Resequencing-based QTL mapping for yield and resistance traits reveals great potential of Oryza longistaminata in rice breeding

Weixiong Long,Nengwu Li,Jie Jin,Jie Wng,Dong Dn,Fengfeng Fn,b,Zhiyong Go,b,Shoqing Li,b,

a State Key Laboratory of Hybrid Rice,Key Laboratory for Research and Utilization of Heterosis in Indica Rice of Ministry of Agriculture,Engineering Research Center for Plant Biotechnology and Germplasm Utilization of Ministry of Education,College of Life Sciences,Wuhan University,Wuhan 430072,Hubei,China

b Hongshan Laboratory of Hubei Province,Wuhan 430072,Hubei,China

Keywords:Oryza longistaminata High yield High disease and pest resistance QTL

ABSTRACT As a natural genetic reservoir,wild rice contains many favorable alleles and mutations conferring high yield and resistance to biotic and abiotic stresses.However,there are few reports describing favorable genes or QTL from the AA genome wild rice O.longistaminata,which is characterized by tall and robust habit and long tassels and anthers and shows high potential for use in cultivated rice improvement.We constructed a stable BC2F20 backcross inbred line(BIL)population of 152 lines from the cross of 9311×O.longistaminat.Some BILs showed large panicles,large seeds,and strong resistance to rice false smut,bacterial leaf blight,rice blast spot,and brown planthopper.Genomic resequencing showed that the 152 BILs covered about 99.6% of the O.longistaminata genome.QTL mapping with 2432 bin markers revealed 13 QTL associated with seven yield traits and eight with resistance to brown planthopper and to four diseases.Of these QTL,12 for grain yield and 11 for pest and disease resistance are novel in Oryza species.A large-panicle NIL1880 line containing QTL qPB8.1 showed a nearly 50%increase in spikelet number and 27.5%in grain yield compared to the recurrent parent 9311.These findings support the potential value of O.longistaminata for cultivated rice improvement.

1.Introduction

Domestication and improvement of rice in the last thousand years have underlain the development of Asian civilization.Rice is now a staple food for over 50%of the world’s population,providing ＞23% of the calories consumed by humans [1].The growth of the human population will require the doubling of global agricultural production by 2050 [2].However,the rice genetic base has been narrowed by 6000 years of continuous selection by humans and by intensive breeding in recent centuries [3],resulting in a yield plateau in rice production and yield instability under biotic and abiotic stresses.These challenges call for the exploitation of new genetic resources from other plants,such as wild rice.

Wild rice species are a reservoir of many favorable genes that could be exploited to improve modern cultivars [3-5].Of the 24 species in theOryzagenus,the worldwide-grown Asian riceO.sativa,the African riceO.glaberrima,and the other six wild speciesO.rufipogon,O.nivara,O.barthii,O.longistaminata,O.meridionalis,andO.glumaepatulaall belong to the AA-genome group of the genusOryzaand descend from a common ancestor[6].They intercross readily and represent the primary gene pool for rice improvement.

In recent decades,efforts have been made to introgress favorable genes from AA-genome wild species into cultivars by interspecific hybridization [7,8].Many agronomic genes,including resistance to nematodes and African gall midge fromO.glaberrima[9],brown planthopper resistance fromO.officinalis[10],high grain yield [7,8,11,12],and cytoplasmic male sterility fromO.rufipogon[13],have been identified and transferred to cultivated rice since the first report[5]of introgression of grassy stunt virus resistance fromO.nivarain 1977.

Oryza longistaminata is a perennial African AA-genome wild rice with self-incompatibility,long anthers,and strong tolerance to biotic and abiotic stresses [14].Few successful introgressions fromO.longistaminata intoO.sativahave been reported other thanXa21for bacterial leaf blight [15] andqSPP2.2for grain number [16].The objective of this study was to investigate the potential value ofO.longistaminata for rice improvement by resequencing and phenotyping of a BIL population derived from a cross ofO.longistaminata toO.sativa.

2.Materials and methods

2.1.Plant materials

A set of 152 BILs were derived from an initial cross of the eliteO.sativacultivar 9311 ×O.longistaminata(IRGC 103886) (Fig.S1).The F1hybrid was backcrossed with 9311 as the recurrent parent,and 67 BC1plants were backcrossed again to generate 176 BC2plants.All of the BC2plants were then bagged and self-pollinated for 20 generations without intentional selection.Removal of sterile BC2F20plants left 152 fertile BILs.For agronomic character evaluation,9311 and the BILs lines were planted in 20 cm rows with 16.5 cm spacing under conventional crop management in various experimental fields during 2012-2017.A resequencing-based bin-map construction workflow was developed withO.longistaminatacore BILs,and QTL mapping is described in Fig.S1B.

2.2.Characterization of disease and pest resistance

Rice blast disease infection was characterized during 2013-2014 in Enshi,Hubei,China,and Taojiang,Hunan,China,both typical blast-affected areas in the Middle Yangtze River region.9311 was planted as a negative check every five entries and on the borders to maximize disease incidence.Disease score was recorded on the 0-9 scale of the Standard Evaluation Scale (SES) of IRRI,2002.The scoring of Bacterial leaf blight (Xanthomonas oryzaepv.oryzaeorXoo)resistance was performed following inoculation with type-IV and type-VXoostrains cultured in a peptone sucrose agar(PSA)medium at 28°C for 48 h.Bacterial suspensions(OD600=0.6)were used for inoculation of fully expanded rice leaves at the tillering stage.Five plants of each line were inoculated using the leaf-tipclipping method.Disease symptoms were recorded three weeks after inoculation and measured as lesion length [17].An artificialUstilaginoidea virensinoculated panicle was conducted to identify the resistance score of Rice false smut for each plant.U.virenswas inoculated on PSA plates at 28°C until the emergence of mycelium.Pure culture ofU.virenswas multiplied in a 1 L flask and incubated in an incubator shaker at 180 r min-1at 28°C for 6 days.Hyphae and conidia were harvested and resuspended in fresh peptone sucrose broth(PSB)to 106conidia mL-1as inoculum.Disease symptoms were recorded three weeks after inoculation and measured as smut balls per plant.For the resistance score of brown planthopper(BPH)identification,about 20 seeds of each line were randomly sown in separate plastic boxes in 26 cm-long rows,with 2.5 cm between rows.At the third-leaf stage,each seedling was infested with 10 s to third-instar BPH.When all of the susceptible 9311 or Taichun native 1 control plants had died(scored as 9),the positive control Luoyang 69(Bph6andBph9,scored as 0)and each of the BILs were scored as 0,1,3,5,7 or 9 as previously described[18].

2.3.Characterization of yield-related and general agronomic traits

The recurrent parent 9311 and the 152 BILs were evaluated for agronomic and yield-related traits in a randomized complete block design with three replications.Phenotypes were recorded for five plants from each of the entries for plant height (PH),days to 50%flowering (DTF),effective panicle number (EPN),panicle length(PL),grain number per panicle(GNPP),filled grain number per panicle (FGNPP),spikelet setting rate (SSR),thousand-grain weight(TGW),grain yield per plant (GY),primary branch number (PB),second branch number (SPB),seed setting rate (SSR),grain width(GW),grain length (GL),and grain thickness (GT).PH was measured from the soil surface to the apex of the tallest panicle at maturity.HD was calculated as days from sowing to time when spikes were extruded from boots for more than half of the plants.EPN was evaluated when grains were fully matured.The length of the longest panicle was recorded as PL.The total grain number and total filled grain number of spikelets produced on the main panicle were recorded.The panicle and grain traits were measured in the laboratory after plants were dried.PL was measured with a ruler and PPB and SPB on the main panicle were determined manually.GNPP and FGNPP were estimated with an automatic seed counter (SLY-C,Hinotek,Riverside,CA,USA).SSR was calculated as FGNPP/GNPP.Grain were mixed and 10 were randomly sampled for phenotypic analysis,GL and GW were recorded at each grain’s maximum values using an electronic digital caliper.TGW value was calculated from the weight of 300 grains.

2.4.Construction of a near-isogenic line (NIL)

Large-panicle line 1762,with GNPP of about 315 and TGW of 25 g,was selected to cross with 9311 as recipient,after which the F1hybrid was backcrossed again to 9311 until all the characters but panicle characters were identical to those of 9311.The resulting NIL was named L1880.Single sequence repeats (SSR) and insertion-deletion(InDel)markers were used to identify the introgression region fromO.longistaminatain NIL1880.

2.5.DNA preparation and population resequencing

Genomic DNA of 9311,O.longistaminata,and the BILs was extracted with a DNeasy Plant Mini Kit (QIAGEN,Shenzhen,Guangdong,China).Sequencing libraries with a 500-bp insert size were constructed and sequenced on a HiSeq2500 instrument(Illumina and Life Technologies/ABI SOLiD) according to the manufacturer’s instructions.The raw re-sequencing data of BILs can be seen in Table S1.Whole-genome resequencing was performed for 9311 andO.longistaminatato 30×depth by BGI(Shenzhen,Guangdong,China),and the BILs were resequenced to ＞15× depth by BEINA (Wuhan,Hubei,China).All sequence data are available under accession ID PRJNA615752 in the Sequence Read Archive of NCBI.For SNP calling,the short sequencing reads were aligned to the rice reference genome of Minghui63 (https://rice.hzau.edu.cn/rice_rs1/download_ext/MH63RS1.LNNK00000000.fsa.tar.gz)using BWA v0.7.12 [19] with default parameters,and population single nucleotide polymorphism (SNP) calling was performed with GATK 4.0 HaplotypeCaller[20].Only homozygous SNPs polymorphic between the parents were reserved for further analysis.SNPs were accepted only if SNP coverage depth was at least 5×and no higher than 80×in either parent and the genotype quality was at least 20.

2.6.Bin map construction

A slightly modified sliding-window approach [21] was used to identify recombination breakpoints and construct a bin map of BILs.For each plant,a window size of 15 SNP without missing data was chosen for genotyping.The SNP genotype from 9311 was coded as ‘‘a(chǎn)” and that fromO.longistaminataas ‘‘b”.The ratio of the number of SNPs with the 9311 to that with theO.longistaminatagenotype was calculated [22].When the ratio a:b exceeded 4:1,the genotype of the window was recorded as‘‘a(chǎn)”;if it was less than 1:4,the genotype was recorded as‘‘b”;otherwise,the window was called as heterozygous [22].Breakpoints were determined as previously described[23]for high-throughput genotyping by Next generation sequencing (NGS) with some modification,and the Kosambi mapping function was used to calculate map distance(cM).

2.7.QTL mapping and new QTL identification

Bin markers were generated from 363,978 high-quality SNPs.A linkage map was constructed from high-density bin markers using QTL IciMapping software.The threshold for declaring the presence of a QTL for each trait was calculated by 1000 permutations atP＜0.05 and a minimum of significance threshold LOD=3.0 was chosen[24].QTL were identified with the internal composite interval mapping (ICIM) algorithm for additive gene effects implemented in QTL IciMapping.The chromosomal locations of QTL were drawn by physical distance with Circos software.As with previous yield-associated and resistance-associated genes and QTL identified inOryza sativa,the sequence corresponding to the region of tight linked two bin markers was aligned to the Nipponbare(NIP) reference genome (https://rice.plantbiology.msu.edu/).A novel QTL was defined as one not within 200 kb of any gene or QTL already labeled.

2.8.Identification of SNPs,InDels,structural variation (SV) and copy number variation (CNV)

To better understand the variation between 9311 andO.longiatminatagenome in sub-species background,we used two reference genomes includeOryza sativaindica (MH63) andOryza sativa japonica(NIP).SNPs and short (less than50 bp) indels were called using the HaplotypeCaller in GATK.SNPs were filtered using Variation Filtration in GATK,according to the following criteria:variant quality(QD)＞2.0,quality score(QUAL)＞40.0,and mapping quality(MQ)＞30.0.SnpEff software was used to annotate the effects of SNPs and InDels.A sliding-window method(window size,200 kb)was used to calculate the distribution of SNPs and InDels in each genome.Structure variation calling was performed with LUMPY v0.2.13 [20].The pipeline operated as follows.i) Read lengths and insert sizes were extracted from BAM files for each sample using SAMtools 1.9 [25].ii) SVs were genotyped with Breakdancer 1.1.2 [26].CNVs among 9311,O.rufipogon,andO.longistaminatawere identified with CNVnator (version 0.3) [27] based on read depth.The parameter setting for CNVnator was ‘‘-call 100”.SVs with a minimum length of 500 bp and read depth ＜1.2 or ＞1.8 of the mean genome depth were assigned as candidate CNVs.Bedtools coverage(with window size 200 kb)was used to calculate the distribution of SVs in each genome.

3.Results

3.1.Construction and characterization of a BIL population of O.Longistaminata

To exploit favorable traits inO.longistaminata,we crossed 9311 as a maternal parent with anO.longistaminataaccession that shows contrasting genetic variation with 9311 (Fig.S2;Table S1).To characterize the phenotypes of theO.longistaminataBIL population,we investigated domestication traits: shattering,awn length,hull,and hull and seed coat color of the BILs from 2012 to 2014 in Wuhan and Hainan.Of the 152 BILs,111 shattered before maturation,98 had awns with a mean length of 0.6 cm,and the longest awn reached 5.2 cm.For hull color,148 BILs showed yellow straw,three red,and one black.For seed coat color,117 lines showed white seed,19 light red,13 deep purple,and three black in Hainan or Wuhan(Figs.S3,S4),meaning that these characters were highly consistent in different environments.

The agronomic traits plant height,heading date,and stem width were also recorded.Plant height ranged from 63.7 to 224.7 cm,with over 70% of the lines taller and only 15% shorter than the recurrent parent 9311 (Fig.S5;Table S2).Heading date ranged from 73.9 to 126.2 days,with ＞60% of the BILs heading later than that of 9311 and most BILs heading earlier in Hainan than in Wuhan(Table S2;Fig.S6).Around 68%of the lines had wider stems than 9311,a trait associated with stronger lodging resistance.

3.2.Characterization of grain-yield traits in O.Longistaminta BIL population

Although wild rice has low grain yield in natural environments,the long panicles and sturdy and large culms ofO.longistaminatasuggest that this plant may have high yield potential.We investigated grain yield traits: effective panicles,grain number,numbers of primary and secondary branches of panicles,thousand-grain weight,seed length,and seed width in theO.longistaminataBIL population.

Grain number ranged from 82.0 to 390.9,with the highest value almost double that of 9311.The numbers of primary and secondary branches reached 17.3 and 73.2,respectively (Fig.1).Thousandgrain weight ranged from 15.8 g to 35.6 g,the highest grain length and width reached respectively 11.4 and 3.1 mm,(Table S3;Fig.S7),and grain yield per plant reached 67.3 g.BIL 1795 showed not large seeds,but large panicles (Fig.1),with the same mean seed-setting rate and panicle number as 9311,indicating a relationship among the yield characters and showing high potential for yield improvement.

Fig.1.Phenotypic variation of yield traits in O.longistaminata BILs.(A)Panicle size of BILs.Scale bar,10 cm.(B)Primary branch number.Scale bar,10 cm.(C)Second branch number.Scale bar,10 cm.(D) Grain length.Scale bar,1 cm.(E) Grain width.Scale bar,1 cm.

EPN ranged from 2.5 to 16.3,with about 10%of the BILs carrying 7-9 more EPN than 9311 (Figs.1,S7).SSR ranged from 17.2% to 98.4%,with over half of the BIL lines showing a seed-setting rate of over 80% similar to that of 9311.Among the four traits,SSR showed the largest variation among sites and years (Table S3),reflecting its high sensitivity to environmental effects.

3.3.Characterization of disease and pest resistance in O.longistaminta BIL population

To characterize genes for resistance to rice diseases and insect pests inO.longistaminata,the resistances of BILs to rice false smut,leaf blast spot,bacterial leaf streak,and rice brown planthopper were repeatedly investigated from 2012 to 2017 at the growing sites.

For brown planthopper,seven lines showed high resistance,retaining the green color of Luoyang 69 (positive control carryingBph6andBph9) when 9311 had been killed by the insect (Fig.2).For rice leaf blast spot disease,the BIL population was tested in two disease epidemic regions: Tao-Jiang,Hunan province (2012-2013) and En-Shi,Hubei province (2014-2015).Over 30% of the BILs showed elevated blast resistance,with the highest resistance reaching grade 2 in contrast to the grade 5 of 9311 (Figs.2,S8,S9;Table S4).For bacterial leaf blight,over half of the BILs were highly resistant toX.oryzaeinfection,with the highest resistance toXooIV andXooV reaching grades 0.5 and 1.2,respectively(Figs.2,S9).Nine BILs showed strong resistance to rice false smut in Dong-Kou(Hunan province),and E-Zhou(Hubei province),with BIL 1809 showing no false smut ball at either site(Fig.2;Table S4).The BIL population showed strong resistance to these epidemic and destructive diseases and pests.Grain yield showed a positive and highly significant correlation with effective panicle number,grain number per plant,grain number per panicle,seed setting rate,and thousand grain weight.BPH and rice bacterial blight resistance were positively correlated,and false smut disease resistance was positively correlated with heading date,plant height,seed width,and seed setting rate,suggesting that false smut infection is readily affected by environments and by plant and panicle architecture(Fig.S10).

Fig.2.Phenotypic variation of disease and pest resistance in O.longistaminata BILs.(A) Identification of resistance against brown planthopper,Luoyang 69 (LY69) carrying Bph6 and Bph9 genes as a positive control.Taichun native 1(TC1)was used as negative control.(B)Phenotype of the BIL lines for rice false smut.(C)Bacterial blight resistance of Xoo strain V type.Scale bar 10 cm.(D) Bacterial blight resistance of Xoo strain IV type.Scale bar,10 cm.(E) Rice leaf blast.Scale bar,10 cm.

3.4.Genotyping of O.longistaminta BIL population by resequencing

Compared to 9311,the alien chromosome introgression lines showed distinctive characteristics,including strong plant stature,high grain yield,and strong resistance to epidemic diseases and pests,implying that many valuable genes fromO.longistaminatawere introgressed into BILs.To characterize the genotype of the BIL population,each line was sequenced to～15×genome coverage to identify the genomic fragments fromO.longistaminata(Tables S5,S6).Genotypes of the 152 BILs were identified by a slidingwindow approach along the 12 chromosomes.There were 2432O.longistaminatarecombinant blocks with an average length of 156.05 kb identified from 1091.6 Gb of clean reads (Fig.3).Each line contained on average 22.5 bins of 7327 kb,with bin size ranging from 301 to 21,825 kb (Figs.3A,S11;Table S7).

3.5.QTL mapping of the grain yield and resistance-related characters

In order to further identify the favorable genes for rice grain yield and resistance to disease and pests,potential QTL were identified using the high-density linkage map fromO.longistaminataBILs and QTL IciMapping.Thirteen QTL for seven grain yield traits were detected inO.longistaminata(Fig.4).These QTL comprised five QTL for seed length and two for seed width,explaining respectively 6.5%-13.0%,and 10.0%-11.8% of phenotypic variance.Two QTL for grain number,three for primary branch number,and one QTL for secondary branch number explained respectively 14.7%-18.1%,14.7%-35.9%,and 13.8%of phenotypic variance(Table S8).qPB10.1,qSL1.1,qSL3.1were repeatedly detected in the same or different sites(Table S8),implying these QTL were little affected by environment.

Fig.4.QTL mapping of grain yield,disease,and brown planthopper resistance in O.longistaminata.There were respectively 13 QTL associated with grain yield and 12 QTL associated with disease and brown planthopper detected in the BIL population.The plot (A) shows QTL identified in O.longistaminata. Different colors represent different trait-associated QTL,Xoo,blight bacterial disease in red triangle;Bph,rice brown planthopper in green triangle;RFS,rice false smut in blue triangle;Blast,blast disease in yellow triangle.The circos(B)represents yield-associated QTL detected in wild rice.SPP,the number of spikelets per panicle in blue rectangle;PB,the primary branch number in green rectangle;SB,secondary branch number in red triangle;SL,seed length in blue triangle;SW,seed width in green triangle.The circos(C)stands for SNP coverage per 100 kb.The circos(D)stands for the number of SNPs per 100 kb between 9311 and O.longistaminata.The circos(E)stands for gene density per 100 kb along the chromosomes of the reference genome Minghui 63.

For rice disease and pests,eight QTL for resistance to brown planthopper (BPH),one QTL for resistance to bacterial leaf blight(BB),two QTL for resistance to rice false smut(RFS)and one QTL for resistance to leaf blast spot(Pi) were detected inO.longistaminata(Fig.4).Of these,two Bph QTL were repeatedly detected in different years at the same site,suggesting that these QTL were highly stable and might function effectively in rice defense.The QTL for brown planthopper and rice false smut resistance explained phenotypic variances of respectively 0.78%-9.47% and 4.40%-34.55%.The QTL responsible for leaf bacterial blight and rice blast accounted for respectively 36.4% and 8.3% of phenotypic variance(Table S9).

3.6.Validation of high-yield QTL

To determine whether the newly identified QTL from wild rice overlapped with previously reported QTL,we compared the 13 high grain yield-associated QTL and 12 pest-and diseaseresistance QTL with those in previous reports.Only one QTLqBph12.1 wasoverlapped withBph10[28[,implying that most of theO.longistaminataQTL or genes for grain yield and disease and planthopper resistance are novel in rice.

To determine whetherO.longistaminata-specific QTL would be valuable in rice breeding practice,we crossed line 1762 carryingqPB8.1as donor parent to 9311 and then backcrossed for six rounds to develop a near-isogenic line named NIL-1880 (Fig.5A),which carried only a short alien fragment of 84.5 kb on the short arm of chromosome 8 (Fig.5B).In comparison with 9311,NIL-1880 had similar stature,heading date and seed-setting rate,plant height,1000-grain weight,and effective panicles,but grain number per panicle was about 100 greater than that of 9311,resulted in a roughly 27.5% increase in grain yield (Fig.5C).

Fig.5.Validation of grain yield-associated qPB8.1.(A)Plant stature and panicles of 9311 and qPB8.1 near-isogenic line 1880.(B)Genetic background of NIL-1880.(C)Yieldassociated traits in 9311 and NIL-1880.PB,primary branch number;SB,secondary branch number;SSR,seed setting rate;SSP,spikelet number per panicle;TGW,thousandgrain weight;PL,panicle length;PY,plant yield.Values are means ± SD.

4.Discussion

4.1.O.longistaminta genome holds more genetic variation than common wild rice

Genetic variation is considered the basis of crop breeding.To breed elite rice as rapidly as possible,commercial breeding programs tend to cross genetically similar elite parents,a practice that results in a narrow range of genetic variation in rice and renders modern cultivars vulnerable to biotic and abiotic stresses [29].It is desirable to exploit distant wild resources in search of alien genetic variation[30].Wild rice species have survived over a thousand centuries of pest attacks and extreme environments and accumulated a wide range of beneficial mutations,representing a potential reservoir of genetic variation influencing fertility,grain number,grain size,and tolerance to biotic and abiotic stresses.Because the AA-genome wild relatives share a genome withO.sativa,they are the most accessible genetic resource in view of the relative ease of hybridization.

TheO.longistaminataconsistently showed differentiation from the otherOryzaAA-genome species.This differentiation is mirrored not only by the unique morphological feature but also by the distinctively differentiated genomes [14].Many reports have documented that bothO.longistaminataandO.glumaepatulaare the most ancestral species.TheO.rufipogonancestor diverged fromO.longistaminata5-7 million years ago [31-33],meaning thatO.longistaminatais genetically more distant from cultivated rice thanO.rufipogon.

Theoretically,a higher genetic distance between two species will support more genetic variation and potentially create more vigorous heterotic descendants in optimal mating designs [34].Genetic variations can be well characterized by genomic indexes,including SV,SNP,InDel,and copy number variation (CNV).We compared the genomic variations betweenO.longistaminataandO.rufipogongenomes relative to the cultivars Minghui 63 and Nipponbare and found that SNP,InDel,and CNV variations were all richer inO.longistaminatathan inO.rufipogon(Fig.6).SNP,InDel,and CNV were 50%,120% and 20% more frequent inO.longistaminatathan inO.rufipogon(Table S10).This genetic feature suggests thatO.longistaminatamay contain many more valuable variations thanO.rufipogon,a finding consistent with that of QTL with potential to improve rice performance (Fig.4).

Fig.6.Genetic variation among O.longistaminata,O.rufipogon,and Oryza cultivars.Left,Minghui 63-based genomic variations;Right,Nipponbare-based genomic variations.SNP,InDel,SV,and CNV represent respectively single-nucleotide polymorphism,insertion and deletion,structural variation,and copy number variation.Inside and outside of each circos represent respectively O.rufipogon and O.longistaminata variations.

4.2.O.longistaminata is a valuable resource for future rice improvement

Full exploitation of the wild rice gene resources relies on an understanding of the genetic variation and genomic character ofOryzaspecies[31].Genetic variation usually leads to sexual segregation betweenO.longistaminataand cultivated rice[35,36],limiting its utility in rice breeding.Unlike the wild relatives ofO.rufipogon,O.nivara,O.glaberrimaandO.meridionalis,which serve as excellent resources for rice improvement [37-39],no valuable genes have been identified inO.longistaminataexcept forXa21[40].Using distantO.longistaminatato incorporate genetic variation is an option for developing elite cultivars with high grain yield,strong stress tolerance,and disease resistance.

In this study,we developed a BC2F20BIL population covering about 99.6% of theO.longistaminatagenome using rice 9311,an elite high yield rice widely used for the restorer of the three and two-line hybrid rice in China,and found that a series of lines show high grain yield characters (Fig.1) or strong resistance to brown planthopper,bacterial blight,leaf blast,or rice false smut (Fig.2).This is the first report of 12 QTL for high grain yield (Table S8),and 11 for resistance to brown rice planthopper and epidemic rice diseases (Table S9) inOryzaspecies.Line 1795 showed not only large grain size but large panicle and strong stem strength(Fig.1),implying that genes for grain size and large panicle can be combined to increase both grain yield and lodging resistance.To investigate whether these high-yield QTL could be useful in rice breeding,we combined the large-panicle alleleqPB8.1fromO.longistaminatawithRf3-Rf6,Bph6,andBph9to develop a novel high-yield restorer line with strong resistance to Bph[41],showing that theqPB8.1allele can be combined with alleles against this insect pest.Eight QTL for resistance to brown planthopper were detected inO.longistaminata(Table S9).some of the BILs showed the same strong resistance as the positive control Luoyang 69(Fig.2),which carry theBph6andBph9alleles [42].These discoveries further highlight the potential value ofO.longistaminatain rice germplasm improvement.

Her delicious meals and quiet smile graced the cabin with a wonderful woman s touch. But the wrong woman, Edward mourned as he collapsed9 onto his cot each night. Why did they send Marta? Would he ever see Ingrid again? Was his lifelong dream to have her as his wife forsaken10?

We identified two QTL associated with resistance to rice false smut,a severe disease that is increasingly epidemic worldwide.Line 1809 was almost immune to this disease at every season and site(Fig.2).In the last decades.rice breeders have striven to identify sources of resistance to rice false smut in modern cultivars,landrace,and even the common wild riceO.rufipogonandO.nivara,with unsatisfactory results [43].The discovery ofO.longistaminataBILs resistant toU.virenscould reduce the threat ofU.virensto rice production.A new QTL for strong resistance to bacterial leaf blight was identified (Fig.4;Table S9),suggesting thatO.longistaminatamay provide new genes to improve rice resistance againstXoostrains in breeding practice.With the release of a high-resolution genome[14] and increasing studies in search of valuable genes fromO.longistaminata[16,44-48],more valuable genes hidden inO.longistaminatawill be detected and cloned.CombiningO.longistaminatahigh-yield QTL with disease-and pest-resistance alleles in elite rice cultivars will contribute to future food security.

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

The authors thank Professor Jinshan Tian for his kind offering the strains ofXanthomonas oryzae.This work was partly granted from the National Natural Science Foundation of China(U20A2023,31870322),the Creative Research Groups of the Natural Science Foundation of Hubei Province,China(2020CFA009),and the Hubei Hongshan Laboratory (2021hszd010).

CRediT authorship contribution statement

Weixiong Long:Data curation,Investigation,Writing-original draft.Nengwu Li:Resources.Jie Jin:Resources,Data curation,Investigation.Jie Wang:Data curation.Dong Dan:Investigation.Fengfeng Fan:Investigation.Zhiyong Gao:Writing -review &editing.Shaoqing Li:Conceptualization,Project administration,Writing -review &editing.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2023.03.017.