Amir Hmid Khn,Yizn M,Yunlong Wu,Adnn Akbr,Muhmmd Shbn,Abid Ullh,b,Jinwu Deng,Abdul Sboor Khn,Hubin Chi,Longfu Zhu,Xinlong Zhng,Ling Min,*

a National Key Laboratory of Crop Genetic Improvement,Huazhong Agricultural University,Wuhan 430070,Hubei,China

b Department of Botany,University of Malakand,Dir Lower 18800,Khyber Pakhtunkwa,Pakistan

Keywords:Cotton(Gossypium hirsutum)Jasmonic acid Allene oxide cyclase 2 ROS CRISPR/Cas9 High-temperature stress

ABSTRACT Cotton(Gossypium spp.)yield is reduced by stress.In this study,high temperature(HT)suppressed the expression of the jasmonic acid(JA)biosynthesis gene allene oxide cyclase 2(GhAOC2),reducing JA content and causing male sterility in the cotton HT-sensitive line H05.Anther sterility was reversed by exogenous application of methyl jasmonate(MeJA)to early buds.To elucidate the role of GhAOC2 in JA biosynthesis and identify its putative contribution to the anther response to HT,we created gene knockout cotton plants using the CRISPR/Cas9 system.Ghaoc2 mutant lines showed male-sterile flowers with reduced JA content in the anthers at the tetrad stage(TS),tapetum degradation stage(TDS),and anther dehiscence stage(ADS).Exogenous application of MeJA to early mutant buds(containing TS or TDS anthers)rescued the sterile pollen and indehiscent anther phenotypes,while ROS signals were reduced in ADS anthers.We propose that HT downregulates the expression of GhAOC2 in anthers,reducing JA biosynthesis and causing excessive ROS accumulation in anthers,leading to male sterility.These findings suggest exogenous JA application as a strategy for increasing male fertility in cotton under HT.

1.Introduction

High-temperature(HT)stress induces male sterility in cotton,leading to yield reduction[1].In response to biotic and abiotic stresses,plants have evolved complex defense mechanisms[2,3].Jasmonic acid(JA),a stress-responsive signaling molecule,is synthesized from α-linolenic acid(α-LeA),which is oxygenated by 13-lipoxygenase and converted to the fatty acid hydroperoxide 13S-hydroperoxy-(9Z,11E,15)-octadecatrienoic acid(13-HPOT)[4].Then 13-HPOT is dehydrated to an unstable allene oxide by allene oxide synthase(AOS)and then hydrolyzed to α-or γketols or racemic oxophytodienoic acid(OPDA).In the presence of allene oxide cyclase(AOC),only the cis-(+)-enantiomer(9S,13S)of OPDA is formed.Knockout or silencing of AOC genes creates a JA-deficient mutant in rice[5,6].

The AOC gene was initially reported in tomato and is found as a single-copy gene[7].Homologs have since been cloned from numerous species,including barley[8],Arabidopsis[9],Medicago truncatula[10],Hyoscyamus niger[11],Camptotheca acuminata[12],Jatropha curcas[13],Physcomitrella patens[14],Glycine max[15]and Leymus mollis[16].Studies of various organisms have shed light on the biochemical and physiological roles of AOC genes in the adaptation to a variety of biotic and abiotic stresses.Overexpression of the JcAOC gene in E.coli conferred resistance to salt and low-temperature stress[13].Overexpression of the AOC gene of C.acuminata in tobacco plants conferred tolerance to lowtemperature and salt stress conditions[12].GmAOC1-and GmAOC5-overexpressing transgenic tobacco lines showed increased tolerance to salinity and oxidative stresses,respectively[15].However,in rice,aoc mutants reduced sensitivity to salinity stress conditions by reducing Na+ion concentrations and increasing those of reactive oxygen species(ROS)[17].Manipulating JA metabolism increased salt tolerance under salt-stress conditions in rice[18].

ROS activity has been recorded in plants under various biotic and abiotic stress conditions[19–21]and plays signaling roles in many organisms and signaling pathways[22].In plants,NADPH oxidases,also known as respiratory burst oxidase homologs(RBOHs),produce ROS in numerous developmental and physiological activities[23,24].Among 10 Arabidopsis RBOH proteins,RBOHD activity is mainly involved in ROS production[25],and among the ROS,hydrogen peroxide(H2O2)is the most studied because its half-life is relatively long[26].

HT exposure increases the level of ROS and leads to a decline in pollen viability[27–29].Anther dehiscence occurs in mature stamens.During this process,lignin starts accumulating in the endothecium walls of anthers and is the major compound involved in the thickening of secondary walls.The polymerization of lignin is associated with the concentration of hydrogen peroxide(H2O2),the major form of ROS in plant cells[30].In a previous study[31],a role for the NAC transcription factors NAC SECONDARY WALL THICKENING PROMOTING FACTOR 1(NST1)and NST2 was also found in the regulation of lignin synthesis in anther secondary walls.The downstream relationship between ROS and JA under stress conditions in pea leaves has been reported[19].

In a previous study[32],high day and night temperatures reduced JA content in the anthers of the HT-sensitive line H05,leading to male sterility.However,it is still unknown how the regulatory mechanism of JA,JA-Ile and JAZ proteins is associated with ROS during anther development in Gossypium hirsutum.Moreover,the role of JA-pathway genes under HT stress conditions(both day and night)as well as in the HT-tolerant line 84021 has not been studied.In the present study,our main objective was to find the gene having role in the biosynthesis of JA during HT stress and then knocked out that gene through CRISPR/Cas9 technology to disrupt JA biosynthesis to investigate the function of the GhAOC2 gene in the link between JA synthesis and HT-induced male sterility in cotton.Our current research lays a strong foundation for further research on cotton male sterility during anther development under HT and provides an important strategy for improving male fertility under HT by exogenous JA application.

2.Materials and methods

2.1.Plant materials and HT treatment

The cotton(G.hirsutum)cultivar H05 is sensitive to high temperature[33].Plants of H05 were grown in a glasshouse.At the bud stage,two temperature treatments were applied:1)normal temperature(NT)[28–35 °C/22–27 °C(day/night)]and 2)high temperature(HT)[39–41 °C/29–31 °C(day/night)].After 10 days,buds at the tetrad stage(TS;6–7 mm),the tapetum degradation stage(TDS;9–14 mm),and the anther dehiscence stage(ADS;greater than 24 mm)under NT and HT were collected and stored in liquid nitrogen or at-70 °C.

2.2.CRISPR/Cas9 mediated mutation

To investigate the function of GhAOC2,the CRISPR/Cas9 genome editing system was used to create a Ghaoc2 mutant in Jin 668,an easily transformed cultivar[34].The pRGEB32R-GhU6.7-NPT IIGhAOC2 plasmid was constructed as described by Wang et al.[35].Two single-guide(sg)RNA targets(sgRNA1 from 74 to 97 bp;sgRNA2 from 465 to 488 bp)were prepared using CRISPRP online software(https://crispr.hzau.edu.cn/CRISPR2/).A fragment containing a tRNA-sgRNA1(CCTTCGTTTAAGTCAGAAACCTG)fusion and a gRNA-tRNA-sgRNA2(CCAGCTGTTGAAGCCGC TCCGGG)fusion were prepared from a template(pGTR)with the help of GhAOC2-CRISPR primers(Table S1)and cloned into the pRGEB32R-GhU6.7-NPT II vector as described by Wu et al.[36].The prepared vector was transferred to Agrobacterium(GV3101),which was used to infect cotton hypocotyls to obtain transgenic plants via tissue culture following Li et al.[34].Positive plants were selected on the basis of PCR confirmation with the Cas9 or NPTII primers(Table S1).Positive plants were checked for editing efficiency using the High-throughput tracking of mutations(Hi-TOM)technique as described by Liu et al.[37].

Owing to the male sterility of the mutant plants,wild-type(WT)pollen grains was placed manually onto the stigmas of mutant plants and the seeds were collected.The F1plants were confirmed by PCR with the Cas9 primers and Hi-TOM;the F1plants containing Cas9 showed the same male-sterile phenotypes as Ghaoc2-2 and-6 and had the same mutation at sgRNA1(4-bp deletion and 1-bp insertion);these plants were selected and treated with external MeJA(500 μmol L-1dissolved in 0.05% Tween 20)to obtain seeds.The seeds were grown on MS medium supplemented with antibiotic(kanamycin,50 mg L-1)to yield F2(Ghaoc2 mutant lines×WT)Cas9-negative plants(confirmed by Cas9 primer amplification and Sanger sequencing).

2.3.Microscopy and histology

Floral organs were photographed with a Canon EOS 80D digital camera.To determine the effect of the GhAOC2 gene on anthers,tissue sectioning was performed.Fresh buds were collected from WT and Ghaoc2-mutant plants and fixed in FAA solution(50% ethanol,10% formalin,and 5% acetic acid).Slides were prepared following Min et al.[33].To examine anther shape and lignification,the slides were dyed with 1% toluidine blue and 1% aniline blue solution and observed under a Zeiss Axio scope A1 microscope for imaging under a bright field at 20-,50-,and 100-μm scales.

For scanning and transmission electron microscopy,the anthers of WT and Ghaoc2-mutant plants at three stages(TS,6–7 mm;TDS,9–14 mm;and ADS,greater than 24 mm)were collected and scanning electron microscopy studies were performed following Wu et al.[38]using a JSM-6390/LV scanning electron microscope(Jeol,https://www.jeol.co.jp/en/).Transmission electron microscopy studies were performed with a transmission electron microscope(H-7650;Hitachi,Japan)following Wu et al.[38].

2.4.Quantification and exogenous application of jasmonic acid

For measurement of JA and JA-Ile,50 mg(fresh weight)of anthers was homogenized in 600 μL of 80%cold methanol.Samples were covered with black cloth and left at 4 °C with shaking for at least 12 h(overnight).The samples were centrifuged at 4 °C and 12,000 r min-1for 5 min,and JA was quantified on an ABI 4000Q-TRAP HPLC–MS system(Applied Biosystems,http://www.appliedbiosystems.com),with(±)-9,10-dihydrojasmonic acid(Olchemlm Ltd.,CAS:3572–64-3,Olomouc,CZ)used as an internal standard[32].

To confirm the role of JA in male sterility,a MeJA solution(500 μmol L-1in 0.05% Tween 20 v/v)was prepared following Khan et al.[32].During HT stress,MeJA solution was sprayed on early buds at 2-day intervals until the flower opened,while for mutant plants,MeJA solution was sprayed on early buds 3 times at 2-day intervals,with 0.05%Tween 20 solution spraying as mock treatment.Anthers and pollen grains were examined on the day of flower opening.

2.5.Reactive oxygen species(ROS)staining and H2O2 quantification

Anthers from the WT and F2plants of Ghaoc2 mutant lines×WT before and after MeJA treatment were collected for ROS detection following Ma et al.[1].For ROS staining,10 μmol L-12′,7′-dichlorodihydrofluorescein diacetate(2′,7′-DCFDA)dissolved in phosphate buffer solution,pH 7(PBS)was used.Pollen grains were incubated in 2′,7′-DCFDA and left in the dark for 30 min.After incubation,they were washed twice with PBS and then observed under a microscope at an emission wavelength of 522 nm and an excitation wavelength of 488 nm.

For H2O2quantification,anthers at several stages were collected,weighed to 50 mg,added to 200 μL of acetone(80%)and ground.To the mixture was added 400 μL of 80% acetone after grinding,followed by mixing and centrifugation.The supernatant was collected,and H2O2was quantified using an H2O2Quantitative Assay Kit(Sangon Biotech,# C500069-0250,Shanghai,China).For each sample,at least three biological and three technical replicates were used[1].

2.6.Quantitative reverse transcriptase PCR(qPCR)

Total RNA was extracted from all the samples,and complementary DNA(cDNA)was prepared using 3 μg RNA reverse-transcribed using the M-MLV(Promega)reverse transcript protocol.Genespecific primers were prepared(Table S1),and qPCR was performed using the MonAmp SYBR Green qPCR Mix(Low Rox)(REF:MQ10201S)kit and an ABI 7500 real-time PCR system[(Stage 1:95°C for 30 s(1 Cycle)and Stage 2:95°C for 5 s+60°C for 35 s(40 cycles)].The cotton gene GhUBQ7(UBIQUITIN7)was used as a reference gene,and the relative expression levels of the genes were calculated using the 2-ΔΔCTformula[33].

3.Results

3.1.High temperature(HT)reduced JA accumulation in the anthers of an HT-sensitive cotton line and caused male sterility

Under HT,the flower size was smaller(38.6±0.63 mm vertically and 54.61±1.28 mm horizontally)than under NT(63.40±1.21 mm vertically and 95.11±1.95 mm horizontally)(Fig.1A,B,K,L),and all anthers showed indehiscence under HT,whereas under NT,all dehisced normally(Fig.1C,D,M).Almost all pollen grains were normal in shape and showed high viability under NT,while all pollen grains were abnormal in shape and were inviable under HT(Fig.1E,F,N).

JA and JA-Ile(jasmonoyl-L-isoleucine)contents were measured in anthers of H05 under NT and HT,and HT significantly reduced JA and JA-Ile accumulation in H05 anthers(Fig.1G).To confirm that the reduced JA under HT caused male sterility,methyl jasmonate(MeJA)was exogenously applied to early buds under HT.The sizes of flowers treated with external MeJA(Fig.1H)remained nearly the same(50.61±0.76 mm vertically and 60.80±1.20 mm horizontally)as those of HT-treated flowers(Fig.1B,K,L),but all anthers showed dehiscence on the flowering day(Fig.1I,M),unlike HTtreated anthers(Fig.1D,M).Almost all pollen grains of MeJAtreated flowers were normal in shape and fertile(Fig.1J,N),unlike HT-treated pollen grains(Fig.1F,N).Accordingly,we assessed the expression levels of various JAZ gene.The steady-state levels of JAZ6,JAZ7,JAZ9,JAZ12,JAZ17,JAZ19,JAZ21,JAZ22,JAZ26,and JAZ28 transcripts were increased in HT-treated H05 anthers(Fig.S1;Table S2).Most of the other JAZ genes showed no expression changes during HT stress conditions.It seems that HT reduced JA accumulation in cotton HT-sensitive line anthers and inhibited the JA response pathway by activating the expression of JAZ genes,leading to male sterility.

3.2.HT reduced JA accumulation by suppressing the expression of allene oxide cyclase 2(AOC2)in anthers of the HT-sensitive line H05

Some JA biosynthesis genes(allene oxide synthase,acyl-coA oxidase,and allene oxide cyclase)showed expression differences under high day and night temperatures in H05[32].Among these genes,allene oxide cyclase(AOC)functions in the biosynthesis of JA by converting 12,13-epoxy-octadecatrienoic acid(12,13-EoTrE)to 12-OPDA.It affects the biological activity and stereoisomerism of JA molecules and functions in plant stress resistance[39].In Arabidopsis,four isoforms of AOC are present,among which AOC2 has been found to be the most active form,and among the four isoforms,AOC2 has been described as the most active form in terms of structure–function analysis by Otto et al.[40].Keeping in mind the multiple roles of AOC genes,the phylogenetic relationships between AOC genes were characterized in G.hirsutum and Arabidopsis using MEGA 7 with the neighbour joining method(Fig.2A).

In upland cotton(Gossypium spp.),the AOC gene family contains five members(GhAOC1–GhAOC5)that are possibly involved in JA biosynthesis[41].We named these genes GhAOC1–5,as shown in Fig.2A.To investigate how HT reduced JA accumulation in H05 anthers,the fragments per kilobase of transcript sequence per million base pairs sequenced(FPKM)expression values of GhAOC genes(GhAOC1–5)at TS anthers were collected from our data previously submitted to National Center for Biotechnology Information(PRJNA393079)and showed that the expression of all GhAOC genes was downregulated under HT stress compared to normal conditions(Fig.S2;Table S3).Among these GhAOC genes,the

GhAOC2 genes(Ghir_D08G004260.1 and Ghir_A08G004050.1)showed the highest expression at TS under normal conditions(Fig.S2;Table S3),and the expression of the GhAOC2 genes was downregulated in TS anthers when the HT-sensitive line H05 was treated with HT(Figs.1O,S2;Table S3).In the HT-tolerant line 84021,GhAOC2 genes showed the opposite trend(upregulation)in TS anthers during HT stress conditions compared to the HTsensitive line H05(Fig.1O).Considering that HT reduced the amount of JA in the H05 anthers at TS(Fig.1G),the GhAOC2 genes were chosen for further study.Sequence analysis showed that the GhAOC2 genes have two introns with an open reading frame of 537 base pairs encoding 178 amino acids(Fig.S3A).The GhAOC2 proteins have one conserved domain and four distinct motifs(Fig.S3B,C).Sequence alignment of Ghir_D08G004260.1 and Ghir_A08G004050.1 revealed a difference of seven nucleotides(Fig.S4A)and two amino acids(Fig.S4B).We named Ghir_D08G004260.1 as GhAOC2D and Ghir_A08G004050.1 as GhAOC2A.

3.3.Knockout of GhAOC2 by CRISPR/Cas9 caused male sterility

Fig.1.Effects of HT on H05 flowers,anthers and pollen and the recovery of the HT-induced phenotype by exogenous application of MeJA.(A–F)Phenotype comparison of H05(HT-sensitive line)under normal temperature(NT)(A)and HT stress(B)flowers.Anthers(C,D)and pollen(E,F)were observed on the day of flowering.The flowers of the HT group were smaller than those of the NT group(A,B);the anthers showed indehiscence in HT-treated anthers(D);the viability of pollen in HT-treated anthers(F)was lower than that in NT(E),as shown by TTC staining.(G)Quantification of JA and JA-Ile content in H05 anthers at TS(tetrad stage)under NT and HT.HT significantly reduced the JA and JA-Ile content in the H05 anthers at TS.Significant differences were assigned by Student’s t-test(**,P＜0.01).(H–J)Phenotype evaluation of MeJA-treated HT-stressed flowers(H).Anthers(I)and pollen grains(J)were observed on the flowering day.The flowers of the MeJA+HT group(H)were smaller than those of the NT group but almost identical to the flowers of the HT group(A,B,H);the anthers showed dehiscence under HT stress with MeJA application(I);the viability of pollen grains was similar to the NT(E)and in MeJA+HT stress(J),as shown by TTC staining.(K–N)Graphical representation of flower size,both horizontal(K)and vertical(L),anther dehiscence(M)and pollen fertility(N)at normal temperature(NT),HT stress and MeJA-treated HT-stressed flowers.(O)Expression analysis of the GhAOC2 gene in HT-sensitive line(H05)and HTtolerant line(84021)anthers at TS under NT and HT.HT significantly repressed GhAOC2 in the H05 anthers,whereas it was expressed in 84021 anthers at TS.JA,Jasmonic acid;JA-Ile,jasmonoyl-isoleucine;NT,normal temperature;HT,high temperature;H05→NT,H05 under normal temperature;H05→HT,H05 under high-temperature stress;H05+MeJA→HT,H05 with MeJA solution application during high-temperature stress(H–J);TTC,2,3,5-triphenyltetrazolium chloride.HN,H05 under normal temperature,HH,H05 under high temperature,8N,84021 under normal temperature,8H,84021 under high temperature.Scale bars,1 cm in(A–D,H–I)and 50 μm in(E,F,J).Rd arrows show indehiscente anthers.Ten plants from each treatment and three flowers from each plant were used in(A–F,J–N).Significant differences in(K–O)were assigned by one-way ANOVA(P＜0.05).Values with different letters are significantly different.

To identify the function of GhAOC2,genome editing was used to create a Ghaoc2 mutant in Jin 668,which is readily genetically transformed).Two sgRNAs were designed to mutate Ghaoc2:sgRNA1 was designed to target GhAOC2D(Ghir_D08G004260.1)and GhAOC2A(Ghir_A08G004050.1)to avoid the functional redundancy of homologous genes,while sgRNA2 was designed to target specifically GhAOC2D.The two sgRNAs were ligated into a CRISPR/Cas9 vector[35](Fig.2B).Transforming the vector into cotton(Fig.2C–I)yielded 60 transgenic plants.Of these,59 were positive by PCR using NPTII-and Cas9-specific primers.Hi-TOM was performed to evaluate the mutation at the selected target site in all transgenic lines.In Hi-TOM,samples are amplified using sitespecific primers(target-specific PCR)and by barcoding PCR;the products of each sample are barcoded.The results are categorized by mutation type and sorted in descending order of the read number for each mutation and sample individually.Mutations with the highest read numbers are extracted for each sample,ensuring the decoding of introduced mutations.The mutation types and positions of each sample are listed in detail,including read number,ratio,mutation type,mutation bases,and DNA sequence[37].Ten plants carried more than 60% mutations at the sgRNA1 target site,and three at sgRNA2 were obtained by Hi-TOM analysis(Fig.2J,K).Plants with more than 60% mutations can show a phenotype,but only plants with more than 80% mutations were used for further analysis,following the recommendation of Li et al.[37].Seven mutant plants with a＞80%mutation rate at the sgRNA1 target site showed male sterility(Fig.2J).Among the seven plants with mutations at sgRNA1,Ghaoc2-2 and-6 both showed only one type of mutation(4-bp deletion and 1-bp insertion),respectively,in the Hi-TOM results(Fig.2L)and were mutated at both target sites(sgRNA1 and sgRNA2)in Ghir_D08G004260.1(GhAOC2D)(Fig.2L,M).The mutations in Ghir_A08G004050.1(GhAOC2A)were checked in Ghaoc2-1 to-7 plants.The mutations were found at the sgRNA1 target site,but no mutations occurred at the sgRNA2 site(Fig.2L,M).The lines Ghaoc2-2 and-6 were selected for further study.To investigae the effect of sgRNAs on isoforms of GhAOC2,we evaluate the sgRNA target sequence in Ghaoc2 and found that sgRNA1 showed some similarity in GhAOC3,4 and 5.We evaluate these genes and found no mutation in GhAOC3,4 or 5(Fig.S5).This result confirmed that these sgRNAs did not affect the other isoforms because the desired nucleotide bases for the target region were not present in the isoforms(Fig.S5).We collected buds containing three HT-sensitive-stage anthers,those at the TS,TDS and ADS from Ghaoc2-2 and Ghaoc2-6 and compared them with WT(Jin 668)buds.The buds and anthers showed no significant differences at TS(Fig.S6A–C),but at TDS and ADS,the stigma was more exposed in both mutant line buds than in WT buds(Fig.S6D–I).However,flower size was reduced in the Ghaoc2-2 and Ghaoc2-6 mutant lines compared to the WT(Fig.3A–C).In the Ghaoc2-2 and Ghaoc2-6 mutant lines,all anthers were indehiscent and pollen grains were shriveled and sterile compared to WT anthers,which were dehiscent and carried pollen grains that were normal in shape and fertile(Fig.3D–I).

Fig.2.CRISPR/Cas9 genome editing and selection of Ghaoc2-mutant plants.(A)Phylogenetic tree of GhAOC2 and homologous genes in G.hirsutum.(B)T-DNA region of the pRGEB32/Cas9 vector containing the U6-7 promoter and two target regions.(C–I)Stages of tissue culture.After infecting the hypocotyl with Agrobacterium,the explants were placed on medium(without antibiotics)containing 2,4-D for 2 days(C).Then the explants were moved to fresh medium containing 2,4-D until callus formation(D,E).The callus was transferred to differentiation medium(F).Plantlets were transferred into the rooting medium(G).Seedlings were cultured in a nutrient solution to obtain strong roots(H)and transferred to a soil pot and kept in the greenhouse(I).(J,K)Percentage of the targeted mutation in T0 plants edited by CRISPR/Cas9.The mutation percentage of the GhAOC2 gene at the sgRNA1 site ranged from 0 to 100%(J),while at the sgRNA2 site,it ranged from 0 to 61.54%(K).The plants with more than 80% mutation rates are indicated by red arrows(J).(L)The sgRNA1-target-site mutations of GhAOC2D(Ghir_D08G004260.1)and its homologous gene GhAOC2A(Ghir_A08G004050.1)in T0 of Ghaoc2-mutant plants.sgRNA1 starts from 74 bp and ends at 97 bp.The numbers Ghaoc2-1 to 7 represent edited lines of GhAOC2D(Ghir_D08G004260.1)and GhAOC2A(Ghir_A08G004050.1).Numbers of reads are shown in detail at left.(M)The sgRNA2-target-site mutations of GhAOC2D(Ghir_D08G004260.1)and its homologous gene GhAOC2A(Ghir_A08G004050.1)in T0 of Ghaoc2-mutant plants.sgRNA2 starts from 465 bp and ends at 488 bp.The numbers Ghaoc2-1 to 7 represent edited lines of GhAOC2D(Ghir_D08G004260.1)and GhAOC2A(Ghir_A08G004050.1).Numbers of reads are shown in detail at left.

3.4.GhAOC2 mutation led to defective microspore development and anther dehiscence

To investigate microspore development in the Ghaoc2 mutant,cross-sections were made of anthers from the Ghaoc2-mutant lines and WT plants at TS,TDS,and ADS(Fig.3J–R).At TS,no marked differences were observed in tetrad formation between mutant lines and WT(Fig.3J–L),but at TDS,microspores were more shriveled in mutant lines than in WT(Fig.3M–O).In TDS,tapetum degradation was greater in Ghaoc2-mutant line anthers than in WT anthers(Fig.3M–O).At ADS,nearly all pollen grains were shriveled in Ghaoc2-mutant anthers,while all pollen grains had normal shape in the WT(Fig.3P–R).SEM showed that the mutant tetrad was covered with layers of unknown substances,while in the WT,the tetrad was clearly visible at TS(Fig.4A–C).At TDS,WT microspores were normal in shape and fully covered with spines,while spines were absent in Ghaoc2 mutant microspores,and the microspores were covered with unknown substances and were shriveled(Fig.4D–I).At ADS,only few pollen grains were found in the anthers of Ghaoc2-mutant plants,with spines with abnormal shapes and covered in unknown substances,while all WT pollen grains had spines with normal shapes(Fig.4J–O).

Fig.3.Comparison and histological study of the WT and the T0 of Ghaoc2-mutant lines in flower,anthers and pollen grains.(A–L)Phenotype comparison of WT(A,D,G)and the T0 plants(B,C,E,F,H,I)in flower(A–C),anthers(D–F),and pollen grains(G–I)on the day of flowering.The flowers of the T0 plants were smaller than those of the WT(A–C);the anthers showed indehiscence in the T0 plants(D–F);and the pollen viability in the T0 lines was lower than that in the WT,as shown by TTC staining(G–I)WT,Wild type(Jin 668);TTC,2,3,5-triphenyltetrazolium chloride.Scale bars,1 cm in(A–F)and 50 μm in(G–I).Red arrows show indehiscent anthers.(J–L)At TS,no clear differences were found in tetrad formation between the WT(J)and mutant anthers(K and L).(M–R)The Ghaoc2 mutant showed more shriveled microspores at TDS(N and O)than WT(M)and more shriveled pollen grains at ADS(Q and R)than WT(P).TS,tetrad stage;TDS,tapetum degradation stage;ADS,anther dehiscence stage;T,tetrad;TP,tapetum;Ms,microspores;P,pollen grains;En,endothecium.Scale bars,20 μm in(J–L)and 50 μm in(M–R).Five buds from each line were collected and 10–15 anthers from each bud were used in(J–R).

Fig.4.Scanning electron microscopy of pollen grains at several anther stages in the Ghaoc2 mutant and WT.(A–C)At TS,there were no clear differences in tetrad formation of WT(A)and mutant(B and C).However,the tetrads of mutant anthers(B,C)were covered with unknown substances,while in the WT(A),the tetrad was visible at TS.(D–I)At TDS,the microspores of mutant plants(E,F,H,and I)were shriveled in comparison with those of WT plants(D and G).There were no spines on the surface of Ghaoc2-mutant microspores(E,F,H,and I),unlike the WT microspores(D and G),which were full of spines.However,unknown substances adhered to the microspore surface in both the WT and Ghaoc2 mutant.(J–O)At ADS,the pollen of mutant plants(K,L,N,and O)was more shriveled compared to WT pollen(J and M).Only a few Ghaoc2-mutant pollen grains had spines on the surface(K,L,N,and O),and unknown substances adhered to the surface,unlike the WT pollen(J and M)surface,which was full of spines and had no adhering substances.TS,tetrad stage;TDS,tapetum degradation stage;ADS,anther dehiscence stage.Scales bars,20 μm in(A–C),100 μm in(D–F,J–L),and 50 μm in(G–I,M–O).Five buds from each line were collected and 10–15 anthers from each bud were used in(A–O).

Transmission electron microscopy(TEM)was used to observe the microspores or pollen grains in WT and Ghaoc2-mutant line anthers at TS,TDS,and ADS(Fig.5).At TS,high-density contents were observed in WT microspores,while Ghaoc2-mutant microspores showed low-density contents(Fig.5A–C).At TDS,almost all contents(including starch and lipid droplets)disappeared in both mutant lines compared to WT.The nexine layer was thicker and the intine was reduced in the pollen grains of mutant lines but not the WT(Fig.5D–F,d–f).At ADS,there was a clear intine in WT pollen grains;however,in Ghaoc2-mutant pollen grains,the intine was lacking.An extra layer wall was present on the spines of Ghaoc2-mutant pollen grains compared to WT pollen grains(Fig.5G–I,g–i).The TEM results were consistent with the cross-sectioning and SEM results,showing clearly defective pollen grains in the Ghaoc2-mutant lines associated with male sterility.

3.5.Ghaoc2 mutants displayed genetic stability of male sterility

Fig.5.Transmission electron microscopy of Ghaoc2 mutant lines and WT pollen grains at several anther stages.(A–C)At TS,there was a clear difference between Ghaoc2 mutant lines(B,C)and WT tetrads(A).Low-density content was found in Ghaoc2 mutant lines(B,C),while WT showed high-density content in microspores(A).(D–F)At TDS,almost all the contents(including starch and lipid droplets)disappeared in the Ghaoc2-mutant lines(E,F),unlike the WT(D).(d–f)The nexine wall was thicker and the intine layer was absent in microspores of Ghaoc2-mutant lines(e,f)compared to WT(d).(G–I)At ADS,almost all the contents disappeared in Ghaoc2-mutant microspores(H,I)compared to WT microspores(G).(g–i)The intine layer was thicker in WT pollen(g)but absent in Ghaoc2-mutant line pollen(h,i).TS,tetrad stage;TDS,tapetum degradation stage;ADS,anther dehiscence stage;In,intine;Ne,nexine;Ba,bacula;Te,tectum;ELW,extra layer wall;S,starch;LD,lipid drops.Scale bars,5 μm in(A to C),10 μm in(D–I),and 2 μm in(d–i).Five buds from each line were collected and 10–15 anthers from each bud were used in(A–I,d–i).

To test the genetic stability of male sterility of Ghaoc2 mutant lines,WT pollen grains were used to pollinate Ghaoc2-2 and-6 to develop F1plants.F1Ghaoc2 mutant lines×WT plants that were positive for Cas9 and showed the same male-sterility phenotypes as Ghaoc2-2 and-6 were subjected to Hi-TOM.Then,the F1plants from Ghaoc2-2×WT and Ghaoc2-6×WT,which showed 4-bp deletion and 1-bp insertion,respectively,at sgRNA1 were treated with external MeJA(500 μmol L-1dissolved in 0.05% Tween 20),and seeds were collected.The seeds of F1(Ghaoc2 mutant lines×WT)plants were grown on MS medium supplemented with antibiotic(kanamycin,50 mg L-1),and on the basis of root germination,Cas9-positive(roots germinated)and Cas9-negative(roots not germinated)plants were separated.Seedlings were transferred into a flask containing nutrient solution for new root development and then transferred into small pots containing soil for further growth.The F2(Ghaoc2 mutant lines×WT)Cas9-negative plants were confirmed by Cas9 primers and Sanger sequencing.Evaluating the two SNPs(at 151 bp and 174 bp,Fig.S4A)between GhAOC2A and GhAOC2D in these Cas9-negative F2plants revealed that the sgRNA1 target had a 4-bp deletion in GhAOC2A but no mutation in GhAOC2D genes in F2of Ghaoc2-2 mutant lines×WT plants.In the F2of Ghaoc2-6 mutant lines×WT plants,a 1-bp insertion occurred at the sgRNA1 site of GhAOC2A and GhAOC2D(Fig.S7A).The F2plants(Cas9 negative)of Ghaoc2-2 and Ghaoc2-6 mutant lines×WT contained respectively 6-bp and 5-bp deletions at the sgRNA2 target site of GhAOC2D(Fig.S7A).However,in neither F2plants of the Ghaoc2-2 mutant line×WT nor in F2plants of the Ghaoc2-6 mutant line×WT was GhAOC2A edited at the sgRNA2 target site(Fig.S7A).Respectively 22 and 25 Cas9-negative plants for Ghaoc2-2 and Ghaoc2-6 were retained for further experiments.The flower size was reduced in F2(Cas9 negative)of Ghaoc2-2 mutant lines×WT(46.62±0.78 mm vertically and 59.20±0.48 mm horizontally)and in the F2of Ghaoc2-6 mutant lines×WT(50.61±1.74 mm vertically and 77.61±1.33 mm horizontally)mutant lines compared to WT(63.40±1.21 m m vertically and 95.11±1.95 mm horizontally)(Fig.S7B–D).All anthers in the WT flowers were dehiscent,normal in shape,and fertile,while the F2of Ghaoc2 mutant lines×WT showed anther indehiscence,pollen shriveling,and sterility(Fig.S7E–J).These results confirmed that sgRNA1 could target GhAOC2D and GhAOC2A to avoid the functional redundancy of homologous genes,while sgRNA2 targeted specifically GhAOC2D and showed that the male-sterility phenotype in Ghaoc2 mutants was heritable.

3.6.GhAOC2 mutation disrupts JA biosynthesis

AOC function in establishing the naturally occurring enantiomeric structure of jasmonate,which acts as a signal during plant development[42].To confirm the effect of GhAOC2 gene knockout on JA biosynthesis,JA content was measured in WT and F2Ghaoc2 mutant lines×WT mutant anthers.The content of JA was significantly lower in F2of Ghaoc2 mutant lines×WT than in WT anthers at all stages(Fig.6A).This finding indicates that GhAOC2 is essential for JA biosynthesis in cotton anthers.Exogenous MeJA was applied to investigate the association between male sterility in the Ghaoc2 mutant and JA.Application of exogenous MeJA(500 μmol L-1dissolved in 0.05% Tween 20)to buds smaller than the TS buds(＜6 mm)caused no further growth,the buds became dry and dropped after 3–4 days,and application of exogenous MeJA solution to TS and TDS buds did not affect the continued growth of buds.More than half of the anthers showed dehiscence on the day of flowering(Fig.6B–M),and flower size remained approximately the same as in mutant flowers without the application of MeJA(Fig.6B–E,N,O).More than 52%of the anthers showed dehiscence(Fig.6F–I,P),and more than 83%of the pollen of MeJAtreated flowers was normal in shape and fertile(Fig.6J–M,Q).These results showed that exogenous MeJA application rescued the occlusive anthers and pollen abortion in the F2of Ghaoc2 mutant lines×WT plants,a result similar to the rescue of male fertility in the HT-sensitive line H05 under HT by exogenous MeJA.

Fig.6.MeJA treatment rescues the phenotype of F2 of Ghaoc2-mutant lines×WT lines.(A)JA contents were measured in anthers of WT and F2 plants at TS,TDS,and ADS.The JA content in anthers of F2 plants was significantly lower than in WT anthers at all three anther stages.Significant differences were found by using one-way ANOVA(P＜0.05).(B–E)Comparison of flowers in F2 plants with(C,E)and without(B,D)MeJA treatment.No significant differences in flower size were observed between the F2 lines with and without MeJA treatment.(F–I)Anthers in MeJA-treated F2 line flowers(G,I)compared to F2 line flowers without MeJA treatment(F,H).Anther dehiscence increased,and more than half of the anthers were dehiscent in MeJA-treated F2 line flowers.Red arrows show the dehiscent anthers(G,I).(J–M)The viability of pollen was observed by TTC staining,and more than 80%of pollen showed a red color after staining by TTC in MeJA-treated F2 line flowers(K,M).Viable pollen showed a red color,while inviable pollen was white.Yellow arrows show viable pollen(K,M).(N–Q)Graphical representation of flower size(horizontal(N),vertical(O)),anther indehiscence(P),and pollen sterility(Q)of MeJA-treated F2 plants compared to the F2 plants without MeJA treatment.Differences were tested with Student’s t-test(*,P＜0.05;**,P＜0.01).(R)H2O2 concentration in anthers at TS,TDS,and ADS.H2O2 was significantly greater in F2 line anthers than WT anthers at all stages.Significant differences were found by one-way ANOVA(P＜0.05).Values with different letters are significantly different.F2,The F2 plants of Ghaoc2-mutant lines×WT with male-sterility phenotype;Scale bars,1 cm in(B–I)and 50 μm in(J–M).Every measurement was performed on three independent replicates(one independent replicate included buds from 5 plants)of F2 of Ghaoc2-2 and Ghaoc2-6 mutant lines×WT plants(without Cas9 but homologous edited plants)in(A and R).Ten F2 plants of Ghaoc2-2 and Ghaoc2-6 mutant lines×WT plants(without Cas9 and homologous plants each)from each treatment[three flowers(MeJA-treated and without MeJA treatment)from same plant in each treatment]were used in(B–M).

3.7.Decreased JA biosynthesis caused accumulation of H2O2 in Ghaoc2 mutant anthers and pollen

HT increases endogenous ROS concentrations and leads to a decline in pollen fertility[27,43],and the oxidative burst is enhanced in JA signaling mutants[44].To investigate the interaction between JA and ROS under HT stress in cotton anthers,H2O2was measured.It was significantly higher in the F2of Ghaoc2 mutant lines×WT anthers than in WT anthers at all stages(TS,TDS,and ADS)(Fig.6R).When stained for ROS,WT anthers and pollen grains showed weaker ROS signals than F2of Ghaoc2 mutant lines×WT mutant anthers and pollen grains(Fig.7A–F).In aniline blue staining,WT anthers showed a stronger signal for lignin compared to F2of Ghaoc2 mutant lines×WT mutant anthers(Fig.7G–L),showing that secondary wall thickening was reduced in F2of Ghaoc2 mutant lines×WT mutant anthers.The expression of RBOH(respiratory burst oxidase homolog)genes was analyzed,and the four RBOH-D genes(Ghir_A05G026340.1,Ghir_D05G026370.1,Ghir_A05G019830.1 and Ghir_D05G019850.1)were significantly upregulated in F2of Ghaoc2 mutant lines×WT mutant anthers compared to WT anthers(Fig.7M–P).These results are consistent with increased H2O2content and stronger ROS signal in Ghaoc2-mutant anthers.However,the concentration of H2O2in anthers of mutant plants treated with 500 μmol L-1MeJA solution was significantly lower than that in mutant plant anthers(Fig.S8A),and the ROS signal was also reduced in anthers and pollen grains(Fig.S8B–I).

4.Discussion

Fig.7.ROS and aniline blue staining in the F2 lines of Ghaoc2-mutant lines×WT and WT anthers.(A–C)ROS staining of anthers of WT(A)and F2 lines(B and C).In the anthers of F2 lines,ROS accumulated more than in the WT anthers.Red arrows(B,C)show ROS accumulation.(D–F)ROS staining of pollen of WT(D)and F2 lines(E and F)showed that more ROS accumulated in the F2 line pollen than in WT pollen.(G–L)Aniline blue staining of anthers of the WT(G,J)and F2 lines(H,I,K,L).The lignification decreased in the F2 line anther wall compared to that in the WT anther wall.Yellow arrows(J)show lignification,and red arrows(K,L)show reduced lignification in anther walls.(M–P)Expression of RBOH-D genes in anthers of the WT and F2 lines.WT,Wild type(Jin 668);F2,F2 plants of Ghaoc2-mutant lines×WT with male sterility phenotype.Significant differences were found by one-way ANOVA(P＜0.05).Values with different letters are considered statistically significant.Scale bars,100 μm in(A–C),100 μm in(D–F),50 μm in(G–I),and 20 μm in(J–L).Yellow boxes show the zoom area of anthers(G–I).Ten plants from F2 of Ghaoc2-2 and Ghaoc2-6 mutant lines×WT plants without Cas9 and homologous plants each,and 10 anthers from three flowers in each plant were used in(A–L).

Allene oxide cyclase are believed to be of special importance to JA biosynthesis and regulation in plants[41,45].In cotton,the AOC gene family contains five members(GhAOC1–GhAOC5),among which GhAOC4 has been found[41]to be highly expressed in anthers.In the present study,the GhAOC2 gene was knocked out using the CRISPR/Cas9 system to elucidate its role in JA biosynthesis and to identify its contribution to HT-induced male sterility.Thus,Ghaoc2 gene knockout in cotton disrupts the JA biosynthesis pathway,in turn leading to male sterility,which may mimic plant male sterility under high-temperature stress.In the present study,JA and JA-Ile amounts were reduced,but the expression of GhJAZ transcripts was increased in HT-treated H05 anthers(Fig.S1;Table S2).In the previous study[46],it was found that GhJAZ2 acts as a repressor in JA signaling in cotton and several studies[46–49]have indicated that JA and JA-Ile play a role in the degradation of the jasmonate-zim domain(JAZ)transcriptional repressor.Furthermore,Chen et al.[50]also found that during insect attack on Arabidopsis,HARP1 directly attached to JAZ repressors,preventing COI1-mediated JAZ degradation and ultimately reducing JA signals.It appears that HT reduced JA accumulation in cotton HT-sensitive line anthers and inhibited the JA response pathway by activating the expression of JAZ genes and,as a result,causing male sterility.The JA signal was shown[51]to be necessary for male fertility.In Arabidopsis,there was antagonism between JA and IAA in anther dehiscence[52].In our previous study[33],the IAA content was increased in TDS and ADS under HT in H05,and exogenous IAA application in 84021(an HT-tolerant line)before HT increased anther indehiscence[33,53],suggesting that excessive auxin signaling may suppress JA biosynthesis,leading to anther indehiscence under HT.However,exogenous auxin completely rescued male sterility caused by HT in barley and Arabidopsis[54],and in miRNA157-overexpressing cotton anthers,suppressed auxin signaling also caused male sterility under HT[53].These results indicate that higher or lower levels of auxin are both detrimental to male fertility.Increased JA biosynthesis in Arabidopsis auxin perception mutants also results in male sterility[52].Thus,a systematic regulated JA and auxin signaling pathway must be essential for male fertility under HT.

The H2O2concentration increased markedly in the F2of Ghaoc2 mutant lines×WT lines compared with the WT;after exogenous application of MeJA,anthers and pollen grains of the F2of Ghaoc2 mutant lines×WT showed little accumulation of H2O2,a finding consistent with JA reduction of H2O2function[19].The RBOH protein is responsible for the synthesis of H2O2[24,25,55,56].The RBOHD genes were upregulated,ultimately increasing the amount of H2O2.This finding is in accord with previous studies,such as in grapevine[57],in which VvRBOHD genes were highly expressed during SA hormone treatment.The promoter of RBOH genes contained the binding site(TGACG and CGTCA)for the conserved domains(Tify and CCT-2)of JAZ proteins(Fig.S9).These two domains(Tify and CCT-2)function in abiotic stress response[58],suggesting that HT may induce the expression of JAZ genes and activate the expression of RBOH genes by binding of JAZ proteins to the promoter of RBOH genes,ultimately increasing H2O2.During anther dehiscence,a key process is the thickening of secondary walls in the endothecium cells of anthers.NST1,NST2,and NST3 play roles in secondary wall thickening in Arabidopsis thaliana[59,60],and an imbalance of ROS affects NST1 and NST2,leading to anther indehiscence[4].Wang et al.[30]also found that disruption in the thickening of secondary walls arrests anther dehiscence and pollen release and ultimately leads to male sterility.Zhang and Xing[61]found the accumulation of ROS to be due to increased MeJA during single-cell analysis.In our study,the F2of Ghaoc2 mutant lines×WT showed reduced JA biosynthesis and accumulated more ROS in indehiscent anthers,a finding similar to that of Abouelsaad and Renault[62],who reported increased ROS and H2O2in JA-mutant tomato plants compared to WT tomato plants under salt stress conditions.The reduced JA biosynthesis,may have functioned to reduce lignification by disrupting ROS accumulation in peroxisomes,ultimately disabling the thickening of secondary walls in the endothecium cells of anthers and producing indehiscent anthers because the highly expressed RBOHD participates in the mechanism of cell death and in cell wall damage-induced lignification[63].But we found no direct evidence supporting this speculation.In Arabidopsis,single mutants of jar1 and jat1,with partial deficiencies in JA,became fertile,and complete or nearly complete deficiency of JA led to male sterility[64].In cotton,in contrast to Arabidopsis,partial JA deficiency caused male sterility(Figs.3A–I,6).We speculate that there is a lack of interconnectedness between the lignification of the anther wall and JA in Arabidopsis[64].However,the interconnectedness of lignification and JA in cotton provides defense against Verticillium dahliae,the cotton bollworm,and the cotton aphid[65].

In plants,the microspores are protected by two layers,the exine and intine,which form the cell wall of pollen grains[60]{Blackmore,2007 #245;Li,2017 #246}[66,67].The exine is further divided into two layer,the sexine and nexine,in which sporopollenin functions.Sporopollenin is composed of fatty acids and phenolics,and the tapetum is considered to be its precursor[68].The tapetum contains mitochondria,which are metabolically active[69]and lead to ROS production during normal development[70].Previous studies[71–73]found that mutation in lipid metabolism genes resulted in tapetum abnormalities and produced defects in exine development.In further study[72],the role of gibberellin(plant hormone)was also found in tapetum differentiation as well as in programmed cell death(PCD).In the Ghaoc2 mutant,abnormal spines,thicker nexine,and reduced intine in pollen grains were observed and associated with male sterility,as described previously[38,74].It seems that the Ghaoc2 mutation disturbs ROS homeostasis in the tapetum,leading to untimely PCD.Undesirable PCD leads to accumulate more sporopollenin in the exine and disturbs the normal pattern of exine development[75,76].As the amount of sporopollenin starts increasing,it disrupts the fatty acid pathway and fails to create spines on the pollen walls of the male sterile line,as described by Wu et al.[38].The intine is composed of cellulose,hemicellulose,pectin polymers,hydrolytic enzymes,and hydrophobic proteins derived from the microspores[68,77].In our study,the nexine was thicker in the mutant,inhibited the development of microspores,caused the microspores to be degraded following the degradation of intine,and finally,at ADS,the mutant pollen grains lacked the intine layer(Fig.5).

We propose that HT suppresses GhAOC2 in cotton anthers,reduces JA content,upregulates the expression of RBOHD genes,increases the amount of ROS,and disrupts the lignification process in the endothecium wall of anthers,disrupting the thickening of secondary walls,ultimately producing indehiscent anthers and male-sterile cotton(Fig.S10).

CRediT authorship contribution statement

Aamir Hamid Khan:Writing–original draft,Software,Data curation,Methodology.Yizan Ma:Software,Data curation.Yuanlong Wu:Writing–review & editing.Adnan Akbar:Writing–review&editing.Muhammad Shaban:.Abid Ullah:Formal analysis.Jinwu Deng:Software,Data curation.Abdul Saboor Khan:Formal analysis.Huabin Chi:Formal analysis.Longfu Zhu:Funding acquisition.Xianlong Zhang:Funding acquisition,Supervision.Ling Min:Conceptualization,Funding acquisition,Supervision,Writing–review & editing.

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

We are grateful for funding support from the National Natural Science Foundation of China(32072024),the Fundamental Research Funds for the Central Universities(2021ZKPY019),and the National Key Research and Development Program of China(2018YFD0100403,2016YFD0101402).We thank Dr.Hongbo Liu(Huazhong Agricultural University,China)for help in hormone analysis and Mr.Qinghua Zhang(Huazhong Agricultural University)for help with Hi-Tom sequencing.We thank Jianbo Cao,Limin He,and Fangmei Zhang(Huazhong Agricultural University)for their help with scanning and transmission electron microscopy.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2022.05.009.