Yongqing Yang,Qiao Su,Yurong Li,Zengshu Cheng,Yahui Song,Xinxin Jin,Jin Wang

Institute of Cereal and Oil Crops,Hebei Academy of Agricultural and Forestry Sciences,the Key Laboratory of Crop Genetics and Breeding of Hebei,Shijiazhuang 050035,Hebei,China

Keywords:Arachis hypogaea Candidate gene E3 ubiquitin ligase Yield-related QTL region

ABSTRACT High yield is a major objective for peanut (Arachis hypogaea L.) breeding worldwide.However,fewer yield-related quantitative trait loci (QTL) have been reported in peanut than in other staple food crops such as rice(Oryza sativa),wheat(Triticum aestivum),and maize(Zea mays).This study aimed to identify stable major-effect QTL associated with pod yield per plant,hundred-pod weight for double-seeded pods,hundred-seed weight,shelling percentage,and pod number per plant,allowing us to predict candidate genes by means of transcriptome and genome sequencing.To this end,we used a population of recombinant inbred lines comprising 192 F9:11 families derived from a JH6×KX01-6 cross to construct a highresolution genetic map (1705.7 cM) consisting of 2273 polymorphic SNPs,with 0.75 cM (on average)between adjacent SNPs.We identified two high-confidence,yield-related QTL, qHYF_A08 and qHYF_B06,explaining 5.78%-31.40% of phenotypic variation and with LOD values of 5.10-24.48,in six environments. qHYF_A08 mainly explained the variation in shelling percentage,whereas qHYF_B06 explained variation in hundred-pod weight and hundred-seed weight and accounted for 8.77%-31.40%of the variation in effective pod number per plant,pod number per plant,and shelling percentage.We narrowed down qHYF_B06 to an 890-kb interval using an advanced mapping population.Transcriptome and genome analyses revealed that only Arahy.129FS0 and Arahy.3R9A5K in the candidate mapping interval were differentially expressed between JH6 and KX01-6,with substantial structural variations in their promoter and coding regions.Genotypes of 208 peanut accessions determined using a diagnostic CAPS marker suggested that the two haplotypes of Arahy.3R9A5K were highly associated with hundred-seed weight and hundred-pod weight;this diagnostic CAPs marker could therefore be useful for selecting high-yielding lines during peanut breeding.Overall,our results provide valuable information for cloning alleles with favorable effects on peanut yield.

1.Introduction

Peanut (Arachis hypogaeaL.) is a widely consumed food crop with high seed protein and oil contents.It is cultivated in ＞100 countries,producing a total annual yield of approximately 50 Mt(https://faostat.fao.org/).However,the global demand for peanut is increasing rapidly.If population growth continues at the current rate,the limited arable land worldwide will be insufficient for producing enough food (including peanut) to satisfy human needs by 2050[1].Breeding high-yielding cultivars and optimization of field management practices are realistic strategies for increasing peanut yields to meet human demand[2].Consequently,yield and quality traits have long been targeted by peanut breeding programs worldwide.However,as in most crops,yield and quality are complex quantitative traits affected by genetic and environmental factors[3-7].Environmental conditions can significantly affect a particular trait,which is problematic for breeders.For example,low heritability (0.3-0.5) of shelling percentage (SP) was recorded in 160 advanced breeding lines over six seasons [8],with a combined analysis of variance suggesting that the environment was the main contributor to total variation.Indeed,pod yield,seed yield,oil yield,and oil content differ significantly between rainy and postrainy seasons [8].Therefore,highly heritable traits are typically used for selective breeding because they tend to be more predictable than low-heritability traits.

Compared to other staple food and oil crops,such as rice(Oryza sativa)and soybean(Glycine max),exploration of quantitative trait loci (QTL) and cloning of functional genes in peanut have lagged.The PeanutBase database (https://peanutbase.org) lists ＞90 QTL associated with yield,pod,and seed traits,identified mainly using simple sequence repeat(SSR)markers[9-12].Limited markers and their uneven distribution across the genome greatly restrict the efficiency of map-based cloning in peanut.High-throughput sequencing has recently been introduced for gene discovery in peanut,enabling rapid identification and fine mapping of candidate genes,particularly for yield-related QTL,based on the substantial abundance of single nucleotide polymorphism (SNP) markers.For example,a major and stable QTL explaining 36%-66% of seed and pod trait variation was localized to a 1.5-Mb genomic region on chromosome A05 [13].From 27 QTL with high effects (74%-83%)on seed size and weight,two stable loci were narrowed down to 7.06-Mb and 12.21-Mb genomic regions on chromosomes 2 and 16,respectively [2].However,genes encoding proteins with functions that affect pods or seeds have not been reported.

In this study,we constructed a high-resolution genetic map based on a bi-parental recombinant inbred line (RIL) population comprising 192 F9:11families using a 48K SNP microarray.Our main objectives were (1) to identify stable major-effect loci,(2)to fine-map candidate genes by incorporating transcriptome and genome sequencing data of the two parental lines,and(3)develop and validate a gene-specific marker for selecting high-yield breeding lines.Our results provide a basis for cloning functional genes for increasing yields in peanut.

2.Materials and methods

2.1.Plant materials and field trials

Peanut parents ‘Jihua 6’ (JH6) and ‘Kaixuan 01-6’ (KX01-6),which differ in terms of yield-related traits,were used to construct a RIL population consisting of 192 F9lines via single-seed descent.JH6,a large-pod peanut cultivar with a low oleic acid content,was released in 2009 by the Institute of Industrial Crops,Hebei Academy of Agricultural and Forestry Sciences.KX01-6,a small-pod peanut cultivar that is extensively used as a backbone parent for breeding because of its high oleic acid content,was provided by the Kaifeng Academy of Agriculture and Forestry Sciences.An additional 208 peanut accessions used for gene-specific marker validation were provided by the Oil Crop Research Institute,Chinese Academy of Agricultural Sciences.The following peanut yieldrelated traits were included in correlation and QTL analyses: pod yield per plant (PYPP),hundred-pod weight (HPW) for doubleseeded pods,hundred-seed weight (HSW),shelling percentage(SP),total pod number per plant (PNPP;including rotten pods resulting from disease or unknown reasons),and healthy pod number per plant without rotten pods (i.e.,effective pod number per plant,EPNPP).The RIL population was grown under six field environments,and yield-related traits were evaluated after harvest.The six environments were 2019 Dishang (E1),2019 Nongchang_3502 (E2),2020 Dishang (E3),2020 Nongchang_3502(E4),2021 Dishang (E5),and 2021 Nongchang_3502 (E6).Dishang(38.03°N,114.48°E) and Nongchang_3502 (38.21°N,114.31°E)experimental stations belong to the Institute of Cereal and Oil Crops,Hebei Academy of Agriculture and Forestry Sciences (Shijiazhuang,Hebei province,China).The HSW and HPW of the 208 peanut accessions were evaluated under routine field management conditions in 2022.Compound fertilizer (15:15:15 N:P2O5:K2O)was applied as the base fertilizer at a rate of 600 kg ha-1,but fertilizers were not applied during the peanut growth period.Water was supplied using porous pipes as needed during the peanut growth stage.The RILs and the two parents were cultivated using a complete randomized block design,with three replicates per genotype.Approximately 15 plants were grown in each row (3 m in length and separated by 0.4 m).

2.2.Analysis of yield-related traits

The pods on each plant were counted manually to determine PNPP.PYPP,HPW,and HSW were measured using a centesimal balance.SP was calculated as follows: weight of seed (kg)/weight of pods (kg) × 100%.The Performance Analytics package in R[14]was used for genetic analysis and traits correlation with Pearson method.The generalized heritability(H2)of yield-related traits in different environments was estimated using QTL IciMapping v4.1 [15] and the following formula:whereis the genetic variance component among the RILs,is the RIL × environment variance,is the residual variance,nis the number of environments,andris the number of replicates.Student’st-test in SPSS 19 [16] was used to determine the significance of yield-related traits.

2.3.Detection of genotypes

Genomic DNA was extracted from fresh leaf tissue collected from each RIL and the two parents and then genotyped using a peanut-specific Affymetrix genotyping array as described by Clevenger[17].Genotype data generated by GeneTitan were analyzed using Affymetrix Power Tools(e.g.,clustering and genotyping),and a final customized report (PED and MAP files) was generated for subsequent analysis.

2.4.Genetic linkage map construction,QTL analysis,and marker development

After removing partially segregating SNPs detected by Chisquared tests,the remaining markers were used to construct a genetic linkage map using IciMapping v4.1 [15] as previously described [18].Markers were grouped according to the following:(i)logarithm of the odds(LOD)values exceeding 3,(ii)recombination frequency ＞0.3,(iii) marker distance ＜50 cM,(iv) predefined group number,and(v)anchored marker information.Marker order was initially determined using nnTwoOpt algorithms and then checked using SER and RCORD methods.Composite interval mapping (CIM) was used for mapping QTL with data derived from six environments;the algorithm was implemented in the freely available package R/qtl [19].A permutation test was conducted with 1000 replications to determine the threshold for assessing QTL significance.Derived cleaved amplified polymorphic sequence(dCAPS) or CAPS markers based on closely linked polymorphic SNP markers flanking major-effect QTL were developed as described by Yang et al.[20].F5individuals with homozygous and heterozygous genotypes at the deduced loci were identified using dCAPS markers.Sub-F2:3and sub-F3:4progenies included in an advanced mapping population were derived from self-cross of heterozygous F5individuals and harvested individually.

2.5.Genetic variation analysis after resequencing of parental genomic DNA

The CTAB protocol[21] was used to extract genomic DNA from fresh leaves collected from JH6 and KX01-6 seedlings.A TruSeq Library Construction Kit was used to construct libraries for sequencing on an Illumina HiSeq platform.Default parameters of the BWA software (https://bio-bwa.sourceforge.net/) were used to align clean reads to the polyploid peanut ‘Tifrunner’ v1.0 reference genome(https://phytozome-next.jgi.doe.gov/).SAMtools and Picard(https://picard.sourceforge.net)were used to remove duplicated reads,call variants,and screen for high-confidence SNPs and insertions/deletions (InDels).The following criteria were used to identify high-confidence SNPs and InDels: mapping quality ＞10;variant position depth ＞5;and homozygous genotype at the variant position.

2.6.RNA extraction,library construction,sequencing,and data analysis

TRIzol reagent (Invitrogen,Carlsbad,CA,USA) was used to extract RNA from JH6 and KX01-6 seed at the late pod-filling stage.Oligo-(dT)beads were used to enrich for mRNA,which was cut into short fragments in fragmentation buffer.An Illumina NEBNext Ultra RNA Library Prep Kit (New England Biolabs,Ipswich,MA,USA) was used to synthesize cDNA from mRNA.cDNA libraries were sequenced on an Illumina NovaSeq 6000 system by Gene Denovo Biotechnology Co.,Ltd.(Guangzhou,Guangdong,China).After constructing the reference genome index,HISAT2.2.4 was used to map paired-end clean reads to theTifrunner v1.0reference genome[22];default values were used for the‘-rna-strandness RF’option and all other parameters.Fragments per kilobase of exon model per million mapped fragments (FPKM) values were used to quantify gene expression.Specifically,FPKM values were calculated in RSEM [23] as follows: FPKM=total exon fragments/[mapped reads(million)×exon length(kb)].The DESeq2 program was used to identify differentially expressed genes according to the following criteria: false discovery rate ＜0.05 and absolute foldchange ≥2.

3.Results

3.1.Phenotypic evaluation of parents and RILs

There were significant differences in the appearance of the two parents (Table 1;Fig.S1),with PYPP,HPW,HSW,SP,PNPP,and EPNPP differing between JH6 and KX01-6.In addition,JH6 had a higher yield(159.9%)than KX01-6 because it produced more pods and larger seeds.HPW,HSW,PNPP,and EPNPP were 88.3%,90.6%,36.4%,and 39.3%higher for JH6 than for KX01-6.However,SP was significantly lower(2.7%)for JH6 than for KX01-6(P＜0.001).These clear differences between the two parental lines were deemed sufficient for determining the genetic mechanism underlying peanut yield increases.

Table 1Summary information of phenotypic variation and genetic analysis for the RIL population across six field environments.

We observed transgressive segregation among the phenotypes collected from the 192 RILs in six environments,with the mean value of each trait falling within the parental range.The generalized heritability (H2) was 0.82-0.94 for the six traits (Tables 1,S1).These results indicate that the six traits are mainly affected by genetic factors and that the phenotyping data are suitable for QTL analysis.With the exception of SP,the traits analyzed followed normal distribution curves;their absolute kurtosis and skewness values calculated using data collected from six environments were ＜1,strongly suggesting that these traits are controlled by multiple loci.Because most of the observed traits (except for SP)varied significantly between the parents,the yearly coefficient of variation (CV) for these traits also varied significantly (18.14-30.51);the CV for SP varied slightly (3.83-6.24) (Table S1).These findings strongly suggest that genetic variation is responsible for the observed traits.

3.2.Correlation analysis of the peanut yield traits

Correlation analysis of the yield-related traits in our RIL population(Fig.1)revealed that PYPP is significantly correlated with most of the pod number and weight traits,except for SP.The weight traits HPW and HSW were positively correlated with PYPP,with correlation coefficient values of 0.37 (P＜0.001),whereas the correlation coefficient between PYPP and pod number traits(i.e.PNPP and EPNPP) were relatively high,with values of 0.75 and 0.70(P＜0.001),respectively.These correlation coefficient values suggested that pod number traits may be more important than weight traits for increasing peanut yield.Notably,although SP is considered to be an important trait influencing peanut yield[24],our correlation analysis of yield-related traits in the RIL population did not support this notion,likely because there is a lack of significant genetic variation underlying the SP trait in this RIL population and therefore SP was easily affected by other traits.

Fig.1.Correlation analysis among the 6 observed traits.EPNPP,effective pod number per plant;PNPP,pod number per plant;PYPP,pod yield per plant;HPW,hundred-pod weight;HSW,hundred-seed weight;SP,shelling percentage;The histograms with fitting curve of traits are put in diagonal.Above the diagonal is correlation coefficient with significant level,below the diagonal is scatter plot with fitting curve.**, P ＜0.01;***, P ＜0.001 (Student’s t-test).

3.3.Construction of genetic linkage maps

After a stringent screening,we selected 2273 recombinant markers that were polymorphic between the two parents and distributed over the 20 peanut chromosomes.We used these markers to construct a genetic linkage map (1705.7 cM with an average of 85.3 cM per linkage group).Chromosome A03 had the most polymorphic markers (258),spanning 120.1 cM of the chromosome,whereas chromosome B03 had the fewest markers (27),spanning 113.1 cM of the chromosome.The average distance between adjacent SNPs was 0.75 cM,ranging from 0.35 cM for B10 to 4.19 cM for B03 (Table S2;Fig.S2).

3.4.Identification of high-confidence QTL for yield-related traits

We detected 14 significant QTL for the six traits.These QTL had LOD scores of 3.18-24.48 and accounted for 3.39%-31.40% of the phenotypic variation among the 192 F9peanut RILs grown in six field environments(Table S3).Among these,QTL on chromosomes A08 and B06,namedqHYF_A08andqHYF_B06,respectively,had the most stable effects on the observed traits,as we detected them under all six environments.These two loci explained the major variation in the yield-related traits(Table 2).Specifically,qHYF_A08was mainly associated with variation in SP and HPW,with percentage of variance explained (PVE) values ranging from 5.78% to 23.20% and LOD scores between 5.10 and 15.40.By contrast,qHYF_B06was mainly related to variation for HPW and HSW,with PVE ranging from 13.38% to 31.29%,as well as variation in EPNPP,PNPP,and SP (PVE of 8.77%-31.40%).These results strongly indicate that yield-related traits are controlled genetically by two major QTL as well as many minor QTL in our RIL population.

Table 2High confident quantitative trait loci (QTL) for yield and yield-related traits in six environments.

3.5.Fine mapping and prediction of candidate genes for qHYF_B06

BecauseqHYF_A08was mainly associated with variation in SP but poorly associated with variation in PYPP,therefore,we mainly performed fine mapping ofqHYF_A06.Given the high LOD and PVE values ofqHYF_B06on HPW and HSW at the six environments,we used the mean HPW and HSW values for the RIL population to detect candidate genes forqHYF_B06via map-based cloning.Genetic analysis showed that the 95% confidence intervals for HSW and HPW at theqHYF_B06locus define a genomic region corresponding to a genetic distance of 116.2-120.5 cM (Fig.2A).Tovalidate the QTL results and more precisely map the genes related to the target weight-related traits,we generated an advanced mapping population from a single F5residual heterozygous plant that was heterozygous atqHYF_B06.We inferred the crossover points from genotypes at dCAPs markers M1-M5 (Table S4),which covered the 95% confidence interval ofqHYF_B06,determined by examining seed-size segregation patterns among F2:3and F3:4progenies.Analysis of progenies derived from five recombinant homozygous plants revealed that those with a genotype in the M3-M4 region consistent with that of KX01-6 have relatively small seeds (i.e.,HSW ＜70 g).By contrast,progenies with a genotype in the M3-M4 region consistent with that of JH6 produced larger seeds (i.e.,HSW ＞83 g).Basing on these results,we delimitedqHYF_B06to an approximately 890-kb interval (genetic distance of approximately 0.3 cM) between markers M3 and M4 (Fig.2B).

Fig.2.Fine mapping of qHYF_B06.JH6,Jihua 6;KX01-6,Kaixuan 01-6.(A) The qHYF_B06 locus was mapped into a 4.3 cM genetic region on chromosome B06 using mean values of HPW and HSW for the RIL population which observed under six environments.The separate QTL are represented by bars(1-LOD interval)and extended lines(2-LOD interval).(B) Segregation of seed size in the sub-F2:3 inbred family derived from a single F5 residual heterozygous individual plants.Black,white and grey boxes indicate homozygosity for the allele from the JH6 and KX01-6 parents,and heterozygosity,respectively (C)The delimited 890-kb genomic region for qHYF_B06 contains 1 predicted gene with a significant structural variation between parents.

To identify the functional genes inqHYF_B06,we detected genomic and transcriptomic variation in the 890-kb interval via highthroughput sequencing.According to the annotated Tifrunner v1.0 genome in Phytozome 13 (https://phytozome.jgi.doe.gov/pz/portal.html),this region contains 50 annotated genes.However,29 of the 50 annotated genes were barely expressed in peanut seeds,suggesting that they are unlikely to be involved in seed development (Table S5).Of the remaining annotated genes,12 and 9 were respectively expressed at relatively low and high levels in seeds (Fig.S3).Transcriptome and genome analyses detected two and five annotated genes with differential expression between JH6 and KX01-6 at the 5% and 1% significance levels,respectively,but only Arahy.129FS0 and Arahy.3R9A5K showed substantial structural variation in their promoter and coding regions.Therefore,we considered Arahy.129FS0 and Arahy.3R9A5K as candidate genes forqHYF_B06because of the InDels present in these genes.A 1-bp insertion between nucleotides 657 and 658 of Arahy.3R9A5K in KX01-6 may be a loss-of-function mutation because of the associated incorrect translation from amino acid p.M219fs onwards(Fig.2C).An InDel in the predicted Arahy.129FS0 promoter sequence may explain differential expression in JH6 and KX01-6 seeds (Fig.2C).These findings strongly suggest that one or both of these candidate genes may control the seed-size trait associated withqHYF_B06.

3.6.Development and validation of a gene-specific marker

To evaluate the application value of the strong candidate gene Arahy.3R9A5K for breeding programs,we developed a genespecific CAPS marker that can specifically recognize the 1-bp insertion between nucleotides 657 and 658 of Arahy.3R9A5K.Use of this marker to determine the genotypes of 208 peanut accessions revealed that 131 and 77 peanut accessions harbor the homozygous genotypes G/G and -/-,respectively (Table S6).The HSW of the two genotype groups ranged from 24.5 g to 93.0 g and from 21.3 g to 101.5 g,with means of 57.8 g and 71.6 g,respectively(Fig.3A).The HPW of the two genotype groups ranged from 64.9 g to 260.8 g and from 62.8 g to 348.5 g,with means of163.1 g and 203.3 g,respectively (Fig.3B).One-way analysis of variance (ANOVA) suggested that the HSW and HPW of peanut accessions harboring the-/-genotype are significantly higher than those of peanut accessions harboring the G/G genotype(P＜0.001).Therefore,these results suggest that the two haplotypes at Arahy.3R9A5K are highly associated with the observed traits and that the diagnostic CAPS marker could be useful for selecting high-yield lines during the peanut breeding process.

Fig.3.Phenotypic difference between two genotypes of Arahy.3R9A5K in 208 peanut accessions.Phenotypic difference of hundred-seed weight(A)and hundred-pod weight(B).The G/G and-/-homozygous alleles were consistent with KX01-6 and JH6,respectively.***,P ＜0.001;‘‘n”means peanut accession number.JH6,Jihua 6;KX01-6,Kaixuan 01-6.

4.Discussion

Genetic diversity and genomic recombination are the main factors underlying the development of outstanding cultivars through traditional crop breeding [25-27].Crossing two outstanding varieties to combine alleles for ideal traits remains the primary method for breeding new cultivars.Unfortunately,if the genetic background of the parents is unclear,limited genetic diversity can make this procedure inefficient and time-consuming.Therefore,the genetic background of a given trait should be characterized before attempting to improve the trait.Increasing yield and quality has long been a high priority for crop breeders;however,yield and quality traits are complex and easily affected by environmental conditions,especially yield traits [5,28].Breaking complex traits into simpler related traits that are relatively unaffected by the environment will help breeders to define the criteria used for selecting outstanding cultivars.In this study,we selected six yield-related traits with a relatively highH2(0.82-0.94) according to observations recorded in six environments.These traits were only slightly affected by the environment,making them suitable candidate traits for breeding.Correlation analysis revealed that the yield-related traits PNPP,EPNPP,HSW,and HPW were significantly associated with PYPP.These four relatively simple traits not easily affected by the environment may therefore be useful for selecting outstanding cultivars during breeding.

Shelling percentage (i.e.,ratio of the seed weight to the pod weight) varies significantly among peanut cultivars and breeding populations.More specifically,SP varied from 40.5% to 80% and from 65%to 83%in two independent and unrelated RIL populations comprising 816 and 195 lines,respectively,in multiple environments [24,29].Considering its importance,SP was unsurprisingly significantly associated with seed yield and was previously identified as an important trait for optimizing yield[30,31].Notably,our correlation analysis did not find a significant association between SP and PYPP.This apparent inconsistency between our findings and those of earlier studies may be explained as follows:1)the relatively small genetic variation for SP,as indicated by 4.84%phenotypic variation in the RIL population,may result from environmental changes;or 2) the minor effect of SP on yield,as indicated by 28.36%phenotypic variation for PYPP and 4.84%phenotypic variation for SP in the RIL population,suggests that PYPP is primarily influenced by other highly variable yield-related traits(e.g.,HPW and HSW),and differences caused by these yieldrelated traits may obscure the correlation between SP and PYPP.Therefore,although our results did not confirm that SP significantly affects peanut yield,we still consider SP to be a critical trait for breeding peanut varieties with increased yields.

Peanut is an allopolyploid that likely originated from hybridization between two wild diploid species,Arachis duranensisandArachis ipaensis,followed by chromosome doubling [32,33].The relatively recent divergence ofA.duranensisandA.ipaensisapproximately 2.5-3.5 million years ago [34] resulted in the peanut subgenomes being very similar[35].Moreover,peanut has a very narrow genetic base,with relatively few genetic variants among cultivars.This lack of variability greatly impedes the development of polymorphic markers useful for constructing SNP-based saturated genetic maps and may help explain the uneven distribution of genetic markers in peanut genetic maps [36-38].For example,Hu et al.[39]identified 2334 polymorphic markers on 20 chromosomes,but these are unequally distributed among linkage groups(13-369 per linkage group).Unfortunately,we encountered similar problems in our study.For example,four linkage groups(A01,A10,B03,and B09) harbour ＜40 polymorphic markers,whereas A03,B04,and B08 contain ＞200 polymorphic markers.Hence,valuable genes were likely missed because of a lack of closely linked polymorphic markers.

Cultivated peanut originated more recently than other oil crops[32] and has been the focus of fewer studies relevant to modern breeding programs,especially analyses of important genes using map-based cloning strategies.Nevertheless,some of the previously reported findings regarding QTL for yield-related traits are consistent with our results.For example,Li et al.[31]used an SSR-based genetic map and detected 15 QTL that explained 3.4%-31.7%of the phenotypic variation in four environments.One of these QTL(qSPA08.3) spanned a physical interval from 32.35 Mb to 49.40 Mb on chromosome A08,which overlaps with one of our QTL (qHYF_A08).However,possibly because of relatively low SSR marker density and a limited genetic background,Li et al.[31]consideredqSPA08.3to be a minor-effect locus owing to small additive effects (0.84-1.21) and relatively low LOD values (3.57-4.15).In the current study,LOD values forqHYF_A08were 5.10-15.40 and PVE ranged from 5.78% to 23.20%.Notably,by using high-density SNP markers,we delimitedqHYF_A08to a physical interval ＜5 Mb,which is important for map-based cloning of candidate genes.Several earlier studies have explored the genetic basis of yield-related traits[2,13,36,40],but none of the reported QTL share a common candidate region with the QTL identified in our study.For example,Wang et al.[36] detected a QTL on B06 (119.8-128.8 Mb),which is close toqHYF_B06(143.95-144.84 Mb)detected in this study.Chen et al.[40]reported two QTL associated with HSW and seed width on B06 (10.6-21.6 Mb and 12.2-74.9 Mb,respectively),which is far fromqHYF_B06.Thus,the lack of related QTL in the same genomic interval suggests thatqHYF_B06is a previously uncharacterized locus.

Although the low genomic diversity in cultivated peanut greatly limits the effectiveness of map-based cloning,genetic variation in large candidate intervals can be identified rapidly via highthroughput resequencing.Moreover,the number of candidate genes can be reduced by analyzing their expression patterns using transcriptome data.In the present study,we delimitedqHYF_B06to an interval covering approximately 890 kb and containing 50 annotated genes.Most of these genes were not expressed or lacked genetic variation in their coding or promoter regions that would result in a loss of function or functional impairment.We only identified two strong candidate genes associated withqHYF_B06,namely Arahy.129FS0,which encodes an IQ-domain 21 protein,and Arahy.3R9A5K,which encodes an E3 ubiquitin ligase.Because transgenic technology applicable for cultivated peanut has not been thoroughly established,we were unable to directly verify the identity of the functional gene;however,genes with similar domains have been identified in other crops.Earlier research confirmed that gibberellins are plant growth regulators associated with seed development,cell elongation and expansion,and germination [41-43],and that E3 ubiquitin ligases contribute to gibberellin perception and signal transduction [44-46].A previous study found that the RING-type E3 ubiquitin ligase BIG BROTHER is a central negative regulator of seed size inArabidopsis thaliana[47].In wheat (Triticum aestivum),TaGW2-6Aencodes a RING E3 ubiquitin ligase that negatively regulates grain size [48],possibly because the cross-talk between gibberellin and E3 ubiquitin ligase has critical effects on wheat grain development [49].In our study,phenotypic differences between two genotypes of Arahy.3R9A5K also supported Arahy.3R9A5K as being highly associated with HSW and HPW.Considered together,these findings suggest that the E3 ubiquitin ligase encoded by Arahy.3R9A5K is very likely responsible for controlling peanut seed and pod size.However,this hypothesis remains to be confirmed experimentally,and the underlying molecular mechanism will need to be elucidated.

5.Conclusions

We used a previously developed RIL population to construct a high-resolution genetic map containing 2273 SNP markers.We detected two stable QTL(qHYF_A08andqHYF_B06)across six environments,withqHYF_B06having the largest and most stable effect on HSW and HPW.We narrowed downqHYF_B06to an approximately 890-kb interval between markers M3 and M4 (0.3 cM genetic distance).By analyzing whole-genome resequencing and transcriptome data for the parents,JH6 and KX01-6,we identified a strong candidate gene (Arahy.3R9A5K) that is differentially expressed between the parents.The two haplotypes of Arahy.3R9A5K are highly associated with HSW and HPW,and our diagnostic CAPS marker could be used for selecting high-yielding lines during peanut breeding.Our findings may therefore be useful for cloning favorable alleles in peanut.

CRediT authorship contribution statement

Yongqing Yang:Formal analysis,Writing-Original Draft.Qiao Su:Investigation,Visualization.Yurong Li:Resources.Zengshu Cheng:Investigation.Yahui Song:Investigation.Xinxin Jin:Investigation.Jin Wang:Supervision,Writing-Review&Editing,Funding acquisition.

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 jointly supported by the Earmarked Fund for CARS-13,the Modern Agricultural Industrial Technology System of Hebei Province (HBCT2018090101 and HBCT2018090201),the Science and Technology Innovation Team of Modern Peanut Seed Industry (21326316D),the Technology Innovation Special Project(2022KJCXZX-LYS-11),the Basic Research Funds of Hebei Academy of Agriculture and Forestry Sciences(2021060201),and the Talents Construction Project of Science and Technology Innovation,Hebei Academy of Agriculture and Forestry Sciences (C22R0311).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2023.04.004.