|Home | About | Journals | Submit | Contact Us | Français|
Flowering time is an important agronomic trait, and wide variation exists among Brassica rapa. In Arabidopsis, FLOWERING LOCUS C (FLC) plays an important role in modulating flowering time and the response to vernalization. Brassica rapa contains several paralogues of FLC at syntenic regions. BrFLC2 maps under a major flowering time and vernalization response quantitative trait locus (QTL) at the top of A02. Here the effects of vernalization on flowering time in a double haploid (DH) population and on BrFLC2 expression in selected lines of a DH population in B. rapa are descibed. The effect of the major flowering time QTL on the top of A02 where BrFLC2 maps clearly decreases upon vernalization, which points to a role for BrFLC2 underlying the QTL. In all developmental stages and tissues (seedlings, cotyledons, and leaves), BrFLC2 transcript levels are higher in late flowering pools of DH lines than in pools of early flowering DH lines. BrFLC2 expression diminished after different durations of seedling vernalization in both early and late DH lines. The reduction of BrFLC2 expression upon seedling vernalization of both early and late flowering DH lines was strongest at the seedling stage and diminished in subsequent growth stages, which suggests that the commitment to flowering is already set at very early developmental stages. Taken together, these data support the hypothesis that BrFLC2 is a candidate gene for the flowering time and vernalization response QTL in B. rapa.
In flowering plants, the change from vegetative to reproductive development is a major transition that is sensitive to various seasonal climatic signals (Koornneef et al., 2004). Controlling the timing of this transition is especially important in crop plants to ensure high agricultural productivity.
Many genes that control flowering time have been identified by analysing Arabidopsis mutants (Boss et al., 2004; Koornneef et al., 2004). These studies have shown that in Arabidopsis multiple pathways are involved in controlling flowering time, including the vernalization, photoperiod, autonomous, and gibberellin pathways (Mouradov et al., 2002; Jack, 2004; Schmitz and Amasino, 2007; Alexandre and Hennig, 2008; Seo et al., 2009). The largest difference in flowering time among Arabidopsis ecotypes appears to be due to allelic variation at the FLC (FLOWERING LOCUS C) and FRI (FRIGIDA) loci (Koornneef et al., 2004; Engelmann and Purugganan, 2006). FRI acts upstream of FLC to regulate FLC expression positively (Michaels and Amasino, 1999). FLC encodes a MADS-box transcription factor that functions as a repressor of flowering time in the vernalization pathway (Sheldon et al., 1999, 2000, 2008; Schmitz and Amasino, 2007). Genetic analyses have revealed that the difference in flowering time between early and late flowering ecotypes is largely dependent on allelic variation in FLC (Michaels and Amasino, 1999; De Lucia et al., 2008).
Vernalization represses the expression of FLC and promotes flowering in vernalization-responsive late flowering Arabidopsis, and this repressed state is stable during plant development, while the gene is reset to an active transcriptional state in the next sexual generation (Michaels and Amasino, 1999; Sheldon et al., 2000; Schmitz and Amasino, 2007). The extent of promotion of flowering by cold is proportional to the duration of cold exposure; shorter periods of cold exposure result in a smaller acceleration of flowering time and only partial repression of FLC expression (Sheldon et al., 2000).
The Brassica genus, comprising a large and diverse group of important oil, vegetable, fodder, and condiment crops, is closely related to Arabidopsis thaliana, which is a member of the Brassicaceae family. Cultivated varieties of the diploid species Brassica rapa can differ based on morphological appearance and consumed organs, which include leafy types, turnip types, and oil types, and exhibit a diverse range of developmental and morphological traits (Zhao et al., 2005). Flowering time is a very important developmental trait, and wide variation exits among B. rapa accessions, and their flowering habits are generally controlled by temperature and/or day length. Vernalization is the acquisition of the competence to flower by exposure to prolonged periods of cold or low temperatures. Brassica rapa, in contrast to B. oleracea (Lin et al., 2005), can be vernalized at the germinated seed stage, which can shorten the generation time and thus speed up genetic research. However, the vernalization response is different among the different cultivar groups. In general, oil types and several pakchoi cultivars flower very early even under non-vernalized conditions. The very late flowering types are mainly Chinese cabbages and turnips, which need long-term vernalization to accelerate flowering.
In previous QTL (quantitative trait locus) analyses of flowering time in Brassica, evidence has been presented for a role for FLC genes as candidates underlying the flowering time QTL in B. napus, B. oleracea, and B. rapa (Teutonico and Osborn, 1995; Osborn et al., 1997; Tagede et al., 2001; Schranz et al., 2002; Lou et al., 2007; Okazaki et al., 2007). A number of FLC paralogues in B. rapa (BrFLC1, BrFLC2, BrFLC3, and BrFLC5) have been cloned, and mapped both using in situ hybridization and genetically (Schranz et al., 2002; Kim et al., 2006; Yang et al., 2006). The study of BrFLC expression in Chinese cabbage (Kim et al., 2007) and B. napus (Tadege et al., 2001) indicates that the Brassica FLC genes act similarly to Arabidopsis FLC. A major QTL with BrFLC2 as the candidate gene on A02 was identified in several B. rapa populations evaluated in different environments, locations, and seasons (Lou et al., 2007). In another study, the analysis of sequence variation of BrFLC1 on A10 suggested that a naturally occurring splicing mutation in the BrFLC1 gene is associated with flowering time variation in B. rapa (Yuan et al., 2009).
From the above it is postulated that the different FLC paralogues in B. rapa play a role in modulating flowering time and the vernalization response. This paper addresses the effect of vernalization on flowering time in a double haploid (DH) population from a cross between an early flowering yellow sarson and a pakchoi accession and on BrFLC2 expression in selected lines of this DH population. The possible roles of BrFLC2 in flowering time regulation and vernalization response are discussed.
DH population 38 was established from a cross between pakchoi PC-175 (cultivar: Nai Bai Cai; accession number: VO2B0226) and yellow sarson YS-143 (accession number: FIL500) (Lou et al., 2008). A total of 71 lines from population DH38 were evaluated for flowering time under vernalized conditions. Seedlings were germinated on wet filter paper at 25°C, thereafter vernalized in the dark at 5°C for 18d or 31d, and all transplanted to soil (pots of Ø 17cm) at the same time. Plants were grown at 24/18°C (day/night) with a 16h photoperiod in a greenhouse during spring (February to May) of 2007 in Wageningen University. The two sets of 71 DH lines (vernalized for 18d and 31d) were grown in a randomized complete design across three blocks.
The same set of DH lines was also evaluated for flowering time under non-vernalized conditions during winter 2004 (FL04wi), autumn 2005 (FL05au), and spring 2005 (FL05sp). Wide variation existed for flowering time within population DH38 and between the different seasons (Lou et al., 2007). Based on these flowering time data and QTL analyses together with the graphical genotype data of DH lines of population DH38, five early and five late DH lines with and without the A02 QTL, respectively, and the two parental lines (PC-175 and YS-143) were selected for further analysis (Table 1). The expression of BrFLC2 mRNA was studied in these genotypes after six different vernalization treatments. The germinated seeds of selected DH lines and the parents of the population were vernalized at 5°C in the dark for 31, 21, 11, 5, 2 and 0d. Then six seeds per genotype were simultaneously transferred to soil (pots of Ø 17cm), placed in a climate room, and grown under a 12h photoperiod at 20/18°C (light/dark).
In the experiment of flowering time QTL analysis, the number of days to flowering (FL) of the DH38 plants after vernalization for 18d (FL07sp-v18) or 31d (FL07sp-v31) was measured from transplantation into the greenhouse until the first flower opened. The vernalization response (VR) was calculated based on the difference in days to flowering between this experiment (FL07sp-v) and the experiment of spring 2005 (FL05sp) under non-vernalized conditions (Lou et al., 2007), VR-v31=FL05sp–FL07sp-v31, VR-v18=FL05sp–FL07sp-v18.
After collecting samples for RNA isolation from plants of six vernalization treatments (from germinated seeds until 7d after the seventh leaf stage), 2–4 plants per DH line and parental line were transferred from the climate room to a greenhouse and grown under a 16h photoperiod at 24/18°C (day/night) during autumn 2007 (September–October). The flowering time was evaluated for each plant from transplantation into soil until the first flower opened.
The linkage maps of DH38 have been constructed by the JoinMap program 4.0 based on amplified fragment length polymorphism (AFLP) and simple sequence repeat (SSR) markers (Lou et al., 2008). BrFLC2 was mapped on A02 using SSR markers, and co-localized with the major flowering time QTL. Candidate QTLs that affect traits of interest were identified by interval mapping (IM) and verified by multiple-QTL model mapping (MQM) methods using the software MAPQTL 5.0 (Van Ooijen, 2004). The analysis started with the IM test to find putative QTLs. MQM analysis was then performed to locate QTLs precisely after the automatic selection of cofactors in the vicinity of the QTL. Only significant markers at P <0.02 were used as cofactors in the multiple QTL detection. A map interval of 5cM was used for both IM and MQM analyses. A permutation test was applied to each data set (1000 repetitions) to decide the LOD (logarithm of odds) thresholds (P=0.05). In this study, a LOD value of 3.00 was used as a significant threshold. QTLs were graphically displayed using Map chart 2.2 (Voorrips, 2002).
Material for RNA isolation was harvested from several tissues at several developmental stages (12h photoperiod at 20/18°C). Supplementary Table S1 available at JXB online lists the different tissue samples and collection times/plant developmental stages that were harvested for RNA extraction. Plant samples were collected at different developmental stages: seedling stage (germinated seeds with sprout before transplanting, after vernalization treatment); third leaf stage (cotyledons and first and second leaf), fifth leaf stage (first and second leaf and third and fourth leaf), seventh leaf stage (first and second leaf and third and fourth leaf), and the latest stage, which is 7d after the seventh leaf stage (first and second leaf and third and fourth leaf). For the seedling stage, three germinated seeds with sprout were collected for each genotype after vernalization. For other developmental stages, the samples were collected from one different individual plant per DH line or parental genotype. At the latest collection stage 7d after the seventh leaf stage, plants of early DH lines and YS-143 already had flower buds.
Total RNA was extracted using TRIZOL reagent (Invitrogen) starting with ~300mg of leaf powder. RNA concentration and purity were determined with Nanodrop, and quality was checked by agarose gel electrophoresis. Total RNA (5μg) of all samples was treated with DNase I (Invitrogen) according to the manufacturer's instructions. DNA-free total RNAs of the early and late DH lines were pooled in equal concentrations and thereafter both parental and pool RNA were converted into cDNA using oligo(dT) primers, dithiothreitol (DTT; 0.1M), RNaseOUT, and Superscript II (Invitrogen). cDNA was quantified and used for semi-quantitative reverse transcription-PCR (semi-quantitative RT-PCR) and for real-time RT-PCR.
For semi-quantitative RT-PCR, two pairs of BrFLC2-specific primers (FLC2e4-7, forward 5′-AGTAAGCTTGTGGAATCAAATTCTG-3′, reverse 5′-TAATTAAGYAGYGGGAGAGTYAC-3′; FLC2e1-4, forward 5′ CAAGCGAATTGAGAACAAAA-3′, reverse 5′-GAGTCGACGCTTACATCAGA-3′) were selected based on previous reports by Schranz et al. (2002) and Kim et al. (2007), respectively (Fig. 1). The PCR program used was: 94°C for 3min, 30 cycles of 94°C for 1min, annealing at a temperature of 55°C for 1min, 72°C for 1min 30s, and a final extension at 72°C for 7min.
BrFLC2 gene expression was quantified in the early and late pools of DH lines and the parents of population DH38 using quantitative real-time PCR (qRT-PCR). FLC coding sequences from B. rapa (AY115675, AY115677, AY115678, AY205317, and AY205318) were joined and aligned with the A. thaliana mRNA sequence (AF116527). The presence of different homologues of FLC (BrFLC1, BrFLC2, BrFLC3, and BrFLC5) in B. rapa limited the choice for BrFLC2-specific primer selection to exon 4, which is variable (Schranz et al., 2002) between different FLC paralogues. In this study, BrFLC2 gene-specific primers (FLC2e3/4–5/6: forward 5′-CTTGTCGAAAGTAAGCTTGTGGAAT-3′; reverse 5′-GCATTTTCTCCTTTTCTTTGAGG-3′) were designed in the junction of exon 3–4 and exon 5–6 with an amplicon size of 150bp (Fig. 1). The efficiency of cDNA synthesis was assessed by using the constitutively expressed gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (forward 5′-AGAGCCGCTTCCTTCAACATCATT-3′, reverse 5′-TGGGCACACGGAAGGACATACC-3′) which is frequently used as the reference gene in expression studies (Broekgaarden et al., 2007).
QRT-PCR analysis was done with the MyiQ single-colour real-time PCR detection system (Bio-RAD, Veenendaal, The Netherlands) using SYBR green to monitor double-stranded DNA synthesis. Each reaction contained 10μl of 2× IQ SYBR Green super mix reagent (Bio-RAD), 10ng of cDNA, and 300nM of gene-pecific primer in a final volume of 20μl. The thermal cycling consisted of 95°C for 30s, 95°C for 3min, and 40 cycles of 95°C for 10s and 52°C for 45s. After the PCR a melting curve was generated to check the specificity of the amplified fragment. CT (threshold cycle) values were calculated using optical system software, version 2.0 for MYIQ (Bio-RAD). Subsequently, CT values were normalized for differences in cDNA synthesis by subtracting the CT value of GAPDH from the CT value of BrFLC2. Again these values are normalized with a calibrator sample, which was the CT value of BrFLC2 from the fifth leaf stage (0 day vernalization, late line third and fourth leaf together). Fold change (2n) in gene expression was calculated by using the –ΔΔCT method given by Livak and Schmittgen (2001).
Semi-quantitative RT-PCR with primers FLC2e1-4 resulted in an amplification product of the expected size (~320bp) and a larger extra band (~400bp) in pools of early DH lines. Intron 3 of FLC2 from early and late pools was amplified from genomic DNA by using the primers in exon 3 and exon 4 (FLC2e3-4: forward 5'-AAGTATGGTTCACACCATGAG-3'; reverse 5'-GAGTCGACGCTTACATCAGA-3'). The qRT-PCR product of the unexpected smaller size (79bp) that was amplified with qRT-PCR primers in samples of pools of early DH lines at the third leaf stage was also sequenced.
The PCR products were excised from agarose gels, purified by QIAquick gel extraction kits (Qiagen), and cloned into PGEM-T easy vector (Promega Corp.). Escherichia coli strain DH5α (Invitrogen) was transformed with these constructs and the positive colonies were selected. Plasmid DNA was isolated and sequenced using corresponding primers and aligned in CLASTALV.
The variation in flowering time of vernalized plants in this experiment (FL07sp-v), of non-vernalized plants in the experiment of spring 2005 (FL05sp), and the variation in the vernalization response for parental lines and corresponding populations of DH38 is shown in Table 2. As can be seen, the mean value of flowering time is enormously reduced in DH38 after vernalization for 18d and 31d. After vernalization for 18d or 31d, respectively, the variation in days to flowering of DH population was still large (27–70 or 27–61), while the difference between parental lines was only 4d (32 and 36) or 1d (29 and 30). Transgression within the population for days to flowering was observed mainly towards late flowering. The vernalization response was also calculated, which had a similar range after 18d (5–40) compared with after 31d (5–45).
In total, three QTLs for flowering time (FLQTLs) and two QTLs for vernalization response (VRQTLs) were detected in DH38 on two linkage groups (Table 3). One major QTL, FLQTL-1, on the top of A02 for flowering time in non-vernalized experiments (FL05sp) was detected and explained 64% of the variation (Table 3). After 18d vernalization, FLQTL-2 was detected at the same position on A02, explaining 30.9% of the variation for flowering time. After 31d of vernalization the FLQTL on A02 disappeared (Fig. 2). One minor flowering time QTL explaining 15% of the variation (FLQTL-3) in the experiment with 18d vernalization was detected on A08; this QTL was not found in non-vernalization experiments. Two QTLs for the 18d and 31d vernalization response (VRQTL-1 and VRQTL-2) were detected on the top of A02, which explain 72% and 77% of the variation, respectively, and appear to be the main loci responsible for the vernalization response in this population.
For further analysis of the effect of vernalization on flowering time, five early and five late DH lines of population DH38 were selected (Table 1); early lines had yellow sarson alleles at the flowering time QTL region on A02, while late lines had pakchoi alleles. Without vernalization the early lines flowered after an average of 58d, while late lines flowered after 98d (Fig. 3). Upon vernalization for 31d, early lines flowered after 32d and late lines after 56d, which is a decrease of ~30d. The yellow sarson YS-143 parent and early pools showed a similar reduction in flowering time, whereas flowering time in late lines is transgressive over the pakchoi PC-175 parent, with a similar reduction upon vernalization. The biggest decrease in flowering time occurred between 5d and 11d of vernalization in late DH lines and the PC-175 parent.
Semi-quantitative RT-PCR was conducted with BrFLC2-specific primers (FLC2e4-7 and FLC2e1-4) in pools of early and late DH lines and parental lines to quantify BrFLC2 expression in different developmental stages and tissues (seedlings, cotyledons, or leaves) after different vernalization periods. As seen in Fig. 4, BrFLC2 transcript was more abundant in late pools compared with early pools in all stages/tissues. Reduction in flowering time after vernalization also correlated with the decrease in BrFLC2 transcript as indicated by semi-quantitative RT-PCR (Fig. 4A).
In RT-PCR products using primers FLC2e1-4 of first and second leaves harvested from fifth leaf stage plants, an extra larger band (400bp) appeared in addition to the expected band of 320bp in the early pools (Fig. 4B). This 400bp fragment was cloned and sequenced and the sequences were aligned to known FLC sequences. This showed that intron 3 of 86bp was retained, as a result of alternative splicing with intron retention (Supplementary Fig. S1 at JXB online). Alternative splicing was clear in early lines in all developmental stages, whereas in the seedling stage alternative splicing was also observed in late pools and in the pakchoi parent (PC-175) (Supplementary Fig. S2 at JXB online). Intron 3 was amplified from both early and late pools using primers FLC23-4 and cloned. Sequence analysis of intron 3 showed that the sequences were identical in both late and early pools (Supplementary Fig. S3 at JXB online).
Regulation of BrFLC2 gene expression and flowering time by vernalization was further studied by qRT-PCR. To avoid amplification and quantification of several alternatively spliced transcripts, specific primers were designed over exon–exon junctions (FLC2e3/4–5/6: forward, 11bp in exon 3 and 14bp in exon 4; reverse, 14bp in exon 5 and 9bp in exon 6) to amplify only BrFLC2 mRNA with properly spliced intron 3. The expected semi-quantitative RT-PCR product size and sequence data indicated that FLC2e3/4–5/6 primers are indeed FLC2 specific.
Quantitative BrFLC2 expression in different tissues and from different developmental stages after different vernalization treatments of parents and pools of late and early DH lines is shown in Fig. 5 and Supplementary Fig. S4 at JXB online. The differences in BrFLC2 expression between vernalized and non-vernalized plants are the largest in the seedling stage and slowly diminish. In the late pools of DH lines, the level of BrFLC2 expression decreases with vernalization time, and this effect is indeed the clearest in the seedling stage, with a 32- (25) fold decrease in vernalized seedlings for 31d compared with non-vernalized seedlings. BrFLC2 expression is correlated with vernalization duration. During plant development, the difference in BrFLC2 levels in similar tissues exposed to different durations of vernalization decreases. In leaves of plants in the third leaf stage, the difference in BrFLC2 expression between non-vernalized plants and plants vernalized for 31d, is >16- (23) fold and at later stages these differences diminish to only 2-fold. Figure 5 also shows that during plant development from seedling to seventh leaf stage BrFLC2 levels increase 16- (24) fold in plants vernalized for 31d, while for plants vernalized for 0–5d there is almost no increased expression.
In the pools of early lines the same trends can be detected. In the early DH pools the level of BrFLC2 expression also decreases with vernalization time and this effect is again clearest in the seedling stage. BrFLC2 expression in seedlings vernalized for 31d is 64 (26) times lower than that in non-vernalized seedlings. Also here BrFLC2 expression is correlated with vernalization duration, even though the seedlings of the 21d vernalization treatment have a rather high BrFLC2 expression. In the early lines the difference in BrFLC2 levels in similar tissues exposed to different periods of vernalization again decrease during plant development. In leaves of plants in the third leaf stage, the difference in BrFLC2 expression between non-vernalized plants and plants vernalized for 31d is decreased to 32- (24) fold but at later stages the decrease in expression still varies between 4- and 16-fold, which is more than in the late pools of DH lines. Again during plant development from seedling to seventh leaf stage BrFLC2 levels increase 8- (23) fold for plants vernalized for 31d, while in non-vernalized plants there is no increased expression during plant development. In Supplementary Fig. S4 at JXB online, the RNA levels of parents YS-143 and PC-175 are also presented; the results are very similar to those of the pools of DH lines.
There were some unexpected expression levels: for example unexpected high values in the third leaf stage of early pools (cotyledons, 11d vernalization; first and second leaves, 5d vernalization) and unexpected low values in the seventh leaf stage of late pools (first and second leaves, 5d vernalization).
The melting curve of samples collected from third leaf stage plants of early DH pools was irregular compared with that of other stages and pointed to the existence of yet another type of alternative splicing. The sequencing of this transcript (79bp) from early pools showed that 59bp of exon 4 and 22bp of exon 5 were excised; so yet another alternative splicing type (Supplementary Fig. S5 at JXB online).
The evaluation of flowering time in different seasons and growing conditions allows evaluation of the expression of QTLs in different environments. To identify the genetic regulation of variation in flowering time and the effect of vernalization, QTL mapping was performed for flowering time and vernalization response in a DH population. The major flowering time QTL, FLQTL-1, on A02 clearly decreases upon vernalization. It explains 63.8% of phenotypic variation in the experiment of spring 2005 in DH38 compared with 30.9% in spring 2007 after 18d vernalization (FLQTL-2), and was not detected after 31d vernalization (spring 2007). A flowering time QTL explaining 45% of the variation was also detected at the same position on A02 in a large DH population DH68 (165 individuals) derived from a reciprocal cross between YS-143 and PC-175 (data not shown). Since FLC is described to be repressed upon vernalization (Koornneef et al., 2004; Sheldon et al., 2008) and BrFLC2 co-segregates with the flowering time and vernalization response QTL on the top of A02 in this study, it is suggested that BrFLC2 is a candidate gene for this QTL in the DH population used in this study. To study the role of BrFLC2, pools of late and early flowering DH lines were composed based on flowering time and only PC-175 or YS-143 alleles at the A02 QTL, respectively, while the lines in each pool had either pakchoi PC-175 or yellow sarson YS-143 alleles at the BrFLC1 locus on A10. To investigate the function of alleles of other flowering time QTLs or alleles of flowering time genes on vernalization, different pools have to be formed selected from larger populations, such as the DH68 population.
Earlier experiments of Sheldon et al. (1999, 2000) showed that vernalization promotes flowering by reducing the level of FLC transcript and protein. We observed a clear down-regulation of BrFLC2 by vernalization, which is in agreement with results obtained with FLC not only in A. thaliana, but also in other Brassicaceae species such as B. napus, B. oleracea, B. rapa, and Thellungiella halophila (Tadege et al., 2001; Lin et al., 2005; Fang et al., 2006; Kim et al., 2007) and, more recently, outside this plant family (Reeves et al., 2007). Vernalization reduced the expression of the BrFLC2 gene by reducing the quantity of transcript, which correlated to a delay of flowering in both the early and late pools. In all developmental stages late pools showed more BrFLC2 transcript than early pools.
The expression of genes that indicate commitment to flowering is already apparent at a very early stage, with transcription peaks in the seedling stage (Kobayashi et al., 1999). Gendall et al. (2001) and Levy et al. (2002) described two phases in the repression of FLC by vernalization; an initial reduction of FLC activity that occurs during the cold exposure, and a subsequent maintenance of the repressed state during growth and development of the plant after the end of the cold exposure. In the present experiment, quantitative reduction of BrFLC2 expression upon seedling vernalization is strongest in the seedling stage and diminishes in subsequent growth stages, which indicates that in B. rapa the decision for flower initiation is already set in the earliest developmental stages (seedling). Maintenance of the repressed state is not observed in the present experiment, since BrFLC2 levels already increased significantly in plants of the third leaf stage and the difference in BrFLC2 level between leaves of vernalized and non-vernalized plants decreased in later stages, especially in the late flowering pools.
The difference in flowering time under non-vernalized conditions between both early and late lines and their parents was greatly reduced after 11d of vernalization, which was caused by a larger reduction in flowering time in late pools and PC-175. After 31d of vernalization, in the late pool flowering time was reduced by 42d while the BrFLC2 transcript decreased 32-fold, whereas in the early pool flowering time was reduced by 30d with a 64-fold decrease in BrFLC2 transcript.
Analysis of Arabidopsis expressed sequence tags (ESTs) showed an unusually high fraction of retained introns (>30%) that may play a regulatory role (Ner-Gaon et al., 2004). The distribution of the transcripts with retained introns is skewed towards stress responses, like a cold treatment as used to vernalize seedlings. Macknight et al. (2002) showed that the alternative splicing of the FCA transcript has functional significance related to its role in the promotion of floral transition. Short transcripts due to alternative splicing of FLC were observed in Arabidopsis haplogroup FLCA and FLCB after a vernalization treatment of 15d, which were not observed at high levels in plants grown at normal temperature for 15d (Caicedo et al., 2004). The repression of FLC by cold treatment is also associated with changes in histone modification (Sung and Amasino, 2004; Finnegan et al., 2005). Previous studies reported regulation of FLC by the regions in its first intron (Sheldon et al., 2002; He et al., 2003), polymorphisms in which led to differential expression and differential splicing patterns. In the present experiment vernalization/cold stress resulted in differences in BrFLC2 expression that may be partly regulated by alternative splicing, which was detected for intron 3 in all stages in early pools and only in the very early stages of late pools. Real-time PCR using a forward primer in exon 4 with a reverse primer over the exon 5–6 junction amplifies both transcripts with a correctly spliced and retained intron 3. In contrast to the real-time PCR results using a forward primer over the exon 3–4 junction, real-time PCR using this exon 4 forward primer detected very little change in BrFLC2 transcript levels upon vernalization in pools of early lines (data not shown). In these early lines alternative splicing was obvious, which strongly contributed to the observed decrease in BrFLC2 transcript with correctly spliced intron 3. In late lines the intron 4 forward primer gave results comparable with those obtained with the primers over the exon 3–4 junction (data not shown), which agrees with the observation that in late lines intron 3 retention hardly occurs. Alternative splicing was not looked at in other parts of the gene not covered by the primers used in this study.
Multiple copies of B. rapa genes homologous to flowering time genes of Arabidopsis exist, and these multiple functional loci may contribute to the wide variation in flowering time in this species. There are four FLC paralogues in B. rapa; in this study BrFLC2 was found as a candidate gene for the large effect flowering time QTL on A02 in population DH38 from a cross of an early oil type and a middle late leafy type. Yuan et al. (2009) reported that the BrFLC1 gene also contributes greatly to flowering time variation in non-vernalized B. rapa. In a previous study (Lou et al., 2007), co-localization of the flowering time QTL with other possible flowering-related genes in multiple segregating populations was discussed. The roles of the different FLC paralogues and other flowering time-related genes and their interactions in B. rapa are presently being investigated further.
Supplementary data are available at JXB online.
Table S1. The different samples that were harvested for RNA extraction.
Figure S1. Sequencing of RT-PCR products in early and late pools using the BrFLC2-specific primers FLC2e1-4.
Figure S2. Semi-quantitative RT-PCR using BrFLC2-specific primers (FLC2e1-4) in seedling of early and late pools of DH lines and parent lines (P1, YS-143; P2, PC-175) vernalized for 0, 2, 5, 11, 21, and 31d.
Figure S3. Sequencing of intron 3 from early and late pools using specific primers FLC2e3-4.
Figure S4. Fold change (2n) in BrFLC2 expression (fold change) after different vernalization treatments (day) for pools of late and early DH lines.
Figure S5. Sequencing of unexpected smaller size (<100bp) qRT-PCR products in early pools at the third leaf stage using specific primers FLC2e3/4–e5/6.
This research was funded by the Programme Strategic Scientific Alliances (PSA) of the Royal Netherlands Academy of Arts and Sciences (KNAW) and the Chinese Ministry of Science and Technology (MOST). We thank staff at the Plant Sciences Experimental Centre of Wageningen for taking care of the plants, with special thanks to Andre Maassen.