Transcriptome analysis reveals the mechanism and key role of CdPER47 in bermudagrass (Cynodon dactylon) adaptation to submergence and post-submergence stresses Authors: Weitao Jia, Sihan Hou, Songyi Yang, Yuhan Zou, Wenqing Dou, Kai Zhu, Cunfeng Zhao, Shihui Zou, Huanhuan Liu, Lijiao Ai, Shengjun Wu Published: 2026 ABSTRACT AbstractAs an important turfgrass and forage grass, bermudagrass exhibits high tolerance to submergence stress, however, the molecular mechanisms are poorly clarified and required further elucidation. In this study, physiological analysis and transcriptome sequencing of bermudagrass under submergence stress and post-submergence recovery stage were conducted. Results showed that submergence exerted adverse effects on bermudagrass with decreased stolon length and photosynthetic pigment content. Many differentially expressed unigenes/genes (DEGs) were identified in shoots and roots under submergence and post-submergence stresses. Among these DEGs, 830 DEGs were annotated as transcription factors (TFs) and 104 TFs belong to APETALA 2/ethylene response element binding protein (AP2-EREBP) family. KEGG pathway enrichment analysis revealed that 42 and 4 (11 and 19) KEGG pathways were significantly enriched after submergence and post-submergence recovery in roots (shoots), including plant hormone signaling, sulfur metabolism, carbohydrate metabolism and phenylpropanoid biosynthesis. Moreover, protein-protein interaction (PPI) network was conducted, several hub genes were identified, including those encoding ribosomal proteins, translation initiation factor, persulfide dioxygenase ETHE1 and Enolase 1. Finally, a peroxidase gene CdPER47 were selected and cloned, CdPER47 exhibited significantly differential expression in roots under both submergence and post-submergence stresses. Heterogeneous overexpression of CdPER47 in Arabidopsis increased its survival rate significantly after submergence stress. Based on the above results, a hypothetical model was proposed for bermudagrass responses to submergence and post-submergence stresses. Our study reveals that bermudagrass responds to submergence and post-submergence stresses through extensive transcriptional reprogramming involving AP2-EREBP transcription factors, multiple metabolic pathways, and key hub genes including the functionally validated peroxidase gene CdPER47. Together, this study will provide valuable references for developing submergence-tolerance germplasm in the future.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12870-026-08811-4.Keywords: Submergence, Post-submergence recovery, Bermudagrass, Transcriptome, CdPER47 Introduction Flooding events have shown a dramatic upward trend since the 1950s and often cause partial or complete submergence, which is detrimental to many crops [1, 2]. Only in Southeast Asia, flooding damaged 3.72 million ha of cropland during 2015–2019, resulting in a loss of 3.26 million tons of rice [3]. Under submergence stress, limited carbon dioxide (CO2) and light availability reduce photosynthetic rates in shoots, while oxygen (O2) deprivation enhanced anaerobic respiration in respiring tissues [4–7]. Reduced photosynthate production and high carbohydrate consumption lead to an energy crisis in submerged plants [8]. Moreover, after water recedes, plants face additional post-submergence stresses, such as reoxygenation stress, reillumination stress, dehydration stress and leaf senescence. These stresses trigger the overproduction of reactive oxygen species (ROS), and ultimately impair the normal growth and development of plants [9]. Therefore, it is essential to unravel the mechanisms of plants resistance to submergence and post-submergence stresses for developing submergence-tolerance crops. Plants have evolved multiple mechanisms to mitigate the adverse effects of submergence and post-submergence stresses [9, 10]. Under complete submergence, plants restrict their growth to conserve energy for recovery upon water recession [7]. Previous study showed that ethylene response factor SUB1A prevented shoot elongation and reduced carbohydrate consumption by regulating Slender Rice-1 (SLR1) and SLR1 Like-1 (SLRL1) in rice [11]. Additionally, two other rice genes with antagonistic functions, ACCELERATOR OF INTERNODE ELONGATION 1 (ACE1) and DECELERATOR OF INTERNODE ELONGATION 1 (DEC1), were reported to regulate internode elongation in coordination with gibberellin (GA)‌ [12]. During the post-submergence stage, plants minimize the effects of excessive ROS through activating the generation of antioxidants [9, 13]. For instance, Arabidopsis increases glutathione and ascorbate antioxidant levels to restrict ROS accumulation after submergence stress, which contributed to the recovery of photosynthesis and facilitates plants regrowth upon water recession [9, 14]. Furthermore, plant shoots typically experience dehydration stress after water recedes [9]. In response, plants promote the accumulation of abscisic acid (ABA) and initiate stomatal closure to prevent water loss [14, 15]. Bermudagrass (Cynodon dactylon), as a perennial C4 herb, has been considered as an excellent warm-season turfgrass and forage grass [16, 17]. Previous studies have shown that bermudagrass is extremely tolerant to submergence stress and can resume normal growth after approximately 6–8 months of complete submergence [18, 19], indicating that it may have special submergence-resistance mechanism. In this study, the effect of submergence stress on bermudagrass growth, as well as the transcriptome profiles under submergence stress and post-submergence recovery stage were investigated, respectively. DEGs, the enriched GO categories and critical KEGG pathways response to submergence or post-submergence stresses were identified, and hub genes were revealed through PPI network analysis. Additionally, a DEG CdPER47 was cloned and confirmed to function in bermudagrass resistance to submergence and post-submergence stresses. Exploring the mechanisms of bermudagrass tolerance to submergence and post-submergence stresses paves the way for future breeding of submergence-resistant crops and horticultural plants through genetic engineering. Materials and methods Plant growth and treatment Bermudagrass (Cynodon dactylon) used for cuttings comes from the hydro-fluctuation zone of the Pengxi River (31◦08′59″N, 108◦33′40″E), which is one of the tributaries of Yangtze River in Kaizhou district, Chongqing municipality, China. This bermudagrass germplasm is identified as Cynodon dactylon according to the description in the Flora of China and cultivated in the greenhouse of Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences. For submergence and post-submergence recovery experiments, C. dactylon was planted with cuttings from greenhouses and cultured in pots with humus-vermiculite mixtures (1:1, v/v) for 30 days in a cultivate room with 25 °C temperature and 16 h photoperiod. Bermudagrasses were fully submerged in plastic containers (60 × 40 × 35 cm) and the submergence depth of shoots was approximately 20 cm, then experienced submergence treatment for 24 h, 7 d, 14 d, 30 d and 24 h recovery (after 24 h submergence stress). The distance from the bottom of shoot to the shoot apex was measured as shoot length. For phenotype analysis, Arabidopsis overexpressing CdPER47 together with wild-type lines were grown on 1/2 MS media for about 3 weeks, then subjected to dark submergence treatments in a tank (60 × 40 × 35 cm) filled with deionized water. Leaves were positioned 10 cm below the water surface referred to Liang et al. [20]. After being submerged for 3 d, all the plants were taken out and recovered for 5 d under normal conditions with 16 h photoperiod at 22 °C. Survival rates and dry weights of surviving plants were recorded. The assay kits from Beijing Boxbio Science & Technology Co., Ltd. were used to determine POD activity (CAS No. AKAO005C) and hydrogen peroxide content (CAS No. AKAO009C) according to the manufacturer’s instruction. For the determination of chlorophyll and carotenoid content, the third and fourth leaves from the apical bud were collected and weighed, then immersed in 95% ethanol for 4 days. Optical density (OD) of the extracts was measured at 665, 649, and 470 nm for chlorophylls a and b and carotenoid, respectively [21]. Total chlorophyll is the sum of chlorophyll a and chlorophyll b. RNA isolation and cDNA library construction The roots and shoots of bermudagrass that experienced submergence treatment for 24 h and 24 h recovery from 24 h submergence were harvested, respectively and stored at -80 °C for sequencing. In order to ensure that C. dactylon was in the same biological age and growth rhythm, bermudagrasses in the treatment groups were harvested together with control samples at the same time referred to the methods described by Fan et al. [22]. Each treatment had three biological replicates with roots and shoots separately, resulting in a total of 18 samples. Total RNA was extracted from roots and shoots of bermudagrass using ethanol precipitation protocol [23] and CTAB-pBIOZOL reagent (Bioer, Hangzhou, China) according to the manual instructions. The quality and quantity of the RNA were analyzed using a NanoDrop 1000 (Thermo Fisher Scientific, Waltham, MA USA) and Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, USA). mRNA was purified from total RNA using oligo (dT)-attached magnetic beads and purified mRNA was fragmented into small pieces with fragmentation buffer. Then random hexamer-primed reverse transcription was used to generate the first-strand cDNA and form the second-strand cDNA synthesis simultaneously. These cDNA fragments were taken as a template for PCR, and products were purified and validated on the Agilent 2100 bioanalyzer for quality control. Furthermore, the double stranded PCR products from previous step were denatured and formed the single strand circle DNA, which was formatted as the final libraries. Ultimately, eighteen libraries, including three biological replicates for each group, were constructed and sequenced with BGIseq500 platform (BGI-Shenzhen, China). De novo transcriptome assembly and function annotation To get clean reads, raw reads were filtered by removing sequences with adapters, and low-quality transcripts, then the clean reads were de novo assembled into transcripts using Trinity software [24]. The assembled transcripts were further clustered and removed redundancy using TGICL software to obtain unigenes [25]. These unigenes were annotated with the databases of Non-Redundant Protein Sequence Database (NR), Nucleotide Sequence Database (NT), Gene Ontology (GO), Clusters of orthologous groups for eukaryotic complete genomes (KOG), Pfam, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Swiss-Prot. Differential gene expression analysis Clean reads were blasted to unigenes using Bowtie2 software [26] and gene expression levels were evaluated by a software package, RNA-Seq by Expectation Maximization (RSEM) [27]. DESeq2 software based on negative binomial distribution was used to identify differentially expressed genes and the sequences with fold change > 2 and the adjusted P-value ≤ 0.05 was considered as the transcripts with significantly differential expression [28]. GO and KEGG pathway analyses Based on the GO annotation results, DEGs were classified into three distinct categories (biological process, molecular function and cellular component). Likewise, they were classified into different KEGG pathways according to the KEGG annotation of unigenes. Both GO and KEGG pathway enrichment analysis were performed using the phyper function in R software (version 4.3.1). The GO terms and KEGG pathways with FDR (false discovery rate) ≤ 0.01 were defined as significantly enriched. PPI network analysis and hub genes identification The PPI pairs based on differentially expressed genes were obtained according to the STRING (version 10) protein interaction database (https://cn.string-db.org/), and Cytoscape (version 3.10.1) was used to visualized the interaction network. Furthermore, PPI pairs were arranged by scores in descending order and the top 100 pairs were selected to find the hub genes. Notably, if PPI pairs above the 100th have the same score as the 100th PPI pair, select them all together. The top 10 hub genes were screened out from the PPI network according to Maximal Clique Centrality (MCC) algorithm via the Cytoscape plugin cytoHubba. Cloning of CdPER47 Roots of C. dactylon that experienced 24 h recovery after 24 h submergence stress was used to extract total RNA with the TransZol Up Kit (TransGen Biotech, Beijing, China), then the total RNA was converted to cDNA using the SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, USA). We looked up the phylogenetic relationships of the Cynodon genus on the website (http://www.timetree.org/) and found that common bermudagrass (Cynodon dactylon) and African bermudagrass (Cynodon transvaalensis) are closely related. Fortunately, the genome of African bermudagrass has been sequenced [15]. We aligned the CDS sequence of CdPER47 obtained from transcriptome sequencing with the African bermudagrass genome. The alignment showed the highest scores with two segments on chromosome 7 (possibly separated by an intron). Both segments exhibited a similarity of 99%. Subsequently, forward and reverse primers were designed based on the upstream and downstream sequences of aligned CdPER47 at the genome of African bermudagrass. The full-length cDNA of CdPER47 was amplified by the PCR using primers 5’-CGCTCATCTTCTTCCTCTT-3’ and 5’-CATTATTTGCCTTCGTCTT-3’. Quantitative real-time PCR analysis Quantitative Real-Time PCR was conducted in ViiA 7 Real-time PCR system (Applied Biosystems, USA) and used ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nangjing, China) according to the manufacturer’s instruction. The actin gene (GenBank accession no. ES295104) of C. dactylon was used as a constitutive control [29]. Primer sequences for RT-qPCR are listed in Table S17. Each reaction was repeated four times and the expression levels of unigenes in different samples were calculated by the 2–ΔΔCt method [30]. Statistical analysis The statistical analysis was conducted by the SPSS 17.0 statistical package (SPSS Inc., Chicago, IL, USA). Differences were tested by Duncan’s test unless otherwise specified, and the values were considered statistically significant at P < 0.05. Results Effect of submergence stress on bermudagrass The growth of bermudagrass was progressively inhibited as submergence duration increased. Shoot length exhibited no significant reduction during the first two weeks of submergence (Fig. 1A). However, when submergence extended to 14 d or 30 d, shoot length decreased significantly compared to control (Fig. 1A). Moreover, leaves of bermudagrass become severely wilted after 30 d of submergence stress (Fig. S1). Simultaneously, H2O2 content in roots initially increased significantly first when exposed to submergence stress for 7 d and 14 d, then returned to normal levels at 30 d (Fig. 1B). In addition, the photosynthetic capacity of bermudagrass was impacted by submergence stress. The contents of chlorophyll a, chlorophyll b, and total chlorophyll showed a downward trend with prolonged submergence, especially after 14 days or longer of submergence stress (Fig. 1C-E). In contrast, carotenoid content first increased and then decreased (Fig. 1F). Effect of submergence stress on bermudagrass. Bermudagrass were treated under submergence stress for 0, 7, 14 and 30 days, and shoot length (A), H2O2 content in roots (B), chlorophyll a (C), chlorophyll b (D), total chlorophyll (E) and carotenoid content (F) were determined. Values are means ± SE (n = 10 for (A), n = 4 for (B), n = 4–5 for (C-F)). Different letters above the bars indicate significant differences at P < 0.05 Sequencing analyses To explore the molecular mechanisms under submergence stress, roots and shoots of C. dactylon experienced 24 h submergence treatment and 24 h recovery (after 24 h submergence treatment) were harvested for transcriptome sequencing. As shown in Table S1, an average of 52.37 M (million) raw reads were generated for each sample, and after removing low-quality reads, 41.79 M clean reads were retained averagely (Table S1). Subsequently, differentially expressed genes (DEGs) were identified. In roots, 22,017 DEGs were found in roots of C. dactylon after 24 h submergence stress, including 12,003 upregulated and 10,014 downregulated unigenes (Fig. 2A, Table S2). And numbers of DEGs fell to 2412 (1557 upregulated, 855 downregulated) when experienced 24 h recovery from submergence stress (Fig. 2A, Table S2). 503 common unigenes were upregulated when experienced 24 h submergence and then downregulated after 24 h recovery from submergence stress while 951 common unigenes existed in upregulated DEGs of 24 h/CK comparison (24 h submergence treatment compared to control) and downregulated DEGs of Re/24 h comparison (24 h recovery compared to 24 h submergence treatment) (Fig. 2B). Only two common unigenes were upregulated and one common gene was downregulated under both 24 h/CK and Re/24 h comparisons (Fig. 2B). Contrastingly, a total of 792 unigenes (511 upregulated, 281 downregulated) were found to be DEGs under submergence stress while the DEG numbers reached 1109 (407 upregulated, 702 downregulated) after recovery in shoots (Fig. 2C, Table S2). Among these, 101 common unigenes were initially upregulated under submergence stress and then downregulated during the recovery stage, whereas 78 common unigenes showed the opposite expression pattern (Fig. 2D). No unigenes were found to be upregulated or downregulated under both submergence stress and the recovery stage simultaneously (Fig. 2D). Summary of DEGs. Numbers of DEGs in roots (A) and shoots (C) after submergence or post-submergence recovery; Venn diagrams of upregulated or downregulated unigenes in roots (B) and shoots (D) of 24 h/CK and Re/24 h. “24 h/CK” represents “24 h submergence treatment compared to control”. “Re/24 h” represents “24 h recovery compared to 24 h submergence treatment” To evaluate the validity of the high-throughput sequencing data, six unigenes were selected for expression analysis via RT-qPCR. The expression levels obtained from transcriptome sequencing showed good consistency, confirming the reliability of the sequencing data (Fig. S2). Transcription factors distribution To analyze transcriptional regulation during submergence stress and the post-submergence recovery stage, a total of 830 DEGs annotated as transcription factors were identified in roots and shoots of C. dactylon, belonging to 45 families (Table S3). As shown in Figs. 3 and 758 TFs were significantly differentially expressed genes in roots under submergence stress, with the majority belonging to the MYB-related (106), AP2-EREBP (104), bHLH (100), NAC (56), GRAS (44), WRKY (34) and TUB (32) families. During the post-submergence recovery stage, 62 TFs were identified as DEGs in roots, including members of the MYB-related (13), NAC (9), AP2-EREBP (6) and bHLH (5) family. In contrast, only 17 DEGs encoding TFs were detected in shoots under submergence stress, including members of the AP2-EREBP (5), WRKY (4), and other families. After 24 h recovery, 44 TFs were identified in shoots, most of which belonged to the AP2-EREBP (17), bHLH (6), C3H (6), and Trihelix (4) families. Notably, roots under submergence stress possessed the largest number of differentially expressed TFs. Transcription factors distribution in roots and shoots during submergence stress and post-submergence recovery stage. R-24 h/CK and R-Re/24 h (S-24 h/CK and S-Re/24 h) represent samples of roots (shoots) under submergence stress, roots after post-submergence recovery, shoots under submergence stress and shoots after post-submergence recovery, respectively Gene ontology enrichment analysis of DEGs To identify the major functional categories of DEGs, GO enrichment analysis was carried out on DEGs of roots and shoots. In the cellular component category, 23 and 12 GO terms were significantly enriched under submergence and post-submergence stresses, respectively, of which 11 GO terms were common ones, including cell periphery, plasma membrane, extracellular region, and so on (Fig. 4A, Table S4). For DEGs of shoots, 34 and 1 GO terms were significantly enriched under submergence and post-submergence stresses, respectively (Fig. 4A, Table S5). The top 10 GO terms with the highest number of enriched genes were similar between roots under submergence stress and those under post-submergence recovery, which is completely different from shoot samples under submergence stress (Fig. 4B, Table S4 and S6). Significantly enriched GO terms in roots and shoots after submergence and post-submergence recovery. Venn diagrams of enriched GO terms (A, C and E) and GO terms with the top ten enriched gene numbers (B, D and F) in three categories: cellular component (A and B), molecular function (C and D) and biological process (E and F). 24 h/CK-Roots and Re/24 h-Roots (24 h/CK-Shoots and Re/24 h-Shoots) represent DEGs in roots (shoots) under submergence stress and during post-submergence recovery, respectively In the molecular function category, 211 and 62 GO terms were significantly enriched in roots under submergence and post-submergence stresses, respectively, with 32 GO terms were common ones, such as oxidoreductase activity, antioxidant activity, peroxidase activity, and so on (Fig. 4C, Table S4). For DEGs of shoots, 36 and 41 GO terms were significantly enriched when experienced submergence and post-submergence stresses, respectively, of which 18 GO terms were shared, including oxidoreductase activity, tetrapyrrole binding, lyase activity, and so on (Fig. 4C, Table S5). Among the top 10 GO terms with the highest number of enriched genes, 3 (5) GO terms are common ones between roots (shoots) exposed to submergence stress and those under post-submergence recovery condition (Fig. 4D, Table S4 and S6). In the biological process category, 349 and 69 GO terms were significantly enriched in roots under submergence and post-submergence stresses, respectively, with 38 GO terms were common ones, including response to stimulus, cell wall organization or biogenesis, carbohydrate transport, phenylpropanoid metabolic process, sulfate reduction, and so on (Fig. 4E, Table S4). In shoots, 25 and 34 GO terms were significantly enriched during submergence and post-submergence recovery, respectively (Fig. 4E, Table S5), among which 2 enriched GO terms (lipid metabolic process and phosphatidylcholine biosynthetic process) were shared. Among the top 10 GO terms with the highest number of enriched genes, roots (shoots) shared 4 (1) GO terms between submergence and post-submergence recovery stages (Fig. 4F, Table S4 and S6). KEGG pathway enrichment analysis of DEGs DEGs in roots and shoots under submergence or following post-submergence recovery were mapped to the reference canonical pathways in the KEGG database for understanding their functions. Under submergence stress, 42 KEGG pathways were significantly enriched in roots, including plant hormone signal transduction, starch and sucrose metabolism, phenylpropanoid biosynthesis, sulfur metabolism and so on (Fig. 5A, Table S6). In contrast, only 4 enriched KEGG pathways were enriched after a 24-hour recovery from submergence—namely sulfur metabolism, carbon fixation in photosynthetic organisms, phenylpropanoid biosynthesis and metabolic pathways—all of which were also enriched KEGG pathways under submergence stress simultaneously (Fig. 5B, Table S6). KEGG analysis of DEGs in roots and shoots after submergence and post-submergence recovery. A-B Enriched KEGG pathways in roots after submergence (A) and post-submergence recovery (B); (C-D) Enriched KEGG pathways in shoots after submergence (C) and post-submergence recovery (D). More information is shown in Table S6 For shoots, 11 and 19 KEGG pathways were significantly enriched under submergence and after post-submergence recovery, respectively, with 7 KEGG pathways were common ones. These included sesquiterpenoid and triterpenoid biosynthesis, linoleic acid metabolism, and carbon fixation in photosynthetic organisms, among others (Fig. 5C and D, Table S6). Based on the KEGG pathway analysis results, we inferred that when exposed to submergence and post-submergence stresses, bermudagrass perceives the external stimulus, initiates the signal transduction (e.g., plant hormone signaling), activates genes involved in translation and regulation, and ultimately triggers a series of defense responses associated with energy metabolism (e.g., sulfur metabolism), carbohydrate metabolism, and biosynthesis of other secondary metabolites ((e.g., phenylpropanoid biosynthesis). To further explore the DEGs involved in bermudagrass response to submergence and post-submergence stresses, several key processes were selected for in-depth analysis. Unigenes involved in plant hormone signaling Ethylene and gibberellin are two key phytohormones involved in plant response to submergence stress [11, 30–33]. In roots, 105 unigenes associated with ethylene were significantly differentially expressed under submergence stress, whereas only 13 ethylene-related DEGs were identified during the post-submergence recovery stage (Table S7). Among these unigenes, 4 ethylene-related unigenes—Unigene337581_All, CL40599.Contig15_All, CL40599.Contig7_All and CL5653.Contig13_All—were common to both submergence and recovery processes. Contrastingly, 4 ethylene-responsive unigenes were found to be upregulated in shoots after 24 h submergence and 17 DEGs were identified during the post-submergence recovery stage (Table S7). Only one common unigene, CL628.Contig7_All, was detected, and its expression increased under submergence stress and decreased during post-submergence recovery. Further analysis showed that most of these ethylene-associated unigenes belong to the ethylene responsive factor (ERF) transcription factors family, which has been widely reported to play important roles in plant adaptation to submergence stress [31, 32]. Gibberellin is another important phytohormone that functions in plant response to submergence stress. Under 24 h submergence stress, 44 DEGs gibberellin-related were identified in roots while 4 such unigenes were significantly differently expressed during the recovery stage (Table S8). In shoots, however, no gibberellin-associated DEGs were detected under submergence stress and only one unigene (Unigene309096_All) was found to be downregulated after post-submergence recovery (Table S8). Unigenes associated with sulfur metabolism Sulfur is an essential element for plants, required for the biosynthesis of sulfur-containing amino acids, numerous proteins, vitamins, and cofactors [34]. It is also involved in the redox reaction of energy metabolism and plant resistance to abiotic stresses [34, 35]. KEGG pathway analysis revealed that under submergence stress and during post-submergence recovery, 139 and 18 differentially expressed genes (DEGs) in roots, respectively, were collectively enriched in sulfur metabolism (Fig. 6A, Table S9). Among them, 18 DEGs were common to both conditions. Specifically, 11 DEGs showed decreased expression under submergence stress but increased expression during recovery, whereas the remaining 7 exhibited the opposite expression pattern (Fig. 6B and C, Table S9). DEGs in roots related to sulfur metabolism, carbohydrate metabolism and phenylpropanoid biosynthesis. A Venn diagram of DEGs in roots involved in sulfur metabolism under 24 h/CK and Re/24 h conditions. B-C Heatmap of common DEGs in sulfur metabolism for 24 h/CK and Re/24 h with the expression pattern of first downregulated and then upregulated (B), and the opposite tendency (C). D Numbers of DEGs in KEGG pathways related to carbohydrate metabolism. E Venn diagram of DEGs in roots involved in phenylpropanoid biosynthesis under 24 h/CK and Re/24 h conditions; (F-G) Heat map of common DEGs in phenylpropanoid biosynthesis with the expression pattern of first downregulated and then upregulated (F), and the opposite tendency (G) “24 h/CK” represents “24 h submergence treatment compared to control”. “Re/24 h” represents “24 h recovery compared to 24 h submergence treatment” Unigenes related to carbohydrate metabolism Submergence stress make plants in a state of hypoxia, leading to increased carbohydrate consumption as respiration shifts from aerobic to anaerobic pathways [7]. Based on the DEGs in roots under submergence stress, 8 KEGG pathways associated with carbohydrate metabolism were enriched, including glycolysis/gluconeogenesis, glyoxylate and dicarboxylate metabolism, and starch and sucrose metabolism (Fig. 6D, Table S10). Pyruvate decarboxylase (PDC), which catalyzes the decarboxylation of pyruvate to acetaldehyde and alcohol dehydrogenase (ADH), which reduces acetaldehyde to ethanol, are two enzymes required in the ethanolic fermentation pathway of anaerobic respiration [36, 37]. Among DEGs associated with carbohydrate metabolism, 70 and 36 unigenes related to PDC and ADH, respectively, were found in roots under submergence stress (Table S11). However, no enriched pathway related to carbohydrate metabolism was found after post-submergence recovery. In shoots, only 1 pathway associated with carbohydrate metabolism (glyoxylate and dicarboxylate metabolism) was enriched under submergence stress, while 3 enriched pathways (pyruvate metabolism, pentose phosphate pathway and glycolysis/gluconeogenesis), were found during recovery stage (Fig. 6D, Table S10). Unigenes involved in phenylpropanoid biosynthesis Phenylpropanoid metabolism plays an important role in plant adaptation to biotic and abiotic stresses [38–40]. KEGG pathway analysis exhibited that DEGs in roots were enriched in phenylpropanoid biosynthesis during both submergence and post-submergence recovery stages. Among these DEGs, 76 unigenes were common ones, and 780 unigenes were specially expressed under submergence stress, while 32 unique DEGs were specifically expressed only during post-submergence recovery stage (Fig. 6E, Table S12). Notably, all 76 common unigenes exhibited opposite expression patterns between the submergence stress and post-submergence recovery stages. Among them, 63 DEGs showed a trend of initial downregulation followed by upregulation, while the remaining 13 displayed the opposite pattern (Fig. 6F and G). Furthermore, among the DEGs involved in phenylpropanoid metabolism, 31 common unigenes were annotated as peroxidase in roots during submergence stress and the post-submergence recovery stage (Table S13). PPI network construction and hub genes identification To clarify the protein interactions of DEGs, a PPI network was constructed according to the STRING database. Following submergence and recovery stages, 218,283 and 2,080 PPI pairs were identified in roots, whereas 399 and 450 pairs were found in shoots, respectively (Table S14). The top 100 PPI pairs from the network were selected and visualized using Cytoscape software (Fig. S4), then the top ten hub genes were screened out via Cytoscape plugin cytoHubba in roots and shoots under different treatments (Fig. 7, Table S15). Under submergence stress, ten hub genes in roots and nine hub genes in shoots encoded ribosomal proteins, while the remaining one in shoots encoded a eukaryotic translation initiation factor 5 A (Fig. 7A and B, Table S15). During the post-submergence recovery stage, the ten hub genes identified in roots encoded proteins such as probable protein phosphatase, enolase 1, and persulfide dioxygenase. In contrast, the ten hub genes in shoots encoded proteins including 1-phosphatidylinositol-3-phosphate 5-kinase FAB1A, enolase 1, and aldehyde dehydrogenase (Fig. 7C and D, Table S15). Notably, enolase 1 was the only hub protein that exists in both roots and shoots simultaneously. Hub genes in roots and shoots after submergence and post-submergence recovery. The top ten hub genes and their related genes in roots (A and B) and shoots (C and D) after submergence (A and C) and post-submergence recovery (B and D). The green solid ellipse represents genes associated with the hub genes. The hub genes were marked with the color gradient background from red to yellow, and the darker the color of the nodes, the higher the rank Heterologous overexpression of CdPER47 in Arabidopsis increased its tolerance to submergence stress Plant peroxidases are enzymes involved in ROS scavenging and phenylpropanoid biosynthesis [41, 42] and may contribute to bermudagrass resistance to submergence stress. Therefore, we selected and cloned CdPER47 (CL32811.Contig47_All), a DEG encoding a peroxidase. CdPER47 consists of 321 amino acids and shares high similarity with PaPER47-1 (82.24%) and PaPER47-2 (82.92%) of Phragmites australis (Fig. 8A and B). Heterologous overexpression of CdPER47 in Arabidopsis showed increased dry weight, although the difference was not statistically significant, and a significantly higher survival rate compared to the wild-type (WT) after 5 days of recovery from 72 h of submergence (Fig. 8C-E). Additionally, POD activity in shoots of transgenic lines was higher than that in the WT after 72 h of submergence stress, whereas the increase in the roots of transgenic lines was not statistically significant (Fig. 8F and G). Meanwhile, H2O2 content in the shoots of transgenic lines decreased compared to WT, although there was no significant difference (Fig. 8H). Phenylpropanoid biosynthesis provides precursors for lignin biosynthesis [39] and thus lignin content in roots and shoots was determined. Results showed that lignin content in both shoots and roots exhibited no significant difference between WT and CdPER47 transgenic lines after 72 h of submergence stress (Fig. S3). Characterization of CdPER47 and phenotype of Arabidopsis heterologous overexpressing CdPER47. A Phylogenetic analysis of CdPER47; (B) Sequence alignment of CdPER47, PaPER47-1 and PaPER47-2; (C) The performance of Arabidopsis CdPER47-overexpressing lines (P5 and P11) after submergence for 72 h and recovery for 5 d; (D-E) Dry weight per plant (D) and survival rate (E) of WT, P5 and P11 after recovery for 5 d from 72 h submergence. F-H POD activity in shoots (F) and roots (G), H2O2 content in shoots (H) after 72 h submergence stress. Values are means± SE (n= 9 for D, n= 3-4 for E-H). * indicates significant difference at P <0.05 by Student’s t-test Discussion Submergence stress induces O2 deficiency and limited CO2 availability for plants, leading to enhanced anaerobic respiration and hampered photosynthesis, and the onset of an energy crisis [43, 44]. When water recedes, plants encounter additional stresses, including reoxygenation, reillumination, dehydration stresses, and accelerated senescence [9, 45]. Previous studies have shown that bermudagrass exhibits extremely strong submergence tolerance [17, 18]. However, the underlying mechanism remains poorly understood. In this study, the impact of submergence stress on bermudagrass was investigated, and transcriptome profiles during both submergence stress and the subsequent recovery stage were analyzed. Several key pathways and hub genes were identified. Based on these findings and functional validation of the focal gene, a working model for bermudagrass adaptation to submergence and post-submergence stresses is proposed. This model spans from signal perception to transduction (plant hormone signal transduction), followed by the activation of key genes, ultimately triggering a series of defensive responses, including energy metabolism (sulfur metabolism), carbohydrate metabolism, and the biosynthesis of other secondary metabolites (e.g., phenylpropanoid biosynthesis). Phytohormone signaling and crosstalk in regulating bermudagrass resistance to submergence and post-submergence stresses Ethylene and gibberellin are recognized as pivotal phytohormones that play essential roles in plant adaptation to submergence stress [31–33]. Ethylene is a gaseous hormone that hardly diffuses out of plants under submergence stress, leading to its rapid accumulation in plant tissues [46]. The increased ethylene level subsequently induces the expression of multiple ERF family genes, which have been widely reported to be involved in plants response to submergence stress [31, 32, 47]. For example, Sub1A, an ERF-type transcription factor previously identified in rice, enhances submergence tolerance when overexpressed [31]. Fan et al. [48] reported that two ERF-VII transcription factors, RAP2.12 and RAP2.3, interacted with and were phosphorylated by calcium-dependent protein kinases CPK12 to regulate Arabidopsis tolerance to submergence stress. The ERF family is a subfamily of the AP2/EREBP-TF family [49]. Our results showed that 104 and 6 (5 and 17) members of AP2-EREBP family were significantly differentially expressed in roots (shoots) after submergence and post-submergence recovery, respectively, and most of them belong to the ERF family. Among them, 4 common unigenes (Unigene337581_All, CL40599.Contig15_All, CL40599.Contig7_All, CL5653.Contig13_All) in roots as well as 1 common unigene (CL628.Contig7_All) in shoots exhibited opposite expression patterns between the submergence stress and post-submergence recovery stages (Fig. 3, Table S7). Except for CL5653.Contig13_All, the other four unigenes are members of the ERF family. Collectively, these ERF genes may play important roles in bermudagrass adaptation to submergence and post-submergence stresses. Gibberellin usually function together with ethylene in plant response to submergence stress. Fukao et al. [11] reported that submergence-inducible Sub1A promoted the accumulation of the GA signaling repressors Slender Rice-1 (SLR1) and SLR1 Like-1 (SLRL1) in rice, thereby reducing GA responsiveness and ultimately preventing underwater shoot elongation. In deepwater rice, the ERF family genes SNORKEL1 and SNORKEL2 regulated internode elongation via GA signaling [32]. In addition, GA has been shown to regulate internode elongation together with two antagonistical genes, ACE1 (accelerator) and DEC1 (decelerator) [11]. In our study, 44 and 4 DEGs associated with gibberellin were identified in roots under submergence stress and during the post-submergence recovery stress, respectively (Table S8). By contrast, only one GA-related DEGs was found in shoots (Table S8). These gibberellin-associated DEGs may be involved in bermudagrass acclimation to submergence stress. Energy homeostasis under submergence and post-submergence stresses Prolonged submergence triggers enhanced anaerobic respiration that rapidly depletes energy reserves. Coupled with inhibited photosynthesis, this plunges plants into a severe energy crisis [5]. As a structural constituent of S-rich proteins, hydrogen sulfide, glutathione, and various secondary metabolites, sulfur is primarily involved in redox reactions associated with energy metabolism, as well as in plant growth, development, and tolerance to abiotic stresses [35]. For instance, when the hyperaccumulator Indian mustard (Brassica juncea) was exposed to cadmium stress, sulfur supply increased the activities of ascorbate peroxidase, glutathione reductase, and catalase, thereby facilitating the scavenging of ROS [50]. In tobacco, overexpression of a maize sulfite oxidase gene (ZmSO) enhanced tolerance to sulfite and drought stresses by regulating sulfur metabolism and ROS homeostasis [51, 52]. Our results showed that the expression levels of two DEGs encoding sulfite oxidase (CL3854.Contig20_All and CL3854.Contig82_All) were upregulated when experienced 24 h submergence stress and then downregulated following 24 h recovery (Fig. 6C, Table S9), suggesting that these two DEGs may function in bermudagrass acclimation to submergence stress. Ethylmalonic encephalopathy protein 1 (ETHE1) encodes a sulfur dioxygenase enzyme that oxidizes sulfides or persulfides into thiosulfate and sulfate [53–55]. Earlier studies have shown that ETHE1 participates in amino acid catabolism under carbohydrate starvation and during embryo development, suggesting its potential role in plant stress responses [54, 56]. In our study, five ETHE1 homologous unigenes (CL993.Contig179_All, CL993.Contig230_All, CL993.Contig65_All, CL993.Contig173_All and CL993.Contig205_All) were differentially expressed under submergence stress or during the post-submergence recovery stage (Fig. 6C, Table S9). Among these unigenes, CL993.Contig65_All and CL993.Contig230_All were identified as two of the ten hub genes in roots at post-submergence recovery stage (Fig. 7B, Table S15), implying that these unigenes may contribute to bermudagrass tolerance to submergence and post-submergence stresses. Changes in carbohydrate metabolism under oxygen deficiency and reoxygenation stage Complete submergence restricts oxygen diffusion in plants and promotes anaerobic respiration, leading to alterations in carbohydrate metabolism [57, 58]. For instance, previous studies have shown that submergence stress reduced total soluble carbohydrate, sucrose and starch contents of C. dactylon [19, 59]. Our results showed that 8 KEGG pathways related to carbohydrate metabolism were enriched, among which 2 pathways, starch and sucrose metabolism, glycolysis/gluconeogenesis (Fig. 6D, Table S10), have been reported in an earlier study about bermudagrass responses to submergence stress [60]. Additionally, anaerobic respiration typically facilitates the establishment of alcoholic fermentation, which relies on two key enzymes, PDC and ADH [36]. Previous study reported that the enzymatic activities of PDC and ADH in leaves of submergence-tolerant rice increased significantly compared with the submergence-intolerant rice during submergence. Moreover, overexpression PDC1 gene in rice enhanced its tolerance to submergence stress [57, 61]. Our results showed that the expression of 70 unigenes associated with PDC and 36 unigenes associated with ADH were induced by submergence stress in bermudagrass roots (Table S11), indicating these unigenes may participate in alcoholic fermentation of bermudagrass under submergence stress. ENO1 (Enolase 1) encodes an enolase that functions in glycolytic-related processes and is involved in glucose metabolism and energy conversion [62–64]. Li et al. [65] identified an enolase gene, OsEE1, that regulates coleoptile elongation in rice under submergence stress by modulating glycolysis and the tricarboxylic acid cycle pathways. In our study, ENO1 was the only hub gene that was differentially expressed in both roots and shoots simultaneously during the post-submergence recovery stage, which may function in bermudagrass recovering from submergence stress. The role of secondary metabolism in bermudagrass responses to submergence and post-submergence stresses Plants naturally synthesize secondary metabolites that are critical for coping with environmental stresses [57]. The phenylpropanoid biosynthesis pathway, derived from phenylalanine, generates a diverse array of aromatic compounds, including benzenoids, coumarins, flavonoids, hydroxycinnamates, and lignins [38, 40, 66, 67]. Peroxidases catalyze the reduction of H₂O₂ and the oxidation of specific secondary metabolites associated with phenylpropanoid biosynthesis (e.g., flavonoids and lignin monomers), thereby contributing to plant tolerance against diverse biotic and abiotic stresses [41, 42, 68, 69]. For example, overexpression of the peroxidase gene ZmPRX1 in maize promotes root development and lignification, thereby enhancing drought stress tolerance [69]. Likewise, overexpression of the wheat peroxidase gene TaPRX-2 A confers salt tolerance by scavenging ROS and modulating stress-related genes [68]. Lin et al. [70] identified 8 DEGs/DEPs in soybean encoding peroxidases that were enriched in the phenylpropanoid biosynthesis pathway under submergence stress. Our results showed that 31 common peroxidase unigenes in roots were enriched in the phenylpropanoid biosynthesis pathway during both submergence and post-submergence recovery stages (Table S13). Notably, all these unigenes were downregulated under submergence stress and subsequently upregulated following post-submergence recovery. Furthermore, heterologous overexpression of the peroxidase gene CdPER47 in Arabidopsis enhanced submergence stress tolerance, and POD activity in shoots of transgenic lines was significantly higher than WT following 72 h of submergence stress (Fig. 8F). However, lignin content in both shoots and roots exhibited no significant difference between WT and CdPER47 transgenic lines (Fig. S2). These results suggest that a single peroxidase gene exerts a limited impact on lignin content. CdPER47 transgenic lines are more likely to enhance submergence tolerance by elevating the activity of the antioxidant enzyme POD. According to the above results, we propose a hypothetical model illustrating the response of bermudagrass to submergence and post-submergence stresses (Fig. 9). When bermudagrass is exposed to submergence or post-submergence stresses, signal perception and transduction are initiated, such as plant hormone signaling and MAPK signal pathways. Subsequently, a large number of related genes are activated, including many genes associated with the translation process, which included 20 hub genes identified in Fig. 8 (Table S16). Additionally, transcription factors involved in the regulation of gene expression begin to function, including MYB-related, AP2-EREBP, bHLH and NAC family members, among others (Table S3). Moreover, a series of defense responses, such as energy metabolism, carbohydrate metabolism and the biosynthesis of other secondary metabolites, are triggered and functioned in resisting submergence and post-submergence stresses. Notably, 2 and 10 hub genes identified in Fig. 7 were found to be involved in energy and carbohydrate metabolism, respectively (Table S16). A hypothetical model of bermudagrass response to submergence and post-submergence stresses. This hypothetical model was constructed based on the results of the KEGG pathway analysis, including signal perception, signal transduction, activation of the related genes and defense responses (energy metabolism, carbohydrate metabolism and biosynthesis of other secondary metabolites). The pathways response to submergence and post-submergence stresses are marked with orange and palegreen backgrounds, respectively. Pathways with both orange and palegreen backgrounds indicate they enriched in both submergence and post-submergence stresses simultaneously Conclusions The growth of bermudagrass was significantly affected following two weeks of submergence treatment. Comparative transcriptome analysis of bermudagrass under submergence and post-submergence stresses revealed that numerous unigenes were differentially expressed, with more DEGs identified in roots than in shoots. Among these, 104 DEGs annotated as transcription factors are identified to be members of AP2-EREBP family. KEGG pathway analysis showed that 42 and 4 (11 and 19) pathways in roots (shoots) were significantly enriched under submergence stress and during post-submergence recovery, including plant hormone signaling, sulfur metabolism, carbohydrate metabolism and phenylpropanoid biosynthesis. Heterogeneous overexpression a peroxidase gene, CdPER47, in Arabidopsis enhanced its tolerance to submergence stress. Furthermore, PPI network identified several hub genes response to submergence and post-submergence stresses, including ENO1 and ETHE1. Finally, a hypothetical model illustrating the response of bermudagrass to submergence and post-submergence stresses was proposed. These findings provide new insights for improving plants tolerance to submergence stress through genetic manipulation in the future. Supplementary Information Supplementary Material 1: Fig S1. Phenotype of bermudagrass under control and submergence stress. (A) Bermudagrass without submergence experience; (B) Bermudagrass after 30 d of complete submergence. Fig S2. Validation of expression profiles of several unigenes by RT-qPCR. (A-D) Differentially expressed unigenes after 24 h submergence in roots (A-C) (R-24 h) and shoots (D) (S-24 h); (E-F) Differentially expressed unigenes after post-submergence recovery in roots (E) (R-Re) and shoots (F) (S-Re). Fig S3. Lignin content of Arabidopsis heterogeneous overexpressing CdPER47 under submergence stress. (A) Lignin content in shoots; (B) Lignin content in roots. Fig S4. The interaction network between proteins encoded by DEGs in roots and shoots after submergence and post-submergence recovery. (A-B) PPI networks under submergence stress (A) and during post-submergence recovery (B) in roots; (C-D) PPI networks under submergence stress (C) and during post-submergence recovery (D) in shoots. The turquoise and orange represent proteins encoded by downregulated and upregulated DEGs respectively. Table S1. Summary of transcriptome sequencing data of C. dactylon. Table S2. DEGs in roots and shoots under submergence stress and during post-submergence recovery. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent DEGs in roots (shoots) under submergence stress and during post-submergence recovery respectively. Table S3. DEGs annotated as transcription factors. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent transcription factors in roots (shoots) under submergence stress and during post-submergence recovery respectively. Table S4. GO terms after submergence and post-submergence recovery in roots. GO terms belonging to “Cellular Component”, “Biological Process” and “Biological Process” were presented respectively. Table S5. GO terms after submergence and post-submergence recovery in shoots. GO terms belonging to “Cellular Component”, “Biological Process” and “Biological Process” were presented respectively. Table S6. KEGG pathway of DEGs in roots and shoots under submergence stress (24 h/CK) and post-submergence recovery stage (Re/24 h). Table S7. DEGs related to ethylene in roots and shoots under submergence stress (24 h/CK) and post-submergence recovery stage (Re/24 h). Table S8. DEGs related to gibberellin in roots and shoots under submergence stress (24 h/CK) and post-submergence recovery stage (Re/24 h). Table S9. DEGs related to sulfur metabolism in roots under submergence stress and post-submergence recovery stage. Table S10. KEGG pathways related to carbohydrate metabolism in roots and shoots under submergence stress and post-submergence recovery stage. Table S11. DEGs annotated as PDC and ADH in roots under submergence stress. Table S12. DEGs related to phenylpropanoid metabolism in roots under submergence stress and post-submergence recovery stage. Table S13. DEGs annotated as peroxidase in roots under submergence stress and post-submergence recovery stage. Table S14. PPI pairs in roots and shoots under submergence stress and post-submergence recovery stage. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent PPI pairs in roots (shoots) under submergence stress and during post-submergence recovery respectively. PPI pairs for the hub genes seeking were highlighted with yellow color. Table S15. The top ten hub genes in roots and shoots under submergence stress and post-submergence recovery stage. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent the top ten hub genes in roots (shoots) under submergence stress and during post-submergence recovery respectively. The top ten hub genes were obtained based on the top 100 (or more than 100) PPI pairs by MCC algorithm. Table S16. KEGG pathway information of the hub genes identified in Table S15. Table S17. The primers used in RT-qPCR analysis. Acknowledgements Not applicable. Abbreviations Differentially expressed genes Reactive oxygen species Gene ontology Kyoto Encyclopedia of Genes and Genomes Peroxidase RNA-Seq by Expectation Maximization Search tool for the retrieval of interacting genes/proteins Pyruvate decarboxylase Alcohol dehydrogenase Gibberellin Protein-protein interaction Wild-Type Authors’ contributions Lijiao Ai and Shengjun Wu conceived and supervised this research, and revised the manuscript. Weitao Jia wrote the original draft, Weitao Jia and Sihan Hou, Songyi Yang and Wenqing Dou performed experiments and analyzed data. Yuhan Zou, Kai Zhu, Cunfeng Zhao, Shihui Zou and Huanhuan Liu helped revise and check the manuscript. All authors read and approved the final manuscript. Funding This work was supported by Chongqing Landscape and Gardening Research Institute (XTZW2023-KF01); the Geological Disaster Patterns and Mitigation Strategies Under River-Reservoir Hydrodynamics in the Three Gorges Reservoir Fluctuation Zone, Chongqing Municipal Bureau of Water Resources (5000002024CC20004); Water Resources Bureau of Chongqing Municipal (CQS23C01036). Data availability The sequencing data has been submitted to the NCBI Sequence Read Archive (SRA) database and the BioProject accession number is PRJNA967854. Declarations Ethics approval and consent to participate The experimental research and all plant materials (either cultivated or wild) were complied with relevant institutional, national, and international guide lines and legislation. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Weitao Jia and Sihan Hou contributed equally to this work. Contributor Information Lijiao Ai, Email: alj01461@foxmail.com. Shengjun Wu, Email: wsj@cigit.ac.cn. References Associated Data Supplementary Materials Supplementary Material 1: Fig S1. Phenotype of bermudagrass under control and submergence stress. (A) Bermudagrass without submergence experience; (B) Bermudagrass after 30 d of complete submergence. Fig S2. Validation of expression profiles of several unigenes by RT-qPCR. (A-D) Differentially expressed unigenes after 24 h submergence in roots (A-C) (R-24 h) and shoots (D) (S-24 h); (E-F) Differentially expressed unigenes after post-submergence recovery in roots (E) (R-Re) and shoots (F) (S-Re). Fig S3. Lignin content of Arabidopsis heterogeneous overexpressing CdPER47 under submergence stress. (A) Lignin content in shoots; (B) Lignin content in roots. Fig S4. The interaction network between proteins encoded by DEGs in roots and shoots after submergence and post-submergence recovery. (A-B) PPI networks under submergence stress (A) and during post-submergence recovery (B) in roots; (C-D) PPI networks under submergence stress (C) and during post-submergence recovery (D) in shoots. The turquoise and orange represent proteins encoded by downregulated and upregulated DEGs respectively. Table S1. Summary of transcriptome sequencing data of C. dactylon. Table S2. DEGs in roots and shoots under submergence stress and during post-submergence recovery. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent DEGs in roots (shoots) under submergence stress and during post-submergence recovery respectively. Table S3. DEGs annotated as transcription factors. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent transcription factors in roots (shoots) under submergence stress and during post-submergence recovery respectively. Table S4. GO terms after submergence and post-submergence recovery in roots. GO terms belonging to “Cellular Component”, “Biological Process” and “Biological Process” were presented respectively. Table S5. GO terms after submergence and post-submergence recovery in shoots. GO terms belonging to “Cellular Component”, “Biological Process” and “Biological Process” were presented respectively. Table S6. KEGG pathway of DEGs in roots and shoots under submergence stress (24 h/CK) and post-submergence recovery stage (Re/24 h). Table S7. DEGs related to ethylene in roots and shoots under submergence stress (24 h/CK) and post-submergence recovery stage (Re/24 h). Table S8. DEGs related to gibberellin in roots and shoots under submergence stress (24 h/CK) and post-submergence recovery stage (Re/24 h). Table S9. DEGs related to sulfur metabolism in roots under submergence stress and post-submergence recovery stage. Table S10. KEGG pathways related to carbohydrate metabolism in roots and shoots under submergence stress and post-submergence recovery stage. Table S11. DEGs annotated as PDC and ADH in roots under submergence stress. Table S12. DEGs related to phenylpropanoid metabolism in roots under submergence stress and post-submergence recovery stage. Table S13. DEGs annotated as peroxidase in roots under submergence stress and post-submergence recovery stage. Table S14. PPI pairs in roots and shoots under submergence stress and post-submergence recovery stage. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent PPI pairs in roots (shoots) under submergence stress and during post-submergence recovery respectively. PPI pairs for the hub genes seeking were highlighted with yellow color. Table S15. The top ten hub genes in roots and shoots under submergence stress and post-submergence recovery stage. R-24 h vs CK and R-Re vs 24 h (S-24 h vs CK and S-Re vs 24 h) represent the top ten hub genes in roots (shoots) under submergence stress and during post-submergence recovery respectively. The top ten hub genes were obtained based on the top 100 (or more than 100) PPI pairs by MCC algorithm. Table S16. KEGG pathway information of the hub genes identified in Table S15. Table S17. The primers used in RT-qPCR analysis. Data Availability Statement The sequencing data has been submitted to the NCBI Sequence Read Archive (SRA) database and the BioProject accession number is PRJNA967854.