Soybean adaption to high‐latitude regions is associated with natural variations of GmFT2b, an ortholog of FLOWERING LOCUS T

ABSTRACT Day length has an important influence on flowering and growth habit in many plant species. In crops such as soybean, photoperiod sensitivity determines the geographical range over which a given cultivar can grow and flower. The soybean genome contains ~10 genes homologous to FT, a central regulator of flowering from Arabidopsis thaliana. However, the precise roles of these soybean FTs are not clearly. Here we show that one such gene, GmFT2b, promotes flowering under long‐days (LDs). Overexpression of GmFT2b upregulates expression of flowering‐related genes which are important in regulating flowering time. We propose a ‘weight’ model for soybean flowering under short‐day (SD) and LD conditions. Furthermore, we examine GmFT2b sequences in 195 soybean cultivars, as well as flowering phenotypes, geographical distributions and maturity groups. We found that Hap3, a major GmFT2b haplotype, is associated with significantly earlier flowering at higher latitudes. We anticipate our assay to provide important resources for the genetic improvement of soybean, including new germplasm for soybean breeding, and also increase our understanding of functional diversity in the soybean FT gene family.

Soybean, Glycine max (L.) Merr., is a typical short-day plant that is sensitive to seasonal changes in day length. Day length has an important influence on soybean flowering and growth habit, and responses to photoperiod sensitivity limit the geographical ranges of soybean cultivars. At present, at least four major flowering pathways, including the vernalization, autonomous, gibberellin (GA) and photoperiod pathways are known to regulate the floral transition process (Cai et al., 2019). At least ten FT gene homologs have been identified in soybean (Kong et al., 2010). These FT genes are the result of tandem and whole genome duplications in soybean (Thakare, Kumudini, & Dinkins, 2011;Wang et al., 2015). Among these genes, GmFT2a and GmFT5a have been found to promote flowering following the expression of the phytochrome PHYA in soybean, and show the same relationship as AtFT and AtTSF (TWIN SISTER of FT) in Arabidopsis (Nan et al., 2014;Takeshima et al., 2016;Yamaguchi, Kobayashi, Goto, Abe, & Araki, 2005). Expression of both GmFT2a and GmFT5a is up-regulated under SD conditions, whereas they are significantly down-regulated under long-day (LD) conditions (Kong et al., 2010;Sun et al., 2011;Xu et al., 2013). In a previous study, we used the CRISPR/Cas9 (Clustered regularly interspaced short palindromic repeat/CRISPRassociated 9) genome editing system to specifically induce targeted mutations in GmFT2a in the soybean cultivar 'Jack.' Homozygous ft2a mutant plants exhibited the late flowering phenotype under both LD and SD conditions . GmFT4 is known to act as a floral repressor, and its function is similar to that of TFL1 (TERMINAL FLOWER 1) in Arabidopsis (Zhai et al., 2014). Transgenic overexpression of GmFT1a delayed flowering and maturation in soybean, confirming that GmFT1a also functions as a flowering inhibitor (FI;Liu et al., 2018). The results suggest that FT proteins have undergone functional divergence in soybean. Currently, the results related to the homologous FT genes (GmFT2a, GmFT5a, GmFT1a and GmFT4) are not sufficient to explain the complex flowering mechanism in soybean.
Further studies of other FT genes are urgently needed to better understand the functions of the 10 FT genes in the soybean genome.
In this study, we isolated a soybean FT homolog, GmFT2b, that is highly homologous to GmFT2a. We obtained transgenic overexpressing lines and also produced lines carrying mutations in GmFT2b using The results of our study contribute important genetic information and provide tools for the genetic improvement of soybean such as new germplasm for soybean breeding, especially in high-latitude regions.

| Plant materials and growth conditions
The soybean cultivars 'Zigongdongdou' ('ZGDD') and 'Jack' were used in the present study. The cultivar 'ZGDD' was used for gene cloning.
The cultivar 'Jack' was used for Agrobacterium-mediated transformation. Soybean seeds were germinated and the seedlings grown in a controlled culture room at 28 C under LD (16 hr light/8 hr dark) and SD (12 hr light/12 hr dark) conditions.

| GmFT2b cDNA cloning
Total RNA was extracted using Trizol reagent from the trifoliolate leaves of soybean cv. 'ZGDD' seedlings. First-strand cDNA was synthesized with Superscript II reverse transcriptase (TransGen Biotech, Beijing, China) and used as a template for further GmFT2b cDNA cloning.
Amplification was performed via PCR using KOD-plus-Neo DNA polymerase (Toyobo, Tokyo, Japan). The sequences of the primers used for amplifying the full-length GmFT2b cDNA are given in Table S1.

| Subcellular localization of GmFT2b
The open reading frame (ORF) of GmFT2b was fused with the 5 0 end of the GFP gene sequence in a construct under control of the CaMV 35S promoter. The GmFT2b gene was cloned into the p16318 plasmid . The recombinant fusion plasmids were introduced into onion epidermal cells by particle bombardment using a CaMV 35S:eGFP vector as control. Transformation was achieved with a PDS 1000/He device (BioRad, Hercules, CA, USA), with a 6 cm shot distance, 25 mmHg vacuum and 1,100 psi rupture disc pressure. Green fluorescent protein (GFP) fluorescence was monitored using a Zeiss LSM710 confocal microscope (Carl Zeiss, OKO, Germany).  Table S1. Statistical analyses were performed using Microsoft Excel. The two asterisks represent significant differences at p < .01, The one asterisk represents significant differences at p < .05.
For construction of the CRISPR/Cas9 expression vector, a 20-bp sgRNA sequence was designed using the web tool CRISPR-P (http:// cbi.hzau.edu.cn/crispr/), and its expression was driven by the Arabidopsis U6 gene promoter. The Cas9 sequence was inserted downstream of the CaMV 2X 35S promoter. The bar gene driven by a CaMV 35S promoter was used as a screening marker. A pair of DNA oligonucleotides for the sgRNA were synthesized by TSINGKE (Beijing) and annealed to generate a double-stranded sgRNA, which was subsequently inserted into the CRISPR/Cas9 expression vector (Cai et al., 2019).

| Flowering time measurements and statistical analyses
The flowering time of each soybean plant was recorded as days from emergence to the R1 stage (the time at which the first flower appears at any node on the main stem; Fehr & Caviness, 1977). GmFT2b (Glyma.16g151000) is located on soybean chromosome 16 and is close to GmFT2a (Glyma.16g150700). GmFT2b shares 90.91% amino acid sequence identity with its paralog GmFT2a  Figure 6a). The haplotypes of the GmFT2b promoter were also for analysis ( Figure S4). Four haplotypes were identified in GmFT2b promoter, which mainly corresponded to the four haplotypes in the coding region of GmFT2b. We further investigated the maturity groups   I G U R E 5 Expression patterns of GmFT genes and flowering-related genes in WT plants and ft2b mutants under LD and SD conditions. RNA was extracted from trifoliate leaves and the shoot apex at 30 (LD) and 15 DAE (SD). Relative transcript levels were quantified by qRT-PCR and normalized to GmActin expression. Average values ± SE (standard error) for three replications are shown for each data point. *, the ft2b mutants exhibit highly significant early flowering (p < .05) GmFT2b haplotypes Hap1 and Hap4 were unable to flower in Heihe due to the high latitude (N50 ).
We also analysed the geographical distribution of varieties with major GmFT2b haplotypes. Hap1 was present in a comparatively wide distribution in the southern and middle region of China. Hap2 was mostly found at higher latitude regions in the central and northern parts of the country. Hap3 occurred at higher latitudes in the northern regions. The geographical distribution of Hap4 was in the southern parts of China at low latitudes ( Figure S5). Notably, all of the varieties with Hap3 can flower in the northern latitudes, which suggests that the Hap3 genotype may contribute to early flowering at high latitudes.

| DISCUSSION
Soybean is a diploid species that evolved from an ancient tetraploid, and its genome has undergone homologous chromosomal recombination and reassortment of the entire genome during its long evolutionary history (Wang et al., 2015). At present, it has been shown that the soybean genome encodes at least ten members of the GmFT gene Based on the results of the present study and previous reports Lu et al., 2017;Yue et al., 2017), we propose a 'weight' model for soybean flowering under SD and LD conditions ( Figure 8). In this model, we consider that the transformation from vegetative growth to reproductive growth is due to the balance of flowering activators and FIs. The FIs in SDs may be less than in LDs.
Under SD conditions, the inhibition of E1 on the expression of GmFT genes is relieved by the J gene (Lu et al., 2017;Yue et al., 2017). Overexpression of GmFT2a promotes early flowering under SD conditions, but overexpression of GmFT5a or GmFT2b does not change flowering time (Cai et al., 2019). The floral activator GmFT2a is more important than GmFT5a and GmFT2b under SD conditions. GmFT2a is sufficient to overcome the effect of FIs to promote the transformation to reproductive growth. The effects of GmFT5a and GmFT2b are less obvious.
Under LD conditions, the FT genes are inhibited by E1. More flowering activators are required to overcome the increased number of FIs. All of the GmFT2a-ox, GmFT2b-ox, or GmFT5a-ox plants will flower under LD conditions (Cai et al., 2019). The ft5a mutants showed much later flowering compared to the ft2a and ft2b mutants (Cai et al., 2019). Of these, GmFT5a is more important than GmFT2a and GmFT2b under LD conditions. By investigating the nucleotide polymorphisms in the GmFT2b coding region in 195 soybean accessions, we found that the GmFT2b haplotypes are associated with flowering time. Soybean varieties carrying GmFT2b haplotype Hap3 exhibited significantly early flowering in all six environments. Most varieties with the Hap1or Hap4 haplotypes were unable to flower normally when they were grown at Heihe (higher latitude). In addition, the geographical distribution and MG F I G U R E 7 Flowering times of soybean accessions carrying the major GmFT2b haplotypes grown at six different latitudes. Each dot represents a soybean variety. DAE, days after emergence. The dotted line represents a flowering time of 130 DAE. The numbers in the dotted boxes indicate the number of soybean varieties that did not flower after 130 DAE. **, the Hap3 varieties exhibit highly significant early flowering times (p < .01) F I G U R E 8 A model for the soybean flowering transition under SD and LD conditions. Gray arrows represent stimulation of gene expression. Gray T-shaped symbols represent inhibition of gene expression. The red '×' symbols represent the release of inhibition. The 'J' gene is shown in an orange box to inhibit E1 under SD conditions. FT2a, FT2b and FT5a are in boxes of different colors and sizes, the box size represents the relative importance of the gene (larger is more important). FIs, flowering inhibitors; SD, short-day (12 hr light/12 h dark); LD, long-day (16 hr light/8 h dark) distribution of the GmFT2b haplotypes showed that Hap3 is only found in varieties grown in higher latitude regions in northeast China that belong to earlier maturing varieties from MGs 0-2. Hap4 is found in varieties grown in lower latitude regions in southern China that belong to later maturing varieties in MGs 000-8. Our previous studies showed GmFT2a also has four major haplotypes with higher frequencies (Cai et al., 2019). FT2a-Hap1 was mostly found in the Huanghuaihai. FT2a-Hap2 was mostly found in higher latitude region in the north and the Huanghuaihai. FT2a-Hap3 was present in the south. FT2a-Hap4 was comparatively wide, but was rare in the northeast. FT2a-Hap2 was mainly distributed in the varieties of MG 1, MG 2 and MG 3. The FT2a-Hap1, FT2a-Hap3 and FT2a-Hap4 genotypes were not found among the earlier maturing varieties (MG 000, MG 00 and MG 0). These results suggest that different GmFT2a and GmFT2b haplotypes have considerable effects on the diversity of flowering time in soybean at different latitudes.