For the diurnal rhythm analysis of GmIDD gene, seeds from “DongNong 42,” a photoperiod sensitive soybean variety, were provided by Soybean Research Institute of Northeast Agricultural University (Harbin, China). Seeds of “DongNong 42” were planted in a greenhouse at 25°C with 250 μmol m^–2 sec^–1 white light under LDs (16 h/8 h light/dark). When the first trifoliate leaves were expanded, part of the seedlings were transferred to SDs (8 h/16 h light/dark) under the same temperature regime. Seedlings were cultured under LDs and SDs for 30 days (transferred at 15 days) and sampled at 3 h intervals for 24 h, and sampled under continuous light (LL) and dark (DD) conditions for 48 h, and then immediately frozen in liquid nitrogen. Samples of different tissues (including roots, stems, leaves, flowers, pods, and seeds) of plants cultured under LDs and SDs were collected for tissue specific expression analysis of GmIDD gene.

In this study, Arabidopsis thaliana (Col-0) was used as control and the background plant of genetic transformation. Seeds of the idd mutant (SALK_129969c) were obtained from the Arabidopsis Biological Resource Center (ABRC, Columbus, United States). The seeds of GmIDD-overexpression (GmIDD-ox), idd mutant, 35S:GmIDD/idd restoration and wild-type (WT) Arabidopsis were surface sterilized with 10% hypochlorite and then planted on MS agar medium. The seeds were placed at 4°C for 72 h and transferred to room temperature (22°C). The 10-day-old seedlings were transplanted into 1:1 vermiculite turbidite soil and cultured under LDs and SDs for flowering phenotype analysis. WT and GmIDD-ox transgenic Arabidopsis plants were cultured under LDs on MS agar medium for 15 days and sampled for quantitative real-time PCR (qRT-PCR) analysis of flowering time-related genes. These experiments were performed in three biological replicates.

For the FLAG and HIS tag construct, we first synthesized the tandem repeats of 3 × FLAG and 6 × Histidine (3F6H) tags with NotI at 5′ end and XbaI at 3′ end (5′-GCGGCCGCCCTGGAGCTCGGTACCCGGG(SmaI)GATCCC AGGATCTGATTACAAGGATCATGATGGTGATTACAAGG ATCACGACATCGACTACAAGGATGACGATGACAAGCAC CATCATCACCACCATTGATCTCTAGA-3′, the sequences encoding 3F6H tag were in bold) (Song et al., 2012). The synthesized products above were cloned into NotI-XbaI sites of pENTRY vector (named pENTRY-3F6H) which contained the 3F6H sequence in the C terminus of the cloning site, and the sequences were verified. The full-length coding region of GmIDD was amplified from total RNA of “Dongnong 42” using GmIDD-3F6H-F and GmIDD-3F6H-R primers (Supplementary Table S1). The PCR product was purified and cloned into pENTRY-3F6H vector linearized by SmaI using In-Fusion cloning system (Clontech, United States) to construct recombinant vector 35S:GmIDD-3F6H-pENTRY. LR reaction was conducted by 35S:GmIDD-3F6H-pENTRY and pB7WG2 to generate 35S:GmIDD-3F6H-pB7WG2 fusion expression vector. The recombinant vector 35S:GmIDD-3F6H -pB7WG2 was introduced into the Agrobacterium GV3101 which used to transform Arabidopsis thaliana (Col-0) and idd mutant using the floral dip method with Agrobacterium tumefaciens strain GV3101 (Clough and Bent, 1998). Transformants were selected on MS agar medium with 8 mg/L phosphinothricin. T₃ transgenic seeds of two homozygous lines were selected for further study.

A cDNA fragment of GmIDD was amplified by PCR with GmIDD-TOPO-F and GmIDD-TOPO-R primers (Supplementary Table S1) from total RNA of “DongNong 42” and cloned into pENTR/D-TOPO (Life technologies) and then transferred to the expression vector pGWB506 through LR reaction to generate the 35S:GFP-GmIDD fusion vector. The recombinant construct was introduced into Agrobacterium GV3101 and subsequently transformed into N. benthamiana (Sheikh et al., 2014). After infiltration, the tobacco leaves were grown for 2 days and the GFP signal was detected by fluorescence microscopy.

One representative GmIDD-ox-1 lines were selected for ChIP-seq. GmIDD-ox-1 transgenic Arabidopsis plants were cultured on MS agar medium under LDs for 14 days, and then 1 g seedlings were quick-frozen in liquid nitrogen. With the purpose of minimizing the deviation between different parallel samples under the same treatment, the sample was retrieved from independent three plots of Arabidopsis seedlings. After crushing, the seedlings were fixed with 1% (V/V) formaldehyde for 15 min at 4°C. Final concentration of 0.125 M glycine was subsequently added to quench the cross-linking reaction. After nuclei were isolated and lysed, and the chromatin solution was then sonicated to approximately 200–1,000 bp DNA fragments. Immunoprecipitation reactions were performed using Anti-Flag antibody (Monoclonal ANTI-FLAG^® M2 antibody produced in mouse, F1804, Sigma-Aldrich). The complex of chromatin antibody was captured with protein G beads (Invitrogen), and DNA was purified using a QIAquick PCR purification kit (QIAGEN). The purified DNA was sent to a sequencing company for library construction and deep sequencing. The ChIP-seq sequencing project was completed on Illumina sequencing platform. Illumina PE library (∼300 bp) was constructed for sequencing and quality control of the obtained sequencing data. Bioinformatics methods were then used to analyze ChIP-seq data. The raw sequence data were aligned to the Arabidopsis thaliana(TAIR10)¹. MACS was used to predict the length of the protein binding sequence through modeling, and the relative abundance of the corresponding peak of the sequence was determined by the length of the sequence and the numbers of pair end reads mapped on the sequence. ChIP-seq was completed by SeqHealth Tech Co., Ltd., Wuhan, China.

ChIP-qPCR was performed using the whole WT and GmIDD-ox-1 seedlings. The relative enrichment of WT was set to 1. IPP2 gene (isopentenyl pyrophosphate: dimethyl allyl pyrophosphate isomerase 2, AT3G02780) was used as a negative control. DNA samples were analyzed by real-time quantitative PCR using the appropriate DNA primers (Supplementary Table S1) and SuperReal PreMix Plus (TIANGEN, Beijing, China). IgG was used as an antibody control. Data are shown as the mean ± SD of three biological replicates.

To generate AGL18 promoter driven LUC constructs proAGL18:LUC, promoter DNA was amplified from genomic DNA of Arabidopsis (Col-0) using proAGL18:LUC-F and proAGL18:LUC-R primers (Supplementary Table S1). The PCR product was purified and cloned into binary vector pGreenII-0800-LUC linearized by SmaI using In-Fusion cloning system. The recombinant constructs were introduced into Agrobacterium GV3101 and subsequently transformed into N. benthamiana (Sheikh et al., 2014). The transient activity of recombinant vectors was assayed using dual-luciferase assay kit (Promega, United States) and Multiscan Spectrum (TECAN Infinite 200 PRO, Männedorf, Switzerland). The construct 35S:GmIDD-3F6H-pB7WG2 and proAGL18:LUC were simultaneously transferred into N. benthamiana to measure transient assay of the AGL18 promoters affected by GmIDD protein. A floral repressor GmRAV effector construct (35S:GmRAV-3F6H-pB7WG2) and proAGL18:LUC were simultaneously transferred into N. benthamiana as the negative control of this experiment. The recombinant vector 35S::GmRAV-3F6H-pB7WG2 was obtained and introduced into Agrobacterium GV3101 by referring to the construction method of 35S::GmIDD-3F6H-pB7WG2 above. Expression vector pB7WG2 and proAGL18:LUC was simultaneously transferred into N. Benthamiana as the blank control of this experiment. Three independent experiments were performed and each experiment was repeated three times to obtain reproducible results. The luminescence signal was captured using Amersham Imager 600 (General Electric Company)² after spraying 1 mM luciferin (Heliosense)³ on N. benthamiana leaves.

RNA isolation has been described previously (Zhao et al., 2013). qRT-PCR amplifications were performed using the SuperReal PreMix Plus (TIANGEN, Beijing, China) on Applied Biosystems^TM 7500 Fast Dx Real-Time PCR Instrument (ABI). The PCR cycling conditions were as follows: 95°C for 15 min; 40 cycles of 95°C for 10 s, 60°C for 20 s, and 72°C for 30 s. GmActin4 (GenBank accession number AF049106) and IPP2 (AT3G02780) were used as endogenous regulatory genes of soybean and Arabidopsis, respectively. Three biological replicates and three technical replicates were applied for the whole assays. The primers used in qRT-PCR analyses were shown in Supplementary Table S1.

Flowering time of Arabidopsis was determined by scoring the numbers of rosette leaves and the numbers of days from germination to bolting time. At least 20 plants were analyzed each time, and the analysis was repeated for three times.

We have identified that both genes such as GmRAV and GmGBP1 induced by SD detected by suppression subtractive hybridization (SSH) functioned in regulating flowering time (Zhao et al., 2008, 2018). In this study, we further detected a zinc finger transcription factor GmIDD (Glyma.14G095900) gene also induced by SD from SSH library. The GmIDD sequence information was obtained from soybean genome database⁴. The cDNA sequence of GmIDD is 2,034 bp and contains 1,227 bp open reading frame, which encoded 408 amino acids with predicted molecular mass of 45,129 kDa. The domain analysis showed that GmIDD had a typical ID domain, which consisted of a nuclear localization signal motif (KKKR), four zinc finger domains including two types of C2H2 (72–92 aa and 114–142 aa), and two types of C2HC (149–169 aa and 176–195 aa) (Figure 1A). GmIDD showed high homology with AtIDD8, ZmID1, and Ehd2, which are typical IDD family members in Arabidopsis, maize and rice (Figure 1B). Therefore, GmIDD is a member of the soybean IDD transcription factor family. A phylogenetic tree containing 10 soybean proteins with complete ID domain and 15 IDD proteins from other species was constructed by MEGA 6.0. Phylogenetic tree analysis showed that GmIDD was located on the same branch with Vigna angularis (NC_030643.1), Cajanus cajan (KYP72220.1), Vigna radiata (XP_017428159.1), Phaseolus vulgaris (Phvul.001G036500), Lupinus angustifolius (XP_019458210.1), and Medicago truncatula (Medtr1g016010.1) (Figure 1C). Soybean IDD proteins Glyma.08G192300 and Glyma.15G024500 were located on the same branch with Nelumbo nucifera (xP_010242234.1) and Juglans regia (xP_018848593.1) (Figure 1C). Soybean IDD protein Glyma.07G158200 was located on the same branch with Zea Mays (GRMZM2G320287) and Oryza sativa (LOC_Os02g31890) (Figure 1C). Therefore, the evolution distance between GmIDD and other soybean IDD proteins was relatively far, indicating that soybean IDD protein might have certain differences in function.

The subcellular localization of the GmIDD protein might be crucial for its function. The expression of 35S:GFP-GmIDD fusion protein under the control of the 35S promoter in tobacco mesophyll cells showed that the GmIDD fusion protein was concentrated in the nucleus (Figure 1D), whereas GFP was dispersed throughout the entire tobacco mesophyll cells in the 35S:GFP control. The results clearly showed that GmIDD was a nuclear-localized protein.

The trifoliate leaves of soybean grown for 30 days (transferred at 15 days) were collected every 3 h to show the diurnal expression patterns of GmIDD in LDs and SDs by qRT-PCR analysis. GmIDD exhibited photoperiod-specific expression patterns under both LDs and SDs. The expression levels of GmIDD mRNA under SDs were significantly higher than those under LDs (Figure 2A). To further determine GmIDD gene expression patterns, we analyzed GmIDD expression under LL and DD conditions after SDs and LDs transfer. GmIDD mRNA level maintained a strong rhythm under SDs-DD and SDs-LL, reaching the peak at 12 h after dawn (Figure 2B), but no strong cycling was detected under LDs-LL (Figure 2C). Therefore, GmIDD gene was induced by SDs in soybean leaves, and regulated by circadian clock.

Samples of different tissues (including roots, stems, leaves, flowers, pods, and seeds) of plants cultured under LDs and SDs were collected at 12 h after dawn for tissue specific expression analysis of GmIDD gene to determine the expression pattern of GmIDD gene during soybean growth and development. The GmIDD mRNA was present in all organs examined, which included roots, stems, leaves, flowers, pods, and seeds (Figure 2D). The mRNA abundance of GmIDD was the highest in pods among all organs under both SDs and LDs. All the mRNA abundance of GmIDD in leaves, stems, pods, and seeds in SDs was higher than that in LDs. However, the mRNA abundance of GmIDD in roots in SDs plants was lower than that in LDs. These expression patterns were different from those of AtIDD8, which was expressed at a relatively high level in vegetative organs, but at a lower level in pods (Seo et al., 2011). Although GmIDD and AtIDD8 have high amino acid homology and belong to IDD transcription factor family, there may be some differences in biological function.

In order to verify the function of GmIDD in the control of flowering time, the GmIDD gene was genetically transformed into WT Arabidopsis under the regulation of cauliflower-mosaic virus (CaMV) 35S promoter to obtain GmIDD-ox plants. We observed the flowering phenotype of WT Arabidopsis, GmIDD-ox, idd mutant plants under LDs and SDs. Under LDs, the flowering time of GmIDD-ox plants were significantly promoted for 3 days, and the idd mutants were significantly delayed for 4 days (Figures 3A,B). In addition, the total number of rosette leaves of GmIDD-ox plants and idd mutants was fewer and more than that of WT plants at bolting, respectively (Figure 3B). Under SDs, the flowering time of GmIDD-ox plants were significantly promoted for 5 days, idd mutants showed no significant difference (Figures 3A,C). A complementation experiment was conducted on idd mutants to further determine the roles of GmIDD in promoting flowering time. The phenotype of idd mutants was rescued by the expression of a 35S:GmIDD fusion gene under the control of the 35S promoter. The results indicated that overexpression of GmIDD could promote flowering in Arabidopsis.

ChIP-seq was performed to identify the DNA binding sites and target genes of GmIDD in overexpressed transgenic Arabidopsis, and further elucidate the potential mechanism of GmIDD in promoting flowering. Among the 588 GmIDD-binding sites detected by ChIP-seq, 446 (75.85%) were located in genic regions. These binding sites comprised gene bodies and their flanking regulatory sequences, including 2 kb upstream regions that were assumed to harbor promoter regions and 2 kb downstream regions that were assumed to contain terminator regions. Among the 446 sites in genic regions, 32.51, 48.88, and 18.61% were located in defined promoter regions, gene bodies, and terminator regions, respectively (Figure 4A).

AGL18 (AT3G57390) related to flowering time might be a GmIDD candidate target. The MADS-domain factor AGL18 acted redundantly as a repressor of the floral transition in Arabidopsis (Adamczyk et al., 2007). AGL18 acted upstream of the floral integrator FT, and a combination of agl18 and agl15 mutations partially suppressed defects in the photoperiod pathway. The GmIDD binding site in AGL18 gene regions was located in the promoter region (Figure 4B). According to the consensus sequences at the detected GmIDD-binding sites, putative GmIDD-binding motifs were predicted using the Hypergeometric Optimization of Motif EnRichment (HOMER) software (Heinz et al., 2010). Based on the prediction, we identified that GmIDD combined with target gene AGL18 motif [(T/G/C/A)(T/G/A)(A/G/C/T)(A/T)(G/T)(T/G)(G/A/T)(C/G/T) (T/C/G)(T/C/G/A)] (P-value = 1e-6) (Figure 4C). The previous study reported that AtIDD8 regulated SUS4 by binding with the conserved CTTTTGTCC motif of its promoter (Seo et al., 2011). Sequence analysis of AGL18 promoter showed that the AGL18 promoter contained the TTTTGGTCC motif, which was similar to the CTTTTGTCC motif of AtIDD8 combined with SUS4 promoter.

The identified GmIDD-binding sites were further validated by ChIP–qPCR. The 14-day-old GmIDD-ox-1 transgenic lines were used to perform ChIP-qPCR to verify potential GmIDD-binding sites. Anti-Flag antibody was used to precipitate chromatin prepared from GmIDD-ox-1 transgenic lines. IgG was used as an antibody control. The fragments harboring identified GmIDD-binding site in the promoter region of AGL18 was highly enriched in DNA chromatin-immunoprecipitated (ChIPed) with Anti-Flag in ChIP-qPCR experiments, indicating GmIDD bound to the promoter of AGL18 (Figure 4D).

It was found that the AGL18 promoter was bound by GmIDD using ChIP-qPCR, and its expression levels was decreased in GmIDD-ox-1 Arabidopsis plants using qRT-PCR (Figure 4E). The reporter proAGL18:LUC was constructed by AGL18 promoter containing GmIDD-binding site TTTTGGTCC motif driving LUC reporter gene. When co-infiltrating Agrobacterium expressing 35S:GmIDD-3F6H-pB7WG2 effectors together with the proAGL18:LUC reporters into tobacco leaves, the activity of LUC was significantly lower than that of blank control of pB7WG2 and proAGL18:LUC co-transformed tobacco leaves (Figures 4F,G), thus demonstrating that GmIDD could inhibit the transcriptional activation activities of AGL18 gene. However, the LUC activity of the floral repressor GmRAV (Zhao et al., 2008) and proAGL18:LUC co-transformed tobacco leaves showed no significant difference compared with the blank control. Together, our results suggested that GmIDD protein significantly repressed the expression of AGL18 by directly binding to its promoter.

Overexpression of GmIDD shortened the flowering time of Arabidopsis. The expression levels of flowering time-related genes (including CO, FT, FLC, SOC1, LFY, and AP1) in the 15-day-old GmIDD-transgenic seedlings were further investigated to determine the mechanism of GmIDD regulating flowering time. The results showed that the mRNA levels of FT, floral homeotic gene AP1, floral meristem identity gene LFY and floral integrator gene SOC1 were obviously increased in the GmIDD-ox transgenic plants compared to WT plants, but they were significantly decreased in idd mutants. In 35 S:GmIDD/idd complementary lines, the down-regulated trend of FT, AP1, LFY, and SOC1 genes were rescued (Figures 5B,D–F). The expression of CO, FLC and the other genes functioning in the autonomous pathway showed no significant difference compared to WT (Figures 5A,C). FLC is a floral repressor of the vernalization and autonomous pathway and GmIDD might not affect flowering time through the autonomous pathway. Taken together, GmIDD inhibited the transcriptional activities of AGL18 gene, and induced the expression of FT gene to promote flowering in Arabidopsis.