Purple Brassica oleracea var. capitata F. rubra is due to the loss of BoMYBL2–1 expression

Water-soluble anthocyanin pigments are important ingredients in health-improving supplements and valuable for the food industry. Although great attention has been paid to the breeding and production of crops containing high levels of anthocyanin, genetic variation in red or purple cabbages (Brassica oleracea var. capitata F. rubra) has not yet been characterized at the molecular level. In this study, we identified the mechanism responsible for the establishment of purple color in cabbages. BoMYBL2–1 is one of the regulatory genes in the anthocyanin biosynthesis pathway in cabbages. It is a repressor whose expression is inversely correlated to anthocyanin synthesis and is not detectable in purple cabbages. Sequence analysis of purple cabbages revealed that most lacked BoMYBL2–1 coding sequences, although a few had a substitution in the region of the promoter 347 bp upstream of the gene that was associated with an absence of BoMYBL2–1 expression. Lack of transcriptional activity of the substitution-containing promoter was confirmed using transgenic Arabidopsis plants transformed with promoter::GUS fusion constructs. The finding that the defect in BoMYBL2–1 expression was solely responsible for purple coloration in cabbages was further demonstrated using genomic PCR and RT-PCR analyses of many other structural and regulatory genes in anthocyanin biosynthesis. Molecular markers for purple cabbages were developed and validated using 69 cabbage lines. Expression of BoMYBL2–1 was inversely correlated to anthocyanin content, and purple color in cabbages resulted from a loss of BoMYBL2–1 expression, caused by either the promoter substitution or deletion of the gene. This is the first report of molecular markers that distinguish purple cabbages. Such markers will be useful for the production of intraspecific and interspecific hybrids for functional foods, and for industrial purposes requiring high anthocyanin content.

increase in anthocyanin content is one of the most important target traits in crop breeding.
In Brassica species, purple or red color by anthocyanin accumulation is closely related to the induction of anthocyanin biosynthetic genes and/or their transcription factors. Increases in expression of TT8 and most anthocyanin biosynthetic genes are associated with purple coloration in pak choi (Brassica rapa var. chinensis) [40] and red mustard (Brassica juncea var. tumida Tsen et Lee) [39]. Similarly, accumulation of anthocyanin is caused by increases of BoMYB2 and bHLH-WD40 expression in purple cauliflower (B. oleracea L. var. botrytis) [44,45]; of BoMYB2 and BoTT8 expression in red cabbage (B. oleracea var. capitata) [46]; of B. oleracea PRODUCTION OF ANTHOCYANIN PIGMENTATION 1 (BoPAP1/BoMYB2) and downstream genes, such as DFR and anthocyanidin synthase (ANS), under the control of BoPAP1 control in purple kale (B. oleracea var. acephala F. tricolor) [47]; and of most of the anthocyanin biosynthesis genes, as well as BoPAP1 and 2 (BoMYB4), in kohlrabi [48]. Other purple or red crops derived by natural mutations in anthocyanin biosynthetic and regulatory genes have also been reported; for example, red apples resulting from a promoter rearrangement of MdMYB10 [49,50]; pink onions resulting from mutation in the ANS or DFR genes [51,52]; beans with a black seed coat caused by deletion of the CHS promoter [53,54]; purple cauliflower resulting from mutation in the BoMYB2 promoter activating its expression [44]; and purple ornamental kale caused by the deletion of DFR [55]. Red or purple cabbage (Brassica oleracea L. var. capitata f, rubra), a crop native to the Mediterranean region of Europe, is now grown all over the world as a fresh market vegetable [46].
The molecular mechanism responsible for the establishment of heading-type purple cabbages is, however, largely unknown. We found that either the promoter substitution of BoMYBL2-1 or deletion of the entire BoMYBL2-1 gene resulted in the establishment of purple cabbages. Molecular markers to discriminate purple cabbage from green ones were developed based on sequence variation. These markers will be used to develop purple vegetable crops containing high levels of anthocyanin that will be valuable for health improvements and industry use.

Plant materials
Cabbages used for BoMYBL2-1 cloning and marker validation were B. oleracea var. capitata f, alba or rubra plants from six green inbred lines, 32 green F 1 cultivars, two purple inbred lines, two near-isogenic lines (NILs), 16 purple F 1 cultivars, and 11 recombinants, as shown in Table 1. Inbred lines of green cabbages and purple cabbages, supplied by Asia Seed Co. (Korea), were selected to distinguish the genetic differences responsible for anthocyanin content.
Cabbages used for sequence analyses were as follows: inbred lines 337 and 154, selected for high anthocyanin content at low temperature; inbred lines 2437 and 09WH-45, selected for low anthocyanin content; inbred lines 2409 and 842, selected for no anthocyanin content; and the US cultivars Green A and B. The purple cabbages selected were the inbred lines 7S4-51 and 7S4-63, the NILs B90 and B98, and Purple A and B. Additional seeds not listed in Table 1 were purchased at local markets. Most plants were cultivated in a greenhouse at Chungnam National University, Daejeon, Korea, although some leaf samples were obtained from Korean Seed companies.    Table S1). Therefore, DNA sequences from three cultivars, kale, broccoli, and cauliflower, were reanalyzed using four primer sets (Additional file 1: Table S1; Additional file 2: Figure  S2A). After that, DNA sequences of heading-type cabbages (B. oleracea var. capitate) were cloned and analyzed. Genomic PCR was performed under the following conditions: denaturation for 5 min at 94°C, 30 cycles of amplification (30 s at 94°C, 30 s at 58°C, and 2-5 min at 72°C), and a final extension period of 7 min at 72°C. PCR products were purified using a LaboPass Gel Extraction kit (Cosmogenetech, Korea) and cloned into the TA-vector using the T&A cloning kit (RBC Bioscience Co., Taiwan). Escherichia coli (DH5α) cells were transformed with plasmid DNA carrying the desired insert. Plasmid DNA was purified using DNA-Spin (Intron Biotech. Inc., Korea) before sequencing. As cabbages might contain multiple MYBL2-1 alleles, at least ten clones from each line were sequenced and analyzed. Any possible PCR and/or sequencing errors were eliminated by aligning independent sequences.

Inverse PCR (iPCR)
To clone DNA sequences from purple cabbages that could not be amplified using any of the primer sets (Additional file 1: Table S1), inverse PCR (iPCR) was performed on Bol010163 gene sequences using the Universal GenomeWalker™ 2.0 (Clonetech, USA). DNA from purple cabbages and control DNA from the kit were digested with DraI, EcoRV, PvuII, and StuI for 16-18 h at 37°C, and purified using the NuceloSpin Gel kit and PCR Clean-Up kit (Macherey-Nagel GmbH & Co, Germany). After ligation with the GenomeWalker adaptor, primary PCR was performed with a Bol010163 gene-specific primer (5'-AGACGTTGATGAGATCAACGGTTGTGA), followed by secondary PCR with an adaptor primer (5'-CATCCAA-TAAAGGCGAGCAAGAAAGGA). PCR products were electrophoresed on 1% agarose gels, and the resulting bands were excised, purified, and cloned using the T&A cloning kit (RBC Bioscience Co., Taiwan). Three sizes of DNA fragments were obtained: 700 bp from the PvuII library, 2.1 kbp from the DraI library, and 4.0 kbp from the StuI library. To minimize sequence error, at least five clones from each library were sequenced and analyzed.

RT-PCR
Leaf samples were collected from at least three individual plants, and total RNA was isolated from liquid nitrogen-ground samples using TRIzol reagent (Invitrogen, USA), and further purified using the NucleoSpin RNA Clean-up Kit (Macherey-Nagel GmbH & Co., Germany). Total RNA (1 μg) was treated with RQ1 RNase-free DNAase (Promega, USA), and cDNA was synthesized using the Ace-α kit with oligo(dT) primers (Toyobo, Japan). Complementary DNA was diluted 10fold, and 1 μl of diluted cDNA was used in a 20 μl PCR mixture. RT-PCR primers are listed in Additional file 3: Table S2; primers for ACTIN 2 (BoACT2) were used as a control. A standard PCR was performed with a 5 min denaturation at 94°C, followed by 20-30 cycles of amplification (30 s at 94°C, 30 s at 60°C, and 30 s at 72°C), and a final extension time of 7 min at 72°C. PCR products were analyzed following electrophoresis through 1.2% agarose gels.

Marker development and validation
To distinguish green and purple cabbages based on their BoMYBL2-1 sequences, three primer sets were designed following sequence alignment using ClustalW2 (http:// www.ebi.ac.uk/Tools/msa/clustalw2/) ( Table 2). The PCR mix consisted of 10-30 ng of genomic DNA, 5 pmol of each forward and reverse primer, 10 pmol of BoACT2 primers, and 1× HiPi Plus PCR premix buffer in 20 μl. To optimize PCR conditions for each primer pair (BoMYBL2-1w-F and BoMYBL2-1w-R; BoMYBL2-1v-F and BoMYBL2-1v-R; BoMYBL2-1sub-F and BoMYBL2-1sub-R), several annealing temperatures, extension times, and cycle numbers were tested and set. For the primer pair BoMYBL2-1w-F and BoMYBL2-1w-R, used to identify the presence of BoMYBL2-1, the PCR conditions were 5 min at 94°C, followed by 28 cycles of 30 s at 94°C, 30 s at 65°C, and 2 min at 72°C, with a final extension phase of 5 min at 72°C. For the primer pair BoMYBL2-1v-F and BoMYBL2-1v-R, used to identify promoter-substituted BoMYBL2-1, the reaction conditions were 5 min at 94°C, followed by 30 cycles of 30 s at 94°C, 30 s at 60°C, and 1 min at 72°C, and a final extension of 5 min at 72°C. For the primer pair BoMYBL2-1sub-F and BoMYBL2-1sub-R, used to identify DNA with a substitution of the whole BoMYBL2-1 gene, the reaction conditions were 5 min at 94°C, followed by 30 cycles of 30 s at 94°C, 30 s at 65°C, and 2 min at 72°C, and a final extension of 5 min at 72°C.

Presence of genes associated with anthocyanin biosynthesis
To examine whether additional genes were defective in purple cabbages, the presence of genes associated with anthocyanin biosynthesis was studied using genomic PCR. The genes selected were transcriptional activators and repressors (Additional file 3: Table S2). PCR conditions were 5 min at 94°C, followed by 20-30 cycles of 30 s at 94°C, 30 s at 60°C, and 30 s at 72°C, and a final extension time of 5 min at 72°C.

Promoter analysis
To confirm whether the substituted promoter found in purple cabbages retained promoter activity, the wildtype and substituted versions of the BoMYBL2-1 promoter were fused to the GUS reporter gene. Arabidopsis plants were transformed with the reporter constructs, and examined for GUS activity and expression. The MYBL2-1 promoter regions from green and purple B. oleracea (1926 bp upstream from the ATG start codon for green B. oleracea and 1804 bp upstream for purple B. oleracea) were amplified by PCR using the specific forward primers (5'-GATCAGGATCCAAGAACACAT-GAACTT for green cabbage and 5'-GATCAGGATC-CAAGAACCAGTGTTC for purple cabbage) and the same reverse primer (5'-GTGAGCCATGGTACGA-GAAGCA). These primers contain the BamHI and NcoI restriction sites (underlined). The amplified fragments were inserted into the T&A cloning vector (Real Biotech Co., Taiwan), and the presence of the MYBL2-1 promoter sequence was confirmed by sequencing. The fragments were liberated by digestion with BamHI and NcoI, and subcloned into the pCambina3301-GUS binary vector, digested with the same enzymes. The resulting constructs were transformed into Arabidopsis plants using the Agrobacterium tumefaciens-mediated floral dip procedure [58]. Transformed plants were selected using 0.1% BASTA herbicide, and their identity was confirmed by PCR analysis of genomic DNA. T 3 homozygous lines containing a T-DNA insertion at a single locus, determined by a 3:1 ratio of segregation of basta-resistance: sensitivity, were selected for the assay of promoter activity.
Arabidopsis thaliana wild-type (Col-0) and transgenic plants were grown in a growth chamber under 16 h light/ 8 h dark photoperiods at 22°C and a light intensity of 100 μmol m − 2 s − 1 . For plate culture, seeds were surfacesterilized with 30% bleach containing 0.1% Triton X-100, stratified for 3 days at 4°C, and plated onto solidified halfstrength Murashige and Skoog (MS) medium, plus or minus 90 mM sucrose. Three plants were sampled for RT-PCR study of GUS expression, and another three plants were used to analyze GUS expression.
For GUS expression studies, plants were incubated in GUS staining solution (1 mM X-GlucA in 100 mM sodium phosphate, pH 7.0, containing 5 mM Na 2 EDTA, 0.5 mM potassium ferrocyanide, 0.5 mM potassium ferricyanide, and 0.1% Triton X-100) for 16 h at 37°C. After staining, the plants were washed in 95% ethanol at room temperature until wild-type (Col-0) Arabidopsis appeared clear.

Quantification of total anthocyanin content
Total anthocyanin content was determined in cabbage leaves using methanol containing 1% HCl [59]; three independent biological replicates from three individual plants were used in each experiment. The tissues were ground under liquid nitrogen, and the powder was resuspended in a 1.5 ml tube containing methanol (1% HCl) at room temperature and centrifuged at 14,000 rpm for 10 min at 4°C. The absorbance of the supernatants was determined spectrophotometrically at 530 nm and 657 nm. Total anthocyanin content was quantified as where Q = total anthocyanins; A530 = absorption at 530 nm; A657 = absorption at 657 nm; FW = fresh weight of tissues (g).

Results
Selection of BoMYBL2-1 gene To dissect the association of BoMYBL2 with anthocyanin biosynthesis, expression of selected key genes in the pathway was examined in six inbred lines of green cabbage (337, 154, 2477, 09WH-45, 2409, and 842) and two inbred lines of purple cabbage (7S4-51 and 7S4-63). Two-monthold cabbage plants were grown for a month, either under greenhouse (G) conditions, which were non-inductive for anthocyanin accumulation, or outside (C), which was inductive for anthocyanin accumulation because it exposed the plants to low temperature (Fig. 1). Expression of BoMYB1 (Bol042409 = Bo3g081880), BoMYB2 (Bol012528 = Bo6g100940), BoDFR1 (Bol035269 = Bo9g058630), BoMYBL2-1 (Bol016164 = Bo6g112670), and BoMYBL2-2 (Bol034966 = Bo2g070770) was analyzed. Two BoMYBL2 genes (one on chromosome 6 and the Fig. 1 Analysis of anthocyanin biosynthesis-related genes from cabbage plants grown either in a greenhouse (G) or exposed to low temperature conditions outside (C). Plants were grown for 1 month between Oct. 13 and Nov. 12, 2014. Average temperatures ranged between 20 and 30°C in the greenhouse and between 6 and 18°C outside other on chromosome 2) matched Arabidopsis MYBL2 (At1G71030), and BoMYBL2 designations were given according to the level of amino acid sequence identity with the Arabidopsis counterpart. As shown in Fig. 1, the expression of BoDFR1 correlated with anthocyanin accumulation (Additional file 4: Figure S1). BoMYB1 and 2 were highly expressed in purple cabbages, whereas BoMYBL2-2 expression was detected in all samples. BoMYBL2-1 expression was not detected in purple cabbages, and attempts to amplify the gene by PCR from the genomic DNA of purple cabbages failed, suggesting either a high level of sequence variation or deletion of BoMYBL2-1.
To investigate whether BoMYBL2-1 had been deleted from the genome of purple cabbages, we examined the presence and expression of the gene in additional cultivars (Fig. 2). Both the BoMYBL2-1 gene and its transcripts were detected in all the green cabbages tested. Two different types of result were obtained from two purple cabbages: (1) no amplification product from either genomic DNA or cDNA templates; and (2) amplification product from genomic DNA but not from cDNA. More specifically, the BoMYBL2-1 transcript could not be detected in the purple B cultivar, although presence of the gene was confirmed from analysis of genomic DNA. These data suggest that two distinct mechanisms for the loss of BoMYBL2-1 expression were responsible for the color of the two purple cabbages tested.
Cloning and sequence analysis of BoMYBL2-1 from various B. oleracea species Genomic cloning of BoMYBL2-1 was undertaken using six primer pairs (Additional file 2: Figure S2; Additional file 1: Table S1) and eight inbred lines of cabbages obtained from the Asia Seed Co. (Korea). An initial comparison of PCR results revealed that high levels of sequence variation were present around BoMYBL2-1, as observed in the distinct reference sequences previously obtained from two different varieties of B. oleracea [56,57] (Additional file 1: Table S1). We therefore re-established a sequence assembly for kale (B. oleracea var. sabellica), broccoli (B. oleracea var. italica) , and cauliflower (B. oleracea var. botrytis) cultivars using one allele from each variety (Additional file 5). Based on this new sequence information, sequencing and analysis were extended to inbred lines and cultivars of green cabbages, as well as purple cabbages (B. oleracea var. capitata) (Additional file 5). Nonetheless, no PCR product could be amplified from several lines of purple cabbages, including 7S4-51, 7S4-63, and B98. We used iPCR with primers designed around Bol016163 to obtain sequence information for these plants.
The simplified structure of the BoMYBL2-1 genes and upstream nucleotide sequences from all of the lines used in our analysis (Additional file 5), which confirmed high levels of sequence variation in this region, are shown in Fig. 3. Kale and cauliflower were both heterozygous for Bol016163 alleles, one with complete and the other with deleted versions, while broccoli appeared to have homozygous alleles. As shown in Fig. 3, all heading-type green cabbages carried BoMYBL2-1 (1208 bp) plus approximately 1900 bp of upstream sequences. By contrast, two types of DNA sequence were observed in purple cabbages: BoMYBL2-1 with a substituted promoter containing 1190 bp of sequence, also found on chromosome 1, 347 bp upstream of the start codon (called BoMYBL2-1 variant or BoMYBL2-1v), and a substitution (or deletion) type of the entire BoMYBL2-1 plus regulatory regions with putative chromosome 3 and chromosome 7 (called BoMYBL2-1 substitution or BoMYBL2-1sub). Most purple cabbages contained BoMYBL2-1sub. To distinguish the wild-type or intact BoMYBL2-1 sequence from BoMYBL2-1v, the wild-type version was designated BoMYBL2-1w. Interestingly, both BoMYBL2-1w and BoMYBL2-1v contained 159 bp repeat sequences upstream of the ATG start codon, which included seven ACCCGA repeats, 11 CGAA repeats, seven AAAT repeats, and so on. The MYB core motif GGATA [55] was detected in this repeat.

Promoter analysis
To test whether the promoter region of BoMYBL2-1v had any transcriptional activity, 1926 bp from BoMYBL2-1w and 1804 bp from BoMYBL2-1v upstream of the ATG start codon were fused to the GUS reporter gene and the constructs were used to transform Arabidopsis plants. ß-glucuronidase (GUS) activity in Arabidopsis transgenic plants was evaluated using both histochemical staining and levels of expression of GUS transcript (Figs. 4 and 5). Two independent pCambia3301 transgenic plants were used as positive controls. Comparable levels of GUS staining (Fig. 4a) and GUS transcript expression (Fig. 4b) were shown by seven independent pMYBL2-1w::GUS transgenic plants. On the other hand, neither GUS staining nor GUS transcripts were detected in wild-type Arabidopsis (Col-0) or in seven independent pMYBL2-1v:: GUS transgenic plants. These results indicated that the pMYBL2-1v promoter was non-functional under normal growth conditions.
Since anthocyanin biosynthesis is induced by various environmental factors, GUS expression in transgenic Arabidopsis plants was also tested under conditions of sucrose and temperature (low and high temperature) stress. As shown in Fig. 6, GUS staining was not detectible in pMYBL2-1v::GUS transgenic plants, even when the plants were subjected to the stress of 90 mM sucrose or exposure to high or low temperature. Taken together, these data strongly suggest that the promoter substitution found in BoMYBL2-1v resulted in the loss of promoter activity and BoMYBL2-1 expression in some types of purple cabbages. In others, the change of color from green to purple may be explained by the complete deletion of the BoMYBL2-1 coding sequence.
As shown in Additional file 6: Figure S3, all these genes other than BoMYBL2-1, which was lost in several lines of purple cabbage, were present in various varieties of B. oleracea. This strongly suggested that, at least for some purple cabbages (B. oleracea var. capitata f, rubra), the deletion of BoMYBL2-1 was responsible for the development of purple coloration.
Since the presence of a gene does not guarantee its expression, expression of key biosynthetic (BoCHS1, BoCHS2, BoF3H, BoF3'H, and BoDFR1) and regulatory genes affecting anthocyanin accumulation was examined in 20 different purple cabbage (Additional file 6: Figure S3). Transcript levels of most genes did not differ between these purple cabbages, but expression levels of several genes (BoCHS1, BoF3H, BoDFR1, and BoMYB2) were higher in those plants than in the green cabbage Daebakna. All the genes tested, other than BoMYB4, were expressed at relatively high levels, possibly as a result of defective BoMYBL2-1 activity.

And validation of molecular markers to distinguish purple cabbage
Markers that are associated with a particular horticultural trait, such as purple coloration, are very important in developing new B. oleracea varieties with the desired trait from crosses of different subspecies. To develop molecular markers able to identify purple B. oleracea var. capitata carrying defective BoMYBL2-1, we designed three sets of primer pairs: (1) to amplify the BoMYBL2-1 coding sequence; (2) to amplify the promoter-substituted variant BoMYBL2-1 (MYBL2v); and (3) to amplify the DNA sequence that replaced the entire BoMYBL2-1 gene (BoMYBL2-1sub) (Table 2; Fig. 3). The optimal PCR conditions for each pair are described above (Methods). The validity of these primer pairs was tested using the 35 green and 33 purple cabbages listed in Table 1.
As shown in Fig. 6, the band corresponding to BoMYBL2-1 coding sequence was amplified by the F1 and R1 primers (BoMYBL2w-F and BoMYBL2w-R) from all the green cabbages tested, as well as from four purple cabbages (B90, Purple B, NP-J-156, and-047). All of the purple cabbages tested, however, other than B90, contained DNA that could be amplified using the F3 and R3 primers (BoMYBL2sub-F and BoMYBL2sub-R) (Fig. 7), implying that these plants contained the substituted version of the entire BoMYBL2-1 gene, including the promoter and coding regions. In addition, only four purple cabbages contained DNA that could be amplified using the F2 and R2 primers (BoMYBL2v-F and BoMYBL2v-R), which were specific for the promoter-substituted BoMYBL2-1 variant (BoMYBL2-1v) (Fig. 8). Taken together, we concluded that purple cabbages contain either BoMYBL2-1v or BoMYBL2-1sub or both. Our data indicated that most of the purple cabbages in this study were homozygous for BoMYBL2-1sub, although B90 was homozygous for BoMYBL2-1v. Purple B, NP-J-0156, and HK-047 were heterozygous, carrying both the BoMYBL2-1v and BoMYBL2-1sub alleles.

Discussion
Purple or red color in Brassica species due to anthocyanin accumulation is usually related to the activation of anthocyanin biosynthesis genes and/or their transcriptional activators. Purple or red color in B. oleracea is associated with an increase in expression of transcriptional activators, such as BoMYB2 and BoTT8 (BobHLH) [44][45][46], and BoPAP1 and BoPAP2 [47,48]. To date, only two studies have reported natural mutations that lead to an increase in anthocyanin biosynthesis: mutation of the BoMYB2 promoter to activate its expression in purple cauliflower [44], and deletion of the DFR gene in purple ornamental kale [55]. There is, however, no information available on the origin of red or purple cabbage (Brassica oleracea L. var. capitate f, rubra), a native of the Mediterranean region of Europe that is now grown as a fresh market vegetable all over the world [46]. The present study found that purple color in heading-type cabbages resulted from defective expression of BoMYBL2-1 caused by two different mechanisms: either a substitution in the BoMYBL2-1 promoter or a deletion of the entire gene. This is the first report showing that purple plants can be spontaneously generated by a defective repressor, rather than by alteration of activator expression by transposon insertion. Arabidopsis seedlings were grown for 10 days after sowing at a light intensity of 100 μmol − 2 s − 1 under long-day conditions. For temperature stress, Col-0 and mutant plants were grown on soil for 12 days at 22°C with light intensity of 100 μmol − 2 s − 1 under long-day conditions. They were transferred to incubators at either 10°C (low temperature) or 34°C (high temperature) with a light intensity of 100 μmol − 2 s − 1 and incubated for 19 h. GUS staining was as described in Fig. 4 MYBL2, a small MYB protein that has two functional transcriptional repressor motifs, EAR and TLLLER motifs, interacts with TT8, a bHLH protein that is a component of the MBW complex, and represses expression of anthocyanin biosynthesis genes (LBGs) [24,29,33,[35][36][37]. MBW complexes control the flavonoid biosynthesis pathway at the transcriptional level developmentally and environmentally, mainly by activating expression of flavonoid LBGs [29]. MYBL2 and other R2R3-MYBs interact with bHLH proteins in a competitive manner; thus other R2R3-MYBs prevent the formation of the MBW complex and so negatively regulate anthocyanin production [24,28,35]. Repression or sequestering of MYBL2 activates anthocyanin biosynthesis in Arabidopsis [41][42][43].
The B. oleracea var. capitata genome contains two MYBL2 genes, BoMYBL2-1 and BoMYBL2-2; however, only BoMYBL2-1 appears to be closely associated with anthocyanin production in cabbages, as a decrease in BoMYBL2-1 expression was observed in some cultivars of purple cabbages (Fig. 1). These expression data, together with studies of BoMYBL2-1 promoter activity in transgenic Arabidopsis (Figs. 5 and 6), supported the conclusion that the absence of BoMYBL2-1 expression, as a result of either promoter substitution or whole gene deletion, was responsible for the production of purple color in cabbage. However, we cannot rule out the possibility that other negative regulators are also involved in the generation of other purple cabbages not included in the current study, as has been observed in poplar [60,61].
The loss of promoter function in the substituted promoter of BoMYBL2-1 present in several purple cabbage cultivars was supported by an analysis of GUS expression in transgenic Arabidopsis (Figs. 5 and 6). As shown in Fig. 9, the BoMYBL2-1w promoter harbors three types of MYB binding motifs: one MYB core motif (GGATA) [59], two MYB recognition sequences (A/ TAACCA), and two MYB recognition sequences (C/ TAACG/TG) [62]. By contrast, the upstream sequences of the BoMYBL2-1v coding sequence found in two purple cabbages contain only one MYB core motif, which is located in a 159 bp repeat sequence. This implies that one MYB core motif alone is not sufficient for the initiation of BoMYBL2-1v expression, at least in B. oleracea We detected high levels of sequence variation around BoMYBL2-1 and the surrounding region ( Fig. 3; Additional file 7: Figure S5). This may be due to genome assembly errors in the '02-12' reference genome of cabbage (http://brassicadb.org) [57] and the TO1000 sequence (http://plants.ensembl.org/Brassica_oleracea) [57]. Similar assembly errors in cabbage reference genomes have been identified and corrected for mapping of clubroot resistance [63] and yellow-green leaf trait [64]. We provide the sequences flanking BoMYBL2-1 in chromosome 6 obtained in this study, which suggest that Bol016163 is absent from B. oleracea var. capitata (Additional file 5; Additional file 7: Figure S5).
As shown in Additional file 6: Figure S3 and Additional file 8: Figure S4, it is clear that all purple cabbages (B. oleracea var. capitata f, rubra) retain most of the genes associated with anthocyanin biosynthesis and regulation. Most studies of the development of purple or red coloration in crops suggest that it results from activation of anthocyanin biosynthesis genes and/ or regulatory genes, and is largely due to mutation, especially mutations within a promoter that confer constitutive expression [44,[49][50][51][52][53][54][55]. The present study demonstrates, however, that mutation within repressor genes provides another route to the establishment of anthocyanin-enriched plants in nature, and suggests that similar types of mutation may be present in crops with high anthocyanin content.
We developed a method for using PCR-based molecular markers that could quickly identify and distinguish purple cabbages that contain a substituted promoter or a complete deletion of BoMYBL2-1 (Figs. 6, 7 and 8). These markers will facilitate the intraspecific and interspecific breeding of B. oleracea to produce anthocyanin-rich crops. Unfortunately, these markers do not predict the anthocyanin levels or profiles of different purple cabbages, even though a previous study reported genotype-dependent variations in anthocyanin profiles [22]. A more detailed understanding of anthocyanin biosynthesis in cabbages is required to develop additional markers to select purple-colored cabbages possessing more desirable anthocyanin profiles.  Results of PCR analysis using the F2 and R2 primer pair (BoMYBL2v-F and BoMYBL2v-R) to amplify BoMYBL2-1v, which contains the substituted region of the promoter. The white arrow indicates the position of the predicted PCR product. Three purple cabbages (Purple B, NP-J-156, and HK-047) were heterozygous (BoMYBL2-1v/BoMYBL2-1sub), but B90 was homozygous for BoMYBL2-1v. All other purple cabbages were homozygous for BoMYBL2-1sub