ORFs) ofBnAHAS1-BnAHAS3from DS3, M196, M342, and wild-type (WT) N131 were amplified and the resulting nucleotide sequences analysis were performed using DNAStar 7.0, Sequencher 4.8, and DNAMAN 9.0.Nucleotide and amino acid multiple-sequence alignment was performed with MUSCLE (https://www.ebi.ac.uk/Tools/msa/muscle/).BioEdit(https://www.mbio.ncsu.edu/bioedit/bioedit.html) was used for alignment display.The sequence of theAHASgene inA.thalianawas used as a reference (GenBank accession NM_114714).

2.4.Development of a marker for BnAHAS1-3R

Kompetitive allele specific PCR (KASP) is a technology developed by KBioscience based on homogeneous, fluorescence-based genotyping, and as evolved to be a global benchmark technology[25].In a previous study, we developed an allele-specific cleavage amplified polymorphic sequence (AS-CAPS) marker targeting the mutation site of Trp556Leu in M342 (the same mutation as Trp574Leu (APSAT) in DS3) for genotype detection [4].Thus, we needed to develop a molecular marker to detect the other mutation site,Pro182Leu,of theAHAS1gene in DS3.Based on the characteristics of the resistance mutation site inBnAHAS1-3R,we obtained a KASP marker named KBnALS1_1968545A (Table S2), which was then synthesized by the Laboratory of the Government Chemist(LGC, London, UK).This marker consisted of two singlenucleotide polymorphism(SNP)-specific primers and one universal primer.Among these three primers, two C/T alleles at the 3′end were linked to the LMC and HEX fluorescent linker-specific sequences of the LGC KASP reagents at the 5′end.This KASP marker was then evaluated in F2populations.

The PCR reaction mixture (5 μL) contained 10 ng of total genomic DNA, a KASP reaction mixture consisting of 2.5 μL of 2× KASP Master mix, and 0.07 μL of a KASP assay mix in a microwell reaction plate.PCR amplification was performed as follows:pre-denaturation at 94 °C for 15 min; 10 cycles at 94 °C for 20 s and 61-55 °C for 1 min, with the annealing temperature reduced by 0.6 °C per cycle; and 26 cycles at 94 °C for 20 s and 57 °C for 1 min.After the reaction was complete, a Pherastar scanner (BMG Labtech Inc., Ortenberg, Germany) was used to read the fluorescence data of the KASP reaction product,and the results were automatically converted into graphical format.

2.5.Transformation with BnAHAS1-3R

ApBnAHAS1-3Rvector withBnAHAS1-3Rwas constructed to validate the function of the candidate geneBnAHAS1-3Rof line M196 in herbicide resistance, as described in our previous study[24].Genomic fragments encodingBnAHAS1-3Rwere amplified using primers AHAS1-F2/R2 and cloned into thepEASY-T1vector(Table S2),such thatBnAHAS1-3Rwas under the control of the constitutive maize ubiquitin promoter.pEASYT1-BnAHAS1-3Rwas double-digested and ligated into the binary vector pCUbi1390.Finally, the recombinant construct was introduced into EHA105 of anAgrobacterium tumefaciensstrain and transformed intoArabidopsisusing floral dip transformation, as previously described[26].TBM solution was then applied toArabidopsisat 0× and 2×RC during a 3-week period,with the aim of assessing herbicide resistance in the transformed plants.

2.6.AHAS gene expression and AHAS enzyme activity assay

2.6.1.AHAS enzyme activity assay

At 3 weeks after germination, herbicides were applied to each line(DS3,M196,M342,and N131)by leaf spraying at 4×RC.Then,leaves were harvested and frozen in liquid nitrogen after periods of 0,12,24,48,and 72 h.AHAS enzyme extraction and detection were performed as previously described[27].Briefly,AHAS activity was calculated colorimetrically(530 nm)by measuring acetoin production.Enzyme activity was defined as acetoin content per unit mass of protein (mg) production product per hour (μmol mg-1h-1).

2.6.2.Quantitative real-time PCR (qPCR)

The harvest time and harvest method of the sample were as described in section 2.6.1.Total RNA was isolated using the RNAprep pure Plant Kit (Tiangen Biotech Co., Ltd., Beijing, China), in accordance with the manufacturer’s instructions.The qPCR reactions were performed using an ABI 7500 real-time PCR system(Applied Biosystems, Foster city, USA) with the ChamQ Geno-SNP Probe Master Mix (Vazyme Biotech Co., Ltd, Nanjing, China), in accordance with the manufacturer’s instructions.Theeukaryotic translation initiation factor 4B(EIF4B) gene was used as an internal control.Expression levels were normalized to the levels ofEIF4Bunder identical conditions.Three pairs of primers and their corresponding TaqMan probes were designed for qPCR assays and synthesized by Genscript Biotech(Nanjing,Jiangsu,China);all primers and probes are listed in Table S2.Three biological replicates and three technical replicates of the RNA samples were used for qPCR analyses.

2.7.Protein structure prediction

Hydrophobic cluster analysis of rapeseed AHAS proteins was performed using HCA 1.0.2 software,available at the Mobyle Portal(https://mobyle.rpbs.univ-paris-diderot.fr/cgi-bin/portal.py?form=HCA%23forms::HCA#forms::HCA).For secondary structure prediction, one-to-one threading was performed using the PHYRE2 protein modeling server (http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index_advanced).The SWISS-MODEL web server(https://swissmodel.expasy.org/interactive/) was used for protein homology modeling after submission of the amino acid sequences of BnAHAS1 and BnAHAS3 [28,29].SwissDock (http://www.swissdock.ch/) was used for protein-ligand docking analysis, and the PyMOL molecular graphics system was programmed for protein structural visualization and superposition.

3.Results

3.1.Identification of the high herbicide-resistance mutant DS3 and molecular cloning of its mutations

DS3 seedlings (Fig.S1A) in the M2generation (Fig.S1B) grew well in the field after 3 weeks of TBM spraying.The DS3 rapeseed phenotype was also obtained after 3 weeks of TBM spraying with 16× RC TBM (Fig.S1C).DNA sequence alignment showed a single C-T transition at approximately nucleotide +545 inAHAS1of DS3,based on the previously published M342 (Fig.S2A) [4].The mutation resulted in a new amino acid substitution from proline (CCT)to leucine (CTT) at amino acid position 182 (P197L APSAT)(Fig.2A).The AHAS3 protein showed an amino acid substitution from tryptophan (TGG) to leucine (TTG) at amino acid position 556 (Fig.2B), which is equivalent to amino acid substitution in M342 [4,16].Sequence analyses of DS3 with our WT N131 and the WT sequences Z11524(BnAHAS1)and Z11526(BnAHAS3)indicated that DS3 was a new mutant,with two mutations(nucleotide position +545 in the coding sequence ofAHAS1and +1667 in the coding sequence ofAHAS3) that conferred high SU resistance(Fig.S2).

3.2.Development and characterization of M196 with the new BnAHAS1 mutation site

Next, we investigated the effects and functions of the new mutation site in DS3 in herbicide resistance.We isolated plants with new mutation sites, distinct from M342 in DS3.We isolated the resistant line M196, with a mutation distinct from that of M342, from the F2segregating population.We further tested the segregation of the F2population by the χ2goodness of fit test.Then we tested the mutation site of M196 by sequencing.DNA sequence alignment revealed a single C-T transition at +545 inBnAHAS1(Fig.S2A), and we designated this new mutant gene asBnAHAS1-3R.These results confirmed that resistance in M196 was inherited as a semi-dominant trait, conferred by a single nuclear-encoded gene (Table S3).

We also constructed and introduced the plasmid pBnAHAS1-3R,which contained the entire ORF ofBnAHAS1-3R,into WTArabidopsisto confirm TBM resistance(Table S2).We selected two homozygous lines from the offspring of two independent transformants,T2A-7a and T2A-7b for subsequent evaluation of SU resistance.All transgenic plants survived without unhealthy symptoms in comparison with WT plants, which died following the application of 100 mg L-1TBM (2× RC) (Fig.2D).This 2× RC resistance level was also verified in M196 rapeseed in a subsequent toxicology experiment (Fig.1).

Given that the Pro182Leu (P197L APSAT) substitution in M196 represented a new mutation site for herbicide resistance in rapeseed, we developed a KASP marker to detect this new mutation.Based on the characteristics of the resistance mutation site inBnAHAS1, a KASP marker (KBnALS1_1968545A) was developed and found to effectively discriminate three genotypes involving theBnAHAS1gene:T/T (red), T/C (green), and C/C of the P182L resistance mutation site (Fig.2C).We obtained the KASP marker primers and developed a system for detecting the P182L resistance mutation site ofBnAHAS1, which will allow identification of the mutant in production trials and breeding of resistant cultivars.

3.3.Assessment of herbicide toxicity

Dose-response curves for SU herbicide were developed to test the herbicide toxicological response of rapeseed among the resistant mutants (M196, M342, and DS3) and WT (N131).Lines M196 (0.25-4× RC), M342 (0.25-4× RC), and DS3 (0.25-16× RC)showed significantly reduced plant phytotoxic responses, compared with N131,at 3 weeks after TBM application(Fig.1A).A similar trend was also found in plant responses to MES treatment,with reduced sensitivity in all accessions,and among mutants,including M196 (0.25-2× RC), M342 (0.25-4× RC), DS3 (0.25-16× RC), and WT (Fig.1B).All plants died when the application concentration was increased to 32-64× RC (Fig.1).

Fig 2.Partial alignment of an amino acid sequence of AHAS proteins between wild-type(WT)and mutant rapeseed,and characterization of the mutant gene BnAHAS1-3R.(A)BnAHAS1 (GenBank accession no.Z11524.1), BnAHAS1_N131, BnAHAS1_M342, BnAHAS1_M196, and BnAHAS1_DS3 (Brassica napus L.cv.N131,M342, M196, and DS3), (B)BnAHAS3(GenBank accession no.Z11526.1),BnAHAS3_N131,BnAHAS3_M342,BnAHAS3_M196,and BnAHAS3_DS3(B.napus L.cv.N131,M342,M196,and DS3).Numbers at the right indicate the position of the last amino acid in the fragment illustrated.Black stars and pink shading represent point mutations in the mutants.(C)Kompetitive Allele-Specific Polymerase chain reaction(KASP)single-nucleotide polymorphism(SNP)genotyping of the AHAS1 mutation site(C/T)in F2 populations.Sulfonylurea(SU)-resistant and SU-susceptible plants including two parents and F1 as positive controls are indicated by circles.Red, green, blue, and gray circles indicate homozygous resistant (T/T),heterozygous resistant(C/T),homozygous susceptible(C/C),and negative control plants,respectively.(D)Plant phenotype symptoms of Col and transgenic Arabidopsis plants T2A-7a and T2A-7b were treated with tribenuron methyl (TBM) for 3 weeks at the 4-6 leaf stage.0× and 2× indicate the herbicide concentration relative to the recommended concentration (RC) for weed control.

3.4.Assay of AHAS activity and gene expression after TBM treatment

The target enzyme of SU herbicides is AHAS, which undergoes inhibition of enzyme activity through biochemical binding reactions[30].Based on herbicide toxicity analysis of the plant phenotypes, a concentration of 4× RC was selected as the appropriate treatment concentration in subsequent experiments.To investigate the differential inhibition response of AHAS among four lines,we plotted a time-course curve for 72 h,following TBM treatments of 4×RC.The AHAS activities of most lines decreased greatly after TBM application at various time points throughout the time-course experiment(Fig.3A).For N131 and M196,AHAS activity decreased significantly from 12 to 72 h after TBM treatment,reaching respective minima of approximately 0.48-fold and approximately 0.58-fold compared with the control samples (Fig.3A).In M342, the decline was observed only at 72 h, reaching approximately 0.625-fold compared with the control lines (Fig.3A).However,there was no decline in AHAS activity in the DS3 mutant during the 72-h TBM treatment period (Fig.3A).These AHAS activity results indicate that the different lines showed different sensitivities to TBM treatment.

To investigate whether TBM treatment induced changes inAHASgene expression in different lines, we examined the expression patterns of two constitutiveAHASgenes,AHAS1andAHAS3.The transcript levels of mostAHAS1andAHAS3genes in the four lines were greatly increased by TBM induction at various times throughout the 72-h period.The expression levels ofAHAS1in N131,M196,M342,and DS3 peaked at 12 h after 4×RC TBM treatment (Fig.3B).Nevertheless, each line had different levels of peak response compared with their controls; the response was highest at approximately 5-fold for DS3 and lowest at 0.8-fold for N131(Fig.3B).The fold-change values of the peaks for M196 and M342 were intermediate among the four lines, with a slightly higher expression level in M342 (Fig.3B).In all four lines,AHAS3expression exhibited a similar trend toAHAS1following TBM spraying at 4× RC (Fig.3C).No changes were detected between TBM-treated and control plants at 72 h for eitherAHASgene in WT N131,whereas significant differences in transcript levels were simultaneously detected in the mutant lines (M196, M342, and DS3) (Fig.3).After TBM treatment, DS3 also showed the highest expression levels in bothAHASgenes at 72 h, compared with controls(Fig.3).These results indicate that TBM induced upregulation ofAHASgenes under our experimental conditions, with greater upregulation in mutants than in WT plants.

Fig 3.Time responses of (A) AHAS enzyme activity and relative expression levels of AHAS genes (B) AHAS1 and (C) AHAS3 in rapeseed leaves of four varieties after TBM treatment at 4× RC.Values are means±SEs of at least three independent measurements from different experiments (n = 18), with at least three replicated measurements.*and**indicate significant difference between distilled water(control) and exogenous tribenuron methyl(TBM)treatments at P <0.05 and P <0.01(Student’s t-test).

3.5.Structural analysis of AHAS proteins

Hydrophobic clustering analysis was performed for AHAS1 and AHAS3 in WT and mutant DS3 (proteins BnAHAS1-3R and BnAHAS3R).The substitution of P182L changed the hydrophobic cluster distribution of the mutant (Fig.S3A) in comparison with WT(Fig.S3B), whereas no such change was observed in W556L of DS3 (Fig.S3).Secondary structural analysis prediction by the PHYRE2 server indicated that the substitution W556L increased the proportion of total protein from 36% to 37% in α-helix in DS3,compared with WT (Fig.S4C, D).The P182L substitution also led to lower secondary structure confidence in DS3 (Fig.S3B), compared with WT(Fig.S3A).These results suggested that the differential response to TBM treatment might be caused partly by changes in hydrophobic clustering and secondary structure.

We investigated the 3D structures of the AHAS proteins to identify the mechanism of differential amino acid substitution responses to SU herbicide.Our previous study [16] demonstrated that the TBM molecule in the AHAS1 protein was located in a pocket structure near W559 that exhibited strong positive electrostatic potential properties relative to the protein surface (Fig.4A).In the present study, the substitution in AHAS1 protein was P182L in the new mutant M196, which was inconsistent with our previous finding.P182 was located distant from the AHAS1 protein active center,near W559(Fig.4B).Using artificial deletion of a single protein aggregate to prevent overlapping among TBM molecules, we found that the positions of P182 and W559 were within a tetramer structure forming by the whole AHAS1 protein(Fig.4C).In the AHAS1 mutant protein, a change to P182L would alter the structure of the entire tetramer,thereby altering the polymerization ability of protein tetramers (Fig.4C).In the WT AHAS3 protein,the TBM binding position was located in a pocket structure near W556, similar to its location in the AHAS1 protein, which showed strong positive electrostatic potential properties relative to the protein surface (Fig.4D, E).The indole heterocyclic plane belonging to W556 was parallel to the benzene ring plane of TBM, exhibiting a stable π-plane stacking effect (Fig.4F).Thus,the spatial orientation of TBM was more favorable and the affinity between AHAS3 and TBM was increased (Fig.4F).In contrast, the mutant had no aromatic groups in the leucine of W556L, which ultimately led to the destruction of the original π-plane stacking effect in the local region of the branched fat chain of the side chain(Fig.4F).

Fig 4.Three-dimensional(3D)structural analysis of AHAS proteins in rapeseed.(A)Hydrophobic potential analysis of the AHAS1 protein surface in WT N131 after tribenuron methyl (TBM) treatment.(B) AHAS1 protein exhibits binding to TBM herbicide in the form of tetramers (one protein aggregate was removed to prevent overlapping with TBM).(C) Position of P182 on the tetramer binding surface of the AHAS1 protein.Dark blue circle indicates the molecular surface of the binding site in the AHAS1 protein active center.Black circle indicates the interface of tetramer binding.Electrostatic potential analysis of the AHAS3 protein surface in(D)WT N131 and(E)WT N131 after TBM treatment.(F) Changes in molecular surface interactions after TBM treatment near W556 in the binding pocket before (WT) and after mutation (BnAHAS3R).Blue and pink ribbons indicate the WT and mutant proteins, respectively.Black circles indicate the molecular surfaces of the binding site.

3.6.Synergistic effects of mutant genes in the DS3 mutant

Isobole analysis of EC50 values and their confidence intervals was performed to determine the zero interaction, synergism, or antagonism betweenBnAHAS1-3RandBnAHAS3Rgenes after a series of TBM treatments.Based on log-logistic analysis, predicted dose-response curves were plotted for the three mutant lines(M196, M342, and DS3) to compare their responses to TBM application(Fig.5).The estimated means of 50%phytotoxicity(95%confidence intervals) for TBM doses were 159.8 mg L-1(107-299 mg L-1) for M196 withBnAHAS1-3Ralone, 210.6 mg L-1(132.2-335.6 mg L-1) for M342 withBnAHAS3Ralone, and 753.2 mg L-1(372-1206 mg L-1) for DS3 withBnAHAS1-3RandBnAHAS3R(Table 1).

The interactions betweenBnAHAS1-3RandBnAHAS3Rgenes were characterized using the isobole method and the EC50 values.Isoeffective doses are doses that can be substituted between genes in the absence of interaction.On the isobole graph, the zerointeraction line is a straight line connecting the isoeffective rates(in this study, EC50 values) of the two mutant genes,BnAHAS1-3RandBnAHAS3R(Fig.5D).An EC50 line with estimated confidence intervals of the rapeseed response for the combination ofBnAHAS1-3RandBnAHAS3Rwas plotted, as previously described [17].Then,the separate contributions of theBnAHAS1-3RandBnAHAS3Rcomponents to the EC50 value of combinedBnAHAS1-3RandBnAHAS3Rwere parsed and plotted on the isobole graph at an assumed 1:1 ratio,because the combination ofBnAHAS1-3RandBnAHAS3Rcontained two independent AHAS genes(BnAHAS1-3R,chromosome 1,C genome;BnAHAS3R,chromosome 1,A genome)that belonged to two different chromosomes located in two separate genomes[16].The zero-interaction line was below the lower confidence interval of the combination of the twoAHASgenes, showing a synergistic effect.

If the combination ofBnAHAS1-3RandBnAHAS3Rcauses a synergistic effect in increasing SU resistance, then more herbicide is required to achieve the same level of phytotoxic response in rapeseed plants that carry bothBnAHAS1-3RandBnAHAS3R, compared with those that carry only one mutant gene.For example, a larger dose of TBM(735.2 mg L-1)was needed to cause 50%phytotoxicity inBnAHAS1-3R+BnAHAS3Rplants than inBnAHAS1-3Rplants(159.8 mg L-1) orBnAHAS3Rplants (210.6 mg L-1) plants (Fig.5).The synergistic effect of the combination ofBnAHAS1-3RandBnAHAS3R, which improved SU tolerance, was also observed in rapeseed after MES treatment (Table 1).The MES dose required to cause 50% phytotoxicity in theBnAHAS1-3R+BnAHAS3Rbiotype was 669.8 mg L-1, which was larger than the MES dose in theBnAHAS1-3R(31.31 mg L-1) orBnAHAS3Rbiotype (179.1 mg L-1)(Table 1).

Table 1 Dose-response and isobole values in rapeseed plants after sulfonylurea herbicide treatments.

Fig 5.Responses of rapeseed plants after 21-day herbicide treatment.(A)M196,(B)M342,and(C)DS3.(D)Isobole plot based on EC50 for BnAHAS1-3R+BnAHAS3R,indicating synergism versus the zero-interaction line.Values are means±SEs of at least three independent measurements from different experiments (n = 18), with at least three replicated measurements.

4.Discussion

In a previous study,we obtained M342 with 4×RC resistance to SU herbicide [4].However, when the resistance trait of M342 was introduced into the cytoplamic male sterile(CMS)restorer line and then crossed with the sterile line,the resistance genotype of the F1hybrid was heterozygous, with a TBM resistance of 1.5-2× RC.Inappropriate concentrations or uneven application of herbicide by farmers in the field could easily exceed the resistance threshold of these resistance traits.Therefore,the application of M342 resistance has limitations in breeding herbicide-resistant rapeseed cultivars.To improve the resistance level of M342 to meet the requirements of practical breeding applications,we performed further seed mutagenesis using EMS and strengthened the selection pressure by increasing the herbicide application concentration.DS3, which was obtained under 8× RC treatment, exhibited high SU resistance (Fig.S1).The use of mutant rapeseed to develop a new mutant after two further rounds of mutagenesis with the imposed selection pressure is a major outcome of this study.Typically, the breeding of herbicide-resistant rapeseed using EMS mutagenesis has been performed in one round, with a selection pressure of 1-2× RC [5,6,8].The successful identification of DS3 with 16×RC herbicide resistance shows the advantage of our strategy for rapeseed breeding.We also tested several important agronomic and quality traits of this mutant including plant height,flowering time (from sowing to flowering), main inflorescence length, number of effective pods per plant, number of seeds per pod, 1000-seed weight, yield per plant, and oil quality (Table S1).The results indicated that some agronomic traits of DS3 rapeseed were poor and require further improvement.

Sequence alignment showed that DS3 harbored two mutant genes:substitutions of W556L inAHAS3and P182L inAHAS1.Because the W556L substitution is identical to the mutant gene in M342, we performed further experiments to separate the P182L substitution from the DS3 material.M196 with the P182L(P197L APSAT) substitution was selected from the F2segregating population.Point mutations ofAHASgenes can confer tolerance to several types of herbicides, with minimal damage to plant growth [11].The Pro-197-His mutation was initially isolated by measurement of seedling growth in the presence of chlorsulfuron herbicide [31].The resistance mutant genecsris widely used to transform crops through genetic engineering, such as in flax [13]and rapeseed [10].The amino acid substitution P197F in SUresistant phenotypes was created via the CRISPR/Cas-base editing system inArabidopsis[32],watermelon[14],wheat[15],and rapeseed [11].Two studies using the P197S substitution for improving herbicide resistance in rapeseed were reported following EMS mutagenesis selection [6,8] and CRISPR/Cas9-mediated cytosine base editing[12].The mutant rapeseed M45 harbored a P197L substitution that conferred a low level of SU tolerance inAHAS3[5].However, the present study is the first to identify rapeseed M196 with the Pro197Leu substitution located inAHAS1.Further functional identification was performed by transferring the mutant gene toArabidopsisand producing an SU-resistant phenotype(Fig.2D).Generally,M196 was a suitable germplasm for producing herbicide-resistant cultivars.

We also developed a KASP marker to detect the P182L substitution for marker-assisted selection breeding.Various SNP genotyping methods have been developed, including site-directed sequencing technology, allele-specific PCR (AS-PCR), CAPS,temperature-switched PCR, and gene resequencing [33-35].However, these methods have limitations including low throughput and high cost [36].KASP, which is a competitive AS-PCR technology, has been widely used for high-throughput SNP typing due to its high stability,accuracy,and low cost[25].KASP is most advantageous when there are few SNP loci [25].In this study, the KASP marker developed to detect aBnAHAS1point mutation at position 182 accurately predicted the phenotypic response to SUs in the tested F2populations (Fig.2C).A previous study developed 70 KASP markers for wheat, which were successfully identified as markers for the phenotypes of the cultivar parents and parental populations, thus revealing the potential applications of KASP molecular markers in wheat breeding[36].The International Maize and Wheat Improvement Center uses KASP, generating more than one million data points per year for crop improvement [37].Thus far, few studies have examined potential KASP markers for herbicide resistance in rapeseed, and only five AS-PCR markers have been reported to assist in selecting herbicide-resistance traits in rapeseed.Among these,three were identified in our previous studies:one indicated IMI resistance at position 653 of theBnAHAS1mutant[24],while the other two showed SU resistance at position 574 in the mutantsBnAHAS3R[4] andBnAHAS1-2R[16].The other two AS-PCR markers for SU-resistant phenotypes were developed to target point mutations of P197L inAHAS3[5] and P197S inAHAS1[12].In the present study, we applied the developed KASP marker to the P182 site (197 APSAT) in an F2segregating population of DS3, demonstrating the accuracy and applicability of this marker.Thus, DNA marker-assisted breeding technology was successfully applied to the site separation of herbicide-resistant double mutation sites in rapeseed.Compared with conventional breeding methods, this technique has the advantages of reducing errors in the selection process, preventing the loss of resistance sites,reducing breeding years and work volume,and lowering production costs.Such optimized markers will facilitate the rapid and precise incorporation of commercially useful mutations into a wide variety of elite rapeseed germplasms with wide applications in crop breeding.

The DS3 mutant with two independent mutations conferred a resistance of 16× RC to both TBM and MEC (Fig.1).Notably,mutants with the single-point mutation showed resistance levels of only 2-3× RC (P197L APSAT substitution in M196) and 4× RC(W574L APSAT substitution in M342) to TBM and 1× RC (P197L APSAT substitution in M196) and 4× RC (W574L APSAT substitution in M342)to MES.The resistance level of 16×RC was markedly higher than the sum of 2-3×RC and 4×RC for TBM application or 1× RC and 4× RC for MES application, suggesting a synergistic effect.The isobole graph,which allowed visualization of the synergistic effect of the combination ofBnAHAS1-3R+BnAHAS3R, indicated greatly improved SU tolerance in rapeseed after both TBM and MES treatments (Table 1; Fig.5).However, synergistic effects have rarely been reported among genes conferring herbicide resistance.We previously reported a synergistic effect of SU resistance in the 5N rapeseed line[16];the zero-interaction line of the 5N isobole graph appeared below the lower confidence interval of theBnAHAS1-2R+BnAHAS3Rcombination, indicating synergy [16].The double mutations of 5N reside independently on chromosome 1 of the A (BnAHAS3R, W574L APSAT) and C (BnAHAS1-2R, W574L APSAT) genomes [16].Phenotypic analyses in a previous study of soybeans [17] demonstrated thatAls1(P197S APSAT) andAls2(W574L APSAT)mutations exhibited synergistic effects of ALS herbicide tolerance; in agreement with our findings of mutant rapeseed DS3 and 5N with double mutant sites,Als1andAls2were two independent genes respectively located in two distinct soybean chromosomes(4 and 6)[17].However,the synergistic effects detected in our DS3 or 5N rapeseed appeared much stronger than the synergistic effects in soybean, such that the combination ofAHAS1andAHAS3led to improved tolerance to SU herbicides in DS3(Fig.1)and 5N[16]for 16×RC of TBM and MES,respectively.The synergistic effect of SU resistance in the soybean study achieved maximum resistance levels of 8, 4, 2, 2, 8, 8, and 8× RC for chlorimuron, nicosulfuron, rimsulfuron, sulfometuron, thifensulfuron, tribenuron, and flucarbazone, respectively [17].Other studies of synergistic effects in plants have been conducted mainly usingArabidopsis, such as the synergistic effects observed during flower development [38].However, analyses of synergistic effects on herbicide resistance in crops (soybean and rapeseed) remain far less advanced.The molecular and genetic mechanisms behind the synergistic phenomenon of DS3 require further investigation for future crop improvement.BecauseAHAS1andAHAS3were located in the A and C genomes, respectively, the interaction of chromatin of different sections between A and C genomes might also have caused their interaction following SU treatment.We may test this hypothesis in a future study using highthroughput/resolution chromosome conformation capture 3D genome sequencing,combined with further experimental verification.

In addition to synergistic effects, we observed another notable phenomenon in which the SU resistance response differed between two mutants with single-point mutations, M196 (P197L APSAT)and M342 (W574L APSAT).These differences were apparent in the herbicide toxicity response(Fig.1),inhibition of AHAS enzyme activity (Fig.3A), and expression ofAHASgenes (Fig.3B, C) after herbicide treatment.In comparison with the WT, the substitution of P182L(P197L APSAT)changed the hydrophobic cluster distribution in the mutant (Fig.S3A), whereas no such change was observed in W556L (W574L APSAT) (Fig.S3B).We also compared the secondary structures (Fig.S4) of both AHAS1 and AHAS3 proteins between the mutant and WT.Because we did not obtain optimal results in this analysis,we focused on the 3D structures of the AHAS proteins.Importantly, the AHAS1 protein sequence showed very high identity (>90%) with the protein sequence of AHAS3[16], and the differences between the two mutants with singlepoint mutation (M196 and M342) were apparent mainly in the type and position of the amino acid substitutions.In the mutant protein of M196,P182L was distant from the AHAS1 protein active center(Fig.4B)and changed the polymerization of the entire tetramer structure,ultimately destroying the molecular binding of TBM and causing the resistance response to TBM treatment(Fig.4A-C).The W556 substitution was located in a pocket structure, which constituted the TBM binding position, thus indicating strong positive electrostatic potential properties in the WT protein surface(Fig.4D, E).In contrast, the W556L in M342 had no aromatic groups in leucine,ultimately leading to the destruction of the original π-plane stacking effect in the local region of the branched fat chain of the side chain and inhibiting the effective binding of TBM in the active center of AHAS3,thereby demonstrating SU tolerance(Fig.4D-F).The structure of an enzyme determines the specificity of its function,and the mutation of a few amino acids in the active center of the enzyme leads to changes in enzyme catalytic activity and substrate selectivity[39].We presume that the W556L substitution in mutant M342 altered its SU resistance by changing the active center of the enzyme, which caused corresponding changes in the AHAS enzyme activities and expression levels ofAHASgenes.Although the amino acid residues in the catalytic pockets of enzymes play important roles in the respective enzyme activities,distal amino acid residues may also affect enzyme catalytic activity[40].For example, proline with a preferred β-turn or random coil structure functions in maintaining protein structural stability owing to the unique rigid structure of its pyrrolidine ring [41,42].The proline theory suggests that enzyme thermostability increases with increasing proline levels[43],as has been confirmed in many enzymes [43-45].In this study, Pro182, which was located in a random coil distant from the AHAS1 protein active center, was found to stabilize the tetramer structure.P182L substitution in M196 led to changes in tetramer polymerization and hydrophobic cluster distribution, which influenced protein folding and ultimately caused SU resistance by reducing AHAS enzyme inhibition andAHASgene upregulation.We speculate that differences in SU resistance between M342 and M196 were due mainly to the different mechanisms of mutant amino acids in the 3D AHAS protein structure.Amino acid mutations in the catalytic pocket of enzymes in the active center appeared to be more influential than the distal position of the enzyme in altering the SU resistance level, regardless of whether proline was in the distal position.This conclusion invites confirmation in future studies.

The development and application of excellent germplasm resources will drive the seed industry and lead to the development of new cultivars.However, chemical weed control is limited in rapeseed cultivation in China due to shortages in commercial herbicide-resistant rapeseed.To escape these bottlenecks,we have developed a new germplasm series with independent intellectual property rights.These lines are resistant to various herbicides and include lines M342,5N,and DS3,which have been distributed to 28 domestic universities and scientific research and breeding institutions to promote the selection and breeding of herbicideresistant rapeseed cultivars in China.Given that this distribution network covers all of the rapeseed growing ecological regions in China, non-transgenic herbicide-resistant rapeseed accessions are involved in breeding efforts in all ecological zones in China.Corresponding non-transgenic herbicide-resistant rapeseed cultivars will be released in various ecological regions in the near future,owing to the stability of our resistant resources,with great practical significance for rapeseed industrial development.We have also used these new herbicide-resistant germplasm lines to transfer resistance genes into rapeseed parental material.The new herbicide-resistant rapeseed lines Ning R101 (registration no.GPD Rapeseed 2018320256) and Ning R201 (GPD Rapeseed 2020320202) were bred in 2018 and 2020, respectively, and are currently applied in large-scale production,with broad application prospects.

CRediT authorship contribution statement

Yue Guo:Data curation, Formal analysis, Funding acquisition,Investigation, Methodology, Visualization, Writing - original draft,Writing - review & editing.Changle Liu:Investigation, Methodology.Weihua Long:Writing-review&editing,Funding acquisition.Jianqin Gao:Investigation, Methodology.Jiefu Zhang:Writing -review & editing, Funding acquisition.Song Chen:Investigation,Methodology, Funding acquisition.Huiming Pu:Conceptualization,Funding acquisition,Project administration,Resources,Supervision.Maolong Hu:Conceptualization, Funding acquisition,Project administration, Resources, Data curation, Supervision,Validation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the National Natural Science Foundation of China(31870519,31901503 and 31671731),the National Key Research and Development Program of China(2016YFD0101300 and 2016YFD0100202-10), the China Agriculture Research System (CARS-12), and the Natural Science Foundation of Jiangsu Province (BK20190267).We thank the Jiangsu Collaborative Innovation Centre for Modern Crop Production,China and thank the efforts of Ms.Li Cheng and Yanfang Yu for this work.We also thank Textcheck (http://www.textcheck.com/certificate/WslW6j) for editing this manuscript.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.05.006.