CRISPR/Cas9-engineered mutation to identify the roles of phytochromes in regulating photomorphogenesis and flowering time in soybean

Fen Zhao,Xiangguang Lyu,Ronghuan Ji,Jun Liu,Tao Zhao,Hongyu Li,Bin Liu,*,Yanxi Pei

a School of Life Science and Shanxi Key Laboratory for Research and Development of Regional Plants,Shanxi University,Taiyuan 030006,Shanxi,China

b The National Key Facility for Crop Gene Resources and Genetic Improvement,Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,Beijing 100081,China

Keywords:Soybean Phytochrome CRISPR/Cas9 Photomorphogenesis Flowering time

ABSTRACT Soybean(Glycine max)responds to ambient light variation by undergoing multiform morphological alterations,influencing its yield potential and stability in the field.Phytochromes(PHYs)are plant-specific red(R)and far-red(FR)light photoreceptors mediating photomorphogenesis and photoperiodic flowering.As an ancient tetraploid,soybean harbors four PHYA,two PHYB,and two PHYE paralogs.Except for GmPHYA2/E4 and GmPHYA3/E3,which have been identified as photoperiod-dependent flowering repressors,the functions of GmPHYs are still largely unclear.We generated a series of individual or combined mutations targeting the GmPHYA or GmPHYB genes using CRISPR/Cas9 technology.Phenotypic analysis revealed that GmPHYB1 mediates predominantly R-light induced photomorphogenesis,whereas GmPHYA2/E4 and GmPHYA3/E3,followed by GmPHYA1 and GmPHYB2,function redundantly and additively in mediating FR light responses in seedling stage.GmPHYA2/E4 and GmPHYA3/E3,with weak influence from GmPHYA1 and GmPHYA4,delay flowering time under natural long-day conditions.This study has demonstrated the diversified functions of GmPHYAs and GmPHYBs in regulating light response,and provides a core set of phytochrome mutant alleles for characterization of their functional mechanisms in regulating agronomic traits of soybean.

1.Introduction

Light not only is the ultimate resource of energy for photosynthesis but serves as a signal to modulate plant morphology throughout the plant life cycle[1,2].In flowering plants,multiple photoreceptors with wavelength absorption spectra spanning from UV-B to FR(280-750 nm)light have been characterized with respect to function and molecular mechanisms,including phytochromes(PHYs)[3],cryptochromes(CRYs)[4],and the UV resistance locus 8(UVR8)[5].Upon exposure of germinating seedlings to light illumination,these photoreceptors perceive light signals and mediate photomorphogenic growth,such as by inhibiting hypocotyl elongation,promoting cotyledon expansion,and increasing anthocyanin accumulation[6,7].They also interpret information from light intensity,quality,diurnal and seasonal rhythmicity,and direction,functioning either distinctly or redundantly to fine-tune the architecture of sessile plants to adapt and survive in the ever-changing environment[8-11].

The R and FR light-absorbing photoreceptor phytochromes have been most extensively characterized in the model plant Arabidopsis thaliana.Phytochromes assume two photoconvertible forms,of which the Pr form absorbs R light(λmax=660 nm)and the Pfr form absorbs FR light(λmax=730 nm)[12].They detect and monitor R and FR light sources and convert between inactive(Pr)and active forms(Pfr)in response to changes in R:FR ratio,enabling phytochromes to function as R/FR-dependent molecular switches[13].The Arabidopsis genome contains five phytochrome genes encoding PHYA,PHYB,PHYC,PHYD,and PHYE[14,15].Previous work has shown that PHYB synchronizes plant development with the light environment with additional functions of PHYD and PHYE[16,17].The absence of PHYB results in elongated hypocotyl and petiole,less chlorophyll,and earlier-flowering phenotypes[18,19].

PhyB mutants exhibit hyponastic movement of leaves and elongate less than the wild type(WT)in response to end-of-day far-red(EOD-FR)or low R:FR-ratio conditions,indicating that PHYB is essential to regulate Arabidopsis shade avoidance syndrome(SAS)[20,21].In contrast to other phytochromes,PHYA exhibits a unique property:light lability in the active Pfr form[22].This property enables PHYA to detect low quantities of light with a so-called very low fluence response[23],to activate germination of seeds and de-etiolation of seedlings.PHYA also mediates far-red highirradiance responses(FR-HIRs),indicative of deep-canopy shade,to prevent excessive growth of the plant under continuous FRrich light conditions[8,24].Arabidopsis mutants deficient in PHYA display a more marked stem elongation in low R:FR(0.05)conditions than WT seedlings,and reduced survival in deep vegetational shade.

Phytochromes function in regulation of agronomic traits.Crop phytochromes show variation in gene-copy number,owing probably to whole or segmental genome duplications in three lineages.Rice and sorghum harbor single copy phytochromes,PHYA,PHYB and PHYC[25,26];the maize genome encodes a pair of phytochromes for each paralog,named PHYA1 and 2,PHYB1 and 2,and PHYC1 and 2[27];barley contains four phytochrome loci[28];and hexaploid wheat contains three copies for each paralog of PHYA,PHYB and PHYC[29].Phytochromes are thought to serve as photoreceptors mediating photomorphogenesis and SAS in crops[30].In rice,PHYA and PHYB redundantly regulate deetiolation under red light,whereas both PHYA and PHYC are required for photoresponse to far-red light[31].In maize,PHYB1 and PHYB2 show functional overlap in inhibiting sheath and stem elongation[32]and PHYA1 functions as a positive regulator of SAS[33],whereas overexpression of PHYC2 modestly reduces maize plant height,suggesting that it may attenuate SAS in maize in response to shade signal[34].In sorghum,phyB-1 mutants show typical SAS including elongated stem[35].Wheat phyB-null and phyC-null mutants displayed lower tiller number,prolonged flag leaves and elongated stem phenotypes[36].The roles of phytochromes in regulation of flowering time vary by crop.Absence of PHYB in sorghum and of PHYB/PHYC in both rice and maize result in early-flowering phenotypes[31,34,37].Loss of function of PHYB/PHYC in wheat and of PHYC in barley delays flowering time[38,39].

Soybean(Glycine max(L.)Merr.),a worldwide source of human food,animal feed,and industrial materials,is hypersensitive to variation of the light environment,including light quality,quantity,direction,and daylength[40].Low blue(B)light or low R:FR ratio resulting from canopy conditions or adjacent vegetation evokes several morphological alterations,including excessive stem growth,increased petiole elongation and hyponastic leaf movement,to escape from severe conditions or potential threats[41-43].A long-day photoperiod can inhibit soybean flowering[44].Switching from short-day to long-day conditions can even reverse the soybean developmental stage from floral to vegetative growth[45].Thus,light signals received by photoreceptors play essential roles in soybean development and growth in acclimation to surrounding environments.

Unlike Arabidopsis,soybean is an ancient polyploid(paleopolyploid)species with a high number of repetitive sequences and diversification of genes[46].The soybean genome encodes multiple homologs of GmPHYs and GmCRYs[47-49].Several photoreceptors have been identified in soybean.The E3 and E4 genes,which encode respectively the GmPHYA3 and GmPHYA2 proteins[47,50],have been reported to control time to flowering and maturity.Absence of E3 and E4 leads to insensitivity to photoperiod,resulting in early flowering under long-day conditions,which enables commercial soybean cultivation to expand to higher latitudes[50,51].The E3 and E4 genes also function in postflowering photoperiod response,such as by stem termination and pod development[52,53].Besides photoperiod sensing,E4 is required for FR light-induced de-etiolation.Compared with plants homozygous for the E4 allele,its NILs(near isogenic lines)with e4 genotype showed elongated hypocotyls under continuous FR light[47].Thus,compared with their roles in Arabidopsis,photoreceptors have conserved and novel functions in soybean.

The aim of this study was to characterize the functional conservation and divergence of GmPHYAs and GmPHYBs genes in soybean by generating a series of CRISPR/Cas9-engineered mutant plants.The phenotypic analysis revealed their distinct roles in regulating seedling photomorphogenesis and flowering time in soybean.

2.Materials and methods

2.1.Plant materials and growth conditions

The wild-type soybean was the cultivar Tianlong 1(TL1)(E1e2nsE3MiE4)carrying functional E1,E3(GmPHYA3),and E4(GmPHYA2)genes[54].For phenotypic analysis,the seedlings of the lines were grown in growth chambers at a constant temperature of 25-26 °C under constant LED W light(100μmol m-2s-1),R light(50μmol m-2s-1),FR light(30μmol m-2s-1),or B light(50μmol m-2s-1)or in dark conditions.Light quality and intensity were measured with a HiPoint HR-350 spectrometer.Origin 2021 was used for drawing spectral distributions(Fig.S3).The hypocotyl and epicotyl lengths were recorded on day 11 after sowing(DAS 11).To record flowering dates,lines were planted under natural conditions on June 15,2021 in a field in Beijing(40.1°N,116.7°E)with two plot replications.The seeds were planted in a 3.0 m row,with 0.6 m between rows and 0.3 m between adjacent plants.The flowering time of each plant was recorded as days from sowing to the R1 stage(one flower at any node)[55].

2.2.Plasmid construction

The CRISPR/Cas9 system was employed to knock out soybean GmPHYAs and GmPHYBs.The Cas9 protein was expressed under the Arabidopsis RPS5A promoter[56].The cassette of the guide RNA(gRNA)was driven by the soybean U6 promoter[57].The DNA sequences of the four GmPHYA genes and the two GmPHYB genes(Fig.1A)were retrieved from Phytozome[58].All gRNAs were designed with CRISPR-P 2.0(https://crispr.hzau.edu.cn/CRISPR2/)[59].The specificity of each gRNA for its target site was tested by BLAST tool in the Soybase database[60].The CRISPR/Cas9-gRNA expression vectors were constructed as previously reported[61,62].All primers used for the constructs are described in Table S1.

2.3.Soybean hairy root transformation and stable transformation

To test gRNA editing efficiency,the CRISPR/Cas9-gRNA expression vectors carrying sgRNA cassettes were transformed into Agrobacterium rhizogenes strain K599,followed by soybean hairy root transformation as previously described with some modifications[63].TL1 seeds were sterilized with chlorine gas,then germinated in germination medium(GM)for 3 days,and the hypocotyls of germinated seedlings were then cut 0.5 cm from the cotyledon.GM contains B5(Sigma)supplemented with 3%sucrose,0.3%phytagel,and 112 mg L-1B5 vitamin(pH 5.8).The cotyledons of TL1 were separated to remove apical buds attached to the cotyledonary node,incubated with the K599 strain for 30 min,and then placed in co-cultivation medium(CCM)with filter paper in the dark at 28 °C for 3 days.CCM contains 1/10 B5 supplemented with 3%sucrose,3.9 g L-1MES,0.5% agar,40 mg L-1acetosyringone(AS),and 112 mg L-1B5 vitamin(pH 5.4).Liquid root induction medium(RIM)wash 4-5 times then inserted into the solid RIM keeping the hypocotyl upward for 2 weeks.RIM contains B5 supplemented with 3% sucrose,0.59 g L-1MES,0.8% agar,50 mg L-1cefotaxime,100 mg L-1timentin,50 mg L-1vancomycin,and 112 mg L-1B5 vitamin(pH 5.7).The transgenic hairy roots were tested by sequencing using specific primers(Table S1).

Fig.1.Phylogenetic tree and expression profiling of the GmPHY genes.(A)Phylogenetic tree of soybean phytochrome homologs.At and Gm represent Arabidopsis thaliana and Glycine max,respectively.Full-length protein sequences of phytochrome obtained from Phytozome(https://phytozome.jgi.doe.gov/pz/portal.html)were used to construct the tree.GmPHYAs and GmPHYBs are marked with red letters.The scale bar indicates the mean number of amino acid substitutions per site.(B)A heat map illustrating expression profiling of the GmPHY genes in multiple tissues.Gene expression values were normalized with Z-scores and cluster analysis by OmicShare 6.3.0(https://www.omicshare.com/tools/).Relative transcript level is indicated on a color scale from red(high)to blue(low).

For stable transformation,the CRISPR/Cas9 plasmid was transformed into Agrobacterium tumefaciens strain EHA105,which was transformed into soybean TL1 by the cotyledonary node method[61].

2.4.Phylogenetic analysis

The amino acid sequences of PHYs in Ara bidopsis were retrieved from the TAIR database(https://www.arabidopsis.org/)and GmPHYs in soybean were retrieved from Phytozome(Wm82.a4.v1,https://phytozome-next.jgi.doe.gov/info/Gmax_Wm82_a4_v1)and used to construct a phylogenetic tree.Because the annotation of the GmPHYA4 gene was incorrect in the Wm82.a4.v1 version,the protein sequence of PHYA4 refers to the Wm82.a2.v1 version(https://phytozome-next.jgi.doe.gov/info/Gmax_Wm82_a2_v1).

Multiple alignments of amino acid sequences were constructed with ClustalX[64].A phylogenetic tree was constructed with MEGA5[65]by the neighbor-joining method with 1000 bootstrap sampling using full-length sequences.

2.5.Expression profiling

An expression heat map was constructed for the phytochrome genes using the gene TPM values from the Soybean Expression Atlas (https://venanciogroup.uenf.br/cgi-bin/gmax_atlas/index.cgi)[66].Gene expression values were normalized with Z-scores and cluster analysis by OmicShare 6.3.0(https://www.omicshare.com/tools/).

2.6.DNA extraction and mutation screening

Genomic DNA was extracted from the roots of transgenic soybean hairy roots and leaves of transgenic plants using the CTAB method[67].The regions spanning the targets of the GmPHYA and GmPHYB genes were amplified using sequencespecific primer sets(Table S1).The purified DNA fragments were sequenced and examined by Chromas(https://technelysium.com.au/wp/chromas/)and VectorNTI software[68].For obtaining homozygous mutants,the transgenic lines were propagated for at least three generations by self-pollination.Homozygous mutations in GmPHY genes were identified by DNA sequencing(Figs.S4 and S5).

2.7.Data analysis

For phenotypic evaluation of seedling plant height,at least seven plants per genotype were used.Two-way ANOVA(with Tukey’s multiple-comparisons test)was used to identify significant differences between pairs of groups.For phenotypic evaluation of flowering time,at least nine plants per genotype were used.Oneway ANOVA(with Tukey’s multiple-comparisons test)was used to identify significant differences between pairs of groups.Statistical analyses were performed with GraphPad Prism 9 software[69].

2.8.Off-target analysis

Possible off-target sites of each gRNA(Bg1-4,Ag1-5)were predicted with CRISPR-P 2.0.Given that two mismatches,particularly those occurring in a PAM(protospacer-adjacent motif)-proximal region,greatly reduce CAS9 activity[70],potential off-target sites with mismatch number≤3 were selected for further analysis(Table S2).Among them,the potential off-target event of phytochrome genes was checked by DNA sequencing with specified primers(Table S1).The verified genotype of each mutant line is shown in Table S3.

3.Results

3.1.Phylogenetic analysis and gene expression pattern of phytochromes in soybean

Previous studies[49]have revealed that the Arabidopsis genome encodes five phytochromes(PHYA,PHYB,PHYC,PHYD,and PHYE).Soybean harbors eight phytochromes,named GmPHYA1,GmPHYA2,GmPHYA3,GmPHYA4,GmPHYB1,GmPHYB2,GmPHYE1,and GmPHYE2,based on their orthologous relationships with respective PHYs in the model plant Arabidopsis[71].Given that a previous phylogenetic analysis was performed with the first version of the soybean reference genome(Wm82.a1.v1),we updated the phylogenetic tree of phytochromes in Arabidopsis and soybean using the latest version(Wm82.a4.v1)(Fig.1A).Given that the P3/GAF domain is associated with the bilin chromophore and is highly conserved among plant phytochromes[12],we executed a BLAST search(https://phytozome-next.jgi.doe.gov/blast-search)using the consensus-typical GAF domain sequence as a query,and verified that soybean harbors eight phytochrome genes.In agreement with the previous report,our visualized phylogenetic tree demonstrated the expansion of PHYA,PHYB and PHYE in soybean,suggesting their subfunctionalization since whole-genome duplication.No PHYC ortholog was found in the latest reference genome,supporting a hypothesis that PHYC had been lost prior to the diversification of legume species.Wu et al.[49]previously named the Glyma20g22160 locus(Wm82.a1.v1)GmPHYA1 and the Glyma10g28170 locus GmPHYA2.Given that an earlier pioneering study by Liu et al.[47]had already designated the E4 gene(Glyma.20G090000,Wm82.a4.v1)as GmPHYA2,we propose to use GmPHYA1 for Glymahttps://doi.org/10g28170/Glyma.10G141400 and GmPHYA2/E4 for Glyma20g22160/Glyma.20G090000 hereafter.

To characterize the functional diversity of phytochrome genes in soybean,we compared their transcriptional levels in multiple organs and tissues with normalized TPM(transcripts per kilobase million)values[72]in the soybean cultivar Williams 82(Fig.1B).All the GmPHYA and GmPHYB genes were expressed in at least one sample examined.Co-expression cluster analysis indicated that paralogous gene pairs(GmPHYA1/2,GmPHYB1/2 and GmPHYA3/4)showed similar expression patterns,a finding consistent with their phylogenetic relationship and the principle that paralogous genes share redundant or similar functions.GmPHYA1/2 genes were expressed at higher levels and in different patterns than GmPHYA3/4 genes,suggesting that they have evolved different functions.Both GmPHYA1/2 and GmPHYB1/2 were expressed at higher levels in juvenile samples including seedling,cotyledon,and hypocotyl than in other tissues,implying their essential roles in mediating de-etiolation and photomorphogenesis at seedling stage.

3.2.CRISPR/Cas9-engineered mutagenesis of GmPHYB genes

Next,we interrogated the functions of GmPHYB1 and GmPHYB2 genes in soybean using CRISPR/Cas9 technology.To generate individual or combined mutations,we designed four gRNAs,including Bg1(the GmPHYBgRNA 1),Bg2 and Bg4 simultaneously targeting the GmPHYB1 and GmPHYB2 genes,and Bg3 specifically targeting the GmPHYB1 gene.All these target sites are located in the first exons of the indicated genes(Fig.2A).The Bg1 and Bg2 expression cassettes targeting GmPHYB1 and GmPHYB2 genes were inserted into the dual-sgRNAs CRISPR/Cas9 vector,while the Bg3 or Bg4 expression cassettes were inserted separately into single sgRNA CRISPR/Cas9 vectors[61,62].The soybean U6 promoter was used to drive the sgRNA expression cassettes.The editing efficiency of each gRNA was verified with the soybean hairy root transformation system[73].We next performed soybean transformation and obtained multiple T0transgenic lines for each construct.DNA sequencing was performed to identify the CRISPR/Cas9-induced mutations at each target site in the transgenic offspring.We identified multiple independent gmphyB1,gmphyB2,and gmphyB1B2 mutants that carried frameshift mutations resulting in premature termination of translated proteins.Among them,the gmphyB1d1mutant harbored a 1-bp deletion in the first exon that resulted in a truncated GmPHYB1 protein containing only 82 amino acids and lacking all domains required for photosensing and signal transduction.The gmphyB2d2mutant harbored a 2-bp deletion in the first exon that created a truncated GmPHYB2 protein containing 468 amino acids and lacking the C-terminal PAS and HK domains.The gmphyB1i1B2d2mutant harbored a 1-bp insertion in the GmPHYB1 gene and a 2-bp deletion in the GmPHYB2 gene,leading to the absence of the C-terminal PAS domain and the HK domain(Figs.2B,C,S1).We investigated the potential off-target sites of each gRNA as described in Methods.DNA sequencing excluded the presence of these potential off-target events in the mutants(Table S3).

3.3.Diversified functions of GmPHYBs in mediating light regulation of seedling elongation

Previous studies[19,22]have revealed that PHYB is the predominant light receptor regulating Arabidopsis de-etiolation under R light.To determine the roles of GmPHYBs in regulating soybean photomorphogenesis,we recorded the hypocotyl and epicotyl lengths of WT plants and gmphyB1,gmphyB2,and gmphyB1B2 mutants grown under continuous white(W),R,FR,and B light conditions(Fig.S3)and in the dark for 11 days.Given that the allelic mutants used in this study behaved similarly under different light conditions,we present images and statistical results for representative lines(gmphyB1d1,gmphyB2d2and gmphyB1i1B2d2mutants)(Fig.3).The hypocotyls of all indicated lines elongated equally in the dark but were differentially inhibited by the light treatments.Under W and R light conditions,both hypocotyls and epicotyls of the gmphyB1d1mutant were significantly longer than those of WT plants.The gmphyB1d1mutant plants showed 38.1% increase in hypocotyl and 35.3% increase in epicotyl length under R light conditions relative to WT plants(P＜0.01),showing that GmPHYB1 is a prominent photoreceptor regulating soybean photomorphogenesis under R light(Fig.3B).The gmphyB2d2mutant behavior was similar to that of the WT plant,and the gmphyB1i1B2d2mutant showed a longer hypocotyl and/or epicotyl than the gmphyB1d1mutant under W(17.9% increase in hypocotyl and 22.0% increase in epicotyl,P＜0.01)and R(8.4% increase in hypocotyl,P＜0.05 and no significant difference in epicotyl)conditions,showing that GmPHYB2 can mediate R light response only in the absence of GmPHYB1.Under FR light conditions,the epicotyls of gmphyB2d2and gmphyB1i1B2d2were longer than those of the WT(respectively 34.7% and 38.3% increases,P＜0.01),indicating GmPHYB2 can function under FR light.The stem length of gmphyB1i1B2d2mutant was slightly greater than that of the WT plant under B light(23.8%increase in hypocotyl and 9.8% increase in epicotyl,P＜0.01),suggesting that GmPHYBs also contribute to B light responses.

Fig.2.CRISPR/Cas9-engineered mutagenesis in GmPHYB genes.(A)Genomic structures of the GmPHYB1 and GmPHYB2 genes.Black boxes indicate exons,gray boxes indicate untranslated regions(UTRs),continuous lines indicate introns.Bg1,Bg2,Bg3,and Bg4 indicate GmPHYB gRNA target sites.Scale bar,500 bp.(B,C)The schematic diagrams show respective mutations in the GmPHYB1 and GmPHYB2 genes.Nucleotides in red indicate the PAM.Black lines on nucleotide sequences indicate deletions.Letters underlined represent insertions.(B)The gmphyB1i1B2d2 mutant harbors a 1-bp insertion in the GmPHYB1 gene and a 2-bp deletion in the GmPHYB2 gene at the Bg4 site;the gmphyB1d1+d8B2d1+d3 mutant harbors a 1-bp deletion in the GmPHYB1 and GmPHYB2 genes at Bg1 site,and an 8-bp deletion in the GmPHYB1 gene and a 3-bp deletion in the GmPHYB2 gene at the Bg2 site.(C)The gmphyB1d1 and gmphyB1d5 mutant harbor a 1-bp and 5-bp deletions in the GmPHYB1 gene at the Bg3 site.The gmphyB2d1 mutant harboring a 1-bp deletion in the GmPHYB2 gene at the Bg1 site was derived from heterozygous transgenic lines of gmphyB1d1+d8B2d1+d3.The gmphyB2d2 mutant harboring a 2-bp deletion at the GmPHYB2 gene at the Bg4 site was derived from heterozygous transgenic lines of gmphyB1i1B2d2.

3.4.CRISPR/Cas9-engineered mutagenesis of GmPHYA genes

To identify the functions of GmPHYA genes in soybean,multiple gRNAs were designed for CRISPR/Cas9-engineered target mutagenesis,including Ag1(GmPHYAgRNA 1)targeting both the GmPHYA1 and GmPHYA2 genes,Ag2 and Ag3 specifically targeting the GmPHYA3 gene,Ag4 simultaneously targeting the GmPHYA3 and GmPHYA4 genes,and Ag5 specifically targeting the GmPHYA2 gene.All these target sites were located in exons of the respective genes(Fig.4A).The gene editing efficiency of each gRNA was verified by soybean hairy root transformation.We performed stable soybean transformation and obtained multiple T0lines for each gRNA construct.DNA sequencing of the transgenic offspring identified a series of homozygous single or double mutants,which were further used for phenotypic analysis.Among them,representative lines were shown with detailed information of mutations(Fig.4B and C).All gmphyA mutants harbored frameshift mutations,resulting in premature termination of translated proteins lacking the critical domains required for photosensing and signal transduction(Fig.S2).The potential off-target assay of each gRNA excluded the presence of off-target mutations in phytochrome genes in the mutants(Table S3).

3.5.Diversified functions of GmPHYAs in mediating light regulation of seedling elongation

To identify the functions of GmPHYAs in soybean photomorphogenesis,we compared the behaviors of the gmphyA1d10,gmphyA2i1,gmphyA3d2,gmphyA4d5,gmphyA1d10A2d2,and gmphyA3d158A4d5mutants with that of the WT plant under several light conditions.Seedlings of these mutants showed no clear differences from WT seedlings in the dark.Under FR light conditions,the gmphyA2i1and gmphyA3d2mutants showed significantly elongated hypocotyl(both increasing by 21.9%,P＜0.01)and epicotyl(with increases of 44.3%and 42%respectively,P＜0.01)phenotypes in comparison to the WT plant.The gmphyA1d10A2d2mutant exposed to FR light displayed a skotomorphogenic-like phenotype resembling the dark-grown seedling characterized by an apical hook and extremely elongated hypocotyl,demonstrating the dominant roles of GmPHYA2(E4),collectively with GmPHYA1,in mediating FR light response(Fig.5A and B).Loss of function of GmPHYA4 only moderately increased the epicotyl lengths in the WT(TL1)background and not in the gmphyA3 mutant background,suggesting a relatively weak role of GmPHYA4 in mediating FR light responses at the seedling stage.The gmphyA1A2 and gmphyA3A4 seedlings showed a mild elongation phenotype under R,B,or W light conditions,suggesting the involvement of GmPHYAs in multiple light responses.

3.6.Diversified functions of GmPHYs in regulation of flowering time

Previous studies[47,50-52]have found that GmPHYA2 and GmPHYA3 are encoded by the soybean flowering and maturity genes E4 and E3 respectively.To identify the roles of GmPHYA and GmPHYB genes in regulation of soybean flowering time,we investigated the flowering days of all the mutants in comparison to the WT plant grown under natural long-day conditions in the field.In agreement with previous studies,loss of function of GmPHYA2/E4 or GmPHYA3/E3 significantly shortened the time to flowering.Although the gmphyA1 and gmphyA4 mutants did not flower earlier than the WT,the gmphyA1A2 and gmphyA3A4 mutants flower earlier than the gmphyA2 and gmphA3 mutants,respectively,showing that GmPHYA1 and GmPHYA4 are functionally additive to GmPHYA2 and GmPHYA3 in delaying soybean flowering.The flowering times of the gmphyB1,gmphyB2,and gmphyB1B2 mutants were not different from that of WT plants,suggesting that GmPHYBs have negligible effect on floral transition at least under the tested conditions.

Fig.3.The elongated seedling phenotype of gmphyB mutants.(A)Representative seedling images of lines grown under continuous W light(100μmol m-2 s-1),R light(50μmol m-2 s-1),FR light(30μmol m-2 s-1),B light(50μmol m-2 s-1)conditions or in the dark at 25-26 °C for 11 days.Scale bar,10 cm.(B)Statistics of hypocotyl and epicotyl lengths of the seedlings shown in(A).Values are mean±SD of at least 9 seedlings.Significant differences between any two groups were evaluated by two-way ANOVA(Tukey’s multiple-comparison test)(P＜0.05).

4.Discussion

4.1.Gene duplication and redundancy of the soybean phytochromes

Our finding that GmPHYA and GmPHYB genes were actively expressed in all tissues examined and displayed different expression levels or patterns in different tissues suggests that their transcriptional variations may be associated with their diversified functions.Consistent with this the finding that the markedly higher expression levels of GmPHYB1 than that of GmPHYB2 were associated with the longer hypocotyl and epicotyl phenotype of the gmphyB1 mutant than that of the gmphyB2 mutant under R light.The higher expression levels of GmPHYA2/E4 and GmPHYA3/E3 than those of GmPHYA1 and GmPHYA4,respectively,also accord with the stronger elongated-stem and early-flowering phenotypes of the gmphyA2 and gmphyA3 mutants than those of the gmphyA1 and gmphyA4 mutants.These results support the hypothesis that gene expression divergence is an evolutionary driving force for the retention of duplicate genes[74].But the shared expression patterns of GmPHYA1 and GmPHYA2,GmPHYA3 and GmPHYA4,as well as GmPHYB1 and GmPHYB2 in different tissues suggest that these duplicated paralogous genes still have some redundant or additive functions.

Fig.4.Homozygous gmphyA mutants generated by the CRISPR/Cas9 system.(A)Genomic structure of the GmPHYA1,GmPHYA2,GmPHYA3,and GmPHYA4 genes:black boxes indicate exons,gray boxes indicate untranslated regions(UTRs),continuous lines indicate introns.Scale bar,500 bp.(B,C)The schematic diagrams show mutations in the GmPHYA genes.Nucleotides in red indicate the PAM.Black lines on nucleotide sequences indicate deletions.Letters underlined represent insertions.(B)The gmphyA1d10A2d2 mutant harbors a 10-bp deletion in the GmPHYA1 gene and a 2-bp deletion in the GmPHYA2 gene at the Ag1 site;The gmphyA3d158A4d5 mutant harbors a 158-bp deletion in the GmPHYA3 and a 5-bp deletion in the GmPHYA4 genes at the Ag4 site.(C)The gmphyA1d10 and gmphyA2i1 mutants harbors a 10-bp deletion or a 1-bp insertion in the GmPHYA1 or GmPHYA2 gene at the Ag1 site.The gmphyA2d6 mutant harbors a 6-bp deletion in the GmPHYA2 gene at the Ag5 site.The gmphyA3d1 or gmphyA3d2 mutant harbors a 1-bp or 2-bp deletion at the GmPHYA3 gene at the Ag2 or Ag3 site.The gmphyA4d5 mutant harbors a 5-bp deletion in the GmPHYA4 gene at the Ag4 site.

To fully address the functions of phytochromes in life cycles,higher-order triple,quadruple,and quintuple mutants have been generated in Arabidopsis,demonstrating the redundant and divergent roles of phytochrome genes in regulating photomorphogenesis,shade avoidance,circadian clock,and flowering time[16,17,75,76].Based on the loss-of-function phenotypes and expression profiles of each phytochrome gene in soybean,the gmphyA1A2B1B2 quadruple mutant would be the most relevant higher-order mutant to generate for elucidating the roles of GmPHYA and GmPHYB genes in soybean growth and development.The GmPHYE1 gene showed higher expression than GmPHYE2 in most tested tissues,suggesting that GmPHYE1 is more influential between duplicated paralogous GmPHYE1/GmPHYE2 genes.It is likely that GmPHYEs are functionally additive to GmPHYAs or GmPHYBs to regulate physiological responses in soybean,a hypothesis for future study.

4.2.GmPHYA and GmPHYB genes orchestrate photomorphogenesis and flowering time in soybean

Among the loss-of-function alleles generated in this study,the gmphyB1 mutation is the only one that showed clear elongated hypocotyl and epicotyl phenotypes under R light conditions.Knockout of the GmPHYB2 gene only moderately increased the hypocotyl length in the absence of GmPHYB1 function,indicating that GmPHYB1 is the major photoreceptor mediating R lightinduced photomorphogenesis in soybean.

Under FR light conditions,both the gmphyA2 and gmphyA3 mutants showed clear elongated hypocotyl and epicotyl phenotypes.The gmphyA1A2 mutant even displayed a skotomorphogenic phenotype under FR light,indicating that GmPHYA1 is additive to GmPHYA2/E4 in regulation of seedling de-etiolation.The seedlings of gmphyA4 and gmphyB2 mutants showed slightly elongated phenotypes under FR light.We propose that GmPHYA2/E4 and GmPHYA3/E3,with moderate roles of other phytochromes(except for GmPHYB1),function redundantly or additively in mediating FR light responses at seedling stage.

We further showed that the GmPHYA2/E4 and GmPHYA3/E3 genes play major roles in regulating photoperiodic flowering,while GmPHYA1 and GmPHYA4 function additively to GmPHYA2/E4 and GmPHYA3/E3 to delay flowering under natural long-day conditions.Unexpectedly,losses of function of GmPHYB1 and GmPHYB2 had no clear effect on flowering time under the natural long-day conditions in this study.Whether GmPHYBs could modulate flowering time under different climatic conditions or genetic backgrounds awaits further study.

The E3/GmPHYA3 and E4/GmPHYA2 genes function in regulating photoperiodic flowering under both natural daylength and artificially induced long-day conditions[77].GmPHYA3 and GmPHYA2 proteins together confer sensitivity to FR-enriched long day conditions[78].The GmPHYA4 protein in the soybean cultivars(Williams 82 and TL1)contains a deletion of eight amino acids(DCCAKNVK)in the N-terminal P3/GAF domain in comparison to the intact GmPHYA4 in the wild soybean accessions[79].We do not know whether the absence of the eight amino acids diminished GmPHYA4 activity.The gmphyA4 mutant plants showed no difference from WT,whereas the gmphyA3A4 mutant plants flowered earlier than the gmphyA3 mutant plants,suggesting that GmPHYA4 functions additively to GmPHYA3 to delay flowering in the TL1 background under natural long-day conditions.Further studies might reveal the behavior of the quadruple gmphyA1A2A3A4 mutant and elucidate the interaction among these genes in regulating flowering and maturity time in response to diverse photoperiods and light qualities.

Fig.5.The elongated seedling phenotype of gmphyA mutants.(A)Representative seedling images of lines grown under the same conditions as in Fig.3.Scale bar,10 cm.(B)Statistical analysis of hypocotyl and epicotyl length of the seedlings is shown in(A).Values are mean±SD of at least 7 seedlings.Significant differences between any two groups were evaluated by two-way ANOVA(Tukey’s multiple-comparison test)(P＜0.05).

The differing photoperiodic flowering behaviors of short-day crops(such as rice,sorghum,and soybean)and long-day crops(such as wheat and barley)are probably due to the opposite roles of their respective phytochromes in regulation of flowering time.In wheat and barley,PHYB/PHYC promotes flowering under LD conditions.In rice and sorghum,PHYB/PHYC inhibits flowering under LD conditions.In agreement with this hypothesis,gmphyA2 and gmphyA3 mutant plants showed early-flowering phenotypes under LD conditions(Fig.6),supporting the notion that GmPHYA2(E4)and GmPHYA3(E3)function as flowering repressors in soybean.

Fig.6.Flowering time of gmphyA and gmphyB mutants under natural field conditions.(A)Representative images of lines grown in the summer of Beijing,China(40.1°N,116.7°E).Scale bar,10 cm.(B)Mean flowering time(DAS)of each line as in(A).The flowering time is given as mean±SD(n≥9).The value of each plant is represented by a dot.Different letters above the bars indicate significant differences(P＜0.05)as determined by one-way ANOVA(Tukey’s multiple-comparison test).

This study has shed light on the diversified functions of GmPHYA and GmPHYB genes in regulating photomorphogenesis or photoperiodic flowering in soybean,using CRISPR/Cas9-engineered mutations in phytochrome genes.These phenotypic observations and mutant alleles will contribute to more extensive studies of the mechanisms by which light regulates agronomic traits and improve the performance of soybean cultivars in diverse cultivation environments in the future.

CRediT authorship contribution statement

Yanxi Pei and Bin Liu:Conceptualization,Project administration,Funding acquisition,Writing-review & editing.Fen Zhao:Data curation,Formal analysis,Investigation,Writing-original draft.Xiangguang Lyu:Formal analysis,Investigation,Writingoriginal draft.Ronghuan Ji,Jun Liu,Tao Zhao and Hongyu Li:Investigation,Resources.

Declaration of competing interest

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

Acknowledgments

This work was supported by the National Natural Science Foundation of China(31871705,32072091),the Agricultural Science and Technology Innovation Program(ASTIP)of the Chinese Academy of Agricultural Sciences,and the Central Public-interest Scientific Institution Basal Research Fund.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2022.03.008.