Bioinformatics-Driven Identification of Flooding-Tolerance Genes and Their Structural and Molecular Characterization in Soybean

Abstract

Flooding stress is one of the important abiotic factors restricting soybean production globally. Despite numerous efforts, the genetic mechanisms of flooding tolerance are not fully understood. This work used a bioinformatics-driven, integrative approach combining structural analysis, molecular docking, expression profiling, and biochemical phenotyping to prioritize and characterize genes associated with flooding tolerance (FTgenes) during the early stages of soybean seedling development. Functional annotation of 83 previously prioritized FTgenes was used to classify them into five functional categories, with one candidate per category identified for further study: enolase (energy supply); phytocyanin (reactive oxygen species; ROS management); glycoside hydrolase family 17 (GH17; cell wall adjustment); ethylene and sulfur compound biosynthesis-related protein (ESCBp; hormonal signaling); and alfin zinc-finger protein (gene regulation). Comprehensive sequence analyses, phylogenetic relationships, motif predictions, and homology modeling were performed for these five candidate genes. Molecular docking analyses demonstrated that enolase, GH17, and ESCBp had conserved residues necessary for catalysis and forming stable complexes with respective ligands. Expression analysis showed distinct but coordinated transcriptional activation of the FTgenes during short-term flooding stress. Levels of salicylic acid and abscisic acid changed as a function of time upon flooding. In addition, protein levels, chlorophyll a content, total phenolics, flavonoids, and antioxidant capacity were modulated in an organ-dependent manner. Overall, this study has provided evidence for a coordinated network of energy metabolism, ROS homeostasis, cell wall adjustment, hormone signaling, and regulation of gene expression that characterizes the early molecular response of soybean seedlings to flooding stress. These genes provide structurally and molecularly characterized candidate genes for future genetic improvement.

1. Introduction

Soybean (Glycine max [L.] Merr.), is one of the world’s most important sources of edible protein and oil; however, its productivity is extremely vulnerable to short-term or long-term flooding due to irregular rainfall events and poor soil drainage. , Flooding can decrease the soybean yield by 40–60%, depending upon the developmental stage and duration of the flooding. The reasons for this include the inhibiting effects of waterlogging on root respiration, reduction in photosynthesis, and an increase in the susceptibility to disease. ,, Waterlogging can occur at any growth stage, and damage to seeds and seedlings alone can markedly reduce grain yield. Because flooding tolerance in plants is an important and a complex quantitative trait influenced by multiple genetic loci, which interact with the environment, it is necessary to identify the key regulatory genes and pathways to enable the development of improved crop varieties using traditional breeding and biotechnology. In a recent association-mapping study of 254 soybean accessions, the loci controlling early season flood tolerance on chromosomes 4, 17, and 20 differed from those reported for midseason tolerance, which indicates that tolerance at different developmental stages has a partly distinct genetic basis. Functional genomics and systems biology have also helped to identify the genes and regulatory networks that control stress resistance in soybean and sorghum. ,

At the physiological and molecular levels, flooding triggers a rapid transition from aerobic respiration to anaerobic metabolism, along with the activation of glycolysis, fermentation and the remobilization of sugars to generate adenosine triphosphate (ATP) when there is insufficient oxygen. , Hypoxia-associated energy deprivation and carbon starvation are sensed in part by the sucrose nonfermenting-related kinase 1 and target of rapamycin kinase, which adjust growth to the limited ATP supply. Studies using proteomics in soybean roots have shown that there was a significant increase in the level of glycolytic enzymes, including enolase, alcohol dehydrogenase, and aldolase, and changes in proteins involved in modifying the cell wall, specifically pectin methylesterase inhibitors. Comparative morpho-physiological and biochemical profiling of waterlogging-tolerant and sensitive soybean introgression lines has further confirmed that antioxidant enzyme activity was modulated in a genotype-dependent manner during flooding stress. Early transcriptomic studies demonstrated that 97 out of 29,388 soybean genes exhibited greater than 25-fold induction following 12 h of flooding, encompassing genes associated with alcohol fermentation, ethylene biosynthesis, pathogen defense, and cell wall loosening; subsequent comparative transcriptome analyses further identified flooding-responsive regulators including methylation and demethylation (MET1, DME), members of the hemoglobin family (GLBs), aquaporins (AQPs), glycolytic genes, NAC transcription factor, and AP2/ERFs transcription factors, with expression patterns differing substantially between tolerant and sensitive cultivars. , A more recent study of soybean under water excess reported 4,412 differentially expressed genes, many of which were linked to energy metabolism, cell wall modification, and ethylene crosstalk. At the same time, excessive reactive oxygen species (ROS) will be generated during the process of reoxygenation; therefore, plants require a strong antioxidant system based on phenolics, flavonoids, and ROS-scavenging enzymes. The integration of metabolic and redox signals by phytohormones, such as ethylene, abscisic acid (ABA), and salicylic acid (SA), will direct growth-defense trade-offs. Phytohormones constitute a central regulatory layer during flooding: ethylene, the most pervasive and rapid flooding signal, accumulates within 1 h of submergence due to impaired gas diffusion and triggers aerenchyma formation and adaptive root remodeling in tight interaction with ROS and nitric oxide; ABA modulates glycolysis and zinc-finger transcription factors as well as stomatal responses; , and SA contributes to defense priming during reoxygenation. , In soybean specifically, exogenous ethylene application has been shown to enhance waterlogging tolerance, while bZIP and AP2/ERF transcription factors represent key regulatory hubs integrating ABA and ethylene signaling under flooding. , Additionally, it has been reported that transcription factors e.g., NAC, MYB, WRKY, bHLH, and ERF/DREB have emerged as pivotal convergence points in the regulation networks responding to diverse stresses, integrating signals from MAPK cascades, phytohormones, calcium, and ROS to orchestrate coordinated transcriptional reprogramming in response to abiotic stresses. It has been noted that soybean stress resilience regulatory genes, including GmARF20 and GmPgb1, which affect root system plasticity and nitric oxide scavenging under waterlogging, can be exploited in precision breeding.

Despite considerable progress in identifying flooding-responsive proteins through transcriptomic and proteomic approaches, prior work has predominantly characterized individual candidates in isolation, either cataloguing expression changes at the RNA level or detecting differential protein abundance, without providing structural or mechanistic validation of the underlying gene functions. For instance, transcriptomic analyses in soybean have identified hundreds of flooding-responsive genes including those encoding fermentative enzymes, ROS scavengers, and hormone biosynthesis proteins, ,, while proteomic studies have highlighted the importance of enolase, alcohol dehydrogenase, and pectin methylesterase inhibitor in early flooding responses. However, such studies typically stop at the detection stage: the three-dimensional structures of the encoded proteins, their ligand-binding properties, the structural basis of their catalytic activity, and their organ-specific temporal expression dynamics under early flooding conditions remain uncharacterized for the vast majority of candidates. Furthermore, candidates such as phytocyanin and alfin zinc-finger proteinhere identified as representative genes for ROS management and gene regulation, respectivelyhave not been previously characterized at the structural or functional level in the context of soybean flooding tolerance, despite their implied roles in copper-based ROS signaling and stress-responsive transcriptional regulation. The prioritized 83 FTgenes in soybean, established through integrated bioinformatics mining, meta-analysis, QTL mapping, and systems biology, represent a rich but largely unvalidated resource for gene discovery; and while integrative network analysis has identified nine candidate FTgenes within hormone signaling, defense, and metabolic pathways, no prior study has coupled this prioritization framework to a comprehensive, gene-to-biochemistry validation pipeline that encompasses structural modeling, molecular docking, organ expression profiling, phytohormone quantification, and antioxidant phenotyping within a single study. Such an integrative framework is essential for distinguishing core regulators from secondary responders and for understanding how distinct functional modules, such as energy supply, ROS management, cell-wall adjustment, hormonal signaling, and gene regulation, are wired into a coherent flooding-tolerance network.

The present study used a bioinformatics-driven integrative strategy combining structural analysis, molecular docking, expression profiling, and biochemical phenotyping to characterize FTgenes in soybean seedlings exposed to short-term flooding. Starting from the 83 prioritized candidates, the analysis selected one representative protein from each major functional categoryenolase, phytocyanin, GH17, ESCBp, and the alfin zinc-finger transcription factorand combined protein family analysis, motif discovery, homology modeling, and molecular docking to deduce their structural and functional properties. These in silico predictions were then linked to in vivo responses through RT-qPCR-based expression profiling, targeted phytohormone quantification, and biochemical assays of antioxidant capacity in shoots and roots under flooding. By bridging gene-level properties with whole-plant physiology, this study aimed to elucidate how selected FTgenes are expressed and regulated during early flooding stress in soybean seedling and to provide structurally and molecularly characterized candidates for future genetic improvement.

2. Materials and Methods

2.1. Protein Sequence Retrieval, Family Identification, and HMMER Analysis

The 83 FTgenes used in this study were adopted from the priority gene list established by Lai et al. in which genes were systematically prioritized from a pool of 36,705 soybean genes using an integrated bioinformatics framework combining meta-analysis, QTL mapping, gene prioritization algorithms, and systems biology approaches, with a defined cutoff threshold for combined genomic feature scores. The mRNA and protein sequences of 83 prioritized FTgenes were retrieved from the National Center for Biotechnology Information (NCBI) database (https://www.ncbi.nlm.nih.gov). Sequences were downloaded in FASTA format for subsequent analyses. Protein family classification was performed using InterProScan version 105.0. Each protein sequence was analyzed to identify conserved domains and family assignments, then categorized into 5 processes, consisting of energy supply, ROS management, cell wall adjustment, hormonal signaling, and gene regulation, according to its functional role. Profile hidden Markov model (HMM) analyses were conducted using HMMER version 3.4 (http://hmmer.org). For each protein sequence, the alignment file with protein identity 80–100% was performed using Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo), and then profile HMMs were constructed using hmmbuild. Then, these HMMs were used to search the target protein data set using hmmsearch and significant hits were filtered based on E-value thresholds (E < 1 × 10^–10).

2.2. Sequence Feature Prediction

SignalP-5.0 (https://services.healthtech.dtu.dk/services/SignalP-5.0/) was used to identify and characterize N-terminal signal peptides and their cleavage sites. ProtParam (https://www.expasy.org/resources/protparam) was used to compute the theoretical isoelectric point (pI) and molecular weight for each protein sequence. WoLF PSORT (https://www.genscript.com/wolf-psort.html) was utilized to predict the probable subcellular localization of each protein, thus aiding functional annotation.

2.3. Candidate Protein Selection and Downstream Analyses

Five candidate proteins (enolase, phytocyanin, GH17, ESCBp, alfin) were selected based on the results of InterProScan, HMMER analysis, and sequence feature prediction using the following hierarchical criteria: (i) the high number of statistically significant HMMER homologous hits (E < 1 × 10^–10) within each functional category, indicating evolutionary conservation; (ii) canonical functional centrality within the category, prioritizing proteins with well-characterized enzymatic or regulatory roles directly relevant to flooding acclimation; (iii) availability of a high-quality structural template in SWISS-MODEL (sequence identity >50%, QMQE score >0.6) enabling reliable homology modeling and molecular docking analysis; and (iv) one representative protein per functional category, ensuring nonredundant coverage of the five flooding-response modules (energy supply, ROS management, cell wall adjustment, hormonal signaling, and gene regulation). Selected candidates underwent further analyses, based on phylogenetic tree construction, motif analysis, and homology modeling. For phylogenetic tree construction, the homologous protein sequences according to the HMMER analysis were retrieved from the NCBI database. Multiple sequence alignment was performed using the MUSCLE algorithm implemented in MEGA12 (https://www.megasoftware.net). The optimal model of protein evolution was determined using MEGA’s built-in model selection tool and the model selected was based on Bayesian information criterion scores. Phylogenetic trees were constructed using the maximum likelihood method in MEGA12. Bootstrapping was performed with 1,000 replicates. TreeCluster (https://github.com/niemasd/TreeCluster) was used to cluster the exported phylogenetic tree. The Newick-formatted tree was provided as input, with clustering being performed using the “Avg Clade” method. A pairwise distance threshold was adjusted to 0.1 substitutions/site. For visualization, the resulting clusters were mapped to tree topology in iTOL (https://itol.embl.de). The homologous protein sequences were identified further for conserved motifs using the MEME Suite 5.5.9 web server (https://meme-suite.org/meme/index.html). The MEME algorithm was configured to search for up to 3 motifs, with motif widths set between 6 and 50 amino acids following MEME Suite default parameters and published conventions for plant protein family analyses, a range sufficient to capture both short active-site loops and extended domain-level conserved regions while maintaining motif interpretability. Additionally, homology modeling of selected protein sequences was performed using the SWISS-MODEL web server (https://swissmodel.expasy.org). The structure template with the highest QMQE score and sequence similarity was selected to build the structure model. Finalized models were visualized using PyMOL (http://www.pymol.org).

2.4. Molecular Docking Analysis

Ligand-binding interactions of homology-modeled enolase, GH17, and ESCBp were predicted using AutoDock 4.3 (ADT version 1.5.6). The substrate analogs 2-phosphoglyceric acid, laminaritriose, and (2-aminooxy-ethyl)-[5-(6-amino-purin-9-yl)-3,4-dihydroxy-tetrahydro-furan-2-ylmethyl]-methyl-sulfonium covalent with pyridoxal-5′-phosphate were used as a ligand for enolase, GH17, and ESCBp, respectively. Protein structures were prepared by the addition of polar hydrogen atoms and Kollman charges were assigned, with the aromaticity criterion defined as 7.5. Ligands were docked into the proteins using a grid box of 80 × 80 × 80 points along the x, y, and z axes, respectively, with a grid spacing of 0.375 Å. Docking calculations utilized a Lamarckian genetic algorithm executed through 100 independent runs under the Cygwin environment (version 3.4.5; https://www.cygwin.com). The conformation having the most favorable binding energy (ΔG) was selected for analysis.

2.5. Plant Growth and Flooding Treatment

The soybean seeds were surface-sterilized in 1% (v/v) sodium hypochlorite for 10 min, rinsed thoroughly with sterile distilled water, and sown in sandy loam, according to the method of Jhan et al. Ten seeds per pot (240 mm sides, 190 mm deep) were germinated, and at the two-true-leaf stage (approximately 8 d postsowing), five uniform seedlings per replicate were retained. Flooding stress was imposed by submerging pots in water-filled trays with the water level maintained 5 cm above the soil surface. Shoot tissues were harvested at 3, 6, 12, and 24 h postsubmergence. Untreated plants served as controls. All harvested shoots were flash-frozen immediately in liquid nitrogen for subsequent RNA extraction.

2.6. RNA Extraction and Quantitative Reverse-Transcription PCR (qRT-PCR)

The total RNA was isolated from the frozen shoot tissues using TRIzol reagent (Thermo Fisher Scientific), following the manufacturer’s instructions. First-strand cDNA synthesis was performed with 5 μg of DNase-treated RNA, 500 ng of oligo­(dT)₁₈ primer, and 200 U of SuperScript III reverse transcriptase (Invitrogen) in a 20 μL reaction at 50 °C for 60 min qRT-PCR amplification used Luna Universal qPCR Master Mix (New England Biolabs) on a real-time thermal cycler with gene-specific primer pairs detailed in Supporting Table S1. Relative gene expression levels were determined using the 2^–ΔΔCT method, with untreated plant cDNA collected at 0 h before initiation of flooding treatment as the calibrator sample. Actin was used as the reference gene for normalization. Expression levels at 3, 6, 12, and 24 h of flooding were expressed relative to this pretreatment baseline. All experiments were carried out using biological triplicates.

2.7. Phytohormone Quantification Using HPLC-MS/MS

Shoot and root tissues (200 mg fresh weight) were ground finely in liquid nitrogen and extracted in 2 mL in a methanol–water-formic acid ratio of 15:4:1 (v/v/v) supplemented with 10 ppm naphthalene acetic acid (NAA) as internal standard. After vortexing for 5 min and incubation at −20 °C (overnight), extracts were centrifuged at 2,000 rpm for 15 min at 4 °C, after which the supernatants were purified using solid-phase extraction. The Supel–Select HLB 30 mg were activated with 1 mL of methanol, then loaded with 1 mL of supernatant, washed with 1 mL of 10% (v/v) methanol, and eluted with 1 mL of a methanol-acetonitrile ratio of 1:1 (v/v) containing 2% (v/v) ammonium hydroxide. Eluates were passed through 0.22 μm cellulose acetate syringe filters prior to LC–MS/MS analysis.

Chromatographic separation was carried out using an Inertsil ODS-3 column (150 mm × 4.6 mm, 5 μm; GL Sciences) maintained at 40 °C on a Nexera UHPLC system (Shimadzu) coupled to an LCMS-8060 triple-quadrupole mass spectrometer. The mobile phases consisted of 0.1% (v/v) formic acid in water (A) and 0.1% (v/v) formic acid in methanol (B), at a flow rate of 0.3 mL/min under a linear gradient from 65% A to 5% A over 45 min. Electrospray ionization in negative mode with multiple reaction monitoring measured SA (m/z 137 → 93, t _R 32.4 min, CE 20 eV), ABA (m/z 263 → 153, t _R 29.2 min, CE 10 eV), and NAA (m/z 185 → 141, t _R 32.9 min, CE 9 eV). Hormone concentrations were calculated from external calibration curves (Supporting Figure S1) and normalized to NAA recovery (linear regression: y = 297,372x, R ² = 0.9978). Five biological replicates were analyzed per treatment.

2.8. Biochemical Assays

2.8.1. Total Protein and Chlorophyll a

Each tissue sample (150 mg) was homogenized in 1 mL of extraction buffer (0.7 M sucrose, 0.5 M Tris-HCl pH 8.0, 50 mM EDTA, 0.1 M KCl, 0.28 M β-mercaptoethanol), then vortexed for 5 min, and centrifuged at 12,000 rpm for 1 min at 4 °C. Soluble protein in 20 μL of supernatant was quantified based on the Bradford assay at A ₅₉₅ using bovine serum albumin as the standard (Supporting Figure S2A). For chlorophyll a, each shoot tissue (1 g) was extracted in 5 mL of 80% (v/v) methanol, then vortexed, filtered through cotton wool, and clarified using centrifugation. The chlorophyll a content was calculated from dual-wavelength absorbance (A ₆₅₂ and A ₆₆₅) according to Sumanta et al.

2.8.2. Phenolics and Flavonoids

Methanolic extracts (50 μL) from the previous step were incubated with 25 μL of Folin-Ciocalteu reagent for 4 min at room temperature, then added with 125 μL of 0.2 M Na₂CO₃, and developed in darkness for 20 min. The total phenolic content (TPC) was determined at A ₇₆₀ relative to gallic acid equivalents (Supporting Figure S2B). The total flavonoid content (TFC) was determined based on 96 μL of methanolic extracts being treated sequentially with 30 μL of 1% (w/v) NaNO₂ for 5 min, then 12 μL each of 10% (w/v) AlCl₃ and deionized water for 6 min, and 40 μL of 1 M NaOH for 10 min, diluted with 22 μL of water, and measured at A ₅₁₀ against catechin standards (Supporting Figure S2C).

2.8.3. Antioxidant Properties

The total antioxidant capacity (TAC) was assayed using phosphomolybdenum reduction, with 100 μL of methanolic extracts being combined with 200 μL of reagent (0.6 M H₂SO₄, 28 mM sodium phosphate, 1% [w/v] ammonium molybdate), then boiled for 10 min, cooled, and determined at A ₆₉₅ using catechin calibration (Supporting Figure S2D). Ferric-reducing antioxidant capacity (FRAC) was determined based using a modified Prussian blue assay, in which 125 μL of extracts, 315 μL of 0.2 M potassium phosphate buffer (pH 6.6), and 315 μL of 1% (w/v) potassium ferricyanide were incubated at 50 °C for 20 min, then added with 315 μL of 10% (w/v) trichloroacetic acid, centrifuged and subsequently, 100 μL of supernatants were mixed with equal volumes of water and 0.1% (w/v) FeCl₃ for measurement at A ₇₀₀ against ascorbic acid standards (Supporting Figure S2E). DPPH radical-scavenging activity was determined by mixing 20 μL of extract with 60 μL of 0.1 mM DPPH in methanol and 120 μL of water, incubating in darkness (30 min), and measuring at A ₅₁₇. The scavenging percentage was calculated as 100 × (A ₀ – A ₁)/A ₀, where A ₀ and A ₁ denote blank and sample absorbances, respectively.

2.9. Statistical Analysis

Data were analyzed using one-way analysis of variance (ANOVA) with SPSS Statistics version 21 (IBM Corp.). When the null hypothesis was rejected, Tukey’s honestly significant difference (HSD) was applied for multiple comparisons. Differences among treatment means were considered statistically significant at P < 0.05.

3. Results

3.1. Bioinformatics-Based Identification and Functional Classification of Flooding-Responsive Proteins

Comprehensive bioinformatics characterization of 83 previously prioritized FTgenes was performed based on InterProScan-mediated domain annotation and HMM analysis. Following retrieval of complete coding sequences from the NCBI database, conserved protein domains were systematically identified and cross-validated against established protein families. This integrative approach enabled functional classification of the FTgene complement into five categories: energy supply (17 proteins, Table ), ROS management (7 proteins, Table ), cell-wall adjustment (10 proteins, Table ), hormonal signaling (8 proteins, Table ), and gene regulation (20 proteins, Table ). An additional subset of 21 proteins lacked unambiguous domain assignments and was designated as functionally unannotated (Supporting Table S2).

1. Flooding-Responsive Proteins Involved in Energy Supply in Glycine max .

Locus tag	RefSeq accession no.	Family	Length (aa)	Signal peptide	pI	Mature Mw	Subcellular localization	HMMER homologous no.
Glyma.01g118000	XP_003516954.1	Pyruvate decarboxylase/indolepyruvate decarboxylase-like	607	0	5.62	65.63	Cytosol	12
Glyma.07g153100	XP_003529168.1		607	0	5.85	65.15
Glyma.04g092100	XP_003522752.1	Mitochondrial carrier protein	300	0	9.86	32.33	Cytosol	136
Glyma.07g253700	XP_040873375.1	Pyruvate phosphate dikinase	953	0	5.56	104.42	Chloroplast	6
Glyma.17g020600	XP_006600329.1		950	0	5.84	104.26
Glyma.08g139100	NP_001356051.1	Photosynthetic NDH subunit/FKBP13-like	209	0	8.03	22.34	Cytosol	17
Glyma.08g199800	XP_003533035.1	ATP-dependent 6-phosphofructokinase	458	0	6.31	50.30	Cytoskeleton	49
Glyma.09g153900	XP_003534043.1	Enolase	445	0	5.92	48.02	Cytosol	50
Glyma.16g204600	XP_003548246.1		445	0	6.06	48.03
Glyma.10g048200	XP_003537241.1	Adenylate kinase/UMP-CMP kinase	285	0	7.58	31.54	Mitochondria	25
Glyma.12g150500	XP_003540091.1	Stem-specific protein TSJT1-like	254	0	6.19	27.97	Cytosol	26
Glyma.12g222400	NP_001235203.2		254	0	6.13	28.03	Nuclear
Glyma.07g049900	XP_006582819.1	plant glycogen synthase	619	0	6.62	68.01	Chloroplast	30
Glyma.16g018500	XP_003548134.1		619	0	6.62	68.11
Glyma.20g218100	XP_003556431.1		605	0	8.31	66.49
Glyma.17g205000	XP_003550159.1	Adrenodoxin-like	161	0	9.08	17.76	Chloroplast	4
Glyma.18g009700	NP_001240046.1	Glyceraldehyde/Erythrose phosphate dehydrogenase	340	0	6.97	37.00	Cytosol	20

2. Flooding-Responsive Proteins Involved in ROS Management in Glycine max .

Locus tag	RefSeq accession no.	Family	Length (aa)	Signal peptide	pI	Mature Mw	Subcellular localization	HMMER homologous no.
Glyma.08g128100	NP_001235375.1	Phytocyanin	121	25	9.9	10.55	Extracellular	48
Glyma.08g176300	XP_040874167.1	Carotenoid oxygenase	611	0	7.7	67.84	Chloroplast	13
Glyma.09g149200	NP_001358565.1	Iron/ascorbate-dependent oxidoreductase	382	0	5.63	43.40	Cytosol	222
Glyma.13g234500	XP_003543019.1	Glutaredoxin-like plant II	133	0	7.05	14.09	Chloroplast	45
Glyma.13g250400	NP_001276223.2	Terpene synthase-like	534	0	5.59	62.34	Cytosol	33
Glyma.02g268200	NP_001241509.2	Plant 2-oxoglutarate-dependent oxidoreductases	356	0	6	40.90	Chloroplast	212
Glyma.14g049000	XP_003545587.1		307	0	5.35	34.90	Cytosol

3. Flooding-Responsive Proteins Involved in Cell Wall Adjustment in Glycine max .

Locus tag	RefSeq accession no.	Family	Length (aa)	Signal peptide	pI	Mature Mw	Subcellular localization	HMMER homologous no.
Glyma.01g037200	XP_003517862.1	Peptidase M10A	379	30	4.96	38.07	Chloroplast	24
Glyma.03g132700	NP_001238474.1	Glycoside hydrolase family 17	347	32	8.47	34.7	Chloroplast	111
Glyma.05g123900	XP_003524771.1	Polygalacturonase-Inhibiting Protein	336	21	9.1	35.26	Extracellular	929
Glyma.05g124000	XP_003524772.1		334	19	8.64	34.83
Glyma.07g036400	XP_003528441.1	Lecithin:cholesterol/phospholipid:diacylglycerol acyltransferase	676	0	6.28	75.57	Nuclear	13
Glyma.08g119400	XP_003532766.1	Fasciclin-like arabinogalactan protein group A	281	22	8.97	26.8	Extracellular	51
Glyma.08g119500	XP_003532767.1		288	24	9.12	27.25	Plasma membrane
Glyma.08g119600	XP_003532768.1		315	27	8.84	29.63	Chloroplast
Glyma.08g128500	XP_003531307.1	UDP-glucose 4-epimerase	416	0	9.07	45.73	Chloroplast	72
Glyma.13g270100	XP_003543212.1	S-locus-specific glycoprotein/EP1	435	26	5.95	45.99	Vacuolar membrane	85

4. Flooding-Responsive Proteins Involved in Hormonal Signaling in Glycine max .

Locus tag	RefSeq accession no.	Family	Length (aa)	Signal peptide	pI	Mature Mw	Subcellular localization	HMMER homologous no.
Glyma.02g148200	XP_003518926.1	Aspartic peptidase A1	435	23	8.7	43.85	Chloroplast	19
Glyma.13g208000	XP_003542883.1		473	0	8.34	51.19	Chloroplast
Glyma.05g108900	XP_003525876.2	Ethylene and sulfur compound biosynthesis-related protein (ESCBp)	480	0	8.95	54.05	Nuclear	28
Glyma.17g158100	XP_006600907.1		475	0	8	53.5
Glyma.10g048100	XP_003536977.1	Two-component response regulator ARR-like	747	0	6.65	81.12	Nuclear	141
Glyma.11g055700	NP_001241348.1	Zeaxanthin epoxidase	654	0	6.75	72.16	Cytosol	14
Glyma.17g174500	NP_001276261.1		669	0	7.14	73.7	Chloroplast
Glyma.11g181200	XP_003538103.1	Cytokinin-induced F-Box protein AT3G44326-like	341	0	9.24	38.9	Cytosol	13

5. Flooding-Responsive Proteins Involved in Gene Regulation in Glycine max .

Locus tag	RefSeq accession no.	Family	Length (aa)	Signal peptide	pI	Mature Mw	Subcellular localization	HMMER homologous no.
Glyma.04g044900	XP_003523996.1	C2H2-type zinc-finger protein AZF1/2/3-like	233	0	8.66	24.99	Nuclear	60
Glyma.07g036100	XP_006582916.1	Alfin	253	0	5.31	28.81	Nuclear	37
Glyma.07g036200	NP_001235387.1	Nuclear transcription factor Y subunit A	348	0	8.86	37.68	Nuclear	102
Glyma.07g036300	XP_014633178.1	Ataxin2-like	632	0	5.41	69.23	Endoplasmic reticulum	28
Glyma.08g119000	XP_040874048.1	Plant disease resistance NBS-LRR	210	0	9.36	20.93	Cytosol	96
Glyma.08g119200	XP_003532765.2		252	0	9.3	28.17	Cytosol
Glyma.08g119100	XP_003531267.1	BTB/POZ domain-containing adapter for CUL3-mediated RhoA degradation	470	0	5.55	51.2	Cytoskeleton	5
Glyma.08g218600	NP_001237558.1	WRKY transcription factor	313	0	8.6	34.48	Nuclear	235
Glyma.10g047800	XP_003536976.1	Rab GTPase-activating TBC domain-containing protein	825	0	4.96	93.71	Cytosol	14
Glyma.10g180800	NP_001241360.1	Myb transcription factor	275	0	5.38	31.47	Nuclear	360
Glyma.11g149900	NP_001235759.1	PAM2-containing CID1/CID2	170	0	4.84	19.38	Chloroplast	9
Glyma.11g180500	XP_003538097.1	Ribonuclease CAF1	284	0	6.12	32.7	Cytosol	12
Glyma.12g093100	XP_003539842.1		278	0	5.71	31.78	Cytosol
Glyma.12g187400	NP_001304595.2		309	0	6.82	89.71	Nuclear
Glyma.13g243600	XP_014621303.1	SNF2/RAD54 Helicase and Transcription Factor	1319	0	5.25	145.96	Nuclear	134
Glyma.13g243700	XP_003543074.1	Ubiquitination and Signaling Pathway Regulator	242	0	4.61	27.33	Chloroplast	18
Glyma.13g250300	XP_003541764.1	Transcription factor AMS/ICE1/BHLH6-like	208	0	4.89	24.33	Nuclear	29
Glyma.13g251100	XP_003541768.1	Kinesin-like protein	799	0	6.82	89.71	Nuclear	252
Glyma.13g251300	XP_006594652.1	Transcription factor ILR3-like	144	0	6.61	16.22	Nuclear	20
Glyma.18g042100	NP_001238594.2	E3 ubiquitin-protein ligase PUB22/23/24-like	431	0	7.41	47.6	Mitochondria	156

3.1.1. Sequence-Derived Structural and Localization Features

Sequence feature prediction revealed important structural characteristics of the identified proteins. The SignalP-5.0 analysis identified signal peptides in 19 proteins, indicating their secretory or membrane-bound functions. The signal peptide-containing proteins were predominantly localized to extracellular, chloroplast, and plasma membrane compartments. The ProtParam analysis determined theoretical pIs ranging from 4.61 to 10.53 and molecular weights from 7.34 to 145.96 kDa. The WoLF PSORT predictions assigned subcellular localizations with the distribution: nuclear (31 proteins), cytosol (19 proteins), chloroplast (16 proteins), and other compartments (mitochondria, plasma membrane, extracellular, vacuolar membrane, endoplasmic reticulum, peroxisome, cytoskeleton; total of 17 proteins), indicating multicompartmental coordination of flooding responses.

3.1.2. Energy Supply Proteins

The energy supply category comprised 17 proteins involved in glycolytic metabolism, fermentation, and carbohydrate mobilization under hypoxic stress (Table ). Key glycolytic enzymes consisted of two enolase isoforms (Glyma.09g153900, Glyma.16g204600), ATP-dependent 6-phosphofructokinase (Glyma.08g199800), and glyceraldehyde/erythrose phosphate dehydrogenase (Glyma.18g009700). Fermentative metabolism was represented by two pyruvate decarboxylase/indolepyruvate decarboxylase-like isoforms (Glyma.01g118000, Glyma.07g153100). Chloroplast-localized proteins consisted of two pyruvate phosphate dikinase isoforms (Glyma.07g253700, Glyma.17g020600), three plant glycogen synthase isoforms (Glyma.07g049900, Glyma.16g018500, Glyma.20g218100), and adrenodoxin-like protein (Glyma.17g205000). Additional proteins consisted of mitochondrial carrier protein (Glyma.04g092100), adenylate kinase/UMP-CMP kinase (Glyma.10g048200), photosynthetic NADH dehydrogenase (NDH) subunit/FKBP13-like protein (Glyma.08g139100), and two stem-specific protein TSJT1-like isoforms (Glyma.12g150500, Glyma.12g222400).

3.1.3. ROS Management Proteins

The ROS management category contained 7 proteins associated with ROS detoxification, redox signaling, and oxidative-stress mitigation (Table ). Phytocyanin (Glyma.08g128100) was involved in redox signaling, while iron/ascorbate-dependent oxidoreductase (Glyma.09g149200) and glutaredoxin-like plant protein II (Glyma.13g234500) were functioned in ROS detoxification. Additionally, oxidative-stress mitigation consisted of carotenoid oxygenase (Glyma.08g176300), terpene-synthase-like protein (Glyma.13g250400), and plant 2-oxoglutarate-dependent dioxygenase isoforms (Glyma.02g268200, Glyma.14g049000). Notably, carotenoid oxygenase and the terpene-synthase-like protein likely contributed to both oxidative-stress mitigation and redox signaling based on their annotated functions.

3.1.4. Cell Wall Adjustment Proteins

The cell wall adjustment category comprised 10 proteins involved in polysaccharide metabolism, pectin protection, and cell adhesion under flooding stress (Table ). Polysaccharide metabolism was represented by GH17 (Glyma.03g132700), UDP-glucose 4-epimerase (Glyma.08g128500), and peptidase M10A (Glyma.01g037200). Two polygalacturonase-inhibiting protein (PGIP) isoforms (Glyma.05g123900, Glyma.05g124000) were involved in pectin protection. Finally, cell adhesion consisted of fasciclin-like arabinogalactan protein isoforms (Glyma.08g119400, Glyma.08g119500, Glyma.08g119600), lecithin:cholesterol/phospholipid:diacylglycerol acyltransferase (Glyma.07g036400), and S-locus-specific glycoprotein/EP1 (Glyma.13g270100).

3.1.5. Hormonal Signaling Proteins

The hormonal signaling category comprised 8 proteins involved in phytohormone biosynthesis, perception, and downstream signaling cascades (Table ). Proteins associated with plant hormone biosynthesis consisted of aspartic peptidase A1 isoforms (Glyma.02g148200, Glyma.13g208000), ESCBp isoforms (Glyma.05g108900, Glyma.17g158100), and zeaxanthin epoxidase isoforms (Glyma.11g055700, Glyma.17g174500). A two-component response regulator ARR-like protein (Glyma.10g048100) was related to perception, while cytokinin-induced F-box protein AT3G44326-like (Glyma.11g181200) was associated with downstream signaling cascades.

3.1.6. Gene Regulation Proteins

The gene regulation category comprised 20 proteins encompassing transcription factors, chromatin remodelers, RNA-binding proteins, and ubiquitin-proteasome components that orchestrate transcriptional and post-translational responses to flooding stress (Table ). Major transcription factor families consisted of WRKY (Glyma.08g218600), Myb (Glyma.10g180800), alfin zinc-finger protein (Glyma.07g036100), C2H2-type zinc-finger protein AZF1/2/3-like (Glyma.04g044900), nuclear transcription factor Y subunit A (Glyma.07g036200), transcription factor AMS/ICE1/bHLH6-like (Glyma.13g250300), and transcription factor ILR3-like (Glyma.13g251300), with the exceptionally high conservation of Myb (360 matches) and WRKY (235 matches). Chromatin-remodeling machinery was represented by SNF2/RAD54 helicase-transcription factor (Glyma.13g243600). RNA-binding proteins consisted of ataxin-2-like RNA-binding protein (Glyma.07g036300), PAM2-containing CID1/CID2 (Glyma.11g149900), and ribonuclease CAF1 isoforms (Glyma.11g180500, Glyma.12g093100, Glyma.12g187400). Ubiquitin-proteasome components consisted of E3 ubiquitin-protein ligase PUB22/23/24-like (Glyma.18g042100) and ubiquitination-signaling pathway regulator (Glyma.13g243700), nucleotide-binding site leucine-rich repeat (NBS-LRR) disease-resistance proteins (Glyma.08g119000, Glyma.08g119200). Additional regulatory proteins consisted of BTB/POZ domain-containing CUL3 adapter (Glyma.08g119100), Rab GTPase-activating TBC-domain protein (Glyma.10g047800), and kinesin-like protein (Glyma.13g251100).

3.2. Phylogenetic Relationships, Conserved Motifs, and Homology-Based Structural Models of Candidate FTgenes

3.2.1. Enolase (Glyma.09g153900)

The phylogenetic tree of 50 enolase homologues resolved 13 clades, reflecting ancient gene duplication and subfunctionalization within this enzyme family (Figure A). Glyma.09g153900 clustered within a soybean-enriched subclade exhibiting minimal branch divergence, consistent with strong purifying selection maintaining catalytic efficiency across the angiosperms. Homology modeling generated a canonical two-domain enolase architecture comprising an N-terminal α/β fold and a C-terminal TIM-barrel domain (Figure B). Three statistically significant 50-amino-acid motifs were identified across the enolase family (Figure C). When mapped onto the three-dimensional model, these motifs localized to the active-site cleft and adjacent flexible loops, implicating their roles in substrate coordination and catalytic-residue positioning during the reversible dehydration of 2-phospho-d-glycerate.

1.

Phylogenetic and structural characterization of soybean enolase. (A) Maximum likelihood phylogenetic tree of enolase homologues (n = 50 isoforms). Colors represent distinct clades. Candidate enolase isoform (Glyma.09g153900) marked with red star. (B) Three-dimensional homology model of Glyma.09g153900 (template: [1MQ89.1A]; GMQE = 0.97). Colors indicate spatial localization of the three conserved motifs. (C) Conserved sequence motifs identified in enolase protein family. P-values indicate statistical significance of motif conservation.

3.2.2. Phytocyanin (Glyma.08g128100)

Phylogenetic analysis of 48 phytocyanin homologues revealed 25 clades, consistent with functional diversification in redox signaling, pathogen defense, and abiotic-stress responses (Figure A). Glyma.08g128100 occupied a derived position within a legume-specific clade characterized by short terminal branches, indicating recent common ancestry and conserved copper-binding motifs. Homology modeling produced a compact β-sandwich fold typical of plastocyanin-like cupredoxins, preceded by a 25-residue N-terminal signal peptide targeting the extracellular space (Figure B). Three conserved motifs corresponded to surface-exposed loops surrounding the predicted type-I copper-binding site (Figure C), underscoring the structural constraints imposed by metal coordination and electron-transfer function.

2.

Phylogenetic and structural characterization of soybean phytocyanin. (A) Maximum likelihood phylogenetic tree of phytocyanin homologues (n = 48 isoforms). Colors represent distinct clades. Candidate phytocyanin isoform (Glyma.08g128100) marked with red star. (B) Three-dimensional homology model of Glyma.08g128100 (template: [G7L9J5.1.A]; GMQE = 0.99). Colors indicate spatial localization of the three conserved motifs. (C) Conserved sequence motifs identified in phytocyanin protein family. P-values indicate statistical significance of motif conservation.

3.2.3. Glycoside Hydrolase Family 17 ( Glyma.03g132700 )

Maximum-likelihood analysis of 111 GH17 sequences generated a complex tree topology with 35 subclades, reflecting extensive gene duplication and neofunctionalization within this cellulase/glucanase superfamily (Figure A). Glyma.03g132700 nested within a subclade of predominantly leguminous endo-1,3-β-glucanases characterized by minimal sequence divergence, supporting retention of ancestral callose-hydrolysis activity. Homology modeling yielded a (β/α)₈ TIM-barrel architecture, the hallmark fold of glycoside hydrolases (Figure B). Three statistically robust motifs mapped to the inner barrel surface and connecting loops, defining the substrate-binding groove and catalytic machinery responsible for β-1,3-glycosidic-bond cleavage (Figure C).

3.

Phylogenetic and structural characterization of soybean GH17. (A) Maximum likelihood phylogenetic tree of GH17 homologues (n = 111 isoforms). Colors represent distinct clades. Candidate GH17 isoform (Glyma.03g132700) marked with red star. (B) Three-dimensional homology model of Glyma.03g132700 (template: [Q03773.1.A]; GMQE = 0.98). Colors indicate spatial localization of the three conserved motifs. (C) Conserved sequence motifs identified in GH17 protein family. P-values indicate statistical significance of motif conservation.

3.2.4. Ethylene and Sulfur Compound Biosynthesis-related Protein (Glyma.05g108900)

Phylogenetic analysis of 28 ESCBp homologues resolved 7 clades, suggesting ancient divergence and lineage-specific expansion (Figure A). Glyma.05g108900 was positioned basally within a legume clade with relatively short branches, indicative of functional constraint on ethylene biosynthesis or sulfur metabolism. Homology modeling produced an extended multidomain protein featuring an N-terminal α-helical region connected to a compact catalytic core reminiscent of 1-aminocyclopropane-1-carboxylate (ACC) synthase (Figure B). Three conserved 50-amino-acid motifs localized to the predicted active-site pocket and adjacent regulatory surfaces (Figure C), consistent with roles in substrate binding, pyridoxal-5′-phosphate (PLP) cofactor coordination, and allosteric regulation of ethylene biosynthesis.

4.

Phylogenetic and structural characterization of soybean ESCBp. (A) Maximum likelihood phylogenetic tree of ESCBp homologues (n = 28 isoforms). Colors represent distinct clades. Candidate ESCBp isoform (Glyma.05g108900) marked with red star. (B) Three-dimensional homology model of Glyma.05g108900 (template: [W9RA06.1.A]; GMQE = 0.91). Colors indicate spatial localization of the three conserved motifs. (C) Conserved sequence motifs identified in ESCBp protein family. P-values indicate statistical significance of motif conservation.

3.2.5. Alfin (Glyma.07g036100)

Maximum-likelihood reconstruction of 37 alfin-family revealed 6 subclades, reflecting diversification of alfin-domain transcription factors in response to varied environmental stresses (Figure A). Glyma.07g036100 clustered with closely related soybean orthologues exhibiting minimal sequence divergence, suggesting conserved DNA-binding specificity and transactivation potential. Homology modeling generated an extended structure comprising an N-terminal PHD-like zinc-finger domain and a C-terminal DUF3594 domain characteristic of plant alfin proteins (Figure B). Three statistically significant motifs localized to the predicted DNA-binding interface and adjacent regulatory loops (Figure C), implicating these regions in recognition of GC-rich cis-elements and recruitment of chromatin-modifying complexes during transcriptional activation or repression under flooding stress.

5.

Phylogenetic and structural characterization of soybean alfin. (A) Maximum likelihood phylogenetic tree of alfin homologues (n = 37 isoforms). Colors represent distinct clades. Candidate alfin isoform (Glyma.07g036100) marked with red star. (B) Three-dimensional homology model of Glyma.07g036100 (template: [Q06A73.1.A]; GMQE = 0.79). Colors indicate spatial localization of the three conserved motifs. (C) Conserved sequence motifs identified in alfin protein family. P-values indicate statistical significance of motif conservation.

3.3. Molecular Docking Reveals Energetically Favorable Ligand Binding by Enolase, GH17, and ESCBp

Molecular docking showed that soybean enolase Glyma.09g153900 bound its ligand with a binding energy of −6.85 kcal/mol, mainly through hydrogen bonds contributed by residues H120, E172, D207, E258, D285, K310, H338, R339, S340, and K361 (Figure A,D), which are clustered around the catalytic pocket, similar to the human neuron-specific enolase template (3UCC). Soybean GH17 (Glyma.03g132700) exhibited a stronger predicted interaction (−9.4 kcal/mol), with the ligand stabilized by multiple hydrogen bonds involving R8, Y34, N57, N94, E95, N174, Y176, E238, E287, and K290 and by nearby hydrophobic residues L9, Y34, Y176, F179, F282, and W298 at the binding interface (Figure B,D), consistent with the environment observed in potato endo-1,3-β-glucanase (4GZI). ESCBp (Glyma.05g108900) had the most favorable docking score (−11.24 kcal/mol), where hydrogen bonds formed with A104, T105, N186, D215, Y218, S256, S258, and K259, and hydrophobic contacts involved Y3, F4, W7, L29, A104, Y129, I217, and Y218 (Figure C,D), closely matching the active-site surroundings of apple ACC synthase (1M4N). Across all three proteins, the ligands were enclosed by clusters of polar and charged residues within 4 Å of the binding site, indicating well-defined pockets that combined hydrogen bonding and hydrophobic interactions to stabilize ligand binding in enolase, GH17, and ESCBp. These docking results established that the selected soybean FTgenes encoded structurally competent enzymes capable of forming energetically favorable complexes with their physiological substrates, thereby providing molecular-level validation of bioinformatics-derived functional assignments.

6.

Molecular docking studies of enolase, GH17, and ESCBp. (A) Superimposition of docked Glyma.09g153900 (gray) with human neuron specific enolase-1-PGA/PEP (magenta, PDB: 3UCC), (B) Glyma.03g132700 (gray) with potato endo-1,3-β-glucanase (purple, PDB: 4GZI), and (C) Glyma.05g108900 (gray) with apple 1-aminocyclopropane-1-carboxylate synthase (green, PDB: 1M4N). Hydrogen bonds (less than 3.5 Å between donor and acceptor atoms) represented as black dashed lines. Amino acids forming hydrogen bond interactions and amino acids with hydrophobic side chains shown as sticks. (D) Molecular interaction analysis of Glyma.09g153900, Glyma.03g132700, and Glyma.05g108900 against their ligand.

3.4. Temporal Transcriptional Dynamics of FTgenes During Short-Term Flooding

The qRT-PCR analysis of representative FTgenes revealed coordinated but functionally distinct transcriptional responses to submergence (Figure ). The energy supply gene Glyma.04g092100 was rapidly induced, showing significant peaks at 3 h before declining toward later time points, whereas Glyma.09g153900 exhibited a strong, continuous up-regulation, reaching its maximum at 24 h. Of the ROS management genes, Glyma.08g128100 was significantly up-regulated at 3 and 24 h, and Glyma.08g176300 was highest at 12 h, indicating enhanced oxidative-stress responses during prolonged flooding. The cell wall adjustment genes responded strongly, with Glyma.03g132700 rapidly induced with a maximum at 6 h followed by a gradual decline, while Glyma.05g124000 remained low at early time points but was massively induced at 12 and 24 h. Among the hormonal signaling genes, Glyma.05g108900 showed a steady increase throughout the treatment, reaching the highest expression at 24 h, whereas Glyma.11g055700 was highest at the control and early time points and was significantly repressed after 12–24 h of flooding. The gene regulation FTgenes displayed diverse patterns, with Glyma.07g036100 consistently induced with a maximum at 6 h, and Glyma.11g149900 being up-regulated strongly at 12 h, with expression remaining higher than for the controls at 24 h.

7.

Relative gene expression of FTgenes under flooding stress; (A) genes related to energy supply [Glyma.04g092100, Glyma.09g153900], (B) ROS management [Glyma.08g128100, Glyma.08g176300], (C) cell wall adjustment [Glyma.03g132700, Glyma.05g124000], (D) hormonal signaling [Glyma.05g108900, Glyma.11g055700], and (E) gene regulation [Glyma.07g036100, Glyma.11g149900], where error bars indicate mean ± SD and different lowercase letters above columns indicate significant differences at P < 0.05.

3.5. Organ-Specific Accumulation of Salicylic Acid and Abscisic Acid under Flooding

Flooding significantly altered salicylic acid (SA) and abscisic acid (ABA) levels in soybean seedlings. In shoots, the SA concentration rose progressively during flooding, with the highest level at 24 h and differing significantly from the control and earlier time points, indicating a sustained activation of SA-mediated responses (Figure A). The root SA content showed smaller, transient changes, with a modest increase at 12 h that was significantly higher than at some other time points, suggesting a weaker or delayed SA response in roots (Figure B). Shoot ABA concentrations fluctuated but were generally elevated under flooding, with significant peaks at 3 h and again at 24 h compared to the lowest level at 6 h, implying dynamic ABA regulation during early and prolonged submergence (Figure C).

8.

Concentration of salicylic acid in soybean seedling shoot (A) and root (B) and abscisic acid in soybean seedling shoot (C) under flooding stress. Concentrations of each hormone investigated based on 5 biological replicates using HPLC-MSMS analysis. Error bars indicate ± standard error. Different lowercase letters above columns indicate significant differences at P < 0.05.

3.6. Spatiotemporal Modulation of Protein Content, Pigments, Phenolics, and Antioxidant Capacity

Flooding stress caused time-dependent changes in the protein content, pigments, and antioxidant profiles of the soybean seedlings (Table ). In the shoots, the total protein content changed only modestly, with a small but significant increase at 3–6 h followed by values comparable to the control at later time points. Chlorophyll a and the TPC were significantly elevated at 6 h, and shoot antioxidant capacity (TAC, FRAC, DPPH) remained higher than the control during 6–24 h, indicating a sustained enhancement of shoot antioxidant defenses under flooding. In roots, the total protein content increased significantly at 3–12 h but had declined significantly by 24 h, while the TPC remained relatively stable, with a slight decrease at the latest time point. The root TFC levels increased markedly, peaking at 12 h, yet TAC, FRAC, and DPPH radical-scavenging activity progressively decreased, with the lowest values at 12–24 h, suggesting that prolonged flooding had weakened the root antioxidant capacity despite transient increases in flavonoids.

6. Biochemical Profiles of Soybean Seedlings after Flooding Stress Treatment .

Flooding stress times	Total protein (μg/mL)	Chlorophyll a (μg/mL)	TPC (μg GAE/g FW)	TFC (μg CAE/g FW)	TAC (mg CAE/g FW)	FRAC (μg AA/g FW)	% DPPH radical scavenging activity
Shoot
Ctrl	43.19 ± 0.49a	0.94 ± 0.11a	16.72 ± 0.72a	38.31 ± 2.96a	0.99 ± 0.10a	10.03 ± 0.88a	42.67 ± 3.16a
3 h	46.24 ± 3.42ab	1.01 ± 0.29a	16.61 ± 1.02a	45.96 ± 7.59a	0.96 ± 0.06a	9.54 ± 0.20a	43.31 ± 5.45a
6 h	49.48 ± 1.37b	1.93 ± 0.16b	18.47 ± 0.23b	43.22 ± 0.90a	1.36 ± 0.08b	10.28 ± 0.34ab	55.87 ± 1.93b
12 h	42.62 ± 1.28a	0.68 ± 0.15a	16.53 ± 0.26a	42.82 ± 4.12a	1.50 ± 0.02b	11.34 ± 0.71c	52.14 ± 2.99b
24 h	43.50 ± 1.64a	0.82 ± 0.17a	16.34 ± 0.23a	36.75 ± 5.95a	1.40 ± 0.09b	11.24 ± 0.27bc	52.17 ± 3.24b
Root		N/M
Ctrl	29.16 ± 0.36b		6.16 ± 0.20b	12.53 ± 0.88c	0.21 ± 0.01b	6.73 ± 0.53c	33.29 ± 1.92bc
3 h	34.70 ± 1.39c		6.19 ± 0.08b	4.88 ± 0.29a	0.19 ± 0.03ab	5.94 ± 0.17b	34.50 ± 1.11c
6 h	33.53 ± 0.85c		5.95 ± 0.18b	9.49 ± 1.22b	0.32 ± 0.05c	5.81 ± 0.30b	32.58 ± 0.80bc
12 h	34.48 ± 3.73c		5.85 ± 0.39b	17.73 ± 2.45d	0.14 ± 0.01a	5.66 ± 0.23b	31.34 ± 1.05b
24 h	24.82 ± 1.09a		5.35 ± 0.09b	7.33 ± 1.22ab	0.15 ± 0.02a	4.55 ± 0.19a	26.93 ± 0.63a

^aN/M = means not measured.

4. Discussion

Flooding imposes combined hypoxia, carbon starvation, and oxidative stress; typically, tolerant genotypes sustain carbohydrate breakdown and ATP production via enhanced glycolysis and fermentation. Under hypoxia, VII ethylene response factors (ERF-VIIs), master regulators of transcriptional reprogramming, are activated downstream hypoxia-responsive genes such as ADH and PDC, and fermentation regenerates the NAD⁺ needed to sustain glycolysis when mitochondrial respiration is limited. , In the present study, the strong and sustained induction of glycolytic FTgenes, particularly enolase Glyma.09g153900, indicated that the soybean seedlings could rapidly reprogram carbon metabolism to maintain energy supply under submergence. The high structural conservation of Glyma.09g153900 with neuron-specific enolase, together with favorable docking to 2-phosphoglycerate, supported the preservation of canonical catalytic function, implicating increased enolase activity as a key component of metabolic acclimation. Comparable up-regulation of enolase and other glycolytic enzymes in waterlogged rice, Arabidopsis, and soybean underscores the central role of glycolysis in ATP generation and survival under oxygen limitation. ,,, In soybean leaves under water excess, β-amylase was induced, which is in line with an increased sugar supply for anaerobic metabolism.

ROS act both as damaging byproducts of impaired aerobic metabolism and as signals that initiate acclimatory responses during flooding. The late, persistent induction of the phytocyanin gene Glyma.08g128100, together with carotenoid oxygenase Glyma.08g176300, suggested activation of a ROS-management module that coupled extracellular electron transfer with chloroplast-based antioxidant pathways. Homology modeling and motif analysis placed Glyma.08g128100 within a conserved copper-binding phytocyanin clade, implying participation in apoplastic redox reactions where ROS-mediated cell-wall loosening and defense signaling occurred. This finding is consistent with the proposed functions of phytocyanins and blue copper proteins in other plants under abiotic stress, , and with a comprehensive bioinformatics analysis in the Setaria italica genome in which phytocyanin genes were associated with redox-active proteins and abiotic stress regulatory networks. Notably, extracellular ROS produced in the apoplast by plasma membrane-localized NADPH oxidases have been recently shown to play a direct mechanistic role in autophagosome biogenesis in response to nutrient deprivation and hypoxia, underscoring that apoplastic ROS signaling pathways constitute an important regulatory layer during environmental stress.

Cell wall remodeling is a common feature of flooding responses, supporting both growth maintenance in partially submerged tissues and barrier reinforcement in vulnerable organs. The rapid, transient induction of β-1,3-glucanase GH17 (Glyma.03g132700) followed by strong up-regulation of the polygalacturonase-inhibiting protein (PGIP) Glyma.05g124000 points to a two-phase adjustment of the soybean cell wall, in which early GH17 expression promotes controlled loosening of β-glucan networks, whereas later PGIP accumulation preserves pectin integrity against pathogen-derived polygalacturonases during prolonged flooding. Structural conservation of Glyma.03g132700 with known endo-1,3-β-glucanases and favorable docking to laminaritriose reinforce this functional assignment, paralleling observations in maize and wheat where β-1,3-glucanases and PGIPs jointly balanced growth, wall strength, and disease resistance. , Cell wall modification was also among the main processes altered in soybean under water excess, and enhanced cell wall metabolism has been linked to improved germination and seedling growth in a waterlogging-tolerant soybean genotype.

Ethylene is a central regulator of flooding responses, integrating oxygen sensing with downstream morphological and transcriptional changes. Identification of ESCBp (Glyma.05g108900) as a nuclear ethylene- and sulfur-compound-related protein, together with its strong transcriptional induction and tight binding to an ACC-synthase-like ligand, suggests that this factor links ethylene biosynthesis to downstream signaling in flooded soybean. This in lines with the strong induction of several soybean ERF-VII genes by both submergence and the ethylene precursor ACC. The pronounced up-regulation of Glyma.05g108900 contrasted with the down-regulation of zeaxanthin epoxidase Glyma.11g055700, indicating a shift from ABA biosynthesis toward ethylene-dominated control as flooding progressed, echoing antagonistic regulation of ethylene and ABA pathways reported for Arabidopsis.

Transcription factors integrate these metabolic and hormonal cues into coordinated gene-expression programs. The alfin zinc-finger protein Glyma.07g036100 and PAM2-containing CID1/CID2 Glyma.11g149900 displayed strong, time-differential induction, suggesting distinct regulatory roles during the early and late flooding phases. The structural models positioned the conserved motifs of Glyma.07g036100 within the DNA-binding region, supporting their role as a transcriptional regulator under hypoxia and extending the reported functions of alfin-like proteins in plant development and stress response. More broadly, transcription factors of multiple families e.g., zinc-finger proteins have been identified as central regulators bridging biotic and abiotic stress responses by integrating phytohormone and ROS signaling, a broader regulatory role that is consistent with the sustained induction of the alfin zinc-finger gene Glyma.07g036100 observed during both early and prolonged flooding in the present study.

Hormone profiling further indicated that SA and ABA contributed differentially to systemic flooding responses. Shoot SA levels gradually increased and peaked at 24 h, whereas root SA changes were modest, consistent with a primary role in systemic signaling or defense priming in aerial tissues. SA has also been reported to reduce oxidative stress in flooded soybean by increasing antioxidant enzyme activity, which may help to explain why the rise in shoot SA was accompanied by higher shoot antioxidant capacity in this study. Shoot ABA exhibited a biphasic pattern, with an early peak at 3 h and a second increase at 24 h, aligning with rapid stomatal regulation, followed by longer-term acclimation. In soybean under water excess, exogenous ABA further reduced gas exchange and induced NCED3, ERF1, and peroxidase genes, supporting a link between ABA, ethylene signaling, and ROS scavenging. The temporal changes in SA and ABA levels observed in this study were consistent with the established model in which multiple gaseous and hormonal signalsincluding ethylene, O₂, CO₂, nitric oxide, and ROSinteract in a temporally dynamic manner to orchestrate adaptive flooding responses, and with prior reports documenting ABA- and SA-mediated modulation of antioxidant capacity in waterlogging-stressed soybean. ,,−

Biochemical profiling supports the spatial and temporal compartmentalization of antioxidant capacity in flooded seedlings. The shoots had increased chlorophyll a and phenolics at 6 h, together with elevated antioxidation activity from 6 h onward, indicating reinforced antioxidant defenses in photosynthetic tissues, whereas roots accumulated flavonoids but had declining TAC, FRAC, and DPPH activities after 12–24 h, implying that root antioxidant capacity had become insufficient to counter sustained oxidative stress. Similar organ-specific patterns have been reported in waterlogged soybean and maize, where shoot antioxidants increased while root redox homeostasis deteriorated. , Flavonoid biosynthesis genes and metabolites (prunin, chalcone 2′-O-glucoside) also accumulate in soybean under combined salinity and waterlogging, where they help to prevent oxidative damage. Stronger antioxidant capacity and root characteristics like adventitious roots and aerenchyma are typically characteristics of waterlogging-tolerant soybean genotypes. Breeding for flooding tolerance should therefore target traits that sustain root antioxidant systems as well as shoot resilience.

Taken together, the bioinformatic, structural, expression, hormonal, and biochemical data place the five characterized FTgenes within a multilayered early flooding response in soybean. Energy supply genes sustain ATP production; ROS management and cell wall adjustment genes mitigate oxidative and mechanical damage; hormonal signaling components coordinate systemic responses; and transcription factors integrate these inputs into adaptive expression programs. Future work should: (i) directly test the functional contribution of these FTgenes through overexpression, gene silencing, or genome editing to assess survival, biomass retention, and yield stability under field waterlogging conditions; (ii) dissect natural allelic variation in diverse germplasm using GWAS and haplotype analysis; and (iii) explore whether pyramiding favorable allelesan approach proven effective for stress tolerance in other cereals such as sorghumcan synergistically enhance flooding tolerance in soybean.

Glossary

Abbreviations

FTgenes

flooding tolerance genes

ROS

reactive oxygen species

GH17

glycoside hydrolase family 17

ESCBp

sulfur compound biosynthesis-related protein

aa

amino acids

pI

isoelectric point

Mw

molecular weight (kDa)

NDH

NAD­(P)H dehydrogenase

FKBP13

FK506-binding protein 13

ATP

adenosine triphosphate

UMP-CMP

uridine monophosphate/cytidine monophosphate

TSJT1

tomato-seedling-jasmonate-treatment 1

UDP

uridine diphosphate

PGIP

polygalacturonase-inhibiting protein

ARR

Arabidopsis response regulator

C2H2

Cys2/His2

AZF

Arabidopsis zinc-finger

NBS-LRR

nucleotide-binding site leucine-rich repeat

BTB

broad-complex, tramtrack, and bric a brac

POZ

poxvirus and zinc finger

CUL3

cullin-3

WRKY

WRKY domain-containing transcription factor (named after its conserved WRKYGQK sequence)

GTP

guanosine triphosphate

TBC

Tre-2/Bub2/Cdc16 domain

PAM2

poly­(A)-binding protein interacting motif 2

CID

C-terminal conserved interacting domain

CAF1

carbon catabolite repressor 4-associated factor

SNF2

sucrose nonfermenting 2/radiation-sensitive 54

RAD54

radiation-sensitive 54

AMS

aborted microspores

ICE1

inducer of C-repeat binding factor expression 1

BHLH6

basic helix–loop–helix 6

ILR3

indole-3-acetic acid-leucine resistant 3

PUB

plant U-box

TPC

total phenolic content

GAE

gallic acid equivalents

FW

fresh weight

TFC

total flavonoid content

CAE

catechin equivalents

TAC

total antioxidant capacity

FRAC

ferric-reducing antioxidant capacity

AA

ascorbic acid equivalents

DPPH

2,2-diphenyl-1-picrylhydrazyl

Ctrl

untreated control

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c03980.

• Standard curves of peak area versus concentration for SA and ABA (Figure S1); standard curves of bovine serum albumin for total protein assay, gallic acid for TPC assay, catechin for TFC assay, catechin for TAC assay, and ascorbic acid for FRAC assay (Figure S2); sequence of the primers used for qRT-PCR (Table S1); flooding-responsive proteins with uncharacterized functional annotations in Glycine max (Table S2) (PDF)

S.B.: Writing–original draft, conceptualization, data curation, formal analysis, investigation, methodology, resources, validation, visualization, and funding acquisition. C.-F.K.: conceptualization, formal analysis, investigation, methodology, validation, and writing–review and editing. P.-H.K., M.J., S.K., and M.P.: formal analysis, investigation, and writing-review.

This work was financially supported by the Fundamental Fund under Kasetsart University Research and Development Institute Bangkok, Thailand [FF­(S-KU)­43.66] and partially supported by the Faculty of Science at Sriracha, Kasetsart University, Sriracha Campus, Thailand (Sci-Src-M01/2566) and the Advanced Plant and Food Crop Biotechnology Center from the Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education in Taiwan.

The authors declare no competing financial interest.