Constitutive Expression of an Apple FLC3-like Gene Promotes Flowering in Transgenic Blueberry under Nonchilling Conditions

MADS-box transcription factors FLOWERING LOCUS C (FLC) and APETALA1 (AP1)/CAULIFLOWER (CAL) have an opposite effect in vernalization-regulated flowering in Arabidopsis. In woody plants, a functional FLC-like gene has not been verified through reverse genetics. To reveal chilling-regulated flowering mechanisms in woody fruit crops, we conducted phylogenetic analysis of the annotated FLC-like proteins of apple and found that these proteins are grouped more closely to Arabidopsis AP1 than the FLC group. An FLC3-like MADS-box gene from columnar apple trees (Malus domestica) (MdFLC3-like) was cloned for functional analysis through a constitutive transgenic expression. The MdFLC3-like shows 88% identity to pear’s FLC-like genes and 82% identity to blueberry’s CAL1 gene (VcCAL1). When constitutively expressed in a highbush blueberry (Vaccinium corymbosum L.) cultivar ‘Legacy’, the MdFLC3-like induced expressions of orthologues of three MADS-box genes, including APETALA1, SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1, and CAL1. As a consequence, in contrast to the anticipated late flowering associated with an overexpressed FLC-like, the MdFLC3-like promoted flowering of transgenic blueberry plants under nonchilling conditions where nontransgenic ‘Legacy’ plants could not flower. Thus, the constitutively expressed MdFLC3-like in transgenic blueberries functioned likely as a blueberry’s VcCAL1. The results are anticipated to facilitate future studies for revealing chilling-mediated flowering mechanisms in woody plants.


Introduction
Exposure to low temperatures (i.e., vernalization) promotes the transition from vegetative to reproductive phase of plant development in most annually flowering species [1][2][3][4]. MADS-box transcription factors FLOWERING LOCUS C (FLC) and APETALA1 (AP1)/CAULIFLOWER (CAL) perform an opposite function in vernalization-regulated flowering initiation in Arabidopsis [5,6]. FLC represses the expression of two major flowering genes, FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) [5][6][7][8]. Constitutive expression of FLC often delays flowering and causes abnormal floral morphological changes, such as short stamens, reduction of pollen amount, and larger carpels in transgenic plants [6,[9][10][11][12]. In contrast, AP1/CAL positively promotes the expression of floral meristem identity genes in synergy with FRUITFULL (FUL) and LEAFY (LFY) [13,14]. LFY was shown to directly activate the expression of AP1/CAL in floral primordial and then produce a positive feedback effect on its own expression [13]. Overexpression of AP1 gene promoted transition from apical or lateral shoots into flowers and early flowering phenotype in transgenic Arabidopsis [15]. Genetic evidence showed that either single or double AP1/CAL mutants led to abnormal floral morphology in flower development [5,16]. Ectopic expression AP1 genes in other plant species also showed early flowering phenotypes. For example, constitutive overexpression of AtAP1 in citrus showed flowering initiation in the first year [17]. Overexpression of apple AP1-like genes, MdMADS2 and MdMADS5, induced early flowering in both tobacco and Arabidopsis [18]. Overexpression of birch BpAP1 in tobacco and birch also generated early flowering and dwarf phenotypes [19].
Chilling requirement refers to fulfilling a minimum period of chill hours to stimulate dormancy release and ensure seasonal growth of vegetative and floral buds in a fruit-bearing tree. The chilling requirement for dormancy-break is species-and genotype-dependent [20,21]. Insufficient chilling prevents bud-break and often leads to reduced fruit production mainly due to abortive flower buds and prolonged bloom. To date, unlike the extensively studied vernalization pathway in annual plants, the chilling-involved flowering mechanism in woody plants is largely not known [22][23][24].
Climate changes in the last 40 years have caused earlier shifts in the onset of growing seasons and increased temperature fluctuations [25]. Consequently, early onset of the growing seasons causes insufficient chilling for fruit trees, and increased temperature fluctuations during plant bloom turn early season frosts into freezing injuries to flowers and young fruits [26]. To secure deciduous fruit production, manipulating chilling requirements and cold/freezing tolerance through plant breeding is considered to be a long-term solution to mitigate the potential threats of climate change [27]. Developing cultivars with a low chilling requirement through modifying the flowering initiation pathway may expand the cultivation areas of temperate deciduous fruit trees in warm areas and may also secure high profitability of protected cultivation in greenhouses [28][29][30].
Apples (Malus domestica) and highbush blueberries (Vaccinium corymbosum L.) need sufficient chilling exposure to release bud dormancy. Modifying the flowering initiation pathway is one of the most important improvement properties for these two fruit crops. In apples, transcription profiling of chilled buds has revealed the potential roles of FLC-like, FT-like and TERMINAL FLOWER 1-like (TFL1-like) genes [31,32]. To date, functional FLC orthologues have not been identified in woody fruit crops. To study the molecular mechanism of chilling-mediated flowering in woody plants, we cloned an FLC-like MADS-box gene from columnar apple trees (MdFLC3-like). Functional analysis of the MdFLC3-like through constitutive expression was conducted in transgenic highbush blueberry cv. Legacy. Surprisingly, we found for the first time that constitutive expression of the MdFLC3-like in transgenic blueberry plants promoted flowering under nonchilling conditions where nontransgenic could not flower. The results are anticipated to facilitate future studies for revealing chilling-mediated flowering mechanisms in woody plants.

Isolation and Sequence Analysis of an MdFLC3-like Gene
In the previous study of early flowering in column apple trees, a putative FLC-like candidate gene was identified from apples using gene-specific primers based on the apple FLC3 sequence MD10G1041100 retrieved from Malus × domestica GDR RefTrans V1 [33]. The low expression of the FLC-like seemed to be responsible for the early flowering and bud dormancy release [31,33]. The cloned FLC-like cDNA sequence codes 200 amino acid residues (GenBank accession number: MK986790); its molecular mass is projected to be 22.42 kDa and isoelectric point is 8.15. It showed 95% identity to MD10G1041100 at protein level. An NCBI Conserved Domain Database search revealed that this FLC-like cDNA possesses several typical conserved domains including MADS-box, I-domain, K-domain and C-domain, which is a MIKC c -type MADS-box transcription factor (Figure 1). In a phylogenetical analysis, the protein of this cDNA was grouped into the FLC clade along with FLC-like proteins of Arabidopsis, grape, pear, and poplar ( Figure 2); for example, it shows a high identity to a Japanese pear (Pyrus pytifolia 'Culta') protein annotated as FLC (up to 92%). Accordingly, the cloned cDNA was designated as the MdFLC3-like. It is interesting that not every protein sequence showing high identity to the MdFLC3-like in the phylogenetical analysis was annotated as FLC-like ( Figure 2). One MdFLC3-like orthologue in a Chinese white pear (Pyrus bretshneideri) shows 88% identity and was annotated as CAULIFLOWER (CAL)-like. Similarly, the MdFLC3-like shows 82% identity to a blueberry CAL (VcCAL1) in a highbush blueberry ( Figure 2). More interestingly, all the annotated FLC-like proteins that showed high similarities to the MdFLC3-like are grouped more closely to Arabidopsis AP1 than FLC clade ( Figure 2). Apparently, the MdFLC3-like and its orthologues in other woody plants are all annotated based on identities of their protein sequences while their functions have not been verified through functional analysis. I-domain, K-domain and C-domain, which is a MIKC c -type MADS-box transcription factor ( Figure  1). In a phylogenetical analysis, the protein of this cDNA was grouped into the FLC clade along with FLC-like proteins of Arabidopsis, grape, pear, and poplar ( Figure 2); for example, it shows a high identity to a Japanese pear (Pyrus pytifolia 'Culta') protein annotated as FLC (up to 92%). Accordingly, the cloned cDNA was designated as the MdFLC3-like. It is interesting that not every protein sequence showing high identity to the MdFLC3-like in the phylogenetical analysis was annotated as FLC-like ( Figure 2). One MdFLC3-like orthologue in a Chinese white pear (Pyrus bretshneideri) shows 88% identity and was annotated as CAULIFLOWER (CAL)-like. Similarly, the MdFLC3-like shows 82% identity to a blueberry CAL (VcCAL1) in a highbush blueberry ( Figure 2). More interestingly, all the annotated FLC-like proteins that showed high similarities to the MdFLC3-like are grouped more closely to Arabidopsis AP1 than FLC clade ( Figure 2). Apparently, the MdFLC3-like and its orthologues in other woody plants are all annotated based on identities of their protein sequences while their functions have not been verified through functional analysis.

Expression Levels of MdFLC3-like in Four Apple Tissues
RT-qPCR analysis was conducted to characterize the expression of the MdFLC3-like in fully chilled flower buds, petals, stamens/pistils, and sepals. The fully chilled flower buds showed a significant lower MdFLC3-like expression than the other three flower tissues ( Figure 3). Unlike in Arabidopsis, where the expression of FLC was barely detected in young tissues of the inflorescence [34], high expressions of the MdFLC3-like were observed in apple petals, stamens/pistils and sepals (Figure 3), suggesting that it is not likely that the MdFLC3-like functions as an FLC function in apples despite being suggested so from its annotation and reduced expression in responding to chill accumulation [35]. (AP1) proteins. A phylogenetic tree of these MADS-box genes was generated by the neighbor-joining (NJ) method with 1000 bootstrap replicates. The proteins were clustered and divided into two distinct clades, FLC and AP1.

Expression Levels of MdFLC3-like in Four Apple Tissues
RT-qPCR analysis was conducted to characterize the expression of the MdFLC3-like in fully chilled flower buds, petals, stamens/pistils, and sepals. The fully chilled flower buds showed a significant lower MdFLC3-like expression than the other three flower tissues ( Figure 3). Unlike in Arabidopsis, where the expression of FLC was barely detected in young tissues of the inflorescence [34], high expressions of the MdFLC3-like were observed in apple petals, stamens/pistils and sepals (Figure 3), suggesting that it is not likely that the MdFLC3-like functions as an FLC function in apples despite being suggested so from its annotation and reduced expression in responding to chill accumulation [35].

Ectopic Expression of MdFLC3-like Promoted Flowering of Nonchilled Blueberry
The MdFLC3-like was cloned into the recombinant plasmid pBI121 to make a 35S-MdFLC3-like construct of constitutive expression (Figure 4a). The 35S-MdFLC3-like construct was transformed into blueberry cultivar 'Legacy'. A total of 7 out of 232 leaf explants produced Km-resistant shoots, of which six PCR-positive transgenic events were identified (Figure 4b

Ectopic Expression of MdFLC3-like Promoted Flowering of Nonchilled Blueberry
The MdFLC3-like was cloned into the recombinant plasmid pBI121 to make a 35S-MdFLC3-like construct of constitutive expression (Figure 4a). The 35S-MdFLC3-like construct was transformed into blueberry cultivar 'Legacy'. A total of 7 out of 232 leaf explants produced Km-resistant shoots, of which six PCR-positive transgenic events were identified (Figure 4b-d).   (Table 1), and the flowers were eventually developed into fruits ( Figure 5). TR2 and TR3 produced fewer flower clusters (Table 1). Overall, the transgenic plants had more flower buds than the NT plants (Table 1). Apparently, constitutive expression of the 35S-MdFLC3-like functioned as a flowering promoter under nonchilling conditions; the results are in contrast to those of Arabidopsis in which an overexpressed FLC delays flowering [6].  (Table 1), and the flowers were eventually developed into fruits ( Figure 5). TR2 and TR3 produced fewer flower clusters (Table 1). Overall, the transgenic plants had more flower buds than the NT plants (Table 1). Apparently, constitutive expression of the 35S-MdFLC3-like functioned as a flowering promoter under nonchilling conditions; the results are in contrast to those of Arabidopsis in which an overexpressed FLC delays flowering [6].   Constitutive expression of the MdFLC3-like was confirmed in all three transgenic events but was absent in NT plants ( Figure 6). The dwarfing TR1 plants showed a higher expression of the MdFLC3-like gene than TR2 and TR3, suggesting that the higher expression was responsible for the dwarf plants with the most promoted flowering (Table 1). Constitutive expression of the MdFLC3-like was confirmed in all three transgenic events but was absent in NT plants ( Figure 6). The dwarfing TR1 plants showed a higher expression of the MdFLC3like gene than TR2 and TR3, suggesting that the higher expression was responsible for the dwarf plants with the most promoted flowering (Table 1). Due mainly to the lack of real biological replicates for each of the three transgenic lines at the time of analysis, the transcript levels of VcCAL1 and its major downstream genes in newly-formed, fully-expanded leaves were quantified using semi-quantitative RT-PCR. Expression of VcAP1 was detected in all three transgenic plants but absent in the NT plants, suggesting that the increased VcAP1 could be the major cause of the promoted flowering in the three transgenic plants. Increased VcCAL1 and VcSOC1 (compared to the NT plant) were detected in TR1 and TR2 but not in TR3 plants. Expression of VcLFY was hardly detectable in either transgenic or NT plants. The TR1 showed higher expressions of both VcAP1 and VcSOC1 than TR2 and TR3, which might contribute to the more significant phenotypic changes in plant size and flowering time in the TR1. Apparently, constitutively expressed MdFLC3-like acted as a positive regulator of flowering by enhancing expression of VcAP1 or VcSOC1.

A Functional FLC Orthologue in Woody Plants is Unknown
FLC functions as a floral repressor during flowering initiation and it is down-regulated after vernalization [1,3,[6][7][8]11]. FLC-like genes are present in a multi-gene subfamily in most of the annual plants and have been identified in several woody plants based on sequence similarities, including four in poplar (PtFLC2-5) [36,37], two in grapevine (VvFLC1 and VvFLC2) [38] and one in peach (Ppe MADS08) [39]. Using the Arabidopsis FLC as a query, orthologues of the FLC have not been identified in apple, apricot [40], cucumber [41], or blueberry [42]. However, when the phylogenetic analysis was Due mainly to the lack of real biological replicates for each of the three transgenic lines at the time of analysis, the transcript levels of VcCAL1 and its major downstream genes in newly-formed, fully-expanded leaves were quantified using semi-quantitative RT-PCR. Expression of VcAP1 was detected in all three transgenic plants but absent in the NT plants, suggesting that the increased VcAP1 could be the major cause of the promoted flowering in the three transgenic plants. Increased VcCAL1 and VcSOC1 (compared to the NT plant) were detected in TR1 and TR2 but not in TR3 plants. Expression of VcLFY was hardly detectable in either transgenic or NT plants. The TR1 showed higher expressions of both VcAP1 and VcSOC1 than TR2 and TR3, which might contribute to the more significant phenotypic changes in plant size and flowering time in the TR1. Apparently, constitutively expressed MdFLC3-like acted as a positive regulator of flowering by enhancing expression of VcAP1 or VcSOC1.

A Functional FLC Orthologue in Woody Plants is Unknown
FLC functions as a floral repressor during flowering initiation and it is down-regulated after vernalization [1,3,[6][7][8]11]. FLC-like genes are present in a multi-gene subfamily in most of the annual plants and have been identified in several woody plants based on sequence similarities, including four in poplar (PtFLC2-5) [36,37], two in grapevine (VvFLC1 and VvFLC2) [38] and one in peach (Ppe MADS08) [39]. Using the Arabidopsis FLC as a query, orthologues of the FLC have not been identified in apple, apricot [40], cucumber [41], or blueberry [42]. However, when the phylogenetic analysis was conducted with the orthologs of poplar, peach, and grapevine MADS-box genes, four apple genes (i.e., MD09G1009100, MD17G1001300, MD05G1037100, and MD10G1041100) were clustered in the same group together with poplar, peach, and grapevine FLC-like genes [35,43]. To date, none of these FLC-like genes have been verified through a functional analysis. Thus, no convincing evidence has shown that a functional FLC-like gene plays a major role in chilling-mediated flowering in woody plants.

MdFLC3-like Does Not Function as an FLC in Blueberry
The MdFLC3-like isolated from apples shows a high identity to MD10G1041100, which was annotated as an FLC-like gene in apples based on sequence analysis and characterization of its expression in apples [36]. Since the expression of Arabidopsis FLC was hardly detected in the young tissues of inflorescence [35], a functional FLC-like is anticipated to have a low or no expression in inflorescence. However, the high expression levels of MdFLC3-like in different apple flower tissue (i.e., sepals, petals, stamen, and pistil) suggest that the MdFLC3-like in apples is likely different from the FLC in Arabidopsis. Additionally, despite of being grouped in the FLC-like clades (Figure 2), not all of the FLC-like orthologues identified in woody plants showed the similar transcriptional expression pattern as FLC. For example, PtFLC2 and PtFLC4 in poplar showed opposite expression patterns during bud dormancy release [37,39]. The PtFLC2 expression decreased in response to chilling during winter dormancy and increased after dormancy release [37]. A similar expression pattern was observed for grapevine VvFLC1 [39]. Our work showed that the MdFLC3-like expression was similar to that of PtFLC2 and VvFLC1 [36]. For the poplar PtFLC4 and grape VvFLC2, their expressions were increased during dormancy and decreased after cold exposure [37,39].
FLC is a repressor of flowering through repressing its downstream SOC1, SPL15, and FT by directly binding to the promoters of SOC1 and SPL15 or the first intron of FT [44,45]. Overexpression of FLC resulted in a late-flowering phenotype in Arabidopsis and Brassica rapa [9]. Silencing BcFLC2, the homolog of FLC in Pak-choi (Brassica rapa ssp. Chinensis) caused early flowering [46]. In this study, the fact that constitutively expressed MdFLC3-like promoted blueberry flowering under nonchilling conditions supports that MdFLC3-like has little FLC functions as a repressor in chilling-mediated flowering. The observed phenotypic changes seem to be well supported by the expression levels of the MdFLC3-like and the enhanced expression of VcAP1 and VcSOC1.

MdFLC3-like Likely Functions in a Positive Role in Flowering Initiation in Transgenic Blueberry
Based on transcriptome reference (GenBank accession number: SRX2728597) of blueberry cultivar 'Legacy', several MADS-box genes were found to show similarity to the FLC; however, none of them were annotated as an FLC-like due to their higher similarities to the other MADS-box genes [24,42]. For example, using the MdFLC3-like as a query to search the blueberry transcriptome reference, the best hit was actually the VcCAL1 (up to 82% identity). Similarly, the MdFLC3-like orthologue in a Chinese white pear was also annotated as a CAL-like (Figure 2).
In nonchilled tissues of blueberry 'Legacy', the expression of VcCAL1 was the highest in flower tissues, followed by leaves and buds [42]. This is consistent with the tissue specificity and developmental stage of the CAL expression in Arabidopsis [5,13]. In addition, from nonchilled to fully chilled flower buds, a significant increase in VcCAL1 expression occurred, and then a decrease in the expression was found in late-pink buds [24]. Apparently, the expression pattern of the VcCAL1 in blueberries is similar to that of the CAL in Arabidopsis but is different from that of either the MdFLC expression in apples or FLC expression in Arabidopsis [1,3,[6][7][8]11,35].
CAL is considered functionally to be partially redundant to AP1 and acts as a positive regulator of flower development in Arabidopsis [5,13]. Constitutive expression of the MdFLC3-like promoted flowering in transgenic blueberries, suggesting that the MdFLC3-like does not function as an FLC in blueberries as expected. This was further supported at transcript levels, where the upregulated expression of VcAP1 and VcSOC1 was observed only in the transgenic plants (Figures 5 and 6).

MADS-Box Genes Play Key Roles in Chilling-Mediated Flowering of Woody Plants
FLC is a MADS-box gene. Another MADS-box gene cluster named DORMANCY-ASSOCIATED MADS-box (DAM) genes in peaches were proposed as candidates for regulation of the terminal bud's formation in response to dormancy-inducing conditions [47]. However, the DAM genes are the orthologues of A. thaliana AGL24 and SVP genes [48,49] rather than the orthologues of FLC. More recently, the SHORT VEGETATIVE PHASE-LIKE (SVP-like) of hybrid aspen, an orthologue of A. thaliana SVP, was reported to be involved in axillary bud dormancy as a mediator of temperature controlled bud break [50]. Hence, it is likely that multiple MADS-box genes, instead of any individual FLC-like or DAM genes, confunction to determine the process of chilling-mediated flowering in woody plants. In blueberries, the changes from nonchilled to chilled and chilled to late-pink buds are associated with transcriptional changes in a large number of differentially expressed (DE) flowering pathway genes (i.e., the orthologues of FT, FD, TFL1, LFY, and other MADS-box genes) [24]. In general, fully chilling upregulates blueberry MADS-box genes [24]. It is interesting that the functional orthologues of FLC and AGL24 were not detected in blueberries [24]. In addition, overexpression of the keratin-like (K) domain of the VcSOC1 promoted flowering of the blueberry cultivar 'Aurora' under nonchilling conditions [51]. It is likely that the interaction of multiple MADS-box genes, rather than a single FLC-like gene, co-regulates chilling-mediated flowering in blueberries as well as in other woody plants.
Further studies are still needed to reveal roles of MADS-box genes in chilling-mediated flowering of woody plants.

Plant Material
Six-year-old columnar apple trees (Malus domestica) from a cross of 'Gala' × 'Telamon' were used for gene cloning and expression analysis in this study. These trees were grown at the Laiyang Experiment Station of Qingdao Agricultural University in Shandong Province, China. Chilled flower buds, 40-50 per tree, were collected in March 2011. Flower tissues (e.g., sepals, petals, stamens, and pistils), about 200 mg per tree, were harvested in early May 2011. Samples of three biological controls from three trees were collected and immediately frozen in liquid nitrogen and stored at −80 • C for later RNA extraction.
In vitro 'Legacy' shoots were cultured on 30 mL WPM2Z in 40 mm × 110 mm glass jars and incubated for 4 weeks at 25 • C, 30 µE/m 2 /s of 16 h/8 h (day/night); rooting of in vitro cultured shoots and greenhouse care of rooted plants were conducted according to the published protocols [52]. All plants were grown in a greenhouse (heated for winter) under natural light conditions with routine management of watering with added 0.2 g/L fertilizer (nitrogen:phosphorus:potassium = 21:7:7) unless otherwise mentioned. All blueberry experiments were conducted at Michigan State University (East Lansing, Michigan, USA).

MdFLC3-like Cloning and Phylogenetic Analysis
Total RNA was isolated from nonchilled apple flower buds using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) [33]. Total RNA (0.5 µg) was reverse transcribed into complementary DNA (cDNA) using SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA). The open reading frame (ORF) of the MdFLC3-like gene was cloned by PCR amplification reaction using GoTaq Green Master Mix (Promega, Madison, WI, USA). Gene-specific primers on the basis of the sequence MD10G1041100 were retrieved from Malus × domestica GDR RefTrans V1 and used to amplify MdFLC3-like ( Table 2). The PCR was carried out with the following protocol: an initial denaturation of 94 • C for 5 min; 35 cycles of 94 • C for 30 s, 58 • C for 30 s, and 72 • C for 1 min; and a final extension of 72 • C for 7 min. The deduced MdFLC protein sequence was characterized using CLC sequence viewer 7 (QIAGEN Bioinformatics, Aarhus, Denmark). Amino acid sequences of MdFLC orthologs were retrieved using the NCBI server (http: //blast.ncbi.nlm.nih.gov/Blast.cgi). In addition, the MdFLC was also used as the query to search the transcriptome reference developed for 'Legacy' using Trinity and Trinotate (GenBank accession number: SRX2728597) [42]. Selected orthologs were aligned using the ClustalX (http://www.clustal.org). A phylogenetic tree was generated using the MEGA5 [53].

Quantitative Reverse Transcription PCR (RT-qPCR) Analysis
To determine the expression levels of MdFLC3-like in different apple tissues, total RNA was isolated from nonchilled and chilled buds, sepals, petals, stamens, and pistils. For RT-qPCR analysis, 1 µg aliquot of total RNA treated with DNase I (Invitrogen) was reverse-transcribed using SuperScript II Reverse Transcriptase (Invitrogen). Real-time PCR using SYBR Green PCR Core Reagents (Life Technologies, Carlsbad, CA, USA) was carried out on an Applied Biosystems StepOne™ thermocycler (Applied Biosystems, Foster City, CA, USA) with three technical replicates per sample. The amplification conditions were 95 • C for 10 min, followed by 40 cycles of amplification (95 • C for 15 s, 60 • C for 1 min) with plate readings after each cycle and performance of a melting curve analysis. Primers used for the reactions are provided in Table 2. Preliminary experiments were conducted to determine the efficiencies of the primers and confirm that the internal control was suitable. Data was analyzed by the 2 −∆∆Ct method using StepOne Software V2.2 (Applied Biosystems).

Construction of a MdFLC3-like Expression Vector
To make a construct for constitutive expression, the MdFLC3-like fragments (5 -BamHI-MdFLC3-like-EcoRV-3 ) containing two added restriction enzyme sites were amplified by specific primers MdFLC-BF and MdFLC-ER (Table 2) from the apple cDNA. The purified PCR fragments were released by double digestion, and the digested 5 -BamHI-MdFLC3-like-EcoRV-3 fragments were purified. Plasmid pBI121 was digested with SacI; the SacI-digested fragments were incubated by DNA Polymerase I Large Klenow Fragment (NEB, Ipswich, MA, USA) to generate blunt-SacI ends. The digested pBI121 containing blunt-SacI ends were digested with BamHI to remove the gusA coding region. The opened BamHI and blunt-SacI sites in the T-DNA region between the cauliflower mosaic virus (CaMV) 35S promoter and the Nos terminator in pBI121 allows the insertion of the digested 5 -BamHI-MdFLC3-like-EcoRV-3 fragments ( Figure 4A). The resulting 35S-MdFLC3-like was sequenced and introduced into Agrobacterium tumefaciens stain EHA105 [54] using the freeze-thaw method.

Transformation of 35S-MdFLC3-like to 'Legacy'
Transformation of 'Legacy' was performed as previously reported [52]. To identify transgenic blueberry events, DNA and RNA were extracted from leaf tissues according to protocols reported previously [55]. Two pairs of primers, NPTII-F and NPTII-R as well as 35S-F and MdFLC-ER (Table 2), were used to detect the presence of nptII and MdFLC3-like genes separately.
Since the tetraploid southern highbush blueberry 'Legacy' needs over 800 chilling units (CU) for normal flowering, it will not flower under the above greenhouse conditions without a chilling treatment. The total CU was calculated according to Norvell et al., 1982 [56]. Both transgenic and nontransgenic 'Legacy' shoots were directly rooted in 48-cell trays containing sphagnum peat moss [52]. Due to the concerns of their vulnerability to freezing, these rooted one-year-old plants were

Gene Expression Analysis of Transgenic 'Legacy'
To compare the expression levels of MdFLC3-like and other flowering genes in transgenic and nontransgenic blueberries, three transgenic events and wild-type blueberries were studied. Fully-expanded leaves near the shoot tips, about one gram from each plant, were randomly collected from TR1, TR2, TR3, and three nontransgenic plants (NT). The leaf samples were grounded in liquid nitrogen and 200 mg per sample were used for total RNA isolation. A two-step semi-quantitative RT-PCR was performed using SuperScript II Reverse Transcriptase (Invitrogen) for reverse transcription and GoTaq Green Master Mix (Promega) for PCR amplification. For the NT plants, cDNA mixtures from three plants were used in semi-quantitative RT-PCR analysis. Since there was only one plant for each transgenic event in this investigation, RT-qPCR was not conducted due to the lack of biological replicates. The sequences of the specific primers to distinguish the expression of MdFLC3-like, VcCAL1, VcAP1, VcSOC1, and VcLFY were listed in in Table 2. VcTIF was used as an internal control and normalizing reference for each gene in all samples. The PCR reactions were performed in triplicate and the PCR-amplified gene products were detected in a 2% agarose gel.

Conclusions
As an initial step to investigate the potential roles of an FLC-like gene in woody plants, we cloned an apple FLC-like gene (MdFLC3-like), which shows a high similarity to blueberry's VcCAL1. Ectopic expression of the MdFLC3-like likely promoted flowering of nonchilled transgenic plants by enhancing expression of VcAP1 and VcSOC1, both of which were upregulated in the chilled flower buds of nontransgenic plants [24]. These results suggest that the MdFLC3-like gene functioned as VcCAL1, a positive regulator of flowering in transgenic blueberries.