Figures
Abstract
SNP markers for QTL analysis of 4-MTB-GSL contents in radish roots were developed by determining nucleotide sequences of bulked PCR products using a next-generation sequencer. DNA fragments were amplified from two radish lines by multiplex PCR with six primer pairs, and those amplified by 2,880 primer pairs were mixed and sequenced. By assembling sequence data, 1,953 SNPs in 750 DNA fragments, 437 of which have been previously mapped in a linkage map, were identified. A linkage map of nine linkage groups was constructed with 188 markers, and five QTLs were detected in two F2 populations, three of them accounting for more than 50% of the total phenotypic variance being repeatedly detected. In the identified QTL regions, nine SNP markers were newly produced. By synteny analysis of the QTLs regions with Arabidopsis thaliana and Brassica rapa genome sequences, three candidate genes were selected, i.e., RsMAM3 for production of aliphatic glucosinolates linked to GSL-QTL-4, RsIPMDH1 for leucine biosynthesis showing strong co-expression with glucosinolate biosynthesis genes linked to GSL-QTL-2, and RsBCAT4 for branched-chain amino acid aminotransferase linked to GSL-QTL-1. Nucleotide sequences and expression of these genes suggested their possible function in 4MTB-GSL biosynthesis in radish roots.
Citation: Zou Z, Ishida M, Li F, Kakizaki T, Suzuki S, Kitashiba H, et al. (2013) QTL Analysis Using SNP Markers Developed by Next-Generation Sequencing for Identification of Candidate Genes Controlling 4-Methylthio-3-Butenyl Glucosinolate Contents in Roots of Radish, Raphanus sativus L. PLoS ONE 8(1): e53541. https://doi.org/10.1371/journal.pone.0053541
Editor: Jinfa Zhang, New Mexico State University, United States of America
Received: August 30, 2012; Accepted: November 29, 2012; Published: January 7, 2013
Copyright: © 2013 Zou et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the Program for Promotion of Basic and Applied Research for Innovations in Bio-oriented Industry (BRAIN), Japan. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Glucosinolates, sulfur-rich plant secondary metabolites derived from amino acids found in Brassicaceae and related families, are important ingredients determining the flavor and taste of Brassicaceae vegetables and have an anti-carcinogenic effect in animals [1]. Glucosinolates have three moieties, i.e., beta-thioglucose, sulfonated oxime, and a variable aglycone side chain derived from an alpha-amino acid. More than 120 glucosinolates are classified into three groups, i.e., aliphatic, aromatic, and indolic glucosinolates, based on the distinct side chain [2], [3].
Recent studies using doubled haploid populations obtained from F1 and BC4F1 plants in oilseed Brassica juncea have revealed some major and minor quantitative trait loci (QTLs) for seed glucosinolate contents [4]. Fine mapping has revealed candidates for the genes controlling seed glucosinolate contents such as GSL-Elong, Myb 28, GSL-ALK, and GSL-PRO located in QTL regions in B. juncea [5]. An integrated linkage map containing amplified fragment length polymorphism (AFLP) markers and simple sequence repeat (SSR) markers has been constructed for QTL analysis of the contents of eight different glucosinolate compounds in leaves of Brassica rapa using a doubled haploid (DH) population, and 16 loci controlling aliphatic glucosinolate accumulation, three loci controlling total indolic glucosinolate contents, and three other loci regulating aromatic glucosinolate contents have been identified [6]. Many important genes participating in the three phases of the glucosinolate biosynthesis pathway, i.e., side chain elongation of precursor amino acids, formation of the core glucosinolate structure, and side-chain decoration, have been identified in Aarbidopsis thaliana and B. rapa 5,7,8. Branched-chain aminotransferase (BCAT) is involved in elongation of methionine, and methylthioalkylmalate synthase (MAM1 and MAM3) is involved in the methionine chain elongation pathway for isopropylmalate synthesis [9]–[13]. Core structure formation of glucosinolate is accomplished in five steps via oxidation of aldoxime by CYP79 and CYP83 families followed by C-S lyase, S-glucosyltransferase, and sulfortransferase [14]–[16]. Isopropylmalate dehydrogenase (IPMDH1) has been identified in the biosynthesis of both glucosinolate and leucine [10], [17], [18].
Radish (Raphanus sativus L., 2n = 2x = 18) is an important root vegetable in the family Brassicaceae and a common crop in East Asia. Genomic research on R. sativus has not progressed as much as that of B. rapa. Several genetic linkage maps of R. sativus have been constructed using restriction fragment length polymorphism (RFLP) [19]. AFLP [20]–[22]. SSR [23], and single nucleotide polymorphism (SNP) markers [23], and some of these have been applied to analysis of QTLs for shape and color of roots, flowering time [25], and disease and pest resistances [20], [22].
Radish roots contain 4-methylthio-3-butenyl glucosinolate (4MTB-GSL) with the common name of glucoraphasatin as a characteristic common glucosinolate [26] and 4MTB-GSL content is an important characteristic of radish cultivars [27]. Although the genes responsible for glucosinolate contents in Brassica seeds have been intensively studied, genetic analysis of glucosinolate contents in radish roots has not been reported. In genetic analysis of some traits influenced by many other factors, use of closely related parental lines is desirable. Since glucosinolate contents in radish roots are influenced by growing conditions and growth stages [28], parental lines having the same cultivation period are suitable for producing F2 populations for genetic analysis. However, for QTL analysis, it is required to construct a linkage map composed of DNA markers without a large gap, and closely related parental lines are not suitable for developing many DNA markers to cover a linkage map. In the present study, we analyzed QTLs for root glucosinolate contents in radish using two F2 populations obtained from a cross between a high 4MTB-GSL inbred line and a low 4MTB-GSL inbred line, both of which are Japanese daikon cultivars having thick roots. For this study, we developed SNP markers using sequence data of bulked PCR products obtained by next-generation sequencing to cover a linkage map. Three QTLs were repeatedly detected, and SNP markers in these QTL regions were developed. Furthermore, three candidate genes were mapped in the QTL regions, and nucleotide sequences and expression of these genes were analyzed.
Materials and Methods
Plant Materials
An inbred line, ‘TBS-2-5-3-2- (3)’ (‘TBS’ hereafter), derived from a radish F1 hybrid cultivar ‘Taibyousoubutori’, which is a major radish cultivar with a green shoulder in Japan, and an inbred line, ‘AZ26H-24-6-5- (3)’ (‘AZ26H’ hereafter), derived from a pungent local cultivar of Akita Prefecture in Japan, ‘Karamijidaikon’, were used as seed and pollen parents. Roots of ‘TBS’ contain lower amounts of glucosinolates than those of ‘AZ26H’. For QTL analysis, two populations of 220 and 180 F2 plants obtained by self-pollination of F1 between ‘TBS’ and ‘AZ26H’ were cultivated in a plastic greenhouse of the NARO Institute of Vegetable and Tea Science, Kusawa 360, Tsu, Mie, Japan, for four months during autumn-spring in 2009–2010 and 2010–2011, respectively.
Analysis of Glucosinolate Contents
A simple, rapid method using a common solvent, methanol/water = 80:20 (V/V), at room temperature was applied to the extraction of crude glucosinolate from the powder of freeze-dried radish roots. Quantification of 4-methylthio-3-butenyl glucosinolate and other types of glucosinolates was done using HPLC [29], [30].
SNP Identification
In our previous study [24], 2,880 primer pairs were designed for specific amplification of coding regions of genes containing 3′-untranslated regions. These primer pairs were divided into 480 groups by MultiPLX 2.0 (six primer pairs in each group) (Table S1) [31]. Genomic DNAs of ‘TBS’ and ‘AZ26H’ were prepared by a modified cetyltrimethyl ammonium bromide (CTAB) method [32]. Multiplex PCR was conducted in a 10-µl reaction mixture with 20 ng of plant genomic DNA, 10 pmol of each primer mixed as a group, 0.5 unit of TaKaRa Ex Taq® DNA polymerase (Takara Biomedicals, Japan), 1 × Ex Taq buffer, and 0.2 mM dNTP each. Thermal cycling conditions were as follows: 94°C, 2-min denaturation, 40 cycles of 94°C for 30 sec, 57°C for 30 sec, and 72°C for 1 min, and a final extension at 72°C for 1 min. All the PCR products of each line were collected and mixed in one tube and then digested with six enzymes separately (AluI, MseI, HaeIII, MboI, MspI, AfaI, New England Biolabs). The digestion products were again mixed and purified using ethanol precipitation. Sequences were determined using Illumina GAIIx. Using the software Bowtie (http://bowtie-bio.sourceforge.net/index.shtml), the obtained sequence reads of ‘TBS’ and ‘AZ26H’ were respectively mapped to reference sequences previously determined by the Sangar method to produce SAM files [24], which were then analyzed by the program SAMtools (http://samtools.sourceforge.net/) to identify SNPs.
SNP Analysis of F2 Plants
Genomic DNAs from 189 and 174 plants of the two F2 populations, which were prepared by the modified CTAB method, were analyzed by a modified dot-blot-SNP genotyping method, the MPMP-dot-blot-SNP method, following Li et al. [24] First, 30 primer pairs of SNP markers were separated into six groups by MultiPLX 2.0 (five primer pairs per group). Next, the probe hybridization conditions were predicted as described by Shiokai et al. [33] based on Tm values estimated by the DINAMelt web server (http://mafold.rna.albany.edu/?q=DINAMelt/hybird2). Multiplex PCR was performed in a 10-µl reaction mixture with 3.5 ng of plant genomic DNA, 10 pmol of each primer mixture, 1x KAPATaq Extra Buffer (without Mg2+), 1.75 mM of MgCl2, 0.3 mM of each dNTP, and 0.25 U of DNA polymerase (KAPATaq Extra). Thermal cycling conditions were as follows: 94°C, 1 min denaturation, 40 cycles of 94°C for 30 sec, 56°C for 30 sec, and 72°C for 1 min, and a final extension at 72°C for 1 min. Amplified PCR products were denatured in a solution of 0.4 N NaOH and 10 mM EDTA and then dot-blotted onto a nylon membrane (Nytran, Pall, NY, USA) using a Multi-pin Blotter (Atto, Tokyo, Japan). DNA fragments on the nylon membrane were hybridized with unlabeled allele-specific oligonucleotides having bridge sequences [34]. Allele-specific signals were detected by hybridization of digoxigenin-labeled oliginucleotides having sequences complementary to the bridge sequences according to Shiokai et al. [33] One SCAR (sequence characterized amplified region) marker was analyzed using 3% agarose gel.
QTL Analysis
A linkage map was constructed using the JoinMap 4.0 software (Kyazma B.V., Wageningen, the Netherlands) for each population [35]. Linked loci were grouped into nine linkage groups (LGs) at high logarithm of odds (LOD) values (≥6). The Kosambi mapping function was used to convert recombination values into map distances. QTL analysis was performed applying a composite interval-mapping analysis with Windows QTL Cartographer v2.5 [36]. The LOD thresholds for QTL significance were determined by a permutation test (1000 replications) with a genome-wide significance level P = 0.05.
Nucleotide Sequence Analysis of Candidate Genes in Glucosinolate Biosynthesis Pathway
By BLASTN analysis of SNP markers in QTL regions with sequences of A. thaliana and B. rapa, homologues in the QTL regions were applied to prediction of genes controlling glucosinolate content [7]. Primers were designed to amplify genes of CYP83A1, CYP79F1, BAT5, BCAT4, MAM3, MAM1, and IPMDH1 from genomic DNA in ‘TBS’ and ‘AZ26H’ (Table S2). The amplification products were purified by UltraCleanTM15 DNA Purification Kit (MO BIO, USA) and were cloned into pGEM-T Easy Vector (Promega). Sequences of six clones for each product were determined with a DNA sequencer (CEQ2000, Beckman Coulter). The sequences were aligned using SEQUENCHER version 4.7 (Gene Codes Corporation, MI, USA). After comparing nucleotide sequences between the parental lines, probes were designed to map these genes on the linkage map using the F2 population. Exons and introns of these genes were inferred by BLAST analysis with EST sequences of R. sativus (http://blast.ncbi.nlm.nih.gov/).
Expression Analysis of Candidate Genes
RNA was extracted from 30 mg roots of ‘TBS’ and ‘AZ26H’ four months after sowing using the SV Total RNA Isolation System (Promega Corp.). One microgram of total RNA from each sample was reverse-transcribed in a 22 µl reaction mixture using SuperScriptTMIII Reverse Transcriptase First-strand cDNA Synthesis (Invitrogen). RT-PCR was performed with specific primers designed from exon regions of BCAT4, MAM3, CYP83A1, CYP79F1, MAM1, and IPMDH1 by 30 cycles of 94°C for 30 sec, 58°C for 30 sec, and 72°C for 30 sec. Real-time PCR was performed to quantify mRNAs of MAM3, IPMDH1, and BCAT4 using a LightCycler (Roche Diagnostics) with SYBR Premix EX Taq (Takara Biomedicals). Specific primer sets were designed for real time PCR and Actin was used as an internal control (Table S2). Thermal cycling conditions were as follows: 30 cycles of 95°C for 5 sec, 60°C for 20 sec, and 72°C for 15 sec.
Results and Discussion
Phenotypic Variation
The contents of 4MTB-GSL of P1 (‘TBS’) ranged from 24.4 to 40.8 µmol/g DW and those of P2 (‘AZ26H’) ranged from 96.3 to 137.7 µmol/g DW. Average contents of 4MTB-GSL of P1 and P2 were 30.5 and 120.0 µmol/g DW, respectively. The percentage of 4MTB-GSL in the total glucosinolates was more than 96%. The 4MTB-GSL contents in radish roots of the F2 plants harvested in 2010 were distributed from 5.5 to 130.6 µmol/g DW with an average of 60.7 µmol/g DW (Figure 1A). The frequency distribution of 4MTB-GSL contents in the F2 population showed a continuous, bell-shaped distribution, the contents from 50 µmol/g DW to 100 µmol/g DW accounting for 76.8%. The 4MTB-GSL contents of the F2 plants harvested in 2011 ranged from 6.6 to 103.6 µmol/g DW and the average content was 45.6 (µmol/g DW) (Figure 1B). Two inbred parental lines contained significantly different 4MTB-GSL content in radish roots that provided ideal material for QTL analysis.
(A) F2 population grown in 2010, (B) F2 population grown in 2011.
Production of SNP Markers by Nucleotide Sequencing of Bulked PCR Products
Nucleotide sequencing of multiplex PCR products amplified with 2,880 primer pairs using an Illumina sequencer determined sequences of 2,301 and 2,328 fragments of ‘TBS’ and ‘AZ26H’, respectively. The numbers of read bases of ‘TBS’ and ‘AZ26H’ were 175 Mb and 140 Mb, respectively. Comparison of sequence data of 1,777 fragments between ‘TBS’ and ‘AZ26H’ revealed 2,655 possible SNPs. Among them, SNPs with more than five repetitive reads were regarded as credible SNPs, and 1,953 SNPs in 750 DNA fragments (Table 1, S1), 437 of which had been previously mapped in a linkage map, were identified. Among them, 133 SNPs were selected as markers, because probes of dot-blot-SNP markers for these SNPs have already been produced (Table S3) [24]. Twenty-five markers were screened by analysis of the polymorphism between the parental lines from the EST-SNP markers on the high-density linkage map of our previous study (Table S4) [24]. To fill gaps in a linkage map, 30 newly identified SNPs in DNA fragments, which had been mapped previously, were used for developing probes of dot-blot-SNP markers. SNPs identified by new-generation sequencing enabled efficient screening of SNP markers for the parental lines used for producing a segregating population. Furthermore, new SNPs can be identified for developing oligonucleotide probes that had been positioned in gap regions of a linkage map.
Next-generation sequencing technologies have been applied to gene discovery, transcript quantification, marker production, and so on [37]. In Brassica napus, for example, about 40,000 SNPs have been discovered by Solexa-based transcript sequencing of two cultivars, and reads have been analyzed by mapping them onto a reference database of unique genes retrieved from publication archives [38]. The number of SNP markers identified by next-generation sequencing of bulked PCR production by 2880 primer pairs, i.e., 750, was more than sufficient for QTL analysis, and 1,152 primer pairs might be sufficient for producing SNP markers only for QTL analysis. The method used in the present study is simpler and more efficient for producing SNP markers than transcript sequencing. Furthermore, primer pairs used for multiplex PCR in sequence analysis can be used for producing SNP markers. The primer mixture for multiplex PCR would be useful for identification of SNPs between other parental lines for QTL analysis of radish and Brassica.
Linkage Map Construction of Radish for QTL Analysis
Using 188 dot-blot-SNP markers (Table S3, S4), 189 F2 plants harvested in 2010 were genotyped. Finally, the 188 SNP markers were mapped on linkage groups. A linkage map with nine linkage groups (LGs) spanned a total length of 856.0 cM with individual group sizes ranging from 61.3 cM (LG6) to 127.8 cM (LG3), and the average distance between neighboring markers was 4.37 cM (Figure 2). The marker positions in this map were in accordance with those in the radish map published by Li et al. [24].
Black bars and gray bars indicate QTL regions of 4MTB-GSL contents detected in 2010 and 2011, respectively.
For the analysis of the 174 F2 plants harvested in 2011, a linkage map was first constructed using 118 markers without use of markers closely linked with other markers. The final map consisted of nine LGs with 127 SNPs and one SCAR locus. The total length was 849.4 cM with an average distance of 6.58 cM between neighboring loci.
QTL Analysis of Glucosinolate Contents in Radish Roots
Four significant QTLs for 4MTB-GSL contents in radish roots were detected on LG1, LG2, LG6, and LG9 by analysis of the F2 population harvested in 2010, and four QTLs were also detected on LG1, LG6, LG7, and LG9 by analysis of the F2 population harvested in 2011 (Figure 2). LOD threshold values for significant QTLs in 2010 and in 2011 were determined to be 4.90 and 3.40, respectively, by the permutation test (Table 2). Among these five QTLs, GSL-QTL-1 on LG1, GSL-QTL-3 on LG6, and GSL-QTL-4 on LG9 were detected in both populations. In the population of 2010, these three QTLs explained more than 50% of the total phenotypic variance, with the GSL-QTL-3 on LG6 having the highest contribution value, i.e., 29.50%. By contrast, GSL-QTL-3 explained the small phenotypic variance, i.e., 16.72%, in 2011. ‘AZ26H’ alleles of GSL-QTL-3 and GSL-QTL-4 played major roles in increasing 4MTB-GSL contents. GSL-QTL-1 located between makers RS2CL6241 and RS2CL7263 on LG1 had a reverse additive effect, which explained the total phenotypic variances of 13.39% in 2010 and 36.7% in 2011. GSL-QTL-2 and GSL-QTL-5 explained the total phenotypic variances of 14.05% and 11.09%, respectively, but were detected only in one year. The average content of 4MTB-GSL in the F2 population harvested in 2010 was higher than that harvested in 2011 (Figure 1), suggesting that environmental factors, e.g., temperature, moisture, and harvesting time, have considerable effects on 4MTB-GSL contents. Therefore, they may have influenced the detection of QTLs in different years.
Seventeen QTLs have been identified as controling seed glucosinolates and four have been revealed to play major roles in glucosinolate accumulation of B. juncea by analysis using an F1DH population. Among them, only three major QTLs have been repeatedly detected in a BC4DH population [3], [4]. In B. rapa, QTL analyses have identified 16 loci controlling aliphatic glucosinolate accumulation in leaves, three loci controlling total indolic glucosinolate contents, and three loci regulating aromatic glucosinolate contents [6]. In the present study, five QTLs associated with 4MTB-GSL contents were identified in the two F2 populations in radish roots, and three of them, i.e., GSL-QTL-1, GSL-QTL-3 and GSL-QTL-4, were repeatedly detected. Although there is a possibility of the presence of more QTLs controlling total glucosinolates in radish as shown in B. rapa leaves, for only the 4MTB-GSL contents, which were found to constitut more than 96% of the total glucosinolate content, these five QTL are considered to play major roles.
To understand the additive effect of the two reproducible QTLs (GSL-QTL-3 and GSL-QTL-4), the F2 plants were categorized by marker genotypes at each QTL, and the 4MTB-GSL contents were compared. The genotypes of BB-BB (genotype of GSL-QTL-3 - genotype of GSL-QTL-4), BB-AB, and AB-BB, in which A is a ‘TBS’ allele and B is an ‘AZ26H’ allele, showed significantly higher contents of glucosinolates than those of BB-AA, AA-BB, and AA-AA in the two populations (Table 3). In a combination of GSL-QTL-1 (positive additive effect of ‘TBS’ alleles) and GSL-QTL-3 (negative additive value effect of ‘TBS’ alleles), BB-AA (genotype of GSL-QTL-1 - genotype of GSL-QTL-3), BB-BB, and BB-AB showed lower glucosinolate contents than those of AA-AB, AA-BB, and AA-AA (Table S5). This indicates that the allele at GSL-QTL-1 from ‘AZ26H’ acts to decrease the 4-MTB-GSL contents and that the alleles at GSL-QTL-3 and GSL-QTL-4 from ‘AZ26H’ act to increase the contents. However, the heterozygous genotypes of GSL-QTL-1, i.e., AB-BB or AB-AB, showed higher glucosinolate contents than the homologous genotypes, i.e., AA-BB and BB-BB or AA-AB and BB-AB. This may be due to a heterosis effect of GSL-QTL-1 on 4MTB-GSL biosynthesis.
DNA markers in GSL-QTL-1, -3, and -4 are considered to be useful for marker-assisted selection of radish. As radish for a salad or other fresh dishes, radish cultivars having low content of 4-MTB-GSL are preferred, while cultivars having high content of 4-MTB-GSL are required as a spice or a functional food. Since dot-blot-SNP analysis is suitable for SNP genotyping of a large number of plants [34], the dot-blot-SNP markers used for detection of GSL-QTL-1, -3, and -4 and SNPs of candidate genes in the QTL regions (discussed later) can be used in radish breeding.
Sequence Variation of Glucosinolate Biosynthesis Genes
Search for the SNP markers in the QTL regions were conducted by BLAST using the genome sequence of A. thaliana and B. rapa. New SNP markers were added to the QTL regions by selecting markers having SNPs between ‘TBS’ and ‘AZ26H’, which have been previously mapped in these regions. Three SNP markers were selected for GSL-QTL-1. Two markers were added to GSL-QTL-2, GSL-QTL-4, and GSL-QTL-5, respectively (Table 4). In the linkage map, SNP markers Rs2CL1297s and Rs2CL1508s of GSL-QTL-4 were shown to have high homology with At5g11670 and At5g24810, respectively [24]. In this region, MAM3 (At5g23020, Bra013009/013011/029356/021947) was selected as a candidate gene by comparison with the B. rapa genome [40]. Moreover, IPMDH1 (At5g14200, Bra023450) and BCAT4 (At3g19170, Bra02248/Bra001716) were found in syntenic regions with GSL-QTL-2 (At5g05260-At5g18380) and GSL-QTL-1 (At3g21770-At3g22600). No candidate gene was detected in GSL-QTL-3, which accounted for the largest contribution to phenotypic variance, and GSL-QTL-5 by synteny between R. sativus and B. rapa. The glucosinolate biosynthesis pathway is a complex network. In previous studies, more than fifty genes have been found to be involved in glucosinolate biosynthesis, including transcription factors, core structure formation, secondary modification, and core-substrate pathways [7], [42]. There may be other genes controlling 4MTB-GSL contents or several genes co-regulating the 4MTB-GSL biosynthesis in this locus.
To confirm effects of these candidate genes on 4MTB-GSL content of radish roots, three other important genes, i.e., MAM1 for side-chain formation, and CYP83A1 and CYP79F1 for core structure formation, were also analyzed (Table S2) [10], [17], [18], [39]. Sequences of these genes were determined and exons were inferred by BLASTN of EST sequences of R. sativus (http://blast.ncbi.nlm.nih.gov/) listed in Table 5. Comparison of R. sativus sequences of these genes, here named RsMAM3, RsIPMDH1, RsBCAT4, RsMAM1, RsCYP83A1, and RsCYP79F1, with A. thaliana and B. rapa sequences revealed that sequences of RsMAM3, RsMAM1, RsCYP83A1, and RsCYP79F1 showed more than 85% nucleotide identities in both exons and introns with homologous sequences in A. thaliana and B. rapa. However, genomic sequences including exons and introns of RsIPMDH1 and RsBCAT4 showed lower identities with homologous sequences in A. thaliana and B. rapa. Even exon sequences of RsIPMDH1 and RsBCAT4 showed low nucleotide identities between R. sativus and A. thaliana, i.e., 61% and 74%, respectively, and between R. sativus and B. rapa, i.e., 68% and 73%, respectively. In all six of these genes, the allocation of exons and introns was almost the same between R. sativus and A. thaliana.
Comparison of the sequences between ‘TBS’ and ‘AZ26H’ showed RsMAM3 to contain one SNP in an exon region, RsIPMDH1 to have 478 bp insertion in ‘AZ26H’ besides six SNPs and two single nucleotide indels in exons and introns, and RsBCAT4 to contain nine SNPs and six indels in the intron regions (Table 5, Figure 3). RsCYP83A1 had one insertion in an intron region between parental lines, and RsMAM1 had five SNPs and three indels throughout intron and exon regions. There was no variation in RsCYP79F1 between the parental lines (Table 5, Figure S1). SNPs from RsMAM3 and RsBCAT4 were used for designing probes for genotyping of the F2 population. RsMAM3 and RsBCAT4 were found to be linked to the GSL-QTL-4 and GSL-QTL-1 regions, respectively. Based on 478 bp insertion, a SCAR marker was designed for genotype analysis of RsIPMDH1 and was revealed to be linked to GSL-QTL-2 (Figure 2). RsCYP83A1 and RsMAM1 were not linked to the map, possibly due to distorted segregation ratios, i.e., 153:35 and 47:68:73, respectively. High similarity of RsCYP79F1 sequences between ‘TBS’ and ‘AZ26H’ did not allow the design of a probe for genotyping. RsMAM3, RsIPMDH1 and RsBCAT4 were considered to be candidate genes which regulate the 4MTB-GSL content in radish roots.
The black and gray boxes indicate exons and introns, respectively. The black arrows show indels. The positions of dashed arrows indicate SNP sites and nucleotide variations. The numbers under the boxes indicate the start and stop sites of exons.
Expression Analysis of Candidate Genes for Enzymes of the Glucosinolate Biosynthesis Pathway
We performed RT-PCR of six related genes using specific primers (Table S2) and compared gene expression levels between the parental lines in roots. Expression levels of RsMAM3, RsIPMDH1 and RsBCAT4 were higher in ‘AZ26H’ than those of ‘TBS’. However, CYP83A1 and MAM1 showed no expression difference although sequence variations were observed between the parental lines. The expression level of CYP79F1 was also similar between the parental lines (Figure 4). Real-time PCR was employed to delineate detailed expression profiles of RsMAM3, RsIPMDH1 and RsBCAT4. All these genes showed significantly higher expression levels in ‘AZ26H’ than those in ‘TBS’ (Figure 5).
RNAs were extracted from radish roots. Actin was used as a control to demonstrate equal RNA loading.
(A) RsMAM3, (B) RsIPMDH1, (C) RsBCAT4. Gray and black bars indicate parental lines ‘TBS’ and ‘AZ26H’. Significant differences (P<0.01, LSD-t test) between the parental lines are indicated by asterisks.
MAM3 catalyzes the condensation reaction of all side-chain elongation cycles. MAM3 contributes to the production of all aliphatic glucosinolates but primarily to that of glucosinolates derived from one, five, and six elongation cycles [11], [12], [41]. RsMAM3 was strongly expressed in the higher 4MTB-GSL content line (‘AZ26H’). 4MTB-GSL is one of the aliphatic glucosinolates (Figure 4, 5A) [27]. One nonsynonymous SNP (G-T) in the first exon changes the amino acid from leucine to phenylalanine in the coding region (Figure 4). QTL results of 4MTB-GSL contents in radish roots suggested that RsMAM3 plays a role in the difference of aliphatic glucosinolate contents between ‘AZ26H’ and ‘TBS’.
IPMDH1 has been identified based on strong co-expression with glucosinolate biosynthesis and leucine biosynthesis, and a knockdown mutant of IPMDH1 has been shown to cause a decrease in glucosinolate accumulation [7]. Mutation of IPMDH1 leads to a significant reduction in the level of free leucine and of glucosinlates with side chains of four or more carbons in A. thaliana [18]. IPMDH2 and IPMDH3 are predicted enzymes that share a similar function with IPMDH1 [9]. The expression of RsIPMDH1 was significantly higher in ‘AZ26H’ than that in ‘TBS’, implying that RsIPMDH1 may play a co-regulation in the 4MTB-GSL biosynthesis pathway with other genes (Figure 4, 5B).
BCAT4 deaminates methionine to produce 2-Oxo-4-methylthio-thiobutanoic acid, which is processed to 2-Oxo-6-methylthio-thiohexanoic acid, and transaminates 2-Oxo-6-methylthio-thiohexanoic acid to produce dihomomethionine, which is an intermediate of the pathway to 4-MTB-GSL. A BCAT4 knockdown mutant, bcat4, has been revealed to cause an approximately 50% reduction in aliphatic glucosinolates of leaves and to increase levels of free methionine, S-methyl-methionine, and the transport form of methionine, indicating that BCAT4 regulates the first deamination of side chain elongation of the glucosinolate biosynthesis pathway [41]. Higher expression of RsBCAT4 in ‘AZ26H’ than that in ‘TBS’ may suggest that RsBCAT4 acts to increase 4MTB-GSL content of the ‘AZ26H’ allele (Figure 4, 5C). However, the additive effect of the ‘TBS’ allele in GSL-QTL-1 was positive, suggesting that the ‘TBS’ allele increases 4-MTB-GSL content. Since the nucleotide sequence of RsBCAT4 is highly different from that of A. thaliana BCAT4, the function of RsBCAT4 might be different from that of A. thaliana BCAT4. Another possibility of this discrepancy may be the presence of another gene enhancing glucosinolate biosynthesis in ‘TBS’ roots. Delimitation of a GSL-QTL-1 region is required to identify the gene responsible for 4-MTB-GSL contents.
CYP79F1 converts all chain-elongated Met-derivatives, and CYP83A1 converts aliphatic aldoximes [43]–[46]. Amino acid chain elongation occurs, in which additional methylene groups are inserted into the side chain (up to six in A. thaliana). The condensation step of the cycle is considered to be critical for side chain variation and is catalyzed by a methythioalkymalate synthase enzyme. MAM1 catalyzes the condensation of side-chain variation in the first three elongation cycles and is involved in the production of glucosinolates derived from two elongation cycles [11], [13], [47]. Although there were sequence variations in RsCYP83A1 and RsMAM1, the expression levels were similar between ‘TBS’ and ‘AZ26H’. Sequences of RsCYP79F1 were the same between the parental lines and there was no significant difference in the expression level between them (Table 5, Figure 4). These genes may not contribute to the difference of 4MTB-GSL contents between ‘TBS’ and ‘AZ26H’.
Conclusions
Using the sequence data obtained by next-generation sequencing of bulked PCR products, a large number of SNP markers were developed for QTL map construction. Five QTLs were detected and three of them were repeeatedly detected. By analyzing synteny of the QTL regions with A. thaliana and B. rapa, three candidate genes were identified. Sequence and expression analyses indicate that they may be responsible for the difference of 4MTB-GSL contents between ‘TBS’ and ‘AZ26H’.
Supporting Information
Figure S1.
Nucletide polymorphsims of RsCYP79F1, RsCYP83A1, and RsMAM1 between ‘TBS’ and ‘AZ26H’. The black and gray boxes indicate exons and introns, respectively. The black arrows show indels. The positions of dashed arrows indicate SNP sites and nucleotide variations. The numbers under the boxes indicate the start and stop sites of exons.
https://doi.org/10.1371/journal.pone.0053541.s001
(TIF)
Table S1.
Primer pair groups for multiplex PCR and amplicons with SNPs detected between ‘TBS’ and ‘AZ26H’ by next generation sequencing.
https://doi.org/10.1371/journal.pone.0053541.s002
(PDF)
Table S2.
Sequences of primer pairs used for genomic DNA amplification, RT-PCR, and real-time PCR of glucosinolate biosynthesis genes.
https://doi.org/10.1371/journal.pone.0053541.s003
(PDF)
Table S3.
Sequences of primer pairs and oligonucleotide probes of SNP markers and the conditions of hybridization and washing.
https://doi.org/10.1371/journal.pone.0053541.s004
(PDF)
Table S4.
Dot-blot-SNP markers previously mapped with newly identified SNPs.
https://doi.org/10.1371/journal.pone.0053541.s005
(PDF)
Table S5.
Comparison of 4MTB-GSL contents between different genotypes of SNP markers in GSL-QTL-1 and GSL-QTL-3.
https://doi.org/10.1371/journal.pone.0053541.s006
(PDF)
Author Contributions
Conceived and designed the experiments: TN. Performed the experiments: ZZ MI FL TK SS HK. Analyzed the data: ZZ MI FL TK SS HK TN. Contributed reagents/materials/analysis tools: MI HK TN. Wrote the paper: ZZ TN.
References
- 1. Holst B, Williamson GA (2004) A critical review of the bioavailability of glucosinolates and related compounds. Nat Prod Rep 21: 425–447.
- 2. Fashey JW, Zalcmann AT, Talalay P (2001) The chemical diversity and distribution of glucosinolates and isothiocyanates among plants. Phytochemistry 56: 5–51.
- 3. Mithen RF, Dekker M, Verkerk R, Rabot S, Johnsom IT (2000) The nutritional significance, biosynthesis and bioavailability of glucosinolates in human foods. J Sci Food Agric 80: 967–984.
- 4. Ramchiary N, Bisht NC, Gupta V, Mukhopadhyay A, Arumugam N, et al. (2007) QTL analysis reveals context-dependent loci for seed glucosinolate trait in the oilseed Brassica juncea: Importance of recurrent selection backcross scheme for the identification of ‘true’ QTL. Theor Appl Genet 116: 77–85.
- 5. Bisht NC, Gupta V, Ramchiary N, Sodhi YS, Mukhopadhyay A, et al. (2009) Fine mapping of loci involved with glucosinolate biosynthesis in oilseed mustard (Brassica juncea) using genomic information from allied species. Theor Appl Genet 118: 413–421.
- 6. Lou P, Zhao J, He H, Hanhart C, Del Carpio DP, et al. (2008) Quantitative trait loci for glucosinolate accumulation in Brassica rapa leaves. New Phytol 179: 1017–1032.
- 7. Wang H, Wu J, Sun SL, Liu B, Cheng F, et al. 2011, Glucosinolate biosynthetic genes in Brassica rapa. Gene 487: 135–142.
- 8. Zang YX, Kim HU, Kim JA, Lim MH, Jin M, et al. (2009) Genome-wide identification of glucosinolate synthesis genes in Brassica rapa. FEBS J 276: 3559–3574.
- 9. Gigolashvili T, Yatusevich R, Rollwitz I, Humphry M, Gershenzon J, et al. (2009) The plastidic bile acid transporter 5 is required for the biosynthesis of methionine-derived glucosinolates in Arabidopsis thaliana. Plant Cell 21: 1813–1829.
- 10. Sawada Y, Kuwahara A, Nagano M, Narisawa T, Sakata A, et al. (2009) Omics-based approaches to methionine side-chain elongation in Arabidopsis: characterization of the genes encoding methylthioalkylmalate isomerase and methylthioalkylmalate dehydrogenase. Plant Cell Physiol 50: 1181–1190.
- 11. Kroymann J, Textor S, Tokuhisa JG, Falk KL, Bartram S, et al. (2001) A gene controlling variation in Arabidopsis glucosinolate composition is part of the methionine chain elongation pathway. Plant Physiol 127: 1077–1088.
- 12. Field B, Cardon G, Traka M, Botterman J, Vancanneyt G, et al. (2004) Glucosinolate and amino acid biosynthesis in Arabidopsis. Plant Physiol 135: 828–839.
- 13. Textor S, Kraker JW, Hause B, Gershenzon J, Tokuhisa JG (2007) MAM3 catalyzes the formation of all aliphatic glucosinolate chain lengths in Arabidopsis. Plant Physiol 144: 60–71.
- 14. Brader G, Mikkelsen M, Halkier B, Tapio Palva E (2006) Altering glucosinolate profiles modulates disease resistance in plants. Plant J 46: 758–767.
- 15. Grubb CD, Abel S (2006) Glucosinolate metabolism and its control. Trends Plant Sci 11: 89–100.
- 16. Wittstock U, Halkier B (2002) Glucosinolate research in the Arabidopsis era. Trends Plant Sci 7: 263–270.
- 17. Hirai MY, Sugiyama K, Sawada Y, Tohge T, Obayashi T, et al. (2007) Omics-based identification of Arabidopsis Myb transcription factors regulating aliphatic glucosinolate biosynthesis. Proc Natl Acad Sci 104: 6478–6483.
- 18. He Y, Mawhiney TP, Preuss ML, Schroeder AC, Chen B, et al. (2009) A redox-active isopropylmalate dehydrogenase functions in the biosynthesis of glucosinolates and leucine in Arabidopsis. Plant J 60: 679–690.
- 19. Bett KE, Lydiate DJ (2003) Genetic analysis and genome mapping in Raphanus. Genome 46: 423–30.
- 20. Kamei A, Tsuro M, Kubo N, Hayashi T, Wang N, et al. (2010) QTL mapping of clubroot resistance in radish (Raphanus sativus L.). Theor Appl Genet 120: 1021–7.
- 21. Tsuro M, Suwabe K, Kubo N, Matsumoto S, Hirai M (2005) Construction of a molecular linkage map of radish (Raphanus sativus L.), based on AFLP and Brassica-SSR markers. Breed Sci 55: 107–11.
- 22. Budahn H, Peterka H, Mousa MA, Ding Y, Zhang S, et al. (2009) Molecular mapping in oil radish (Raphanus sativus L.) and QTL analysis of resistance against beet cyst nematode (Heterodera schachtii). Theor Appl Genet 118: 775–82.
- 23. Shirasawa K, Oyama M, Hirakawa H, Sato S, Tabata S, et al. (2011) An EST-SSR linkage map of Raphanus sativus and comparative genomics of the Brassicaceae. DNA Res 18: 221–232.
- 24. Li F, Hasegawa Y, Saito M, Shirasawa S, Fukushima A, et al. (2011) Extensive chromosome homoeology among Brassiceae species were revealed by comparative genetic mapping with high-density EST-based SNP markers in radish (Raphanus sativus L.). DNA Res 18: 401–411.
- 25. Tsuro M, Suwabe K, Kubo N, Matsumoto S, Hirai M (2005) Construction of a molecular linkage map of radish (Raphanus sativus L.), based on AFLP and Brassica-SSR markers. Breed Sci 55: 107–11.
- 26. Carlson DG, Daxenbichler ME, VanEtten CH (1985) Glucosinolate in radish cultivars. J Amer Soc Hort Sci 110: 634–638.
- 27. Ishida M, Nagata M, Ohara T, Kakizaki T, Hatakeyama K, et al. (2012) Small variation of glucosinolate composition in Japanese cultivars of radish (Raphanus sativus L.) accommodates simple quantitative analysis for breeding of glucosinolate component. Breed Sci 62: 63–70.
- 28. Ishii G, Saijo R (1987) Effect of season, soil type, sulfate level, mulching and plant density on isothiocyanates content in radish root juice (Raphanus sativus L.). J Japan Soc Hort Sci 56: 313–320.
- 29. Ishida M, Kakizaki T, Ohara T, Morimitsu Y (2011) Development of a simple and rapid extraction method of glucosinolates from radish roots. Breed Sci 61: 208–211.
- 30.
Bjerg B, Sorensen H (1987) Quantitative analysis of glucosinolates and HPLC of intact glucosinolates, In: Wathelet J-P (ed) Glucosinolates in Rapeseeds: Analytical Aspects, Martinus Nijhoff Publishers, Dordrecht, Netherlands. 125–150.
- 31. Kaplinski L, Andreson R, Puurand T, Remm M (2005) MultiPLX: automatic grouping and evaluation of PCR primers. Bioinformatics 21: 1701–2.
- 32. Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12: 13–15.
- 33. Shiokai S, Kitashiba H, Nishio T (2010) Prediction of the optimum hybridization conditions of dot-blot-SNP analysis using estimated melting temperature of oligonucleotide probes. Plant Cell Rep 29: 829–34.
- 34. Shiokai S, Shirasawa K, Sato Y, Nishio T (2010) Improvement of the dot-blot-SNP technique for efficient and cost-effective genotyping. Mol Breed 25: 179–85.
- 35.
Van Ooijen JW (2006) JoinMap 4.0, Software for the calculation of genetic linkage maps in experimental populations. Kyazma, B.V.: Wageningen, The Netherlands.
- 36.
Wang S, Basten CJ, Zeng ZB (2007) Windows QTL Cartographer 2.5, Department of Satistics. North Carolina State University, Ralegih, NC.
- 37. Braütigam A, Gowik U (2010) What can next generation sequencing do for you? Next generation sequencing as a valuable tool in plant research. Plant Biology 12: 831–841.
- 38. Trick M, Long Y, Meng JL, Bancroft I (2009) Single nucleotide polymorphism (SNP) discovery in the polyploid Brassica napus using Solexa transcriptome sequencing. Plant Biotechnology J 7: 334–346.
- 39. Halkier BA, Gershenzon J (2006) Biology and biochemistry of glucosinolate, Ann Rev Plant Biol. 31: 303–333.
- 40. Wang XW, Wang HZ, Wang J, Sun RF, Wu J, et al. (2011) The genome of the mesopolyploid crop species Brassica rapa. Nat Genet 43: 1035–1040.
- 41. Schuster J, Knill J, Reichelt M, Gershenzon J, Binder S (2006) BRANCHED-CHAIN AMINOTRANSFERASE4 is part of the chain elongation pathway in the biosynthesis of methionine-derived glucosinolates in Arabidopsis. Plant Cell 18: 2664–2679.
- 42. Ida ES, Fernando GF, Halkier BA (2010) Biosynthesis of glucosinolates – gene discovery and beyond. Trends Plant Sci 15: 283–190.
- 43. Hansen CH, Wittstock U, Olsen CE, Hick AJ, Pickett JA, et al. (2001) Cytochrome p450 CYP79F1 from Arabidopsis catalyzes the conversion of dihomomethionine and trihomomethionine to the corresponding aldoximes in the biosynthesis of aliphatic glucosinolates. J. Biol. Chem 276: 11078–11085.
- 44. Chen SX, Glawischnig E, Jørgensen K, Naur P, Jørgensen B, et al. (2003) CYP79F1 and CYP79F2 have distinct functions in the biosynthesis of aliphatic glucosinolates in Arabidopsis. Plant J 33: 923–937.
- 45. Hemm MR, Ruegger MO, Chapple C (2003) The Arabidopsis ref2 mutant is defective in the gene encoding CYP83A1 and shows both phenylpropanoid and glucosinolate phenotypes. Plant Cell 15: 179–194.
- 46. Naur P, Petersen BL, Mikkelsen MD, Bak S, Rasmussen H, et al. (2003) CYP83A1 and CYP83B1, two nonredundant cytochrome P450 enzymes metabolizing oximes in the biosynthesis of glucosinolates in Arabidopsis. Plant Physiol 133: 63–72.
- 47. Kroymann J, Donnerhacke S, Schnabelrauch D, Mitchell-Olds T (2003) Evolutionary dynamics of an Arabidopsis insect resistance quantitative trait locus. Proc. Natl. Acad. Sci 100: 14587–14592.