Mutation in TERMINAL FLOWER1 Reverses the Photoperiodic Requirement for Flowering in the Wild Strawberry Fragaria vesca^1,[W]

Mutation in the Floral Repressor FvTFL1 Reverses the Photoperiodic Requirement for Flower Induction

SFL, which causes seasonal flowering in F. vesca (Brown and Wareing, 1965; Albani et al., 2004), was initially mapped to F. vesca linkage group 6 (Sargent et al., 2004). In a recent study, association was found between perpetual flowering habit and a 2-bp deletion in the first exon of F. vesca TFL1 homolog, and the gene was roughly mapped between two markers with a physical distance of approximately 11 Mb in the same linkage group (Iwata et al., 2012). We narrowed down the location of SFL within the mapping window of 248 kb, and our extensive linkage analysis in two populations fully supports the association found by Iwata et al. (2012; Supplemental Figs. S1 and S2).

To explore the function of FvTFL1, we overexpressed the SD F. vesca (PI551792 [National Clonal Germplasm Repository]; abbreviated to VES in the figures) FvTFL1 and the allele with a 2-bp deletion under the control of the cauliflower mosaic virus P35S promoter in the LD accession Hawaii-4 (PI551572; abbreviated to H4 in the figures). Primary transgenic lines overexpressing the wild-type FvTFL1 allele from SD F. vesca (P35S:FvTFL1) remained vegetative for at least 10 months under inductive LD conditions, whereas plants overexpressing the mutated FvTFL1 (P35S:FvmTFL1) flowered continuously (Fig. 1, A–C; Supplemental Fig. S3). We further analyzed the flowering time by growing the plants propagated from single-leafed runner cuttings under LD. Both nontransgenic Hawaii-4 and P35S:FvmTFL1 lines flowered early, whereas P35S:FvTFL1 lines and SD F. vesca were still vegetative after about 10 weeks (Fig. 1D). These data indicate that FvTFL1 represses flowering under LD and that the 2-bp deletion in FvTFL1 leads to a nonfunctional FvTFL1 protein.

In order to further test the hypothesis that FvTFL1 is SFL, we produced transgenic lines in the SD F. vesca background and analyzed their flowering phenotypes in LD or after a strong flower induction treatment (4 weeks of SD at 11°C). FvTFL1 RNA interference (RNAi) silencing lines flowered almost at the same time in both treatments (Fig. 2, A and B). In contrast, nontransgenic control plants remained vegetative in LD and required SD induction treatment for flowering, whereas P35S:FvTFL1 lines stayed vegetative in both treatments (Fig. 2, A and B; Supplemental Fig. S4). In line with the role of Arabidopsis TFL1 in the transcriptional repression of floral meristem identity genes (Liljegren et al., 1999; Ratcliffe et al., 1999; Hanano and Goto, 2011), the expression of putative F. vesca AP1/FUL homologs correlated negatively with the expression of functional FvTFL1 in the transgenic F. vesca lines (Figs. 1, E–H, and 2, C and D). In conclusion, our results demonstrate that FvTFL1 is the major floral repressor SFL, which prevents the activation of floral meristem identity genes and flowering under LD conditions in SD F. vesca. Moreover, the absence of functional FvTFL1 reverses the photoperiodic requirements for flower induction.

We also expressed a P35S:FvTFL1 overexpression construct in the Arabidopsis tfl1-2 mutant, which flowers early especially under SD and produces a determinate inflorescence with a few flowers (Shannon and Meeks-Wagner, 1991; Supplemental Fig. S5B). We analyzed three independent P35S:FvTFL1 lines in the tfl1-2 mutant background and showed that FvTFL1 fully complemented the inflorescence defects and the early-flowering phenotype of the mutant in both SDs and LDs (Supplemental Fig. S5), indicating that FvTFL1 is an ortholog of TFL1.

FvTFL1 Is Not Involved in the Photoperiodic Control of Vegetative Development

Under LDs, strawberry vegetative development typically involves the formation of a single leaf rosette and the differentiation of axillary buds to stolons (also called runners), whereas flower inductive conditions inhibit the emergence of stolons and activate the production of branch crowns (i.e. axillary leaf rosettes; Hytönen et al., 2004, 2009; Heide and Sønsteby, 2007). We analyzed the role of FvTFL1 in the control of vegetative development by using transgenic lines in a SD F. vesca background. Similar to wild-type plants, both FvTFL1 overexpression and RNAi lines continuously produced runners under LD, whereas 3 weeks of SD was enough to stop stolon formation in all lines (Fig. 3, A and B). After the cessation of stolon production in SD, the number of branch crowns started to increase in both wild-type plants and transgenic lines, whereas under LD no branch crowns were formed (Fig. 3, C and D). These results suggest that although the photoperiodic control of vegetative and floral development is tightly connected in strawberries (for review, see Hytönen and Elomaa, 2011), FvTFL1 does not affect vegetative development and is only involved in the photoperiodic control of floral initiation.

Photoperiodic Repression of FvTFL1 Expression Correlates with Floral Initiation

In Arabidopsis, TFL1 is weakly expressed in the lower part of the vegetative apical meristem, but its expression is activated in the inflorescence meristem after flower induction (Bradley et al., 1997; Ratcliffe et al., 1999). We analyzed the relative expression of FvTFL1 normalized against the stable control gene FvMSI1 MULTICOPY SUPPRESSOR OF IRA1 (Mouhu et al., 2009). FvTFL1 mRNA was detected in all tissues tested (Supplemental Fig. S6A). Similar to TFL1, it was highly expressed in the shoot apex compared with leaves in SD F. vesca grown under LDs (Fig. 4A), but a more detailed analysis revealed a divergent expression pattern. According to in situ hybridization, under LDs FvTFL1 mRNA was localized in the whole apical meristem, young leaf initials, and vascular tissues (Fig. 4, B and C). Furthermore, strong photoperiodic regulation of FvTFL1 was detected in the shoot apices but not in the leaves of SD F. vesca seedlings (Fig. 4D; Supplemental Fig. S6B). Under LDs, FvTFL1 was highly expressed in the shoot apex, whereas only weak expression was found in plants grown under SDs for 3 weeks (Fig. 4D). However, floral meristem identity genes were not yet activated in SD-grown SD F. vesca at this time (Supplemental Fig. S7), indicating that the plants were still vegetative. Because several Norwegian F. vesca accessions have been shown to require more than 1 month of SD treatment for flower induction (Heide and Sønsteby, 2007), we grew our SD F. vesca plants in SDs for 45 d followed by flowering-time analysis under LDs. We observed flower buds about 1 month after the SD treatment in all plants of SD F. vesca, whereas all LD-grown plants remained vegetative (Fig. 4E).

In contrast to SD F. vesca, LD-grown Hawaii-4 seedlings flowered rapidly; seedlings raised under SDs for 45 d flowered late, and the plants grown continuously under SDs remained vegetative for several months (Fig. 4, E and F), indicating that Hawaii-4 is a LD plant. In contrast to seedlings, plants propagated from the stolons of LD-grown mother plants were day neutral and flowered continuously under both LDs and SDs (Supplemental Fig. S8). In Hawaii-4, FvTFL1 mRNA expression was low in both photoperiods (Fig. 4D), most likely because of nonsense-mediated mRNA decay, a mechanism that prevents the accumulation of mRNA containing premature stop codons (Conti and Izaurralde, 2005). In association with flowering, the putative floral meristem identity genes FvAP1 and FvFUL1 were up-regulated in Hawaii-4 seedlings grown under LDs (Supplemental Fig. S6). Taken together, our data indicate that the SD requirement for flower induction in SD F. vesca is based on the photoperiodic regulation of FvTFL1, and the lack of FvTFL1 function leads to the rapid activation of flowering in LD-grown Hawaii-4.

Regulation of the Perennial Growth Cycle by FvTFL1

We further analyzed the expression patterns of FvTFL1 as well as FvAP1 and FvFUL1 in SD F. vesca plants grown under SDs followed by LDs and observed the flowering cycle of the plants. Our data showed that FvTFL1 was gradually down-regulated in the apices of the primary shoots grown under SDs, with a concomitant up-regulation of floral meristem identity genes 4 to 6 weeks after the beginning of the SD treatment (Fig. 5, A–C). When SD-grown plants were returned to LDs after 45 SDs, high FvTFL1 and low FvAP1 and FvFUL1 mRNA levels were again detected in the apices of axillary shoots that emerged after the end of the SD treatment. The observed gene expression patterns were clearly related to flowering, which started about 5 weeks after the end of the SD treatment and continued for about 1 month before declining again (Fig. 5D). In contrast to SD F. vesca, Hawaii-4 continuously produced new inflorescences only under LDs (Fig. 5D). The observed expression pattern of FvTFL1 in SD F. vesca indicates that the photoperiodic regulation of FvTFL1 mRNA expression is crucial for cycling between vegetative and reproductive phases in SD F. vesca.

FvFT1 Promotes Flowering under LDs in the F. vesca LD Accession

Functional analysis of FT in several plant species indicates its role as a florigen, a general floral activator (Turck et al., 2008; Turnbull, 2011). To establish the role of FT relative to that of SFL, we analyzed the expression patterns of F. vesca FT-like genes. As demonstrated in Figure 6, A and B, FvFT1 and FvFT2 showed contrasting spatial expression patterns in SD F. vesca. FvFT1 had the highest expression level in old leaves, and no expression or weak expression was observed in other tissues, whereas FvFT2 was expressed almost exclusively in flower buds. The tissue-specific expression pattern of FvFT1 follows the consensus that FT is a mobile signal originating in photosynthetically active leaves (Corbesier et al., 2007; Tamaki et al., 2007). Furthermore, genomic synteny is conserved only around Arabidopsis FT and FvFT1 (Shulaev et al., 2011), suggesting that FvFT1 is a likely ortholog of FT. The comparison of FvFT1 expression levels in seedlings of SD F. vesca and Hawaii-4 in different photoperiods revealed that FvFT1 was expressed under LD in both genotypes, whereas no FvFT1 expression was detected in SD-grown seedlings (Fig. 6C). The finding that FvFT1 mRNA levels did not correlate positively with flowering in the SD F. vesca contrasted with the situation in other SD plants (Hayama et al., 2003, 2007; Kong et al., 2010) and prompted us to perform diurnal rhythm and time-course analyses in older plants of SD F. vesca. As shown in Figure 6D, FvFT1 showed a clear LD-specific expression peak during the night. Moreover, FvFT1 was strongly down-regulated before flower induction under short photoperiods (Fig. 6E). Thus, the activation of FvFT1 mRNA expression correlated positively with flower induction only in the LD accession of F. vesca.

To explore the role of FT in the LD flowering response, we produced FvFT1 RNAi silencing (P35S:FvFT1-RNAi) plants in the perpetual flowering LD accession Hawaii-4. We found down-regulation of floral meristem identity genes in the shoot apex of three independent transgenic lines grown under LD (Fig. 7, A–C; Supplemental Fig. S7). Furthermore, all of these RNAi lines were clearly late flowering under LD (Fig. 7D). These results suggest that in LD-grown Hawaii-4, FvFT1 is required for the normal up-regulation of floral meristem identity genes, which marks the beginning of floral initiation at the apex (Mandel and Yanofsky, 1995; Ferrándiz et al., 2000; Mouhu et al., 2009).

TFL1 Homologs Are Major Floral Repressors in Rosaceae

A previous study by Iwata et al. (2012) and our data show that perpetual flowering is associated with a 2-bp deletion in F. vesca TFL1 homolog (Supplemental Figs. S1 and S2). We conducted functional analysis of FvTFL1 by overexpression and RNAi approaches. FvTFL1 strongly represses flowering in F. vesca, because the introduction of a P35S:FvTFL1 construct containing functional FvTFL1 into a perpetual flowering LD accession prevented the activation of putative floral meristem identity genes (FvAP1/FUL) and flower initiation under LD. Furthermore, overexpression of the mutated version of the gene did not change the flowering phenotype, indicating that it does not encode a functional protein. As constitutive overexpression of TFL1 homologs from different species has been shown to cause pleiotropic effects (Böhlenius et al., 2006; Imamura et al., 2011), we used RNAi silencing of FvTFL1 in SD F. vesca to provide further evidence that FvTFL1 is SFL. Silencing of FvTFL1 in SD F. vesca removed the SD requirement for flower induction and changed the plants to LD flowering. As TFL1 homologs prevent LD flowering and cause the seasonal flowering habit in F. vesca (Figs. 1 and 2), control the length of the juvenile phase in apple (Kotoda et al., 2006), and may also cause seasonal flowering in roses (Iwata et al., 2012), they are possibly major floral repressors and regulators of the perennial growth cycle in the Rosaceae family.

Earlier studies have shown that photoperiod acts in an opposite manner to control flower initiation and vegetative reproduction through stolons (Hytönen et al., 2004; Heide and Sønsteby, 2007). The finding that neither overexpression nor silencing of FvTFL1 affected the photoperiodic control of stolon/branch crown formation from axillary buds showed that FvTFL1 is not directly involved in the regulation of vegetative development of the strawberry shoot. This is consistent with the genetic evidence that two separate single loci, SFL and RUNNERING LOCUS, respectively, control flowering habit and the formation of stolons in F. vesca (Brown and Wareing, 1965). However, FvTFL1 may indirectly affect vegetative growth, because the development of a terminal inflorescence promotes the outgrowth of the uppermost axillary buds as branch crowns by reducing apical dominance (Arney, 1953).

Regulation of FvTFL1 mRNA Expression Controls Photoperiodic Flowering and Seasonal Growth Cycles

In Arabidopsis, TFL1 is developmentally regulated (Bradley et al., 1997; Ratcliffe et al., 1999), and to our knowledge, no photoperiodic control of TFL1 homologs has been previously reported. Our results suggest that the photoperiodic control of FvTFL1 mRNA level in the shoot apex is a primary mechanism to control photoperiodic flowering in SD F. vesca. In SD-grown plants, FvTFL1 was strongly down-regulated before the activation of floral meristem identity genes and flower initiation. The functionality of transcriptional regulation was supported by the phenotype of FvTFL1 RNAi silencing lines in the SD F. vesca background. In two independent lines, the reduction of FvTFL1 mRNA levels was enough to induce flowering under noninductive LD conditions. In contrast, none of the P35S:FvTFL1 lines flowered under flower-inductive conditions in either SD or LD F. vesca backgrounds, indicating that plants cannot overcome constitutive overexpression of FvTFL1 mRNA. In conclusion, we have identified FvTFL1 as a component of the perennial photoperiodic pathway in F. vesca. We propose that down-regulation of FvTFL1 mRNA expression in the shoot apex of SD F. vesca under SD is required for flower induction and the activation of FvAP1/FUL genes. In contrast, the lack of functional FvTFL1 in LD accessions reverses the photoperiodic requirement for flowering.

The finding that nonfunctional FvTFL1 causes perpetual flowering of LD accessions suggests that in the SD F. vesca, the regulation of FvTFL1 contributes to the cycling of vegetative and reproductive phases, and our gene expression results support this hypothesis. Whereas FvTFL1 was down-regulated and FvAP1/FUL was up-regulated in the apex of the primary shoot under SD, subsequent LD conditions restored high FvTFL1 and low FvAP1/FUL mRNA levels in the apices of new vegetative axillary shoots. Earlier studies have shown that strawberries are able to initiate flowers at the apex of the primary shoot and axillary shoots, which have reached competence to flower (Arney, 1953; Hytönen et al., 2004; Hytönen and Elomaa, 2011). Therefore, we hypothesize that the down-regulation of FvTFL1 under SD in the autumn allows flower induction to occur in the apex of the primary shoot and older axillary shoots. However, its up-regulation in the LDs of spring is crucial for perennialism, because it prevents further floral initiation in the newly emerged axillary shoots until SD conditions return. Our results on FvTFL1 and earlier findings that seasonal changes in the expression of two Populus FT-like genes regulate the perennial growth cycle in Populus (Hsu et al., 2011) highlight the importance of CETS proteins in the control of perennialism. Also, in A. alpina, TFL1 contributes to perennialism by regulating the age-dependent sensitivity of meristems to vernalization; but transient silencing of the FLC ortholog, PEP1, by vernalization has a major role in the cycling between vegetative and generative phases in this species (Wang et al., 2009, 2011).

In contrast to FvTFL1, which is broadly expressed in the shoot apex (Fig. 4B), Arabidopsis TFL1 is weakly expressed in the axillary meristems and lower part of the apical meristem during the vegetative phase and is activated in the inflorescence meristem (Bradley et al., 1997; Ratcliffe et al., 1999). This intriguing difference is probably associated with differences in inflorescence development. In Arabidopsis, TFL1 is activated in the inflorescence meristem in order to allow indeterminate growth of the raceme (Bradley et al., 1997; Ratcliffe et al., 1999), whereas the production of a cymose inflorescence with a terminal flower in strawberry (Jahn and Dana, 1970) would be expected to require the silencing of FvTFL1 in the apical meristem.

FvTFL1 Overcomes the Function of the LD-Specific Floral Activator FvFT1

The CO/FT module controls photoperiodic flowering in both annual and perennial plants (Suárez-López et al., 2001; Hayama et al., 2003; Böhlenius et al., 2006), and FT is thought to be a universal flowering signal (Turck et al., 2008; Turnbull, 2011; Wigge, 2011). In concordance with studies in other plants, we observed that the expression of FvFT1, the likely ortholog of FT, correlated with flowering in the LD accession Hawaii-4. Furthermore, strong reduction in the expression of FvAP1/FUL genes in FvFT1 RNAi plants as well as their late-flowering phenotype suggested that FvFT1 controls flower induction in F. vesca LD accessions (Fig. 7). However, in the SD F. vesca, FvFT1 was highly expressed in noninductive LD, which contrasts with findings in other SD plants (Hayama et al., 2003, 2007; Kong et al., 2010). Our data suggest that in this genotype, FvTFL1 expression in the shoot apex may overcome the function of FvFT1 as an activator of flowering. FT and TFL1 are closely related proteins whose opposite function is caused by differences in the external loop (Ahn et al., 2006). A recent study showed that FT forms a so-called “florigen activation complex” together with FD and 14-3-3 proteins (Taoka et al., 2011). Because TFL1 homologs in different species can also bind FD and 14-3-3 (Pnueli et al., 2001; Purwestri et al., 2009; Hanano and Goto, 2011), FvTFL1 may inhibit flowering by competing for binding partners with FvFT1. This is consistent with the regulation of flowering in tomato (Solanum lycopersicum), in which the balance between the expression of FT and TFL1 homologs controls flowering (Shalit et al., 2009). However, in SD F. vesca, both FvFT1 and FvTFL1 are down-regulated under flower-inductive conditions. Therefore, we hypothesize that flower induction in SD F. vesca takes place via an FvFT1-independent mechanism, whereas in Hawaii-4, FvFT1 functions as a LD-specific floral promoter in the absence of functional FvTFL1. On the other hand, we cannot exclude the possibility that FT is involved in the LD activation of FvTFL1, which is photoperiodically regulated only in the shoot apex (Fig. 4D; Supplemental Fig. S6B). As the perception of photoperiod is thought to occur in the leaves (An et al., 2004; Corbesier et al., 2007), a systemic signal may be required for the photoperiodic control of FvTFL1. FT is a good candidate for such a signal, because it is emerging as a general photoperiodic signaling molecule in diverse plant species (Wigge, 2011). Furthermore, Hecht et al. (2011) have shown evidence that leaf-expressed FT may mediate the photoperiodic control of another CETS family gene in the pea (Pisum sativum) shoot apex. Further studies are needed to resolve the function of FvFT1 in SD F. vesca and to establish which mechanism activates floral meristem identity genes after the down-regulation of FvTFL1.

A Model of the Regulation of Flowering and Growth Cycles by F. vesca TFL1 and FT Homologs

Based on our studies, we propose a model (Fig. 8) in which the flowering of SD F. vesca accessions is repressed through LD-activated FvTFL1 expression, which overrides the floral activator function of FvFT1. When photoperiod drops below a critical level during autumn, FvTFL1 mRNA levels decrease, floral meristem identity genes are up-regulated, probably through an FvFT1-independent pathway, and floral development begins. In the following spring and summer, FvTFL1 is activated again in new axillary shoots and keeps them vegetative until inductive conditions return. In contrast, the absence of functional FvTFL1 in LD accessions leads to FvFT1-mediated flower induction and continuous flowering under LDs. More studies are needed to explore how this model generated in the diploid Fragaria can be translated to the more complex octoploid cultivated strawberry, which shows similarities in the environmental control of flowering with F. vesca (Heide, 1977; Heide and Sønsteby, 2007; Sønsteby and Heide, 2007, 2008; Bradford et al., 2010).

The timing and duration of flowering season are traits of importance to crop production because they are major factors contributing to yield potential and the length of the harvest season. Natural variation in TFL1 homologs has been important for the domestication and crop improvement of several annual crops (Turnbull, 2011), and our study shows the importance of TFL1 in the perennial context. As TFL1 homologs exist in all angiosperms (Karlgren et al., 2011) and their function as floral repressors is conserved between plant families (Shannon and Meeks-Wagner, 1991; Foucher et al., 2003; Kotoda et al., 2006; Mohamed et al., 2010; Imamura et al., 2011), our findings are of general significance for understanding the regulation of flowering- and perennial-specific traits in a wide range of economically important crops.

Plant Material

Fragaria vesca accession (PI551792) and F. vesca f. semperflorens Hawaii-4 (PI551572) were obtained from the National Clonal Germplasm Repository. The Arabidopsis (Arabidopsis thaliana) tfl1-2 mutant (Nottingham Arabidopsis Stock Centre identifier N3091) in the Landsberg erecta background was used in the genetic complementation test.

Growth Conditions and Phenotyping

F. vesca plants were grown in greenhouse rooms equipped with darkening curtains. Standard LD growth conditions were 18 h of high-pressure sodium light (approximately 150 µmol m⁻² s⁻¹) at 18°C. In photoperiodic experiments at 18°C, 12 h of high-pressure sodium light (approximately 150 µmol m⁻² s⁻¹) was provided for all plants (SD), and under LD, 6 h of day extension was given as low-intensity incandescent light (approximately 8 µmol m⁻² s⁻¹). Arabidopsis plants were grown under LD (16 h of light at 22°C) or SD (10 h of light at 22°C) in a growth chamber. Light was provided by high-pressure sodium lamps (approximately 180 µmol m⁻² s⁻¹). Flowering time was measured either as the number of leaves initiated before the first inflorescence or the number of days from sowing to the first open flower.

Statistics

The χ² goodness-of-fit test was performed on the flowering data from the F2 and BC1 mapping populations. The critical value at α = 0.05 was 3.841 for all χ² tests.

Genotyping and Mapping

DNA extraction was done using a previously described protocol (Albani et al., 2004). The markers used for mapping are listed in Supplemental Table S1. Linkage maps were constructed as described earlier (Sargent et al., 2004). The F2 mapping population was characterized for FvTFL1 using TFL1-6FAM primers (Supplemental Table S1).

Identification of Single-Nucleotide Polymorphism Markers by Genome Resequencing

DNA of SD F. vesca (PI551792) was extracted (Albani et al., 2004) and treated with RNase A. A genomic DNA library was generated using the Genomic DNA Sample Prep Kit (Illumina), and 30-bp reads were sequenced using the Illumina GAIIx flow cell. Image analysis and base calling were done by OLB version 1.6.0 software and sequence alignment by CASAVA GERALD version 1.6.0. A default chastity threshold of 0.6 was used for purity filtering. All sequence reads with two or fewer mismatches were aligned with 801,610 bp of the genomic reference sequence of Hawaii-4 between markers V8p98 and V8p278 (Supplemental Fig. S1) using Bowtie version 0.12.2 and Readaligner version 2010_4rc2 software, and the resulting .sam files (sequence alignment/map files) were viewed by Tablet version 1.10.05.21 in order to find single-nucleotide polymorphisms. Resequencing and related bioinformatics were carried out at Biomedicum Genomics, University of Helsinki.

Expression Analysis

Total RNA was extracted, cDNA was synthesized, and real-time PCR was performed as described earlier (Mouhu et al., 2009). FvMSI1 was used as a stable control gene (Mouhu et al., 2009) for normalization. The number of biological replicates is indicated in each Figure legend. Three technical replicates were performed for each sample. Real-time PCR primers are listed in Supplemental Table S1.

Plasmid Constructs

Plasmid constructs for overexpressing FvTFL1 or FvmTFL1 were created according to Gateway Technology with Clonase II (Invitrogen). For overexpression of TFL1, the primers used to amplify cDNA from SD F. vesca (PI551792) and sfl mutant Baron Solemacher (PI551507) were 5′-AAAAAGCAGGCTCTGTACAACCTTTTCTCTTCTCCCTCT-3′ (attB1) and 5′-AGAAAGCTGGGTCCTCCCTGCAAGGTGCCTA-3′ (attB2). For the double-stranded FvTFL1 RNAi construct, the fragment was amplified from SD F. vesca cDNA with primers 5′-AAAAAGCAGGCTTGTTTGGCCTTGGCATCTCG-3′ (attB1) and 5′-AGAAAGCTGGGTTCTGCAGTCACCGCCAAACC-3′ (attB2). For creating FvFT1 RNAi constructs, cDNA from SD F. vesca (PI551792) was amplified with primers 5′-AAAAAGCAGGCTTGTTTGGCCTTGGCATCTCG-3′ (attB1) and 5′-AGAAAGCTGGGTTCTGCAGTCACCGCCAAACC-3′ (attB2). The destination vectors were p7WG2D for overexpression and PK7GWIWG2(II) for RNAi (Karimi et al., 2002).

Transformation

Arabidopsis plants were transformed with Agrobacterium tumefaciens strain GV3101 by the floral dip method (Zhang et al., 2006). Strawberry transformations were performed as described earlier (Oosumi et al., 2006). Plants of the T1 or T0 generation were analyzed in Arabidopsis and F. vesca, respectively.

In Situ Hybridization

In situ hybridization was performed on the apex of the main shoot as described previously (Kurokura et al., 2006). The probe template fragment was amplified by reverse transciption-PCR using the primers shown in Supplemental Table S1 and cloned into pDrive cloning vector (Qiagen). Sense and antisense probes were synthesized from T7 or SP6 promoters.

Sequence data from this article can be found in the GenBank data library under accession numbers JN172097 and JN172098.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Fine-mapping of SFL.

• Supplemental Figure S2. Deletion in FvTFL1 cosegregates with SFL in the SD F. vesca × F. vesca f. semperflorens BC1 population.

• Supplemental Figure S3. Phenotypes of 10-month-old LD-grown Hawaii-4 and P35S:FvTFL1 plants.

• Supplemental Figure S4. FvTFL1 expression in P35S:FvTFL1 transgenic lines.

• Supplemental Figure S5. FvTFL1 complements the Arabidopsis tfl1-2 mutant phenotype.

• Supplemental Figure S6. Expression analysis of FvTFL1.

• Supplemental Figure S7. Effect of photoperiod on the expression of floral meristem identity genes.

• Supplemental Figure S8. F. vesca LD accession Hawaii-4 flowers continuously after LD induction in both SD and LD conditions.

• Supplemental Figure S9. Expression of FvFT1 in LD-grown P35S:FvFT1-RNAi lines.

• Supplemental Table S1. Primers used in this study.