Wen Chu,Ji Liu,Hongto Cheng,Cho Li,Li Fu,Wenxing Wng,Hui Wng,Mengyu Ho,Desheng Mei,Kede Liu,Qiong Hu,c,*

a Oil Crops Research Institute of Chinese Academy of Agricultural Sciences/Key Laboratory for Biology and Genetic Improvement of Oil Crops,Ministry of Agriculture and Rural Affairs,Wuhan 430062,Hubei,China

b National Key Laboratory of Crop Genetic Improvement,Huazhong Agricultural University,Wuhan 430070,Hubei,China

c Graduate School of Chinese Academy of Agricultural Sciences,Beijing 100081,China

Keywords:Pod shatter Lignified-layer bridge BSA-Seq KASP marker Rapeseed

ABSTRACT Pod shattering causes severe yield loss in rapeseed(Brassica napus L.)under modern agricultural practice.Identification of highly shatter-resistant germplasm is desirable for the development of rapeseed cultivars for mechanical harvesting.In the present study,an elite line OR88 with strong shatter resistance and a lignified-layer bridge (LLB) structure was identified.The LLB structure was unique to OR88 and co-segregated with high pod-shatter resistance.The LLB structure is differentiated at stage 12 of gynoecium development without any gynoecium defects.Genetic analysis showed that LLB is controlled by a single recessive gene.By BSA-Seq and map-based cloning,the resistance gene location was delimited to a 0.688 Mb region on chromosome C09.Transcriptome analysis suggested BnTCP8.C09 as the gene responsible for LLB.The expression of BnTCP.C09 was strongly downregulated in OR88,suppressing cell proliferation in the pod valve margin.KASP markers linked to the candidate gene were developed.This pod shatter-resistant line could be used in rapeseed breeding programs by direct transfer of the gene with the assistance of the DNA markers.

1.Introduction

Rapeseed(Brassica napus L.)is an important oilseed crop whose oil is widely used for human consumption and can also be used as a biofuel[1].Rapeseed pods readily dehisce and disperse their seeds at maturity,resulting in yield losses of 15% to as high as 50% [2]under unfavorable weather conditions,especially with machine harvesting.Development of pod shatter-resistant cultivars is desirable for the modernization of rapeseed production.

The pod structure of rapeseed is similar to that of the model plant Arabidopsis,sharing the same shattering process.In Arabidopsis,the dehiscence zone consists of a separation layer and a lignified layer.After maturity,tensions are produced by the stiffness of lignified tissues (lignified layer and endocarp b) and shrinking of non-lignified tissues in pod walls.At the same time,adhesion among cells in the separation layer is weakened by enzymatic activity.Pod valves then readily detach at the separation layer and seeds are dispersed [3,4].

Dehiscence zone formation has been extensively studied in Arabidopsis.The zone forms at stage 12 of gynoecium development under the control of an intricate genetic regulatory network that involves several transcription factors [4,5].SHATTERPROOF1(SHP1) and SHP2 work redundantly and are essential for the expression of INDEHISCENT (IND) and ALCATRAZ (ALC) [6,7].IND controls the development of the separation and lignified layers[7].ALC is required for separation layer differentiation[8].REPLUMLESS (RPL) and FRUITFULL (FUL) function in the replum and the valves respectively,constricting the dehiscence zone within the valve margin and protecting replum and valve development [9–12].Other factors have also been reported to act in the pod shattering regulation network.The lignification associated factors NAC SECONDARY WALL THICKENING PROMOTING FACTOR1(NST1) and NST3 regulate secondary wall formation of the endocarp b and lignified layer[13].ARABIDOPSIS DEHISCENCE ZONE POLYGALACTURONASE1 (ADPG1) and ADPG2 encode polygalacturonases and act on the separation layer by weakening cell adhesion [14].Auxin,gibberellins and other phytohormones also function in pod shattering[15–17].Recent studies [18–20] have shown that the Arabidopsis regulatory network is conserved in B.napus.For example,knocking out BnaIND or BnaALC increased pod shatter resistance,confirming that BnaINDs and BnaALCs also function in dehiscence zone formation in rapeseed.

Fruit indehiscence is usually accompanied by special tissue morphology.Overexpression of FUL resulted in an abnormal dehiscence zone.A similar phenotype was found in shp1shp2 and ind mutants [7,10].The double NST mutants nst1nst2 could not produce lignified cells in endocarp b and the lignified layer [13].A SHAT1-5 mutation in soybean caused lignification of fiber cap cells(FCC),resulting in pod shatter resistance [21].An association of a larger replum-valve joint area with increased shatter resistance was reported in B.napus [22].Pods of the BolC.GA4.a mutant in Brassica oleracea failed to form a valve margin properly,leading to a valve bridge with high shatter resistance[23].A similar phenotype was observed in the BraA.IND.a mutant of Brassica rapa,the BolA.IND.a mutant of B.oleracea and the BnIND double mutant of B.napus [20,24,25].

A lack of elite pod-shatter-resistant germplasm has always been an obstacle to the genetic improvement of rapeseed.Although there have been extensive studies of the genetic regulation mechanism of pod shatter in A.thaliana [3,4],identification of natural pod shatter-resistance genes has never been reported.Natural variation for shatter resistance in B.napus,a crop relative of the model plant A.thaliana,is also narrow and only a few pod shatter-resistant lines have been identified [22,26,27].Although several genetic loci controlling pod shatter resistance have been mapped in B.napus,the genetic effect of these loci was too low to meet the requirement for developing cultivars with good pod shatter resistance for mechanical harvesting [26–28].But toorobust pods are also undesirable because they interfere with threshing[29].Mutation of some pod shatter genes such as BnIND led to strong shatter-resistant pods that were not easy to thresh[18].It is desirable to identify more pod shatter-resistant B.napus germplasm suitable for mechanical harvesting.

To screen pod shatter-resistant rapeseed material,more than 2000 accessions including breeding lines and cultivars have been collected and assessed shatter resistance index (SRI) by random impact test(RIT)[30]over the last ten years.An Ogura cytoplasmic male-sterility (CMS) restorer line,OR88,was shown to have the highest SRI among these accessions.The Ogura CMS system is widely used in rapeseed hybrid production [31–33].As both the male sterile cytoplasm and nuclear restorer were derived from radish (Raphanus sativus L.),the Ogura CMS restorer line carries on chromosome C09 a radish introgression that harbors the restorer gene Rfo [32].Because of the radish introgression,unfavorable traits such as low number of seeds per pod and high glucosinolate content are also present.Because the radish introgression severely reduces the meiotic recombination of the radish introgression and surrounding area with the rapeseed genome,it is hard to modify traits controlled by genes located in the area [33].

The purpose of the present study was to exploit the application potential of the pod shatter-resistant line OR88 and identify the genetic basis of its resistance.The experimental approach was to characterize the structure of OR88 anatomically,investigate the inheritance of pod shatter resistance in segregating populations,map resistance loci by bulked-segregant analysis sequencing(BSA-Seq),and develop linked DNA markers for marker-assisted selection in breeding programs.

2.Materials and methods

2.1.Materials

OR88 and OR5 are Ogura CMS restorer lines.No.87 is a breeding line and Zhongshuang 11 (ZS11) is a popular cultivar widely used in semi-winter rapeseed production region in China.They were all developed at the Oil Crops Research Institute,Chinese Academy of Agricultural Sciences (OCRI-CAAS).These materials were grown under standard agronomic practice at Yangluo Experimental Station,Wuhan,China.

An F1progeny was obtained by crossing OR88 and No.87 in 2016 and 2017.Then No.87 was used as the female parent to cross with F1plants to generate a BC1population in 2018.The BC1population comprised 2102 plants.An F2population derived from self-pollination of F1plants in 2017,comprised 222 plants.Doubled-haploid (DH) populations were developed from the crosses No.87 × OR88 in 2017.Another F2population was developed from the cross OR5×OR88 in 2018,comprising 1780 plants.The phenotypic segregation ratios of BC1,F2,and DH populations were tested with a χ2goodness-of-fit test using SAS 9.1.3 [34].

2.2.Anatomical analysis

To observe lignified-layer bridge(LLB)structure,pods were harvested at 21–35 days after anthesis.From the middle of a pod,several cross sections were cut along the longitude axis by a razor blade.Fresh sections were observed under a fluorescence microscope with ultraviolet filter at OCRI-CAAS.Photographs were taken with an Olympus CX31 microscope (Olympus,Tokyo,Japan) with Sony digital camera (Sony,Tokyo,Japan) [35].

For scanning electron microscopy(SEM),pods were collected at 14 days after anthesis and fixed in 2% paraformaldehyde and 1%glutaraldehyde in 1× phosphate buffered saline (PBS,pH7) overnight.Samples were subsequently dehydrated through an increasing range of ethanol series.After dried using the Quorum K850 critical point dryer (Quorum,Laughton,East Sussex,UK),samples were coated with gold using Coating Unit MC1000(Hitachi,Tokyo,Japan) and viewed using SU8100 scanning electron microscope(Hitachi,Tokyo,Japan).

Buds of 3–8 mm length of OR88 and No.87 were collected and fixed in FAA solution (50% ethanol,5% acetic acid,3.7% formaldehyde) for 6 h.Samples were then dehydrated through an ethanol series and embedded in paraffin wax.Histological sections were prepared using a Leica RM 2165 rotary microtome (Leica,Mannheim,Germany).Sections were stained with 0.1% toluidine blue following van Gelderen et al.[17] and photographed as described above.Gynoecium development stage identification was performed following Smyth et al.[36].

2.3.Assessment of pod shatter resistance

Pods from the main inflorescence of each plant were harvested at physiological maturity.SRI was measured by RIT [30].The pods were oven dried at 45°C for 8 h and the dried pods were shaken at 220 r min1for 10 min in a drum with 20 cm inner diameter with 8 ball bearings of 14 mm diameter.During shaking,the drum was stopped every 2 min to record the number of cracked pods.Pod SRI was calculated as SRI=1×(6 i)/100(1 i 5),where Xiis the number of cracked pod recorded at the ith time(1 i 5).

2.4.BSA-Seq and mapping of BnLLB

From the F2population,42 plants with LLB pods and 42 plants with normal pods were mixed as LLB pod and normal pod pools.From each plant,young leaves were used for DNA extraction.Genomic DNA samples of 42 plants with and without LLB were mixed equally to construct a bulked DNA pool for each.Genomic DNAs of two parental lines OR88 and No.87 were also extracted.The two DNA pools and parental DNA samples were sequenced with an Illumina NovaSeq 6000 sequencing platform(Illumina,San Diego,CA,USA) (30× for the parental genomes and 42× for the bulked-pool genomes).Sequences were mapped against Darmor-bzh reference genome(http://www.genoscope.cns.fr/brassicanapus/) [37] using BWA software [38].Single nucleotide polymorphisms (SNPs) were called with GATK software [39].For each SNP,the proportion of short reads harboring SNP that different from the reference sequence (sequence of either of the two parents)was estimated as the SNP-index.ΔSNP-index was calculated by subtracting the normal pod pool from the LLB pod pool.The graph of mean ΔSNP-index values were plotted by calculating in a 4 Mb interval using a 10 kb sliding window.The procedure followed a previous study [40].

Based on the sequences of the two parental lines,SNP and InDel information was obtained and used for Kompetitive Allele Specific PCR(KASP)marker design.The 250-bp sequences flanking a target SNP or InDel on either side were extracted from the Darmor-bzh reference genome.The primers were designed with Primer Premier 5 software(Premier Biosoft,San Francisco,CA,USA).KASP-TF Master Mixe was ordered from LGC Company(Teddington,Middlesex,UK).Primer design and SNP assay methods followed Trick et al.[41].

In all 29 KASP markers were developed on chromosome C09 and 10 markers were identified as being tightly linked to BnLLB in the No.87 × OR88 BC1population.To further narrow the candidate region,an F2population from OR5×OR88 was constructed.Finally,markers with polymorphism between OR5 and OR88 were used to construct a linkage map of the target region using JoinMap 4.0[42].

2.5.RNA-Seq and candidate gene analysis

RNA was extracted from stage 11–12 developing gynoecium of OR88 and OR5 using RNA Easy Fast kit (Tiangen Biotech,Beijing,China) to construct independent RNA libraries with three replicates for each line.The RNA libraries were then subjected to digital RNA sequencing [43] with a Unique Molecular Identifiers (UMI)adapter [44] at Novogene (Beijing,China).After removal of adapters and low-quality reads,clean reads were aligned to the Darmor-bzh reference genome with TopHat2 [45].Based on calculation of FPKM(number of fragments per kilobase of transcript per million mapped reads) of each transcript using Cufflinks [46],differentially expressed transcripts between OR88 and OR5 were identified at a false discovery rate(FDR)＜0.05 and|log2(ratio)|1.GO term enrichment analysis was performed with GOseq [47].

To verify the expression of BnTCP.C09,RT-PCR was performed.A 1–2-μg aliquot of each RNA sample was subjected to reverse transcription using a HiScript III 1st Strand cDNA Synthesis Kit(Vazyme,Nanjing,China).PCR amplification was performed using 2×Phanta Max Master Mix(Vazyme).Primers were designed with NCBI Primer-BLAST (http://www.ncbi.nlm.nih.gov/tools/primerblast/) and synthesized by Genecreate (Wuhan,China).Alignment of BnTCP.C09 amino sequence was performed with DNAMAN 5.0(Lynnon Biosoft,San Ramon,CA,USA).A phylogenetic tree of BnTCP.C09 was constructed with MEGA 7.0 [48].Sequences of 2000 bp upstream of the BnTCP.C09 CDS in OR88 and OR5 were retrieved and subjected to cis-element prediction by PLANT CARE(http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [49].Cis-elements were plotted with GSDS2.0 (http://gsds.gao-lab.org/index.php) [50].

3.Results

3.1.A lignified-layer bridge structure was specific to OR88

In 2016,2017 and 2018,477,173,270 rapeseed cultivars and breeding lines were assessed for pod shatter resistance by RIT.The majority of these materials showed SRI ＜0.20 and was classified into the‘‘very easy shattering”group.However,an Ogura CMS restorer line,OR88,showed the highest SRI over three years and four environments.The SRIs of OR88 grown at Wuhan,Hubei in 2016–2018 were 0.78,0.86,and 0.89,and at Pingan,Qinhai was 0.59 in 2017.Besides the high SRI,OR88 showed normal agronomic traits,with seed number per pod,thousand-seed weight,pod number per plant and plant height similar to those of breeding line No.87 and a cultivar ZS11 widely used for rapeseed production in China (Table S1).

Twelve breeding lines with SRI ＞ 0.30 were subjected to anatomical examination of pod structure.It showed that the boundaries of valves vanished in OR88,in sharp contrast to other materials (Fig.1A and B).There were 2–4 rows of lignified cells at the position of valve margin in OR88.We called these lignified cells as lignified-layer bridge (LLB),because they linked two lignified layers as a whole.Severe loss of valve margin formation was observed by scanning electron microscopy(Fig.1G and H).Though an LLB was visible in all pods,a slight boundary between the exocarp of two valves was visible in some OR88 pods (Fig.1C).Apart from the LLB,the lignified layer,separation layer,and endocarp b in OR88 were normally developed as in No.87(Fig.1).Even though the valves of OR88 detach from the replum along the dehiscence zone after full maturity,the whole pod maintains integrity because the two valves are still tightly linked through LLB (Fig.1E,F),preventing seed dispersal.Eight other restorers of Ogura CMS and two radish cultivars did not show the LLB structure.

3.2.Genetic analysis of LLB and correlation with pod shatter resistance

None of the 50 F1plants showed an LLB irrespective of the cross direction.Among 222 plants in the F2population,161 had no LLB and 61 had LLB pods (Table 1).This segregation fits a 3:1 ratio(χ2=0.73 ＜χ2(0.05,1)=3.84).In the BC1population,860 plants were observed,of which 435 plants had normal pods and 425 plants had LLB pods.This phenotypic segregation fits the expected ratio of 1:1(χ2=0.12 ＜χ2(0.05,1)=3.84)(Table 1).These results indicated that the LLB structure was controlled by a single recessive nuclear gene.

Table 1 Genetic analyses of lignified-layer bridge (LLB) structure in parental lines and genetic populations.

Pod shatter resistance was assessed in 222 plants of the F2population.All 61 plants with LLB showed high pod shatter resistance with mean SRI of 0.86,whereas all 161 plants without LLB showed much lower pod shatter resistance with SRI ＜0.5 (Fig.1I).In a DH population derived from the same cross,SRI of DH lines with and without LLB in 2019 differed markedly.The mean SRI of 47 DH lines possessing LLB pods was 0.98 and that of 160 DH lines without LLB was 0.45.Thus,all F2plants or DH lines with LLB pods had high SRIs like that of OR88,and all F2individuals and DH lines without LLB pods had much lower SRIs like that of No.87.

3.3.LLB differentiation occurs at stage 12 of gynoecium development

To identify the development stages of LLB formation,anatomical examination of developing gynoecium of OR88 and No.87 from stage 10 to 13 was performed (Fig.2).The boundary between valves became visible with a hollow in No.87 at stage 12(Fig.2C).In contrast,there was no visible boundary between the two valves in the OR88 gynoecium.Instead,a bulge at the position corresponding to the valve margin was formed (Fig.2G).At stage 13,the difference of valve margin was more evident in OR88 and No.87.The hollow in No.87 and bulge in OR88 at the junction of two valves became distinct.Except for the boundary of lignified layer and valve margin,no other cellular structural difference was detected(Fig.2).Thus,the differentiation of LLB occurs at stage 12 of gynoecium development.

Fig.1.Pod cross-section structure of parental lines and distribution of SRI in F2 population of No.87×OR88.Cross section of stage 18 fruit of No.87(A)and OR88(B,C).Cross section of OR88 fruits from stage 18 (D) to stage 19 (E,F).Scanning electron microscopy of valve margin of stage 17 fruit of OR88 (G) and No.87 (H).In No.87 × OR88 F2 population,plants with LLB have high SRI value while plants without LLB have low SRI value(I).Arrow in A shows the boundary of lignified layer of two valves,and in(B–F)indicates lignified-layer bridge.Arrow head in (G) and (H) indicates valve margin.LL,lignified layer,SL,separation layer,en b,endocarp b.Scale bars,200 μm in (A–F) and 500 μm in (G,H).

Fig.2.Lignified-layer bridge occurs at stage 12 of gynoecium development.Transverse sections of toluidine blue-stained No.87 fruit (A–D) and OR88 fruit (E–H).Development stage is marked on the upper row.Arrows indicate the difference between No.87 and OR88.Scale bars,100 μm.

3.4.BnLLB was mapped to C09 by BSA-Seq

A total of 132.9 Gb clean reads were obtained from sequencing the genomes of the two parental lines and two DNA pools of the No.87×OR88 F2population.A total of 1,744,056 SNPs were identified between the two bulked pools (Table S2).Using the ΔSNP-index strategy,a 36.14-Mb region from 0 to 36,140,000 bp in chromosome C09 was found to contain the BnLLB locus(Fig.3A).

3.5.Narrowing the BnLLB locus region using KASP markers

Twenty-nine KASP markers near the BnLLB locus were developed,covering the whole 36.14-Mb BnLLB candidate region.Only 10 of the 29 markers could be assigned to a linkage group containing BnLLB.As a result,BnLLB was further delimited to the region of S13.56–S21.04 after screening 2102 BC1plants revealed no further recombinants in this region(Fig.3B).This severe meiotic recombination deficiency may result from radish introgression.To create recombinants in the candidate region,another Ogura CMS restorer line OR5(without LLB pods)was crossed with OR88 and an F2was constructed.More recombinants in the region were identified.Seven KASP markers were found to be linked to BnLLB in this population and used to construct a linkage map of OR5-OR88 in this region (Fig.3B).Finally,BnLLB was mapped between the markers S12 and S12.688,which corresponds to a 688 kb region in the Darmor-bzh reference genome and a 236 kb region in the radish reference genome [51].

3.6.RNA-Seq and candidate gene identification

Fig.3.Map position of BnLLB.(A)candidate region of BnLLB gene based on delta SNP by BSA-Seq.Arrow indicates the candidate region.(B)Linkage map of No.87-OR88 BC1 population and OR5-OR88 F2 population using KASP markers.Map distance (cM) is shown at left and markers at right.

A total of 163,992,480 and 159,220,470 clean reads (24.6 and 23.88 Gb) were obtained from OR88 and OR5,respectively.After redundant UMI adapter sequences were removed,more than 84%of clean reads were mapped to the Darmor-bzh genome (Table S3).A set of 16,878 genes showed differential expression(Fig.4A).Differentially expressed genes(DEGs)in OR88 and OR5 were enriched in cellular process (GO:0009987,P=1.244E 24) and cellular metabolic process (GO:0044237,P=1.729E 28) within biological process(Fig.4B).Cell cycle(GO:0007049,P=0.011)and regulation of cyclin-dependent protein kinase activity (GO:0000079,P=0.013)were down regulated in OR88(Figs.4C and S1).

There were 43 genes in the map intervals referred to the Darmorbzh and ZS11 [52] reference genomes including four genes located in mapping intervals referred to ZS11 genome but in the random scaffold of the Darmor-bzh genome,and 45 genes in mapping intervals referred to the radish genome.Among them,16 genes showed significant differential expression and 12 genes showed sequence alteration (Table 2).Only BnaCnng71510D which encodes a TCP8,a member of the class I TCP(TEOSINTE BRANCHED 1[TB1],CYCLOIDEA [CYC],and PROLIFERATING CELL FACTOR [PCF]) family transcription factors was hypothesized to be associated with gynoecium development [53] and involved in the cell cycle pathway[54].Thus,we propose that BnTCP8.C09 is candidate for LLB.

Table 2 Information of genes differentially expressed or showing sequence differences between OR88 and OR5 within the BnLLB mapping interval.

3.7.Sequence and transcriptional analysis of BnTCP8.C09

BnTCP8.C09 was cloned from OR88 and OR5 by PCR amplification.Even though many amino acid changes were identified in BnTCP8.C09 between OR88 and OR5,the conserved TCP domain is identical between OR88 and OR5 (Fig.5A).However,the transcription level of BnTCP8.C09 in OR88 was severely reduced as confirmed by semi-quantitative RT-PCR (Fig.4E).The TATA box and CAAT box profiles of the BnTCP8.C09-OR88 promoter differed greatly from those of OR5(Fig.5B).The CAT box(involved in meristem expression) and MSA-like cis-element,involved in cell cycle regulation,were present only in BnTCP.C09-OR88 promoter.These variations in BnTCP.C09-OR88 promoter may result in severely reduced expression of BnTCP.C09-OR88.The amino acid sequence of BnTCP8.C09-OR88 was more similar to RsTCP8 of radish whereas that of BnTCP8.C09-OR5 was identical to BnTCP8.C09(BnaC09G0197900ZS) of ZS11.This finding indicates that BnTCP8.C09 of OR88 is derived from R.sativus,whereas BnTCP8.C09 of OR5 came from B.napus (Fig.4D).A gene-specific marker,TCP.C09-1,for BnTCP.C09 was developed and lies about 4 kb from the BnLLB-linked marker S12.39 (Fig.3).

4.Discussion

4.1.LLB causes high shatter resistance in OR88

Fig.4.Analysis of differentially expressed genes (DEGs).(A) Volcano plot of DEGs in OR88 and OR5.(B) GO term enrichment analysis of DEGs in OR88 and OR5.(C) Cyclin genes associated with cell cycle are downregulated in OR88.(D)Phylogenetic tree of BnTCP8.C09 from OR88,OR5,B.napus var.ZS11(BnTCP8.C09)and radish (RsTCP8).(E)Semi-quantitative RT-PCR analysis of BnTCP8.C09 in OR88 and OR5.

Fig.5.Comparative sequence analysis of BnTCP.C09 protein and promoter region between OR88 and OR5.(A)Alignment of BnTCP.C09 protein sequence from OR88 and OR5.Red underline indicates TCP domain.(B)Comparative diagram of 2000 bp upstream sequence of BnTCP.C09 CDS from OR88 and OR5.Core promoter elements CAAT box and TATA box are differentially present.CAT box(involved in meristem expression)and MSA-like cis-element(involved in cell cycle regulation)are present only in the BnTCP.C09-OR88 promoter.

The lignified layer,separation layer and endocarp b in OR88 pods were well developed.However,there was LLB structure in OR88 pods.After maturity,the two valves could depart from the replum along the separation layer,but could not depart from each other owing to LLB link.Given that high SRI and LLB co-segregated in F2and BC1as well as DH populations,it is clear that the LLB structure in the lignified layer causes high shatter resistance in OR88 by preventing valve separation after maturity.Pod structure variation associated with shattering has been observed in Arabidopsis,Brassica crops and soybean.Most of the variation are related to cell lignification[13,21]or unusual dehiscence zone formation [7,10].Failure of valve margin formation causing a valve bridge was also reported in high pod shatter resistant Brassica mutants [20,23–25].LLB is a new type of pod structure variation associated to pod shattering trait.

4.2.Application of lignified-layer bridge in shatter-resistance breeding

To our knowledge,this is the first report of LLB structure in rapeseed pod.This pod structural character is very useful because the pod is readily threshed owing to the normal separation layer.Except for slightly less number of seeds per pod,the LLB structure tightly associated with high pod shatter resistance seems to have no severe negative effect on agronomically important traits.BnLLB was mapped within a 0.688-Mb region of the Darmor-bzh reference genome.The KASP markers flanking this region or specific to BnTCP.C09 gene are tightly linked to LLB structure and thus the high pod-shatter resistance phenotype.Marker-assisted selection of LLB using these KASP markers will facilitate the development of shatter-resistant cultivars in oilseed rape using OR88 as a resistance donor.

4.3.BnTCP8.C09 is the key gene contributing to the LLB structure

BolC.GA4a (Bol038154) mutant resulted in semi-dwarfing and LLB like phenotype in B.oleracea[23].We investigated the expression profiles of genes involved in the gibberellin pathway in OR88 and OR5.Most of the genes were expressed equally in OR88 and OR5,with some of them upregulated in OR88 (Fig.S2).External application of GA3to developing pods did not restore LLB to the wild type (data not shown).Thus,BnLLB may act downstream of the GA pathway.Among the genes within the mapping interval,only BnTCP8.C09 was hypothesized to be associated with gynoecium development.TCP8 belongs to class I TCP transcription factors which are key regulators of cell proliferation [55].DELLA proteins such as RGA and GAI interact with the TCP DNA-binding motif of class I TCPs to prevent its binding to promoters of target genes.In this way,DELLAs suppress the expression of many core cell-cycle genes controlling cell division regulating plant height in Arabidopsis [54].GO term enrichment analysis and RT-PCR results showed that cell cycle associated genes especially cyclin genes were downregulated in OR88.Even though the conserved TCP domain of BnTCP8.C09 was identical between OR88 and OR5,the transcription level of BnTCP8.C09 in OR88 was severely reduced.We hypothesize that BnTCP.C09 also regulates cell proliferation in the valve margin and that the severe reduction of BnTCP.C09 expression resulted in LLB structure formation.

In summary,we have identified a high pod-shatter resistant B.napus breeding line and shown the link of an LLB to the shatterresistance phenotype.BnLLB was mapped into a genomic region corresponding to a 0.688-Mb region of the Darmor-bzh genome.A candidate gene BnTCP8.C09 with sequence alteration in the promoter region in the shatter-resistant line was identified.Based on QTL mapping and candidate gene identification,DNA markers linked to shatter-resistance gene were developed and are ready for use in breeding programs.These results will facilitate the exploitation of LLB in the development of rapeseed cultivars suitable for modern agricultural production practice,and also shed light on the genetic mechanism of pod-shatter resistance in B.napus.

CRediT authorship contribution statement

Wen Chu:Writing– original draft,Investigation,Validation,Data curation.Jia Liu:Writing–review&editing.Hongtao Cheng:Writing–review&editing.Chao Li:Investigation,Validation,Data curation.Li Fu:Investigation,Validation,Data curation.Wenxiang Wang:Investigation,Validation,Data curation.Hui Wang:Investigation,Validation,Data curation.Mengyu Hao:Investigation,Validation,Data curation.Desheng Mei:Writing– review &editing.Kede Liu:Writing– review &editing.Qiong Hu:Conceptualization,Project administration,Funding acquisition,Writing–original draft,Writing– review &editing.

Declaration of competing interest

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

Acknowledgments

This work was supported by the National Natural Science Foundation of China(U19A2029),the National Key Research and Development Program of China (2018YFE0108000),Science and Technology Innovation Project of the Chinese Academy of Agricultural Sciences (CAAS-ZDRW202105),and China Agriculture Research System of MOF and MARA.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.09.005.