Interaction between various environmental signals and flowering genes is critical for plants to flower and complete their life cycle, and thus important to humans, who rely upon adequate production of fruit and seeds to feed the world’s growing population. Climate change fluctuations accompanying global warming [1, 2] are requiring plant breeders to elucidate the molecular mechanisms underlying flowering, and to develop strategies for manipulating and optimizing the flowering times to maximize crop yields. Four flowering pathways—autonomous, vernalization, photoperiod and gibberellic acid—have been established in Arabidopsis and partially identified in other species [3, 4]. Vernalization is an adaptive trait in which plants acquire the ability to flower following exposure to cold temperatures. A series of genes in the endogenous network involved in this process, and their regulatory relationships, have been identified; genes from different flowering pathways function together with other integrator genes to control flowering [5, 6]. The MADS-box family gene FLOWERING LOCUS C (FLC) represses flowering [7, 8] by suppressing the expression of FLOWERING LOCUS T (FT), a key flowering integrator and confirmed florigen in plants [911], and other floral integrator genes such as FLOWERING DURATION and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS[9, 12, 13]. Expression of FLC is reduced by vernalization [7, 8]. FLC is up-regulated by FRIGIDA (FRI) and repressed by genes in the autonomous pathway [1416]. FLC expression has also been shown to be regulated via histone acetylation and methylation, which alters the expression of a trans-acting regulator common to FLC and members of the related MADS AFFECTING FLOWERING gene [1720].

The genus Brassica, which diverged from Arabidopsis 14.5 to 20.4 million years ago [2123] includes more crops of agricultural and horticultural importance than any other genus in the family of Brassicaceae. Comparative analysis has revealed that diploid Brassica genomes are composed of conserved segments triplicated from Arabidopsis[24, 25]. The allopolyploid species B. napus (rapeseed, oilseed rape or canola; genomes AACC, 2n = 4x = 38) is a product of natural hybridization between diploid species B. rapa (2n = 2x = 20, genome AA) and B. oleracea (2n = 2x =18, genome CC). Rapeseed originated in southern Europe along the coastline of the Mediterranean Sea 10,000–100,000 years ago, and was domesticated as an oil crop 400–500 years ago. This crop was originally grown as a spring or semi-winter crop in Mediterranean climates; Its cultivation spread rapidly from southern into northern Europe after the development of winter rapeseed varieties, which do not flower during the long and cold winters. Understanding the evolution of flowering time is critical for domestication and introduction of rapeseed into new agroclimatic regions.

Miniature inverted-repeat transposable elements (MITEs) belong to a class of non-autonomous DNA transposable elements known as class II transposons. They are present in high copy number in the genome and contribute to genomic structure variations and intra-species diversity [26, 27]. Differing MITE insertion profiles among varieties of a given species enable tolerance to environmental changes and allow adaptation under selective pressure [26, 28, 29].

Genetic analyses of several mapping populations of Brassica have revealed that both major and minor quantitative trait loci (QTLs) control flowering time. Some of these QTLs have also been shown to collocated with candidate genes for flowering time such as CO, FLC, FT and FRI[3035]. Forty-two QTLs were identified in a doubled haploid (DH) rapeseed mapping population (TN DH) derived from a cross between Tapidor and Ningyou7, but their magnitude and genetic effects varied with growing environment [36]. One major flowering time QTL, qFT10-4, which accounted for more than 50% of phenotypic variation in flowering time in the TN DH populations grown in non-vernalization environments, was colocalized with the ortholog of FLC from Arabidopsis in chromosome A10 and was designated BnFLC.A10[36, 37]. In our study, the candidate gene BnFLC.A10 for qFT10-4 was dissected using a map-based cloning approach, and an association was found between a Tourist-like MITE insertion/deletion in the upstream region of BnFLC.A10 and the stronger vernalization requirement in rapeseed.


Cloning of BnFLC.A10 from qFT10-4and allelic divergence

To construct a high-resolution map of the qFT10-4 locus, we analyzed a large BC5F2 population (9,000 plants) that was derived from a cross between the DH line TN DH043 (winter-type) and Ningyou7 (semi-winter-type). Four molecular markers developed from the sequence of the Bacterial Artificial Chromosome (BAC) clone JBnB75D10, which contains BnFLC.A10, were used for the analysis (Figure 1A). Eight recombinants were identified and the QTL qFT10-4 was delimited to an 80-kb region that showed collinearity with the top of chromosome 5 of Arabidopsis thaliana (Figure 1B and C). None of the genes in this region except FLC are known to be involved in floral transition.

Figure 1
figure 1

Cloning of qFT10-4 and detailed structure and allelic divergence of BnFLC.A10 . (A) Positions of markers used to fine-map qFT10-4 are shown in the BAC clone JBnB75D10 of B. napus ‘Tapidor’. Marker IP1IP2 was developed from a specific sequence of BnFLC.A10. (B) Genotypes of recombinants detected among non-flowering plants of the BC5F2 segregation population derived from the TN DH line DH043 (winter-type) and Ningyou7 (semi-winter-type). T and H represent homozygous and heterozygous genotypes, respectively, for the Tapidor allele. (C) Genes identified in the 80-kb region of JBnB75D10 that was delimited with markers T11 and Niab009. Arrows show the relative positions of predicted open reading frames (ORFs). For each ORF, the orthologous gene in A. thaliana is marked and genes that lacked an ortholog are labelled ‘hypothetical’. (D) Schematic diagram of the DNA sequence of BnFLC.A10. The arrow shows the translation start site. Roman numerals indicate the indels (I–IV) between the alleles from Tapidor and Ningyou7. Vertical bars labeled with Arabic numerals represent SNPs (1–8). For the SNPs, the nucleotide found in the Tapidor allele is given first. (E) BnFLC.A10 expression as detected by quantitative PCR during different stages of vernalization (0 to 7 weeks) at 4°C. Expression of the Ningyou7 allele decreased much more rapidly than that of the Tapidor allele during vernalization.

To analyze the basis of the vernalization requirement in rapeseed, we cloned and compared BnFLC.A10 sequences (approximately 7 kb) from Tapidor and Ningyou7, the parental lines of the mapping population. No polymorphism was found in the coding sequence (CDS) between the two alleles (BnFLC.A10-T and BnFLC.A10-N). However, there were two insertion/deletions (indels I and II) in the upstream region, together with two indels (indels III and IV) and eight single nucleotide polymorphisms (SNPs 1–8) in intron 1 of BnFLC.A10 (Figure 1D). Expression analysis showed that BnFLC.A10-N was markedly down-regulated upon exposure to cold treatment after 1 week, whereas expression of BnFLC.A10-T decreased gradually over 7 weeks of cold treatment (Figure 1E). This observation provides strong evidence that BnFLC.A10 underlies variation for vernalization requirement and that differences in gene expression establish the basis for allelic variation at the qFT10-4 locus.

A 621-bp insertion upstream of BnFLC.A10-Tis associated with winter habit in rapeseed

To determine whether sequence variations in the two BnFLC.A10 alleles contribute to differences in vernalization requirements or winter growth habit among natural rapeseed populations, we conducted an association analysis using a panel of 79 diverse rapeseed cultivars representing winter, semi-winter and spring genotypes. All of the cultivars were planted in spring environments. Because of lack of vernalization, none of the winter-type cultivars flowered; in contrast all of the spring-type and semi-winter-type cultivars (with one exception) flowered normally (Table 1). The three largest indels (I, II and IV) were analyzed first because the alleles could be easily distinguished by PCR (Figure 2A). For indel I, the 621-bp insertion was absent in all spring and semi-winter cultivars, but was present in all 18 winter cultivars except Coma. The 621-bp insertion showed a highly significant correlation with flowering phenotype (r2 = 0.93, Table 2; Figure 2B). In contrast, Indel II (r2 = 0.49) and indel IV (r2 = 0.56) were only weakly associated with flowering phenotype (Table 2).

Table 1 Phenotypic and genotypic data for 79 B. napus accessions
Figure 2
figure 2

Association of polymorphic sites (indels and SNPs) in BnFLC.A10 and flowering phenotype in rapeseed cultivars that were planted in spring. (A) Genotyping of indels by PCR in some of the analyzed cultivars. Because the polymorphic products for indel III could not be distinguished by PCR, results for this indel are not shown. Lanes 1–19 represent the corresponding PCR products amplified from the genomic DNA of accessions of Apache, Bakow, Bienvenu, Bolko, Brutor, Casino, Ceres, Diadem, JeT-Neuf, Coma, Apomix, Chuanyou11, Dac-chosen, Dunkeld, Erglu, Huashuang2, Jiayou1, Huáyou2 and Karoo. For detailed genotypic information, see Table 1. (B) Association between indel I and flowering phenotype.

Table 2 Haplotypes detected with the sequence information of BnFLC.A10 from 24 B. napus accessions

Haplotyping of BnFLC.A10 specific markers for indels I–IV and SNPs 1–6 (Figure 1D) confirmed that most winter rapeseeds had a 621-bp insertion in the upstream region of BnFLC.A10 (haplotype I), whereas the 621-bp fragment was absent in the spring types (Table 2). These results suggest that indel I (with the 621-bp fragment present or absent) in the upstream region of BnFLC.A10 plays a very important role in modulating flowering time in natural rapeseed germplasm and potential development of a winter growth habit.

The 621-bp insertion in the upstream region of BnFLC.A10 is a Tourist-like MITE

To further characterize the 621-bp insertion sequence in winter-type rapeseed accessions, BnFLC.A10 alleles from the eight cultivars that representing haplotype I (Table 2) were sequenced and aligned. All the sequenced genotypes showed 100% identity. The inserted sequence possessed typical characteristics of a Tourist-like MITE [3840], with 14-bp terminal inverted repeat (TIR) sequences flanked by target sequence duplications (TSDs) of TAA (Figure 3A). Between the TIR sequences, an AT-rich (67%) core that contained 12 classes of important motifs (such as the TATA box and CAAT box) was identified (Additional file 1). These motifs might function in transcriptional initiation or promotion, or in response to different stimuli and signals (Additional file 1). At least four homologs of the MITE insertion (BLASTN expected value < 1e-10) were identified in the genomic sequence of B. napus from public databases ( and up to 200 copies (E-value < 1e-10) were identified in the sequenced genome of B. rapa, the ancestral source of the A genome in B. napus. These homologs defined a new family of MITEs, which we named Monkey King (Figure 3B) after the subject of a Chinese myth. (In “Journey to the west”, Monkey King is capable of 72 methods of transformation and can transform hundreds of monkeys with one of his hairs. He also jumps long distances with a cloud somersault).

Figure 3
figure 3

Structure of the 621-bp MITE and its family. (A) Basic structure of the 621-bp MITE and elements in the Monkey King family. The consensus sequences of the TIRs and TSDs are shown. The length and content of the core AT-rich sequence varied among homologs of Monkey King. The numbers marked in the frame of core sequences represent the length (without TAAs and TSDs) of Monkey King upstream of BnFLC.A10 and its homologs in the B. rapa genome. W=(A/T), Y=(C/T), M=(A/C), R=(A/G), and S=(C/G). (B) Flanking sequence of Monkey King at the 5’ upstream end of BnFLC.A10-T and the corresponding Related Empty Sites (RESites) in B.rapa. (C) Three replicates of the GACTGGTT motif scattered near the 5’ end region of Monkey King. The sequence of Monkey King is shown, with dots representing omitted portions. TSDs are underlined and TIRs are marked with arrows. GACTGGTT motifs are framed in red; all of the three duplications are located near the 5’ end region.

Origin and transmission of the BnFLC.A10 621-bp insertion in B. napus and its A genome ancestral species B. rapa

To understand the evolutionary process behind the adaptation associated with the insertion of Monkey King into the upstream region of BnFLC.A10 and to trace its origin and transmission, we investigated an additional 154 spring cultivars of B. napus and 103 cultivars (including the genome sequenced cultivar, Chiifu [41]) belonging to nine subspecies of B. rapa (oilseed, swede and fodder types, Additional file 2). No Monkey King insertion was detected in the upstream region of BnFLC.A10 in any of the accessions, even though the empty site of insertion was almost 100% identical to the sequences that flanking the Monkey King insertion in BnFLC.A10 in winter rapeseed (Figure 3B and Additional file 2). On the other hand, hundreds of copies of Monkey King were detected in the whole genome, but not in the BrFLC.A10 upstream region of B. rapa ‘Chiifu’. This suggests that Monkey King may have pre-existed in the B. rapa genome but was inactive, after the generation of B. napus, it was activated and inserted into the upstream region of BnFLC.A10, giving rise to winter rapeseed.


In this study, we used positional cloning to dissect the major flowering time QTL, qFT10-4, which was detected only in the spring-cropped TN DH population. The QTL qFT10-4 on chromosome A10 was delimited in a narrow 80-kb genomic region and annotation of different genes allowed us to identify BnFLC.A10, an ortholog of FLC, as the candidate gene. We demonstrated for the first time that flowering time variation at the qFT.10-4 locus is conditioned by the major vernalization response gene, BnFLC.A10; the MITE insertion upstream of BnFLC.A10 show significant association with the flowering time variation between winter and spring rapeseed.

Control of flowering time by vernalization has previously been shown to depend on a complex regulatory network, especially in amphidiploid rapeseed. In one study of the relationship between flowering time and FLC orthologs, five BnFLC sequences were isolated from B. napus cDNA library and in another study six FLC paralogs have been identified in B. napus by comparative analysis of B. napus and Arabidopsis genomes [36, 42]. The fact that indel I in the upstream region of BnFLC.A10 cosegregated with flowering phenotype in the TN DH population but in only some of the diverse cultivars might be due to the contribution of other flowering time QTLs, including other BnFLCs, with very small genetic effects. For examples, one of the BnFLC paralogs, which was located in linkage group A3 (BnFLC.A3b), colocalized with the flowering time QTL and thus might contribute to the vernalization response in certain cultivars [43]. In fact, at least nine copies of Bn.FLC genes exist in rapeseed [43]. Other genes, such as FRIGIDA, also regulate FLC expression in rapeseed; BnaA.FRI.a, one of orthologs of FRIGIDA in Arabidopsis, contributes to flowering time variation in rapeseed, which may partly explain why indel I did not totally cosegregated with flowering time in the association analysis [33]. In our research, the cultivar “Coma” that lacked the BnFLC.A10 upstream MITE insertion still exhibited the winter characteristic. It is thus possible that other copies of BnFLCs or related genes from the vernalization pathway may contribute to vernalization response in B. napus. The expression of one of the other BnFLC copies, or of all BnFLC copies acting in concert in the Coma genome, may be sufficient to inhibit flowering transition under spring environmental conditions, therefore enabling the cultivar ‘Coma’ to function as a winter rapeseed. Genetic diversity with respect to BnFLCs and other Arabidopsis vernalization pathway gene homologs has been associated with vernalization in B. napus, but no gene or polymorphic site as strongly associated with vernalization requirement of rapeseed as the BnFLC.A10 upstream MITE insertion has been dissected previously. The MITE insertion into BnFLC.A10 appears to be one of the most important causative factors of vernalization requirement in winter rapeseed cultivars.

It is believed that rapeseed originated from a natural hybridization between B. rapa and B. oleracea that occurred in southern Europe along the Mediterranean coastline approximately 10,000–100,000 years ago. Given the warm climate in this region year-round, naturally occurring rapeseed genotypes and their ancestors may not have needed to develop an adaptation requiring prolonged vernalization. The activation of Monkey King in B. napus genome would have introduced diversity into the germplasm upon which selective pressure could act. The insertion of Monkey King in the upstream region of BnFLC.A10 resulted in strong dependence on vernalization for flowering; this characteristic was then selected by plant breeders during the development of winter-type rapeseed cultivars of rapeseed that could be grown in northern Europe and other temperate regions of the world.

Several studies have examined the effect of MITEs on neighboring gene expression. For example, the DNA methylation level of a MITE can influence expression of neighboring genes. An assay of transient and stably-transformed rice revealed that the MITE Kiddo, when present in the promoter of the rice ubiquitin2 (rubq2) gene, was responsible for up to 20% of neighboring gene expression; most notably, when DNA methylation of Kiddo was blocked, transcript levels of endogenous rubq2 increased threefold [44]. An association has also been reported between a MITE inserted in the upstream regulator region of the gene Vgt1 (Vegetative to generative transition) and early flowering in Northern maize genotypes [45, 46]. In our study, the MITE upstream of BnFLC.A10 was positively associated with gene expression and induced BnFLC.A10 expression during vernalization. The inserted MITE seems to attenuate cold-induced BnFLC.A10 repression rather than increase its expression, in winter rapeseed. This result is very similar to that observed in Arabidopsis, where FLC expression was correlated with flowering time and vernalization requirement in unvernalized or long days, but not as strongly as anticipated [47, 48]. We thus conclude that either decreased rate of FLC expression during vernalization or additional epistatic interaction with other genes is more important for control of flowering time and vernalization requirement than variation of FLC expression under unvernalized conditions. Using motif prediction, motifs associated with gene regulation were found to exist in the Monkey King sequence (Additional file 1). Most of these motifs were located in gene promoter and enhancer regions (TATA box and CAAT box) or were light responsive elements (Sp1) (Additional file 1) associated with response to environmental signals in different organisms. Certain transcriptional factors presumably bind to this region to more efficiently initiate or enhance the expression of neighbouring genes. The actual protein binding ability of the 621-bp insertion was evaluated using electrophoretic mobility shift assays (EMSAs). Nuclear protein(s) extracted from Tapidor before vernalization were able to bind to some fragments from the middle of the 621-bp Monkey King region that contained TATA box motifs (unpublished data). These results suggest that Monkey King can bind to specific transcription factors that may initiate or enhance BnFLC.A10 expression in winter rapeseed cultivars, giving rise to their stronger vernalization requirement.

Our analysis also indicated that Monkey King is involved in gene regulation in many different settings in the genome. For example, we found three copies of the sequence GACTGGTT near the 5’ end of Monkey King (Figure 3C); this motif is conserved in the upstream region of Dsg1 (desmoglein1, which encodes desmosomal cadherin) in mice. The motif in Dsg1 is recognized by GRHL1 (grainyhead-like 1, a homolog of the Drosophila gene grainyhead) and increases Dsg1 expression [49]. Part of the Monkey King sequence is transcribed in Brassica genomes (, Table 3), and has been identified in the 3' untranslated region of the WRKY21-1 gene (EU912394). Other transcripts that share high similarity with portions of the Monkey King sequence have been found in the expressed sequence tag library of rapeseed ( The presence of these transcripts suggests the existence of a novel gene regulatory mechanism that is similar to the method by which exon shuffling generates new genes [50, 51] or overlapping transcripts generate siRNAs to regulate gene expression [52, 53]. It is possible that transcripts derived from Monkey King might regulate native gene expression through siRNA-induced DNA methylation. MITE activities within BnFLC.Al0 may have shaped phenotypic diversity and influenced mechanisms of adaptation to diverse climates during the evolutionary process.

Table 3 20 sequences that show high similarity with Monkey King in the B. napus EST library


This study demonstrated that BnFLC.A10 is the highly likely causative gene underlying qFT10-4, which accounted for most flowering time variation in the TN DH population under spring environmental conditions. Comparision of allelic sequences from Tapidor and Ningyou7 revealed the presence of a Tourist-like MITE insertion in winter-type cultivar Tapidor. Association analysis among winter- and spring-type rapeseeds revealed that the presence of the Tourist-like MITE insertion is very strongly associated with vernalization requirement, and suggested that it appeared after B. napus was generated as a product of natural hybridization between B. rapa and B. oleracea. MITE activity led to genetic and phenotypic diversities among varieties and provided the fuel for evolutionary selection. As a result, winter genotypes may have evolved from spring genotypes; this useful variation has subsequently been used as a genetic resource for the development of winter cultivars enabling worldwide production of rapeseed.


Plant materials

For fine mapping of the BnFLC.A10 locus, we used 9,000 plants derived from four BC5F1 individuals: 8y085-1, 8y086-1, 8y086-2 and 8y086-4. TN DH043 (the 43rd line of the TN DH population) was crossed wiht Ningyou7 (semi-winter recurrent parent) and seeds were collected from the F1 generation. Plants were then backcrossed with Ningyou7 over five successive generations (BC1 to BC5). Molecular markers were used to track the Tapidor allele at the BnFLC.A10 locus in the F1 backcross. The BC5F2 near-isogenic lines were planted in the spring of 2009 for phenotyping with respect to flowering time. A panel of 79 diverse rapeseed cultivars (Table 1) was used for the association analysis. These cultivars were planted in spring during three successive years (2007–2009) for phenotyping. Climatic conditions during the planting season and geographic features of the planting site were as described previously [36]. The spring rapeseed and B. rapa accessions representing nine subspecies that were used to detect the presence of Monkey King upstream of BnFLC.A10 and BrFLC.A10 are listed in Additional file 2. These accessions were obtained from the National Brassica Germplasm Improvement Program (Wagga Wagga, Australia), the Australian Temperate Field Crops Collection (Horsham, Australia), and from the Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences (Beijing, China).

Phenotypic evaluation

Flowering times of the different cultivars used for the association analysis were recorded as the number of days from the day of sowing to the day when 50% of plants in the plot flowered. In the BC5F2 populations, days to flowering (DTF) were recorded as the number of days from the day of sowing to the day when the first flower opened. The phenotype ‘non-flowering’ was assigned when plants showed no visible buds at autumn harvest in middle-October. Phenotypes of the 79 cultivars used for the association analysis are listed in Table 1.

Sequencing of BnFLC.A10alleles from Tapidor and Ningyou7

A fragment amplified from the Tapidor genome with primer pair “Exon 4-7” (Table 4) was used as a probe to screen the Tapidor BAC library [54]. From 12 BAC clones that contained BnFLC.A10, one clone with the code JBnB75D10 was selected and sequenced to obtain the BnFLC.A10-T allele. Primers (P4, P5, exon1-2,exon2-4, exon4-7, Table 4)were designed based on the basis of the BnFLC.A10-T sequence and used to obtain the sequence of BnFLC.A10-N. The amplicons were cloned into a pGEM-T Easy vector (Promega, Madison, WI, USA) and sequenced to determine the BnFLC.A10-N sequence. Information on primers and amplified gene regions is provided in Table 4. BnFLC.A10-N and BnFLC.A10-T sequences were obtained, accession numbers [GenBank: JX901141 and JX901142].

Table 4 Sequence information for primers used for polymorphism and BnFLC.A10 gene expression analysis

Gene annotation for the BAC sequence

Gene annotation was carried out using the FGENESH program by selection of the organism category “Dicot plants (A. thaliana)” and alignment with A. thaliana genes. Simple sequences and transposons were identified using RepeatMasker (, validated 19th September, 2011) followed by manual inspection. We predicted the function of genes that were not aligned with A. thaliana orthologs from their conserved domains.

RNA extraction and q-RT PCR

Plants were grown under long-day conditions (16 h light/8 h dark) at 23°C until they had developed to the six-leaf stage, at which point they were transferred to 4°C for vernalization. BnFLC.A10 expression was analyzed in plants that had been subjected to a (control), 1, 4 and 7 weeks of vernalization. Total RNA was extracted from plant leaves using TRIzol® reagent (Invitrogen, Carlsbad, California, USA). Total RNA (2 μg) was reverse-transcribed using M-MLV Reverse Transcriptase (Promega). An iQ5 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) was used for quantitative RT-PCR to detect levels of BnFLC.A10 expression in the two parents. BnFLC.A10 primers (Prt f/Prt r, Table 4) amplified a 235-bp fragment of the BnFLC.A10 CDS. Two genes, actin2 and 18S rRNA (Table 4), were used to normalize expression levels. Three biological and technical replicates were analyzed.

Natural variation in BnFLC.A10

Allele-specific primers “Itr1f/Itr1r” (Table 4) were used to distinguish the BnFLC.A10 Tapidor allele from the Ningyou7 BnFLC.A10 variant in 24 rapeseed cultivars. The PCR products were cloned into a pGEM-T Easy vector (Promega) for sequencing. Plasmid prepared from two to four colonies from each PCR product was sequenced separately to minimize the contribution of polymerase errors to sequence variation.

Screening of homologous sequences of Monkey King in the B. rapagenome

To identify homologous sequences, the full length MITE sequence was queried against the B. rapa genome in the brassicadb database ( using BLAST. Results were filtered using an E value<1e-10 as the cutoff.