Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
To elucidate the genetic mechanism of flowering in wheat, we performed expression, mutant and transgenic studies of flowering-time genes. A diurnal expression analysis revealed that a flowering activator VRN1, an APETALA1/FRUITFULL homolog in wheat, was expressed in a rhythmic manner in leaves under both long-day (LD) and short-day (SD) conditions. Under LD conditions, the upregulation of VRN1 during the light period was followed by the accumulation of FLOWERING LOCUS T (FT) transcripts. Furthermore, FT was not expressed in a maintained vegetative phase (mvp) mutant of einkorn wheat (Triticum monococcum), which has null alleles of VRN1, and never transits from the vegetative to the reproductive phase. These results suggest that VRN1 is upstream of FT and upregulates the FT expression under LD conditions. The overexpression of FT in a transgenic bread wheat (Triticum aestivum) caused extremely early heading with the upregulation of VRN1 and the downregulation of VRN2, a putative repressor gene of VRN1. These results suggest that in the transgenic plant, FT suppresses VRN2 expression, leading to an increase in VRN1 expression. Based on these results, we present a model for a genetic network of flowering-time genes in wheat leaves, in which VRN1 is upstream of FT with a positive feedback loop through VRN2. The mvp mutant has a null allele of VRN2, as well as of VRN1, because it was obtained from a spring einkorn wheat strain lacking VRN2. The fact that FT is not expressed in the mvp mutant supports the present model.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Floral transition, the phase transition from vegetative to reproductive development, is a critical event in the life cycle of seed-propagated plants. In Arabidopsis, the current consensus is that there are four major flowering promotion pathways: the vernalization, photoperiod, autonomous and gibberellin (GA) pathways (Boss et al., 2004). The vernalization and photoperiod pathways integrate environmental signals into the floral transition, whereas the autonomous and GA pathways act independently of external signals. The vernalization pathway mediates the promotion of floral transition, and is induced by low temperatures. FLOWERING LOCUS C (FLC), a MADS-box gene that acts as a repressor of the floral transition, is negatively regulated by vernalization (Michaels and Amasino, 1999; Sheldon et al., 1999). Vernalization results in the stable reduction of the levels of the flowering repressor FLC by epigenetic regulation (Sung and Amasino, 2006). The autonomous pathway prevents the accumulation of FLC mRNA by epigenetic regulation (He and Amasino, 2005), indicating that the vernalization and autonomous pathways are connected through the flowering repressor FLC (Jiang et al., 2008). The photoperiod pathway, which consists of photoreceptor, circadian clock and circadian clock-regulated genes, promotes the floral transition in response to a long photoperiod (Searle and Coupland, 2004). This pathway includes downstream genes of the circadian clock, such as CONSTANS (CO) and FLOWERING LOCUS T (FT).CO encodes a transcription factor with two B-box zinc fingers, and directly induces FT expression (Putterill et al., 1995). Circadian clock-regulated CO transcription is mediated by controlling factors such as GIGANTEA (GI) proteins (Imaizumi and Kay, 2006). In long-day (LD) conditions CO expression peaks during the light period, resulting in the activation of FT expression, thereby causing early flowering (Yanovsky and Kay, 2003). FT encodes a protein, similar to the animal Raf kinase inhibitor-like protein, that functions as a flowering promoter (Kardailsky et al., 1999; Kobayashi et al., 1999). FT is an integrator of the vernalization and photoperiod pathways: prior to vernalization, the activity of FT is directly suppressed by FLC (Searle et al., 2006). Recent studies indicate that the FT protein functions as a systemic signaling molecule from leaf to apex (Corbesier et al., 2007; Jaeger and Wigge, 2007; Mathieu et al., 2007). The FT protein interacts with the bZIP transcription factor FD at the apex, and activates the floral meristem identity genes APETALA1 (AP1) and LEAFY (LFY) (Abe et al., 2005; Wigge et al., 2005). The growth regulator GA promotes flowering by upregulating the transcription of SUPPRESSOR OF OVEREXPRESSION OF CO1 (SOC1), an activator of LFY (Moon et al., 2003).
In cereal crops such as wheat and barley, the heading time is associated with the timing of floral transition, and is an important character because of its influence on adaptability to various environmental conditions. Bread wheat (Triticum aestivum, 2n=6x=42, genome constitution AABBDD) is grown in a wide range of environments, all over the world, and its wide adaptability results from the variation in heading time among cultivars. Many genetic studies have been performed to clarify the genetic control of heading time in wheat, and the following three component characteristics have been identified: vernalization requirement, photoperiod sensitivity and narrow-sense earliness (earliness per se) (Worland and Snape, 2001).
Vernalization requirement is concerned with the sensitivity of the plant to cold temperatures for accelerating spike primordium formation. The vernalization insensitivity (spring habit) genes, Vrn-A1, Vrn-B1 and Vrn-D1, have been shown to be located on chromosomes 5A, 5B and 5D, respectively, of bread wheat (Worland and Snape, 2001). Bread wheat is a hexaploid species with the genome constitution AABBDD that originated from three diploid ancestral species: the A genome came from Triticum urartu, the B genome came from Aegilops speltoides, or another species classified in the Sitopsis section, and the D genome came from Aegilops tauschii (Feldman, 2001). Consequently, the hexaploid wheat genome contains triplicated homoeologous genes (homoeologs) derived from the ancestral diploid species. Vrn-A1, Vrn-B1 and Vrn-D1 are three homoeologs of the Vrn-1 gene. Yan et al. (2003) isolated Vrn-Am1, an ortholog of Vrn-A1, in diploid einkorn wheat Triticum monococcum (2n=2x=14, AmAm) using a map-based method, and named this gene VRN1. VRN1 was shown to have a high sequence similarity to AP1/FRUITFULL (FUL) of Arabidopsis. Following this study in diploid wheat, the Vrn-1 genes of hexaploid wheat were identified as WAP1 (wheat AP1) (Murai et al., 2003; Trevaskis et al., 2003) or TaVRT-1 (Triticum aestivum vegetative to reproductive transition-1; Danyluk et al., 2003). Transgenic and mutant analyses indicated that VRN1 is a flowering promoter that has an indispensable role in the floral transition pathway of wheat (Murai et al., 2003; Loukoianov et al., 2005; Shitsukawa et al., 2007a). In diploid wheat, the genes Vrn-Am1 and Vrn-Am2 are mainly responsible for the control of the vernalization response (Dubcovsky et al., 1998). Vrn-Am2 encodes a protein (ZCCT1) with the CO, CO-like and TOC1 (CCT) domain that was identified in the CO-like gene family, and named VRN2 (Yan et al., 2004). A transgenic study indicated that VRN2 is a strong repressor of flowering in wheat. As VRN2 is downregulated by vernalization, it is postulated that the VRN2 protein suppresses VRN1 expression. The reduction of the VRN2 protein through vernalization allows VRN1 transcription to increase gradually to enable flowering competence. Although large deletions within the first intron of the VRN1 gene are associated with a spring habit (Fu et al., 2005), there is currently no biochemical evidence that the VRN2 protein directly binds to the regulatory region of the VRN1 gene. TaVRT2, a MADS-box gene with sequence similarity to Arabidopsis SHORT VEGETATIVE PHASE (SVP), has been identified as another flowering repressor in wheat (Kane et al., 2005). TaVRT2 directly binds the CArG motif in the VRN1 promoter, suggesting that TaVRT2 represses the transcription of VRN1 in association with VRN2 (Kane et al., 2007). Recently, another vernalization gene, Vrn-3, located on chromosome 7B, was shown to encode a homolog of Arabidopsis FT, and was named VRN3 (Yan et al., 2006). Analyses of transgenic plants showed that VRN3 functions as a flowering promoter. These findings indicate that the genetic system that determines the vernalization response in wheat is quite different from that in Arabidopsis, a conclusion that is consistent with the absence of an FLC-like gene in the wheat genome.
The photoperiod (LD) response in wheat is determined by the dominant genes Ppd-A1, Ppd-B1 and Ppd-D1, which control sensitivity to photoperiod, and are located on chromosomes 2A, 2B and 2D, respectively (Worland and Snape, 2001). Wheat is a quantitative LD plant, and short-day (SD) conditions delay the heading time. The Ppd gene acts to reduce the delay of heading under SD conditions (Murai et al., 2003). Recently, a barley (Hordeum vulgare, 2n=2x=14, HH) ortholog of the Ppd genes, Ppd-H1, was identified as a pseudoresponse regulator (PRR) that showed most similarity to Arabidopsis PRR7 (Turner et al., 2005). In Arabidopsis, the PRR family consists of five members (PRR9, PRR7, PRR5, PRR3 and PRR1/TOC1) that are involved in circadian clock function (Mizuno and Nakamichi, 2005). This suggests that the barley Ppd gene is likely to be a circadian clock-associated gene. Comparative mapping indicated that the wheat Ppd gene is orthologous to the barley Ppd gene, a conclusion that has been supported by sequence and expression analyses (Beales et al., 2007).
Narrow-sense earliness, which corresponds to the autonomous flowering pathway in Arabidopsis, refers to earliness in the flowering of fully-vernalized plants grown under LD conditions. Several quantitative trait loci (QTLs) have been identified in barley and wheat for this characteristic (Cockram et al., 2007). GA accelerates the transition from the vegetative to reproductive phase in wheat in a similar manner as in Arabidopsis (Weibel, 1960; Pauli et al., 1962). The wheat SOC1 homolog, wheat SOC1 (WSOC1), may function in the GA pathway (Shitsukawa et al., 2007b).
In this study, we performed expression, mutant and transgenic studies to clarify the genetic network of flowering-time genes in wheat, especially the relationship of two flowering promoter genes, VRN1 (an AP1/FUL homolog) and VRN3 (an FT homolog). Expression analysis of a VRN1 deletion mutant suggested that VRN3 (FT) is upregulated by VRN1 together with CO under LD conditions. Furthermore, our transgenic analysis of VRN3 (FT) indicated that VRN3 (FT) suppresses VRN2 expression. Here, we present a model for the genetic network of flowering-time genes in wheat, in which VRN1 is upstream of VRN3 (FT), with a positive feedback loop through VRN2.
Diurnal rhythmic expression patterns of GI, CO and FT in wheat leaves
Wheat GI (TaGI1) and wheat FT (VRN3) have been identified in previous work (Zhao et al., 2005; Yan et al., 2006). In this study, we identified wheat CO (WCO1) by using the PCR method with primers based on the sequence of barley HvCO1 located on chromosome 7H (Griffiths et al., 2003). In the barley genome, nine CO homologs have been identified. On the basis of sequence similarity and chromosomal synteny, HvCO1 located on chromosome 7H may be the counterpart of rice Hd1. The mapping study showed that WCO1 was located on homoeologous group 7 (7A, 7B and 7D) (Figure S1). Wheat chromosomes 7A, 7B and 7D are syntenic with rice chromosome 6 (the location of the Hd1 gene), suggesting that WCO1 is the counterpart of Hd1. It is known that TaHd1, previously identified as a wheat CO-like gene, is located on a group-6 homoeologous chromosome (Nemoto et al., 2003), indicating that TaHd1 is not an ortholog of Hd1. A phylogenetic tree of rice Hd1 and barley CO-like genes, together with two wheat CO-like genes, TaHd1 and WCO1, using deduced amino acid sequences (see Figure 1 for the tree, and Figure S2 for amino acid alignment), reinforced the conclusion that WCO1 is the wheat ortholog of Hd1/CO. However, from their investigations with transgenic rice, Nemoto et al. (2003) reported that TaHd1 has some function in flowering. In this study, therefore, both WCO1 and TaHd1 are used as wheat CO genes.
The diurnal expression patterns of GI (TaGI1), CO (WCO1 and TaHd1) and FT (VRN3) were examined in leaves at the three-leaf stage in spring wheat cv. Triple Dirk (TD) plants with Ppd dominant alleles (Ppd-TD), grown under LD or SD conditions (Figure 2). The expression analysis was performed using real-time PCR with primers that amplify all three homoeologs located on A, B and D wheat genomes. The shoot apical meristem (SAM) of the Ppd-TD plant transits from the vegetative to reproductive phase around at the four-leaf stage when grown under LD conditions. Under SD conditions the timing of the phase transition is a little delayed (occurring at the five-leaf stage). Under LD conditions, GI mRNA accumulated during the light period, and expression peaked late in the light phase (Figure 2a). CO (WCO1 and TaHd1) mRNA accumulated during the dark period (Figure 2c), and FT mRNA accumulated from the beginning of the light phase (Figure 2e). Under SD conditions, GI mRNA expression showed a similar pattern as under LD conditions, although the peak of expression was shifted towards the end of the light period (Figure 2b). CO (WCO1 and TaHd1) expression also showed a similar pattern as under LD conditions, but the WCO1 expression was mostly confined to the dark period under SDs (Figure 2d). In contrast to GI and CO, no expression of FT was detected under SD conditions (Figure 2f). The expression analysis did not conflict with the idea that the functional hierarchy, GI → CO → FT, is conserved in wheat. Analysis of the diurnal expression pattern in barley also indicated conservation of this cascade (Turner et al., 2005).
The diurnal expression patterns of GI, CO and FT were also examined in leaves at the three-leaf stage in TD plants with ppd recessive alleles (ppd-TD), grown under LD or SD conditions (Figure 3). The SAM of the ppd-TD plant transits from the vegetative to the reproductive phase around at the four-leaf stage when grown under LD conditions, whereas the timing of the phase transition is very delayed (occurring at the six-leaf stage) under SD conditions. The expression of these genes in ppd-TD showed similar patterns as in Ppd-TD plants, except for the expression patterns of GI and CO genes under SD conditions (Figure 3). In GI, ppd-TD plants showed a dual peak expression pattern in the light period (Figure 3b). Furthermore, in both WCO1 and TaHd1, no peaks were observed late in the dark period compared with the expression patterns in Ppd-TD plants (Figure 3d). As the Ppd gene should be one of the circadian clock component genes, the results indicate that the GI and CO genes are circadian clock-regulated genes. The difference in the timing of the phase transition between Ppd-TD and ppd-TD under SD conditions may be associated with the difference of diurnal expression patterns of GI and CO genes.
VRN1 also shows a diurnal expression pattern in wheat leaves
At the three-leaf stage, spring wheat Ppd-TD plants showed a diurnal pattern of VRN1 expression under both LD (Figure 2g) and SD (Figure 2h) conditions. VRN1 expression peaked at the beginning of the light period under both LDs and SDs, suggesting that the expression of VRN1 is regulated by the light–dark cycle. Comparison of the expression patterns of VRN1 and FT under LD conditions (Figure 2e, g) showed that the accumulation of FT transcripts in the light period followed the upregulation of VRN1. This suggests that the VRN1 expression is associated with the FT expression in wheat leaves under LD conditions. Although VRN1 was expressed under SD conditions, no FT expression was detected (Figure 2f, h). In ppd-TD plants, VRN1 showed a similar expression pattern as in Ppd-TD: the expression peaked at the beginning of the light period under both LD (Figure 3g) and SD (Figure 3h) conditions. This result supports the idea that the expression of VRN1 is regulated by the light–dark cycle. As in Ppd-TD, ppd-TD plants showed that the accumulation of FT transcripts in the light period followed the upregulation of VRN1, suggesting that the VRN1 expression is associated with FT expression under LD conditions.
FT is not expressed in mvp, a deletion mutant of VRN1
The einkorn wheat mutant, maintained vegetative phase (mvp), was induced by accelerated nitrogen ion irradiation treatment, and was identified by its inability to transit from the vegetative to reproductive phase (Shitsukawa et al., 2007a). The mvp mutant cannot transit to the reproductive phase under any growth conditions (LD, SD, non-vernalized and vernalized). In the previous study, genetic analysis indicated that the mvp phenotype is controlled by one locus, and we demonstrated that the mvp mutation was caused by the deletion of the VRN1 coding and promoter regions (Shitsukawa et al., 2007a). The mvp mutant, mvp-1, used in this study was obtained from einkorn spring wheat strain KU104-2, with dominant alleles of VRN1 and null alleles of VRN2 (Figure S3). Yan et al. (2004) revealed that spring einkorn wheats are classified into three types in the VRN2 allele: wild type, R/W mutant type and deletion type. KU104-2 is a deletion type spring einkorn wheat (Figure S3). Consequently, the mvp mutant has deletion alleles of both VRN1 and VRN2, because it was derived from KU104-2. The expression patterns of VRN1, CO (WCO1) and FT were analyzed by RT-PCR in mvp, the corresponding wild type (WT) and the normal lines, using autumn-sown plants in the experimental field (Figure 4). The normal line comprised M3 plants with a normal phenotype that segregated from an mvp heterozygous M2 plant. VRN1 gene expression was observed in WT and normal lines at both vegetative and reproductive growth stages in spring to summer, but was not detected in the mvp mutants (Shitsukawa et al., 2007a). The reproductive stage is defined here as the growth stage in which the internodes of WT and normal plants elongate. At this stage, the mvp mutants are still in the vegetative phase. In contrast to VRN1, the expression of CO was detected in mvp mutants, indicating that CO acts upstream of or in a different pathway to VRN1. No expression of FT was observed in the mvp mutants. Because the mvp mutant has intact coding and promoter regions of the FT gene in the genome (Figure S4), the present observation indicates that no expression of FT resulted from the null allele of VRN1.
It has been reported that the expression of FT increased significantly in wheat leaves in response to vernalization, just as the expression of VRN1 did (Yan et al., 2006). To examine the effect of vernalization on gene expression patterns of VRN1, FT and VRN2 in the mvp mutant together with the WT, seedlings at the one-leaf stage or at the three-leaf stage were cold-treated, and gene expression levels were analyzed by real-time PCR (Figure 5). At these stages, the SAMs of both the mvp mutant and the WT plants are still in the vegetative phase. In strain KU104-2 plants at the one-leaf and three-leaf stages, VRN1 and FT were clearly upregulated by cold treatment (Figure 5a–d), but no VRN2 expression was observed because of the lack of VRN2 alleles (Figure 5e,f). KU104-2 is an early-heading type, and high expression of FT suggests that it has the dominant allele of FT. In the mvp mutant, no expression of FT as well as of VRN1 was observed at the one-leaf and three-leaf stages (Figure 5a-d), confirming the results obtained from the field-grown plants (Figure 4). We also examined normal winter einkorn wheat strain KT10-1 with recessive vrn1 and dominant VRN2. KT10-1 is a late-heading type, with no expression of FT at the one-leaf and three-leaf stages (Figure 5c,d). At the one-leaf stage of KT10-1 plants, the VRN2 expression level was high in non-vernalized plants, and was dramatically decreased in vernalized plants (Figure 5e). At the three-leaf stage, the VRN2 expression was still at a high level in the non-vernalized plants, and was downregulated by cold treatment (Figure 5f), whereas the VRN1 expression was upregulated (Figure 5b). Contrary to the spring strain KU104-2 and the mvp mutant, the winter strain KT10-1 showed a high expression level of VRN2 at the three-leaf stage (Figure 5f), suggesting that the expression of VRN2 was not affected by aging from the one-leaf to the three-leaf stage in the winter strain.
Overexpression of FT induces the downregulation of VRN2 and the upregulation of VRN1
Using spring wheat cv. Norin 61 (N61), we developed transgenic plants with VRN3-D cDNA driven by P35S. VRN3-D is a homoeolog of wheat FT (VRN3) located on the D genome. In culture in a petri dish, a transgenic T0 plant with overexpression of FT showed extremely early heading (Figure 6a), indicating that FT is a strong activator of the flowering pathway. Transgenic wheat plants were grown in a growth chamber under SDs. mRNAs were extracted from leaves of non-vernalized seedlings at the one-leaf stage, and the expression levels of FT, VRN1 and VRN2 were examined. At this stage, the SAMs of transgenic and non-transgenic (WT) wheat are still in the vegetative phase. In WT plants, the expression of FT and VRN1 were not observed, whereas a high expression of VRN2 was detected (Figure 6c). Overexpression of FT was observed in 10 T1 plants, 1, 2, 4, 5, 9, 11 and 13–16, which showed an early heading trait compared with non-transgenic plants (Figure 6b, c). To investigate whether the RT-PCR patterns of FT showed expression profiles of the VRN3-D transgene, the amplified RT-PCR products of positive transgenic lines 4 and 5 were cloned and sequenced. The sequences of RT-PCR products were identical with those of VRN3-D, and differed from the sequences of VRN3-A and VRN3-B (Figure S5), indicating that the RT-PCR analysis of FT exhibited the expression profile of the VRN3-D transgene.
The positive transgenic T1 plants showed a reduction in VRN2 mRNA levels, but WT plants and T1 null segregants (3, 6, 7, 8, 10 and 12) exhibited a high expression of VRN2 (Figure 6c). Furthermore, the T1 plants with overexpressing FT showed a high expression of VRN1. These findings suggest that the downregulation of VRN2 is associated with the upregulation of FT in the positive T1 plants, and the upregulation of VRN1 in the positive T1 plants probably resulted from the downregulation of VRN2. Real-time PCR analysis confirmed that the positive transgenic plant 1 with high expression of FT also showed a downregulation of VRN2 and an upregulation of VRN1, whereas the negative transgenic plant 3 and the WT showed a high level of VRN2 transcripts and a low level of VRN1 transcripts (Figure 6d). In the previous transgenic studies of wheat FT (VRN3) (Yan et al., 2006; Li and Dubcovsky, 2008), it has also been shown that VRN1 was upregulated in transgenic plants. However, the previous study did not examine the expression of VRN2.
By analyzing the expression patterns of flowering genes (VRN1, VRN2 and FT) in the mvp mutant with null alleles of both VRN1 and VRN2, and transgenic plants with overexpressing FT, we hypothesized that VRN1 acts upstream of FT, and upregulates FT with a positive feedback loop through VRN2 in wheat leaves. This genetic network among VRN1, VRN2 and FT is different from the two other models previously presented (Hemming et al., 2008; Li and Dubcovsky, 2008). In this study, analysis of the mvp mutant provided crucial information on the relationship between VRN1 and FT. The other two models can not explain the present results: no expression of FT was observed in the mvp mutant with null alleles of both VRN1 and VRN2.
VRN1 is upstream of FT in the wheat flowering-gene network
The present study indicated that wheat FT (VRN3) was upregulated by vernalization in a similar fashion to VRN1 (Figure 5a–d), as was also reported in a previous study (Yan et al., 2006). This fact suggests that FT as well as VRN1 is associated with the vernalization pathway in wheat leaves. What is the relationship between VRN1 and FT? Recently, we described an einkorn wheat mutant, mvp, obtained by ion-beam treatment, that does not transit from the vegetative to reproductive phase (Shitsukawa et al., 2007a). The mvp mutant was caused by the deletion of VRN1. This clearly demonstrated that VRN1 is an indispensable gene for phase transition in wheat. No expression of FT was observed in the mvp mutant (Figures 4 and 5c,d), suggesting that expression of FT is related to the expression of VRN1. In this study, we also demonstrated that VRN1 shows a diurnal expression pattern under both LD (Figures 2g and 3g) and SD (Figures 2h and 3h) conditions. Under LD, an accumulation of FT transcripts follows the upregulation of VRN1 (Figures 2e,g and 3e,g). Based on these findings, we hypothesize that VRN1 acts upstream of FT, and activates FT expression under LD conditions. This putative genetic interaction would not be completely unexpected, because in Arabidopsis a MADS-box protein, FLC, interacts directly with FT sequences, and regulates FT expression as part of a protein complex (Helliwell et al., 2006). Under SD conditions, VRN1 is expressed in a similar pattern as under LDs (Figure 2h and 3h), but FT is not expressed (Figures 2f and 3f). This suggests that the upregulation of FT by VRN1 involves other factor(s).
In Arabidopsis, it was demonstrated that FT is expressed in the vascular tissue of leaves, and that the FT protein moves from the leaves to the SAMs as a long-distance signal for flowering induction (Corbesier et al., 2007; Jaeger and Wigge, 2007; Mathieu et al., 2007). The FT protein has also been shown to function as a florigen in rice (Tamaki et al., 2007) and cucurbits (Lin et al., 2007). Although it is not known if the FT protein has a florigenic effect in wheat, our transgenic study demonstrated that FT is a strong activator of flowering in this species (Figure 6a–c). This conclusion is supported by the results of another transgenic study (Yan et al., 2006). Arabidopsis shows an upregulation of FT under LDs, and a downregulation under SDs, in which the CO protein induces the expression of FT in a light-dependent manner at dusk (Yanovsky and Kay, 2003). Recently, the abundance of CO protein is regulated by the ubiquitin-mediated proteolysis in darkness (Jang et al., 2008; Liu et al., 2008). Based on these studies, we can postulate that functionally stable CO proteins (WCO1 and/or TaHd1) could be candidates for the co-factors in wheat flowering.
The FT protein interacts with the bZIP transcription factor FD at the SAM, and activates the floral meristem identity gene AP1 in Arabidopsis (Abe et al., 2005; Wigge et al., 2005). Recently, a wheat ortholog of FD (TaFDL2) was identified (Li and Dubcovsky, 2008). TaFDL2 is expressed in wheat leaves together with VRN1, and TaFDL2 protein can interact with FT protein and binds in vitro with the promoter region of VRN1. Based on these results, a hypothesis that VRN1 is upregulated by FT in wheat leaves was presented (Li and Dubcovsky, 2008). However, TaFDL2 is expressed in SAMs as well as in leaves (Li and Dubcovsky, 2008), and VRN1 is also expressed in SAMs (Preston and Kellogg, 2008). Therefore, it is possible that the FT-TaFDL2-VRN1 complex functions in SAMs rather than in leaves, as in Arabidopsis. The expression analysis supported the idea that VRN1 has discrete roles of flowering competency in leaves and of floral meristem identity in SAMs (Preston and Kellogg, 2008).
VRN1 and FT are upregulated by cold, regardless of VRN2
Based on the reciprocal expression patterns of VRN2 and VRN1 when plants are vernalized, and the knowledge of the epistatic nature that VRN2 has on VRN1, a model of the vernalization pathway has been proposed (Yan et al., 2003, 2004, 2006). In this model, VRN2 encodes a repressor of VRN1, and, as the vernalization process reduces the abundance of the VRN2 gene products, VRN1 transcription gradually increases, leading to competency in flowering. In this study, we demonstrated that VRN1 is upregulated by cold in the spring einkorn wheat strain KU104-2 that has null alleles of VRN2 (Figure 5a,b). This clearly indicates that the expression of VRN1 is regulated by cold, regardless of VRN2. This is also supported by the observation that in barley a cold signal induces VRN1 expression, but not through the repression of VRN2 (Trevaskis et al., 2006, 2007; Hemming et al., 2008). This idea is based on the finding that VRN2 expression is not affected by vernalization under SD conditions (Trevaskis et al., 2006). Furthermore, in barley doubled haploid (DH) lines that lack the VRN2 locus, vernalization can still induce VRN1 and accelerate flowering (Hemming et al., 2008). These findings suggest an additional pathway in which VRN1 is upregulated by vernalization independently of VRN2.
To explain the relationship between VRN1 and FT, two different models were previously presented. In the first model, both FT and VRN1 are suppressed by VRN2. Vernalization downregulates VRN2, leading to the upregulation of both FT and VRN1 (Yan et al., 2006). In this model, a long photoperiod induces FT expression, and then FT upregulates VRN1. However, this model cannot explain why KU104-2 lacking VRN2 shows an upregulation of FT and VRN1 by vernalization (Figure 5a–d). Furthermore, this model cannot explain why the mvp mutant lacking both VRN1 and VRN2 shows no expression of FT (Figure 5c,d). The second model indicated that FT is suppressed by VRN2. Vernalization upregulates VRN1 and VRN1 downregulates VRN2, leading to the upregulation of FT (Hemming et al., 2008). In this model, a long photoperiod induces FT expression. This model can explain why the cold signal upregulates both VRN1 and FT, but does not explain why KU104-2 lacking VRN2 shows an upregulation of FT by vernalization (Figure 5c,d), nor why the mvp mutant lacking both VRN1 and VRN2 shows no expression of FT (Figure 5c,d).
FT, a flowering activator, suppresses the expression of VRN2
In agreement with Yan et al. (2006), we found that transgenic plants overexpressing FT headed early (Figure 6a–c), indicating that FT is an activator of the flowering pathway in wheat. The early-heading transgenic plants with high FT expression showed an upregulation of VRN1 and a downregulation of VRN2, compared with control non-transformants and null segregants (Figures 6c,d). Assuming that the upregulation of VRN1 resulted from the downregulation of VRN2 (Yan et al., 2003, 2004), our results suggest that VRN2 is downregulated by FT in the transgenic plants.
An alternative model in which VRN2 downregulates FT was presented in barley, based on the results showing that transgenic barley plants overexpressing VRN2 showed no expression of FT (Hemming et al., 2008). However, this observation does not conflict with our model, because it is possible that the overexpression of VRN2 induces the downregulation of VRN1, and that the low level of VRN1 expression leads to a decrease in the expression of FT. In the previous study of VRN2 transgenic barley, Hemming et al. (2008) reported that there was no difference of VRN1 expression level between transgenic plants and null segregants. However, the previous study took no account of the diurnal expression pattern of VRN1, and so the expression level of VRN1 could not be accurate. More precise work is necessary to investigate the alternative model.
Feedback regulatory loop models have been developed based on studies using an F2 population of diploid wheat (Loukoianov et al., 2005), or a DH population of barley (Trevaskis et al., 2006), segregating VRN1 and VRN2 alleles. In these models, VRN2 acts as a repressor of flowering, and the expression of VRN2 is directly or indirectly repressed after the initiation of VRN1 expression. Our model proposes that VRN2 expression is controlled by VRN1 through FT, because FT should be upregulated by VRN1, and VRN2 should be downregulated by FT.
The VRN1–FT–VRN2 triangle model for flowering in wheat
According to the present results, one possible model for the regulation of the floral transition in wheat is shown in Figure 7. This model includes the assumption that VRN2 suppresses the expression of VRN1, and that a GI → CO →FT cascade functions in the photoperiod pathway in wheat flowering. Furthermore, Ppd is supposed to function in the circadian clock system, and upstream of GI and CO. Our expression and mutant analyses suggested that VRN1 acts upstream of FT, and possibly acts with CO (WCO1 and/or TaHd1) to activate FT expression under LD conditions. Vernalization downregulates VRN2 and upregulates VRN1, independently of each other. Furthermore, the transgenic study suggested that VRN2 is downregulated by FT. Based on these results, a feedback triangle of VRN1–FT–VRN2 is suggested to be involved in wheat flowering. Our model can explain why VRN1 represses VRN2 expression (Loukoianov et al., 2005; Trevaskis et al., 2006). However, our model does not describe why VRN2 is downregulated by a short photoperiod (Dubcovsky et al., 2006; Trevaskis et al., 2006). This might be caused other photoperiod pathway functions under SD conditions. Note that the gene interactions shown in Figure 7 are events that occur in leaves. As in Arabidopsis and rice, FT proteins could be the florigen that moves from the leaves into the SAMs to determine floral meristem identity in wheat.
The Arabidopsis genome contains three AP1-like genes, AP1, CAULIFLOWER (CAL) and FUL, all of which act redundantly at the SAM (Ferrandiz et al., 2000). The grass family, including wheat, has two paralogues of AP1/FUL in the genome, FUL1/VRN1 and FUL2, both derived from the FUL lineage (Preston and Kellogg, 2007). Expression analyses indicated that both VRN1 and FUL2 are expressed in SAMs as well as in leaves (Preston and Kellogg, 2008). The upregulation of VRN1 is significantly later in SAMs than in the leaves of vernalized plants, suggesting that VRN1 may perform discrete roles in leaves and in SAMs. This idea is highly consistent with our model in which VRN1 functions in leaves to obtain flowering competency, and acts in SAMs to facilitate the transition to flowering. It has been reported that the expression of FUL2 increased following the attainment of flowering competency (Preston and Kellogg, 2008), suggesting that FUL2 functions redundantly in SAMs for the transition to flowering, together with VRN1.
In diploid wheat, an epistatic interaction between VRN1 and VRN2 has been demonstrated (Tranquilli and Dubcovsky, 2000). In contrast to VRN1, which is a dominant allele for spring growth habit, the VRN2 allele is dominant for winter growth habit. The effect of VRN1 on heading time is significant only when the dominant VRN2 allele is present, that is, VRN2 is epistatic to VRN1. Our model can explain the epistatic relationship between VRN1 and VRN2. In plants with vrn2 recessive alleles, high expression of VRN1 induces FT regardless of which VRN1 alleles are present. Thus, in the absence of VRN2 activity, the LD induction of FT causes early flowering, which overcomes the requirement for vernalization (spring habit). This is in good agreement with the observation that FT is expressed at higher levels in einkorn wheat lines that lack VRN2 (Yan et al., 2006; Figure 5c,d in this study). In plants with VRN2 dominant alleles, VRN1 expression is high in the presence of dominant VRN1 alleles because VRN2 cannot suppress the expression of VRN1. The expression of VRN2 is then downregulated by FT, inducing flowering without vernalization (spring habit). In the presence of recessive vrn1 alleles, VRN1 expression is suppressed by VRN2 under non-vernalizing conditions. Vernalization decreases VRN2 expression and induces VRN1 expression, thereby inducing flowering (winter habit). Once VRN1 is upregulated, FT is induced and VRN2 is suppressed.
We conclude that FT is the principal factor that integrates the CO-related photoperiod and VRN1-related vernalization pathways in wheat leaves, leading to the competence of flowering in wheat plants. As flowering is a very complicated characteristic in wheat, as well as in Arabidopsis, the present model of wheat flowering will undoubtedly need to be modified in the future following the cloning and analysis of novel genes.
Bread wheat (T. aestivum, 2n=6x=42, genome constitution AABBDD) cv. Chinese Spring (CS) was used for the cDNA cloning for WCO1. CS is a spring wheat cultivar, and is known to carry the vernalization-insensitive (spring habit) gene, Vrn-D1. Spring wheat cv. Triple Dirk (TD) with Ppd dominant alleles (Ppd-TD) or ppd recessive alleles (ppd-TD) were used for the diurnal expression studies in a growth chamber. Ppd-TD is not completely photoperiod-insensitive, and SD conditions delay the heading time compared with LD conditions (Murai et al., 2003). ppd-TD shows early heading as much as Ppd-TD under LD conditions, but shows extremely late heading under SD conditions (Murai et al., 2003). Both Ppd-TD and ppd-TD carry the vernalization-insensitive (spring habit) genes Vrn-A1 and Vrn-B1. Spring wheat cv. N61, which carries Vrn-D1, was used for the transgenic study.
The einkorn wheat (T. monococcum, 2n=2x=14, AmAm) mutant, mvp, induced by nitrogen ion-beam treatment (Shitsukawa et al., 2007a), was used for expression analyses. The mvp mutant was identified by its inability to transit from the vegetative to reproductive phase under any environmental conditions. The mvp mutation was caused by deletion of the VRN1 promoter and coding regions. The mvp mutant, mvp-1, used in this study was obtained from spring einkorn wheat strain KU104-2 with dominant alleles of VRN1 and null alleles of VRN2. Spring einkorn wheats are classified into three types in the VRN2 allele: WT, R/W mutant type and deletion type (Yan et al., 2004). KU104-2 is a deletion type spring einkorn wheat (Figure S3). Consequently, the mvp mutant has deletion alleles of both VRN1 and VRN2, because it was derived from KU104-2.The mvp mutant and original WT strain, together with winter einkorn wheat strain KT10-1, with recessive vrn1 and dominant VRN2, were used in the expression study.
For the diurnal expression study, non-vernalized plants were grown under LD (16-h light/8-h dark) or SD (10-h light/14-h dark) conditions at 20°C (100 μE m−2 s−1). The mutant study was performed using autumn-sown plants in the experimental field. For the vernalization study of the mutant, seedlings at the one- or three-leaf stage that had been grown in a growth chamber under LD conditions were transferred into a cold chamber at 4°C (20 μE m−2 s−1) for 0 or 5 weeks, and were then used for expression analysis. Transgenic plants were grown in a growth chamber under SD conditions at 20°C (100 μE m−2 s−1) for 5 weeks.
Cloning and phylogenetic analysis of WCO1
The full-length cDNA sequence of WCO1 was amplified with PCR primers designed using the sequence of barley HvCO1 (Griffiths et al., 2003): WCO1-full-3L (5′-TGCATGGTCTTTGTGGTG-3′) and WCO1-full-3R (5′-ATCCAACCATTATTCAGAGCAT-3′). Sequence data of WCO1 are found in the DDBJ/EMBL/GenBank data libraries under the accession number AB361064. The chromosomal location of WCO1 was determined by PCR analysis using nulli-tetrasomic CS lines with homoeolog-specific primers (Figure S1). A phylogenetic analysis was conducted using the neighbor-joining method with the deduced amino acid sequence (Figure S2). The accession numbers of other genes in the phylogenetic tree are as follows: Hd1 (AB041838); TaHd1 (AB094490); HvCO1 (AF490468); HvCO2 (AF490469); HvCO3 (AF490471); HvCO4 (AF490475); HvCO5 (AY082958); HvCO6 (AY082961); HvCO7 (AY082963); HvCO8 (AY082964); and HvCO9 (AY082965).
Sequence analysis of VRN3 in mvp
PCR analyses of genomic DNA from WT (KU104-2) einkorn wheat and mvp plants were performed with PCR primers designed using the sequence of VRN3 (DQ890162): for the promoter region, CArG1-Fw (5′-GCTTTTTTCCTAATACGGCCCGCGTC-3′) and CArG3-Re (5′-CACTTTATATAGGGCCGAAAAG-3′); for the coding region, CArG3-Fw (5′-CTTTTCGGCCCTATATAAAGTG-3′) and CArG4-Re (5′-GTGTTGCCAATATATAGGTAATGC-3′). The amplified PCR products were sequenced (Figure S4).
Total RNAs were extracted from leaves using ISOGEN (Nippon Gene, http://www.nippongene.com), and cDNAs were synthesized from the total RNAs with oligo dT primer in accordance with the protocol for the Ready-To-Go T-Primed First-Strand Kit (GE Healthcare, http://www.gehealthcare.com). Leaves were sampled from the plants 2 h after the start of the light period. The RT-PCR analysis of einkorn wheat was performed using gene-specific primer sets for VRN1 (VRN1-BAC 81655L and VRN1-BAC 82017R), WCO1 (WCO1-L and WCO1-R) and FT (WFT-all-3L and WFT-all-3R). The Ubiquitin gene was used as the control with the primer set Ubi-1L and Ubi-1R. For the expression analysis of transgenic plants, total RNA was extracted from the first leaf of each plant using a Get Pure RNA Kit (Dojindo Laboratories, http://www.dojindo.com). Leaves were sampled from the plants 2 h after the start of the light period. RT-PCR was performed using the RNA PCR Kit (AMV) Version 2.1 (Takara Bio, http://www.takara-bio.com) with AmpliTaq Gold DNA Polymerase (Applied Biosystems, http://www.appliedbiosystems.com). Gene-specific primer sets were used for FT (WFT-FW and WFT-RV), VRN1 (WAP1-556L and WAP1-982R) and VRN2 (VRN2-FW2 and VRN2-RV2). Wheat actin gene (Actin) was used as the control with the primer set ACT-FW and ACT-RV. The sequences of the primer sets and the annealing temperatures are shown in Table S1. In all experiments, the RT-PCR analyses were performed in the exponential range of amplification. To determine the sequence of RT-PCR products of VRN3 amplified from transgenic plants, RT-PCR products from positive transgenic lines 4 and 5 were directly sequenced (Figure S5).
Real-time PCR analysis
In the diurnal expression study and mutant analysis, real-time PCR analyses were performed using a LightCycler 2.0 (Roche Diagnostics, http://www.roche.com/diagnostics) with gene-specific primer sets for GI (TaGI-3L and TaGI-3R), TaHd1 (TaHd1-2L and TaHd1-2R), FT (WFT-F4 and WFT-R4), VRN1 (WAP1-545L and WAP1-698R) and VRN2 (ZCCT1-1Lt and ZCCT1-1Rt). The PCR primers and annealing temperatures for WCO1 were the same as in the RT-PCR study. The quantity of transcripts was determined by the SYBR Green fluorescence of Actin using the primer set actin361-L and actin361-R. The Ubi-1 gene was also used as the internal control using the same primer set as in the RT-PCR study. In the transgenic study, real-time PCR was carried out using the Stratagene MX3000 Real-Time PCR system (Stratagene, http://www.stratagene.com/). The template cDNAs were amplified with Brilliant SYBR Green QPCR Master Mix using the same primer sets for WFT, WAP1 and VRN2 as in the RT-PCR experiment, at and annealing temperature of 60°C. The data were analyzed using Stratagene MXPRO ver. 3.0 software (Stratagene). The sequences of the primer sets and the annealing temperatures are shown in Table S1.
Generation of transgenic wheat lines
Transgenic wheat plants were produced by a particle bombardment method using immature embryos. The expression plasmid 35S:VRN3-D was constructed and co-transformed with the plasmid pUBA (Toki et al., 1992), which contains the bar gene as a selection marker under the control of the maize ubiquitin promoter, according to the method of Pellegrineschi et al. (2002). The integration of the transgene was confirmed by PCR of genomic DNA using primers designed from the 35S promoter sequence (35S-WFT-FW: 5′-CGCACAATCCCACTATCCTT-3′), and from the WFT cDNA sequence (35S-WFT-RV: 5′-GAAGAGCACGAGCACGAAG-3′). One transgenic plant, designated WF3, was selected from six independent transgenic T0 plants. Using 16 T1 plants derived from the WF3 T0 plant and two non-transgenic plants, RT-PCR experiments were carried out to examine the transcription levels of FT, VRN1 and VRN2.
We thank Jorge Dubcovsky, Ben Trevaskis, Shigeo Takumi, Yasunari Ogihara, Kanako Kawaura and Ryo Ishikawa for their valuable suggestions, and Tsutomu Kawasaki and Seiichi Toki for the gift of plasmids for wheat transformation. We are grateful to the National Bioresource Project – Wheat (NBRP-KOMUGI) for providing wheat materials and the expressed sequence tag (EST) database. This work was supported by a Grant-in-Aid from the Ministry of Agriculture, Forestry and Fisheries of Japan (Green Technology Project GD-3005, Comparative Genomics for Understanding the Diversity of Cereal Crops), to KM and HH, and from the Fukui Prefectural Government, to KM.
Sequence data from this article can be found in the DDBJ/EMBL/GenBank data libraries under the accession number AB361064 for WCO1-D.