BackgroundDue to its variegated and colorful leaves, ornamental kale (Brassica oleracea L. var. acephala) has become a popular ornamental plant. In this study, we report the fine mapping and analysis of a candidate purple leaf gene using a backcross population and an F2 population derived from two parental lines: W1827 (with white leaves) and P1835 (with purple leaves).ResultsGenetic analysis indicated that the purple leaf trait is controlled by a single dominant gene, which we named BoPr. Using markers developed based on the reference genome ‘02–12’, the BoPr gene was preliminarily mapped to a 280-kb interval of chromosome C09, with flanking markers M17 and BoID4714 at genetic distances of 4.3 cM and 1.5 cM, respectively. The recombination rate within this interval is almost 12 times higher than the usual level, which could be caused by assembly error for reference genome ‘02–12’ at this interval. Primers were designed based on ‘TO1000’, another B. oleracea reference genome. Among the newly designed InDel markers, BRID485 and BRID490 were found to be the closest to BoPr, flanking the gene at genetic distances of 0.1 cM and 0.2 cM, respectively; the interval between the two markers is 44.8 kb (reference genome ‘TO1000’). Seven annotated genes are located within the 44.8 kb genomic region, of which only Bo9g058630 shows high homology to AT5G42800 (dihydroflavonol reductase), which was identified as a candidate gene for BoPr. Blast analysis revealed that this 44.8 kb interval is located on an unanchored scaffold (Scaffold000035_P2) of ‘02–12’, confirming the existence of assembly error at the interval between M17 and BoID4714 for reference genome ‘02–12’.ConclusionsThis study identified a candidate gene for BoPr and lays a foundation for the cloning and functional analysis of this gene.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-017-3613-x) contains supplementary material, which is available to authorized users.

A splicing site mutation in BrFLC5, a non-syntenic paralogue of FLOWERING LOCUS C, was demonstrated to be related to flowering time variation in Brassica rapa. Flowering time regulation in Brassica rapa is more complex than in Arabidopsis, as there are multiple paralogues of flowering time genes in B. rapa. Brassica rapa contains four FLOWERING LOCUS C (FLC) genes, three of which are syntenic orthologues of AtFLC, while BrFLC5 is not. BrFLC1, BrFLC2, and BrFLC3 have been reported to be involved in flowering time regulation. However, BrFLC5 has thus far been deemed a pseudogene. We detected two alternative splicing patterns of BrFLC5 resulting from a nucleotide mutation (G/A) at the first nucleotide of intron 3 (named as Pi3+1(G/A)). Genotyping of BrFLC5Pi3 + 1(G/A) for 301 B. rapa accessions showed that this single nucleotide polymorphism was significantly related to flowering time variation (p < 0.001). In the collection, the frequency of the functional G allele (35.2%) was much lower than that of the nonfunctional A allele (59.1%); however, the frequency of the G allele was very high among the turnips (83.6%). An F2 population segregating at this locus was developed to analyze the genetic effect of BrFLC5. The result showed that the G allele individuals began to bolt two days later than the A allele individuals, indicating that BrFLC5 is a weak regulator of flowering time. BrFLC5 was expressed at the lowest level among the three analyzed BrFLCs. The late allele (G allele) was dominant to the early allele (A allele) at the BrFLC5 locus, which was in contrast to that of BrFLC1 and BrFLC2. This characteristic suggests that BrFLC5 would be more efficient for breeding premature bolting resistance in B. rapa.

Flowering time is important for Brassica rapa vegetables because premature bolting before harvest can lower yield and quality. FLOWERING LOCUS C (FLC) acts as a key repressor of flowering. In this study, we identified a nonsynonymous mutation at the 58th nucleotide of exon1 in BrFLC1 (named as Pe1+58 (A/C)) by screening resequencing data of 199 B. rapa accessions and verified this mutation as being related to flowering time variation. Strong linkage inheritance was detected between this locus and a previously reported splicing site mutation at intron 6 of BrFLC1 (Pi6+1 (G/A)), showing as co-occurrence of BrFLC1Pe1+58(A) and BrFLC1Pi6+1(G), named as haplotype H1: AG, or co-occurrence of BrFLC1Pe1+58(C) and BrFLC1Pi6+1(A), named as haplotype H2: CA. The frequency distribution of BrFLC1 haplotypes skewed to the haplotype H1 in turnip, broccoletto, mizuna, komatsuna, and taicai, while it was skewed to the haplotype H2 in caixin, pak choi, zicaitai, and wutacai. The frequencies of the two haplotypes were comparable in Chinese cabbage. This indicated that BrFLC1 haplotypes were related to B. rapa intraspecific diversification. Further analysis of a Chinese cabbage collection revealed that accessions from the spring ecotype preferred to keep H1: AG and almost all accessions from the summer ecotype were H2: CA. The early flowering haplotype of BrFLC1 was purified in summer Chinese cabbage, indicating that BrFLC1 had been strongly selected during genetic improvement of summer Chinese cabbages. A significant difference in flowering time of F2 individuals with the homologous BrFLC1Pi6+1(G) allele but different BrFLC1Pe1+58 (A/C) alleles, indicated that this locus had independent genetic effects on flowering time. The newly identified allelic diversity of BrFLC1 can be used for breeding of resistance to premature bolting in B. rapa vegetables.