Transcriptome-Wide Survey of LBD Transcription Factors in Actinidia valvata Under Waterlogging Stress and Functional Analysis of Two AvLBD41 Members

Abstract

Actinidia valvata, a promising rootstock for kiwifruit cultivation, demonstrates superior waterlogging tolerance compared with commercial cultivars. Lateral organ boundaries domain (LBD) transcription factors (TFs) are known to be pivotal in plant responses to abiotic stress. Nevertheless, the characterization of the LBD family under waterlogging stress in A. valvata remains limited. In this study, 26 AvLBD genes were identified from a transcriptome dataset, with the majority classified into phylogenetic Class II. Under waterlogging stress, transcript accumulation of most AvLBD41 members, particularly AvLBD41_11 and AvLBD41_7, was markedly increased in roots. Bimolecular fluorescence complementation (BiFC) assays indicated that AvLBD41_7 heterodimerizes with both the AP2/ERF activator AvERF75 and the trihelix repressor AvHRA1, whereas AvLBD41_11 only interacts with AvERF75. Neither AvLBD41 isoform interacts with AvERF73, thereby defining distinct components of a waterlogging-responsive module. Yeast-based assays revealed an absence of transactivation activity for AvLBD41_7, and transient expression analyses confirmed its exclusive nuclear localization. The promoters of both AvLBD41_11 and AvLBD41_7 harbor numerous cis-elements responsive to hormones and abiotic stresses. An AvLBD41_7-derived PCR marker could be used to distinguish A. valvata from A. deliciosa accessions. Collectively, these findings provide a comprehensive functional annotation of the LBD gene family in A. valvata and establish AvLBD41_7 as a potential molecular target for future kiwifruit breeding programs aimed at waterlogging resilience.

1. Introduction

Kiwifruit, also known as Chinese gooseberry, originated in China and is now prized worldwide for its unique flavor and high nutritional value [1]. As the global demand has grown, China has expanded kiwifruit plantings into diverse agro-climates, including high-humidity regions [2]. Climate change, however, has increased the frequency of extreme rainfall events that leave soils waterlogged and oxygen-deficient for days or even weeks [3]. Roughly 16% of the world’s cropland is afflicted by waterlogging, cutting yields by about one-fifth [4]. Waterlogging stress suppresses root function, reduces fruit quality and yield, and causes huge economic losses in kiwifruit production [5]. Hence, breeding tolerant cultivars is necessary for safeguarding the long-term sustainability of the kiwifruit industry.

Most commercial kiwifruit cultivars belong to A. chinensis var. deliciosa or A. chinensis var. chinensis. These cultivars are highly sensitive to waterlogging stress because the resultant hypoxia rapidly suppresses root aerobic respiration, triggering an acute cellular energy crisis that precedes root death [6]. Consequently, grafting elite but waterlogging-sensitive scions onto tolerant rootstocks has become a key strategy for mitigating yield losses imposed by transient soil saturation [7]. In current orchards, A. chinensis var. deliciosa seedlings are the most commonly used rootstocks; however, they remain sensitive to waterlogging stress [8]. By contrast, A. valvata develops a robust, woody, highly aerated root system and survives prolonged waterlogging far better than A. deliciosa [9,10]. Previous studies demonstrated that A. valvata maintains carbohydrate reserves and restricts oxidative damage in roots during prolonged waterlogging stress [11,12]. One adaptive morphology trait is the rapid development of adventitious roots, structures that enable internal oxygen diffusion under hypoxic conditions [12,13]. The superior waterlogging tolerance is maintained when A. valvata is used as a rootstock for commercial scions [9]. A. valvata rootstock can also boost scion vigor, fruit size and quality, highlighting its potential to raise yields [14]. As A. valvata rootstocks are increasingly adopted in kiwifruit orchards, a comprehensive dissection of the molecular mechanisms underlying their waterlogging tolerance becomes imperative.

In Arabidopsis thaliana, Group VII ERFs (ERF-VII) orchestrate the core hypoxia response by trans-activating genes such as ADH1, PDC1, and LBD41 via the hypoxia-responsive promoter element (HRPE) [15]. Transcriptome surveys in A. valvata implicate AP2/ERF, C3H, LBD, MYB, Trihelix, and WRKY TFs in the waterlogging response [16]. Among the TFs, the ERF-VIIs AvERF73 and AvERF75 were strongly up-regulated under waterlogging stress, and their overexpression improves kiwifruit survival [17,18]. Strikingly, several AvLBD41 unigenes are co-induced, physically interact with AvERF75 [18], and thus represent their immediate involvement in the ERF-VII pathway. Yet, the roles of LBD proteins during waterlogging stress remain largely uncharacterized in kiwifruit. Transcriptome-wide identification and functional screening of key LBD genes are helpful for elucidating the waterlogging-adaptive signaling in kiwifruit.

The LBD TFs are characterized by a conserved N-terminal lateral organ boundaries domain and a highly divergent C-terminal region [19]. It consists of a zinc finger-like motif (CX₂CX₆CX₃C) responsible for DNA binding, a Gly-Ala-Ser stretch that stabilizes promoter interaction, and a leucine zipper-like repeat (LX₆LX₃LX₆L) that mediates protein dimerization [20]. Based on structural features, LBD TFs are classified into two subfamilies: Class I, which retains all three domains, and Class II, which contains only the zinc finger-like motif [21,22]. Genome-wide analyses have identified LBD members in apple, grape, rice, tomato, and other crops, demonstrating extensive conservation that facilitates functional inference across taxa [23,24,25,26]. In kiwifruit, however, the LBD family remains not systematically characterized.

LBD proteins regulate a wide range of plant processes, including growth, development, and stress adaptation [27]. In Arabidopsis, AtLBD16 and AtLBD18 interact with auxin response factors to drive lateral root emergence [28]. In peach, LBD genes (PpBSBRLs) function as positive regulators of lateral and adventitious root initiation [29]. In rice, LBD12-1 protein is essential for adventitious root formation and was regulated by auxin signaling under submergence stress in coleoptiles [30]. Collectively, these data underscore the pivotal roles of the LBD family in regulating root system architecture. In Arabidopsis, transcript levels of LBD4, LBD37, LBD39, LBD40 and LBD41 were significantly altered under hypoxic stress [31]. Among them, AtLBD41 regulates the expression of AtHB1, a non-symbiotic hemoglobin gene, under hypoxia [32]. Nevertheless, the homozygous T-DNA insertional mutants of AtLBD41 (SALK_078678C) display no visible phenotype relative to Col-0 in low-oxygen assays [33]. Although constitutive over-expression of AtLBD41 alters floral architecture and root nitrogen metabolism [34,35], its influence on waterlogging tolerance remains essentially unexplored. In soybean and Populus deltoides, LBD41 and LBD1 are rapidly induced in leaves and roots under waterlogging stress, respectively [36,37]. Likewise, AcLBD was strongly up-regulated in leaves of the waterlogging-tolerant A. polygama genotype ‘Zhemizhen 1’, upon waterlogging stress [38]. LBD proteins can form functional complexes with diverse partners to coordinate plant growth and development [39,40]. Despite these findings, the functional roles and regulatory network of LBDs under waterlogging stress, particularly in kiwifruit, remain limited.

Because LBD genes control adventitious rooting and low-oxygen signaling across taxa, we hypothesized that AvLBD members are key players in the waterlogging tolerance of A. valvata. Therefore, a transcriptome-wide identification of the AvLBD gene family was conducted, and their expression dynamics under waterlogging stress were quantified. Two unigenes, AvLBD41_11 and AvLBD41_7, exhibited the strongest up-regulation after 12 h of waterlogging, and were subjected to detailed functional dissection, including subcellular localization, transactivation activity, protein–protein interaction (PPI), and promoter features. This study provides a comprehensive functional annotation of the AvLBD gene family and establishes AvLBD41_7 as a central regulator within a waterlogging-specific transcriptional module, thereby providing new insights into the molecular framework underlying waterlogging tolerance in kiwifruit. An AvLBD41_7-based PCR marker was used to accurately distinguish A. valvata from A. deliciosa across an initial panel of 11 accessions, underscoring a prime candidate gene target for genotyping and marker-assisted breeding of waterlogging-tolerant cultivars.

2. Materials and Methods

2.1. Identification of A. valvata LBD Genes

A chromosome-scale genome assembly with systematic annotation of A. valvata remains unavailable; consequently, full-length transcriptome sequencing can partially characterize the LBD gene family. The full-length transcriptomic dataset of A. valvata genotype KR5 (accession number PRJNA796628) was obtained from a previous study [41]. Putative LBD protein sequences were retrieved from annotation results and screened for the presence of the LOB domain (Pfam PF03159) using the ScanProsite online tool (accessed on 4 October 2025, https://prosite.expasy.org/scanprosite/) with default parameters. After removing duplicate (100% identity) and incomplete sequences (lacking a full open reading frame, pfam score < 15.0, or domain coverage < 83.0%), the final candidate sequences were validated with the NCBI Conserved Domain Search (CD Search v3.21, accessed on 4 October 2025, https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) with default parameters using an E value cut-off of 10 -2. The CD Search output was then imported into TBtools v2.363, and the conserved LOB domain was visualized using the ‘Visualize Domain Pattern’ command [42]. Fundamental properties of AvLBD proteins, including amino acid length, molecular weight (MW), and theoretical isoelectric point (pI), were calculated using ExPaSy-ProtParam (accessed on 5 October 2025, https://web.expasy.org/protparam/) with default parameters. Subcellular localization was predicted with Cell-PLoc 2.0 (accessed on 5 October 2025, http://www.csbio.sjtu.edu.cn/bioinf/Cell-PLoc-2/) using the web-server for Plant-mPLoc predictor with default parameters. Detailed information on the AvLBD proteins, including their ID, nucleotide sequence, and protein sequence, is provided in Supplementary Table S1.

2.2. Phylogenetic Tree Construction

LBD sequences from Arabidopsis and rice were retrieved from the Plant Transcription Factor Database (accessed on 5 October 2025, http://planttfdb.gao-lab.org/index.php), and their ID and protein sequence are provided in Supplementary Table S2. Multiple sequence alignment of all LBD proteins from A. valvata, Arabidopsis, and rice was conducted using ClustalW with default parameters (gap open penalty equal to 10.0, gap extend penalty equal to 0.10, and the Gonnet protein weight matrix). A phylogenetic tree was generated in MEGA 11.0 with the maximum likelihood (ML) method and 1000 bootstrap replicates. The Jones-Taylor-Thornton model was used as the amino acid substitution model, with 5 discrete gamma categories for rate variation among sites, and partial deletion was applied to handle gaps and missing data. The random seed and algorithm used for the bootstrap were set to the default values. The resulting unrooted evolutionary tree was visualized with iTOL (accessed on 5 October 2025, https://itol.embl.de/) [43]. The AvLBD genes were systematically renamed AvLBD1-AvLBD41 based on their homology with the corresponding LBD proteins of Arabidopsis.

2.3. Multiple Sequence Alignment and Conserved Motif Analysis

Conserved domains of multiple LBD protein sequences were aligned in DNAMAN 9.0 (Lynnon BioSoft) using the ClustalW algorithm (Feng-Doolittle and Thompson) for optimal alignment. Putative motifs were subsequently identified with the MEME online suite (accessed on 5 October 2025, http://meme-suite.org/tools/meme) under the following parameters: minimum width 6, maximum width 50, and a maximum of 10 motifs. The zero or one occurrence per sequence model was used to expect motif sites which are distributed in sequences. The MEME XML output was then imported into TBtools v2.363, and the motifs were visualized using the ’Visualize Motif Pattern’ command [42].

2.4. Transcriptomic Data Analysis of AvLBDs Under Waterlogging Stress

In a previous experiment, asexually propagated KR5 plants at the five-to-six leaf stage were subjected to waterlogging for 0, 12, 24 or 72 h. The 0 h treatment served as the non-waterlogged control. Two potted plants were immersed in each plastic tray (45 cm × 35 cm × 16 cm) and the water level was kept 2–3 cm above the soil surface. Every time point comprised four pots, and the experiment was replicated three times. Total RNA was isolated from root samples with the RNeasy Mini Kit (Qiagen, Germantown, MD, USA) and quantified for downstream library construction. Strand-specific libraries were prepared with the NEBNext^® Ultra™ RNA Library Prep Kit (San Diego, CA, USA) following the manufacturer’s instructions. Paired-end sequencing (2 × 150 bp) was performed on an Illumina HiSeq 2000 (San Diego, CA, USA). Raw reads were quality-assessed with FastQC v0.11.9 and technical sequences were trimmed using Cutadapt (version 1.9.1, TU Dortmund University, TU Dortmund, Germany). RNA-seq raw reads have been deposited in the NCBI Sequence Read Archive under BioProject PRJNA792211 [41]. Clean reads were then aligned to the reference full-length transcriptome (accession number PRJNA796628) with HISAT2 v2.0.4, and transcript abundances were quantified with RSEM v1.2.12 [41]. The differential expression analysis was performed using DESeq2 v1.4.5 at a Q value ≤ 0.05. A heatmap of AvLBD expression was generated in TBtools v2.363 using log2-transformed Fragments Per Kilobase per Million mapped fragments (FPKM) values.

2.5. PPI Prediction Analysis

Homologous proteins from A. chinensis were employed to predict the PPI network of AvLBD and other waterlogging-induced proteins using the STRING online tool version 12 (https://string-db.org/) [44]. Default parameters were applied: full STRING network type with network edges representing evidence. Functional enrichment visualization of the STRING network displayed 10 terms in the Biological Process category (GO, Gene Ontology), with terms clustered at a similarity threshold of ≥0.8.

2.6. Transactivation Assay in Yeast Cells

The coding sequence of AvLBD41_7 (819 bp) was inserted into the pGBKT7 vector at the EcoRI/BamHI sites under control of the T7 promoter, and the resulting recombinant plasmid was employed to assess transcriptional activation. The BD empty vector and the BD-AvERF73 fusion plasmid served as negative and positive controls, respectively [45]. All plasmids were transformed into the yeast strain AH109 and cultured on SD/-Trp and SD/-Trp/-His/-Ade media for 3 days at 30 °C. Transcriptional activation activity was evaluated based on the growth status of yeast cells. Primers are provided in Supplementary Table S3.

2.7. Subcellular Localization Analysis

The AvLBD41_7 CDS was amplified with gene-specific primers (Table S3) using 2 × Phan-Q5 SuperMix (KERMEY, Zhengzhou, China) and ligated into pCAMBIA-super1300-GFP at the KpnI/SalI sites with the ClonExpress^® II One-Step Cloning Kit (Vazyme Biotech, Nanjing, China). The resulting 35S::AvLBD41_7-GFP construct was introduced into Agrobacterium tumefaciens strain EHA105 by chemical transformation, and the bacteria were infiltrated into Nicotiana benthamiana leaves. After 3 days of greenhouse recovery, epidermal strips of leaves were mounted in water and imaged using a laser-scanning confocal microscope (Zeiss LSM 900, Jena, Germany).

2.8. BiFC Assay

The full-length CDSs of AvLBD41_7, AvLBD41_11, AvHRA1, AvERF73, and AvERF75 were cloned into the pCAMBIA1300-nYFP and pCAMBIA1300-YFPc vectors at the BamHI/SalI sites, generating AvLBD41_7-NY/YC, AvLBD41_11-NY/YC, AvHRA1-NY, AvERF73-YC, and AvERF75-YC. The recombinant plasmids were introduced into Agrobacterium tumefaciens strain EHA105, and all plasmid combinations were co-infiltrated into tobacco leaves for 48–72 h. Yellow fluorescence was subsequently visualized using a Leica TCS SP8 confocal laser-scanning microscope (Leica, Bensheim, Germany). The relevant primers are provided in Supplementary Table S3.

2.9. Promoter Cloning, Cis-Elements and Phylogenetic Analysis

By mining the genome database [46], AvLBD41_11 and AvLBD41_7 were found to share significant homology with Amadzw1x02g026630 and Amadzw1x03g047470, respectively, and primers were subsequently designed from these aligned sequences (Table S3). Fragments containing the promoter regions upstream of the initiation codons of AvLBD41_11 and AvLBD41_7 from A. valvata genomic DNA were amplified, respectively. The purified PCR products were cloned into the pMD18-T vector (TaKaRa, Dalian, China) and sequenced. Promoter analysis was performed with PlantCARE [47], and the results were visualized with TBtools v2.363 using the ‘Basic Biosequence View’ command [42]. Using the AvLBD41_7 promoter as the query, BLASTX 2.17.0+ searches against the 30 kiwifruit genome assemblies in the Kiwifruit Genome Database retrieved the most similar homologous promoters (Table S4). Promoter sequences were aligned with ClustalW, and a phylogenetic tree was built in MEGA 11.0 as described above.

2.10. Quantitative Real-Time PCR (qRT-PCR) Analysis of Transcript Levels

To validate the RNA-seq differential-expression results, we quantified six representative LBD genes by qRT-PCR using the same root samples collected at 0, 12, 24 and 72 h of waterlogging (RNA-seq BioProject PRJNA792211). First-strand cDNA was synthesized with the High-Capacity cDNA Reverse Transcription Kit (TOYOBO, Osaka, Japan). Actin primers were taken from a previous study [48]. Reactions were run with 2 × SYBR Green qPCR Premix Kit (KERMEY, Zhengzhou, China) and followed the procedure: initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. For tissue-specific expression of AvLBD41_11 and AvLBD41_7, total RNA was isolated from leaf, petiole, stem and stem-end tissues with the RNAprep Pure Plant Plus Kit (TIANGEN, Beijing, China) and subjected to the same qRT-PCR protocol. The relative expression levels of the tested genes were calculated via the 2^−ΔΔCt method. Primer sequences are provided in Supplemental Table S3. qRT-PCR was performed with three biological replicates and three technical replicates per sample, and each replicate comprised pooled material from eight plants. Expression differences across the four time points were evaluated by one-way ANOVA test (p < 0.05) using SPSS v19.0.

2.11. PCR Marker Development and Test

Polymorphic amplicons were predicted from the AvLBD41_7 DNA sequence and species-specific primers (Table S3) were designed with Primer Premier v5.00. The initial test includes 4 accessions of A. valvata and 7 accessions of A. deliciosa. Reactions were run with 2 × Mega PCR mix (KERMEY, Zhengzhou, China) and followed the procedure: initial denaturation of 3 min at 95 °C, followed by 35 cycles of 98 °C for 10 s, 58 °C for 30 s and 72 °C for 2 min, followed by a final 5 min elongation step. PCRs were conducted with a total reaction volume of 50 µL (1 × PCR mix, 0.25 µM of each primer, 50–100 ng of template DNA, and make up to 50 µL with nuclease-free water). Amplified products were separated on a 1.5% TAE/ethidium bromide agarose gel for 60 min at 85 V and imaged with UV transillumination.

3. Results

3.1. Transcriptome-Wide Identification of LBD Genes in A. valvata

By analyzing the annotated sequences from the full-length transcriptome dataset, 48 putative LBD candidate genes were initially identified. After removing redundant and incomplete protein sequences lacking the conserved LOB domain (Pfam PF03195), a final set of 26 AvLBD family members was confirmed. All 26 AvLBD proteins possess a highly conserved LOB region at the N-terminus of approximately 100 amino acids, and the positions of the LOB domains were validated using the CD Search (Figure 1).

The amino acid length, MW, and pI of AvLBD proteins ranged from 94 to 295 amino acids, 10.16 to 32.86 kDa, and 5.44 to 9.33, respectively. Instability index analysis indicated that 25 of the 26 AvLBD proteins were unstable in vitro. The grand average of hydropathicity (GRAVY) values were negative for all 26 AvLBD proteins, suggesting that they are predominantly hydrophilic. Cell-PLoc subcellular localization predictions indicated that all AvLBD proteins are localized in the nucleus.

3.2. Phylogenetic Analysis of AvLBD Genes

To investigate the evolutionary relationships between AvLBDs and LBDs from other species, the amino acid sequences of 26 A. valvata LBDs, 43 Arabidopsis LBDs, and 36 rice LBDs were used to construct an ML phylogenetic tree (Figure 2). Based on this analysis, a total of 105 LBD proteins from the three species were classified into seven subgroups: Class Ia, Ib, Ic, Id, Ie, IIa, and IIb. Seventeen of the 26 AvLBD proteins clustered within Class II, which was larger than Class I. Within Class I and Class II, the Ia and IIb subgroups contained the largest proportion of AvLBD members. The AvLBD genes were systematically renamed AvLBD1-AvLBD41 according to their homology with the corresponding LBD proteins of Arabidopsis (Table S1).

3.3. Multiple Sequence Alignment and Conserved Motif Analysis of AvLBD Genes

The conserved domains of 26 AvLBD proteins were examined through multiple sequence alignment using DNAMAN software. The analysis revealed that all AvLBD proteins contain the canonical CX₂CX₆CX₃C zinc finger-like motif, whereas the GAS block and the LX₆LX₃LX₆L leucine zipper-like domain were restricted to Class I members (Figure 3a). To further investigate the functional architecture of AvLBD proteins, conserved motifs were identified with the MEME tool. Ten conserved motifs, designated Motifs 1–10, were detected across the AvLBD proteins (Figure 3b). Motif 3 corresponded to the CX₂CX₆CX₃C zinc finger-like signature of the LBD family and, consistent with the alignment results, was present in every AvLBD protein. Motif 8 was specific to Class I, while Motifs 2, 4, 7, and 9 were restricted to Class II.

3.4. Expression Patterns of AvLBDs in Waterlogged Roots

RNA-seq profiling of 26 AvLBDs in waterlogged roots revealed three distinct expression clusters (Figure 4a). Within Cluster I, seven members were orthologous to AtLBD41 and one to AtLBD39. All Cluster I AvLBDs were strongly up-regulated at 12, 24, and 72 h, indicating their involvement as early responders to waterlogging. Most Cluster II AvLBDs exhibited down-regulation under waterlogging, except for AvLBD41_4, AvLBD41_5, and AvLBD41_6, which were markedly induced. Cluster III AvLBDs were expressed at nearly undetectable levels, apart from AvLBD41_1 and AvLBD16, which displayed specific up-regulation at 24 and 72 h. To corroborate the RNA-seq findings, qRT-PCR analysis was conducted on six representative AvLBDs using the same root samples employed for transcriptome sequencing (Figure 4b). The qRT-PCR profiles closely paralleled the RNA-seq results, thereby confirming the robustness of the transcriptomic dataset. Collectively, these observations identified Cluster I AvLBDs as pivotal regulators of root acclimation to waterlogging stress in A. valvata.

3.5. PPI Prediction for Cluster I AvLBDs

Previously, 143 waterlogging-induced co-expressed DEGs annotated as TFs were identified, among which 16 were assigned to the LBD family. Subsequent re-analysis revealed that seven of these 16 fragments lacked a complete LBD and were therefore excluded as authentic AvLBDs. A Venn comparison confirmed that all eight Cluster I AvLBDs were encompassed within the final set of 136 co-expressed TFs (Figure 5a). To elucidate potential interaction partners, PPIs among the 136 TFs were predicted by aligning them with A. chinensis proteins. Within Cluster I, AvLBD41_2 and AvLBD41_3 were found to be orthologous to A. chinensis Acc21845; AvLBD41_10 and AvLBD41_11 corresponded to Acc04000; AvLBD41_7, AvLBD41_8, and AvLBD41_9 matched Acc02240; whereas AvLBD39 was paired with Acc23802 (Figure S1). The predicted network indicated that Acc21845 co-occurred with an auxin-responsive protein, a trihelix TF GT-2-like, and a NAC domain-containing protein (Figure S1a). Putative orthologues of Acc04000 and Acc02240 have been reported to co-express with trihelix TFs in other plant species (Figure S1b). GO enrichment analysis of the STRING network demonstrated significant over-representation of ‘cellular response to hypoxia’, and ‘response to anoxia’ (FDR < 0.05, Figure 5b). Within these categories, eight proteins were identified, including two trihelix, three ethylene-responsive RAP2-3-like, and one WRKY TF, all predicted to interact with the two LBD proteins (Figure 5c).

3.6. AvLBD41_7 Interacts with AvERF75 but Not AvERF73

Among the AvLBD41 paralogues corresponding to Acc04000 and Acc02240, AvLBD41_11 and AvLBD41_7 exhibited the strongest up-regulation after 12 h of waterlogging, respectively (Figure S2). Therefore, the interaction partners of these two LBD proteins were further examined. It had been previously demonstrated that AvLBD41_11 interacts with AvERF75 (Acc29730). Therefore, BiFC assays were employed to determine whether AvLBD41_7 also interacts with AvERF75. A strong YFP signal was detected in tobacco leaves co-expressing AvLBD41_7-nYFP and AvERF75-cYFP, resembling the positive control, whereas all other combinations lacked fluorescence, confirming that AvLBD41_7 forms a heterodimer with AvERF75 (Figure 6a). In addition to AvERF75, AvERF73, another Acc29730 homolog, has also been implicated in the positive regulation of kiwifruit waterlogging tolerance. To assess whether AvERF73 interacts with AvLBD41s, BiFC assays were conducted. Moreover, no YFP fluorescence was detected in leaves co-expressing either AvLBD41_11-nYFP/AvERF73-cYFP or AvLBD41_7-nYFP/AvERF73-cYFP (Figure 6b,c), indicating that AvERF73 does not interact with these two AvLBD41 proteins. Collectively, these findings indicate that only specific AvLBD41-AvERFVII pairings participate in the waterlogging stress response of A. valvata.

3.7. AvLBD41_7, but Not AvLBD41_11, Interacts with AvHRA1

PPI prediction suggested potential interactions between LBD and trihelix TFs (Figure 5c). A previous weighted gene co-expression network analysis (WGCNA) of the waterlogging response identified the top 60 hub genes, including the trihelix TF i2_LQ_K_c82159/f1p6/2230 [41]. As its sequence contains a conserved Myb_DNA_bind_4 domain (Pfam 13837; Table S5) and clusters with AT3G10040.1 (AtHRA1) in the phylogenetic tree (Figure S3), it was designated AvHRA1. BiFC assays revealed a strong YFP signal exclusively when AvLBD41_7-cYFP was co-expressed with AvHRA1-nYFP, whereas AvLBD41_11-cYFP produced no fluorescence, indicating that AvLBD41_7, but not AvLBD41_11, forms a heterodimer with AvHRA1 (Figure 7). These findings demonstrate that AvLBD41_7 functions as a specific heterodimeric partner of the trihelix factor AvHRA1.

3.8. Transcriptional Activation Detection and Subcellular Localization of AvLBD41_7

It had been previously verified that AvLBD41_11 lacks transactivation capacity. In this study, the transactivation potential of AvLBD41_7 was further examined using a yeast assay system. The coding sequence of AvLBD41_7 was fused to the GAL4 DNA-binding domain (BD-AvLBD41_7), with empty BD and BD-AvERF73 serving as negative and positive controls, respectively. All transformants grew normally on SD/-Trp medium. However, growth on SD/-Trp/-Ade/-His medium was observed only in AvERF73-expressing cells (Figure 8). BD-AvLBD41_7 failed to grow under these conditions, confirming that AvLBD41_7 lacks transactivation activity. Transient expression of AvLBD41_7-GFP in tobacco epidermal cells revealed exclusive nuclear localization, overlapping with the nuclear marker (Figure 9), consistent with the subcellular distribution previously reported for AvLBD41_11 [18].

3.9. Analysis of Cis-Elements in AvLBD41_11 and AvLBD41_7 Promoters

Promoter fragments of 1115 bp upstream of AvLBD41_11 and 1498 bp upstream of AvLBD41_7 (Table S6) were cloned by referring to the reference genome of kiwifruit and scanned for cis-acting elements using the PlantCARE web server. The analysis revealed that both AvLBD genes contained a diverse set of stress- and hormone-responsive cis-elements as well as multiple TF binding sites (Figure 10). AvLBD41_7 uniquely harbors the ARE motif essential for anaerobic induction, whereas both promoters contain drought-responsive MBS and ABA-responsive ABRE elements. The AvLBD41_11 promoter carries a TGA-element (auxin-responsive), while AvLBD41_7 possesses a TCA-element (salicylic acid-responsive). Furthermore, multiple TF-binding sites, including MYB, MYC, and the W-box (WRKY), were identified across both promoters.

3.10. Phylogenetic Analysis of the AvLBD41_7 Promoter and Its Homologs

To explore the phylogenetic relationships between the AvLBD41_7 promoter and its homologs, promoter sequences from 30 different kiwifruit accessions were aligned with the AvLBD41_7 promoter and subjected to an ML analysis with MEGA 11.0 (Figure 11). Homologs were grouped strictly by species, and the AvLBD41_7 promoter formed a well-supported clade with A. macrosperma and A. polygama, two taxa that consistently display superior waterlogging tolerance. This phylogenetic signature positions the AvLBD41_7 promoter as a potential marker for rapidly discriminating waterlogging-tolerant from waterlogging-sensitive kiwifruit germplasm.

3.11. Development and Preliminary Test of a PCR Marker Based on the AvLBD41_7 DNA Sequence

The polymorphic amplicons between A. valvata and A. deliciosa were predicted from the AvLBD41_7 DNA sequence and species-specific primers were designed. Agarose gel electrophoresis of PCR products showed that all 11 accessions generated products of 1000–2000 bp (Figure 12). 7 A. deliciosa samples consistently yielded Band 2, whereas 4 A. valvata samples generated only Band 1. Hence, this PCR marker can be used for the distinction of the two taxa.

3.12. Differential Expression of AvLBD41_11 and AvLBD41_7 Across Tissues During Waterlogging Stress

To assess organ-specific expression, AvLBD41_11 and AvLBD41_7 transcript levels were examined in leaf, petiole, stem, and the submerged stem base (stem end) under waterlogging (Figure 13). In leaves, both genes displayed comparable low-level expression that remained largely unaffected by waterlogging. In the petiole, AvLBD41_11 transcripts increased significantly at 24 h, whereas AvLBD41_7 was significantly induced at 12 h. In the stem, both AvLBD41_11 and AvLBD41_7 transcripts peaked sharply at 12 h before returning to baseline levels. In the stem end, both genes were significantly up-regulated at 24 h, but AvLBD41_7 accumulated to substantially higher levels than AvLBD41_11, suggesting a predominant role for AvLBD41_7 in waterlogging responses.

4. Discussion

A. valvata is a waterlogging-tolerant kiwifruit species valued for its vigorous root system, physiological resilience, and superior performance as a rootstock under waterlogging stress [12]. Although a chromosome-scale genome assembly with systematic annotation is not yet available, full-length transcriptome sequencing has enabled transcriptome-wide characterization of multiple gene families, including CIPK, ERF, and NAC [45,49,50]. By contrast, the LBD TF family, central to plant development and abiotic stress responses, has remained uncharacterized in A. valvata [27]. In this study, 26 AvLBD genes were identified in A. valvata (Figure 1), which were classified into two major classes and seven subclasses (Figure 2). Compared with Arabidopsis and rice, the AvLBD family comprises fewer genes and lacks Class Ie members, indicating interspecific divergence [21]. Another plausible explanation is that the LBD family is still incompletely annotated in A. valvata at the transcript level. Phylogenetic analysis facilitated functional annotation of several AvLBDs (Figure 2), while MEME motif analysis revealed conserved, class-specific protein architectures (Figure 3).

Transcriptome and qRT-PCR analyses revealed a pronounced and coordinated up-regulation of the AvLBD41 clade, orthologs of AtLBD41, in waterlogged kiwifruit roots (Figure 4), suggesting pivotal roles in waterlogging tolerance. In Arabidopsis, LBD41 functions as a core hypoxia-responsive gene together with PCO1/2, ADH1, PDC1, and ERF-VII factors, with RAP2.2/2.12 directly activating it via the HRPE cis-element [51]. Although the HRPE motif is absent from kiwifruit promoters, both AvLBD41_11 and AvLBD41_7 harbor multiple ERF-binding GCC-core and WRKY-binding W-boxes (Figure 10), indicating possible regulation by waterlogging-induced ERF and WRKY TFs [52]. Consistent with soybean [36], LBD41 is also induced under waterlogging stress in A. valvata roots.

Previous studies showed that multiple AvLBD41 unigenes were sharply up-regulated under waterlogging [18], pointing to their possible roles in A. valvata. However, the regulatory network of LBDs under waterlogging stress in kiwifruit remains limited. PPI analysis enables the prediction of functional protein interconnections and the pathways they collectively enrich [53]. In this study, LBD, ethylene-responsive RAP2-3-like, trihelix, and WRKY TFs were enriched in the ‘cellular response to hypoxia’ pathway, indicating a multi-TF regulatory network orchestrating the waterlogging response in kiwifruit (Figure 5). ERF-VIIs, such as AvERF73 and AvERF75, are key regulators of kiwifruit tolerance to waterlogging stress [17,18]. AvLOB41, renamed AvLBD41_11 in this study, was previously identified as an AvERF75 interactor [18], and here AvLBD41_7 was added as another AvERF75 partner (Figure 6). As AvERF75 enhances kiwifruit waterlogging tolerance, its LBD partners are likely co-regulators of the hypoxia response. By contrast, AvERF73, although implicated in waterlogging tolerance through the activation of AcNAC022 and AcHMGS1 [17], did not interact with either AvLBD41_11 or AvLBD41_7 in BiFC assays (Figure 6). AvERF73 and AvERF75 proteins share 84.53% amino acid identity. Sequence differences within their AP2 domain or adjacent motifs presumably specify their distinct interaction partners. Collectively, these findings suggest that only specific AvLBD41-AvERFVII heterodimers contribute to the regulation of waterlogging adaptation in kiwifruit. Trihelix TFs are plant-specific regulators known to play essential roles in abiotic stress responses [54,55]. Over-expression of the Salix matsudana GT-1 member TTF30 enhances submergence tolerance in Arabidopsis, and seven maize GT genes have been identified by RNA-seq as candidates for waterlogging tolerance [56,57]. In this study, AvLBD41_7 was shown to interact with both the trihelix factor AvHRA1 and AvERF75 (Figure 6 and Figure 7), revealing a three-TF regulatory module. Because Arabidopsis AtHRA1 attenuates hypoxia signaling by repressing RAP2.12 [58], it remains essential to determine whether AvHRA1 similarly regulates the kiwifruit homolog AvRAP2.3/AvERF75. How this AvLBD41_7-centered network contributes to waterlogging tolerance in kiwifruit needs to be further clarified. WRKY and ERF TFs are well-established determinants of waterlogging tolerance in Arabidopsis and sesame [59]. In kiwifruit, WRKYs have been identified as pivotal regulators of the waterlogging response [15,60]. Whether WRKYs physically associate with LBD proteins to co-regulate waterlogging responses in A. valvata remains to be investigated. Interestingly, AvLBD41_7, i1_LQ_K_c67155/f1p0/1459 (ADH1) and i1_LQ_K_c38965/f1p0/1342 (ADH2) all showed an abrupt peak at 12 h of waterlogging and declined thereafter [41], suggesting that AvLBD41_7 may act as a transcriptional rheostat that sustains rather than initiates ADH expression. Its lower induction amplitude may curb excessive ethanol fermentation, and this dosage-dependent hypothesis needs more direct functional testing.

Consistent with AvLBD41_11 [18], AvLBD41_7 displayed no transactivation activity in yeast assays. This observation aligns with previous findings in several Arabidopsis and rice LBDs [39,61], indicating that the LOB domain lacks intrinsic activation capacity and may require interacting partners. The exclusive nuclear localization of AvLBD41_7 (Figure 9) parallels that of its paralog AvLBD41_11 and conforms to the canonical expectation for LBD TFs [18,62,63]. Liquid–liquid phase separation is emerging as a key mechanism by which plants sense and respond to environmental stress [64]. In A. valvata, AvLBD41_7 formed protein condensates during subcellular-localization assays, implying that its response to waterlogging is possibly mediated by liquid–liquid driven compartmentalization.

PCR-based markers can help to identify and categorize genotypes resistant to abiotic stress [65]. Phylogenetic tree of AvLBD41_7 promoter mirrors species taxonomy (Figure 11). Based on its tight clade with waterlogging-tolerant A. macrosperma/A. polygama accessions, AvLBD41_7 promoter is a candidate marker for kiwifruit waterlogging resilience selection. Complementarily, an AvLBD41_7 PCR assay generates a single 1–2 kb amplicon, yet yields taxon-specific band patterns (Band 2 for A. deliciosa, Band 1 for A. valvata), allowing immediate discrimination between the two species (Figure 12). Together, AvLBD41_7 offers growers a molecular tool for predicting waterlogging-tolerant kiwifruit accessions or genotyping in practice.

The tissue-specific expression of a gene is closely associated with its functional characteristics. In this study, qRT-PCR revealed that both AvLBD41_11 and AvLBD41_7 were rapidly up-regulated at 12 h of waterlogging not only in roots but also in petiole, stem, and stem end, indicating that above-ground organs serve as additional response sites (Figure 13). Notably, AvLBD41_7 transcripts continued to accumulate in the stem end through 24 h of waterlogging stress (Figure 13). This region represents the primary site of adventitious root formation, which enables Actinidia to survive root-zone hypoxia [12]. Elevated LBD expression is correlated with increased adventitious root emergence in tolerant soybean [66]. Whether AvLBD41_7 mediates this anatomical remodeling thus requires functional validation. Similarly to Brassica napus, in which LBD41 is markedly more responsive in roots than in leaves after 24 h of submergence [67], both AvLBD41_11 and AvLBD41_7 were scarcely differentially regulated in leaves under waterlogging stress.

In this study, AvLBD41_7 was established as a master regulator of waterlogging tolerance in A. valvata. It physically links the hypoxia sensors AvHRA1 and AvERF75, revealing more about the waterlogging transcriptional network in kiwifruit. The AvLBD41_7 promoter and its PCR marker are helpful tools for molecular-assisted kiwifruit breeding. The main demerit is the lack of in vivo function validation. In the future, stable transgenic lines that over-express or silence AvLBD41_7 are required to prove if it directly affects waterlogging survival.

5. Conclusions

This study identified 26 AvLBD genes in A. valvata and classified them into seven phylogenetic subclasses. Through integrated analyses of conserved domains, stress-responsive expression profiles, transactivation capacity, subcellular localization, promoter cis-element composition, phylogenetic analysis and PCR marker development, a comprehensive functional overview was obtained. AvERF75 and AvHRA1 are the direct partners of AvLBD41_7, delineating a core transcriptional module underlying waterlogging stress adaptation. The AvLBD41_7 promoter and PCR-based markers can help to identify and categorize genotypes resistant to waterlogging stress. These findings advance the functional annotation of LBD genes in A. valvata and provide tools for the breeding of waterlogging-tolerant kiwifruit cultivars.