Feifei Ma,Ranzhe Li,Guanghui Guo,Fang Nie,Lele Zhu,Wenjuan Liu,Linlin Lyu,Shenglong Bai,Xinpeng Zhao,Zheng Li,Dale Zhang,Hao Li,Suoping Li,Yun Zhou,Chun-Peng Song

State Key Laboratory of Crop Stress Adaptation and Improvement,College of Agriculture,School of Life Sciences,Henan University,Kaifeng 475004,Henan,China

Keywords: A-WI Aegilops tauschii Wheat QTL mapping Yield-related traits

ABSTRACT To break the narrow diversity bottleneck of the wheat D genome,a set of Aegilops tauschii-wheat introgression (A-WI) lines was developed by crossing Ae.tauschii accession T015 with common wheat elite cultivar Zhoumai 18 (Zhou18).A high-density genetic map was constructed based on Single Nucleotide Polymorphism (SNP) markers and 15 yield-related traits were evaluated in 11 environments for detecting quantitative trait loci (QTL).A total of 27 environmentally stable QTL were identified in at least five environments,20 of which were derived from Ae.tauschii T015,explaining up to 24.27% of the phenotypic variations.The major QTL for kernel length(KL),QKl-2D.5,was delimited to a physical interval of approximately 2.6 Mb harboring 52 candidate genes.Three Kompetitive Allele Specific PCR (KASP)markers were successfully developed based on nonsynonymous nucleotide mutations of candidate gene AetT093_2Dv1G100900.1 and showed that A-WI lines with the T015 haplotype had significantly longer KL than the Zhou18 haplotype across all 11 environments.Four primary valuable A-WIs with good trait performance and carrying yield-related QTL were selected for breeding improvement.The results will facilitate the efficient transfer of beneficial genes from Ae.tauschii into wheat cultivars to improve wheat yield and other traits.

1.Introduction

Hexaploid wheat (Triticum aestivumL.,AABBDD,2n=6x=42)originated from hybridization of diploidAegilops tauschii(DD,2n=14) with tetraploidT.turgidum(AABB,2n=4x=28) about 8000 years ago in the Fertile Crescent [1-3].The genetic diversity of the wheat D genome is significantly lower than that of the A and B genomes [4,5].The low-level genetic diversity of the wheat D-genome has been a severe genetic bottleneck for selecting desirable agronomic traits in wheat breeding[6].Therefore,it is necessary to increase genetic diversity in wheat breeding pools to overcome this genetic constraint and incorporate novel alleles or genes into modern cultivars.

Aegilops tauschiihas significantly higher genetic diversity than the D-genome of common wheat [5,6],and possesses many desirable genes for wheat improvement.Many genes/quantitative trait loci (QTL) fromAe.tauschiihave been identified,such as diseaseresistance genesSrTA1662[7],Pm58[8,9],Cmc4[10,11],andLr42[12].Some studies also reported the improved environmental stress resistance,including tolerance to phosphorus (P)-deficiency [13],cadmium (Cd) [14],and drought stress [15].Comparatively,there are only limited reports related to grain productivity traits.Thus,exploring moreAe.tauschiiaccessions is required to widen the genetic diversity of wheat.Wild populations ofAe.tauschiiare naturally distributed in central Eurasia,spreading from northern Syria and Turkey to western China [16].Aegilops tauschiifrom China was reported to show a distinctive population structure based on consensus phylogenetic relationships ofAe.tauschiiaccessions [16,17].Therefore,Ae.tauschii,especially those from China,is a valuable wild resource for enriching the genetic diversity of common wheat.Introgression and utilization of beneficial genes fromAe.tauschiiwill have great potential for improving wheat cultivars in breeding programs.

Numerous efforts have been made to harness a large share of theAe.tauschiigene pool for wheat improvement.Those efforts have followed two distinct approaches: indirect,crossingAe.tauschiiwith tetraploid wheat;and direct,crossingAe.tauschiiwith hexaploid wheat.The direct approach provides an alternative strategy to transfer desirable D genomic regions(carrying target alleles) without disrupting adaptive allelic combinations located in the A and B genomes.Some genes have been transferred fromAe.tauschiito common wheat,for example,SrTA1662[7],pre-harvest sprouting(PHS)resistance gene[6],and wheat curl mite resistance geneCmc4[10,11].Nyine et al.[18] assessed the impact of the introgression fromAe.tauschiiinto wheat cultivars on yield component traits using the direct approach and found that up to 23%of the introgression lines produced more grains than their parents.

Direct hybridization can generate synthetic octoploid wheat(SOW,AABBDDDD) if the chromosome sets of the F1hybrids were successfully doubled [6].Then,SOW will be backcrossed with a receptor parent to developAe.tauschii-wheat introgression (AWI) populations for identifing QTL fromAe.tauschii.For instance,Olson et al.[19] made a direct cross using a stem rust-resistantAe.tauschiiaccession and a stem rust-susceptible wheat breeding line.Using the BC2F1mapping population,a stem rust resistance gene from anAe.tauschiiaccession was located on the chromosome arm 1DS.Zhang et al.[20] detected two major QTL for seed dormancy on chromosomes 2D and 3D by the SOW approach,which explained 10.3% and 20.4% of the phenotypic variations,respectively.These results indicated that SOW is an efficient way to transfer favorable genes into elite wheat germplasm and detect promising QTL derived fromAe.tauschiiby developing an A-WI population containingAe.tauschiichromosomal segments.However,few studies on detecting QTL of yield-related traits using AWI populations have been reported.

In this study,we genotyped an A-WI population derived fromAe.tauschiiT015 and Zhou18 and identified QTL for 15 yieldrelated traits evaluated in 11 environments.Our objectives were to: (i) construct a high-density genetic linkage map of the A-WI population;(ii) identify environmentally stable QTL that were significant for yield-related traits in five or more environments;(iii)predict candidate genes for the key QTL;and(iv)screen elite introgression lines with good performance.The results contribute environmentally stable yield-related QTL and facilitate the transfer of favorable yield-related genes fromAe.tauschiiinto elite wheat germplasm.

2.Materials and methods

2.1.Plant materials and field trials

Aegilops tauschiiaccession T015 was originally collected from Henan province of China.Zhoumai 18 (Zhou18) is an elite variety registered in the same province.An A-WI population,consisting of 322 BC1F9lines,was developed by directly crossingAe.tauschiiaccession T015 and wheat cultivar Zhou18 (Fig.S1) [21].The T015/Zhou18 A-WI lines and two parents were grown at Zhongmou (ZM) and Huixian (HX) during the growing season 2015-2016,and at Huixian (HX),Kaifeng (KF) and Shangqiu (SQ) during 2016-2017 and 2017-2018,forming 11 year×location trials,designated ‘16ZM’,‘16HX’,‘17HX’,‘17KF’,‘17SQ’,‘18HX-1’,‘18HX-2’,‘18KF-1’,‘18KF-2’,‘18SQ-1’,and ‘18SQ-2’.A split-plot design with 2 replicates was used in each trial.Each plot consisted of threerows with 20 plants per row.The row length was 150 cm with 30 cm between rows.

2.2.Non-denaturing fish detections

The karyotype based on non-denaturing fluorescentin situhybridization(FISH)was conducted on the A-WI lines and the parental lines,according to previously described methods [6,17].The oligonucleotide probes were oligo-pTa535 (labelled with 6-carboxytetramethylrhodamine at the 5′end),oligo-(GAA)10(labelled with 6-carboxyfluorescein at the 5′end),oligo-pSc119.2(labelled with CY-5 at the 5′end),and oligo-pTa71(labelled with CY-5 at the 5′end).

2.3.Trait assessments

For each A-WI line of T015/Zhou18,ten plants in the middle of the internal rows were sampled to investigate the following traits:plant height (PH),spike length (SL),tiller number (TN),kernel number per spike(KNS),total spikelet per spike(TSS),fertile spikelet number per spike(FSS),heading date(HD),flowering date(FD),flag leaf length (FLL),and flag leaf width (FLW).All traits were determined by the mean of the ten plants.Seeds were harvested from five randomly selected plants from each line.Kernel length(KL),kernel width (KW),kernel length-width ratio (KLWR),and kernel perimeter (KP) were calculated from 100 kernels using SCG software (WSeen,Hangzhou,Zhejiang,China).Thousand-kernel weight (TKW) was evaluated by weighing three samples of 500 kernels per sample from each of the eight trials (Table S1).

2.4.Statistical analysis of field data

Mean values of yield-related traits (Mean),standard deviations(SD),coefficients of variation (CV),and an analysis of variance(ANOVA)were determined using IBM SPSS Statistics 20.0 software(SPSS,Chicago,USA).Broad-sense heritability was calculated using the ANOVA model to evaluate the variance components on an accession mean basis:,whereinis the genotypic variance=(MSg-MSge)/(re),is the genotype × environment interaction variance=(MSge-MSe)/r,andis the error variance;MSe,MSg,and MSgecorrespond to the mean squared error,genotype,and genotype by environment mean square,respectively;r is the number of replications,and e refers to the number of environments [22].Pearson’s correlation coefficients between traits were calculated with the GGally packages in the R package (https://www.rdocumentation.org/packages/GGally/versions/1.5.0).

2.5.Genotyping and genetic map construction

The T015/Zhou18 A-WI population and the two parents were genotyped using the Wheat 55 K SNP array and 12,892 SNP markers from the D genome were retained.SNP markers were removed if they have more than 10% missing values or exhibit minor allele frequencies of lower than 5%.Flanking sequences of SNPs were used to BLAST against the reference genome sequences ofAe.tauschii(accession T093) [6] and Chinese Spring (IWGSC RefSeq v1.1,https://wheatomics.sdau.edu.cn/) to identify their physical positions.To reduce the complexity of calculation,redundant markers(i.e.,co-segregating markers)in the A-WI population were removed using the BIN function in IciMapping 4.1(https://www.isbreeding.net/) [23].The retained markers were tested for significant segregation distortion using a Chi-squared test.SNPs were sorted into linkage groups using the MAP function in IciMapping 4.1.The Kosambi mapping function was used to calculate genetic distances in centimorgans (cMs) with a Logarithm of odds (LOD)score of 3.5 and a recombination fraction of 0.3[24].Markers without linkage or linkage groups with less than 5 markers were discarded in the subsequent analysis.MapChart 2.2 (https://mybiosoftware.com/mapchart-2-2-graphical-presentation-linkage-maps-qtls.html) was used to draw the genetic map.For the redundant loci that were co-segregated in the A-WI population,only one unique informative marker was shown in the genetic map.

2.6.QTL analysis

Analysis of QTL was conducted using the BIP function in IciMapping 4.1.The inclusive composite interval mapping of additive(ICIM-ADD) QTL method was chosen.The phenotypic values of the A-WI population in each environment were used for individual environment QTL mapping analysis.For QTL detection,the walk speed was 1.0 cM,and the stepwise regression probability was 0.001.The LOD scores was calculated using 1000 permutations with a type I error of 0.05.Digenic epistasis and environment interaction of QTL were analyzed using QTL IciMapping V4.1 through inclusive composite interval mapping of epistatic QTL (ICIM-EPI).The LOD score to detect the digenic epistasis QTL was set at 5.0.The QTL × environment interactions were scanned with QTL IciMapping V4.1 through ICIM-ADD.For a given trait,it was assumed that QTL with overlapping confidence intervals were treated as equivalent.All QTL were named according to Fan et al.[25].

2.7.Cloning the candidate genes

Reference genomic sequences (Ae.tauschiiaccession T093 and Chinese Spring RefSeq v1.1) of the candidate genes were used to design a set of conserved primers.PCR products amplified from the two parents were gel-purified and ligated into the pGEM-T Easy Vector (Promega,Beijing,China),and positive clones were selected and sequenced.Nucleotide sequence alignments of parents were conducted using DNAMAN 8 (8.0.8.789,Lynnon BioSoft,San Ramon,CA,USA).

2.8.Conversion of SNPs to KASP markers

Based on the nonsynonymous nucleotide mutations of the candidate gene between T015 and Zhou18,we developed a set of Kompetitive Allele Specific PCR (KASP) markers (LGC Genomics LLC,Beverly,MA,USA).Newly designed KASP markers were evaluated for polymorphisms in reaction mixtures containing 5.0 μL water,5.0 μL 2× KASPar reaction mix,0.14 μL assay mix,and 50 ng dried DNA.The PCR profile was set at 94 °C for 15 min (activation),followed by 10 cycles of 94°C for 20 s,61-55°C for 60 s(drop 0.6 °C per cycle),and then 26 cycles of 94 °C for 20 s and 55 °C for 60 s in a Roche LightCycler 480 Real-Time PCR system(F.Hoffmann-La Roche Ltd.,Basel,Switzerland) and fluorescence signal was analyzed using the LightCycler software.

3.Results

3.1.Chromosome constitutions of T015/Zhou18 A-WI lines

To evaluate chromosome stability of the T015/Zhou18 A-WI population,22 A-WI lines were randomly selected and screened by ND-FISH.The karyotype analysis revealed that all A-WI lines were euploid with 42 chromosomes.Comparatively,chromosomes from A and B subgenomes showed minimal difference between the A-WI lines and Zhou18,while chromosomes from the D subgenome carried high levels of homozygous variations,indicating successful introgression of multiple alleles fromAe.tauschiiT015 into A-WI lines (Fig.1).It was thus indicated that A-WIs inherited unique alleles fromAe.tauschiiT015,and chromosomes were almost stable in the BC1F9generation.

Fig.1.ND-FISH karyotype of T015/Zhou18 A-WI lines and parent lines.Four representative A-WI lines were screened by ND-FISH.Karyotype of common wheat Zhou18 was placed on the left panel and karyotype of Ae.tauschii T015 was placed on the right panel.Green,oligo-(GAA)10;red,oligo-pTa535;white,oligo-pSc119.2 and oligo-pTa71.Arrowheads indicate chromosomes with polymorphic signals.

3.2.Phenotypic variation and correlation analysis

Phenotypic performance of the A-WI population and their parents for the 15 traits are shown in Fig.2 and Table S2.Among the A-WI lines,phenotypic values for each trait showed broad and continuous variations(Figs.2,S2).Coefficients of variation of the seven traits were greater than 10%,indicating that these trait values differed among lines in the population.These results indicated that the introgression ofAe.tauschiiT015 into bread wheat could generate a high level of phenotypic variations in the derived A-WIs.

Fig.2.Phenotypic variation of the T015/Zhou18 Ae.tauschii-wheat introgression (A-WI) population.

Estimated correlation coefficients among the 15 traits are shown in Table S3.For the three yield traits,TKW had a positive correlation with kernel-size related traits such as KL,KW,and KP.Trait KNS was positively correlated with spike-related traits and flag leaf-related traits(FLL and FLW).Trait TN had positive correlations with KLWR and FLL,and had negative correlations with KW,FLW,and HD.Significant correlations were observed among spike-related traits.Trait TSS had the highest positive correlation with FSS (r=0.89).For the five kernel-related traits,KL had a strong positive correlation with KP (r=0.94).Trait KLWR had a negative correlation with KW and did not significantly correlate with TKW.

3.3.Linkage map construction

The T015/Zhou18 A-WI population was genotyped with the Wheat 55 K SNP array and a final set of 3131 polymorphic markers on the D genome were retained.After removing unlinked markers,the resulting genetic map consisted of 1426 markers mapped within 506 bins.Thus,one unique marker was chosen to represent each bin to construct the genetic map (Fig.3).

Fig.3.High-density genetic linkage map of the T015/Zhou18 A-WI population based on the Wheat 55K SNP array.For redundant loci that showed co-segregation in the A-WI population,only one unique informative marker is shown.The positions of the marker loci are indicated using a ruler on the left side.The names of the marker loci are listed to the right of the corresponding chromosomes.

3.4.QTL mapping

A total of 346 significant QTL were detected for the 15 traits examined in 11 environments(Tables S4,S5;Fig.S3).Chromosome 7D had the most QTL(76),followed by 2D(56),6D(54),5D(52),1D(47),3D (46),and 4D (15).Twenty-seven environmentally stable QTL were detected in at least five environments,explaining 1.36%-24.27% of the phenotypic variations.Twenty stable QTL had positive alleles provided by T015,and the remaining seven stable QTL were contributed by alleles from Zhou18 (Table S4).

3.4.1.Kernel-related traits

One hundred and ten QTL were detected for kernel-related traits (KL,KW,KP,KLWR,and TKW) and explained 1.42%-21.99%of the phenotypic variations(Fig.S3;Tables S4,S5).Seven environmentally stable QTL were identified on chromosomes 2D,5D,6D,and 7D(Fig.4;Table S4).Four environmentally stable QTL had positive alleles contributed by T015,and they were designatedQKl-2D.5,QKlwr-2D.5 QKlwr-5D.3,andQKlwr-7D.8,respectively.QKl-2D.5was detected for KL in seven environments,explaining 5.91%-19.83% of the phenotypic variations.The QTL was also significant for KLWR (QKlwr-2D.5) in nine environments,explaining 5.43%-21.99% of the phenotypic variations.QKlwr-5D.3andQKlwr-7D.8were detected in five to seven environments,explaining as much as 8.62% of the phenotypic variations.

Fig.4.Chromosomal locations of 27 environmentally stable QTL for yield-related traits.Acentimorgan(cM)scale is shown on the left.The short arms of the chromosomes are located at the top.The names of the marker loci and the QTL are listed on the right side of the corresponding chromosomes.The positions of the marker loci are listed on the left side of the corresponding chromosomes.Orange triangles indicate QTL with positive alleles from parent T015;green triangles indicate QTL with positive alleles from parent Zhou18.

3.4.2.Spike-related traits

Among the 81 QTL for spike-related traits (SL,TSS,FSS,and KNS),four environmentally stable QTL were mapped on chromosomes 2D and 7D(3)(Figs.4,S3;Tables S4,S5).Two environmentally stable QTL (QSl-2D.2andQSl-7D.3) for SL had positive alleles contributed by T015,explaining up to 17.44% of the phenotypic variations.

3.4.3.Plant architecture-related traits

In total,98 QTL were detected for plant architecture-related traits(PH,TN,FLL,and FLW) and explained 1.36%-24.27% of the phenotypic variations(Fig.S3;Tables S4,S5).Five environmentally stable QTL were identified on chromosomes 2D,4D,6D,and 7D,and they were designatedQPh-2D.4,QPh-7D.1,QFlw-4D.1,QFlw-4D.2,andQFlw-6D.2,respectively(Fig.4;Table S4).All the five environmentally stable QTL had positive alleles contributed by T015.QPh-2D.4andQPh-7D.1were detected in five to six environments,explaining as much as 24.27% of the phenotypic variations.QFlw-4D.1,QFlw-4D.2andQFlw-6D.2were detected in six to eight environments,explaining up to 16.22%of the phenotypic variations.

3.4.4.Flowering time-related traits

A total of 57 QTL were detected for flowering time-related traits(HD and FD) and eleven environmentally stable QTL were identified on chromosomes 1D,2D,4D,6D,and 7D (Figs.4,S3;Tables S4,S5).Four environmentally stable QTL for HD had positive alleles contributed by T015 and in some cases explained more than 10%of the phenotypic variations.They were designatedQHd-2D.1,QHd-6D.3,QHd-7D.1,andQHd-7D.5,respectively.QHd-6D.3was significant in five environments and explained 7.65%-18.42%of the phenotypic variations.QHd-2D.1,QHd-7D.1andQHd-7D.5were significant in six to nine environments,explaining as much as 18.52% of the phenotypic variations.These QTL were also significant for FD in five to eight environments,which were designatedQFd-2D.2,QFd-7D.1andQFd-7D.6,respectively,explaining up to 16.48% of the phenotypic variations.

3.5.Epistasis and QTL × environment interaction

A total of 77 pairs of epistasis QTL for the 15 yield-related traits were detected,explaining 0.18%-2.53% of the phenotypic variations,indicating that the interactions between those QTL had no significant major effect on the evaluated traits (Table S6).Moreover,QTL × environment (QE) interactions were identified to be overlapped with 263 QTL,including 27 environmentally stable QTL and 236 putative QTL (Table S7),indicating that these evaluated traits were affected by the environmental factors.The environmental effects [PVE (AbyE)] of environmentally stable QTL ranged from 0.46% to 10.06%,while those of putative QTL ranged from 0.08%to 4.16%.Among them,the largest environmental effect was detected in the intervalAX-110611968-AX-94988821[PVE(AbyE)=10.06%],indicating that the environmentally stable QTLQTss-7D.1for TSS was significantly affected by environments(Table S7).The environmental effect of ten stable QTL (QKl-6D.4,QKlwr-5D.3,QTkw-6D.6,QSl-7D.4,QFlw-4D.1,QFlw-4D.2,QFlw-6D.2,QHd-4D.1,QFd-4D.2,andQFd-7D.1)was less than 1%,indicating that these QTL were almost uninfluenced by environments.For the major QTLQKl-2D.5for KL,which was also significant for KLWR(QKlwr-2D.5) and SL (QSl-2D.2),the environmental effect ranged from 2.40% to 2.89%,indicating that the QTL was less affected by environments (Table S7).

3.6.QTL clusters and pleiotropic loci

To profile the QTL hot-spot regions distributed on the chromosomes,seven QTL clusters containing at least two environmentally stable QTL for different traits were identified on chromosomes 1D,2D,6D(2),and 7D(3)(Table 1).Cluster 2 on chromosome 2D contained six environmentally stable QTL and showed a positive effect with the T015 allele(Table 1).Specifically,this cluster was related to KLWR,KL,and HD in nine environments,to PH and SL in five environments,and to FD in eight environments.Three clusters were detected on chromosome 7D.The first one,Cluster 7,contained five environmentally stable QTL.This cluster was identified for FD,SL (2),and KL in five environments,and for HD in six environments.The second one,Cluster 5,was significant for PH in six environments,and for TSS in five environments.The third one,Cluster 6,was significant for HD and FD in seven and eight environments,respectively,and showed positive effect with the T015 allele.Cluster 1 on chromosome 1D was significant for HD and FD in six and nine environments,respectively.Cluster 3 on chromosome 6D was significant for FLW and HD in eight and five environments,respectively.Cluster 4 on chromosome 6D was significant for KL and TKW in six and five environments,respectively.Additionally,some QTL showed pleiotropic effects on correlated traits.For example,QKl-2D.5(AX-89576820-AX-110276364)was identified for KL in seven environments,KLWR in nine environments (QKlwr-2D.5),SL in five environments (QSl-2D.2),KP in four environments (QKp-2D.2),TKW in two environments (QTkw-2D.2),and TSS in one environment (QTss-2D.1) (Table 1).

Table 1The seven QTL clusters harboring environmentally stable QTL for yield-related traits in T015/Zhou18 A-WI population.

Furthermore,some of the clusters were located on relatively small chromosome regions.For example,by blasting against the genome sequences of Chinese Spring (RefSeq v1.1),we located six QTLQFd-2D.2,QKlwr-2D.5,QHd-2D.1,QKl-2D.5,QSl-2D.2,andQPh-2D.4of Cluster 2,which showed a positive effect with the T015 alleles,in the intervals 32.8-47.9 Mb,26.3-28.6 Mb,32.8-47.9 Mb,22.2-24.5 Mb,22.2-24.5 Mb,and 29.2-32.0 Mb in chro-mosome 2D,respectively.Collectively,Cluster 2 was located in the 22.2-47.9 Mb of 2D and was speculated to be a QTL hot-spot region.

3.7.Prediction of candidate genes of QKl-2D.5

QKl-2D.5was used as an example to identify the potential candidate genes of the environmentally stable QTL and confirm their genetic effects.This QTL was identified for KL in seven environments and its positive effect was contributed by the T015 allele(Fig.5A).The confidence interval ofQKl-2D.5was flanked by SNP markersAX-89576820andAX-110276364(Fig.5B),corresponding to a physical distance of～2.6 Mb (23,757,314-26,401,714 bp,Ae.tauschiiT093),which contained 52 high-confidence annotated genes(Table S8).Of these,AetT093_2Dv1G100900.1was annotated as an ortholog of a rice gene for kernel size[26,27].Analysis of the complete coding sequences (CDSs) of T015 and Zhou18 revealed six nonsynonymous nucleotide mutations at positions +52,+100,+1709,+2438,+2809,and +4673 (Fig.5C,D).Protein sequence alignment showed that these SNPs caused the replacement of a glycine (G) with a serine (S);a proline (P) with an alanine (A);a valine(V)with an asparticacid(D);a glutamine(Q)with a glutamicacid(E);a histidine(H)with an arginine(R);and a valine(V)with an isoleucine(I).Moreover,three of the six SNPs were successfully transferred into KASP markers to genotype the T015/Zhou18 A-WI population (Table S9).All these KASP markers showed cosegregation in the A-WI population and two-tailedt-test analysis indicated that the A-WI lines with the T015 haplotype had significantly(P＜0.01)longer KL than those carrying the Zhou18 haplotype across all the 11 environments (Fig.5E).

Fig.5.QTL mapping and candidate gene analysis of QKl-2D.5.(A)QTL mapping of QKl-2D.5 for kernel length(KL)in seven environments.(B)Genetic map of the interval on 2D containing QKl-2D.5.The red line represents the genetic location of the candidate gene.(C)Schematic diagram of the candidate gene.Red bars represent the relative position of the nucleotide polymorphisms.Exons are indicated by black boxes,flanking regions and introns are indicated by solid black lines.(D)Sequence alignment of the candidate gene showing six nucleotide polymorphisms between T015 and Zhou18.(E) Effect of the candidate gene after dividing the T015/Zhou18 A-WI population into two classes based on KASP markers.**, P ＜0.01.

3.8.Four primary elite A-WI lines valuable for wheat breeding programs

To provide user-friendly materials for yield improvement in wheat breeding,we set standards of pH between 70 and 80 cm,TKW greater than 50 g,KNS greater than 60,TN greater than 15,and FD less than 189 d to screen the A-WI lines.Four primary elite A-WI lines were identified,which had higher KL,KLWR,KP,and TKW,more TN,and earlier HD and FD than their receptor parent Zhou18 (Fig.6A,B;Table S10).Genetic background analysis showed that the four lines carried most of Zhou18-specific markers and a limited number of T015-specific markers (Fig.6C).Furthermore,we analyzed the QTL harbored in the four lines.Line TZ9 possessed seven QTL derived from T015,includingQKlwr-2D.5,QKlwr-2D.7,QKl-2D.5,QKp-2D.1,andQTkw-2D.1for kernel-related traits,QTn-1D.1for TN,andQHd-3D.4for HD.Line TZ26 carried five T015-derived QTL,includingQKlwr-2D.7,QKp-2D.1,andQTkw-2D.1for kernel-related traits,QTn-1D.1for TN,andQFd-6D.5for FD.Line TZ94 possessed six QTL derived from T015,including

Fig.6.Anchoring four elite A-WI lines based on evaluation of the T015/Zhou18 A-WI population.(A)Anchoring elite lines based on plant height(PH),thousand-kernel weight(TKW),kernel number per spike (KNS),tiller number (TN),and flowering date (FD).(B) Heatmap and cluster representing phenotypic variation z-scores using the phenotyping data matrix of the four elite T015/Zhou18 A-WI lines and Zhou18.The blue and pink bars indicate high and low phenotype values,respectively.PH,KL,KLWR,TN,KP,TKW,HD,KW,KNS,FLL,SL,FLW,TSS,FD,and FSS are the same as in Fig.2.(C) Chromosome background of four elite lines colored according to T015 and Zhou18-specific SNP markers.

QKlwr-2D.5,QKl-2D.5,QKp-2D.1,andQTkw-2D.1for kernel-related traits,QTn-1D.1for TN,andQFd-6D.5for FD.Line TZ220 possessed seven T015-derived QTL,includingQKlwr-2D.5,QKlwr-2D.7,QKl-2D.5,QKp-2D.1,andQTkw-2D.1for kernel-related traits,QTn-1D.1for TN,andQFd-6D.5for FD.These results show that successful introduction of yield-related loci fromAe.tauschiiT015 contributed to the improvement of modern wheat cultivars.

4.Discussion

4.1.Improved genetic map facilitates QTL mapping for the A-WI population

Previously,a number of genetic maps of wheat have been reported and successfully applied to identify QTL for various yield-related traits [25,28-30].However,some genetic maps were constructed using PCR-based molecular markers with restricted marker density and the QTL were mapped within relatively large confidence intervals [31],greatly limiting the applications of the detected QTL in wheat breeding programs.With the development of next-generation sequencing technology and low-cost genome sequencing,high-density maps were reported using the highthroughput microarray genotyping method,such as the Wheat 9 K,90 K,and 660 K SNP arrays [31-33].These maps were then used for QTL mapping of kernel-related traits,spike-related traits,and KNS in common wheat.Due to the flanking SNP markers with accurate physical positions,the QTL detected in these studies could be efficiently used for comparative analysis and candidate gene prediction [31,33].We constructed a high-density genetic map for the T015/Zhou18 A-WI population based on the Wheat 55 K SNP array.Using this genetic map,we detected a total of 346 QTL for 15 yield-related traits in 11 environments.Of these,27 environmentally stable QTL were identified in at least five environments,explaining up to 24.27%of the phenotypic variations.These QTL and their flanking markers will be valuable for marker-assisted selection in wheat high-yield breeding programs.Zhang et al.[21]genotyped the same T015/Zhou18 A-WI population using 261 simple sequence repeat (SSR) markers and detected only 14 QTL for the agronomic traits of TKW,SL,and PH.These results indicated that a high-density genetic map based on SNP arrays with abundant markers significantly improved the efficiency of QTL mapping.

Additionally,the genetic map of chromosome 4D in this study had only five unique markers.Similar results were reported by multiple studies using diverse bi-parental populations and different types of markers [20,33-35].Moreover,Pont et al.[36] used a wheat exome-based target enrichment sequencing assay to capture variations of 435 wheat genotypes,and detected 595,939 high-confidence genetic variants.The fewest variants were observed on chromosome 4D with a total of 2341,which was much lower than the other six D chromosomes,further indicating the lower genetic diversity of chromosome 4D.Remarkably,although the total number (15) of QTL detected on 4D was minimal,four environmentally stable QTL were identified.These results suggested that the low diversity of a chromosome did not mean that it has no important QTL.

4.2.Comparison of the major QTL with previous observations

We detected QTL for 15 yield-related traits in 11 environments using the T015/Zhou18 A-WI population.Among these QTL,QKl-2D.5(AX-89576820-AX-110276364) was detected for KL in seven environments,KP in four environments (QKp-2D.2),KLWR in nine environments (QKlwr-2D.5),and TKW in two environments(QTkw-2D.2) (Table S5).The QTL was in the interval 22.2-24.9 Mb on chromosome 2D based on the genome sequence of cultivar Chinese Spring(RefSeq v1.1),which corresponded to reported QTL for KL,KP,and kernel dimension in the interval 15.6-23.0 Mb on chromosome 2D [37].Previous studies have mapped QTL for kernel weight at the same chromosome region [38-40].Huang et al.[41] identified a QTL for kernel weight in the interval 19.6-23.0 Mb on chromosome 2D and investigated its effects on KL and KW.These studies showed that the region ofQKl-2D.3was a major locus affecting kernel weight and size.

Moreover,QKl-2D.5was significant for SL in five environments(QSl-2D.2) (Table S4).The famousRhtgeneRht8was reported to be located on the same genomic interval,which could reduce PH and SL significantly [42,43].A reported major QTL for SL on chromosome 2DS was identified in the Nanda 2419 × Wangshuibai RIL population and was precisely mapped near the position 23.0 Mb[44,45].In the similar region,many previous studies have reported QTL for SL in different genetic backgrounds[28,39,46,47].These results indicated thatQKl-2D.5/QSl-2D.2derived from T015 was an important locus affecting kernel-related traits and spike length.It could be a promising locus for wheat yield improvement.

For the flowering time-related traits,QHd-2D.1/QFd-2D.2(AX-109838359-AX-109755068) was significant for HD and FD in eight to nine environments and located in the interval 32.8-47.9 Mb of 2D.A reported QTL for photoperiod sensitivity was identified in the interval 44.7-48.2 Mb on chromosome 2D in a wheat RIL population[48].In the similar position,two QTL for HD were detected by QTLmeta-analysis [49].More importantly,previous studies reported that the photoperiod genePpd-D1was located at 33.95 Mb on chromosome 2D [50-52].Therefore,QHd-2D.1/QFd-2D.2containing thePpd-D1gene is a major locus for HD and FD in wheat.QHd-7D.1/QFd-7D.1(AX-110422885-AX-94755613) was identified for HD and FD in seven to eight environments and located in the interval 63.9-79.6 Mb on chromosome 7D.VRND3/FT-D1,a gene related to vernalization and flowering,was located at 68.4 Mb on chromosome 7D and allelic variation was associated with large differences in flowering time [53].QHd-6D.3(AX-111263460-AX-110553854) was detected for HD in five environments and located in the interval 378.7-390.6 Mb on chromosome 6D,which is at the same position asTaHD1-6Dreported at 379.6 Mb of chromosome 6D [54].QHd-7D.5/QFd-7D.6(AX-111586126-AX-111606484) was detected for HD and FD in five to six environments and located in the interval 59.9-70.1 Mb on chromosome 7D.The QTL was overlapped with a SNP cluster for HD and FD in the interval 66.2-70.2 Mb [55].QHd-1D.2/QFd-1D.1(AX-111306420-AX-109972702) was significant for HD and FD in six to nine environments and located in the interval 449.2-457.2 Mb on chromosome 1D,which is located at similar position to the QTL for HD reported in the interval 455.8-482.2 Mb on chromosome 1D [56].

In addition to the QTL mentioned above,some QTL were also reported in previous studies,such asQKlwr-5D.3[57],QKl-7D.4[58],andQPh-2D.4[59].These results showed that the environmentally stable QTL detected using the T015/Zhou18 A-WI population were also significant in different genetic backgrounds,and these QTL with linkage markers are valuable genetic resources for wheat breeding.Several environmentally stable QTL (QFlw-6D.2,QKl-7D.4,QKlwr-7D.8,QSl-7D.3,QSl-7D.4,andQTss-7D.1)have not been reported in previous studies and might be novel QTL,which deserve further study including positional cloning and marker-assisted selection for wheat breeding programs.

4.3.Utilization of Ae.tauschii introgression lines for wheat improvement

With rapid growth of the global population,wheat yield must be sustainably increased to ensure food security.One strategy for improving productivity is to utilize the wild relatives of wheat to increase diversity in the breeding pool.Ae.tauschiipossesses many desirable genes and could be used for wheat genetic improvement.Direct and indirect approaches could be used to develop wheat accessions with improved traits fromAe.tauschii.The indirect approach requires crossingAe.tauschiito tetraploid wheat and may introduce unwanted loci from the A and B genomes,while the direct approach introduces genetic variation from only the D genome by direct crossing between diploid and hexaploid species.Previously,curl mite resistance,PHS resistance,and grain size of wheat accessions were successfully improved through direct approach [6,7,10].

A total of 27 environmentally stable QTL were anchored using the A-WI population derived by crossingAe.tauschiiaccession T015 to wheat cultivar Zhou18.Of these,20(74%)stable QTL showed a positive effect with the T015 allele,suggesting the contribution of the additive effect of the introgressed fragments.Four primary elite AWI lines were screened and successfully anchored.Compared with Zhou18,these lines had better performance in kernel size,TKW,TN,HD,and FD,which were highly consistent with their corresponding T015-derived QTL and showed the successful transfer of several loci for important yield-related traits fromAe.tauschiiT015 to Zhou18.These four A-WI lines with good trait performance were valuable for wheat breeding improvement.For example,ZT9 could be used for longer KL,more TN,and earlier HD and FD;ZT26 could be used for longer KL,higher TKW,more TN,and earlier FD;ZT94 could be used for higher TKW,more TN,and earlier FD;and ZT220 could be used for longer KL,higher TKW,and more TN.In addition to the four elite lines,several other A-WI lines showed significantly superior performance to Zhou18 in at least one trait,such as higher TKW,more KNS,or earlier FD.All these lines showed great potential to facilitate gene transformation for wheat breeding.Furthermore,we successfully developed anAe.tauschii-wheat SOW pool derived from more than 80 differentAe.tauschiiaccessions and developed the corresponding A-WI populations [6],allowing effective introduction of beneficial variations fromAe.tauschiito wheat.

5.Conclusions

We constructed a high-density genetic map using an A-WI population with the Wheat 55K SNP array and conducted QTL mapping for 15 yield-related traits in 11 environments.Twenty-seven environmentally stable QTL were identified in at least five environments,20 of which were derived fromAe.tauschiiT015,explaining up to 24.27% of the phenotypic variations.The major QTLQKl-2D.5for KL was detected in seven environments and showed a positive effect with the T015 allele.AetT093_2Dv1G100900.1was speculated to be the candidate gene forQKl-2D.5.Four primary valuable introgression lines with good trait performance and carrying yield-related loci derived fromAe.tauschiiT015 were selected for improving wheat in breeding programs.These findings could facilitate efficient transfer of beneficial genes fromAe.tauschiiinto elite wheat germplasm for wheat improvement.

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.

CRediT authorship contribution statement

Feifei Ma:Data curation,Software,Visualization,Writing -original draft.Ranzhe Li:Data curation,Formal analysis,Visualization,Validation.Guanghui Guo:Data curation,Methodology,Funding acquisition.Fang Nie:Data curation,Investigation,Formal analysis.Lele Zhu:Data curation,Investigation,Formal analysis.Wenjuan Liu:Data curation,Investigation.Linlin Lyu:Data curation,Investigation,Formal analysis.Shenglong Bai:Data curation,Investigation,Visualization.Xinpeng Zhao:Data curation,Investigation.Zheng Li:Data curation,Validation.Dale Zhang:Resources.Hao Li:Resources,Data curation,Funding acquisition,Project administration,Writing-review&editing.Suoping Li:Resources,Funding acquisition.Yun Zhou:Resources,Project administration,Writing -review &editing.Chun-Peng Song:Resources,Project administration,Writing -review &editing.

Acknowledgments

This research was financially supported by the National Natural Science Foundation of China(32230079,32001492,31871615,and 31901547) and Natural Science Foundation of Henan Province(222301420102).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2023.05.001.