Comprehensive genome-wide mining and characterization of the NAC transcription factor family in Gynostemma pentaphyllum identifies GpNAC68 as a positive regulator of drought tolerance in transgenic Arabidopsis

Abstract

Background

NAC (NAM, ATAF1/2, CUC2) transcription factors (TFs) serve as pivotal regulators in plant growth, development, and abiotic stress response, exhibiting great application potential in the stress-tolerant breeding of crops. Gynostemma pentaphyllum (Thunb.) Makino is a medicinally important species in the Cucurbitaceae family, whose medicinal quality and yield are directly determined by its stress resistance capacity. However, the genome-wide systematic characterization and functional validation of the NAC TF family has not yet been reported in G. pentaphyllum, which severely limits the understanding of its stress resistance regulatory mechanism and the development of molecular breeding.

Results

In this study, 86 full-length GpNAC genes were identified from the G. pentaphyllum genome, which were phylogenetically divided into 21 conserved subgroups. The reliability of this classification was verified by subsequent gene structure analysis and conserved motif characterization. Collinearity analysis revealed that segmental duplication, rather than tandem duplication, was the core driving force for the expansion of the GpNAC family, and all duplicated gene pairs were subjected to strong purifying selection during evolution. Tissue-specific expression profiling showed that GpNAC genes were widely involved in the growth and development of G. pentaphyllum, with most genes exhibiting preferential expression patterns in roots, flowers, and fruits. Phytohormone responsiveness analysis demonstrated that the majority of GpNAC genes were responsive to abscisic acid (ABA), methyl jasmonate (MeJA), ethylene (ETH), and melatonin (MT), among which eight genes were simultaneously induced by all four tested hormones. Furthermore, heterologous overexpression of GpNAC68 significantly enhanced drought tolerance in transgenic Arabidopsis. This functional phenotype was achieved by suppressing reactive oxygen species (ROS) accumulation, preserving cell membrane integrity, and upregulating the expression of core drought-responsive genes.

Conclusions

In conclusion, this study provides the first comprehensive genome-wide characterization of the NAC TF family in G. pentaphyllum, screens out key candidate genes for the stress resistance breeding of this medicinal species, and offers novel insights into the evolutionary dynamics of the NAC family in Cucurbitaceae plants.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-09020-9.

Background

Drought stress is the most pervasive abiotic constraint limiting plant growth, agronomic yield, and specialized metabolite accumulation worldwide, with climate change exacerbating the frequency and severity of drought events globally [1, 2]. For plant species with dual medicinal and edible uses, drought stress exerts a dual agronomic and economic constraint: it not only restricts vegetative biomass accumulation, but also alters the biosynthesis of bioactive metabolites, two core traits that directly define the medicinal efficacy and commercial value of these crops [3]. G. pentaphyllum, a perennial medicinal herb in the Cucurbitaceae family, is widely cultivated across East and Southeast Asia. Its well-documented pharmacological activities are exclusively conferred by abundant gypenosides, a class of dammarane-type triterpenoid saponins that define its medicinal properties [4–6]. However, G. pentaphyllum cultivation is highly vulnerable to drought stress, which severely reduces yield and medicinal quality. Breeding drought-tolerant G. pentaphyllum germplasm is thus a critical priority for sustainable production, and this process is fundamentally dependent on the identification and functional validation of key stress-resistance regulatory genes.

The NAC TF family is one of the most ancient and functionally diverse TF families, originating in early land plants and undergoing extensive lineage-specific expansion in angiosperms [7–9]. NAC proteins are defined by a highly conserved N-terminal NAC DNA-binding domain, which mediates sequence-specific binding to NAC recognition cis-elements (containing CATGT and CACG) in target gene promoters, and a variable C-terminal transcriptional regulatory domain that confers functional specificity [9–11]. Genome-wide phylogenetic analyses in the model plants Arabidopsis and Oryza sativa have classified the NAC family into 21 conserved subgroups, and extensive functional studies have established NAC TFs as master regulators of diverse core plant processes, including organ development, cellular senescence, phytohormone signaling crosstalk, and abiotic/biotic stress responses [10, 12, 13]. Notably, NAC TFs are central integrators of the plant drought stress response network, functioning in both ABA-dependent and ABA-independent signaling pathways [14, 15]. For instance, Arabidopsis ANAC072/RD26 is a well-characterized transcriptional activator that positively regulates ABA-mediated dehydration stress responses [16]. However, the functional conservation of such well-characterized NAC regulators across non-model dicot lineages remains poorly understood.

Existing studies on the NAC gene family in Cucurbitaceae have predominantly focused on edible fruit and vegetable crops, with functional investigations largely centered on their regulatory roles in fruit development, nutritional and flavor quality formation, and stress tolerance. For instance, CmoNAC1 from pumpkin rootstocks has been shown to enhance salt tolerance in grafted cucumber plants [17]; CsNAC22 modulates soluble sugar accumulation in cucumber fruit [18]; and ClNAC100 acts as a positive regulator of plant height and fruit size in watermelon [19]. However, our understanding of the evolutionary history and functional diversification of the NAC gene family in non-model Cucurbitaceae lineages remains limited. The recent release of the high-quality G. pentaphyllum genome has provided a solid foundation for genome-wide gene family characterization and stress-related gene mining [6]. Nevertheless, the majority of stress-regulatory gene families in G. pentaphyllum have not been systematically analyzed, and the evolutionary dynamics, expression characteristics, and stress-related functions of the NAC family in this species remain entirely unknown. This critical knowledge gap not only hinders our understanding of NAC family evolution across Cucurbitaceae lineages, but also severely limits the development of drought-tolerant G. pentaphyllum varieties via molecular breeding.

Here, we present a systematic genome-wide characterization and integrative analysis of the NAC gene family in G. pentaphyllum, identifying 86 full-length GpNAC genes phylogenetically clustered into 21 conserved subgroups. We systematically analyzed their gene structure, conserved motif architecture, intraspecific and interspecific collinearity, and evolutionary selection pressure, demonstrating that segmental duplication is the dominant driver of GpNAC family expansion under purifying selection. Spatial expression profiling and phytohormone responsiveness assays further revealed functional diversification among GpNAC paralogs, providing a reference for stress-related functional screening. We further demonstrate that GpNAC68, the G. pentaphyllum ortholog of Arabidopsis ANAC072, positively regulates drought tolerance. Its heterologous overexpression in Arabidopsis enhanced drought resistance by attenuating ROS burst, preserving plasma membrane integrity, and activating core drought-responsive genes. Collectively, our study not only sheds new light on the evolutionary trajectory and functional diversification of the NAC family in Cucurbitaceae, but also identifies promising candidate genes for molecular breeding and genetic improvement of drought-resistant G. pentaphyllum.

Methods

Identification of NAC family proteins

Candidate NAC TF sequences were initially screened from the G. pentaphyllum genome using the Hidden Markov Model (HMM) profile of NAC domain (PF02365), as implemented in TBtools software [20], followed by iTAK verification [21]. Redundant sequences were removed at thresholds of ≥ 95% sequence identity and ≥ 90% alignment coverage, retaining only the longest transcript. For alternative splicing isoforms from the same gene locus, only the transcript with an intact NAC domain and full-length complete open reading frame was kept for downstream analysis. Final high-confidence GpNAC genes were obtained after triple domain validation using the SMART, Pfam, and NCBI-CDD databases. Physicochemical parameters of the identified GpNAC proteins, including molecular weight (MW) and isoelectric point (pI), were predicted using ProtParam, an online bioinformatics tool available on the ExPASy server. In parallel, the subcellular localization of each GpNAC protein was predicted using the ProtComp v9.0 on the Softberry web server.

Phylogenetic-based gene structure analysis of GpNAC genes

Phylogenetic reconstruction was performed on the aligned sequences using the Maximum Likelihood (ML) algorithm in FastTree v2.1.1 [22]. The final phylogenetic tree was then visualized using the Interactive Tree of Life online web server [23]. Additionally, a second ML phylogenetic tree incorporating the full-length amino acid sequences of all 86 identified GpNAC proteins was constructed using the same methods described above.

The exon–intron architecture of GpNAC genes was visualized by mapping the predicted coding sequences of each gene against their corresponding full-length genomic sequences, with all visualization operations performed using the TBtools toolkit. The MEME v5.5.9 tool [24] was employed to predict conserved motifs in GpNAC protein sequences, where the upper limit of predicted motifs was specified as 20, with all other parameters kept at default settings. An integrated visualization of the GpNAC phylogenetic tree, gene structures, and conserved protein motifs was finally generated using TBtools software.

Gene duplication and evolutionary analysis of GpNAC genes

Chromosomal mapping of the identified GpNAC genes was visualized using Circos software, with each gene locus precisely anchored to the corresponding chromosomes of the G. pentaphyllum reference genome. Interspecific and intraspecific collinearity analyses were performed using the MCscanX toolkit [25], and the resulting syntenic relationships and collinearity networks were visualized with TBtools. To clarify the selection pressure experienced by the duplicated GpNAC gene pairs, we adopted KaKs_Calculator v3.0 [26] to calculate the non-synonymous substitution rate (Ka), synonymous substitution rate (Ks), and the Ka/Ks ratio for each pair of duplicated genes.

Plant material preparation and treatments

The G. pentaphyllum germplasm used in this study belongs to the genotype for which a high-quality reference genome has been previously reported [6]. Plant materials were collected from the Xianhu Medicinal Botanical Garden of Guangxi University of Chinese Medicine, Nanning, Guangxi Zhuang Autonomous Region, China (geographic coordinates: 22° 49′ 12″ N, 108° 21′ 36″ E) on May 15, 2022. Seedlings were pre-cultivated in a climate-controlled growth chamber (constant 24 °C, 16-h light/8-h dark photoperiod) to ensure synchronized and uniform vegetative growth. For exogenous phytohormone treatment assays, six-week-old morphologically uniform seedlings were transferred to hydroponic culture with Hoagland’s nutrient solution, which was separately supplemented with 100 μM MeJA, 100 μM ABA, 100 μM ethephon (ETH donor), and 100 μM MT for independent treatment groups. Leaf samples were collected on 2 days after treatment to quantify the transcript levels of GpNAC genes in response to each phytohormone stimulus. Each independent biological replicate comprised pooled leaf tissues from three randomly selected individual seedlings.

The wild-type (WT) plant material used in the heterologous expression assays was Arabidopsis ecotype Columbia-0. Seeds of WT plants and GpNAC68-overexpressing transgenic Arabidopsis lines underwent sequential surface sterilization: first with 75% (v/v) ethanol, followed by sodium hypochlorite solution, before being sown on solid half-strength Murashige and Skoog (1/2 MS) medium. To ensure uniform and synchronous germination, sown seeds were subjected to dark cold stratification at 4 °C for 2 to 4 days. Germinated seedlings were transplanted into standardized potting mixture composed of peat soil, vermiculite, and perlite at a ratio of 3:1:1 (v/v/v). All plants were cultivated under controlled conditions: 24 °C, a 16 h/8 h light/dark photoperiod, 120 μmol·m^-2·s^-1 light intensity, and 70% relative humidity. Four-week-old soil-grown seedlings were subjected to drought stress by complete cessation of irrigation for 15 consecutive days. Following the drought period, plants were re-watered for 3 days to assess recovery.

Generation of GpNAC68-overexpressing transgenic Arabidopsis

We first constructed the GpNAC68 overexpression vector driven by the 35S constitutive promoter, and introduced the validated recombinant plasmid into Agrobacterium tumefaciens strain GV3101. The floral dip method was adopted to perform genetic transformation on WT Arabidopsis plants, to obtain the primary transgenic materials. For transgenic line screening, harvested T₁ generation seeds were sown on 1/2 MS solid medium containing kanamycin resistance pressure, positive resistant seedlings were transferred to nutrient soil, cultivated until the completion of the full growth cycle, and then T₂ seeds were collected from mature individual plants. After continuous multi-generation screening and genetic stability verification, we finally obtained stable homozygous T₃ transgenic lines. Two independent homozygous lines (L5 and L9) with high GpNAC68 expression levels were screened out for subsequent drought resistance functional verification experiments.

Transcriptomic profiling and heatmap analysis of GpNAC gene expression

Total RNA was digested with DNase I, and high-quality RNA (RIN ≥ 8.0) was used to construct cDNA libraries, which were sequenced (2 × 150 bp) on an Illumina HiSeq 2500 (Illumina, San Diego, CA, USA) after quality validation. For transcriptomic profiling, reads were quality-trimmed and pseudo-aligned to the G. pentaphyllum coding sequence annotation in Kallisto [27], with expression quantified as length-normalized transcripts per million (TPM). Heatmaps of GpNAC gene expression patterns were generated using TBtools software. Hormone-responsive differentially expressed genes were identified according to stringent filtering criteria: an absolute log₂ fold change (|log₂FC|) ≥ 1, and a false discovery rate-adjusted p-value < 0.05.

Subcellular localization

Subcellular localization of GpNAC68 was performed using a yellow fluorescent protein (YFP) fusion system. The GpNAC68 open reading frame was cloned into the pHB-YFP vector, and the recombinant construct was introduced into A. tumefaciens GV3101 for infiltration into Nicotiana benthamiana leaves. YFP signals were observed with a confocal microscope (Zeiss LSM 510 Meta) at 3 days post infiltration.

Dual luciferase assay

A transactivation activity assay was conducted as described previously [28]. The full-length coding sequence (CDS) of GpNAC68 was cloned into the pBD vector to generate the recombinant vector pBD-GpNAC68, with the empty pBD vector as the negative control and pBD-VP16 as the positive control.

Phenotypic and physiological assays

All physiological assays were performed using developmentally matched rosette leaves harvested from WT and GpNAC68-overexpressing transgenic Arabidopsis plants after 15 days of drought treatment. Each assay included three independent biological replicates (six leaves from six individual plants per replicate) and three technical replicates per biological replicate.

Anthocyanin content was determined via a modified acidified methanol extraction method. Briefly, fresh Arabidopsis leaves were extracted with 1 mL 0.1% (v/v) HCl-methanol at 4 °C for 30 min with ultrasonication, then centrifuged at 13,000 rpm for 10 min. The supernatant absorbance was read at 530 nm and 657 nm, and anthocyanin content was calculated as (A₅₃₀ − 0.25 × A₆₅₇) per gram fresh weight.

Relative electrolyte leakage (REL) and malondialdehyde (MDA) content were measured according to a previously reported protocol [29]. REL was calculated as the ratio of initial conductivity to total conductivity of leaf leachates, before and after boiling to rupture plant cells. MDA content was quantified via the thiobarbituric acid colorimetric method, with absorbance corrected for soluble sugar interference, and normalized to sample fresh weight (nmol/g FW).

Histochemical staining was performed to visualize ROS accumulation. In brief, 3,3'-diaminobenzidine (DAB) staining was used to detect H₂O₂, and nitroblue tetrazolium (NBT) staining was applied to monitor the distribution of superoxide anions (O₂·^-) in leaf tissues. H₂O₂ content was determined using the titanium sulfate colorimetric method. Fresh plant tissues were homogenized in pre-chilled acetone and centrifuged. The supernatant was mixed with titanium sulfate and ammonia solution to form an H₂O₂–titanium complex precipitate, which was then washed, dissolved in sulfuric acid, and measured for absorbance at 415 nm. The H₂O₂ concentration was calculated based on a standard curve and normalized to fresh weight.

O₂·^- content was quantified via the hydroxylamine oxidation method. Fresh plant samples were homogenized in pre-chilled sodium phosphate buffer and centrifuged at 12,000 × g for 15 min at 4 °C. The supernatant was reacted with hydroxylamine hydrochloride, p-aminobenzenesulfonic acid and α-naphthylamine for chromogenic reaction, and the absorbance was measured at 530 nm. O₂·^- content was calculated against a sodium nitrite standard curve and normalized to sample fresh weight.

RT-qPCR analysis

For RT-qPCR, 1 μg of DNase I-treated total RNA was reverse-transcribed to cDNA with the Maxima H Minus Kit (Thermo Scientific). Gene-specific primers were designed in Primer Express 5.0 (Table S1). Transcript levels of stress-responsive genes AtDREB2A (AT5G05410), AtRD29B (AT5G52300), AtNCED3 (AT3G14440), AtRD22 (AT5G25610), and AtCOR47 (AT1G20440) were analyzed, with AtActin2 (AT3G18780) as the internal reference. Relative expression was calculated via the 2^-ΔΔCt method on a CFX96 Touch system (Bio-Rad, Hercules, CA, USA).

Statistical analysis

To ensure the stability and repeatability of all experimental results, we set at least three independent biological replicates for each assay in this study. All quantitative data obtained from the experiments are expressed as the mean value ± standard deviation. For pairwise comparisons, significant differences were analyzed using Student’s t-test. For multiple comparisons of expression differences across different tissue samples, one-way analysis of variance followed by Duncan’s multiple range test was applied.

Results

Identification and phylogenetic analysis of the GpNAC genes in G. pentaphyllum

In this study, we adopted the published reference genome data of G. pentaphyllum to carry out systematic genome-wide identification and integrated bioinformatics analysis, aiming to clarify the basic sequence characteristics of the NAC family in G. pentaphyllum. We finally identified 86 NAC family genes with complete conserved domains from the G. pentaphyllum genome, and named them GpNAC1 to GpNAC86 according to their arrangement order on the 11 chromosomes of the species. The full annotation information of these GpNAC genes and their encoded proteins, including gene ID, amino acid length, MW, pI, predicted subcellular localization and conserved domain features, was collated and summarized in Table S2. Further analysis showed that the protein sequences of the identified GpNAC members varied greatly: the encoded amino acid length ranged from 65 to 651 amino acids, the corresponding MW was distributed between 7.58 kDa (GpNAC79) and 74.09 kDa (GpNAC81), and the theoretical pI spanned from 4.29 (GpNAC76) to 10.62 (GpNAC2). Subcellular localization prediction results indicated that 80 of the 86 GpNAC proteins were located in the nucleus, which is in line with the typical functional properties of TFs as gene expression regulators.

We constructed a phylogenetic tree to dissect the evolutionary relationships of the GpNAC proteins identified in this study. The analysis results revealed that all 86 GpNAC members could be classified into 21 subgroups, with high confidence support for each phylogenetic node (Fig. 1). This classification result was completely consistent with the classic NAC family classification system established in the model plants O. sativa and Arabidopsis, reflecting the high evolutionary conservation of the NAC family across different plant species. Further statistics on member distribution in each subgroup revealed that the OsNAC7 subfamily had the largest number of members, with 11 GpNAC proteins clustered in this clade, while the TIP subfamily had the fewest members, containing only one GpNAC protein. In addition, none of the GpNAC proteins were grouped into the unclassified1 subfamily, which implied that gene loss events may have occurred in the G. pentaphyllum genome throughout its evolutionary history.

Fig. 1.

Unrooted phylogenetic tree illustrating the evolutionary relationships of NAC proteins from G. pentaphyllum and Arabidopsis. All identified GpNAC proteins were clustered into 21 subgroups in the tree, which are visualized as distinct colored clades. Numbers on the tree branches indicate the bootstrap support values for each corresponding node

Phylogeny-guided analysis of gene structure and conserved motif composition of the GpNAC gene family

To further clarify the evolutionary conservation and structural features of GpNAC genes, we combined the phylogenetic clustering results of GpNAC members, and carried out a systematic analysis on the exon–intron structure and conserved motif composition of the 86 GpNAC genes identified in this study (Fig. 2A). Exon–intron structural analysis revealed that the number of exons per GpNAC gene ranged from 1 to 7, and the specific distribution was as follows: 4 genes with 1 exon, 8 genes with 2 exons, 48 genes with 3 exons, 7 genes with 4 exons, 4 genes with 5 exons, 13 genes with 6 exons, and 2 genes with 7 exons (Fig. 2B). It is worth noting that GpNAC genes belonging to the same phylogenetic subgroup had highly similar exon–intron arrangement patterns, while the gene structures of members from different subgroups were significantly different. For example, GpNAC genes in the NAM, NAP, ANAC3, ATF and SENU5 subgroups all contained 3 exons, while most members of the ANAC001 subgroup had 6 exons.

Fig. 2.

Integrated visualization of phylogenetic clustering, exon–intron structural features, and conserved motif distribution characteristics of the GpNAC gene family. A Phylogenetic tree constructed with the full-length sequences of the identified GpNAC proteins, where different subgroups of the NAC gene family are marked with unique colors for clear differentiation. B Distribution pattern of conserved motifs in the amino acid sequences of GpNAC proteins, Scale bar indicates amino acid sequence length. Each box with a unique color corresponds to a specific conserved motif, and the position of the box is fully consistent with the actual location of the motif in the corresponding GpNAC protein sequence. C Exon–intron structural composition of each GpNAC gene. The structural elements are annotated with color-coded boxes: green boxes indicate untranslated region (UTR), orange boxes represent CDS, and gray lines refer to intron sequences. Sequence length is indicated by the scale bar at the bottom

Conserved motif analysis via the MEME suite identified 20 conserved motifs across the 86 GpNAC proteins (Fig. 2C). In line with the findings of the gene structure analysis, GpNAC proteins clustered in the same phylogenetic clade had highly consistent motif types and linear arrangement orders. Specifically, the NAC2 subgroup was distinctly characterized by the presence of motifs 1, 3, 4, 6, and 9; the TERN subgroup specifically contained motifs 2, 3, 4, 5, 6 and 8. In addition, motifs 2, 3, 4, 6, 7 and 8 were unique to the ANAC003 subgroup, further distinguishing this group from other subgroups in the GpNAC family. The existence of these subgroup-specific motifs reflected the evolutionary conservation of these genes.

In general, most GpNAC genes assigned to the same phylogenetic subgroup had highly conserved gene structure and motif distribution characteristics, which strongly supported the reliability of the phylogenetic classification results in this study. Meanwhile, even for members belonging to the same phylogenetic clade, there were still slight differences in gene structure and motif composition, suggesting that minor variations in motif architecture or other additional regulatory elements may lead to functional differentiation among these NAC family members.

Collinearity and evolutionary analysis of GpNAC gene family

Gene duplication events are well established as the primary evolutionary force governing the expansion of TF gene families throughout plant genome evolution. To clarify how different gene duplication events affect the evolutionary expansion of the GpNAC gene family, we carried out genome-wide intraspecies collinearity analysis, and systematically identified the duplication types of all GpNAC members in the G. pentaphyllum genome (Fig. 3). Among the 86 GpNAC genes, we found 16 gene pairs generated by segmental duplication events, and seven gene pairs derived from tandem duplication events. These findings clearly indicated that the expansion of the GpNAC gene family was predominantly driven by segmental duplication events, rather than tandem duplication. From the perspective of chromosomal distribution, segmental duplication events were most concentrated on chromosome 1, followed by chromosome 4; while only 1 duplicated gene pair was detected on chromosome 5 and chromosome 10, respectively, suggesting that the distribution of duplication events on each chromosome was extremely uneven. For these screened duplicated gene pairs, we further performed the calculation of the ratio between Ka and Ks, with all duplicated GpNAC pairs showing values < 1 (Table S3). The results indicated that these duplicated GpNAC genes have experienced strong purifying selection pressure during the evolution of G. pentaphyllum.

Fig. 3.

Visualization of the interchromosomal duplication relationships of GpNAC genes in the G. pentaphyllum genome. Red lines represent the identified segmental duplication events between GpNAC genes, and the tandemly duplicated gene pairs are marked in red on the chromosomes

For an in-depth dissection of the evolutionary trajectory and phylogenetic relationships of the NAC TF family across different plant lineages, we performed genome-wide interspecies collinearity analysis between G. pentaphyllum and two representative reference species: Arabidopsis and Cucumis sativus (Fig. 4). At the whole-genome level, we identified far more syntenic blocks between G. pentaphyllum and C. sativus than between G. pentaphyllum and Arabidopsis (marked with gray lines). In contrast, the number of collinear orthologous NAC gene pairs was very close between the two comparison groups (marked with blue lines), which revealed that the NAC gene family has maintained a high degree of collinearity and evolutionary conservation among different dicotyledonous plants, and this conservation is not affected by the level of genome-wide synteny between species.

Fig. 4.

Interspecific genome-wide collinearity analysis of NAC family genes among G. pentaphyllum, Arabidopsis and C. sativus. Gray lines represent all the genome-wide collinear blocks identified in the analysis, while the syntenic orthologous NAC gene pairs are highlighted with red lines

Expression patterns imply versatile roles of GpNAC genes in development and hormone signaling

To clarify the functional differentiation and potential regulatory roles of GpNAC genes, we systematically analyzed their transcriptional expression characteristics in different tissues and their response patterns to exogenous phytohormone treatments (Table S4). We also constructed a comprehensive expression heatmap to intuitively display the transcript levels of all GpNAC genes in seven different tissues of G. pentaphyllum (including mature leaf, young leaf, stem, flower, fruit, tendril and root), as well as their expression changes in leaves treated with four stress-related phytohormones (ABA, MeJA, ETH and MT) (Fig. 5).

Fig. 5.

Expression profiles of the GpNAC gene family across diverse tissues and in response to exogenous hormone treatments. The color intensity gradient spans from red (high transcript abundance) to blue (low transcript abundance), which is used to visualize the relative differential transcript levels of each gene. Gray shading denotes transcript expression below the limit of detection

The results of tissue expression analysis showed that 82 of the 86 identified GpNAC genes had detectable expression in at least one tested tissue, and these genes were highly expressed in flowers, fruits, young leaves and roots. Significant differences in expression across tissues were analyzed by Duncan’s multiple range test, and full statistical results are listed in Table S5. These results indicated that the GpNAC gene family is widely involved in the regulation of vegetative and reproductive growth and development of G. pentaphyllum, and these four tissues are the main sites for GpNAC genes to perform their regulatory functions. The remaining four GpNAC genes (GpNAC57, GpNAC60, GpNAC76, and GpNAC79) exhibited no detectable expression in any of the tested tissues. These genes are likely putative pseudogenes that lost transcriptional activity during evolution, or may have highly specific spatiotemporal expression patterns not covered by our current sampling conditions and tissue libraries. Interestingly, GpNAC genes belonging to the same phylogenetic subgroup showed significantly different tissue expression patterns, which further supported the functional differentiation of closely related family members after the expansion of the NAC gene family.

In terms of response to exogenous phytohormones, 26 GpNAC genes showed no detectable transcript accumulation in leaf tissues after treatment with at least one of the four phytohormones. This result suggested that these genes may not participate in the regulatory network mediated by phytohormones in leaf tissues, or they may respond to other environmental signals or developmental factors not involved in this study. On the contrary, a group of GpNAC genes showed significant induced expression under multiple hormone treatments. Specifically, 8 GpNAC genes (GpNAC2, GpNAC3, GpNAC22, GpNAC31, GpNAC35, GpNAC45, GpNAC63, GpNAC64) were transcriptionally induced by all four tested phytohormones, indicating that these 8 genes may be the core regulators in the process of coordinating stress response in G. pentaphyllum. In addition, most GpNAC genes were significantly induced by at least one hormone treatment, which confirmed that the GpNAC family is widely involved in the phytohormone signal transduction process of this medicinal plant.

GpNAC68 is a drought-responsive nuclear-localized transcriptional activator

To identify candidate GpNAC genes involved in drought stress response, we noticed that GpNAC68 has high sequence similarity with the Arabidopsis ANAC072 gene, which has been previously confirmed to act as a transcriptional activator in the ABA-dependent drought stress response pathway. RT-qPCR analysis confirmed that GpNAC68 was significantly induced by drought stress, with transcript levels peaking at 12 h post-treatment (16.73-fold upregulation relative to the 0 h control) and remaining elevated at 24 h (10.62-fold upregulation) (Fig. 6A). Subcellular localization assay in N. benthamiana leaves showed that the GpNAC68-YFP fusion protein was exclusively localized to the nucleus, with empty 35S:YFP as the negative control (Fig. 6B). A GAL4-based dual-luciferase reporter assay further verified that GpNAC68 possessed strong transactivation activity, with significantly higher firefly luciferase (LUC)/renilla luciferase (REN) ratios observed in pBD-GpNAC68-expressing samples compared with the empty vector control, consistent with the activity of the positive control pBD-VP16 (Fig. 6C-D). These data confirm that GpNAC68 is a drought-responsive, nuclear-localized transcriptional activator.

Fig. 6.

Expression patterns, subcellular localization, and transcriptional activation of GpNAC68. A Expression of GpNAC68 following treatment with drought stress. B Confocal microscopy images of tobacco epidermal cells expressing 35S: GpNAC68-YFP or 35S:YFP (control). Scale bar = 20 μm. C Schematic representation of the GAL4-based dual-luciferase reporter assay used for transcriptional activation activity. D Dual-luciferase assay showing transcriptional activation activity of GpNAC68

Functional verification of GpNAC68 in regulating plant drought resistance via heterologous overexpression in Arabidopsis

To investigate the biological function of GpNAC68 in plant drought tolerance, we generated a 35S promoter-driven GpNAC68 heterologous overexpression construct, and obtained 12 stable transgenic Arabidopsis lines ectopically overexpressing this gene (Figure S1). Two independent homozygous T₃ transgenic lines (L5 and L9) with the highest GpNAC68 expression level were screened for subsequent drought tolerance functional assays (Figure S2). After 15 days of drought stress treatment, we found that GpNAC68-overexpression (OE) plants had significantly better growth status, and their anthocyanin accumulation was significantly lower than that of WT plants (Fig. 7A, B). After 3 days of rehydration, the majority of WT and GpNAC68-OE seedlings recovered growth, accompanied by a marked reduction in anthocyanin accumulation in WT seedlings. Since anthocyanin accumulation is a widely recognized physiological indicator of drought-induced oxidative stress in plants, this result suggested that overexpression of GpNAC68 can effectively reduce the oxidative damage caused by drought stress in plants.

Fig. 7.

Enhanced drought tolerance in GpNAC68-overexpressing Arabidopsis. A Phenotype of WT and two independent GpNAC68-overexpressing lines (L5 and L9) under control, drought stress, and rehydration conditions. Scale bar = 1 cm. B Anthocyanin contents in WT and 35S:GpNAC68 lines under control, drought, and rewatering conditions. C-D Histochemical staining with DAB (C) and NBT (D) was performed to visualize the accumulation of H₂O₂ and O₂·^-. Scale bar = 1 cm. E–H Quantitative determination of H₂O₂ content (E), O₂·^- content (F), REL (G) and MDA content (H) in WT and transgenic lines under control and drought conditions. Student’s t-test was used to assess significant differences relative to the WT (*p < 0.05, **p < 0.01, and ***p < 0.001)

To further explore the underlying mechanism of GpNAC68 improving plant drought resistance, we detected the accumulation of ROS by histochemical staining, because the abnormal burst of ROS under drought stress will cause serious oxidative damage to plant cells. Under normal conditions, weak DAB and NBT staining signals (reflecting H₂O₂ and O₂·^- accumulation, respectively) were detected in both WT and OE leaves; after drought stress, the staining intensity was significantly enhanced in WT, while much weaker signals were observed in OE lines (Fig. 7C, D). Consistently, the quantitative contents of H₂O₂ and O₂·^- in OE lines were significantly lower than those in WT under drought stress (Fig. 7E, F).

Consistent with the phenotypic results, physiological index detection showed that GpNAC68-OE lines had stronger ability to maintain cellular homeostasis under drought stress. Specifically, the REL level and MDA content in transgenic lines were significantly lower than those in WT plants, indicating that overexpression of GpNAC68 effectively alleviated plasma membrane damage and lipid peroxidation caused by drought (Fig. 7G, H). In summary, the above phenotypic and physiological results consistently demonstrated that overexpression of GpNAC68 enhances the drought tolerance of transgenic Arabidopsis. This functional effect is mainly achieved by inhibiting the excessive accumulation of ROS under drought stress, reducing cell membrane damage and lipid peroxidation, and thus maintaining the stability of intracellular environment.

To explore the molecular regulatory mechanism of GpNAC68 in enhancing plant drought resistance, we performed RT-qPCR assays to detect the expression levels of five key drought stress-responsive marker genes with validated biological functions, namely AtDREB2A, AtRD29B, AtNCED3, AtRD22 and AtCOR47. After 15 days of continuous drought stress treatment, the results revealed that the expression of all five marker genes was significantly upregulated in the two independent homozygous GpNAC68-OE lines, compared with the WT control plants (Figure S3). Taken together, these findings confirmed that the improved drought tolerance mediated by GpNAC68 in Arabidopsis is achieved, at least in part, by positively regulating the transcriptional expression of these classic drought stress response functional genes.

Discussion

Phylogenetic reconstruction clustered the 86 GpNAC proteins into 21 well-supported subgroups, consistent with the canonical classification system defined for Arabidopsis and O. sativa [10]. This high degree of phylogenetic conservation across monocot and dicot lineages indicates that the core functional diversification of the NAC family occurred prior to the divergence of these major angiosperm lineages, consistent with previous evolutionary analyses of the family [30, 31]. The total number of GpNAC genes (86) is comparable to that of other diploid Cucurbitaceae species, which typically encode 82–91 NAC proteins [32, 33], suggesting that the NAC family has not undergone extensive lineage-specific expansion in G. pentaphyllum. This aligns with the absence of recent whole-genome duplication events in the G. pentaphyllum genome, as whole-genome duplication is a major driver of TF family expansion in plants [6, 31].

Consistent with the phylogenetic classification results, gene structure and conserved motif analyses further independently validated the evolutionary reliability of our subgroup classification. GpNAC genes within the same phylogenetic subgroup exhibited nearly identical exon–intron architectures and motif composition and linear arrangement, a signature of evolutionary constraint acting on functionally related family members [10, 34]. The majority of GpNAC genes harbored 2–3 exons, with subgroup-specific exon number patterns that mirror those observed in other plant species [18, 33]. Furthermore, we identified 20 conserved motifs across the GpNAC family, with motifs 1–5 corresponding to the core subdomains of the conserved NAC DNA-binding domain. The strict conservation of these core motifs across all subgroups underscores their essential role in sequence-specific DNA binding [9, 12]. In contrast, subgroup-specific C-terminal motifs likely mediate subgroup-specific functional specialization, as the C-terminal domain of NAC TFs is responsible for transcriptional activation/repression and protein–protein interactions [9, 13].

Gene duplication events are the primary drivers of TF family expansion and functional diversification in plant genomes [35, 36]. Our intraspecific collinearity analysis revealed that the GpNAC family has expanded predominantly via segmental duplication events, with 16 segmental duplication gene pairs identified, compared to only seven tandem duplication pairs. This finding is consistent with the expansion pattern of NAC families in most Cucurbitaceae species [37]. The uneven chromosomal distribution of segmental duplication events, with the highest density on chromosomes 1 and 4, further reflects the dynamic nature of genome evolution in G. pentaphyllum. Furthermore, all duplicated GpNAC gene pairs exhibited Ka/Ks ratios < 1, indicating that these duplicated genes have undergone purifying selection pressure during evolution. Purifying selection acts to remove deleterious mutations, maintaining the functional conservation of duplicated genes over evolutionary time [38], which aligns with the conserved structural and expression features we observed within phylogenetic subgroups. Moreover, genome-wide syntenic blocks across G. pentaphyllum, Arabidopsis, and C. sativus highlight the strong evolutionary constraint acting on the NAC family, likely due to their essential roles in core plant developmental and stress response pathways [12, 13].

Tissue-specific expression profiling provides fundamental insights into the physiological functions of gene families, as expression pattern is tightly linked to biological function [39, 40]. Our expression analysis revealed that the majority of GpNAC genes exhibited preferential expression in reproductive organs (flowers and fruits), young leaves, and roots. The root-preferential expression of a subset of GpNAC genes is particularly noteworthy, as roots are the primary organ for water and nutrient uptake, and the first tissue to perceive drought and osmotic stress signals [41]. This expression pattern suggests that these root-enriched GpNAC genes may play pivotal roles in root development, stress signal perception, and belowground abiotic stress adaptation in G. pentaphyllum, consistent with the well-documented roles of NAC TFs in root system architecture regulation and drought stress response [13]. Meanwhile, the preferential expression of many GpNAC genes in flowers, fruits and young leaves, suggests their potential involvement in reproductive development, fruit ripening and secondary metabolite biosynthesis, processes that are tightly regulated by NAC TFs in other plant species [13, 42]. For the four GpNAC genes (GpNAC57, GpNAC60, GpNAC76, GpNAC79) with no detectable expression in tested tissues, we will perform full-length genomic sequence fine annotation to identify pseudogene-specific sequence features, and expand expression detection to cover more tissues and developmental stages throughout the G. pentaphyllum growth cycle for further validation.

Phytohormones including ABA, MeJA, ETH, and MT are core signaling molecules that integrate plant stress responses and developmental programming [43, 44], and NAC TFs are well-established key nodes in these hormone signaling networks [45, 46]. Our data revealed that the majority of GpNAC genes were responsive to at least one of the four tested phytohormones, with 8 GpNAC genes significantly induced by all four hormones. These multi-hormone-responsive genes likely act as critical integrators of multiple phytohormone signaling pathways in G. pentaphyllum, coordinating complex stress responses and developmental trade-offs. This finding aligns with previous functional studies demonstrating that NAC TFs act as key regulators to orchestrate the crosstalk among ABA, jasmonic acid, and ETH signaling pathways during abiotic and biotic stress responses [45, 46], highlighting the functional conservation of NAC TFs as hormone signaling cascades across plant lineages.

Drought stress is a major abiotic constraint that limits plant growth and productivity worldwide, triggering a cascade of physiological and molecular responses including excessive ROS accumulation, oxidative damage to cellular membranes and macromolecules, and transcriptional reprogramming of stress-responsive genes [47]. Here, we functionally characterized GpNAC68, which shares high sequence homology with Arabidopsis ANAC072, a previously characterized transcriptional activator in ABA-dependent dehydration stress responses [16]. Heterologous overexpression of GpNAC68 in Arabidopsis significantly enhanced drought tolerance, with OE lines exhibiting better growth and less anthocyanin accumulation relative to WT under drought. Physiological assays revealed that OE lines had significantly lower drought-induced ROS accumulation, reduced MDA content, and lower REL, indicating GpNAC68 maintains redox homeostasis and membrane integrity under stress. These findings align with the conserved role of NAC TFs in regulating ROS scavenging to boost abiotic stress resistance [48, 49].

At the transcriptional level, we found that overexpression of GpNAC68 significantly upregulated the transcript abundance of five key drought-responsive marker genes: AtDREB2A, AtRD29B, AtNCED3, AtRD22, and AtCOR47. These five genes constitute central nodes of the plant drought stress response network, with well-defined functional roles in stress signaling: AtNCED3 encodes the key rate-limiting enzyme in ABA biosynthesis [50], AtDREB2A acts as a positive regulator of the drought stress tolerance [51], while AtRD29B, AtRD22, and AtCOR47 are well-characterized downstream effector genes of both ABA-dependent and ABA-independent stress signaling cascades [52]. Strikingly, this regulatory mode is deeply conserved across plant NAC TFs. Compelling evidence from prior studies has demonstrated that ectopic expression of stress-responsive NAC genes, including ANAC019, ANAC055, ANAC072, ANAC096, PeNAC036, FtNAC31 and TaNAC6-3B, drives upregulation of these same drought-responsive genes [11, 52–54], further supporting the conserved functional role of GpNAC68 as a positive regulator of plant drought stress responses.

Combined with our expression analysis and functional validation results, we compared the functional conservation and interspecific divergence between GpNAC68 and its Arabidopsis homolog ANAC072. The two genes are highly conserved in that both encode nuclear-localized transcriptional activators involved in drought response regulation. Unlike ANAC072, a well-characterized core regulator in the ABA-dependent drought response pathway that is directly induced by ABA [16], GpNAC68 itself is not responsive to exogenous ABA, but may modulate the ABA-mediated drought response pathway by upregulating key ABA biosynthesis and signal transduction genes. The loss of direct ABA responsiveness of GpNAC68 may be an adaptive evolutionary trait to avoid excessive growth inhibition caused by sustained ABA activation, which is more suitable for the long-term growth strategy of perennial plants. Given that heterologous overexpression of GpNAC68 significantly enhanced plant drought tolerance without causing obvious growth defects under normal conditions, GpNAC68 has important application potential in the drought resistance molecular breeding of Cucurbitaceae cash crops, which are generally sensitive to drought stress throughout the whole growth period. The root-preferential expression characteristic of GpNAC68 also makes it an ideal candidate gene for root-specific expression modification, which can further improve crop drought resistance without affecting vegetative growth and yield through tissue-specific promoter driving.

Conclusions

In summary, this study provides a comprehensive genome-wide characterization of the NAC TF family in G. pentaphyllum, establishing a robust framework for understanding the evolutionary dynamics, structural conservation, and functional diversification of the NAC family in Cucurbitaceae. We demonstrate that the GpNAC family has expanded predominantly via segmental duplication under strong purifying selection, and exhibits deep evolutionary conservation of phylogenetic classification, gene structure, and motif organization across angiosperms. Spatial expression and phytohormone responsiveness profiling identified key candidate GpNAC genes involved in developmental regulation and stress signaling. Furthermore, we demonstrate that GpNAC68 functions as a positive regulator of drought stress tolerance by modulating ROS homeostasis and activating core drought-responsive gene expression. Collectively, these findings advance our understanding of NAC family evolution and function in non-model medicinal plants, and provide a valuable genetic resource for engineering drought-resilient crops.

Abbreviations

NAC

NAM, ATAF1/2, CUC2

TFs

Transcription factors

ABA

Abscisic acid

MeJA

Methyl jasmonate

ETH

Ethylene

MT

Melatonin

ROS

Reactive oxygen species

HMM

Hidden Markov Model

MW

Molecular weight

pI

Isoelectric point

ML

Maximum Likelihood

Ka

Non-synonymous substitution rate

Ks

Synonymous substitution rate

WT

Wild-type

1/2 MS

Half-strength Murashige and Skoog

TPM

Transcripts per million

YFP

Yellow fluorescent protein

CDS

Coding sequence

REL

Relative electrolyte leakage

MDA

Malondialdehyde

DAB

3,3'-Diaminobenzidine

NBT

Nitroblue tetrazolium

H₂O₂

Hydrogen peroxide

O₂·^-

Superoxide anion

UTR

Untranslated region

LUC

Firefly luciferase

REN

Renilla luciferase

OE

Overexpression

Authors’ contributions

HD and LX conceived and designed the research. WYP, ZH, LJM, and HML carried out the experiments and related investigations. WJ and XSQ contributed to sample collection, resource sorting and experimental preparation. MRH and LJM performed formal analysis and bioinformatic processing; WYP and ZH completed data curation and experimental validation. HD wrote the original draft of the manuscript, which was reviewed and edited by YSC and LX. HD and LX acquired the funding for this study; LX and YSC supervised the whole project. All authors have read and approved the final version of the manuscript for submission.

Funding

This project was supported by the National Natural Science Foundation of China (U24A202868), the Natural Science Foundation of Guangxi Zhuang Autonomous Region (2026GXNSFAA00641295), the Guangxi Peak Discipline: Traditional Chinese Medicine (Chinese Materia Medica) (Document No. 2 [2026] of the Department of Education of Guangxi for Scientific Research) and the Project supported by the Shanghai Committee of Science and Technology, China (23QA1410300).

Data availability

The raw transcriptome data have been submitted to NCBI SRA under the project number: PRJNA720501, PRJNA631355 and PRJNA1442663.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.