Xio Chen,Suxin Yng,Yohu Zhng,Xiobin Zhu,Xinjing Yng,Chunbo Zhng,Hiyn Li,Xinzhong Feng,*

a Key Laboratory of Soybean Molecular Design Breeding,Northeast Institute of Geography and Agroecology,The Innovative Academy of Seed Design,Chinese Academy of Sciences,Changchun 130102,Jilin,China

b University of Chinese Academy of Sciences,Beijing 100049,China

c School of Life Science,Jilin Agricultural University,Changchun 130118,Jilin,China

d Jilin Academy of Agricultural Sciences,Changchun 130000,Jilin,China

Keywords:

A B S T R A C T Soybean[Glycine max(L.)Merr.]provides a rich source of plant protein and oil worldwide.The commercial use of transgenic technology in soybean has become a classical example of the application of biotechnology to crop improvement.Although genetically modified soybeans have achieved commercial success,hybrid soybean breeding is also a potential way to increase soybean yield.Soybean cytoplasmic malesterile(CMS)lines have been used in three-line hybrid breeding systems,but their application to exploiting soybean heterosis has been limited by rare germplasm resource of sterile lines.The generation of various genetic diversity male-sterile soybean lines will help to overcome the shortcoming.In this study,we used targeted editing of AMS homologs in soybean by CRISPR/Cas9 technology for the first time to generate stable male-sterile lines.Targeted editing of GmAMS1 resulted in a male-sterile phenotype,while editing of GmAMS2 failed to produce male-sterile lines.GmAMS1 functions not only in the formation of the pollen wall but also in the controlling the degradation of the soybean tapetum.CRISPR/Cas9 technology could be used to rapidly produce stable male-sterile lines,providing new sterile-line materials for soybean hybrid breeding systems.

1.Introduction

Heterosis,or hybrid vigor,refers to a phenomenon in which heterozygous hybrid progeny are superior to both homozygous parents in one or more traits[1,2].Heterosis breeding,which is one of the most effective measures to increase crop yield,has increased yields by 3.5%–15%in wheat,55%in rice,47%in common bean,and 200% inBrassicaoilseeds[3].Heterosis breeding has become the main breeding method for rice,maize,and rapeseed[4–6].The success of hybrid breeding depends greatly on the strength of heterosis as well as the selection and application of male-sterile lines.Male sterility is an important precondition for heterosis breeding and hybrid seed production in crops,especially sexually propagated crops.Compared with artificial emasculation and chemical mutagenesis,using male-sterile lines to produce hybrids can reduce the production cost of hybrid seeds,improve the quality of hybrids,and even expand the range of heterosis use[3–5].

With the development of genomics and biotechnology,genome-editing technology,in particular the CRISPR/Cas9 system with its characteristic simple operation and high efficiency,has become a powerful tool for crop improvement[7].This system can not only precisely improve the agronomic characteristics of crops,but also accelerate crop breeding cycles[8].For example,Li et al.[9]used CRISPR/Cas9 to target the endogenousCSAgene in rice to produce a photosensitive male-sterile line,providing a straightforward method for producingcsa-based rP(T)GMS lines in the two-line hybrid rice system.Zhou et al.[10]used CRISPR/Cas9 to edit specifically the temperature-sensitiveTMS5gene in rice,and developed new‘‘transgene-clean”TGMS lines,accelerating the breeding of sterile lines.Okada et al.[11]used CRISPR/Cas9 to knock outMs1,allowing the rapid generation of malesterile hexaploid wheat lines that could be used for commercial hybrid breeding.Du et al.[12]used CRISPR/Cas9 to modify a stamen-specific gene,SlSTR1,and created a new tomato malesterile line that could be used for hybrid breeding.Zhang et al.[13]used CRISPR/Cas9 to edit the key meiotic geneLpDMC1in ryegrass and caused a completely male-sterile phenotype.Xu et al.[14]used CRISPR/Cas9 to knock outOsROS1to investigate its role in the regulation of rice fertility.Thus,CRISPR/Cas9 technology provides a novel and convenient method for producing male-sterile lines.

Cells of the tapetum,the layer of somatic cells closest to the pollen,regulate sporogenesis and pollen wall development[15].TheABORTED MICROSPORES(AMS)gene is a bHLH transcription factor that affects tapetal development inArabidopsis thaliana[16].Theaborted microspores(ams)mutant displays male sterility with aborted microspores and reduced filament length.The putative rice homolog ofAMS,TDR,causes male sterility when mutated,owing to delayed degradation of the tapetum leading to pollen abortion[17–19].Thus,mutations of theAMSgene in bothArabidopsisand rice lead to a male-sterility phenotype,and it can be concluded that pollen and anther development are highly conserved across monocots and dicots[17–20].

It has been reported[21,22]that soybean has strong heterosis,and using soybean heterosis to increase soybean yields has been an efficient strategy to increase soybean yield.The bottleneck in commercial application of soybean heterosis is a lack of suitable male-sterile lines.Although CRISPR/Cas technology has been applied to improve soybean agronomic traits,such as altering the plant architecture of soybeans[23],the creation of late-flowering mutants[24]and the generation of seed lipoxygenase-free soybeans[25],creating nuclear male sterility by CRISPR/Cas9 technology has not been reported.In this study,we used CRISPR/Cas9 technology to generate new male-sterile lines through editingAMSorthologs in soybean.By comparing pollen fertility and tapetum development of edited lines,the most suitable ortholog for the generation of male sterile lines was identified.This work provides a new method to produce new sterile lines in diverse genetic background and will help to accelerate the hybrid soybean breeding.

2.Materials and methods

2.1.Plant materials and growth conditions

The soybean cultivar,Williams 82,was used for transformation.Seeds of wild-type Williams 82(as a control)and all seeds collected from CRISPR/Cas9 transgenic plants were planted in growth chambers under conditions with a 14 h light and 10 h dark photoperiod at 26 °C.

2.2.Construction of phylogenetic trees

Based on the sequence of theAMSgene inArabidopsis thaliana,BLAST alignment with the soybean genome reference sequence from Phytozome database(www.phytozome.net/)indicated that there were twoAMSgenes in soybean,which we namedGmAMS1andGmAMS2.The full-length protein sequences of soybeanAMSgenes and their homologs were identified for the following species:Phaseolus vulgaris,Vigna radiata,Lotus corniculatus,Lupinus angustifolius,Arachis hypogaea,Medicago truncatula,Arabidopsis thaliana,andOryza sativa.These sequences were retrieved from the Phytozome database and were used to construct a phylogenetic tree to study the evolutionary relationships among them.MEGA 7.0[26]was used to construct a neighbour-joining unrooted tree.The protein conserved motifs fromAMShomologs were then identified using MEME and TBtools software[27].Multiple sequence alignment of them was performed using DNAMAN software[28].

2.3.Guide RNA(gRNA)design

The sequences ofGmAMSgenes derived from Williams 82 were retrieved from Phytozome database and used for gRNA design.Target sites were designed using the web tool CRISPR-P[29].All gRNAs were designed to possess a canonical PAM(Proto-spacer adjacent motif).

2.4.Construction of the CRISPR/Cas9 system and soybean transformation

The Cas9/sgRNA plasmid construction kit(Viewsolid Biotech Co.,Ltd.,Beijing,China)was used to construct the CRISPR/Cas9 plasmids.These plasmids contained a dicotyledon codonoptimized dpCas9(under the GmUbi3 promoter)and a gRNA scaffold(under the GmU6-2 promoter).Construction of the CRISPR/Cas9 system followed the manufacturer’s instructions.

The resulting CRISPR/Cas9 plasmids carrying gRNAs were transformed into theAgrobacterium tumefaciensstrain EHA105,followed by a soybean transformation protocol previously described by Yamada et al.[30]with minor modifications.Sterilized Williams 82 seeds were germinated in B5 medium and cultured overnight.They were then cut along the navel to divide the embryo and cotyledon into two and the large buds and seed coats were removed.The explants were wounded with a scalpel and immersed in an infection solution(B5:0.31 g L-1,MES:3.9 g L-1,B5 Vitamin:112 mg L-1,GA3:0.25 mg L-1,6-BA:1.67 mg L-1,AS:80 mg L-1,DTT:180 mg L-1,Silwet:200 μL).After 5 days in co-cultivation medium,the explants were transferred to a shootinduction medium without glufosinate.Seven days later,they were transferred to a shoot induction medium containing 6 mg L-1glufosinate to induce cluster shoots and cultivated for 2 weeks.They were then transferred to shoot elongation medium containing 6 mg L-1glufosinate to induce the elongation of the cluster shoots.After the clustered shoots elongated to 5 cm,they were transferred to rooting medium.When the roots grew to 3–5 cm in length,they were transplanted into soil and cultured in growth chambers under 14 h light and 10 h dark at 26 °C.

2.5.Detection of CRISPR/Cas9-induced transgenic lines in T0 plants

The leaves of T0generation CRISPR/Cas9-induced transgenic plants were placed in 1.5 mL centrifuge tubes,and crushed with a pestle.Detection buffer(200 μL)was added and mixed with each sample.The PAT/bartest strips(Youlong Biotech Co.,Ltd,Shanghai,China)were inserted into the above centrifuge tubes and the positive red bends appeared in the strips after 10 min.

2.6.Identification of the genotypes of edited plants

Genomic DNA was extracted from the seedlings of wild-type and all positive transgenic plants,and DNA fragments spanning target sites were amplified by PCR.The target site forGmAMS1-gRNA was amplified with the primersGmAMS1-F(5′-GTTCCACTAT CATCATCCAGGTC-3′)andGmAMS1-R(5′-ATCCCAACCACTCAA GAAACCTA-3′).The target site forGmAMS2-gRNA was amplified with the primersGmAMS2-F(5′-CCCAACTCACCCATTGAACATCC-3′),GmAMS2-R(5′-GCACAACAGCATCCCAACCACTC-3′).The target sites forGmAMS1-GmAMS2-gRNA were amplified with the primersGmAMS1-F,GmAMS1-R andGmAMS2-F,GmAMS2-R.The editing mutations were visualized by DNA sequencing peak chromatogram of the target sites with the SeqMAN software.The wild type and homozygous mutations showed singlets at each target site with no overlapping peaks.Heterozygous mutations showed overlapping peaks from the target sites to the end.The mutation types were identified by sequence differences between wild-type and mutations.The CRISPR/Cas9-induced mutation frequency(%)was calculated from the number of lines with mutations among all the positive transgenic plants.

2.7.Evaluating pollen viability

Pollen viability was assessed by I2–KI and Alexander staining[31].Pollen grains were mounted on glass microscope slides and imaged using a microscope equipped with a SC2000 digital CMOS camera(DM2500,Leica,Germany).Images were analyzed with CapStudio software(LASX,IMG,China).

2.8.Scanning electron microscopy(SEM)

The fresh anthers of wild-type andGmamsmutants were fixed in 2.5%glutaraldehyde solution overnight at 4°C as described previously[32].Observation and recording of the SEM images were performed with both a HITACHI S-3400 and an electron microscope(JSM-IT500,JEOL,Japan).

2.9.Histological analysis

Flower buds at various developmental stages ofGmamsmutants and wild-type parental lines were fixed overnight at 4 °C in FAA buffer(5% formaldehyde,10% glacial acetic acid,and 50% ethanol)after two rounds of vacuum infiltration for 20 min.The tissues were successively incubated in 70% ethanol(2 h),85% ethanol(2 h),95% ethanol(2 h),100% ethanol(3×2 h),ethanol/xylene(75%/25%,2 h)and ethanol/xylene(50%/50%,overnight)at room temperature.The next day,these tissues were sequentially incubated in ethanol/xylene(25%/75%,2 h)and 100% xylene(3×2 h)at room temperature,and then were transferred to a hybridization oven(60 °C)and successively incubated in paraplast/xylene mix(25%/75%,2 h)and paraplast/xylene mix(50%/50%,overnight).The samples were then sequentially incubated in paraplast/xylene mix(75%/25%,2 h)and 100% paraplast(1 day,with the solution being changed three times)at 60 °C.100% paraplast was then used for embedding of the tissues,and a microtome(RM2245,Leica,Germany)was used for sectioning[33].Sections were colored using toluidine blue and observed under the microscope(DM2500,Leica,Germany).

2.10.RNA isolation and real-time quantitative PCR analysis

Total RNA was extracted from wild-type soybean flower buds at several developmental stages using Trizol reagent(Tiangen,DP424)according to the manufacturer’s instructions.RT-PCR was performed using theTransScriptOne-Step gDNA Removal and cDNA Synthesis SuperMix(TRAN,AT3311).Real-time quantitative PCR(RT-qPCR)was performed using a 2*RealStar Green Fast Mixture(GenStar,A301).GmActin11was used as a reference gene and the relative expression level was calculated using the 2-ΔΔCtmethod.Three independent biological replications were performed for each sample[33].The primers are listed in Table S1.

3.Results

3.1.AMS gene orthologs identified in soybean

Using the protein sequence of AMS inArabidopsis thaliana,two soybean orthologs of theAMSgene,Glyma.10G281800andGlyma.20G107500,named respectivelyGmAMS1andGmAMS2,were identified and showed respectively 60.4% and 56.4% amino acid sequence identity withArabidopsis AMS.Phylogenetic analysis showed that GmAMS proteins were most closely related toPhase-olus vulgarisandVigna radiatahomologs(Fig.1A).GmAMS1 was more closely related toArabidopsisAMS than to GmAMS2.MEME analysis showed that GmAMS1 contained 10 and GmAMS2 contained 9 conserved motifs(Fig.1B).The functions of GmAMS protein homologs in these species were relatively well conserved.GmAMS proteins shared a very conserved helix-loop-helix DNAbinding domain with 50 amino acid residues with AMS and TDR(Fig.S1).Besides this conserved domain,GmAMS1andGmAMS2also contained another bHLH-MYC-N domains(Fig.1C).Although the functional domains ofGmAMS1andGmAMS2are very similar,the positions of the bHLH-MYC-N domain were quite different in these two proteins,and the domains also differed by five amino acids.Comparison of the amino acids sequences revealed an extra 48 residues in the N-terminal of the bHLH-MYC-N domain ofGmAMS2and 44 residues in a 63-residue lower-similarity region in the middle ofGmAMS1(Fig.1C).

The relative expression level ofGmAMS1was higher than that ofGmAMS2at several flower-bud developmental stages(Fig.1D).The expression levels ofGmAMS1were respectively 20.8 and 17.2 times higher than that ofGmAMS2in 1.5-mm and 2-mm flower buds.

3.2.Targeted mutagenesis of GmAMS genes mediated by CRISPR/Cas9 system

We choose the bHLH-MYC-N domain as the mutagenesis target for the mutation of bothGmAMS1andGmAMS2,and gRNA scaffold was driven by the soybean GmU6-2 promoter,the dicotyledon codon-optimized dpCas9 by soybean GmUbi3 promoter(Fig.2A).The first target site was designed to create a specific single knockout mutant ofGmams1located at bp 59–77 in the first exon ofGmAMS1on chromosome 10(Fig.2B),and the second target site was designed to create a specific single knockout mutant ofGmams2,located at bp 64–83 of the second exon ofGmAMS2on chromosome 20(Fig.2C).The third target site was designed to create a double knockout mutant ofGmams1andGmams2,targeting regions located at bp 28–46 of the first exon ofGmAMS1and 33–51 of the second exon ofGmAMS2(Fig.2D).

3.3.CRISPR/Cas9-induced mutation frequency for GmAMS genes

The mutants ofGmAMSinduced by CRISPR/Cas9-mediated genome editing at the three target sites were namedGmams1,Gmams2,andGmams1,2.In total,179 T0transgenic lines were identified by test strips used for the detection of the selectable marker genebar,and 52%(32 of 61),40%(21 of 52),and 45%(30 of 66)T0lines were positive forGmams1,Gmams2,andGmams1,2at these three target sites.DNA was extracted from the 83 positive plants and wild-type,and the DNA fragments flanking the target site were amplified by PCR(Fig.2E)to determine the mutation frequency in T0lines.The results of the amplified sequence analyses from these target sites in T0transgenic plants are summarized in Fig.2F.In the T0generation,among the 32 positiveGmams1single mutants,the mutation frequency at the target site was 25%(8/32).The mutation frequency of the 21 positiveGmams2single mutants was 14%(3/21)and that of the 30 positiveGmams1,2double mutants was 23%(7/30).

3.4.Inheritance of targeted mutations in T1 generation

Sequencing of the progeny of T0knockout lines demonstrated that the targeted sites were edited in the 8Gmams1,3Gmams2,and 7Gmams1,2edited lines of the T1generation plants.The genotypes of the CRISPR/Cas9-induced homozygous mutations are summarized in Fig.3.Among these edited lines,three mutation types were detected:deletions,insertions,and substitutions(Fig.3).Among the 8 male-sterile mutants in the T1generation,six wereGmams1single mutants and the other two wereGmams1,2double mutants(Fig.3A,C).Homozygous edited plants ofGmAMS2with base deletion at the target site in the T2generation plants were obtained(Fig.3B),but they were all fertile(Fig.S2).

Fig.1.Phylogenetic tree and multiple sequence alignment of GmAMS.(A)Phylogenetic tree of GmAMS and its protein homologs from Phaseolus vulgaris,Vigna radiata,Lotus corniculatus,Lupinus angustifolius,Arachis hypogaea,Medicago truncatula,Arabidopsis thaliana,and Oryza sativa.(B)The conserved motif of GmAMS and its homologs,which are represented by colored boxes.(C)Alignment of the protein sequences of GmAMS1 and GmAMS2.Dark blue marks the same amino acid in GmAMS1 and GmAMS2.Light blue marks amino acids differing between GmAMS1 and GmAMS2.The bHLH-MYC-N domain is represented by the green line and the HLH domain is represented by a red line.(D)Expression levels of GmAMS1 and GmAMS2 relative to GmActin11 at several developmental stages of flower buds in Williams 82.Values are means of three biological replicates and error bars represent standard deviation(SD).Statistical contrasts were performed by t-test,and significant differences are marked with asterisk(**,P＜0.01;***,P＜0.001).

Fig.2.Structure diagram of the CRISPR/Cas9 vector and the target sites in GmAMS genes.(A)Structure diagram of CRISPR/Cas9 vector.(B)Single target site in GmAMS1.(C)Single target site in GmAMS2.(D)Double target site in GmAMS1 and GmAMS2.(E)PCR amplification results of DNA fragments flanking the target sites of GmAMS1 and GmAMS2 in the T0 generation.M,Trans2K DNA marker;lane 1 shows the PCR amplification result from wild-type plants,and the other lanes show the PCR amplification results from partially edited plants in the T0 generation.(F)Summary of mutation frequency at three target sites in the T0 generation.

In the sixGmams1male-sterile mutants in the T1generation,one(Gmams1-16–1)harbored a homozygous mutation with a 3-bp deletion that caused a single amino acid deletion in the bHLH-MYC-N domain.Three mutants(Gmams1-143–2,3,and 4)harbored homozygous mutations with 2-bp deletions that caused frameshifts and premature translational termination during the protein translation.One mutant(Gmams1-148–4)harbored a biallelic mutation with a 7-bp deletion in one allele that caused frameshifts and premature translational termination,and the other allele was a 9-bp deletion that caused a 3-amino acid deletion in the bHLH-MYC-N domain.Another mutant(Gmams1-148–5)harbored a homozygous mutation with a 9-bp deletion,resulting in a 3-amino acid deletion in the bHLH-MYC-N domain.

Of twoGmams1,2double male-sterile mutants in the T1generation,one(Gmams1,2–29-1)was a double mutant,with a Gto-A substitution atGmAMS1that changed the amino acid sequence in the bHLH-MYC-N domain,and a 4-bp deletion inGmAMS2,resulting in frameshift mutations that may change the function of the translated protein.The other mutant(Gmams1,2–123-2)also harbored mutations in bothGmAMS1andGmAMS2:a 1-bp insertion inGmAMS1that caused frameshifts and premature translational termination,and a 3-bp deletion inGmAMS2,resulting in an amino-acid deletion in the bHLHMYC-N domain.

3.5.CRISPR/Cas9-induced mutant plants exhibit male sterility in soybeans

The phenotype of T1homozygous edited plants was observed after genotyping by DNA sequencing.The T1male-sterile mutants isolated from the linesGmams1-143–2,3,and 4 were namedGmams1-1,because all of them contained the same homozygous mutation,a 2-bp deletion at the target site.Gmams1-1displayed complete male sterility with small pods but normal vegetative growth(Fig.4 A,B).With I2–KI staining,the pollen grains of a wild-type line appeared round and stained black,while those ofGmams1-1were shrunken and stained brown(Fig.4C).When the whole anther was dyed with Alexander’s stain,the pollen grains in wild-type anthers were dyed pink,whereas the pollen grains in theGmams1-1anthers were not dyed,indicated that they were immature(Fig.4D).SEM showed that the pollen grains ofGmams1-1mutant plants presented a shriveled morphology compared with wild-type pollen(Fig.4E).Thus,the pollen development ofGmams1-1mutant plants was abnormal,resulting in a completely aborted phenotype.Among theGmams1,2double mutants were two male-sterile mutants,and their phenotypes were consistent with that of aGmams1single mutant(Fig.S3).

3.6.Knockout of GmAMS1 affected the development of microspores and tapetum cells

We divided the development process of soybean anther into 14 stages according to appearance of characteristic cell structures,including the development of stamen primordium,formation of sporogenous cells,pollen mother cell,the meiosis and tetrad of pollen mother cell,microspore maturation and mature pollen release(Fig.S4).The tapetum cells appeared in the 5th period of anthers development,and completely disappeared in the 12th period.The microspores emerged in the 8th period,and developed into mature pollen grains in the 11th period(Fig.S4).

Before the tetrad stage,pollen grains ofGmams1-1mutant plants were not different from those of wild-type plants.However,at the microspore stage(stage 8),the outer-wall morphologies ofGmams1-1microspores were markedly different from those of wild-type spores(Fig.5).At stage 10,the tapetum of wild-type spores began to degrade,while inGmams1-1mutants,the tapetum did not initiate degeneration,but rather thickened and wrapped around the microspores.By stage 11,the tapetum of theGmams1-1mutant was abnormally enlarged and vacuolated,and the microspores did not develop into mature pollen grains as did those of the wild type.Further,the micropores were squeezed together,and the individual morphologies of the microspores were indistinct.In stage 13,the tapetums of wild-type lines were completely degraded,and the septum became cracked.However,the tapetum inGmams1mutants was still present,and pollen grains and tapetum continued to degenerate.Eventually,both the microspores and tapetum disintegrated completely,resulting in no mature pollen grains inGmams1-1mutants.

4.Discussion

Currently,although the CMS system has been successfully applied in soybean hybrid breeding,the yield of hybrid soybean has not drastically increased as hybrid rice and corn.One the main problem is from the rare number of identified CMS lines,which restricts the full utilization of heterosis in the three-line commercial hybrid seed production system.Besides discover CMS lines from soybean germplasm resources,the utilization of nuclear male sterility becomes possible for commercial utilization due to the rapid development of new hybrid breeding technology.SPT technology as a representative approach of nuclear male sterility is the predominant method for commercial maize hybrid seed production[34].This method effectively solves the problem of maintaining and reproducing the recessive nuclear male sterile line of crops,which has important practical value in cross breeding and hybrids production.Combination SPT technology with nuclear male sterility,the third-generation has been widely applied in rice[4,35]and corn[5,34].It has demonstrated the advantages of freedom of combination,stable fertility,and distinguishability,which can realize the commercial production of crop hybrids.In this study,we use the CRISPR/Cas9 technology to create soybean male sterile lines,which not only can expand the germplasm resources of soybean male sterility,but also lay the foundation for the third-generation soybean hybrid breeding system.

As a diploid plant evolved from palaeotetraploids,soybeans have a highly repetitive genome,in which about 75% of its genes exist in the form of multiple copies,resulting in a high degree of genetic redundancy[36].In this study,we identified twoAMShomologous genes in soybean and performed knockout experiments targeting these twoAMSgenes in soybean using CRISPR/Cas9.In order to detect whether theGmams1mutants carry theGmAMS2mutation,we amplified the DNA fragments flanking the target sites ofGmAMS1andGmAMS2in theGmams1mutants.Among theGmams1mutants,mutations were detected only at theGmAMS1target site and not at theGmAMS2target site,indicating that the male-sterile phenotype of theGmams1mutants was caused only by theGmAMS1mutation(Fig.S5),and we infer thatGmAMS1is the key gene regulating tapetal development in soybean.

Fig.4.CRISPR/Cas9-induced mutant plants exhibit male sterility in soybeans.(A)The whole-plant phenotypes of WT and Gmams1-1 plants.Scale bar,5 cm.(B)Pods of WT and Gmams1-1 plants at maturity.Scale bar,1 cm.(C)I2–KI staining of WT and Gmams1-1 pollen grains.Scale bars,100 μm.(D)Alexander’s staining of WT and Gmams1-1 pollen grains.Scale bars,100 μm.(E)Scanning electron microscopy of WT and Gmams1-1 pollen grains.Scale bars,5 μm.

Fig.5.Semithin sections of anther lobes at four developmental stages from WT and Gmams mutants after staining with toluidine blue.Tapetum layer(T),microsporocytes(Msp),septum(St),and pollen grains(PG)indicated by arrows.Scale bars,50 μm.

It has been reported[37]thatAMShas been shown to act as a master regulator of pollen wall development,and its mutation could influence the accumulation of sporopollenin precursors.In this study,the mutation ofGmAMS1affected the synthesis of sporopollenin precursors involved in the formation of the pollen primary outer wall during the tetrad period,affecting the development of the primary outer wall of pollen.The delayed degradation of the tapetum also led to the abortion of pollen grains inGmams1-1lines,as shown by the male-sterile phenotype.This result was similar to the phenotype of male sterility caused by the mutation ofTDRin rice[18,38].Finally,it can be confirmed that targeted editing ofGmAMS1in soybean by CRISPR/Cas9 is an efficient method for generatingAMS-based male-sterile lines.GmAMS1could thus be used to generate male-sterile lines in soybean as in other crops.

5.Conclusions

CRISPR/Cas9-based genome editing technology can be used to generate nuclear male sterility lines in soybean by generating mutants that show complete male sterility in the first generation.Targeted editing ofGmAMS1with CRISPR/Cas9 can produce a male-sterile phenotype in soybeans.Loss of function of theGmAMS1gene in soybean not only caused abnormal morphology of the microspore outer wall,but also delayed the degradation of tapetum cells,eventually leading to pollen abortion and resulting in a completely male-sterile adult-plant phenotype.GmAMS1thus functions in regulating the development of the soybean tapetum.CRISPR/Cas9 technology can rapidly produce male-sterile lines and could accelerate the establishment of commercially viable soybean hybridization platforms.

CRediT authorship contribution statement

Xianzhong Feng and Haiyan Liconceived the project and designed the research.Xiao Chen and Suxin Yangprepared figures,and wrote the paper.Xiao Chenperformed most of the experiments.Xiaobin Zhu and Chunbao Zhanghelped in histological analysis.Xinjing Yang and Yaohua Zhangperformed SEM.Xianzhong Fengsupervised the project.

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 Key Research and Development Program of China (2016YFD0101900,2016YFD0100401).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.05.003.