Molecular characterization of a soybean FT homologue, GmFT7

Soybean (Glycine max) is a vital oilseed legume crop that provides protein and oil for humans and feedstock for animals. Flowering is a prerequisite for seed production. Floral transition, from vegetative to reproductive stage, in a plant, is regulated by environmental (light, temperature) and endogenous factors. In Arabidopsis, Flowering Locus T (FT) protein is shown to be a mobile signal that moves from leaf to shoot apical meristem to induce flowering. However, FTs role in soybean is not fully resolved due to the presence of multiple (ten) homologs in the genome. Two of the ten FT homologs (GmFT2a and GmFT5a) have a role in the floral transition while GmFT1a and GmFT4 suppress soybean flowering. Recent deep sequencing data revealed that six FT homologs are expressed in shoot apical meristem and leaves during floral transition. One FT homolog, GmFT7 showed strong expression during soybean floral transition. Though bioinformatic analyses revealed that GmFT7 had high similarity with GmFT2a, ectopic GmFT7 expression in Arabidopsis could not promote flowering or rescue the late-flowering phenotype of Arabidopsis ft-10 mutant.

Soybean seed is one of the major sources of high-quality proteins and oil for human consumption and soybean seed meal for livestock feed. Further, soybean fixes atmospheric nitrogen through interaction with rhizobia reducing industrial fertilizer use and contributing to sustainable agriculture. Soybean is leading in oil production among the major oilseed crops with the world soybean production 361.0 million tons in 2018, accounting for 61% of the world oilseed 1 . Flowering is a prerequisite to yield of crops where seeds are the desired products. Moreover, with the climate change conditions, flowering is a critical factor for maintaining crop productivity.
Soybean flowering and seed maturity are determined by day length and temperatures during the growth and development stages. A wide range of soybean cultivars is grown worldwide across a range of latitudes with each cultivar adapted to a narrow latitude band restricting its cultivation. Maturity groups classification system (MGs) has been introduced to group soybean varieties based on their adaptation to specific environmental conditions. There are 13 maturity groups, namely, MG 000, 00, 0, and I to X. In North America, maturity groups MG 000, 00, and 0 are the earliest to flower and set seeds, and these maturity groups are mainly cultivated in south Canada 2 . In contrast, MGs IX and X are late flowering types and primarily grown in the southern US 3 .
Plants depend on their ability to measure the seasonal changes in photoperiod and ambient temperatures for their development and reproduction. Photoperiod is among important cues, which plants use to know the time of a year 4 . In 1920, Garner and Allard introduced photoperiod concept when they performed experiments with Maryland Mammoth cultivar of tabacoo 5 . Photoperiod is the extent of light and darkness in a 24 h cycle 6 . According to the responses under different photoperiods, flowering plants can be grouped as short-day (SD) plants, long-day (LD) plants, and day-neutral plants. The SD and LD plants flowering process can be accelerated if the day length is shorter or longer than a threshold. The flowering time of day-neutral plants is not regulated by day length. When photoperiod and temperature signals are perceived, shoot apical meristem (SAM) transition from vegetative growth to the reproductive stage to initiate flowering 7 .
Soybean is a short day (SD) plant implying floral induction will occur when plants are grown in a 12 h or less day length. Due to the critical role that flowering plays in determining seed yield, recent studies have focused on unravelling the molecular mechanisms underlying floral transition in soybean [8][9][10][11][12][13][14][15][16][17][18] . Most of our knowledge on flowering is based on a model plant, Arabidopsis. However, Arabidopsis is an LD plant requiring the longer than a critical day length for flowering.
In Arabidopsis, three main environmental pathways, namely photoperiod, vernalization, and autonomous control of flowering 19 . Of these, the photoperiod pathway being the most ancient and conserved. CONSTANS (CO), a transcriptional regulator, integrates day length information in the flowering photoperiod pathway. Furthermore,  20 . In brief, CO protein is stabilized by light and can only accumulate under inductive long days while degrading quickly in darkness 21 . CO is expressed in leaves, and it activates the expression of FLOWERING LOCUS T (FT). FT encodes a protein which is a mobile flowering signal 19 . FT protein, rather than FT mRNA, is the florigen which moves from leaf to shoot apex [22][23][24] . At SAM, FT interacts with FD (FLOWERING LOCUS D), a bZIP transcription factor to form FT-FD complex leading to initiation of flowering by activating floral integrator genes such as SUPPRESSOR OF OVEREXPRES-SION OF CO 1 (SOC1) and LEAFY (LFY) and enhances the expression of floral identity gene APETALA1 (AP1) and represses the expression of repressor TERMINAL FLOWER (TFL). AP1 is a homeotic gene that encodes a transcription factor to specify the floral meristem identity and determines sepal and petal development 25 . On the contrary, rice (Oryza sativa) is an SD plant and Heading date 3a (Hd3a), the Arabidopsis FT homologue in rice, interacts with 14-3-3 proteins first in SAM. The Hd3a-14-3-3 complex translocates to the nucleus to bind to OsFD1, an Arabidopsis FD homologue in rice, to activate floral transition 26,27 highlighting differences between Arabidopsis (a model plant) and a crop plant. Ten FT homologues found in the soybean genome and genetic studies showed that GmFT2a and GmFT5a promote flowering 12 while GmFT1a and GmFT4 function repressors 13,18 . Functions of GmFT2a and GmFT5a are non-redundant, and GmFT2a is more critical under SD condition, while GmFT5a is more critical under LD condition 9 . Mutagenesis of GmFT2a resulted in delayed flowering under both LD and SD conditions 8 , ft2aft5a double mutants exhibited late flowering even under SD condition 9 . Besides, GmFT2a and GmFT5a are also involved in post-flowering stem termination with GmFT5a being more effective than GmFT2a by interacting with a bZIP transcription factor GmFDL06 16 . In contrast, GmFT4 was primarily expressed in leaves and showed induced and reduced expression under LD and SD conditions, respectively 18 . Highly elevated expression of GmFT4 was observed in transgenic soybean plants over-expressing soybean maturity gene E1, which is a legume specific gene with a highly significant effect on flowering time and photoperiod sensitivity. In addition, ectopic expression of GmFT4 resulted in late flowering phenotype in transgenic Arabidopsis plants. Like GmFT4, GmFT1a was primarily expressed in leaves, GmFT1a showed induced and reduced expression under LD and SD conditions, respectively 13 . Over-expression of GmFT1a in soybean delayed flowering, confirming that it as a flowering repressor 13 . Furthermore, Chen et al. 10 examined GmFT2b sequence variations in 195 soybean cultivars and found that a Hap3, a major halophyte correlated with significant earlier flowering in higher altitude soybean cultivars.
Previous deep sequencing 28 studies in our lab to characterize leaf and shoot apical meristem transcriptome in soybean after short day treatment to evoke floral transition revealed a soybean FT homologue, Glyma02g07650, named to GmFT7 here and after. Differential gene expression analysis identified the genes involved during soybean floral transition. GmFT7 was found to be highly expressed both in leaf and shoot apical meristem tissues upon short-day treatment. This study aimed to understand the function of GmFT7 during floral transition.

Results
Identification, phylogenetic and expression analysis of GmFT7. FT is a member of the Phosphatidylethanolamine-binding protein (PEBP) gene family encoding a 175 amino acids (aa) in Arabidopsis (AtFT). Most of the soybean homologues (GmFT homologues) are predicted to be between 172 to 176 aa. A database search showed that GmFT7 (Glyma02g07650) has a CDS region of 195 bp, corresponding to 64 amino acids. The predicted GmFT7 protein with a molecular weight of 7.50 kDa is much smaller than that of GmFT2a (19.74 kDa), GmFT5a (19.09 kDa) and Arabidopsis FT (19.81 kDa), but with a higher isoelectric point (theoretical pI) of 10.38 than GmFT2a, GmFT5a and AtFT (5.59, 7.85 and 7.75 respectively). The instability index ranging from 41.27 (GmFT2a) to 48.81 (AtFT) indicated that all the four FT proteins are unstable. GRAVY (Grand Average of Hydropathy) values, which indicate hydropathicity of proteins, are all negative implying that all the four FT proteins are hydrophilic. Protein subcellular location analysis showed that GmFT7 is located in cytoplasm and nucleus similar to GmFT2a/5a and Arabidopsis FT.
Chromosomal distribution of ten soybean FT homologues showed that all FTs are localized on five of the 20 soybean chromosomes, GmFT1a/1b on chromosome 18, GmFT2a/2b/3a/5a on chromosome 16, GmFT3b/5b on chromosome 19, GmFT4 on chromosome 8 and GmFT7 on chromosome 2 ( Fig. 1). Centromeric coordinates for each chromosome were retrieved from Golicz et al. 29 . Among soybean FT homologues, eight FT homologues found in pairs (GmFT1a/GmFT1b; GmFT2a/GmFT2b; GmFT3a/GmFT5a; and GmFT3b/GmFT5b) except GmFT4 and GmFT7 which are located on chromosomes 8 and 2, respectively ( Fig. 1). Gene structures of GmFTs showed that all the FTs have four exons except that GmFT7 has only one (fourth) exon (Fig. 2). Previous studies showed a significant difference between FT2c from domesticated soybean (Glycine max) and wild soybean (Glycine soja), where GsFT2c is an intact FT homologue while GmFT2c harboured a structural rearrangement in domesticated soybean 17,30 . This variation suggested that there is no intact FT2c in domesticated soybean. A 13.6 kb insertion separated GmFT2c into two portions with the C-terminus containing only the fourth exon and 3′-UTR 17 . Our analysis showed that GmFT7 has only one exon in the C-terminus, that shows 100% of sequence similarity to the fourth exon of GsFT2c ( Supplementary Fig. S1).
To characterize phylogenetic relationships between GmFT7 and other FT homologues in soybean, protein sequences of GmFT7 and other GmFT homologues were retrieved from Phytozome. We also included Arabidopsis FT in the analysis. Ten GmFT homologues were classified into four groups and one singleton (GmFT4; Supplementary Fig. S2) under the Maximum Likelihood tree. FT homologues in the same clade showed a higher percentage of identity: 98.41% between GmFT2a and GmFT7, 96.51% between GmFT2a and GmFT5a, 94.29% between GmFT3a and GmFT3b, 90.91% between GmFT2a and GmFT2b, while a relatively lower identity between GmFT1a and GmFT1b.
Gene structure analysis showed that GmFT7 only has the fourth exon compared to other FT homologues with four exons (Fig. 2). The fourth exon had been shown to harness a critical role 31  www.nature.com/scientificreports/ GmFT7 and the four well studied GmFT homologues (flowering promoters and repressors) showed that GmFT7 has a conserved 14-amino acid segment B and the LYN in segment C 12 (Fig. 3). The second last amino acid in segment B was indicated as a critical amino acid that was regarded as FT proteins 31 , suggesting that GmFT7 can function as an FT. LYN (triad) is a three amino-acid region which distinguishes FT and TFL1 proteins. This triad is more conserved in FT compared to TFL1. The first residue of the triad is leucine (L) or isoleucine (I), followed by tyrosine (Y) and asparagine (N) 31 . In Arabidopsis FT and soybean GmFT5a this triad is IYN. However, it is LYN for GmFT2a and GmFT7. Our earlier RNA-seq experiment highlighted transcriptional dynamics of leaf and the shoot apical meristem of soybean during different short-day treatment 28 . Our data showed that among the six soybean FTs identified,  www.nature.com/scientificreports/ GmFT7 displayed the highest expression in leaf following two short-day treatment (Fig. 4). To confirm the dynamics of GmFT7 expression during the floral transition, we performed RT-PCR analysis using leaf and shoot apex samples collected at SD0 (0 short-day treatment) to SD4 (4 short-day treatment) at a 4 h interval. Four short-day treatment was chosen based on the appearance of AP1 expression (a floral identity gene) in the SAM 32 .
RT-PCR analysis showed that GmFT7 mRNA had a consistent expression in leaf and shoot apex tissues in a 24 h cycle upon short-day treatment (Fig. 5). These results were consistent with the previous study, which suggested that GmFT7 is ubiquitously expressed across different tissues 11 . In contrast, it has been shown that GmFT2a and GmFT5a mRNA expression follow circadian rhythm 12 .
Cis-elements analysis. It is well established that cis-regulatory elements play essential roles in regulating gene expression. In eukaryotes, temporal and spatial gene expression is governed by the binding of transcription factors to cis-elements 33 . To compare the cis-elements between promoters of GmFT7 and other flowering regulators such as GmFT2a, GmFT5a and repressor GmFT4, light-or circadian clock-related cis-elements were identified using NewPLACE webserver in 1.8 kb promoters upstream of ATG (Fig. 6). Unlike the FT promoter in Arabidopsis where a distal (approximately 5 kb upstream of ATG) promoter region contained critical elements for FT activity 34 , GmFT7 had a promoter region of around 1.8 kb as determined by the presence of an upstream gene Glyma.02G069400.1 (Supplementary Fig. S3). Twelve main cis-elements related to light or circadian-clock are summarised in Table 1. These include GATA, CIRCADIAN, GT1, I-Box, CCAAT, REalpha, E-Box, -10PE, INR, SORLIP1, CCA1, and ASF-1. Our analysis identified 11 out of 12 main cis-elements related to light or circadian-clock motifs in the promoter region of GmFT7 (Table 1); however, motif ASF-1 could not be located in the promoter region of GmFT7. It is worth noting that ASF-1motif only appeared once in GmFT2a and GmFT4 promoter region. Among the 12 motifs, CCAAT 35,36 and E-Box (CANNTG) 37 had been suggested to have important roles in regulating the flowering process. CCAAT motifs are recognized by NUCLEAR FACTOR Y (NF-Y) proteins and its associated DNA complex 38 . NF-Y complex has been shown to participate in the flowering process in the photoperiodic regulation of flowering 39 . Research has shown that Arabidopsis cryptochrome 2-interacting basic helix-loop-helix 1 (CIB1) interacts with the chromatin region of FT promoter, which contains only E-Box elements other than G-Box (CAC GTG ) elements to mediate flowering control 37 . The CCAAT motifs are the most frequent elements in the promoter region of eukaryotic genes typically found between 80 to 300 bp region upstream of the transcriptional start site. In this study, three CCAAT motifs were located in the proximal promoter region. Three CCAAT motifs in GmFT7 promoter were located in a region around 1100-1500 bp upstream of the transcriptional start site. In contrast, four CCAAT motifs in GmFT2a promoter were scattered between 600 to 1800 bp upstream of ATG, while five and four CCAAT motifs were found in 1300-1650 bp region and 800-1800 bp region for GmFT5a and AtFT promoters, respectively. Interestingly, there were 26 E-Box motifs in GmFT7 promoter, approximately two-fold than that of GmFT2a/5a/4 promoters.  www.nature.com/scientificreports/ E-box was found to be distributed throughout the promoter of GmFT7, while three CCAAT motifs exist in a 297 bp region (Fig. 6).  (Fig. 7b,c). Further, transgenic lines harbouring GmFT7 retained the late-flowering phenotype of Arabidopsis ft-10 mutant background regardless of photoperiod (Fig. 7b,c), indicating that GmFT7 could not complement the late-flowering ft-10 mutants.

Over-expression of
Over-expression of GmFT7 in soybean. Transgenic soybean plants over-expressing GmFT7 were generated to investigate the function of GmFT7 in the floral transition of soybean. The transgenic soybean plants were produced using Agrobacterium-mediated transformation described by Li et al. 40 using glufosinate as a selectable agent. The transgenic status of the plants (T2 generation soybean) was confirmed by genomic PCR and expression of the Bar, and GmFT7 genes were confirmed using RT-PCR (Fig. 8a,b). These experiments showed that the transgenic plants were harbouring and expressing the introduced genes (Fig. 8a,b). Further, GmFT7 expression levels in these transgenic plants were determined using quantitative PCR. The qPCR results showed elevated expression of GmFT7 in all the transgenic lines compared with WT plants (Fig. 8c). These plants were grown along with WT and flowering time was identified by recording the day of the first flower appearance. The flowering time for the transgenic plants and WT was recorded (Fig. 8d). The statistical analysis showed no significant difference (P > 0.05) in the flowering time between transgenic (56.2 ± 4.6, 56.7 ± 3.9 and 56.4 ± 3.3 in three transgenic lines) and wild type (56.0 ± 3.0) soybean plants (Fig. 8d). It is well established that different flowering pathways are integrated by downstream genes such as LFY (LEAFY), SOC1 (SUPPRESSOR OF CONSTANS1), FT (FLOWERING LOCUS T). These integrators convey the outcomes to floral meristem identity gene, AP1 (APETALA 1) to evoke the floral initiation process at the shoot apical meristem. Accordingly, we studied the expression of GmFT2a, GmFT5a, GmSOC1, GmLFY and GmAP1 Figure 6. Light-and circadian clock-related cis-regulatory elements in GmFT promoters (1841 bp upstream of ATG). The cis-regulatory elements were identified by searching New PLACE database (https ://www.dna.affrc .go. jp/PLACE /?actio n=newpl ace). Distribution of cis-regulatory elements was generated using TOUCAN (http:// touca n.aerts lab.org/softw are/touca n.php) software.  Fig. 8c. Higher expression of GmFT7 was observed in the transgenic soybean lines. However, the expression of floral identity genes GmAP1 and GmLFY2 remained unaltered in the GmFT7 over-expression lines (Fig. 8c).

Discussion
Soybean is an important legume crop used for food and feed. Since soybean cultivars have a narrow latitude band that keeps it from adapting to the higher or lower latitudinal area, it would be desirable to breed soybean varieties adapted to a broader region. Such adapted lines would assist in increasing the soybean production to meet growing worldwide demand. Under the current changing climate scenario, it is especially crucial to ensure that soybean plants flower at the optimal time since flowering is a prerequisite for its seed production. Among the environmental factors that influence flowering, photoperiod is the most well-investigated. Several studies recently investigated soybean floral transition via bioinformatics and functional genomic approaches highlighting the roles of histone modifiers 41 or circadian clock or photoperiod-related genes 42,43 . A comprehensive understanding of flowering signals such as FT, the florigen in the plant, could provide the crucial information on how legume plant like soybean coordinate both external and internal cues for the flowering process. Soybean is a paleopolyploid crop plant with two genome duplications occurring about 59 and 13 million years ago, leading to multiple copies of three-fourths of its protein-coding genes 44 . Investigations on flowering time regulation in legumes, including soybean, have highlighted unique and conserved features 45 . As a result of these duplication events, the soybean genome contains at least ten FT homologues in soybean. Among them, GmFT1a and GmFT4 were shown to be floral repressors 13,18 . On the other hand, GmFT2a and GmFT5a are flowering promoters 8,9,12 . GmFT7 is a truncated portion (C-terminus) of FT2c. However, the evolution of FT2c is a soybean lineage-specific event which occurred after the separation of cultivated and wild soybeans 17 .
GmFT7 shows high sequence similarity with GmFT2a, a soybean FT homologue which could promote flowering 12 . The GmFT7 only harbours the fourth exon, and it retained the critical segment B and LYN in segment C 12,31 . Prediction of subcellular localization showed that GmFT7 located in the same compartment as other FT homologues. Our analysis of the regulatory elements in its promoter region indicated similarities between other FTs-our data showed all the 12 cis-elements related to light or circadian-clock are present in the promoter region of GmFT7. Especially, CCAAT 35,36 and E-Box that have been shown to play essential roles in flowering regulation. Though C-terminus of GmFT7 resembles GsFT2c (Supplementary Fig. S1), GsFT2c could promote flowering in wild-type Arabidopsis and rescue the ft-10 mutant 17 . In comparison, GmFT7 could not promote flowering in wild-type Arabidopsis or compliment the late flowering ft-10 mutants under inductive SD or noninductive LD conditions. Transgenic soybean plants also showed that the overexpression of GmFT7 could not promote flowering compared to wild type soybean plants. However, qPCR results showed that relative expression of GmFT2a, GmFT5a and GmSOC1b were decreased in transgenic soybean lines, but further work is required to show other possible players in the flowering pathways. To this end, CRISPR generated soybean plants would help unravel the floral network controlling the transition to flowering in soybean. However, our results indicate that GmFT7, which only carries the fourth exon, could not function as a floral signal in the soybean flowering www.nature.com/scientificreports/ process. We speculate that all the four exons in an FT homologue are required to promote the floral transition in soybean. It is also possible that due to soybean genome duplications and evolution GmFT7 acquire the new role that needs further study. Total RNA was isolated from leaf and shoot apex tissues using TRIzol reagent (Invitrogen, USA) as per the manufacturer's instructions. The quality of RNA was examined by NanoDrop 1000 (Thermo Scientific, USA) and gel electrophoresis. According to the manufacturer, 1 μg of total RNA was used for first-strand cDNA synthesis using SuperScript III Reverse Transcriptase (Thermo Fisher, USA) in a 20 μL reaction's protocol. After reverse transcription, 80 μL of nuclease-free water was added to dilute the cDNA. 2 μL was then used for PCR reaction. GmActin was used as a reference gene.

Methods
Bioinformatics analysis. Soybean FT homologues and Arabidopsis FT protein sequences were retrieved from Phytozome (https ://phyto zome.jgi.doe.gov/pz/porta l.html; version 12.0) database. Sequence alignment was performed using MUSCLE 46 program in MEGA7 47 with default parameters. The phylogenetic tree was inferred using the Likelihood method with cpREV + G substitution model; partial deletion was used for Gaps/ Missing data treatment. Reliability was examined using the Bootstrap method with 1000 replicates. A phylogenetic tree was drawn using FigTree (http://tree.bio.ed.ac.uk/softw are/figtr ee/; version 1.4.3). Theoretical pI, molecular weight, instability, and GRAVY values were calculated using the ProtParam tool (https ://web.expas www.nature.com/scientificreports/ y.org/protp aram/). Protein subcellular localization was analyzed using a predictive tool, ProtComp 9.0 (http:// www.softb erry.com). Gene structures of soybean and Arabidopsis FT genes were displayed by Gene Structure Display Server (GSDS) 2.0 48 . Cis-regulatory elements were examined using NewPLACE web server 49 . Location of cis-elements was highlighted using TOUCAN 2 50 . Protein sequences of GmFT1a/2a/4/5a/7 and AtFT were aligned using Clustal Omega 51 . The alignment was imported to GeneDoc 52 (version 2.7.000) for shading.

Plant constructs and transformation.
For gene expression construct, the coding sequence of Glyma02g07650 (GmFT7) was amplified (with XhoI and BamHI restriction enzyme sites on 5′ and 3′ end, respectively) and ligated to pGEM-T Easy vector (Promega). The vector was digested, and gene fragment was ligated to pRT101 at Xho I and Bam HI sites, to obtain 35S promoter and polyA terminator. Then the 35S:GmFT7:polyA cassette was excised using Hind III and ligated to pUQC10255 expression vector (Supplementary Fig. S4). Transformation of Arabidopsis Col (WT) and ft-10 mutant was performed using the floral dip method 53 .
Bragg cultivar was used for soybean transformation following the protocol by Li et al. 40 with modifications; selection on SIM medium containing 4 mg/L glufosinate (Sigma, USA) was started during the second subculture step SIM and continued till SEM medium. 100 mg/L Cefotaxime (GoldBio, USA), 50 mg/L Vancomycin (GoldBio, USA) and 50 mg/L Ticarcillin (GoldBio, USA) were used instead of carbenicillin on SIM, SEM and RM medium. Five independent experiments were conducted using an average of 110 seed explants per experiment and recovered five transgenic shoots. However, one-line produced seeds. The transgenic seeds were multiplied and selected for the Bar gene. Three T2 lines were used for analysis.

Confirmation of transgenic soybean plants.
Trifoliate leaf (21st day after sowing) was collected and used for genomic DNA and total RNA extractions. Genomic PCR and RT-PCR were performed to detect Bar gene and GmFT7 expression in the putative transgenic soybean plants. Genomic DNA was extracted using the CTAB method (https ://agric ultur e.uq.edu.au/files /5650/plant -genom ic-dna-extra ction -by-ctab-2-fiona .pdf). 100 ng of genomic DNA was used for PCR reaction. Genomic PCR and RT-PCR were performed at 94 °C for 3 min, followed by 35 cycles of 94 °C for 45 s, 50 °C for 30 s and 72 °C for 60 s, with a final extension for 10 min at 72 °C. Supplementary Table S1 lists all the primers used in this study.
Flowering time determination. Flowering time in Arabidopsis was determined by counting rosette leaves both under LD and SD conditions. Ten independent Arabidopsis transgenic lines (5 plants from each independent line) were used for analysing transgenic plants, while 12 Arabidopsis plants were used as control (12 Col-0 plants and 12 ft-10 mutant plants). Soybean plants were grown under LD for 4 weeks before shifting to SD conditions. Flowering time in soybean was identified by recording the day of the first flower appearance. Three T2 lines (8 plants per line) were used for counting the number of days to first flower appearance. Significant differences were tested using Student's t-test in Microsoft Excel.
Quantitative PCR (qPCR). Five independent plants for each of Col-0, ox-GmFT7 in Col-0 and ox-GmFT7 in ft-10 were used for qPCR analysis with two technical replicates. For soybean qPCR analysis, two biological replicates of trifoliate leaf and shoot apex tissues (pool of tissues from three plants were counted as one biological replicate) and two technical replicates were used. Total RNA isolated from a trifoliate leaf from plants (21st day after sown) were used for the relative expression of GmFT7, GmFT2a and GmFT5a, while total RNA extracted from shoot apex were used for examining the expression of GmAP1, GmSOC1b and GmLFY2. qPCR was performed using Brilliant III Ultra-Fast SYBR Green QPCR Master Mix (Agilent Technologies, USA) on a Stratagene Mx3005P qPCR platform (Agilent Technologies, USA). 2 µL cDNA was used as a template in a 10 µL qPCR reaction (5 µL 2xSYBR Green Master Mix, 0.3 µL of each primer (final concentration of 300 nM), 0.3 µL diluted (1:500) reference dye and 2.1 µL nuclease-free water). The qPCR program was 95 °C for 3 min, followed by 40 cycles of 20 s at 95 °C and 20 s at 60 °C, then finished with 1 min at 95 °C, 30 s at 55 °C and 30 s at 95 °C for melting curve. Primers (Supplementary Table S1) were retrieved from previous studies (GmTUB, GmFT2a, GmFT5a, GmAP1, GmSOC1b, and GmLFY2) 14 . GmTUB and AT4G34270 54 were used as reference genes for soybean and Arabidopsis plants, respectively. Relative gene expression was calculated using 2 −ΔΔCT method 55