Expression of a maize SOC1 gene enhances soybean yield potential through modulating plant growth and flowering

Yield enhancement is a top priority for soybean (Glycine max Merr.) breeding. SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) is a major integrator in flowering pathway, and it is anticipated to be capable of regulating soybean reproductive stages through its interactions with other MADS-box genes. Thus, we produced transgenic soybean for a constitutive expression of a maize SOC1 (ZmSOC1). T1 transgenic plants, in comparison with the nontransgenic plants, showed early flowering, reduced height of mature plants, and no significant impact on grain quality. The transgenic plants also had a 13.5–23.2% of higher grain weight per plant than the nontransgenic plants in two experiments. Transcriptome analysis in the leaves of 34-day old plants revealed 58 differentially expressed genes (DEGs) responding to the expression of the ZmSOC1, of which the upregulated FRUITFULL MADS-box gene, as well as the transcription factor VASCULAR PLANT ONE-ZINC FINGER1, contributed to the promoted flowering. The downregulated gibberellin receptor GID1B could play a major role in reducing the plant height. The remaining DEGs suggested broader effects on the other unmeasured traits (e.g., photosynthesis efficiency and abiotic tolerance), which could contribute to yield increase. Overall, modulating expression of SOC1 in soybean provides a novel and promising approach to regulate plant growth and reproductive development and thus has a potential either to enhance grain yield or to change plant adaptability.

Plant transformation. Mature seeds of soybean cultivar Jack were used to produce cotyledonary node explants for transformation. Transformation and regeneration were conducted essentially by following the protocol by Pat and Wang 39 . About 400 explants were used in each transformation. The co-cultivation time was five days. After co-cultivation, the total number of shoots produced in 4 weeks and the total number of shoots elongated in 4-14 weeks were counted. After transferred to rooting medium, the number of elongated shoots produced roots was counted. All T 0 plants were transferred to soil and grown in the greenhouse to produce T 1 seeds for phenotyping T 1 plants. Nontransgenic (NT) seedlings were used as a control. The plants were grown under a 16-h photoperiod at 21-30 °C. Phenotypic variations (e.g., date of the appearance of the first flower, the node number where the first flower appeared, the date of the appearance of the first pod, flower shape and structure, leaf color and leaf shape, plant architecture, and plant height) among the T 0 plants were recorded. Young leaves, about 0.2 g for each greenhouse-growing plant, were harvested for DNA isolation and polymerase chain reaction (PCR) analysis of the transgenes. The transformation frequency was calculated as the percentage of the inoculated explants which produced the PCR positive T 0 plants. The transformation experiment was repeated twice.

Phenotypic evaluation of T 1 and T 2 plants.
Early-and late-planted T 1 plants were evaluated in two experiments. In experiment #1, the seeds were sowed on May 04, 2020, and the plants were growing in the greenhouse and moved to a secured field under natural environment conditions at East Lansing, Michigan on June 01, 2020. In experiment #2, the seeds were sowed on June 04, 2020 and the plants were growing in the field shared with the plants of experiment #1. Depending on the availability of the seeds, 1-40 seeds per line for all the T 0 line/plant that produced seeds were sowed to evaluate T 1 transgenic plants and to produce T 2 seeds. 150 mg L −1 glufosinate ammonium (GS) was initially used to screen herbicide resistant transgenic plants by painting a half of a leaf along the midvein for each of the T 1 seedlings. GS-resistant plants were transferred to one-gallon pots (20 cm diameter × 16 cm height). DNA from the GS-resistant plants was isolated for PCR analysis of the transgenes.
For phenotypic evaluation, the data collected for each plant included: (1) Date of seed germination; (2) Date of the appearance of first flower; (3) Node position where the first flower appeared; (4) Date of the first pod set; (5) Plant height measured and branch number and pod number counted at 141 for experiment #1 and 110 day for experiment #2 after sowing; (6) Pictures taken at different developmental stages; and (7) Seed protein, oil, and fiber contents were measured using a Grain Analyser (Infratec™ 1241, FOSS Analytical AB, Denmark). Fatty acids were extracted as described by Bubeck et al. from two seeds of each plant 40 . Fatty acid composition was determined by gas chromatography.
Detection of transgenes. DNA was isolated from leaf tissues, 50-200 mg for each sample, using the cetyltrimethylammonium bromide (CTAB) method 41 . Three pairs of primers, bar-F and bar-R for the bar gene, and forward primers 35S-F (3′ portion of the CaMV 35S promoter) or MK-F and reverse primer MK_R for the ZmSOC1 gene, were used to detect the presence of transgenes in each sample (Figure S1A), and GmAct11_F and GmAct11_R primers were used as a DNA quality control to detect soybean's actin gene 42 (Table S1). PCR reaction conditions for all primer pairs started with an initial denaturation for two min at 94 °C, 30 cycles of 45 s at 94 °C, 60 s at 60 °C and 90 s at 72 °C, and a final extension for 10 min at 72 °C. All amplified PCR products were separated on 1.0% agarose gel containing ethidium bromide and visualized and photographed under UV light.
RNA sequencing and transcriptome analysis. The 5th young leaves near the shoot tips and exposed to sunshine at noon, 3-5 per plants, were harvested from 34-day old plants of three nontransgenic and six transgenic plants (i.e., 3 plants for each of the two transgenic lines) in experiment #2. The leaf samples were frozen immediately in liquid nitrogen, brought to lab, and stored at − 80 °C in a freezer for RNA isolation. Total RNA of each sample was isolated from about 500 mg leaf tissues using a CTAB method 43 and was purified using RNeasy Mini Kit (Qiagen, Valencia, CA, USA). On-Column DNase digestion with the RNase-free DNase Set was used to remove DNA in the RNA samples (Qiagen, Valencia, CA, USA). RNA quality was determined using the High  The RNA samples were sequenced using the Illumina HiSeq4000 to generate 32-42 million 150 bp-paired end reads per sample. FastQC program (http:// www. bioin forma tics. babra ham. ac. uk/ proje cts/ fastqc/) was used to assess the quality of sequencing reads for the per base quality scores. The reads with average scores greater than 38 were obtained and used for transcriptome analysis. A total of 72 million reads (MR) combined from a portion of the reads of the nine samples, 6.6-9.8 MR/sample, were assembled to develop a transcriptome reference using Trinity/2.8.5 44 . The paired reads were aligned to the transcriptome reference to estimate and the abundance for each of a single read. The differentially expressed transcripts (DETs) with the false discovery rate (FDR) value below 0.05 were identified using the Trinity command "run_DE_analysis.pl --method edgeR" 44 . Pfam proteins were used to annotate the transcriptome reference.
Cytoscape 3.8.2 was used to construct gene networks of overrepresented gene ontology (GO) terms for the selected DETs under BiNGO's default parameters with selected ontology file 'GO_Full' and selected organism 'A. thaliana' 45,46 . Statistical analysis. Statistical analysis of the phenotypic data was conducted using ANOVA and Tuk-eyHSD in RStudio (Version 1.0.136).

Results
Cloning of the ZmSOC1. The pTF101.1-ZmSOC1 contains a streptomycin/spectinomycin aminoglycoside adenylyltransferase gene (aadA) for bacterium selection and a bialaphos resistance (bar) gene under the CaMV 35S promoter conferring resistance to herbicide glufosinate (GS) for selection of transformed plant cells. In the 696-bp ZmSOC1 sequence, 694 bp are identical to the published 696-bp reference deposited in the GenBank (accession numbers HQ858775.1 and NM_001111682.1). The protein sequence of the cloned ZmSOC1 is identical to that derived from either the HQ858775.1 or a part of the NM_001111682.1. It has a 54.8% identity to the soybean's SOC1 gene (GmSOC1) (iTAK: Glyma.18G224500.2_CDS), which is the sequence detected in the soybean transcriptome reference (cultivar Jack) assembled in this study (DN17539_c0_g1) (Fig. 1B).
Phenotypic variations in T 0 transgenic plants. 'Jack' was used to produce transgenic soybean plants because it was a transformable soybean cultivar using the bialaphos resistance (bar) gene as a selectable marker 39,47 . Herbicide-resistant plants were induced for half-seed explants after 2-4-week selection, 6-12-week elongation, and 2-4-week rooting ( Figure S1). Of the total of 770 explants inoculated, 38 T 0 transgenic lines/plants were produced in two experiments with transformation frequencies of 3.9% and 5.9%, respectively (Table S2). Remarkably, three of the 38 lines flowered and further formed seed pods during in vitro cultures ( Fig. 2A,B).
Of the 38 T 0 transgenic lines growing in the greenhouse under a 16-h photoperiod, 11 lines had no visible difference from the nontransgenic control and the rest 27 lines showed various phenotypic changes, including abnormal flowers or early flowering in seven lines, chimeric or increased numbers of branches in eight lines, and dead plants of two lines ( Fig. 2C-I, Table S3; Figure S2). The lines with abnormal flowers or early flowering, which were rarely observed in our previous soybean transformations using other genes, were likely induced by the expression of the ZmSOC1, although the effect of tissue culture could also contribute to these phenotypic changes. Overall, 33 T 0 transgenic lines produced seeds, and the lines with abnormal flowers produced no or a few seeds.
Phenotypic evaluation of T 1 transgenic plants. Seedlings from 21 lines were screened in two experiments to identify transgenic plants for phenotyping. Herbicide-painting was effective in detecting the barexpressing transgenic plants ( Figure S3). Polymerase chain reaction (PCR) analyses was reliable to detect the transferred ZmSOC1 gene. Fourteen transgenic lines produced at least one transgenic seedling, and all the tested seedlings from the other seven lines were nontransgenic (Table S4). As anticipated, not all T 0 transgenic lines were able to produce transgenic seeds due possibly to the chimeric nature of the transgenic plants produced from cotyledonary node explants. Unsurprisingly, only two lines showed a segregation rate of about 3:1 between transgenic and nontransgenic seedlings in the 14 transgenic lines (Table S4).
Plants from 10 transgenic lines were grown for phenotyping in each of the two experiments, including five transgenic evaluated in both experiments (Table S4). According to the local light length (https:// www. timea nddate. com/ sun/ usa/ lansi ng), the early and late-planting experiments were conducted by sowing the seeds on May 4th and June 4th, respectively, because soybean plants are photoperiod-sensitive for flowering. In fact, the variations in the environmental conditions (i.e., light and temperature) of the two experiments affected several traits of both transgenic and null segregants (hereafter: nontransgenic). For example, the early-planted nontransgenic plants in experiment #1, compared to the late-planted nontransgenic plants in experiment #2,  Table S5). In both experiments, transgenic plants, compared to nontransgenic plants, showed an early reproductive phase with early flowering (i.e., 6 and 7 days earlier in the experiment #1 and #2, respectively), early pod set (i.e., 24 and 8 days earlier), and lower node positions (~ 2 nodes lower in both experiments) where the first flower appeared (Figs. 3B, 4). Additionally, the mature transgenic plants were 9.5% and 13.0% shorter than the nontransgenic ones in the experiment #1 and #2, respectively, but in each of the two experiments there was no difference in the total number of nodes between the transgenic and the nontransgenic plants, suggesting the reduced plant height was due to the reduced internode length (Fig. 4). This consistency suggested that expression of the ZmSOC1 was able to enhance reproductive production and reduce plant height. On the other hand, the early-planted plants in the experiment #1 were taller and had more seed pods that the late-planted plants in the experiment #2 due to the difference in photoperiod and maybe temperatures too. In both experiments, the transgenic plants had more seed pods and seed production, including a nonsignificant (P = 0.053) 23.2% of increase of the grain weight per plant in the experiment #1 and a significant (P = 0.040) 13.5% of increase in the experiment #2 (Figs. 3C,D, 4). Seed quality of nontransgenic and ten transgenic lines was evaluated by measuring eight quality parameters. When the ten transgenic lines were compared as one transgenic group to the nontransgenic seeds, no significant difference between the transgenic and the nontransgenic groups was found for all of the 10 seed quality parameters, suggesting that the expression of ZmSOC1 had little impact on grain quality (Table 1).
Phenotypic evaluation T 2 transgenic plants. T   In the leaves of six plants from the two transgenic lines, high expressions of the bar and the ZmSOC1 were found; in contrast, no sequence reads of the transgenes were detected in any of the three nontransgenic plants ( Table 2). The results verified the expression of the two transgenes. Meanwhile, it showed that the expression of  Table 2; Figure S2; Table S6). 410 and 782 DETs were identified in the two comparisons between each of the two transgenic lines (i.e., #20tr and #60tr) and the nontransgenic line, respectively (Fig. 5A). Of the annotated soybean genes, 58 DE genes (DEGs), including 31 upregulated, 25 downregulated DEGs and two other DEGs were shared in the two comparisons; the two DEGs (i.e., FUCO2_ARATH and FTSH2_ORYSJ) showed opposite changes in the two transgenic lines (Table 2). Using regular RT-PCR, expression of both the bar and the ZmSOC1 was detected in  48,49 , and it also binds to the cis-acting region of the PYROPHOSPHATE-ENERGIZED VACUOLAR MEMBRANE PROTON PUMP 1 gene (AVP1) which regulates auxin-mediated organ development and enhance NaCl and drought tolerance 50,51 . The upregulated VOZ1 resulted in an increased expression of the AVP1 in the #60tr but not in the #20tr line (Table S4) Table 2). CYTOPLASMIC MALATE DEHYDROGE-NASE (MDHC_MEDSA) is involved in carbohydrate metabolic process. FRUCTOSE-BISPHOSPHATE ALDO-LASE 5 (ALFC5_ARATH) hydrolyzes the fructose 1-6-bisphosphate to fructose 6-phosphate. Receptor protein kinase TMK1 (TMK1_ARATH) functions in auxin signal transduction 55 . WRKY DNA-binding transcription factor 70 (WRK70_SOLLC) modulates various phytohormones signals and affects senescence, biotic and abiotic stress responses 56 . Gibberellin receptor GID1B (GID1B_ARATH) is a soluble gibberellin (GA) receptor redundant with GID1A and GID1C that control root growth, seed germination, and flower development through GA signaling; loss-of-function of these receptors can lead to plant dwarfing [57][58][59] . CHLOROPLASTIC PALMITOYL-MONOGALACTOSYLDIACYLGLYCEROL DELTA-7 DESATURASE (ADS3_ARATH) catalyzes desaturation of fatty acid 60,61 (Table 2). 60S RIBOSOMAL PROTEIN L10 (RL10_EUPES) is a component of the small ribosomal subunit. The 58 upregulated and downregulated DEGs could be responsible, at least in part, for the phenotypic changes of the ZmSOC1-expressing plants. The other two shared DEGs along with those non-shared DETs likely contributed to the differences between the two transgenic lines ( Table 2, Table S6).
Gene ontology (GO) analysis of the 58 shared DEGs, which were consistently upregulated or downregulated in both transgenic lines, resulted in a total of 41 overrepresented GO terms (P < 0.05) in three networks, including 26 in "biological process", 18 in "molecular function", and six in "cellular component" (Fig. 5B, Figure S4). These overrepresented GO terms in the three networks revealed the multi-faceted effect of ZmSOC1 expression in the two soybean transgenic lines.

Discussion
In this study, a maize MADS-box gene ZmSOC1, which has a 54.8% identity to the soybean's GmSOC1 at the protein level, was constitutively expressed in soybean cultivar Jack. We found that the constitutively expressed ZmSOC1 promoted flowering and reduced plant height, suggesting that the monocot-derived ZmSOC1 functions similarly to the SOC1 in Arabidopsis as a flowering pathway integrator 21 . Furthermore, we found that the Table 1. Effect of ZmSOC1 expression on seed quality. Protein, fiber and oil contents were measured using a Grain Analyzer. Fatty acids were determined by gas chromatography. The mean value (± STDEV) for each transgenic and nontransgenic (NT) line, were for the seeds from 2-4 transgenic plants except that all_tr is the average of all transgenic lines. Each transgenic line and all_tr were compared with the nt. NA not available. Signif. codes: ***P < 0.001, **P < 0.01, and *P < 0.05.   www.nature.com/scientificreports/ transgenic plants had a 23.2% and a 13.5% of higher grain weight per plant than the nontransgenic plants in the two experiments, respectively. The DEGs responding to the expression of the ZmSOC1 suggested broader effects of the ZmSOC1 on plant height, flowering time, and the other unmeasured traits (e.g., photosynthesis efficiency and abiotic tolerance), which could contribute to yield increase. This is the first report that a constitutively expressed SOC1 gene was evaluated in soybean. We used the ZmSOC1 instead of the GmSOC1 in this study because it was easier to detect integration and expression of the transgene ZmSOC1 than that of a transferred endogenous GmSOC1 in transgenic soybean plants.
ZmSOC1 expression hastens soybean flowering. This is the first demonstration that a major flowering pathway integrator SOC1 from maize functioned in soybean. Unsurprisingly, the expression of the ZmSOC1 promoted soybean plant flowering because the ZmSOC1 is similar to the GmSOC1, which is a SOC1 homology that functions as a positive regulator for plant initiation and flowering 21,29,30,33,38 . Interestingly, more phenotypic diversities were observed in the T 0 plants than the T 1 plants because of the differences in the genetic background of different transgenic lines as well as the culture condition for the plants. Remarkably, the type of the altered flowers appeared only in T 0 transgenic plants although these abnormal flowers produced no seeds. In this particular case, the abnormal flowers could be caused by the transgene expression because none of the T 0 nontransgenic plants had similar abnormal flowers. In general, T 0 transformants of soybean obtained through the transformation of half-seed explants were often chimeric, which made the T 0 transformants not convincing transgenic targets for phenotyping.
Overexpression of MADS-box gene GmAGL1 was more effective in hastening soybean plant (cv. Jack) flowering under long-day than short-day conditions 32 . In this study, regardless of the photoperiod variations in the two experiments, the transgenic plants flowered earlier than the nontransgenic plants. The plants grown under short-day conditions in experiment #1 flowered earlier and set pods much later than those under the long-day conditions in experiment #2. The low temperatures during the early development stage in the experiment #1 could have delayed pod set, which could in turn contribute to taller and more branches for each plant observed in the experiment #1 than those in the experiment #2.
ZmSOC1 expression reduces soybean size. ZmSOC1 overexpression reduced the height of maize transgenic plants 38 . Similarly, the ectopic expression of the ZmSOC1 resulted in the reduction of the transgenic soybean plants. This provides further evidence to demonstrate that manipulation of the expression of a SOC1 homolog is an effective approach to change plant architecture.
Effects of the ZmSOC1 expression on soybean yield and grain quality. Both yield and seed quality are determined by the interactions of genetic and environmental factors. Overexpression of the GmAGL1 in two transgenic lines promoted plant maturity but had no trade-off of yield and grain quality 32 . In this study, T 1 transgenic plants from ten transgenic lines were used to expand the genetic diversity of the transgenic plants, meanwhile, the plants were evaluated under two environmental. It was exciting and convincing that the transgenic plants had a higher yield than the nontransgenic plants grown in the pots although field trials are still needed to evaluate homozygous transgenic plants.
Genes and gene networks responding to the expression of the ZmSOC1. We used RNA sequencing data of two transgenic lines to identify the differentially expressed genes associated with the expression of the ZmSOC1 expression. Although the data only represented the transcriptome at the moment of sampling, the information of the DEGs was very useful to reveal the molecular mechanism that underpinned the phenotypic changes in the transgenic plants ( Table 2, Fig. 6). For example, in addition to the upregulated expression of the APRR3 and VOZ1, and PAB, the upregulated FUL gene (FULL_VITVI) was a direct molecular evidence to support the enhanced floral initiation and early flowering in the transgenic plants. The increased expressions of APRR3 and VOZ1 were not associated with the increased CO or FT due possible to the strong expression of both genes in the leaf tissues harvested for sequencing at noon when the sunshine was strong. The downregulated GID1B was the best indicator for the reduced plant size of the transgenic plants because of the known function of the GID1B as a GA receptor in regulating plant height, and flower development [57][58][59] . The associations among the expressed ZmSOC1, the upregulated expression of FUL gene, the downregulated GID1B, the promoted flowering, and reduced plant height suggested that the GA signaling pathway (e.g., GID1B) had involved in ZmSOC1 expression-induced flowering and plant dwarfing. This provides a new insight to elucidate the pathway of an overexpressed SOC1 in reducing plant height.
In addition, five DEGs (i.e., HSF24, CHLD, WRK70, TMK1, and PER31) involved in plant photosynthesis, phytohormone signaling, or abiotic tolerance ( Table 2, Fig. 6) had impact on yield potential in the transgenic plants, and more studies are still needed to investigate photosynthetic efficiency, phytohormone types and contents, and abiotic tolerance.

Conclusion
Soybean is a photoperiod-sensitive crop. Both plant height and days to flowering and maturity are of importance for soybean adaptability and yield. We demonstrated that expression of the ZmSOC1 in soybean was very effective in promoting flowering and reducing plant height. Remarkably, the expression of the ZmSOC1 could increase grain production. Transcriptome analysis of two transgenic lines revealed the genes underpinning the potential phenotypic changes driven by the expression of the ZmSOC1 in the transgenic plants. Overall, the results provided new information to understand SOC1-mediated flowering in soybean. Most importantly, both the phenotypic and transcriptome data of the transgenic plants suggest that modulating expression of a MADS-box Figure 6. Potential effects of the expression of the ZmSOC1 on soybean plant growth and development revealed by the shared differential expressed genes (DEGs) identified in leaves of two transgenic lines (i.e., #20tr and #60tr) ( Table 2).