Introduction

Brassica rapa (2n = 20, AA genome) has the diploid Brassica A genome and includes important species of vegetables that are widely cultivated around the world. The plant shows diverse morphological variations, with three well-defined cultivated groups: the leafy-type (Kimchi cabbage and pakchoi), turnip-type (vegetable turnip), and oil-type (yellow sarson) cultivars. In addition to its agricultural and economic importance, B. rapa is one of six Brassica species that have been excellent models for studying genome hybridization and polyploidy evolution (UN 1935). Because of the importance of B. rapa as a crop and a scientific model, a variety of genetic studies has been reported so far. In the early generations of the genetic mapping, random amplified polymorphic DNAs (RAPDs), restriction fragment length polymorphisms (RFLPs), and amplified fragment length polymorphisms (AFLPs) markers were used. More recently, markers such as simple sequence repeats (SSRs), sequence-tagged sites (STS), insertions and deletions (InDels), and single nucleotide polymorphisms (SNPs) have also been used to generate genetic maps. Using these markers and genetic maps, multiple traits of B. rapa including disease resistance, flowering time, plant height, agronomic and morphological traits, root and seed coat color, self-incompatibility, microspore culture ability, mineral accumulations, fatty acid composition, and glucosinolates have been surveyed (reviewed in Ramchiary and Lim 2011). Furthermore, the recent completion of the sequencing of B. rapa genome allows for high-throughput marker development (The Brassica rapa Genome Sequencing Project Consortium 2011).

Even with these large numbers of genetic linkage maps, there are few maps that mainly consist of gene-based or sequence-tagged markers. Thus, to perform quantitative trait locus (QTL) analysis and comparative mapping for map-based cloning, the extensive development of gene-based markers is required. In this regard, SNPs identified in the coding sequences are good targets for high-throughput gene-based marker development. SNPs are the most abundant type of DNA polymorphisms. For example, several million SNPs were characterized in humans (Frazer et al. 2007) and more than 140,000 SNPs were identified in Arabidopsis thaliana (Zhang et al. 2011). With recent advances in sequencing technology, including next generation sequencing (NGS), the amount of SNPs that has been identified is rapidly increasing. One of the major obstacles in using SNP markers has been the efficiency and cost of genotyping compared to assessing other types of polymorphisms. Recently, cost-effective and high-throughput SNP genotyping methods such as the GoldenGate assay have been developed. The GoldenGate assay was proved to be efficient in high-throughput genotyping as shown in barley, soybean, wheat, maize, and rice (Rostoks et al. 2006; Hyten et al. 2008; Akhunov et al. 2009; Yan et al. 2009; Chen et al. 2011). The construction of SNP genetic linkage maps with high-throughput genotyping technology can greatly facilitate molecular breeding.

One of the biggest challenges in breeding B. rapa has been finding ways to introduce durable disease resistance into agronomically important cultivars because its susceptibility to multiple pathogens hinders its cultivation and reduces its produce yield and quality (Mun et al. 2009). Brassica crops are usually infected by a mixture of strains and TuMV-C4 is the most prevalent among the infecting pathotypes in Korea and Taiwan (Suh et al. 1995). Turnip Mosaic Virus (TuMV), which belongs to the Potyviridae family, causes serious economic damages and has long been the focus of disease resistance studies of Brassica crops. Multiple TuMV pathotypes, such as C1-4 (Provvidenti 1980), C5 (Green and Deng 1985), and Tu1-7 (Liu et al. 1990a, b) have been reported. Jenner and Walsh (1996) examined 124 TuMV isolates and divided them into 12 groups (Jenner and Walsh 1996). Several TuMV resistance loci have been mapped in Brassica. In B. napus, TuRB01 and TuRB02 (N6, dominant; Walsh et al. 1999), TuRB03 (N6, dominant; Hughes et al. 2003), and TuRB04 and TuRB05 (dominant; Jenner et al. 2002) have been identified. In B. rapa, TuRB01b (A6, dominant), ConTR01 (R8, dominant) and retr01 (R4, recessive; Rusholme et al. 2007, Rnt1 (A6, recessive; Fujiwara et al. 2011), and retr02 (A4, recessive; Qian et al. 2013) have also been reported. Through molecular genetic analyses, numerous disease resistance (R) genes that confer dominant disease resistance against viruses have been identified. Among the R genes, the largest class is the NBS–LRR genes, which have conserved motifs of a nucleotide-binding site (NBS) and a leucine-rich repeat (LRR) region. Previously reported NBS–LRR class dominant plant virus R genes include HRT, N, RCY1, Rx1, Rx2, and Sw5 (Kang et al. 2005). In addition to these R genes, several recessive resistance genes have also been isolated. Among these recessive genes, eukaryotic initiation factor 4E (eIF-4E) plays an important role during the infection of viruses in the Potyviridae family. In A. thaliana and B. rapa, disrupting the functions of eIF-4E and eIF-4G resulted in resistance to TuMV (Duprat et al. 2002; Lellis et al. 2002; Yoshii et al. 2004; Gallois et al. 2010; Kim et al. 2013). However, there have been few reports on the molecular identity of the dominant TuMV resistant loci.

Previously, we reported the genome-wide discovery of DNA polymorphisms in B. rapa through the re-sequencing of STS among eight varieties including TuMV-resistant VC40 and TuMV-susceptible SR5 (Park et al. 2010). The goal of the current study was to develop co-dominant genetic markers for the genes of the B. rapa genome and to apply these markers for the genetic mapping of the TuMV resistance locus (TuMV-R) in B. rapa. We have emphasized the development of gene-based SNP markers for high-throughput genotyping along with InDel and SSR markers to construct a genetic map of the VCS-13M doubled haploid (DH) population, a DH population of a VC40 and SR5 cross. The resulting genetic map successfully mapped a novel dominant TuMV resistance locus to the short arm of chromosome A6. The mapped region contains several disease resistance-related genes such as NBS–LRR genes and pathogenesis-related-1 (PR-1) genes. This information will be helpful in facilitating the map-based cloning of the target gene.

Materials and methods

Plant materials

DH lines carrying TuMV resistance were generated by microspore cultures. In brief, crosses were performed to generate plants harboring multiple disease resistance. Plants showing desired resistant traits were selected and were subjected to microspore cultures. A plant from the DH population without self-incompatibility, VC40, was chosen and crossed with SR5. The characteristics of VC40 and SR5 are described in Park et al. (2010). The basic mapping population VCS-13M DH consisted of 83 DH progeny derived from a F1 plant of the VC40 and SR5 cross. A Tb1M DH population (n = 106) was also made from microspore cultures of a cross between VCS3M-12 (a sibling line of VC40 and TuMV-resistant plants) and Bulam 1M-68 (which is TuMV-susceptible), and was also genotyped for the mapping of TuMV-R. Relationships of these populations are shown in Supplementary Figure S1.

SNP discovery and genotyping

For SNP discovery from genes, we used genome-wide polymorphisms identified in eight varieties of B. rapa (Park et al. 2010) which contain VC40 and SR5, parental lines of the VCS-13M DH population. SNPs suitable for the GoldenGate assay were selected following the work flow shown in Fig. 1. The GoldenGate assay method was used for the genotyping of two DH populations. Loci with designability scores >0.9 as calculated by the Assay Design Tool (ADT) were selected for the GoldenGate assay. Genomic DNA was extracted using a DNeasy Plant DNA kit according to the manufacturer’s protocol (Qiagen, Hilden, Germany) and 250 ng of genomic DNA was used for the assay. The GoldenGate genotyping assay was conducted by Illumina Inc. according to the manufacturer’s recommended protocol. Allele calling for each locus was performed with the GenomeStudio software (Illumina, San Diego, CA, USA).

Fig. 1
figure 1

A workflow to select markers for genetic map construction. Single nucleotide polymorphism (SNP) markers for GoldenGate assay were selected from the SNP information (marked with asterisk) in Park et al. (2010). Gel-based markers include InDels and SSRs. The resulting 661 polymorphic markers were used for construction of the genetic map of the VCS-13M DH population

Full size image

High-resolution melt (HRM) analysis

A total of 172 primer pairs used in HRM analysis were designed using Beacon Designer 7.5 software (Premier Biosoft International, Palo Alto, CA, USA). PCR was carried out on a CFX96 cycler (Bio-RAD Laboratories, Hercules, CA, USA) in white-wall 96 well PCR plates with 20 μl of reaction mixture containing 0.5 μM of each primer, SYBR GreenER (Bio-Rad Laboratories), and 20 ng of genomic DNA. The cycling conditions were one cycle at 95 °C for 2 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 45 s. The melting curves of the PCR amplicons were measured from 65 to 95 °C. Data acquisition was conducted at every 0.2 °C increase for 10 s. Melting curves were analyzed with high-resolution melting analysis software (Bio-Rad Laboratories). Each test was replicated three times and automatically clustered according to their melting profiles.

InDel and SSR marker development and application

PCR-based InDel and SSR markers were developed based on the re-sequencing data of VC40 and SR5. Approximately 50X sequences of 100 bp paired end reads were generated from each line by Illumina GA IIX (Illumina). The paired-end sequence reads were aligned to the reference sequences of B. rapa cv. Chiifu using CLC Assembly Cell 3.2 (CLC Bio, Aarhus, Denmark), with a minimum similarity fraction of 0.9, a minimum length fraction of 0.9, a maximum of two mismatches, and paired-read mode options. The reads that mapped to multiple locations were excluded. InDels and SSRs were identified by the alignment of the reads to the draft genome sequences of B. rapa cv. Chiifu. Amplicons with more than 6 bp differences in length between parental lines were selected for PCR amplification. Primer 3 (Rozen and Skaletsky 2000) was used to design primers with the following parameters; (1) amplicon size 100–250 bp; (2) Tm 55–60 °C; (3) GC content 40–60 %. Each PCR was performed using Tenuto 2X premix (Enzymonics, Daejeon, Korea), 0.5 μM of each primer, and 20 ng genomic DNA. The PCR cycle conditions were 94 °C for 5 min for one cycle, followed by 35 cycles of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 1 min. The PCR products were separated on 3 % agarose gel in 1X sodium borate buffer. In addition, 239 SSR markers developed by KBGP were also used to screen for polymorphisms between the VC40 and SR5 lines. The fragment analysis was performed on an ABI 310 Genetic Analyzer (Life Technologies, Carlsbad, CA, USA) using the POP-4 polymer and a 47 cm capillary. Allele sizing was analyzed with GeneMapper software 3.7 (Life Technologies).

Evaluation of TuMV resistance and the genotyping of TuMV-R

The isolate of TuMV-C4 that was stored in liquid nitrogen was revived by inoculating it into B. juncea leaves. Inoculated leaves that showed disease symptoms were homogenized in a mixer and re-suspended in inoculation buffer (1 % potassium phosphate, pH 7.0). Carborundum powder was used to break the leaf surface for the facilitated infection of the virus. Disease resistance was evaluated with the third and fourth leaves of the 20-day-old young plants. The second inoculation was performed 1 week after the first inoculation. The disease response of the inoculated plant was visually assessed 25 days after the second infection and was scored as either resistant (R) or susceptible (S). Susceptible plants showed symptoms such as mosaic light- or dark-green coloration, whereas resistant plant showed no symptoms compared to the uninfected plant. When the disease symptom was too ambiguous to record, the enzyme-linked immunosorbent assay (ELISA) test was performed to verify resistance and susceptibility (Walsh et al. 1999; Zhang et al. 2008). For the genetic mapping of TuMV-R, TuMV resistance was examined for the plants from the VCS-13M DH population as described above. These results were used as the genotype of TuMV-R and its linkages were analyzed together with the markers developed in this study. The same method was used for the mapping of TuMV-R in the Tb1M DH population.

Genetic map construction and alignment of linkage groups to the reference sequences

Using genotypes of each marker in the VCS-13M DH population, a genetic map was constructed by JoinMap ver. 4.0 (van Ooijen 2006). A LOD threshold of 4.0 was used for the assignment of linkage groups. The Kosambi map function (Kosambi 1944) was used to calculate the genetic distance between markers. The physical position of the markers linked on the genetic map of the VCS-13M DH population was determined using sequence information of the markers on the reference genome sequence of cv. Chiifu. Consensus sequences of the SNPs and primer sequences of the gel-based markers were analyzed using a BLASTN search. The best match with an E-value cut-off of E−20 and sequence coverage >90 % was chosen to match the genetic markers to the genome sequence. The TuMV-R region, as defined by flanking markers, was analyzed using the reference sequence information of cv. Chiifu. Detailed lists of the marker attributes developed in this study are given in Supplementary Tables S1 to S3.

Results

SNPs selection strategy and genetic marker development

With the goal of constructing a core genetic map of the VCS-13M DH population to map a TuMV-R locus based on high-throughput genotyping, we focused on SNPs in the genome. We previously reported the identification of genome-wide polymorphisms in the genic regions of eight varieties of B. rapa (Park et al. 2010). Among 1,398 STS, we examined 21,311 SNPs suitable for high-throughput genotyping by the GoldenGate assay. The workflow of the application of the markers is presented in Fig. 1. Consensus sequences with more than two SNP/InDels within 80 bp were filtered out and the resulting 15,285 SNPs were analyzed by ADT. There were 10,699 SNPs remaining which had SNP scores >0.9 and designability ranks of 1.0. Among these, 3,741 SNPs were polymorphic between the VC40 and SR5 parent lines of the VCS-13M DH populations. Finally, we screened novel SNPs which were not reported in any B. rapa genetic maps, and the resulting 447 SNPs were selected for the GoldenGate assay. The GoldenGate assay successfully genotyped 438 SNPs suitable for genetic map construction. The nucleotide types of the SNPs were A/T (72), A/G (230), A/C (90), and G/C (46). In addition, 10 SNPs were developed as HRM markers.

To assign linkage groups, public SSR markers anchored in JWF3p and VCS3M maps (Supplementary Table S3) were applied. In total, 64 out of 239 SSRs showed polymorphisms between VC40 and SR5 in agarose gel, and were genotyped by fragment analysis. Using these markers, we could construct a skeleton version of the genetic map of the VCS-13M DH population. We also developed 96 primer sets targeting SSRs in chromosome A10 using in silico sequence analysis. Among these, 30 (31 %) primers were polymorphic in 3 % agarose gel between parental lines and their genotypes were scored in the mapping population. Thus, each linkage group (A1–A10) was assigned based on at least two previously published SSR markers.

During the construction of the genetic map of the VCS-13M DH population, the draft genome sequence of B. rapa cv. Chiifu was published (The Brassica rapa Genome Sequencing Project Consortium 2011). We developed InDel markers from genes based on the sequence comparison of VC40 and SR5 with the reference sequences. In B. napus, two TuMV resistance-related genes of TuRB01 and TuRB03 were reported in A6 (Walsh et al. 1999; Hughes et al. 2003), whereas a recessive locus for resistance was also reported in A4 (reviewed in Rusholme et al. 2007; Ramchiary and Lim 2011). Thus, to identify possible TuMV-R loci in the VCS-13M DH population, we developed InDel markers from A4 and A6 based on re-sequencing data of VC40 and SR5. A total of 765 InDels in A4 and 1,246 InDels in A6 were identified. Among them, 86 InDels in A4 and 136 InDels in A6 with amplicon sizes of 120–210 bp with 12–48 bp differences between the genotypes were subjected to agarose gel electrophoresis analysis. A total of 204 primers (92 %) amplified a single band in both genotypes and 110 InDels were identified as co-dominant and polymorphic. We integrated all the genotypes to construct a genetic map.

Genetic map construction and comparison with the reference genome sequences

A total of 661 markers including 438 SNP, 110 InDel, 94 SSR, and 10 HRM markers were used to construct a genetic map of the VCS-13M DH population. Markers that showed the same genetic position or remained as singletons after linkage grouping were removed. As a result, the genetic map shown in Fig. 2 was derived from the analysis of unique markers including 221 SNP, 10 HRM, 62 SSR, and 31 InDel markers. A complete list of all markers and their sequence information used in this study are provided in Supplementary Tables S1 to S3. The 314 genetic markers spanned 1,115.9 cM with an average distance between markers of 3.6 cM (Table 1). The biggest linkage group A9 contained 45 markers spanning 157.5 cM, while the smallest group A7 contained 24 markers covering 75.0 cM. Considering that the physical length of B. rapa is 529 Mb (Johnston et al. 2005), the estimated correlation between the physical and genetic distance was 474 kb/cM. To examine whether the genetic map showed genetic distortion, we compared the genetic map with the B. rapa reference genome sequences. The physical position of the markers was determined by BLASTN analysis. Most of the markers in the linkage groups showed co-linearity with their physical positions, especially with A6, which is where the TuMV-R has been reported to exist (Fig. 3). However, the markers were unevenly distributed in chromosome A2 and A3, indicating that there may be a biased distribution of SNP loci or a rearrangement of genome blocks in the VCS-13M DH population.

Fig. 2
figure 2

Genetic map of the B. rapa VCS-13M DH population. Ten linkage groups were assigned based on the SSR markers previously reported. Map distances calculated by Kosambi map units (cM) are shown on the left margin and the marker names are shown on the right margin of each linkage group. Marker types are indicated with the following prefixes: N, SNP; H, SNP using HRM; KS and BrSSR, SSR; CUK, InDel marker. In A6, markers close to TuMV-R are indicated in a blue bar

Full size image
Table 1 Summary of the genetic map constructed using the B. rapa VCS-13M DH population
Full size table
Fig. 3
figure 3

Correlation between the genetic map of the VCS-13M DH population and the draft genome sequence of B. rapa cv. Chiifu. The physical position of each marker was identified in the reference genome based on a BLASTN search

Full size image

Inheritance of TuMV resistance in the VCS-13M DH population

The VCS-13M DH population was generated by a cross between VC40 and SR5 to survey multi-disease resistance in B. rapa. VC40 is resistant to TuMV and club root (CR) but susceptible to soft rot, while the opposite is true for SR5. An ancillary goal of this study was the mapping of a TuMV-R locus to the VCS-13M DH population. Toward this end, we examined the inheritance of TuMV resistance in the VCS-13M DH population. As shown in Table 2, the TuMV resistance was surveyed in parental lines, F1, backcross lines, and F2 populations and was scored as either resistant (R) or susceptible (S). In the case of the parental lines, VC40 plants showed resistance to TuMV, while the SR5 plants were scored as susceptible. All of the F1 plants were scored as resistant, indicating that TuMV resistance is inherited in a dominant manner. In the case of the F2 population, a 3:1 segregation was observed (χ 2 = 0.27, p = 0.05), suggesting that TuMV resistance is caused by a single dominant locus (Table 2).

Table 2 Genetic analysis of TuMV-C4 resistance using P, F1, F2, and BC of VCS-13M population
Full size table

Mapping of a TuMV resistance locus in the VCS-13M DH population

To map the TuMV-R locus, TuMV disease resistance was examined for the plants of the VCS-13M DH population and the genotypes of the population were denoted as VC40-type (R) or SR5-type (S). Because the DH line is homozygous for all loci, the phenotype of the plant can be regarded as the genotype. Using the genotyping data, we analyzed linkage between the markers and TuMV-R using the JoinMap 4.0 program at LOD scores of 4.0. With these methods, TuMV-R was mapped to the short arm of linkage A6, which has 40 framework InDel markers. We could find two makers closely linked to TuMV-R, CUK_0023i (6.0 cM) and N0343 (9.6 cM) in the short arm of A6 (Fig. 2). However, we could not anchor a region of the locus due to the insufficient and unknown phenotype data of the TuMV resistance screen. Therefore, to define a region that contains TuMV-R, we used the Tb1M DH population, which consisted of 106 individuals derived from a cross between VCS3M-12 (TuMV-resistant) and Bulam 1M-68 (TuMV-susceptible) DH lines. We examined whether the Tb1M DH population could be used to define a TuMV-R locus. First, we applied the SNP markers that were mapped to the VCS-13M DH population using the GoldenGate assay and then constructed a genetic map. Second, we used TuMV resistance data of the Tb1M DH population to map a TuMV-R locus. As shown in Fig. 4 and Supplementary Table S4, N0343 was identified as the closest marker linked to the TuMV-R locus on A6 of the Tb1M DH population, which was similar to the case of the VCS-13M DH population (Figs. 2, 4). This finding highly suggested that the TuMV-R locus may be identical in both populations. We applied InDel markers which were mapped to A6 of the VCS-13M DH to narrow down the region in the Tb1M DH population and found that CUK_0040i was the flanking marker of the TuMV-R locus together with N0343. Sequence comparison revealed that the physical position of N0343 was 1.66 Mb from the top of chromosome A6 and that of CUK_0040i was 2.00 Mb and a total of 56 genes were identified in the 0.34 Mb intervals between the two markers. Interestingly, this region contains four genes encoding CC–NBS–LRR and two PR-1 genes in the B. rapa reference genome (Fig. 4 and Supplementary Table S4). Sequence comparison between the genotypes (Chiifu, VC40, and SR5) identified non-synonymous sequence substitutions from all six genes (Supplementary Figure S2). The putative function of these genes might be related with resistance against TuMV. We compared the sequences and physical positions of the flanking makers with the previously reported TuMV resistance genes but could not found any relationships, suggesting that the TuMV-R locus mapped in this study is novel.

Fig. 4
figure 4

Detailed mapping of the TuMV-R locus on chromosome A6 using the Tb1M DH population. Genetic distance in cM was calculated using the Kosambi function. The genetic map of the Tb1M DH population is shown on the left panel and the corresponding physical position of the marker is indicated on the right panel. Genes related with disease resistance within the marker-defined region are shown on the right of the physical map

Full size image

Discussion

B. rapa is one of the most valuable vegetable crops worldwide and there has been considerable effort invested into introducing durable disease resistance into agronomically important cultivars. In this study, we generated a genetic map of the VCS-13M DH population which segregated two important disease resistance traits, TuMV and CR resistance. Because cv. Chiifu, the reference species of B. rapa genome sequencing, is susceptible to various pathogens including TuMV, the genetic map constructed in this study using markers developed from the genic regions will be useful for diverse populations of B. rapa as well as for functional studies of disease resistance. Furthermore we defined a region that contains a novel TuMV-R locus along with putative resistance-related gene candidates.

Several genetic maps of B. rapa have been reported so far. For example, the JWF3p (Kim et al. 2006), CK (Kim et al. 2009), and VCS3M (Seo et al. 2013) maps were used for a genome sequencing project with B. rapa. In addition, Liu et al. (2013) developed a genetic map with sequence-based InDel markers. Most of the markers used in these maps were PCR-based markers which have limited throughput. One of the main objectives of this study was to construct a genetic map based on the high-throughput genotyping of gene-based markers which can be easily applied to other populations. For this purpose, we used SNP markers in the genic sequences that were identified in our previous genome-wide study of polymorphisms in eight varieties of B. rapa (Park et al. 2010). We found that 98 % (438 out of 447) of SNPs genotyped by the GoldenGate analysis were identical to the re-sequencing data of the parental genotypes, suggesting that the high-throughput SNP genotyping was precise. However, the constructed map of the VCS-13M DH population contained only 221 loci, mainly because redundant markers were clustered or removed from the map. Other possible reasons for the small number of loci include the small population size and the genomic structures of VC40 and SR5. It has been reported that genic regions in the B. rapa genome were not evenly distributed and comprised only 29.0 % of the 283 Mb genome assembly (The Brassica rapa Genome Sequencing Project Consortium 2011). Ramchiary et al. (2011) also reported that EST–SSR markers were not evenly distributed and were less polymorphic compared with genomic SSRs. However, with the completion of the B. rapa genome project, the uneven distribution of markers can be overcome by developing markers based on the genomic information. In our results, the order of the markers showed good agreement with the physical map (Fig. 3). Moreover, we have documented the successful genotyping of markers to the Tb1M DH population without analysis of sequence variation between the parental lines. This result suggests that the gene-based markers developed in this study have high potential for application to the genetic study of B. rapa. Therefore, we anticipate that the polymorphisms identified in our previous study (Park et al. 2010) and the genetic markers developed in this study will be easily applied to other populations, which can provide valuable resources for the marker-assisted breeding (MAB) of B. rapa.

To date, several loci that confer TuMV resistance in Brassica species have been reported. Among these, TuRB01 and TuRB03 in B. napus (Walsh et al. 1999; Hughes et al. 2003) and TuRB01b in B. rapa (Rusholme et al. 2007) are known to exist on A6. In this study, we mapped a TuMV-R locus in the VCS-13M DH and Tb1M DH populations, which exists in a 1.66–2.00 Mb region from the top of A6 (Fig. 4). Comparisons of marker sequences with the reference genome indicated that the TuMV-R was novel and not allelic to known resistance loci. In the case of TuRB01 (Walsh et al. 1999), the physical position of the linked marker pO120b is 18.8 Mb based on the BLASTN analysis against B. rapa genome ver 1.5. In addition, sS1949a and sORE87, which are flanking markers of TuRB03 (Hughes et al. 2003), were located in the 18.5 and 20.9 Mb regions of A6, respectively. We could not define the position of TuRB01b due to the lack of marker sequences. However, we anticipate that TuRB01b is possibly allelic to TuRB01 or that they are members of a resistance gene cluster described in a previous report (Hughes et al. 2003). Therefore, the three known loci are not likely allelic to the TuMV-R of the VCS-13M DH population. Although it is still possible that unmapped loci such as TuRB04 and TuRB05 could be identical to TuMV-R, the locus identified in this study is likely to be a novel TuMV resistance locus in B. rapa.

Molecular mechanisms of TuMV resistance have been studied in the context of eukaryotic initiation factor 4E (eIF-4E). The protein interaction between Potyvirus genome-linked protein (VPg) and eIF-4E was shown to be important for viral infection (Ruffel et al. 2002). In B. rapa, ConTR01, retr01 (Rusholme et al. 2007), and retr02 (Qian et al. 2013) were mapped to the regions containing eIF-4E/(iso)4E genes. However, the roles and molecular mechanisms of NBS-LRRs in TuMV resistance have not been well studied in B. rapa. In Arabidopsis, TuNI causes systemic veinal necrosis upon infection with TuMV and among the 15 candidate genes for TuNI, five of them were CC–NBS–LRRs (Kaneko et al. 2004). Kaneko et al. (2004) reported that the induction of systemic necrosis was observed in the lines containing TuRB01 and TuRB03 in B. napus, providing a clue for understanding the pathogenesis and resistance of TuMV. Moreover, a novel microRNA that was induced by TuMV infection was reported to target TIR–NBS–LRR in chromosome A3 of B. rapa (He et al. 2008). In our study, the 0.34 Mb region of TuMV-R was found to contain disease resistance-related gene candidates such as four CC–NBS–LRRs genes and two PR-1 genes. Given that many NBS–LRRs belong to dominant R genes in the disease response, CC–NBS–LRR genes can be plausible candidates for TuMV resistance. Meanwhile, it was reported that the expression of PR-1 proteins enhanced resistance to the Tobacco mosaic virus, Tobacco necrosis virus, and Tomato bush stunt virus in Nicotiana plants (Cole et al. 2004). Thus, the PR-1 genes in this region can also be related with TuMV-R. Interestingly, this region showed approximately two-times higher sequence variation (one SNP per 129.1 bp for VC40 and one SNP per 98 bp for SR5) than other genome regions (one SNP per 272.9 bp for VC40 and one SNP per 219.8 bp for SR5) in re-sequencing data of VC40 and SR5 compared to the reference genome sequences. Sequence divergence might result in the differentiation of gene expression or function in this region. Further study is needed to reveal the molecular nature of TuMV disease resistance in the VCS-13M DH population.

Application of SNP markers and high-throughput genotyping successfully mapped a novel TuMV-R locus in the genome of B. rapa. The current work has contributed to an increased knowledge of gene-based marker development and rapid map construction using a high-throughput genotyping strategy. Based on the genome sequence information, we can narrow down the possible candidate genes for TuMV resistance. We anticipate that this genetic map can have a crucial utility for the map-based cloning of genes that confer TuMV resistance in the VCS-13M population.