Physiological and Transcriptomic Insights into Waterlogging Responses of Liriodendron Hybrids

Abstract

Waterlogging is a major abiotic stress that restricts plant growth, productivity, and survival by disrupting root aeration and altering hormonal homeostasis. To elucidate the physiological and molecular responses associated with flooding tolerance in Liriodendron hybrids (Liriodendron chinense × Liriodendron tulipifera), this study investigated its morphological, physiological, and transcriptomic changes under 0, 1, 3, and 6 days of waterlogging. Roots exhibited rapid decay, while leaves showed delayed chlorosis and reduced chlorophyll content. Changes in antioxidant enzyme activities reflected enhanced antioxidant capacity, with superoxide dismutase (SOD) activity decreasing and peroxidase (POD) and catalase (CAT) activities increasing. Hormone measurements indicated organ-specific patterns, including abscisic acid (ABA) accumulation in leaves and decreased indole-3-acetic acid (IAA) and gibberellin (GA) levels in both roots and leaves. Transcriptome profiling revealed extensive transcriptional adjustments in hormone biosynthesis, signaling, and stress-responsive pathways, including divergent regulation of ABA-associated genes in leaves and roots and broad downregulation of auxin- and gibberellin-related genes. Key ABA biosynthetic genes (NCED1, ABA2) and signaling components (PYL4, PP2C, ABF) were upregulated in leaves but downregulated in roots, whereas auxin (YUC6) and gibberellin (GA20ox) genes were generally suppressed. These coordinated physiological and molecular responses suggest organ-differentiated adaptation to waterlogging in Liriodendron hybrids, highlighting candidate pathways and genes for further investigation and providing insights for improving flooding tolerance in woody species.

1. Introduction

Water is essential for plant survival, but excessive water resulting from waterlogging or flooding disrupts soil–atmosphere gas exchange and imposes severe stress. Climate change in the 21st century has increased the frequency and intensity of extreme weather events [1,2], thereby raising the risk of unprecedented heat waves and flooding [3,4] and threatening ecosystem stability, plant growth, and survival.

Waterlogging severely constrains the growth, distribution, and survival of vegetation worldwide, exerting profound effects not only on natural ecosystems but also on agricultural and forestry productivity [5]. Excessive rainfall often leads to soil waterlogging, which suppresses gas exchange between plant roots and the soil. Under severe hypoxia, plants shift their respiratory pathways, resulting in the accumulation of toxic compounds in roots, root decay, and ultimately impaired aboveground growth [6]. As a major abiotic stress factor, flooding disrupts normal physiological processes through root hypoxia and triggers complex molecular responses at both cellular and genetic levels. Waterlogging stress profoundly alters hormone biosynthesis, transport, and signaling, thereby shaping adaptive responses [7]. The decline in leaf water potential induced by submergence rapidly activates NCED3/5, leading to increased Abscisic acid (ABA) accumulation and stomatal closure [8]. Conversely, root hypoxia promotes ethylene accumulation, which suppresses ABA biosynthesis while enhancing its catabolism, thereby releasing the inhibition of auxin efflux carriers (PIN2/3). This results in local auxin accumulation at the stem base, driving cortical programmed cell death and aerenchyma formation [9]. In addition, ABA suppresses gibberellins (GA) biosynthesis while enhancing its catabolism, leading to DELLA accumulation and inhibition of GA-GID1-SCF signaling. This shift not only restrains growth processes such as internode elongation but also redirects metabolic flux toward stress-responsive pathways, including the antioxidant system [10]. Through this highly integrated hormonal network, plants balance survival and growth under flooding stress. The synergistic interactions of ABA, indole-3-acetic acid (IAA), and GA ultimately shape the flooding tolerance threshold.

Although extensive progress has been made in understanding flooding stress responses, current research has largely focused on crops such as rice [11], maize [12], and soybean [13]. In contrast, studies on trees remain limited, with most investigations centered on morphological and physiological traits, whereas the underlying molecular mechanisms are far less explored. Therefore, elucidating the molecular basis of flood tolerance in trees through transcriptomic approaches is both urgent and essential for advancing the genetic improvement in stress resistance in forest species.

Liriodendron L. (Magnoliaceae) represents a small genus of temperate deciduous trees characterized by straight trunks and distinctive, showy foliage and flowers. As a relic lineage originating from the Tertiary period, the genus comprises only two extant species: Liriodendron chinense native to East Asia and L. tulipifera distributed in North America, forming a well-recognized intercontinental disjunct distribution pattern [14]. In China, L. chinense is classified as endangered due to its low seed germination rate and limited natural regeneration capacity. In contrast, interspecific Liriodendron hybrids (L. chinense × L. tulipifera) exhibits pronounced heterosis, including improved germination success, rapid growth, and enhanced resistance to waterlogging stress [14,15]. These superior traits suggest that Liriodendron hybrids represents a promising germplasm resource for identifying and characterizing stress-responsive genes, particularly those associated with waterlogging tolerance.

Previous studies have mainly focused on morphology and physiology [15,16], leaving the underlying molecular mechanisms largely unexplored. This study integrates morphological observations, physiological assays, and transcriptome sequencing of leaves and roots in Liriodendron hybrids. Our objectives are to (1) elucidate the physiological and molecular responses to flooding and (2) identify key differentially expressed genes involved in waterlogging adaptation, thereby providing a basis for understanding flood tolerance mechanisms.

2. Materials and Methods

2.1. Plant Materials and Waterlogging Treatments

The experimental materials consisted of interspecific hybrids between L. chinense (female) and L. tulipifera (male), from which waterlogging-tolerant hybrids lines were selected. Two-year-old clonal seedlings were propagated through cuttings. Seedlings with uniform growth, averaging 110 cm in height and 1.25 cm in basal stem diameter, were used as experimental materials. Seedling cultivation and waterlogging experiments were conducted at the experimental station in Nanchang, Jiangxi Province, China (E116.027189, N28.378126).

In late March 2021, seedlings were transplanted into permeable nonwoven fabric containers filled with a substrate composed of peat, perlite, and vermiculite. No additional fertilizer was applied during the cultivation or experimental period. In June, 30 Liriodendron hybrids seedlings with similar growth status were selected for waterlogging treatment. During the experimental period, the mean daily air temperature was 29.8 °C, relative humidity was 62%, and photoperiod was 13.9 h. Waterlogging lasted for 6 days, with water maintained at 5 cm above the soil surface by replenishment. Destructive sampling was conducted at 0, 1, 3, and 6 days of waterlogging. Because root sampling required destructive harvest and the availability of uniform hybrid seedlings was limited, a higher-resolution sampling interval could not be implemented. For each time point, three seedlings were sampled, with leaves and roots collected separately. Non-waterlogged seedlings (0 d) served as controls. Samples were labeled as Leaf_CK, Leaf_1d, Leaf_3d, Leaf_6d, Root_CK, Root_1d, Root_3d, and Root_6d. After collection, tissues were immediately frozen in liquid nitrogen and stored at −80 °C for subsequent physiological measurements and transcriptome sequencing.

2.2. Determination of Chlorophyll Content and Antioxidant Enzyme Activities

Chlorophyll (Chl) content was determined using the dimethyl sulfoxide (DMSO) extraction method [17]. Commercial assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) were used to determine the activities of superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT), as well as the contents of malondialdehyde (MDA), proline, and soluble sugars [18]. SOD activity was measured using the hydroxylamine method, POD activity by spectrophotometry, CAT activity by ammonium molybdate method, MDA content by thiobarbituric acid (TBA) assay, proline content by acid-ninhydrin colorimetry, and soluble sugar content by the anthrone method.

2.3. Measurement of Endogenous Hormones

The levels of indole-3-acetic acid (IAA), gibberellins (GA), and abscisic acid (ABA) were quantified using enzyme-linked immunosorbent assay (ELISA) kits (Shanghai Enzyme-linked Biotechnology Co., Shanghai, China). A double-antibody sandwich ELISA was performed following the manufacturer’s protocol. Samples and standards were incubated in microplates pre-coated with specific antibodies, followed by the addition of horseradish peroxidase (HRP)-conjugated detection antibodies. After incubation and washing, tetramethylbenzidine (TMB) substrate was added for color development. The reaction was terminated with acid, and absorbance was measured at 450 nm using a microplate reader. Hormone concentrations were calculated based on standard curves. The kits exhibited acceptable analytical performance: mean spike recoveries were 93%–94% (n = 5, RSD < 7%) and cross-reactivity with structurally related compounds was <5% (manufacturer’s datasheet). Although ELISA is less specific than LC–MS/MS, its sensitivity and reproducibility were sufficient to capture the hormone changes induced by the treatments.

2.4. RNA Extraction, Library Construction, and Sequencing

Total RNA from leaf and root tissues was extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) at a ratio of 1 mL TRIzol per 100 mg fresh tissue, following the manufacturer’s instructions. RNA purity, concentration, and integrity were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and agarose gel electrophoresis. Only RNA samples with an RNA integrity number (RIN) ≥ 7.0 were used for library construction. For each library, 1 µg of high-quality total RNA was used as input material. Eukaryotic mRNA was enriched using oligo(dT) magnetic beads, and no rRNA depletion was performed. The enriched mRNA was fragmented into short fragments using fragmentation buffer, followed by first-strand cDNA synthesis using random hexamer primers and second-strand cDNA synthesis using DNA polymerase I and RNase H. After end repair, A-tailing, and adapter ligation, cDNA fragments of appropriate size were selected using AMPure XP beads (Beckman Coulter, Brea, CA, USA) and amplified by PCR to generate the final sequencing libraries. Sequencing libraries were generated using the NEBNext^® Ultra™ RNA Library Prep Kit (NEB #7530; New England Biolabs, Ipswich, MA, USA) and sequenced on the Illumina NovaSeq 6000 platform (Gene Denovo Biotechnology Co., Guangzhou, China) with paired-end 150 bp reads. Raw reads were processed using fastp v0.18.0 to remove adaptor-containing reads, reads with >10% unknown bases, and low-quality reads (Q ≤ 20 for >50% of bases). The resulting high-quality clean reads were used for downstream analyses.

2.5. De Novo Transcriptome Assembly and Functional Annotation

Clean reads from all samples were pooled for de novo transcriptome assembly using Trinity (v2.14.0), consisting of Inchworm, Chrysalis, and Butterfly modules. Assembly quality was evaluated using N50, GC content, average transcript length, and BUSCO completeness (v5.5.0) against the embryophyta_odb10 database. Clean reads from each sample were mapped back to the assembled transcriptome using Bowtie2 v2.5.1 to calculate mapping ratios, ensuring assembly reliability. According to the de novo assembly report, redundant transcripts were processed and unigenes were defined as the longest transcripts within Trinity gene clusters (as described in the provider’s assembly pipeline). Functional annotation of unigenes was performed using BLASTx implemented in BLAST+ suite (v2.12.0), with an E-value threshold ≤ 1 × 10⁻⁵, against public databases including NCBI Nr, SwissProt, KEGG, and KOG. Gene Ontology (GO) terms were assigned using Blast2GO. Protein domains (Pfam and SMART), CDS prediction (TransDecoder), transcription factor classification (PlantTFDB), and R-gene identification (PRGdb) were also performed to enable comprehensive annotation.

2.6. Gene Expression Quantification and Differential Expression Analysis

Gene expression abundance was quantified using RSEM (v1.3.3) with Bowtie2 alignment. Expression levels were reported as FPKM for descriptive and comparative analyses. Raw counts generated by RSEM were used for differential expression analysis to ensure accurate statistical inference. Differentially expressed genes (DEGs) were identified using DESeq2 (v1.48.1) in R (v4.2.2). DESeq2 modeled count data using a negative binomial generalized linear model with empirical Bayes shrinkage for dispersion estimation. p-values were adjusted using the Benjamini–Hochberg procedure to control the false discovery rate (FDR). Genes with FDR < 0.05 and |log2 fold change| ≥ 1 were considered significantly differentially expressed. To assess overall gene expression patterns and sample reliability, we performed principal component analysis (PCA) and sample-to-sample Pearson correlation analysis. GO and KEGG enrichment analyses of DEGs were conducted using clusterProfiler, with adjusted p-values < 0.05 considered significant.

2.7. Quantitative Real-Time PCR Validation

To validate the RNA-seq results, twelve representative differentially expressed genes (DEGs) were selected for quantitative real-time PCR (qRT-PCR) analysis. Gene-specific primers were designed using Primer-BLAST (NCBI, https://www.ncbi.nlm.nih.gov/tools/primer-blast/ 6 November 2025) and synthesized by Sangon Biotech (Shanghai, China) (Table S1). To ensure specificity within the de novo transcriptome context, primers were designed to target unique unigene regions. The same RNA samples used for RNA sequencing were reverse-transcribed into cDNA using EasyScript^® One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Co., Ltd., Beijing, China), following the manufacturer’s protocol for two-step RT-qPCR. Quantitative PCR reactions were performed using Hieff UNICON^® Universal Blue qPCR SYBR Green Master Mix (Yeasen Biotechnology (Shanghai) Co., Ltd., Shanghai, China) on a CFX Opus 96 Real-Time PCR System (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The thermal cycling program consisted of an initial denaturation at 95 °C for 2 min, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. Melting curve analysis was performed from 65 °C to 95 °C to verify amplification specificity. The expression stability of 18S rRNA [19,20] across all treatments and tissue types was evaluated by analyzing Ct variance and performing one-way ANOVA. No significant difference in Ct values was detected among treatments (p > 0.05), suggesting that 18S rRNA is suitable reference gene under the tested waterlogging conditions. Further validation using geNorm or NormFinder is recommended for more rigorous assessment. Relative expression levels were calculated using the $2^{-\mathsf{\Delta }\mathsf{\Delta }{C}_T}$ method. qRT-PCR expression trends were highly consistent with RNA-seq profiles (Figure S1), and correlation analysis showed a strong linear relationship between the two datasets (R² = 0.926; Figure S2), confirming transcriptome reliability.

2.8. Data Analysis

Statistical analyses of physiological and biochemical data were performed using SPSS Statistics 25 and Origin 2021. Each treatment included three independent biological replicates, and data are presented as mean ± standard error (SE). Significant differences among treatments were determined using one-way analysis of variance (ANOVA), followed by Duncan’s multiple range test. Statistical significance was defined at p < 0.05, and different lowercase letters indicate significant differences among treatments. Degrees of freedom (df), F values, and exact p-values were obtained directly from the ANOVA output in SPSS. For transcriptomic analyses, FPKM values were used for expression visualization and clustering analysis. Heatmaps were generated using hierarchical clustering, with red indicating higher expression levels and blue indicating lower expression levels.

3. Results

3.1. Morphological Responses of Liriodendron Hybrids to Waterlogging Stress

Waterlogging stress exerted progressive effects on the external morphology of Liriodendron hybrids seedlings, with symptoms intensifying as the treatment duration increased. Root systems were the first to be affected, subsequently influencing aboveground growth. At day 1 of waterlogging, root morphology showed no obvious changes. By day 3, mild stress symptoms were observed, with some root tips turning black. After 6 days, roots were extensively darkened and decayed. In contrast, aboveground parts responded more slowly. After 1 day of waterlogging, seedlings remained largely healthy, with only slight yellowing along the margins of some upper leaves. By day 3, apical buds began to droop slightly, accompanied by aggravated chlorosis and wilting in upper leaves. After 6 days, extensive chlorosis developed, petioles exhibited severe drooping, and apical buds wilted, indicating pronounced stress damage (Figure 1).

3.2. Effects of Waterlogging Stress on Photosynthesis (Chlorophyll Content)

Under waterlogging stress, the contents of Chl a, Chl b, total chlorophyll (Chlt), and carotenoids (Car) in Liriodendron hybrids leaves exhibited an overall downward trend (Figure 2). Notably, Chl a and Car showed a significant reduction on day 1 of waterlogging (Figure 2, p < 0.05), whereas Chl b and Chlt did not differ significantly at this time (Figure 2, p < 0.05). At day 6, Chl a, Chl b, and Chlt decreased significantly compared with the control, whereas Car showed a downward trend without statistical significance. Most leaves exhibited visible chlorosis after 6 days of waterlogging.

3.3. Changes in Endogenous Hormones and Antioxidant Enzyme Activities

The ABA content in leaves increased gradually with the duration of waterlogging and was significantly higher than the control on day 6 (Figure 3A, p < 0.05). In contrast, root ABA levels decreased significantly on day 1, indicating a more immediate response in roots, while leaf ABA showed only a slight early increase. Leaf IAA content declined significantly during the early stage of waterlogging (day 1–3) and recovered slightly by day 6 (Figure 3B, p < 0.05), whereas root IAA levels exhibited minimal changes but a gradual downward trend throughout the treatment. GA levels in leaves followed a similar downward trend but with a delayed response compared with IAA (Figure S3, p < 0.05). Collectively, these results suggest that hormone contents in leaves were relatively stable during the early stage of waterlogging, but early reductions in root ABA and dynamic changes in leaf IAA and GA during the mid-to-late stages were associated with alterations in growth and metabolic activity, indicating differential responses between leaves and roots under waterlogged conditions.

The antioxidant enzyme system also displayed significant changes under waterlogging (Figure 3C–E, p < 0.05). SOD activity declined in both leaves and roots, whereas POD and CAT activities exhibited fluctuating upward trends, suggesting a compensatory response to scavenge excessive reactive oxygen species (ROS). MDA content in roots increased steadily throughout the treatment, remaining consistently higher than the control (Figure 3F). In contrast, leaf MDA content decreased significantly on day 3 and remained lower than both the control and early-stage levels by day 6. These results indicate that membrane lipid peroxidation was aggravated under waterlogging, but leaves suffered relatively milder oxidative damage compared with roots during the later stages.

3.4. Transcriptome Data Quality, Assembly, and Annotation

A total of 24 libraries were sequenced, generating high-quality datasets appropriate for de novo transcriptome analysis. After filtering, all samples retained Q20 and Q30 values above 96% and 91% (Table S2). The N50 was 1334 bp, the longest contig reached 18.8 kb, and the GC content was 41.8% (Table S3). BUSCO analysis against the embryophyta_odb10 lineage indicated 72.9% completeness (1050 of 1440 BUSCOs), with 69.0% present as single-copy and only 3.9% duplicated; 13.7% were fragmented and 13.4% missing (Table S4). Functional annotation revealed that 42,277 unigenes (37.12%) matched at least one public database. Overall, the high sequencing quality, robust assembly, and comprehensive annotation jointly provide a reliable dataset for exploring transcriptional regulation under waterlogging stress in Liriodendron hybrids.

3.5. Differential Gene Expression Dynamics Under Waterlogging

Differentially expressed gene (DEG) numbers varied between tissues and over time (Figure 4). In roots, DEGs rose progressively from 1081 at 1 d to 17,119 at 6 d, whereas leaves exhibited only 888 DEGs at 1 d but peaked at 13,035 by 6 d. In general, root tissues exhibited more DEGs than leaf tissues and the number of DEGs increased with prolonged treatment. KEGG enrichment revealed that plant hormone signal transduction, phenylpropanoid biosynthesis and starch and sucrose metabolism were consistently enriched in both leaf and root tissues, with enrichment intensity increasing over treatment time (Figure S4).

3.6. Expression of Key Genes in Hormone Biosynthesis

To investigate transcriptional changes in hormone-related genes under waterlogging stress, a heatmap was constructed based on the z-score–normalized expression profiles of 13 selected unigenes associated with ethylene, ABA, auxin, and GA biosynthesis and transport in roots and leaves (Figure 5). Under waterlogging, ethylene biosynthesis genes (ACO1, ACS1, ACS3, ACS7) were strongly upregulated in roots at early time points (1–3 d), whereas their expression in leaves was relatively moderate, suggesting an early ethylene response in roots. For ABA biosynthesis, ABA2 was upregulated in leaves at 1 and 3 days of waterlogging while showing slight downregulation in roots during the same period. 9-cis-epoxycarotenoid dioxygenase 1 (NCED1) expression increased in leaves at day 3, whereas it showed a decreasing trend in roots. Auxin biosynthesis genes of the YUCCA(YUC) family exhibited divergent expression patterns: YUC4 was upregulated in roots, whereas YUC5 and YUC6 were downregulated. Gibberellin 20-oxidase genes (GA20ox2 and GA20ox3) showed reduced expression in roots under waterlogging.

3.7. Auxin Signaling Pathway

KEGG enrichment analysis revealed that 25 differentially expressed genes (DEGs) in leaves and 31 DEGs in roots were enriched in the auxin signaling pathway (Figure 6). Most auxin signaling genes were downregulated during waterlogging, with only a few showing upregulation. In leaves, small auxin up RNA 71 (SAUR71) was upregulated on days 1 and 3, SAUR36, GH3.1, and GH3.2 were upregulated on days 3 and 6, and GH3.6 was significantly upregulated on day 6. In roots, IAA13 and IAA30 were consistently upregulated with increasing stress duration. Overall, most auxin signaling genes were suppressed under waterlogging, in line with the observed decrease in endogenous auxin levels, suggesting that inhibition of auxin signaling contributed to growth suppression.

3.8. ABA Signaling Pathway

A total of 12 DEGs in leaves and 7 DEGs in roots were enriched in the ABA signaling pathway (Figure 7). In leaves, most genes in this pathway were gradually upregulated with prolonged waterlogging, consistent with the increasing ABA levels. For example, PYL4, several protein phosphatase 2C (PP2C) genes, SAPK genes, and ABF transcription factors were generally upregulated, except SAPK10, which was downregulated. In roots, most ABA signaling genes were downregulated over time, although a few genes (e.g., PP2C and PYL4) showed transient or sustained upregulation. These results suggest that ABA signaling was more active in leaves than in roots, indicating that leaves played a more central role in ABA-mediated waterlogging responses.

3.9. GA Signaling Pathway

Compared with auxin and ABA, fewer DEGs were enriched in the GA signaling pathway, with five identified in both leaves and roots (Figure 8). In leaves, the GA receptor gene GID1C showed an overall upward trend, and the DELLA gene GAI1 maintained high expression levels during waterlogging. In roots, GID1A and GID1C were both upregulated under stress. These findings suggest that GA signaling was suppressed at the biosynthetic level but partially compensated by receptor-level regulation.

4. Discussion

4.1. Organ-Specific Differences Between Leaves and Roots

Under waterlogging stress, the leaves and roots of Liriodendron hybrids exhibited distinct physiological, biochemical, and transcriptional responses. Roots, as the primary organs for water and nutrient uptake, are directly exposed to hypoxic and oxidative stress, as well as restricted nutrient acquisition. Oxygen deficiency suppresses mitochondrial aerobic respiration, thereby impairing energy metabolism. This may lead to changes in cellular redox balance, as suggested by altered antioxidant enzyme activities, potentially disrupting cellular homeostasis [21]. By contrast, leaves, as the major sites of photosynthesis, display regulatory mechanisms that differ markedly from roots, particularly in the metabolism and signaling of endogenous hormones such as ABA, ethylene, and IAA. However, ethylene content was not directly measured in this study, representing a limitation in fully characterizing hormonal responses.

For example, ABA in leaves primarily functions to regulate stomatal closure and reduce transpirational water loss [22,23], whereas in roots it plays a more critical role in hypoxia responses and antioxidant defense [24]. In the present study, ABA, IAA, and GA levels in roots declined significantly under waterlogging, consistent with the transcriptional downregulation of their biosynthetic genes. This suggests that rapid onset of hypoxia and osmotic stress may have inhibited hormone biosynthesis and metabolism in roots. By contrast, leaves displayed a delayed but more pronounced hormonal response: ABA levels increased significantly while IAA and GA contents decreased, accompanied by the upregulation of ABA biosynthetic genes such as ABA2 and NCED1. Conversely, the expression of GA20ox and YUC genes was markedly downregulated in roots. Similar organ-specific transcriptional reprogramming under waterlogging has been reported in Hordeum marinum and rice [22]. Collectively, these findings indicate that roots suffer immediate suppression of hormone synthesis due to hypoxia, whereas leaves have greater temporal and metabolic flexibility to adjust hormone biosynthesis and signaling. However, the functional consequences of these hormonal changes at the physiological level require further validation. Therefore, identifying stress-responsive genes in leaves may provide more promising targets for improving waterlogging tolerance in Liriodendron hybrids.

4.2. Changes in Phytohormone Contents and Key Differentially Expressed Genes in Hormone Biosynthesis Under Waterlogging Stress

ABA regulates multiple physiological processes in plants and plays a central role in responses to water-related stresses [25,26]. In soybean, a reduction in ABA levels promotes aerenchyma formation [27]. In this study, ABA content in Liriodendron hybrids leaves increased after 1 day of waterlogging, slightly decreased at day 3, and reached its highest level at day 6, a trend consistent with findings in other species [28]. Although gas-exchange parameters such as stomatal conductance and transpiration were not directly measured, the overall increase in leaf ABA suggests a potential involvement of ABA signaling in regulating leaf physiological responses under prolonged flooding. It should be noted that hormone quantification was performed using ELISA, which may have limited specificity, particularly for auxin and gibberellin; therefore, mechanistic interpretations based on these measurements should be considered with caution. Therefore, any inference regarding ABA-mediated stomatal regulation should be considered indirect.

In contrast, root ABA levels declined rapidly under waterlogging, similar to patterns reported in citrus [29]. As the duration of flooding increased, root damage impeded normal physiological activity. Moreover, the decline in root ABA may be associated with developmental adjustments such as adventitious root formation, which has been reported in other species under flooding stress [30]. In the ABA biosynthetic pathway, 9-cis-epoxycarotenoid dioxygenase (NCED) acts as a key rate-limiting enzyme whose expression is induced by abiotic stress. For example, PbrWRKY53 in pear enhances drought tolerance by upregulating NCED1/3 [31], while overexpression of PdNCED3 in Populus davidiana increases ABA accumulation and osmotic stress tolerance [32]. This mechanism is conserved across species, as NCEDs is strongly upregulated under osmotic stress in Poplar [33], Stipa purpurea [34] and watermelon [35]. Our results show that waterlogging strongly induced NCED expression in Liriodendron leaves, where transcript levels were much higher than in roots. In contrast, NCED expression in roots declined. In addition, the ABA biosynthetic gene ABA2 was also highly expressed in leaves. These findings suggest that transcriptional regulation of ABA biosynthesis in Liriodendron under waterlogging predominantly occurs in leaves rather than roots.

Gibberellins (GA) are key regulators of plant growth and development [36]. Under abiotic stress, GA levels generally decrease to limit growth and enhance stress tolerance [37,38]. In this study, GA content and GA20-oxidase (GA20ox) gene expression in both leaves and roots initially increased and then decreased under flooding. GA20ox genes catalyze late steps of GA biosynthesis and directly regulate active GA levels [39,40]. In Arabidopsis, drought stress significantly suppressed GA20ox expression [41], while in poplar, the drought-induced transcription factor PagKNAT2/6b suppressed GA20ox1 expression to reduce GA synthesis and remodel plant architecture for water-deficit adaptation [42]. Notably, crosstalk exists between hormone pathways: auxin homeostasis can influence ABA synthesis, and the balance between these hormones is critical for stress adaptation [43].

Auxin itself plays essential roles in growth and stress responses [37,38]. In Liriodendron hybrids, leaf IAA levels declined during the early stage of waterlogging (day 1–3) and recovered slightly by day 6, whereas root IAA exhibited minimal changes but a gradual downward trend. The YUCCA (YUC) family of flavin monooxygenases (e.g., YUC1, YUC2, YUC4, and YUC6) encode rate-limiting enzymes in auxin biosynthesis, exerting spatiotemporal control of auxin production [44,45,46]. Overexpression of YUC genes enhances auxin synthesis [44], whereas loss-of-function mutants exhibit severe developmental defects [47]. Moreover, YUC-mediated auxin biosynthesis contributes to stress adaptation: heterologous expression of Arabidopsis YUC6 in poplar improved drought tolerance by reducing ROS levels [48]. Similarly, FUSCA3 (FUS3) and LEAFY COTYLEDON2 (LEC2) activate YUC4 to mediate local auxin biosynthesis and regulate lateral root development [49]. In our study, YUC4 was strongly upregulated in roots, which may be associated with local auxin biosynthesis during root system adjustment. However, without direct anatomical or auxin transport evidence, its role in regulating lateral root formation under waterlogging remains speculative. Overall, changes in endogenous hormone contents suggest that Liriodendron hybrids gradually adjust hormone homeostasis under prolonged flooding, characterized by increased ABA and reduced GA and IAA levels in leaves. This hormonal pattern is consistent with growth inhibition and metabolic downregulation, which are commonly regarded as adaptive strategies to conserve resources under abiotic stress [50].

4.3. Differentially Expressed Genes in Hormone Signaling Pathways

In the ABA signaling pathway of leaves, genes encoding the ABA receptor PYL4, type 2C protein phosphatases PP2C (PP2CA and PP2C06), sucrose non-fermenting 1-related protein kinases SAPK (SAPK1, SAPK2, and SAPK3), and transcription factors ABF (ABF2 and ABF4) were all gradually upregulated with increasing waterlogging duration. In roots, however, only PP2CA was consistently upregulated, while PYL1 and PYL4 was induced at day 3, and other signaling genes were generally downregulated over time. The rapid accumulation of ABA in leaves under flooding allows for binding to PYR/PYL/RCAR receptors, inducing conformational changes that enhance interaction with PP2Cs, thereby suppressing their phosphatase activity and releasing the inhibition of SnRK2 kinases [51]. Activated SnRK2s subsequently induce ABF expression, enhancing stress tolerance [52]. Overexpression of the AP2/ERF gene RAP2.6L in Arabidopsis enhanced ABA biosynthesis and signaling during waterlogging, promoting antioxidant defense, stomatal closure, and improved flooding tolerance [53]. Previous studies showed that PP2C genes participate in drought responses in Arabidopsis, maize, and tomato [54,55]. Overexpression of PP2C in Arabidopsis improved abiotic stress tolerance and ABA sensitivity. In our study, PP2Cs was upregulated in both leaves and roots during waterlogging, suggesting its critical role in flood stress adaptation. Overall, most ABA signaling genes in leaves were upregulated under waterlogging, indicating that Liriodendron hybrids enhances flood tolerance through positive regulation of ABA signaling.

In the auxin signaling pathway, auxin perception by the receptor TIR1 promotes ubiquitination of AUX/IAA repressors, thereby releasing auxin response factors (ARFs) to activate downstream genes [56]. In leaves, GH3.1 and GH3.2 were upregulated at day 3 and day 6, while GH3.6 was strongly induced at day 6. Since leaf IAA levels remained consistently lower than controls, the induction of GH3 genes suggests enhanced IAA conjugation, which may contribute to shoot growth suppression under waterlogging. In roots, the auxin-responsive genes IAA13 and IAA30 were upregulated with flooding duration, potentially reflecting localized auxin signaling adjustments [57]. Collectively, most auxin signaling genes in both leaves and roots were downregulated under waterlogging, consistent with the overall decline in IAA content. Although PIN-mediated polar auxin transport has been reported to contribute to root adaptation to hypoxia in other plant species, the expression or localization of PIN and AUX/LAX transporters was not directly assessed in the present study. Accordingly, the potential involvement of directional auxin transport in Liriodendron hybrids is discussed here in a hypothesis-driven context rather than as a confirmed mechanism.

5. Conclusions

This study compared hormone dynamics and the expression of related genes in leaves and roots of Liriodendron hybrids under waterlogging stress, revealing clear organ-specific response patterns to hypoxic conditions. Roots exhibited rapid suppression of hormone biosynthesis and signaling, whereas leaves showed a delayed but sustained activation of ABA accumulation and ABA-related gene expression during prolonged stress. Together, these results indicate an organ-partitioned waterlogging response strategy in Liriodendron hybrids, characterized by rapid root-level inhibition and delayed leaf-level hormonal regulation. The hormone-associated genes identified in this study should be regarded as candidate genes linked to waterlogging tolerance. While their expression patterns suggest potential regulatory roles, functional validation will be required to confirm their direct involvement. This work therefore provides a foundation for future functional studies and targeted improvement in flood tolerance in Liriodendron.