Chun Li,Jingwen Wng,Zhoyong Hu,Yunyn Xi,Qing Hung,To Yu,Hongyng Yi,Ynli Lu,b,Jing Wng,Moju Co,b,*

a Maize Research Institute,Sichuan Agricultural University,Chengdu 611130,Sichuan,China

b State Key Laboratory of Crop Gene Exploration and Utilization in Southwest China,Chengdu 611130,Sichuan,China

Keywords:

A B S T R A C T

1.Introduction

Chloroplast is one of the two extranuclear membrane-bound and semiautonomous organelles in plants.The plant chloroplast is not only the site of photosynthesis,but the main site of metabolite synthesis[1].The formation of normal chloroplasts is important to plant growth.Chloroplast development is a dynamic process regulated by environmental and intracellular signals[2].Functions of chloroplast genes are generally divided into three categories:transcription/translation functions,photosynthesis functions,and synthesis of proteins involved in other metabolic processes[3].Chloroplast transcription machinery is complex and plays important roles in the regulation of chloroplast development[4].Multiple RNA polymerases(RNAP)have been identified in various higher plants and divided into two types:a bacterialtype multi-subunit plastid-encoded polymerase(PEP)and nuclear-encoded polymerase(NEP)[4].Chloroplast-encoded genes are transcribed mainly by either PEP or NEP[3,4].NEP is represented by the single subunit phage-type RNAP encoded by RpoT genes[4,5].In contrast,PEP consists of four core subunits:α,β,β′and β′′,which are encoded by the genesrpoA,rpoB,rpoC1,andrpoC2,respectively[5].In the process of chloroplast development,NEP is first activated for transcription of genes including chloroplast housekeeping genes and genes encoding PEP core subunits and ribosomal proteins,and then replaced by PEP for the transcription of photosynthesis-associated genes[4].

Like bacterial RNA polymerases,nuclear-encoded sigma(σ)factors are responsible for the promoter recognition and transcription initiation of the PEP holoenzyme in higher plants[5].To date,six σ factors including SIG1,SIG2,SIG3,SIG4,SIG5,and SIG6 have been identified in the model dicotyledonous plantArabidopsisand their roles in response to ever-changing developmental and/or environmental conditions have been clarified.SIG1 initiates transcription of a specific subset of chloroplast genes:psaA,psbB,psbE,rbcL,andrpoB,and participates in the photosynthetic control of PSI reaction center gene transcription[6].Disruption of SIG2 protein causes impaired chloroplast development and results in a pale green leaf phenotype[7].SIG3 initiatespsbNandatptranscription[8,9].Plastid NDH activity is regulated at the transcriptional level by anndhF-specific plastid sigma factor,SIG4[10].SIG5 recognizes a blue light-responsive promoter and initiates the transcription ofpsbD,which encodes the D2 protein of the PSII reaction center[11].In addition,SIG5 controls circadian rhythm in chloroplast gene expression at the transcription level[12].The null mutant of SIG6 exhibited a cotyledon-specific pale green phenotype as a result of delayed light-dependent chloroplast development[13].In the model monocotyledonous plant rice,six σ factors including OsSig1(OsSigA),OsSig2A,OsSig2B,OsSig3(OsSigC),OsSig5,and OsSig6 have been isolated or predicted[14–17].OsSig1 maintains PSI activity by regulating the expression of thepsaAgene in rice chloroplasts and its deficient mutant displays reduced chlorophyll content and lighter green leaf[16].OsSIG2A is required for early chloroplast differentiation under low temperatures by regulation of plastid gene expression[14].

In maize,five putative σ factors:ZmSig1A,ZmSig1B,ZmSig2A,ZmSig2B,and ZmSig6,have been identified by database searches[18–20].Transcription ofZmSig1AandZmSig1Bis light-inducible and their transcripts were abundant in greening leaves.The expression of the ZmSig1A protein can be detected in both chloroplasts and etioplasts,but transcripts ofZmSig1Bare barely detectable in etiolated leaves and neither are detectable in roots[19].ZmSig2Ais expressed specifically in seedling green leaves and accumulates in mature chloroplasts,behavior consistent with the light-activated chloroplast development program[21].Transcripts ofZmSig6are accumulated in non-green tissues,including roots,etiolated seedling leaves,and the basal region of greening seedling leaves[21].Unlike the other four σ-like factors,ZmSig2B is located and appears to function in both chloroplasts and mitochondria[20].Although the expression patterns of the five maize putative σ factors have been investigated,their functions are still unknown.Identifying their roles in organelle development will require identifying deficient mutants.

In the present study,a maize revertible leaf-color mutant namedetiolated/albino leaf 1(eal1)was identified in progeny of spaceflight-treated seeds.During the early development stage of seedlings,eal1displays etiolated or albino leaves,reduced photosynthetic pigment content,and retarded chloroplast development and plant growth,but the leaf color gradually returns to green at later stages.We cloned theeal1gene using a map-based approach and demonstrated that the disruption of plastid-localized σ factor ZmSig2A causes the phenotype ofeal1.This study is the first to establish the important role of σ factor in maize chloroplast development using forward genetics.

2.Materials and methods

2.1.Plant materials and growth conditions

Seeds of the maize hybrid Chuandan No.9 were treated by spaceflight in 1996.After eight generations of selfing,seedlings with etiolated or albino leaves were obtained from one of the recombinant inbred lines.These plants were self-pollinated and all of their progeny displayed the mutant phenotype at seedling stage,whether planted in Sichuan or Yunnan.Here we refer to this mutant aseal1and its wild-type sibling,whose progeny displayed normal green leaf color,as WT.To create near-isogenic lines carryingZmSig2AΔVallele in B73 inbred line genetic background,BC5F2(-eal1)plants were generated by repeated backcrosses ofeal1with B73 for five generations.EMS4-110eeb seeds derived by ethylmethanesulfonate mutagenesis of B73 were obtained from the seed stock of the maize ethylmethanesulfonate-induced mutant database(MEMD)(http://www.elabcaas.cn/memd/).TheArabidopsisecotypes Columbia(Col-0)and σ factor SIG2 mutantsig2(SALK_045706C)in the Col-0 background were used in this study.

For phenotypic observation and photography,seeds ofeal1,WT,EMS-6S1,and BC5F2(eal1)were sown in flowerpots,after which theeal1,WT and EMS-6S1 seedlings were transplanted to the field at seven days after sowing(DAS).For measurement of agronomic traits,the seeds ofeal1,WT and F1hybrids were sown directly into the field.Sowing-anthesis interval and plant height were recorded at the anthesis stage.Ears used for the assessment of yield traits including kernel number per ear,ear diameter,ear length,ear row number,and ear grain weight were harvested from openpollinated plants.Plants were cultivated in experimental fields in Chengdu,Sichuan in the spring season and Jinghong,Yunnan in the winter season.For temperature-sensitivity testing of theeal1mutant,seeds of WT andeal1were sown separately in nursery trays containing a mixture of soil and vermicultite.The growth conditions were 16 h light and 8 h dark at 20 °C or 28 °C and 70% relative humidity.

Arabidopsisseeds were surface-sterilized and then treated at 4 °C in the dark for three days.Seeds were sown on half-strength Murashige and Skoog(MS)plates containing 0.7%(w/v)agar and 2%(w/v)sucrose for 10 days.The seedlings were then transplanted and grown at a density of four plants per pot containing a mixture of soil and vermiculite.For phenotype observation and photosynthetic pigment extraction,seeds were sown in nursery trays containing a mixture of soil and vermiculite.The growth conditions were 16 h light and 8 h dark with a light intensity of 100 μmol photons m-2s-1at 22–24 °C and 70% relative humidity.

2.2.Determination of photosynthetic pigment content

Photosynthetic pigments were extracted from leaves ofeal1,WT,Col-0,sig2,and transgenic lines with 80% acetone.The absorbance of the extract was measured with a BIOMATE 160 UV–Vis Spectrophotometer(Thermo Scientific,Waltham,Massachusetts,USA)at 663,646,and,470 nm.The concentrations of chlorophylla(Chla),chlorophyllb(Chlb),and carotenoid(Car)were calculated following Shi et al.[22].

2.3.Transmission electron microscopy(TEM)assay

Sampling for TEM observations of leaves ofeal1and WT plants was performed in parallel with the pigment content analysis.Fresh leaves were rapidly cut into 5×5 mm2pieces and immediately prefixed in precooled 3% glutaraldehyde solution containing 0.1 mol L-1potassium phosphate(pH 7.2)at 4 °C in the dark overnight,rinsed with phosphate buffer(pH 7.2)three times,and post-fixed in 1% osmium tetroxide.The samples were then dehydrated in an ethanol series and embedded in Epon resin 812 prior to thin sectioning.The sections were stained with methylene blue and ultrathin sections were prepared by cutting with a diamond knife,stained with uranyl acetate and lead citrate,and examined by TEM(HITACHI,H-600IV,Chiyoda-ku,Tokyo,Japan)[23].

2.4.Genetic analysis and fine mapping of the eal1 locus

F2populations [B73×eal1]F2, [B104×eal1]F2and[eal1×Huangzaosi]F2,and BC1populations[B73×eal1]×eal1,[B104×eal1]×eal1and[eal1×Huangzaosi]×eal1were used for genetic analysis ofeal1.Seeds of these populations were sown in nursery trays and seedlings were cultivated in a greenhouse.Phenotyping was conducted at the 10 DAS stage.For initial mapping,leaf DNA of 10 mutant and 10 wild-type plants derived from[B73×eal1]F2were pooled separately for bulked-segregant analysis(BSA).Simple sequence repeats(SSR)markers derived from the maize genome database(http://www.maizegdb.org/)were used for screening for codominant polymorphic markers.Insertion and deletion(InDel)and Single nucleotide polymorphism(SNP)markers were designed for fine mapping.Primers used in this experiment are listed in Table S1.The genotypes of individual plants were confirmed by either 6% SDS-PAGE,4% agarose gel electrophoresis,or PCR-based sequencing.

2.5.Candidate-gene analysis

The data source Gramene(http://ensembl.gramene.org/Zea_-mays/Info/Index)was used to predict protein-coding genes in the fine-mapped region.cDNA sequences containing open reading frames(ORFs)of candidate genes were amplified with specific primers(Table S1).The PCR products were sequenced and aligned.For conservedness analysis of the deleted nucleotides ofZmSig2Aineal1,the DNA sequences ofZmSig2Afrom 24 inbreds including 18-599,48-2,87-1,478,698-3,B73,B104,Chang7-2,Fengke1,Huangzaosi,Hui313,LH8012,ZH02,ZH04,ZH91,Zi330,Zifeng1,ZJ102,ZJ202,ZJ302,ZJ401,ZJ402,Mo17,and ZNC4-4-2 were amplified.WoLF PSORT[24]and Plant-PLoc[25]were used for protein subcellular location prediction.Protein structure predictions were run on the Phyre2 server(http://www.sbg.bio.ic.ac.uk/phyre2)[26].Sequence alignment and phylogenetic analysis of the amino acid sequences of σ factors were performed using MUSCLE[27]and MEGA X[28]with default parameters.An unrooted phylogenetic tree was built by the neighbor-joining method with the following parameters:pairwise deletion option,1000 bootstrap replicates,and Poisson model.

2.6.Subcellular localization

ZmSig2A+(the wild-type allele ofZmSig2Ain WT plants)was amplified from first-leaf cDNA of WT plants at 7 DAS with primer pair Sig2A-F/Sig2A-R.The PCR products were verified by sequencing and used as template for the amplification ofZmSig2A+ORF with primer pair Sub2A-F/Sub2A-R.The PCR product was purified by agarose gel DNA extraction kit(Tiangen,Beijing,China)and cloned into the pEASY-Blunt cloning vector(Transgen,Beijing,China).The modified binary vectorpCAMBIA2300was used for the construction of the subcellular localization cassetteP35S::ZmSig2A+::eGFP.TheZmSig2A+ORF fragments were cloned into thepCAMBIA2300expression vector by restriction enzyme digestion and ligation.P35S::ZmSig2A+::eGFPrecombinant plasmids were transformed intoAgrobacterium tumefaciensstrain EHA105 competent cell.KOD DNA polymerase(Toyobo,Osaka,Japan)was used for PCR amplifications.Primers are listed in Table S1.The fusion proteins were expressed inN.benthamianaleaf by agroinfiltration[29].Fluorescence was examined under a confocal microscope LSM 800(Zeiss,Oberkochen,Germany)after 48 h of agroinfiltration.

2.7.Genetic complementation tests

The ORFs ofZmSig2A+andZmSig2AΔV(the mutant allele ofZmSig2Aineal1plants)were amplified with primer pairs 35S-2AF/R and Ubi-2A-F/R from WT andeal1plants.The promoters ofArabidopsis SIG2and maizeubiquitin1genes were amplified with the primer pairs PSIG2-F/PSIG2-R and Pubi-F/Pubi-R.The PCR products were cloned into the pEASY-Blunt cloning vector and sequenced.The modified binary vectorspCAMBIA1391andpCAMBIA1300were used to generate the cassettesPSIG2::ZmSig2AΔV,P35S::ZmSig2A+,andP35S::ZmSig2AΔVby restriction enzyme digestion and ligation.

TheP35S::ZmSig2A+andP35S::ZmSig2AΔVrecombinant plasmids were transformed into theArabidopsis sig2mutant by the floraldip method usingAgrobacterium tumefaciensstrain GV3101competent cells[30].Transformed seeds were selected on half-strength MS medium containing 30 mg L-1hygromycin.Hygromycinresistant seedlings were transplanted into pots containing a mixture of soil and vermiculite and homozygous lines were generated by selfing.All transgenic lines were verified by PCR amplification at the DNA and RNA levels.

For genetic complementation test witheal1,the modified binary vectorpCAMBIA3301was used to generate the cassettePubi::ZmSig2A+.ThePubi::ZmSig2A+recombinant plasmids were introduced into inbred B104 calli by infection withA.tumefaciensstrain EHA105 competent cells.Maize transformation was performed as described[31].T0plantlets were retested for the presence of the transformation construct by genomic PCR with the primer pair Ubi2A-F/Ubi2A-R.T1seeds were generated by the selfing of PCRpositive T0plants.Crosses between PCR-positive T0plants andeal1plants were performed to generate[B104T×eal1]F1ears.Plantlets of[B104T×eal1]F1lines were genotyped by genomic PCR with the Ubi2A-F/Ubi2A-R primer pair.Positive and negative transgenic[B104T×eal1]F1plants were self-pollinated to generate their respective F2populations.Here we use[B104T×eal1]F2Pand[B104T×eal1]F2Nto denote the respective progenies of positive and negative[B104T×eal1]F1plants.Transgenic complementation tests were finally completed by a combination of genotyping and phenotyping of F2plants.EMS4-110eeb was used for an allelism test.ZmSig2ATis a nonsense allele ofZmSig2Aidentified in EMS4-110eeb seedlings.A heterozygote EMS-6(ZmSig2AC/T)was crossed with aneal1plant.KOD DNA polymerase was used for PCR amplifications.Primers are listed in Table S1.

2.8.Total RNA isolation and quantitative PCR

Leaves ofeal1,WT,EMS4-110eeb,sig2,Col-0,and transgenic lines were sampled at designated stages,quickly frozen in liquid nitrogen,and stored at-80 °C.Total RNA was extracted with RNAiso Plus reagent(TAKARA,Kusatsu,Japan)according to the product description.RNA quality and concentration were determined using 1.5% agarose gel electrophoresis and NanoDrop 2000(Thermo Scientific,Waltham,Massachusetts,USA),respectively.cDNA was synthesized using PrimeScript 1st Strand cDNA Synthesis Kit(TAKARA,Kusatsu,Japan)following the manufacturer’s protocol.The primer pairs ZmACT1-F/R and AtACT2-F/R(Table S1),which span introns of the housekeeping genesActin1of maize andACTIN2ofArabidopsisrespectively,were used for assessment of cDNA quality.Quantitative real-time PCR(qRTPCR)was conducted using a CFX9 Real Time system(Bio-Rad,California,USA)and performed with TB Green Premix Ex Taq(Tli RnaseH Plus)(TAKARA,Kusatsu,Japan)following its operating manual.Actin1andACTIN2were used as internal controls.Primers are listed in Table S1.Relative gene expression levels were calculated by the 2-ΔΔCtmethod[32].

2.9.Accession numbers

Sequence data from this study can be searched by the following gene IDs:SIG2(837376),ubiquitin1(103626648),Actin1(100282267),ACTIN2(821411),psaA(845195),psaC(1466379),psaI(845197),psaJ(845198),psbA(845199),psbE(845203),psbF(845204),psbJ(845207),psbL(845209),psbM(845210),atpA(845169),atpB(845170),rbcL(845212),rbcS1(542212),rbcS2(100279574),LHCB1(100282054),LHCB3(542530),LHCB4(100281795),LHCB9(103643653),cah1(100275493),cah3(100274597),cah6(100382491),ploc1(103629356),andploc2(100281113).

3.Results

3.1.Phenotypic characterization of the eal1 mutant

Theeal1seedlings displayed etiolated or albino leaves at early seedling stages,after which the leaf color gradually recovered and all leaves turned green about 25 DAS(Fig.1A).Although the growth ofeal1plants was relatively delayed(Figs.1A,S1A),no marked difference in plant height between WT andeal1plants was detected(Fig.S1B,C).The sowing–anthesis interval ofeal1plants was 68.8±2.8 days,significantly longer than that of WT plants,63.3±1.7 days(Fig.S1C).Although the ear diameter,kernel number,and kernel weight per ear ofeal1plants were lower than those of WT plants(Fig.S1C),theeal1mutation displayed no negative effects on the yield traits of tested F1hybrids(Fig.S1D).

The contents of Chla,Chlb,and Car extracted from second leaves were markedly lower ineal1than in WT plants,at only 7.3%,6.1%,and 13.6% of WT contents at 7 DAS(Fig.1B;Table S2).Although the pigment contents ineal1plants were still lowers at 10 DAS,they rose to 50.0%,28.6%,and 80.0% of WT plant(Fig.1B;Table S2).The pigment contents ineal1plants further increased to 83.6%,55.3% and 81.8% of the WT contents at 13 DAS(Fig.1B;Table S2).After regreening,all the pigment contents were nearly the same ineal1as in WT plants,with their relative contents reaching 97.8%,94.4%,and 94.6% at 25 DAS(Fig.1B;Table S2).

Under TEM,grana stacks of chloroplasts in WT plants were dense and well-developed at 7,10,13,and 25 DAS(Fig.1C).Ineal1plants,there were only retarded chloroplasts with severely abnormal internal structure at 7 DAS;rudiments of chloroplast with less dense and fuzzy grana stacks were observed at 10 DAS,and more distinct internal structures of the chloroplasts were formed and the thylakoid membranes were markedly increased at 13 DAS.After regreening,the chloroplast ofeal1plants displayed normal ultrastructure,containing normal distributed and well-developed thylakoid grana stacks at 25 DAS(Fig.1C).These results suggested that abnormal chloroplast development accounted for the pigment contents and leaf color variation ofeal1plants.

Some revertible leaf color mutants are temperature-sensitive[33–35].As shown in Fig.S2,the leaf color ofeal1seedlings grown at 28 °C turned green faster than that of seedlings grown at 20 °C.Leaf photosynthetic pigment contents showed no significant differences between 20 °C and 28 °C in WT plants(Fig.S2E,F).Ineal1plants,the pigment contents of seedlings grown at 28°C were significantly higher than those grown at 20 °C(Fig.S2E,F).Thus,the leaf color reversion ofeal1was temperature-sensitive and the color-recovery process was accelerated at higher temperature.

3.2.Genetic analysis and fine mapping of eal1

The F1plants of crosses betweeneal1plants and the inbred lines B73,B104,and Huangzaosi displayed normal green leaves at the seedling stage.The segregation ratios of green to etiolated/albino plants were 1:1 for each BC1population and 3:1 for each F2population(Table S3),suggesting that the mutant phenotype ofeal1is controlled by a single recessive nuclear gene.Theeal1locus was initially mapped to the short arm of chromosome 4 between markers InDel-2 and InDel-19 with 96 etiolated/albino plants of the[B73×eal1]F2population(Fig.2A).Another 1116 etiolated/albino plants were used for fine mapping.The interval was narrowed to the region between InDel-12 and ssr2176-1;one and eight recombinants were detected by these two markers,respectively(Fig.2A).Two SNP markers,160SNP and 161SNP,were developed based on genomic PCR and sequencing and used to genotype the nine recombinants.Theeal1locus was finally localized to a 174.9-kb interval containing three protein-coding genes:Zm00001d049158,Zm00001d049159,andZm00001d049160(ZmSig2A)annotated in the B73 reference genome(RefGen_v4)(Fig.2A).

3.3.Identification of the key candidate gene of eal1

RT-PCR and sequence alignment reveald no difference in the amino acid sequences of Zm00001d049158 and Zm00001d049159 betweeneal1and WT plants.ForZmSig2A,seven same-sense single-base substitutions and a 3-bp deletion were identified ineal1plants compared with WT plants(Fig.S3).The 3-bp deletion changed Val480-Val481-Val482into Val480-Val481(Fig.2B).PCR amplification and sequencing ofZmSig2Afrom another 24 maize inbreds revealed that the 3-bp deletion was specific to theeal1mutant(Fig.S4).ZmSig2Aencoded a σ-like factor,which harbors the highly conserved σ-factor domains σ2,σ3,and σ4[21,36].The domain σ4structural core,which consisted of compactly folded α helices,contains a binding β-subunit flap domain of core polymerase and a-35 element of the promoter[36,37].Protein modeling showed that domain σ4of ZmSig2A contained four α helices and the 3-bp deletion leading to clear structural variation within the second α helix(Fig.2C).Here,ZmSig2A+andZmSig2AΔVrepresent the alleles of theZmSig2Alocus in WT andeal1plants,respectively.

Subcellular localization showed that the green fluorescence signals of ZmSig2A+-GFP were detected only in the chloroplasts,suggesting ZmSig2A was a nuclear-encoded and chloroplast-located protein(Fig.2D).In the phylogenetic tree of 20 σ or σ-like factors from higher plants,the five identified maize σ-like factors were clustered into three groups,indicating some differentiation in molecular function among them(Fig.S5A).ZmSig2A shared the highest similarity with rice OsSig2A(77%)andArabidopsisSIG2(45%)(Fig.S5A,B).OsSIG2Ais required for chloroplast development in rice at low temperature by regulating plastid gene expression[14].Disruption ofArabidopsisSIG2 protein led to impaired chloroplast development and pale green leaves[7].ZmSig2A+was transformed into theArabidopsis sig2mutant,and six single-copy homozygous transgenic lines were selected randomly for further study.All six transgenic lines displayed normal green leaves like Col-0(Fig.3A).Three of the six transgenic lines were randomly selected for measurement of photosynthetic pigment content.As shown in Fig.3B,the Chla,Chlb,and Car contents ofsig2were significantly lower than those of Col-0,whereas the contents of the transgenic lines were restored to the levels of Col-0.The transcript ofZmSig2A+could be detected only in the six transgenic lines and not in Col-0 andsig2(Fig.3C).The finding that ectopic expression ofZmSig2A+complemented the deficiency of SIG2 and rescued the pale green leaf phenotype ofsig2suggested thatZmSig2Awas very likely the gene responsible foreal1.

3.4.Function complementation of ZmSig2A for eal1 mutation

Fig.1.Phenotype characterization of the eal1 mutant.(A)WT and eal1 plants at four development stages.WT and eal1 were planted in Jinghong,Yunnan in late September 2018.(B)Pigment contents of WT and eal1 plants.Chl a,chlorophyll a;Chl b,chlorophyll b;Car,carotenoids.Values are mean±SD from three measurements.Student’s t-test was used to determine the significance of differences,**and*indicate statistical significance at the 0.01 and 0.05 probability levels.NS denotes no significant difference.(C)The second-leaf chloroplast ultrastructure of WT and eal1 from TEM observation.DAS,days after sowing.

Fig.2.Map-based cloning of the eal1 locus and candidate-gene analysis.(A)Map-based cloning of the eal1 locus.The locus was mapped to an interval between markers InDel-2 and InDel-19 on the short arm of chromosome 4(Chr.4)using 96 recessive plants.The map was constructed based on the sequence of Chr.4 from the B73 reference genome(RefGen_v4).SSR markers umc1288,umc2281 and umc2176 were obtained from MaizeGDB;InDel markers InDel-2,InDel-12,InDel-19 and SNP markers 160SNP,161SNP were developed based on genomic PCR amplification and sequence alignment.Numbers indicate recombinants detected by the corresponding markers.Recessive plants(N=96 and N=1116)of the[B73×eal1]F2 population were used for mapping.(B)Amino acid sequence alignment of ZmSig2A between WT and eal1 plants.The mutation site is indicated by a black triangle.(C)Protein structure prediction of ZmSig2A for WT and eal1.The mutant site is indicated by a black triangle.(D)Subcellular localization of ZmSig2A+protein.The ZmSig2A+-GFP fusion protein and GFP protein were expressed separately in N.benthamiana.Fluorescence was observed by fluorescence microscopy 48 h after transformation.Green fluorescence represents GFP and red fluorescence represents autofluorescence of chlorophyll.

Fig.3.Overexpression of ZmSig2A in Arabidopsis sig2 mutant.(A)Leaf color of Col-0,sig2,and six transgenic lines at 10 days after sowing.(B)Pigment contents of Col-0,sig2,and three transgenic lines at 10 days after sowing,with 0.06 g of Arabidopsis seedling leaves was used for pigment extraction.62-1-1,52-6-3,82-8-2,39-2-3.44-5-4,and 48-4-1 represent the P35S::ZmSig2A+transgenic lines;values are means±SD from three measurements;Student’s t-test was used to determine the significance of differences,**and*indicates statistical significance at the 0.01 and 0.05 probability levels,respectively;NS refers to no significant difference;Scale bars,1 cm.(C)The expression of ZmSig2A+was detected by qRT-PCR in transgenic lines,with Col-0 and sig2 plants as negative controls.

ThePubi::ZmSig2A+overexpression cassette was constructed and introduced into maize inbred line B104.Thirty of the herbicideresistant T0transgenic plants labeled with OE1-OE30 were identified by PCR as positive transgenic lines and named B104T(Fig.S6A).Twenty B104T plants were crossed with theeal1mutant,and positive(P)and negative(N)transgenic plants derived from the[B104T×eal1]F1ears were identified by genomic PCR and named[B104T×eal1]F1Pand[B104T×eal1]F1N,respectively.These plants were self-pollinated to generate the[B104T×eal1]F2Pand[B104T×eal1]F2Npopulations.The proportions of plantlets with the mutant phenotype in the[B104T×eal1]F2Ppopulations were sharply reduced in comparison with those in[B104T×eal1]F2Npopulations(Fig.4A).The ratios of green plants to etiolated/albino plants in[B104T×eal1]F2Npopulations were consistent with a 3:1 segregation like that in the[B104×eal1]F2population(Table S4).But the ratios in the[B104T×eal1]F2Ppopulations deviated widely from 3:1,indicating that the exogenousZmSig2A+disturbed these segregations(Table S4).

Sequencing ofZmSig2Aalleles in B104 andeal1plants revealed a SNP(T＞A),which was developed into a cleaved amplified polymorphic sequence marker BeCaps1 for use in identifyingZmSig2A+/+,ZmSig2A+/ΔV,andZmSig2AΔV/ΔVgenotypes(Fig.S6B).All the plants from the four[B104T×eal1]F2Ppopulations were phenotyped and genotyped(Fig.S6C).There were 429 green plants and 14 etiolated/albino plants(Table 1).All 14 etiolated/albino individuals were genotyped asZmSig2AΔV/ΔVby BeCaps1 and identified as negative transgenic plants using Ubi2A primers(Table 1).Among the 429 green plants,94 were identified not only asZmSig2AΔV/ΔVat theZmSig2Alocus,but as PCR-based positive transgenic plants(Table 1).RT-PCR and sequencing of three of the 94 revealed the two transcripts ofZmSig2A+andZmSig2AΔV(Fig.S7).These results suggested thatPubi::ZmSig2A+was successfully expressed inZmSig2AΔV/ΔVplants and that the expression of theZmSig2A+allele rescued the mutant phenotype ofZmSig2AΔV/ΔVplants.

EMS4-110eeb seeds carrying a nonsense mutation(1174C＞T)inZmSig2Acoding region,which resulting in a premature stop codon(Fig.4B).Protein structure prediction showed that the nonsense mutation caused the loss of conserved domains σ3and σ4(Fig.4C).Ten heterozygotes(ZmSig2AC/T)and 24 homozygotes(ZmSig2AC/C)were identified in seedlings from 34 EMS4-110eeb seeds(Fig.4D).All 34 seedlings displayed normal green leaves.Albino seedlings appeared in the progenies of self-pollinatedZmSig2AC/Tplants(Fig.4E).The segregation ratios of green plants to albino plants were 3:1 in each self-pollinated population,suggesting that the albino phenotype is controlled by a single recessive nuclear gene(Table S5).For an allelism test,EMS-6,one of the tenZmSig2AC/Tplants,was crossed witheal1(ZmSig2AΔV/ΔV)plants to generate[eal1×EMS-6]F1seeds.Among the 192[eal1×EMS-6]F1seedlings,91 were etiolated/albino and 101 were normal green plants(Fig.4F),fitting a ratio of 1:1(χ2=0.51＜χ2(0.05,1)=3.84).This result indicated that the mutations in EMS-6 andeal1were allelic.The mutation in EMS-6 affected leaf development more strongly than that ineal1plants(Figs.1A,4E,G),and theZmSig2ATmutant allele was lethal in homozygous plants(Fig.4H).Collectively,the deletion of Val in the domain σ4of ZmSig2A was responsible for theeal1mutant phenotype.

Table 1Genotypes of[B104T×eal1]F2P populations.

3.5.Expression and functional analysis of the mutated ZmSig2A allele

First leaves of WT andeal1seedlings were used for expression analysis ofZmSig2Aat different development stages.The expression ofZmSig2A+gradually decreased along with the growth of WT plants,suggesting that a higher mRNA level ofZmSig2A+was necessary at the initial stage of leaf development(Fig.5A).ForZmSig2AΔV,similar trends were detected ineal1plants;but the expression level ofZmSig2AΔVineal1plants was significantly higher than that ofZmSig2A+in WT plants at the corresponding development stages(Fig.5A).The differing expression patterns inZmSig2A+andZmSig2AΔVin second leaves at 7 DAS resembled those in first leaves(Fig.5B).In sharp contrast to theeal1mutant,the expression levels ofZmSig2ATinZmSig2AT/Tplants were significantly lower than those ofZmSig2ACinZmSig2AC/Cplants at 7 DAS not only in first but in second leaves(Fig.5C).

Fig.4.Genetic complementation test for eal1 mutant and phenotypic comparison.(A)The phenotypes of[B104T×eal1]F2N and[B104T×eal1]F2P populations.B104T represents B104 positive transgenic plants.Positive(P)and negative(N)transgenic plants derived from[B104T×eal1]F1 ears are represented as[B104T×eal1]F1P and[B104T×eal1]F1N,respectively.These plants were self-pollinated to generate[B104T×eal1]F2P and[B104T×eal1]F2N populations.(B)Scheme of a nonsense mutation(1174C＞T)in ZmSig2A coding region that changes an Arg codon into a stop codon.(C)Protein structure prediction of ZmSig2AC and ZmSig2AT alleles.(D)Genotyping for EMS4-110eeb seedlings.Ten heterozygotes(ZmSig2AC/T)(EMS-1,EMS-2,EMS-6,EMS-8,EMS-11,EMS-12,EMS-16,EMS-19 EMS-21,EMS-26)and 24 homozygotes(ZmSig2AC/C)were identified in 34 EMS-110eeb seeds.(E)Seedling phenotype of EMS-6S1 at 7 days after sowing.EMS-6S1 represents the progenies of self-pollinated EMS-6(ZmSig2AC/T)plants.(F)Seedling phenotype of[eal1×EMS-6]F1 plants at 10 days after sowing.(G)Phenotypic comparison between homozygote(ZmSig2AT/T)of EMS-6S1 and homozygote(ZmSig2AΔV/ΔV)ofBC5F2(eal1).To create near-isogenic lines carrying ZmSig2AΔV allele in B73 inbred genetic background,BC5F2(eal1)population was developed by repeated backcrosses of eal1 with B73 for five generations.1 and 2 represent the first and second leaves of the homozygote(ZmSig2AC/C)of EMS-6S1;1′and 2′represent the first and second leaves of the homozygote(ZmSig2AΔV/ΔV)of BC5F2(eal1).(H)Phenotypic comparison between homozygote(ZmSig2AT/T)and heterozygote(ZmSig2AC/T)of EMS-6S1 in the field in Jinghong,Yunnan in September 2020.Scale bars,5 cm.

To determine whether the Val deletion within the domain σ4abolished the function of ZmSig2A, two cassettes driven by the CaMV35S andSIG2promoters were generated to transiently expressZmSig2AΔVin theArabidopsissig2mutant. Seventeen and 11 single-copy homozygous transgenic lines were obtained for the cassettesP35s::ZmSig2AΔVandPSIG2::ZmSig2AΔV, respectively. The leaf color ofP35s::ZmSig2AΔVtransgenic lines was similar to that of Col-0, but thePSIG2::ZmSig2AΔVtransgenic lines still dis-played pale green leaves (Fig. 5D). Three lines of each cassette were randomly selected for pigment content measurements and qRT-PCR analyses.The Chla, Chlb, and Car contents inP35s::ZmSig2AΔVtransgenic plants were all consistent with those in Col-0 plants and significantly higher than those insig2plants, but the pigment contents inPSIG2::ZmSig2AΔVtransgenic plants were the same as those insig2plants (Fig. 5E). The relative expression levels ofZmSig2AΔVdriven by the CaMV35S promoter in the threeP35S::ZmSig2AΔVtransgenic lines were 1.39 ± 0.01, 1.40 ± 0.03, and 2.81 ± 0.30,but the values were only 0.19 ± 0.01, 0.18 ± 0.01, and 0.22 ± 0.01 under theSIG2promoter in threePSIG2::ZmSig2AΔVtransgenic lines(Fig. 5F), suggesting a reason for the phenotypic differences betweenP35S::ZmSig2AΔVandPSIG2::ZmSig2AΔVtransgenic lines.These findings suggested that the Val deletion within the domain σ4does not completely abolish the function of ZmSig2A and that the increased expression ofZmSig2AΔVineal1plants partly compensates for the loss of function.

3.6.Expression of sigma factors and genes associated with chloroplast development and photosynthesis

cDNAs derived from second leaves of WT andeal1seedlings at 7 DAS were used for qRT-PCR analysis.For the expressions of the other four putative σ factorsZmSig1A,ZmSig1B,ZmSig2B,andZmSig6[18–20],no marked difference was detected betweeneal1and WT plants(Fig.6A),indicating that the Val deletion ofZmSig2Ahad no effect on the expression ofZmSig1A,ZmSig1B,ZmSig2B,andZmSig6.Becauseeal1mutant plants displayed abnormal chloroplast development at the early seedling stage,the expressions of chloroplast development-associated genes were quantified.The expressions ofpsaCandpsaI(encoding PSI reaction-center polypeptides),andpsbA,psbE,psbF,psbJ,psbLandpsbM(encoding PSII reaction-center polypeptides),atpAandatpB(encoding respectively the α and β subunits of chloroplast ATP synthase),andrbcS1(encoding the small subunit of Rubisco)ineal1plants were significantly lower than those in WT plants(Fig.6B,C and D).But the expressions ofpsaAandpsaJ(encoding PSI reaction-center polypeptides)were significantly increased ineal1compared with WT plants(Fig.6B).Thus,the expressions of genes associated with chloroplast development were affected in theeal1mutant at the seedling stage.The expressions of the 10 reaction-center polypeptide genes that were differentially expressed between WT andeal1plants were also quantified in second leaves ofZmSig2AC/CandZmSig2AT/Tplants at 7 DAS.As shown in Fig.6E,all 10 were severely suppressed inZmSig2AT/Tplants.

As a deletion in a nuclear-encoded σ factor,the Val deletion in ZmSig2A resulted in increased expression of the nuclear alleleZmSig2AΔV(Fig.5A,B).The transcriptions of photosynthesisassociated nuclear genes(PhANGs)were also analyzed.In comparison with WT plants,the expression levels ofLHCB1andLHCB9were markedly reduced,with no transcript ofLHCB1as detected ineal1plants.In contrast,the expression levels ofLHCB3,LHCB4,

carbonic anhydrase1(cah1),cah3,cah6,plastocyanin homolog1

(ploc1),andploc2were significantly increased ineal1plants(Fig.6F).

Fig. 6. Expression analysis of maize sigma factors, chloroplast development-associated genes, and photosynthesis-associated nuclear genes (PhANGs) by qRT-PCR. (A)Expression of the putative sigma factors ZmSig1A, ZmSig1B, ZmSig2B, and ZmSig6 in WT and eal1 plants. Expression of genes encoding PSI and PSII reaction-center polypeptides(B), genes encoding the α and β subunits of chloroplast ATP synthase (C), and genes encoding large and small subunits of Rubisco (D) in WT and eal1 plants. (E) Expression of genes encoding the PSI and PSII reaction-center polypeptides in ZmSig2AC/C and ZmSig2AT/T plants of EMS-6S1. (F) Expression of PhANGs in WT and eal1 plants. cDNAs derived from second leaves of WT and eal1 plants at 7 days after sowing were used for A, B, C, D, and F; cDNAs derived from second leaves of ZmSig2AC/C and ZmSig2AT/T plants at 7 days after sowing were used for E. Values are derived from three biological replicates and presented as means ± SEM. Student’s t-test was used to determine the significance of difference, ** and * indicates statistical significance at the 0.01 and 0.05 probability levels, respectively; NS denotes no significant difference.

4.Discussion

4.1.ZmSig2A encodes a σ factor essential for chloroplast biogenesis at the early seedling stage of maize

Fig.5.Expression and functional analysis of ZmSig2A alleles.Relative expression levels of ZmSig2A+and ZmSig2AΔV in first(A)and second(B)leaves of WT and eal1 plants.(C)Relative expression levels of ZmSig2AC and ZmSig2AT in first(L1)and second(L2)leaves of homozygous ZmSig2AC/C and ZmSig2AT/T plants at 7 days after sowing.For(A)to(C),values are derived from three biological replicates and presented as means±SEM;Student’s t-test was used to determine the significance of differences,**and*indicates statistical significance at the 0.01 and 0.05 probability levels,respectively.Leaf color phenotype(D)and pigment contents(E)of Arabidopsis seedlings.Two cassettes,P35S::ZmSig2AΔV and PSIG2::ZmSig2AΔV were generated for transient expression of ZmSig2AΔV in sig2.Col-0 is the wild type of sig2.Lines labeled‘‘P35-1,P35-2,P35-3”and‘‘PSIG2-1,P SIG2-2,P SIG2-3”represent single-copy homozygous transgenic lines of P35S::ZmSig2AΔV and PSIG2::ZmSig2AΔV,respectively.Arabidopsis seedling leaf samples of 0.06 g were used for pigment extraction.Values are means±SD from three measurements;Student’s t-test was used to determine the significance of difference,**and*indicate statistical significance at the 0.01 and 0.05 probability levels,respectively;NS denotes no significant difference;Scale bars,1 cm.(F)Relative expression levels of ZmSig2AΔV in Col-0,sig2,and transgenic lines.Values are derived from triplicate reactions of a representative experiment and presented as means±SEM.

Plastids of higher plants are derived from bacterial endosymbionts and accordingly contain σ factors[38].Five putative σ factors in maize have been identified by database searches[18,19],systematic study of their functions in plant development has not been performed.In the present study,a single Val residue deletion in domain σ4of ZmSig2A led to etiolated or albino leaves and retarded development of chloroplasts at the early development stages,with gradual reversion to a wild-type phenotype during further growth ofeal1seedlings.A nonsense mutation in domain σ3of ZmSig2A generated a lethal mutant with completely albino leaves and much slower regreening at the very early seedling stage.The expressions of PEP-dependent chloroplast genes including

psaA,psaC,psaI,psaJ,psaL,psbA,psbE,psbF,psbJ,andpsbMwere strongly affected in these two mutants.T-DNA insertion in theArabidopsisσ factorSIG2led to lower transcript levels of PEPdependent genes and aberrant chloroplast development[39].The mutant phenotype ofsig2was restored by overexpression ofZmSig2A,suggesting that ZmSig2A and SIG2 are not only homologous in sequence but similar in protein function.The difference is that theArabidopsisnonsense mutantsig2survives because of the genetic redundancy ofSIG2andSIG6[39],but the maizeZmSig2ATmutant allele is lethal in the homozygous state.This finding suggests that there is no genetic redundancy among ZmSig2A and the other four σ factors with respect to regulation of chloroplast development in maize seedling leaves,even though ZmSig2A,ZmSig2B and ZmSig6 were clustered into one phylogenetic group withSIG2andSIG6.We conclude that the nuclear geneZmSig2Aencodes a chloroplast-located σ factor and plays a critical role in early chloroplast development.

4.2.The Val deletion in the domain σ4 does not abolish the function of ZmSig2A

Although PEP-dependent transcription is guided by the interactions of domain σ2with the-10 element and of domain σ4with the-35 element,domain σ4is dispensable when the promoters contain an extended-10 element,which can interact with domain σ3to substitute for the interactions between domain σ4and the-35 element[36,37,40,41].In this study,the Val residue deletion in domain σ4of ZmSig2A resulted in theeal1mutant.Mutant plants displayed abnormal chloroplasts at the early development stages of seedlings,and the chloroplasts gradually developed into normal chloroplasts.As with the wild-type alleleZmSig2A+,overexpression of the mutant alleleZmSig2AΔVin theArabidopsis sig2mutant driven by CaMV35S rescued the pale green leaf phenotype.The relatively low expression ofZmSig2AΔVunder theSIG2promoter did not rescue the phenotype of thesig2mutant.This difference may be attributed to the different expression levels under the CaMV35S andSIG2promoters.In comparison with WT plants,the expression ofZmSig2AΔVineal1plants was strongly upregulated.In contrast,in the presence of the nonsense mutation ofZmSig2A,the expression of theZmSig2ATmutant allele was down-regulated inZmSig2AT/Tplants.If domains σ2and σ3are sufficient for the basic σ factor function[37],it is reasonable to infer that the mutated ZmSig2AΔV still has the basic functions for PEPdependent transcription but with reduced efficiency compared with the wild-type ZmSig2A.We speculate that theeal1plants need to express moreZmSig2AΔVto lessen the negative impact of ZmSig2AΔV.We propose that in comparison with the nonsense allele ofZmSig2A,the Val deletion resulted in the weakly functioning alleleZmSig2AΔV,and that some regulatory pathways can detect this mutation and up-regulate the expression of this allele,accounting for the faster chloroplast development and leaf color self-recovery of theeal1mutant.

4.3.Nuclear-derived ZmSig2A protein may play an important role in nucleus–chloroplast communication in chloroplast development

Fig.7.The workflow for using eal1 as a seedling-specific selective marker for increasing seed purity in hybrid seed production.(A)The emasculation system for maize hybrid seed production.The maternal inbred line can be provided with the eal1 mutation via conventional backcross breeding with molecular marker-assisted selection or a genome-editing technology such as the CRISPR/Cas9 system.The resulting maternal line is self-pollinated for seed propagation.When the seeds are grown for hybrid seed production,off-type contaminants arising from hybridization display normal green seedling leaves(as shown by the red arrow).Although removing the green seedlings would ensure the purity of the maternal line,false hybrid seeds would still arise from incomplete emasculation.With the help of the eal1 mutation in the maternal line,false hybrids displaying etiolated and/or albino leaves(as shown by the yellow arrow)can be easily identified.(B)The SPT or MCS system for maize hybrid seed production.The male-sterile line can also be provided with the eal1 mutation as described in(A).For the maintainer line,using the SPT model as an example,there are three linked expression modules:MS(a male fertility gene),P(a pollen-disruption gene)and S(a seed color marker gene)[47,48].Here,a fourth module LC(PUbi::ZmSig2A+)can be introduced into the cassette to produce a new version of the maintainer line(MS:P:S:LC).During the propagation of male-sterile line seeds,transgene transmission by the transgenic maintainer line would occur in consequence of incompletely inactivated transgenic pollen,but the resulting seedlings displaying the normal green phenotype(as shown by the blue arrow)can be easily identified.This system can ensure that hybrid plants grown in farmers’fields are non-GMO(genetically modified organism).

Most chloroplast proteins are encoded in the nucleus.In consequence,chloroplast multiprotein complexes are composed of subunits encoded by the nuclear and chloroplast genomes[42].To regulate the multiprotein complexes in time and meet the metabolic and energy demands,communication between the two genomes is necessary[42].Blocking of chloroplast gene transcription inArabidopsis,barley,wheat,or mustard can down-regulate the transcription of PhANGs[39,43–45].ZmSig2A is a nuclearencoded and chloroplast-located σ factor.The Val deletion in domain σ4led to up-regulation ofZmSig2AΔVexpression and the nonsense mutation in domain σ3led to down-regulation ofZmSig2ATexpression.Further,we found that the expression of some PhANGs were significantly different between WT andeal1plants.It suggests that the Val deletion might stimulate the generation of chloroplast retrograde signals that are then transported into the nucleus and regulate nuclear gene expression.Woodson et al.[39]described σ factor-mediated plastid retrograde signals and the transcriptions of the PhANGs tested were all decreased in thesig2mutant.Despite the similarity of protein function between SIG2 and ZmSig2A,only two of the nine differentially expressed PhANGs,LHCB1andLHCB9,were significantly decreased ineal1plants.The difference may be associated with the upregulation ofZmSig2AΔV.These findings suggest that the nuclearderived ZmSig2A protein plays important roles in the coordination of the two genomes during chloroplast development.As the main organ of photosynthesis,the chloroplast plays a crucial role in improving crop yield potential[46].Given that theeal1mutant shows a transition from albino to normal green during early leaf development,it is an ideal material for studying chloroplast development,as well as the regulatory mechanism of the σ factor

ZmSig2A.

4.4.Potential advantages of using eal1 as a seedling-specific selective marker for increasing seed purity in maize hybrid seed production

Maize is one of the most successful examples of heterosis exploitation,and single-cross hybrids are used worldwide for this purpose.The purity of hybrid seeds directly influences the value of heterosis.Ensuring hybrid seed purity while reducing seed production cost are essential goals.To produce pure hybrid seeds,the pollen recipient(maternal parent)must be prevented from selfing.Emasculation is the commonly used method to ensure that the maternal parent receives only paternal pollen,but false hybrid seeds are difficult to avoid owing to incomplete emasculation.Although the mutant alleleZmSig2AΔVin the homozygous state may cause a slight(about 7%)decrease in the seed-setting rate,ZmSig2AΔVhad no negative effect on hybrid yield in the heterozygous state.The discovery of theeal1mutant with visible seedlingspecific etiolated or albino leaves provides a screening marker for increasing seed purity in maize hybrid seed production.As shown in Fig.7A,ZmSig2AΔVcan be introduced into maternal inbreds by backcross breeding or genome-editing technologies.When the maternal parent with theeal1mutation is used for hybrid seed production,seedlings with etiolated and/or albino leaves must be false hybrids derived from selfing of the maternal parent,whereas green seedlings must be true hybrids generated by hybridization.Removal of false hybrids with the help of seedling-color screening could greatly increase the purity of hybrids in the field and lead to increased yield.

The development of seed production technology(SPT)[47]in maize has brought hope for applying genic male sterility in hybrid seed production.Although many advantages of SPT for hybrid seed production have been suggested,transgene transmission resulting from incompletely inactivated transgenic pollen of the transgenic maintainer line persisted[48].To reduce this transmission rate,the multicontrol sterility(MCS)system[48]was developed.Although MCS features two pollen-disruption modules and the transmission rate is greatly reduced,transgene transmission was still observed[48].However,if the maternal lines carry theeal1mutation,the seed purity of the male-sterile line produced by the SPT or MCS systems can be further increased(Fig.7B).The modulePUbi::ZmSig2A+connected tightly with the modules of SPT or MCS systems can visually reveal mixed transgenic plants resulting from the transmission of transgene pollen during male-sterile line propagation after the male-sterile line has been provided with theeal1mutation,resulting in high purity of male-sterile line plants and hybrid seeds.Thus,using theeal1orZmSig2AΔVallele for hybrid seed production and male-sterile line propagation in the SPT or MCS systems,off-type contaminants arising from biological mixtures could be readily visually identified.

5.Conclusions

A revertible leaf-color mutanteal1displaying etiolated/albino leaves that gradually return to normal green at the seedling stage was identified.Abnormal chloroplast development appears to lead to the leaf-color mutation ofeal1plants and the revertible leaf color results from the gradual formation of normal chloroplasts.A single Val residue deletion in the σ4domain of the nuclearencoded and chloroplast-located σ factor ZmSig2A accounts for theeal1mutation.ZmSig2A is the first σ factor functionally characterized in maize.TheZmSig2Alocus would be an ideal genome editing target for the study of chloroplast gene regulation and chloroplast-nuclear communication,and theeal1mutation offers potential advantages for increasing seed purity in maize breeding as a seedling-specific selective marker.

CRediT authorship contribution statement

Chuan Li and Moju Caoparticipated in the conceptualization and writing-original draft.Chuan Li,Jingwen Wang,Zhaoyong Hu,Yuanyan Xia,Tao Yu,Hongyang Yi and Qiang Huangparticipated in the investigation and methodology.Jing Wang and Yanli Luparticipated in the 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

This research was supported by the National Key Research and Development Program of China(2016YFD0102104),Platform for Mutation Breeding by Radiation of Sichuan(2016NZ0106),and Applied Basic Research Program of Sichuan Provincial Science and Technology Department(2020YJ0249).We thank Professor Jirong Huang from the Institute of Plant Physiology and Ecology,Shanghai Institutes for Biological Sciences,Chinese Academy of Sciences for the gift of thesig2mutant.The EMS4-110eeb seeds used in this study were provided by Professor Chunyi Zhang,Institute of Biotechnology,Chinese Academy of Agricultural Sciences and Professor Xiaoduo Lu,Qilu Normal University.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.01.005.