Ubiquitin ligase HcPUB30 targets HcWRKY1 to regulate monoterpenoids synthesis in Hedychium coronarium Authors: Ting Gao, Di Huang, Yunyi Yu, Hao Liu, Yuechong Yue, Shuting Zhou, Yanping Fan, Xinyue Li Published: 2026 ABSTRACT Hedychium coronarium, a perennial herb belonging to the genus Hedychium Koenig within the family Zingiberaceae, is renowned for its pleasant fragrance. The volatile compounds of flowers are primarily terpenoids, which are catalyzed by terpenoid synthase (TPS). Earlier studies have shown that HcWRKY1 transcription factor can bind to the promoter of HcTPS1, regulating the metabolism of terpenoids. To further investigate the upstream molecular mechanisms that regulate the release of volatile compounds in Hedychium, we focused on a crucial U-box type of E3 ubiquitin ligase involved in regulating transcription factors. This study utilized genomic data to identify HcPUB gene family. In combination with transcriptome data, seven candidate HcPUB genes were identified and cloned with subsequent functional analysis. Yeast two-hybrid assay demonstrated that HcPUB30 was the sole interactor of HcWRKY1 among the seven HcPUB candidates. In vivo and in vitro ubiquitination assays demonstrated that HcPUB30 ubiquitinates and promotes the degradation of HcWRKY1 via the 26S proteasome pathway. Multi-alignment analysis revealed that HcPUB30 possesses a conserved U-box domain and ARM motifs, which are implicated in plant growth and development. Subcellular localization indicated that HcPUB30 is localized in both the nucleus and cytoplasm. Quantitative real-time PCR analysis revealed that HcPUB30 exhibited the highest expression in petal tissues, and its expression peaked during floral senescence stage. Virus-induced gene silencing of HcPUB30 in Hedychium petals resulted in a significant decrease in monoterpenoid content, accompanied by a significant reduction in the relative expression levels of HcWRKY1 and HcTPS1. These findings indicate that HcPUB30 participates in the regulation of monoterpenoid biosynthesis by mediating HcWRKY1 in Hedychium petals.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09181-7. Introduction The discovery of the ubiquitination pathway has demonstrated its essential role in protein degradation and cell cycle progression. This process requires three classes of enzymes: ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), and ubiquitin ligase (E3). Typically, a single E1 interacts with a limited number of E2, each of which can associate with one or multiple E3. The E3 ligases form a large family responsible for the specific recognition of substrate proteins. While some substrates are directly bound by E3 ligases, most require prior modifications, such as phosphorylation or association with accessory proteins, to facilitate recognition [1]. Ubiquitin is a highly conserved protein composed of 76 amino acids. The ubiquitination pathway involves a complex enzymatic cascade in which a polyubiquitin chain is assembled by the sequential transfer of multiple ubiquitin molecules to a lysine residue (for example, lysine 48) on a covalently bound target. This polyubiquitin chain serves as a recognition marker for the 26S proteasome, leading to substrate degradation. Meanwhile, E3 ligases can be deubiquitinated and recycled for further rounds of reaction [1, 2]. E3 ligases are classified into four major families based on structural domains: HECT (Homologous to the E6-AP Carboxyl Terminus), RING (Really Interesting New Gene), U-box, and CRLs (Cullin–RING ligases) [3–5]. Among these, the U-box type represents one of the smallest E3 families. The U-box domain comprises over 70 amino acid residues and is a highly conserved functional module. Structurally similar to the RING-finger domain, the U-box domain lacks Zn²⁺ coordination, instead achieving stability through salt bridges and hydrogen bonds. The Plant U-box (PUB) E3 ligase gene family has been widely found in plants and plays pivotal roles in regulating responses to various abiotic and biotic stresses, including disease resistance [6–14], immune defense [15, 16], drought tolerance [17–24], salt tolerance [25, 26], and unfavorable temperature tolerance [27–29]. PUB ubiquitin ligases are also implicated in cell development, signal transduction, photoperiod responses, and the regulation of plant growth and development. In Arabidopsis thaliana, PUB genes regulate growth across various tissues. For instance, AtPUB25 and AtPUB26 negatively regulate petal dimensions by promoting the ubiquitination and degradation of RBE [30, 31]. AtPUB30 and AtPUB31 are associated with inflorescence and pedicel growth, as well as stomatal development [32]. PUB14 influences flowering time and floral reproductive development; the pub14 mutant exhibits delayed flowering under long-day conditions [33]. AtPUB18 and AtPUB19 negatively regulate seed germination under ABA and salt stress, with corresponding mutants showing enhanced stress tolerance during germination [34]. In contrast, AtPUB41 acts as a positive regulator of ABA-mediated seed germination [24]. AtPUB13-mediated protein degradation is important for pattern-triggered immunity, flowering time control, cell death, and hormone responses. The pub13 mutant flowers early under long days, and AtPUB13 interacts with HFR1 within the salicylic acid pathway to regulate photoperiod-dependent flowering [35]. Furthermore, AtPUB12 and AtPUB13 ubiquitinate the brassinosteroid receptor BRI1, promoting its endocytosis [36]. AtPUB10 responds to jasmonic acid signaling by ubiquitinating and degrading MYC2 in the nucleus [37]. The PUB protein AtUSR1 promotes leaf senescence; its restoration in mutants rescues delayed senescence, while overexpression accelerates aging [38]. In rice, expression of circadian rhythm–associated flowering genes (hd1, Hd3a, RFT1, Hd1, OsMADS51) is significantly reduced in the osgi mutant, accompanied by decreased OsPUB4 expression [39, 40]. Subsequent studies confirmed that OsPUB4 participates in OsGI-mediated circadian regulation and responds to light conditions, whereas OsPUB63 appears to have an opposing function [41]. OsPUB14 negatively regulates fertility by interacting with OsMTD2 [42], and OsPUB9 modulates leaf angle and grain size [43]. In grape, VvPUB26 ubiquitinates and degrades VvWRKY24, influencing proanthocyanidin content [10]. The strawberry FaU-box127 promotes fruit ripening and increases organic acid and soluble sugar content [44]. PUB17 and PUB16 participate in Brassica self-incompatibility [45]. In wheat, TaPUB4 regulates pollen grain metabolism by affecting sucrose-starch pathway genes in anthers [38]. Heterologous expression of TaPUB4 in the Arabidopsis atpub4 mutant partially restores fertility, pollen quantity, and viability, and enables normal tapetum degradation [38]. Although PUB E3 ligases have been extensively studied in plants growth, development, and stress responses, their functions in regulating floral scent metabolism in ornamental crops remain largely unexplored. Investigating PUB E3 ubiquitin ligases in Hedychium coronarium and their impact on floral scent metabolism will elucidate post-transcriptional regulatory mechanisms underlying late developmental stages and provide insights into secondary metabolism in fragrant plants. The WRKY gene family plays crucial roles in secondary metabolism and stress responses in horticultural plants. WRKY transcription factors regulate stress responses, growth, development, and the biosynthesis of secondary metabolites. For example, WRKYs are involved in stilbenoid biosynthesis in grape [46], promote vitamin C production in pear [47], positively regulate tropane alkaloid biosynthesis in Hyoscyamus albus [48], costunolide synthesis in Saussurea costus [49], ginsenoside biosynthesis in Panax ginseng [50], squalene synthesis in camellia [51], and artemisinin biosynthesis in Artemisia annua [52–54]. Conversely, a WRKY transcription factor negatively regulates sucrose, glucose, and fructose content in grape [55]. Furthermore, WRKYs influence cis-α-bergamotene emission in wild tobacco (Nicotiana attenuata) [56], participate in betalain biosynthesis regulation in pitaya (Hylocereus monacanthus) [57], and in luteolin biosynthesis in dandelion (Taraxacum antungense) [58]. Simultaneously, the WRKY gene family is widely involved in post-translational modifications, including phosphorylation and ubiquitination. MPK3/MPK6 phosphorylate WRKY33 in vivo, leading to metabolic reprogramming and increased phytoalexin production upon pathogen attack [59]. In grape, VvPUB26 ubiquitinates and degrades VvWRKY24, affecting proanthocyanidin content [10]. However, upstream regulatory mechanisms of WRKY transcription factors in ornamental crops remain poorly understood. Floral scent plays significant roles in plant development and defense. Hedychium coronarium (ginger lily) is an important aromatic plant resource in South China. Its unique floral fragrance and butterfly-like flower morphology make it not only a popular ornamental cut flower but also a plant widely used in landscape gardening. Due to the antioxidant properties of its essential oils, the industrial production of H. coronarium essential oil holds significant commercial potential for applications in aromatherapy and daily cosmetics. Terpenes, especially monoterpenes, constitute the most abundant volatile compounds in H. coronarium flowers. This characteristic makes it a suitable model system for studying the biosynthesis of floral terpenoid volatiles. Terpene synthases (TPSs) are key enzymes catalyzing terpenoid formation [60]. Preliminary research from our group identified HcWRKY1 as an important transcription factor regulating the emission of linalool, ocimene, methyl benzoate, and ethyl benzoate in petals. This study identified HcPUB E3 ubiquitin ligases associated with petal development in Hedychium by integrating genomic and transcriptomic data. Interactions between candidate HcPUBs and HcWRKY1 were validated through yeast two-hybrid assays, surface plasmon resonance (SPR) assay, in vitro ubiquitination reactions, and in planta ubiquitination experiments. Finally, virus-induced gene silencing (VIGS) of candidate HcPUB genes were employed to investigate their effects on floral volatile profiles and terpene synthase gene expression. This research explored the molecular mechanism of HcPUB genes in the post-transcriptional regulation of floral scent metabolism, providing insights into metabolic processes during late flower development and establishing a theoretical foundation for future studies on plant aroma metabolism. Materials and methods Plant materials Plant materials included Hedychium coronarium and Nicotiana tabacum cv. Wisconsin 38 (W38). Hedychium coronarium were cultivated and preserved at the Flower Research Center field at South China Agricultural University (Guangzhou, China). Professor Yanping Fang from South China Agricultural University undertook the formal identification of the plant material. W38 is a cultivar bred and promoted by the University of Wisconsin-Madison. It is currently preserved in the Herbarium of South China Agricultural University, with the seed deposition number Nt-W38. W38 seeds were provided by professor Yaoguang Liu from South China Agricultural University. Chromosomal localization and characterization of the HcPUB gene family Arabidopsis PUB protein sequences were acquired from TAIR (https://www.arabidopsis.org/). The Hedychium coronarium genomic data (unpublished) were sequenced and supplied by Novogene Bioinformatics Technology Corporation (Beijing, China). Using Arabidopsis AtPUB gene IDs as a reference, HcPUB proteins were identified using the “Quick Gene Family Identification” module in TBtools-II software [32]. The U-box HMM domain (PF04564) was downloaded from the Pfam database. The “Simple HMM Search” tool in TBtools was used to extract U-box proteins from the Hedychium genome. All protein sequences were verified for conserved domains via NCBI, yielding gene fragments with specific and complete U-box domains. Identified U-box genes were systematically ordered and named according to chromosomal positions. Visualization of GC content, N ratio, gene density, and chromosomal distribution was conducted using the “Advance Circus” module in TBtools [32]. Analysis of the HcPUB gene family Conserved domain data for all HcPUB genes were obtained from NCBI. PUB protein sequences were submitted to MEME (https://meme-suite.org/meme/tools/meme), yielding information on 20 conserved motifs. Combined with the genomic GFF3 file, visualization was performed using the “Gene Structure View” module in TBtools [32]. cis-Acting element analysis of the HcPUB family Promoter sequences (2000 bp upstream of the start codon) were extracted using TBtools. These sequences were submitted to PlantCARE (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) for cis-acting element prediction. Visualization was achieved with the “Simple BioSequence Viewer” module in TBtools. Transcriptomic data analysis Relative expression abundance data for the HcPUB gene family were extracted from Hedychium transcriptomic datasets, covering three tissues (petal, leaf, rhizome) and three petal developmental stages (D1: green bud stage; D4: full bloom stage; D6: senescence stage). Data normalization was performed using log2(FPKM + 1). A heatmap was generated with the “HeatMap” module in TBtools. Phylogenetic analysis of PUB proteins A cluster analysis involving 96 HcPUBs and 58 Arabidopsis PUB genes was conducted using TBtools. Multiple sequence alignment was performed with MUSCLE, followed by trimming with trimAl. A Maximum Likelihood (ML) phylogenetic tree was constructed using the “One Step Build a ML Tree” module in TBtools, with 1000 bootstrap replicates. The tree was visualized and classified using iTOL [61]. Yeast two-hybrid assay The pGADT7 and pGBKT7 plasmids were constructed with HcWRKY1 and HcPUBs, respectively. Plasmids of pGADT7-WRKY1 and seven distinct pGBKT7-PUB vectors were co-transformed into yeast strain Y2H Gold. The transformation mixture comprised: Y2H Gold competent cells, 2–5 µg of target plasmid, 10 µL of denatured carrier DNA (100 °C, 5 min), and 500 µL of PEG/LiAc. The mixture was incubated at 30 °C for 30 min in a water bath, followed by a heat shock at 42 °C for 15 min. After centrifugation (5000 rpm, 40 s), the pellet was resuspended in 400 µL ddH₂O, centrifuged again (5000 rpm, 30 s), and resuspended in 50 µL ddH₂O. The suspension was plated on SD/-Leu-Trp medium and incubated at 29 °C for 48–96 h. Colonies were streaked onto SD/-Leu-Trp-Ade-His plates. After 48–96 h, X-α-gal was used to assess blue coloration, indicating positive interactions. Amino acid sequence alignment PUB protein sequences from 10 plant species (Asparagus officinalis; Arabidopsis thaliana; Musa acuminata; Sorghum bicolor; Ananas comosus; Dendrobium catenatum; Phalaenopsis equestris; Cocos nucifera; Zea mays; Oryza sativa) were downloaded from NCBI. Multiple sequence alignment of all 11 proteins was conducted using MEGA X with ClustalW. The alignment file was analyzed for conserved domains using GeneDoc. Subcellular localization The pEAQ-GFP vector, which was preserved in our laboratory, was used to construct the pEAQ-GFP-HcPUB30 recombinant plasmid. The full-length coding sequence (CDS) (excluding the stop codon) of HcPUB30 was fused in-frame with pEAQ-GFP vector via homologous recombination. The primers are listed in Table S1. Then the pEAQ-GFP-HcPUB30 vector was transformed into Agrobacterium tumefaciens strain EHA105. Recombinant agrobacteria were cultured overnight at 28 °C in LB liquid medium containing 50 µg/mL kanamycin. A 50 µL aliquot was inoculated into fresh LB medium with kanamycin and grown overnight. Bacteria were harvested (5000 rpm, 5 min) and resuspended in infiltration buffer to OD₆₀₀ = 1.0, followed by incubation at room temperature for 3 h. Agrobacterium carrying pEAQ-GFP alone was the control. Uniformly grown tobacco leaves were selected for infiltration. Using a needleless syringe, the bacterial suspension was slowly infiltrated. Each treatment included three technical and three biological replicates. Infiltrated plants were maintained in a growth chamber. Fluorescence was detected and imaged 4 days post-infiltration. Total RNA extraction, cDNA synthesis, qRT-PCR analysis, and target gene cloning Total RNA of different tissues (petals, rhizomes and leaves) and six petal development stages (F1, F2, F3, F4, F5, F6) was extracted using the HiPure Plant RNA Mini Kit (Megen, Guangzhou, China). cDNA for qRT-PCR was synthesized using the Evo M-MLV RT Mix Kit with gDNA Eraser (TaKaRa, Japan). qRT-PCR was performed using SYBR Green Master Mix (Yeasen, Shanghai, China) on an ABI 7500 Fast Real-Time PCR System (Applied Biosystems, USA). HcGAPDH served as the internal reference gene. Relative expression levels were calculated using the 2–ΔΔCT method [62]. Plant DNA was extracted using the DNA Quick Plant System (Tiangen, Beijing, China). cDNA for gene cloning was synthesized with the PrimeScript RT Reagent Kit with gDNA Eraser (TaKaRa, Japan). Specific full-length primers were designed using Primer 5.0. Full-length sequences were amplified via PCR using cDNA from petals at the full bloom stage. Primer sequences are listed in Table S1. Prokaryotic expression and purification The GST, pMAL-MBP and pET-30a fusion expression vectors were preserved in our laboratory. The cloned MBP-HcPUB30 and GST-HcWRKY1 recombinants were transformed into Escherichia coli strain DH5α. Both the fragments of HcPUB30 and HcWRKY1 were cloned into the pET30a vector, respectively. Sequence-verified recombinant plasmids were transformed into E. coli ROSETTA DE3 cells. Optimal induction conditions were determined by testing IPTG concentration, temperature, and induction time. After disrupted by ultrasonication on ice using a probe-type sonicator, the lysate was centrifuged at 12 000 ×g for 30 min at 4 °C. The supernatant was collected and filtered through a 0.45 μm ultrafiltration tubes. Then soluble proteins were purified by Ni affinity chromatography using different columns (Thermo, USA) (His-Tag column for pET30a recombinant proteins, GST-Tag column for GST-HcWRKY1, MBP-Tag column for MBP-HcPUB30). SDS-PAGE analysis of the supernatant and pellet fractions was conducted to assess release efficiency and solubility of the recombinant proteins. Surface Plasmon Resonance (SPR) assay SPR binding experiments were carried out using a BIAcore X100 device at 25 °C. All SPR measurements for kinetic analysis were performed in PBS buffer (10 mM Na2HPO4; 2 mM KH2PO4, pH 7.4; 137 mM NaCl; 2.7 mM KCl). Purified HcPUB30 protein was immobilized on a CM5 sensor chip using the standard EDC/NHS coupling method to achieve approximately 1000 response units (RU). Serial dilutions of HcWRKY1 were prepared and flowed over the chip surface and RU changes were monitored over time. After each cycle, the sensor surface was regenerated with a brief treatment of 10 mM NaOH. The apparent binding affinity (KD) was calculated using BIAcore X100 Evaluation software and replotted using GraphPad Prism 8. Each set of equilibrium binding responses was fitted to a 1:1 binding model. In vitro ubiquitination assay and detection Ubiquitination reaction components [Reaction Buffer (20x), ubiquitin, substrate protein (GST-HcWRKY1), E1 (human, UBE1), E2 (human, UBE2D2), E3 (MBP-HcPUB30)] were combined in sterile microcentrifuge tubes, mixed, and incubated at 30 °C for 2 h [63]. A 30 µL aliquot was mixed with 5x protein loading buffer, denatured at 100 °C for 10 min, and subjected to SDS-PAGE (80 V for 1 h, then 120 V for 1 h). Proteins were transferred to a PVDF membrane using a Bio-Rad standard wet transfer apparatus. The membrane was blocked for 1 h, washed, and incubated with BeyoECL Star A and B solutions. Immunodetection used anti-Ub, anti-MBP, and anti-GST antibodies. Bands were visualized using a gel documentation system (Bio-Rad, USA). In planta ubiquitination in tobacco Agrobacterium strains harboring pEAQ-GFP-PUB30 and pEAQ-HIS-WRKY1 were cultured overnight at 28 °C in LB medium with kanamycin. A 50 µL aliquot was subcultured in fresh medium and grown overnight. Bacteria were harvested (5000 rpm, 5 min) and resuspended to OD₆₀₀ = 1.0 in infiltration buffer, followed by 3 h incubation. Uniform tobacco leaves were pre-treated for 2 h with MG132 (25 and 50 µM) [64, 65] or DMSO (control), with three replicates of each treatment. Treatment groups: (1) pEAQ-GFP + pEAQ-HIS-WRKY1, DMSO; (2) pEAQ-GFP-PUB30 + pEAQ-HIS-WRKY1, DMSO; (3) pEAQ-GFP-PUB30 + pEAQ-HIS-WRKY1, 25 µM MG132; (4) pEAQ-GFP-PUB30 + pEAQ-HIS-WRKY1, 50 µM MG132. All bacterial combinations were mixed equally and infiltrated. After four days, leaf samples were collected, ground in liquid nitrogen, and lysed in RIPA buffer with 1 mM PMSF. The homogenate was vortexed for 1 min, incubated on ice for 30 min, and centrifuged at 12,000 rpm for 5 min. The supernatant was used for SDS-PAGE and Western Blot. Virus-induced gene silencing in ginger lily petals A 200 bp fragment from HcPUB30 was used to construct the pCaBS-γ-HcPUB30 vector based on the Barley Stripe Mosaic Virus (BSMV) vector pCaBS-γ. The reaction was performed at 22–25 °C for 30 min, then 75 °C for 15 min. A mixture containing 200 ng of the target fragment and 20 ng of linearized pCaBS-γ plasmid was incubated at 66 °C for 2 min. The pCaBS-γ-HcPUB30 plasmid was transformed into Agrobacterium EHA105. VIGS experiments were conducted as the following method. Agrobacterium strains carrying pCaBS-α, pCaBS-β, pCaBS-γ, or pCaBS-γ-HcPUB30 were cultured overnight in YET medium (Kan 50 mg/mL, Rif 25 mg/mL) at 28 °C. The cultures were harvested by centrifugation at 5000 rpm for 10 min and resuspended in infiltration buffer (10 mM MgCl2, 0.1 mM acetosyringone, 10 mM MES, pH 5.6) to a final OD₆₀₀ of 0.4–0.6, then incubated for 2–4 h. The control group contained pCaBS-α, pCaBS-β, and pCaBS-γ (1:1:1, v:v:v). The experimental group contained pCaBS-α, pCaBS-β, and pCaBS-γ-HcPUB30 (1:1:1, v:v:v). Silwet-77 was added. Flowers at the uniform bud stage were selected for VIGS. After petal separation, the intact flowers were immersed in the bacterial suspension and vacuum-infiltrated at 0.8 MPa for 5 min. Following vacuum release, the flowers were rinsed with deionized water, transferred to liquid MS medium, and cultured under low-light conditions with a 12 h/12 h light/dark cycle at 16 °C for 5 days [66]. Volatile compounds from full-bloom flowers with consistent growth were subsequently collected for GC-MS analysis. Each treatment group included three biological replicates, and the entire experiment was independently repeated four times. Flower volatile collection and GC-MS analysis After weighing, each intact flower was placed in a 250 mL beaker. We used ethyl caprate (0.87 g/mL, ≥ 99%; Sigma-Aldrich) as the internal standard. After being diluted 1:1000, 8 µL solution was added to the beaker, resulting in a final concentration of 0.87 g/L. Then the beaker was immediately sealed with aluminum foil and parafilm. After 30 min of headspace accumulation, a SPME fiber–polydimethylsiloxane (PDMS, with 50/30 µm divinylbenzene/Carboxen, Supelco) fiber was injected into the beaker for 30 min to trap the volatiles, positioned 3–5 cm above the flower. The collected samples were then injected into a GC-MS (Agilent, USA) system for volatiles analysis. The GC–MS system, comprising an Agilent 7890A GC and an Agilent 5975C MSD (Agilent, USA), was equipped with an AgilentDB-5MS capillary column (30 m × 0.25 mm). Helium was used as the carrier gas at a constant flow rate of 1.0 mL/min. The GC injection port temperature was initially maintained at 40 °C for 3 min, followed by a ramp of 5 °C/min to 250 °C. The column effluent was ionized by electron ionization (EI) at 70 eV, with a transfer temperature of 280 °C and a source temperature of 170 °C. Mass spectrometry parameters were as follows: scan rate 1 s/scan, mass range m/z 35–500, and detector voltage 1 kV. Ion chromatograms were analyzed using the NIST08 mass spectral library. Relative quantification of volatiles was performed using the Agilent ChemStation Data Analysis software based on the peak area ratio relative to the internal standard. Data analysis Genomic and transcriptomic data analyses used TBtools [32]. Other analyses and graphing used Origin 2021, GraphPad Prism 8 and Excel 2021. Differences between the two groups were assessed using a two-tailed Student’s t-test (asterisks indicate significant differences: *, P < 0.05; **, P < 0.01) in Fig. 9B and C. For multiple comparisons (Fig. 8C and D), statistical significance was determined by one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test (P < 0.05). All primer sequences are listed in Table S1. Results Identification and domain analysis of the PUB gene family in Hedychium coronarium genome A total of 96 PUB family members were identified from the Hedychium coronarium genome, designated HcPUB1 to HcPUB96 based on their chromosomal locations (Table S2). Chromosomal localization revealed that 78 HcPUB genes are unevenly distributed across 16 chromosomes, while 18 are on unassembled scaffolds (Fig. 1; Table S2). Chromosomes 14 and 4 harbor the most HcPUBs (10 and 9 genes, respectively). Except for chromosome 13, which lacks HcPUB genes, the remaining chromosomes possess 3–7 members each. Gene density analysis shows clustering near chromosome termini. GC ratio primarily ranges from 0.3 to 0.4. Synteny analysis revealed connections between genes on different chromosomes (e.g., HcPUB50 and HcPUB51 on chromosome 11 with HcPUB42 and HcPUB43 on chromosome 9; HcPUB75 on chromosome 16 with HcPUB2 on chromosome 1) (Fig. 1). Fig. 1Genome features of Hedychium coronarium. Circles from inner to outer represent the following information. Inner lines: the grey lines indicate the gene link regions in Hedychium coronarium, colored lines represent gene link regions where HcPUBs are located on different chromosomes. Circle 1: N ratio. Circle 2: GC ratio. Circle 3: gene density. Circle 4: chromosome distribution and locations of HcPUBs Genome features of Hedychium coronarium. Circles from inner to outer represent the following information. Inner lines: the grey lines indicate the gene link regions in Hedychium coronarium, colored lines represent gene link regions where HcPUBs are located on different chromosomes. Circle 1: N ratio. Circle 2: GC ratio. Circle 3: gene density. Circle 4: chromosome distribution and locations of HcPUBs Structural analysis of HcPUB gene family To elucidate the structural characteristics of the HcPUB family, analyses of conserved domains, motifs, and gene structures were performed. Motif analysis identified 20 motifs; nearly all HcPUB proteins contain Motif 1, Motif 3, and Motif 4, indicating high conservation. Exceptions include HcPUB5 and HcPUB72 (only Motif 3 and Motif 19), HcPUB44 and HcPUB66 (only Motif 4), and HcPUB40 and HcPUB49 (only Motif 1), suggesting potential functional divergence. Conserved domains classified HcPUB proteins into Ubox, RING_Ubox, ARM (armadillo repeat), UPS (ubiquitin-proteasome system), and PKc_like superfamily. Among these, Ubox, RING_Ubox, and ARM domains represent functional binding regions. Gene structure analysis showed HcPUB27 has the longest intron. HcPUB5 and HcPUB44 have abundant non-coding regions, while most HcPUB genes are primarily exonic (Fig. 2). Fig. 2The motifs, conserved domains and gene structures analysis of HcPUB gene family The motifs, conserved domains and gene structures analysis of HcPUB gene family Analysis of cis-acting elements of HcPUB gene family Promoter analysis (2000 bp upstream) revealed six primary cis-element types. Light-responsive, ABA-responsive, MYB binding sites (drought inducibility), and meristem expression elements were widespread, indicating HcPUB roles in growth, development, and stress responses. Additionally, palisade mesophyll cell differentiation elements were found in HcPUB5, HcPUB22, HcPUB47, and HcPUB70, suggesting involvement in cellular differentiation. Protein-binding site elements were detected in HcPUB42, HcPUB6, HcPUB65, HcPUB48, HcPUB43, etc., implying protein-protein interaction roles (Fig. 3). Fig. 3Promoter responsive elements in the HcPUB gene family Promoter responsive elements in the HcPUB gene family Phylogenetic analysis of PUB gene family A phylogenetic tree of 58 AtPUB and 96 HcPUB proteins classified them into 10 groups. Most HcPUB members were in Group 10 (42 genes), followed by Group 3 (15 genes) and Group 9 (9 genes) (Fig. 4). HcPUB85, HcPUB79, HcPUB87 and HcPUB52 were closely positioned to AtPUB12 and AtPUB13 (involve in photoperiod-mediated flowering time and pattern-triggered immunity) [35], demonstrating development-related and immune function. HcPUB86 and HcPUB3 showed high similarity to AtPUB25 and AtPUB26 (regulate petal dimensions) [31], suggesting floral organ development roles. HcPUB70 and HcPUB89 clustered with AtPUB18 and AtPUB19 (influence seed germination), indicating growth roles. HcPUB30 exhibited 98% homology with AtPUB10 and AtPUB11; since AtPUB10 ubiquitinates MYC2 [37], HcPUB30 may similarly targets transcription factors as an E3 ubiquitin ligase. Fig. 4Phylogenetic analysis of PUB family from Hedychium coronarium and Arabidopsis thaliana Phylogenetic analysis of PUB family from Hedychium coronarium and Arabidopsis thaliana Transcriptome expression analysis of HcPUBs Using FPKM values, relative expression of 96 HcPUB genes was determined across three tissues (petals, leaves, rhizomes) and three petal stages (D1, D4, D6). After excluding three non-expressed genes (HcPUB10, HcPUB45, HcPUB74), 93 genes were analyzed for tissue-specific expression. After excluding nine non-detected genes, 87 were analyzed for developmental expression. Leaves had the most highly expressed HcPUBs (46 genes), followed by rhizomes (28), then petals (25) (Fig. 5A), demonstrating tissue-specific expression pattern of HcPUB genes and suggesting their specialized biological functions in different organs. During petal development, senescence stage (D6) had the most highly expressed HcPUBs (43 genes), indicating primary regulatory roles of HcPUBs in late development. Only 25 HcPUBs were highly expressed in early bud stage (D1) (Fig. 5B). These findings show spatiotemporal expression specificity of HcPUBs. Fig. 5Heatmap expression of HcPUB gene family at different tissues and different development stages of petals. A Heatmap depicting the expression profiles of HcPUB genes across three distinct tissues of Hedychium coronarium: petal, leaf, and rhizome. B Heatmap illustrating the expression patterns of HcPUB genes during three developmental stages of flowers. D1 represents bud stage, D4 represents full-bloom stage, and D6 represents senescence stage Heatmap expression of HcPUB gene family at different tissues and different development stages of petals. A Heatmap depicting the expression profiles of HcPUB genes across three distinct tissues of Hedychium coronarium: petal, leaf, and rhizome. B Heatmap illustrating the expression patterns of HcPUB genes during three developmental stages of flowers. D1 represents bud stage, D4 represents full-bloom stage, and D6 represents senescence stage Screening and cloning of target genes The PUB family has been widely reported to be associated with plant growth and stress tolerance as E3 ubiquitin ligases. During Hedychium coronarium flower development, the floral scent originates from aromatic essential oils that accumulate in specialized secretory structures called oil cells. The labellum and lateral staminodes (collectively referred to as petals in this study) are the primary sites for oil cell distribution. Therefore, HcPUBs that are predominantly expressed in petal tissues (compared to rhizomes and leaves) were selected for investigating floral scent emission. As a pivotal transcription factor, HcWRKY1 has been reported to positively regulate biosynthesis of volatile terpenes and stress responses in Hedychium coronarium [67]. Previous results showed that HcWRKY1 was expressed mainly in petals. Its expression was relatively low at bud stage, peaked at full-bloom stage and then decreased at sentence stage [67]. To investigate the post-translational regulation of HcWRKY1, we aimed to identify potential regulatory E3 ubiquitin ligases among the HcPUBs. Gene expression data were standardized using lgFPKM (Fragments Per Kilobase of exon per Million mapped fragments). The candidate HcPUBs were primarily expressed in petals. Furthermore, all candidates showed increasing trends from the full-bloom to senescence stages—opposite to the expression pattern of HcWRKY1—suggesting a potential regulatory role as E3 ubiquitin ligases in flower development and VOC emission. Therefore, seven putative HcPUBs were selected (Fig. S1). These candidate genes were successfully cloned and their full CDS sequences are shown in Table S3. Gel electrophoresis and DNA sequencing confirmed the full-length of target sequences [HcPUB7 (1245 bp), HcPUB30 (1348 bp), HcPUB50 (2301 bp), HcPUB77 (1071 bp), HcPUB87 (1992 bp), HcPUB78 (1956 bp), and HcPUB13 (1191 bp)], which are suitable for subsequent functional experiments (Fig. S2). HcPUB30 interacts with HcWRKY1 To deepen our understanding of post-transcriptional control mechanism of HcWRKY1, we then used a yeast two-hybrid (Y2H) system to test the protein interaction between HcWRKY1 and seven HcPUB candidates. Y2H screening showed no autoactivation for seven HcPUB-BD constructs. On quadruple-dropout (QDO) medium, yeast co-transformed with AD/HcPUB13-BD failed to grow; others grew normally. No colonies turned blue with X-α-gal, confirming no autoactivation (Fig. S3). On double-dropout (DDO) medium, all ten co-transformation combinations grew. However, on QDO medium, negative controls of AD/BD, HcPUB30-BD/AD and HcWRKY1-AD/BD showed no yeast growth on QDO medium (Fig. 6A). Furthermore, the Y2H results of four combinations (HcWRKY1-AD/HcPUB7-BD, HcWRKY1-AD/HcPUB50-BD, HcWRKY1-AD/HcPUB78-BD, and HcWRKY1-AD/HcPUB87-BD) showed no yeast growth. Although the undiluted cells combination of HcWRKY1-AD/HcPUB13-BD and HcWRKY1-AD/HcPUB77-BD exhibited minimal growth on QDO medium without blue color, no growth at higher dilutions. In contrast, the undiluted cells combination of HcWRKY1-AD/HcPUB30-BD grew normally on QDO medium, with decreasing density at higher dilutions. Colonies turned blue after X-α-gal added (Fig. 6A). The Y2H experiment revealed that only HcPUB30 interacted with HcWRKY1 in yeast cell. Consequently, we chose HcPUB30 for further analysis. Fig. 6Proteins interaction between HcWRKY1 and HcPUB30. A Y2H assay of pGADT7-HcWRKY1 and seven candidate pGBKT7-HcPUBs. Combinations of pGADT7-T/pGBKT7-lam, HcPUB30-BD/pGADT7 and HcWRKY1-AD/ pGBKT7-lam are shown as negative controls. B BIAcore sensorgram of HcWRKY1 protein binding to HcPUB30. The HcPUB30 protein was immobilized on the CM5 chip and tested for binding with gradient concentrations of HcWRKY1 using multi-cycle mode on a BIAcore X100. The sensorgram displays time (s) on the x-axis and response units (RU) on the y-axis Proteins interaction between HcWRKY1 and HcPUB30. A Y2H assay of pGADT7-HcWRKY1 and seven candidate pGBKT7-HcPUBs. Combinations of pGADT7-T/pGBKT7-lam, HcPUB30-BD/pGADT7 and HcWRKY1-AD/ pGBKT7-lam are shown as negative controls. B BIAcore sensorgram of HcWRKY1 protein binding to HcPUB30. The HcPUB30 protein was immobilized on the CM5 chip and tested for binding with gradient concentrations of HcWRKY1 using multi-cycle mode on a BIAcore X100. The sensorgram displays time (s) on the x-axis and response units (RU) on the y-axis To further validate the protein interaction between HcWRKY1 and HcPUB30, we employed SPR to quantify their binding affinity (KD). First, the purity of the recombinant proteins was verified by SDS-PAGE to ensure compatibility with the CM5 chip loading process (Fig. S4). The purified HcPUB30 protein was then immobilized on the CM5 chip, and a series of HcWRKY1 dilutions were flowed over the surface to measure binding kinetics. SPR analysis revealed that the HcPUB30-HcWRKY1 interaction exhibited a binding affinity of KD = 68.4 µM (Fig. 6B). This measurable affinity strongly suggests a physiological interaction. HcPUB30 mediates ubiquitination of HcWRKY1 An in vitro ubiquitination assay was conducted to determine whether HcPUB30 can ubiquitination HcWRKY1. Fusion vectors GST-HcWRKY1 and MBP-HcPUB30 were obtained, and proteins were purified under optimized conditions (0.1 mM IPTG, 16 °C, 28 h) (Fig. S5). Concentrations were 3.242 mg/mL (MBP-HcPUB30) and 0.455 mg/mL (GST-HcWRKY1), confirming their suitability for subsequent ubiquitination experiments. Ubiquitination system components were combined. Lane 1 had empty MBP (control); lanes 2, 3, 4 lacked E1, E2, and MBP-HcPUB30, respectively. Lane 5 had the complete system. Western blot with anti-ubiquitin antibody detected the bands of ubiquitin (9.6 kDa) in lanes 1–4, but no polyubiquitin chains. In contrast, lane 5 showed polyubiquitin smears (10–130 kDa), indicating MBP-HcPUB30 mediated polyubiquitination of GST-HcWRKY1. Similarly, anti-GST and anti-MBP showed polyubiquitinated chains only in lane 5 (Fig. 7A). These results demonstrate that MBP-HcPUB30 effectively ubiquitinates GST-HcWRKY1 in vitro. Fig. 7HcPUB30 ubiquitinates and degradates HcWRKY1 in vitro and in vivo. A In vitro ubiquitination assay of MBP-HcPUB30 and GST-HcWRKY1. Top: polyubiquitinated HcWRKY1 was detected with an anti-ubiquitin antibody. Middle: polyubiquitinated HcPUB30 was detected with an anti-MBP antibody. Bottom: polyubiquitinated HcWRKY1 was detected with an anti-GST antibody. Lanes 2, 3, and 4 (left to right) lack E1, E2, E3 enzymes, respectively. MBP is the empty vector control for MBP-HcPUB30. Ub: ubiquitin. B Phenotypes of tobacco leaves under different treatments. MG132 is a proteasome inhibitor used to block the 26S proteasome-mediated degradation. C In planta interaction assay of HcPUB30 and HcWRKY1 in tobacco leaves. Proteins were extracted from leaves shown in B. Western blotting were performed with anti-His, anti-GFP and anti-β-actin antibodies HcPUB30 ubiquitinates and degradates HcWRKY1 in vitro and in vivo. A In vitro ubiquitination assay of MBP-HcPUB30 and GST-HcWRKY1. Top: polyubiquitinated HcWRKY1 was detected with an anti-ubiquitin antibody. Middle: polyubiquitinated HcPUB30 was detected with an anti-MBP antibody. Bottom: polyubiquitinated HcWRKY1 was detected with an anti-GST antibody. Lanes 2, 3, and 4 (left to right) lack E1, E2, E3 enzymes, respectively. MBP is the empty vector control for MBP-HcPUB30. Ub: ubiquitin. B Phenotypes of tobacco leaves under different treatments. MG132 is a proteasome inhibitor used to block the 26S proteasome-mediated degradation. C In planta interaction assay of HcPUB30 and HcWRKY1 in tobacco leaves. Proteins were extracted from leaves shown in B. Western blotting were performed with anti-His, anti-GFP and anti-β-actin antibodies HcPUB30 ubiquitinates and degrades HcWRKY1 in planta To further investigate whether HcPUB30 mediates ubiquitination of HcWRKY1 in plant, a transient expression assay in tobacco was performed. Control was empty pEAQ-GFP co-infiltrated with pEAQ-His-HcWRKY1. For the proteasome inhibitor treatments, leaves infiltrated with pEAQ-GFP-HcPUB30 and pEAQ-His-HcWRKY1 were pretreated with MG132 (25 or 50 µM) for two hours (Fig. 7B). Protein extracts were prepared after two days post-infiltration. Western blot analysis using anti-β-actin confirmed equal loading, with clear bands detected at 42 kDa. Anti-GFP confirmed GFP expression in all infiltrated leaves. The band of GFP protein (27 kDa) was detected only in the control lane (HcWRKY1-His + GFP; left lane, Fig. 7C), whereas the HcPUB30-GFP fusion protein bands (77 kDa) were detected in the other three experimental combinations (Fig. 7C). Anti-His showed strong His-HcWRKY1 bands in controls, but undetectable in experiments without MG132, demonstrating that HcPUB30 mediated ubiquitination and subsequent degradation of His-HcWRKY1 by the 26S proteasome. MG132 pretreatment partially rescued His-HcWRKY1 accumulation in a dose-dependent manner. The 50 µM MG132 treatment resulted in stronger His signals compared to the 25 µM, while subtile weaker band compered to GFP control (Fig. 7C). These results confirm that HcPUB30 ubiquitinates HcWRKY1 in planta, leading to its proteasomal degradation, a process that can be effectively suppressed by proteasome inhibition in a concentration-dependent manner. Analysis of PUB conserved domains across different species Protein functional domains determine their biological roles in plants. Ten PUB proteins with high homology to HcPUB30 were selected for alignment: AtPUB10 (Arabidopsis thaliana), MaPUB10 (Musa acuminata), AcPUB10 (Ananas comosus), AoPUB10 (Asparagus officinalis), DcPUB10 (Dendrobium catenatum), PePUB10 (Phalaenopsis equestris), SbPUB11 (Sorghum bicolor), CnPUB11 (Cocos nucifera), ZmPUB (Zea mays), and OsPUB (Oryza sativa). Among of these, AtPUB10 (At1g71020) was reported to function as an E3 ubiquitin ligase, mediating the ubiquitination and degradation of MYC2, with its ARM repeats identified as the functional interaction domains [37]. Alignment revealed that PUB proteins from various species share conserved N-terminal RING finger (without Zn) and U-box domains, and five C-terminal ARM repeats (armadillo repeat motif) (Fig. 8A). These features demonstrate high conservation of the PUB gene family across species. Both the U-box domain and ARM motifs represent characteristic conserved regions within the PUB family, serving critical roles in protein binding. These findings indicate that HcPUB30 is a U-box-type E3 ubiquitin ligase, recognizing and binding specific substrates for degradation through the 26S proteasome pathway. Fig. 8Multiple sequences alignment, subcellular localization and expression patterns of HcPUB30. A Multiple sequences alignment of PUB proteins from 11 species. NCBI accession numbers are shown as below. AoPUB10: XM_020413777.1 from Asparagus officinalis; AtPUB10: At1g71020 from Arabidopsis thaliana; MaPUB10: XM_018818354.2 from Musa acuminata; SbPUB11: XM_002454289.2 from Sorghum bicolor; AcPUB10: XM_020255445.1 from Ananas comosus; DcPUB10: XM_020848800.2 from Dendrobium catenatum; PePUB10: XM_020741411.1 from Phalaenopsis equestris; CnPUB11: KAG1334903.1 from Cocos nucifera; ZmPUB: NP_001131869.2 from Zea mays; OsPUB: EAZ24511.1 from Oryza sativa. B Subcellular localization of HcPUB30 in tobacco epidemic cells. Bar = 50 μm. C-D Relative expression level of HcPUB30 during different tissues (C) and various petal development stages (D). F1: green bud, F2: bud break, F3: half-open, F4: full-bloom, F5: initial senescence, F6: advanced senescence. The statistical significance was determined by one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test (P < 0.05) Multiple sequences alignment, subcellular localization and expression patterns of HcPUB30. A Multiple sequences alignment of PUB proteins from 11 species. NCBI accession numbers are shown as below. AoPUB10: XM_020413777.1 from Asparagus officinalis; AtPUB10: At1g71020 from Arabidopsis thaliana; MaPUB10: XM_018818354.2 from Musa acuminata; SbPUB11: XM_002454289.2 from Sorghum bicolor; AcPUB10: XM_020255445.1 from Ananas comosus; DcPUB10: XM_020848800.2 from Dendrobium catenatum; PePUB10: XM_020741411.1 from Phalaenopsis equestris; CnPUB11: KAG1334903.1 from Cocos nucifera; ZmPUB: NP_001131869.2 from Zea mays; OsPUB: EAZ24511.1 from Oryza sativa. B Subcellular localization of HcPUB30 in tobacco epidemic cells. Bar = 50 μm. C-D Relative expression level of HcPUB30 during different tissues (C) and various petal development stages (D). F1: green bud, F2: bud break, F3: half-open, F4: full-bloom, F5: initial senescence, F6: advanced senescence. The statistical significance was determined by one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test (P < 0.05) Subcellular localization of HcPUB30 and its expression patterns pEAQ-GFP-HcPUB30 was transiently expressed in tobacco. Fluorescence was observed in nucleus and cytoplasm for both control (pEAQ-GFP) and pEAQ-GFP-HcPUB30, indicating a distribution of HcPUB30 throughout the nucleus and cytoplasm (Fig. 8B). qRT-PCR on nine tissues (stigma, stamen, petal, stalk, root, sepal, leaf, stem, rhizome) and six petal stages (F1: green bud, F2: bud break, F3: half-open, F4: full-bloom, F5: initial senescence, F6: advanced senescence) showed that HcPUB30 expression was highest in petals, followed by rhizomes and roots, and lowest in stamens (Fig. 8C). During development stages, the relative expression of HcPUB30 significantly increased at the F5 (initial senescence) stage, peaking at the F6 (advanced senescence) stage (Fig. 8D). Compared with other tissues, HcWRKY1 exhibited the highest expression in petal tissues (Fig. S6, a). Furthermore, the expression of HcWRKY1 peaked at the full-bloom stage and then sharply decreased during senescence, showing an opposite trend to HcPUB30 from the full-bloom to late senescence stages (Fig. S6, b). HcTPS1 showed a similar expression pattern to HcWRKY1, with its highest expression level in floral organs and a peak at the F4 (full-bloom) stage [68]. The expression patterns suggest that peak E3 ligase activity of HcPUB30 during petal senescence stage might regulate the expression and function of HcWRKY1 and HcTPS1, thereby altering floral VOCs emission. Transient silencing of HcPUB30 in petals increases monoterpenoid contents and terpene synthase gene expression Multiple sequence alignment revealed less than 1% sequence similarity between the VIGS-targeted fragment of HcPUB30 and the other 95 HcPUBs, confirming that this fragment was specifically selected to avoid off-target silencing (Fig. S7). Then VIGS was performed to investigate the functional role of HcPUB30. After vacuum infiltration, petals exhibited uniform infiltration between treatment (pCaBS-γ-HcPUB30) and control (pCaBS-γ) (Fig. 9A). At 3–5 days post-infiltration, flowers were selected for GC-MS detection. Results demonstrated that terpenes (monoterpenoids, sesquiterpenoids) and aromatic esters (methyl benzoate) are the two major classes of volatile compounds in Hedychium coronarium flowers. Silencing of HcPUB30 significantly increased the content of most monoterpenoids, except for β-pinene. Compared to control (pCaBS-γ), linalool content markedly increased 1.98-fold, eucalyptol rose significantly 1.16-fold, β-ocimene increased by 46.39%, and allo-ocimene showed a highly significant elevation of 70.06% (Fig. 9B). However, no significant effect was observed on sesquiterpenoids (α-farnesene) or aromatic esters (benzoic acid, methyl ester) (Fig. S8). Conclusively, HcPUB30 negatively regulates key monoterpenoids among the detected volatiles compounds, specifically reducing the contents of linalool, ocimene, and eucalyptol. Fig. 9Analysis of volatile content and relative gene expression levels in HcPUB30 transient silencing and control flowers. A The phenotypes of ginger lily flowers after transient silencing HcPUB30 and control. B The content of monoterpenoids between pCaBS-γ-HcPUB30 and pCaBS-γ. C Relative expression levels of HcPUB30, HcWRKY1 and HcTPS1 genes. Differences between the two groups were assessed using a two-tailed Student’s t-test (asterisks indicate significant differences: *, P < 0.05; **, P < 0.01) Analysis of volatile content and relative gene expression levels in HcPUB30 transient silencing and control flowers. A The phenotypes of ginger lily flowers after transient silencing HcPUB30 and control. B The content of monoterpenoids between pCaBS-γ-HcPUB30 and pCaBS-γ. C Relative expression levels of HcPUB30, HcWRKY1 and HcTPS1 genes. Differences between the two groups were assessed using a two-tailed Student’s t-test (asterisks indicate significant differences: *, P < 0.05; **, P < 0.01) Among the candidate HcPUBs, the relative expression levels of HcPUB13 and HcPUB30 increased most dramatically from D4 to D6 during flower development, reaching 27.97-fold and 39.74-fold, respectively (Fig. S1). To confirm that the observed increase monoterpenoids is specific to HcPUB30 VIGS plants, we conducted VIGS on HcPUB13 as a paralogous control. The results showed that silencing of HcPUB13 decreased monoterpenoid emissions in flowers. Compared to the control, the contents of eucalyptol, β-ocimene, and linalool in pCaBS-γ-HcPUB13 decreased to some extent, though not significantly, whereas allo-ocimene showed a significant decline (Fig. S9). These results indicate that different HcPUBs exert distinct functions in the late development of Hedychium coronarium floral organs. VIGS of HcPUB30 increases monoterpenoid emissions, while VIGS of HcPUB13 has an opposite effect. Conclusively, HcPUB30 function as a specific negative regulator of monoterpenoid contents. qRT-PCR was performed to confirm silencing efficiency. The results showed that HcPUB30 expression was significantly reduced in silenced petals, while expression level of HcWRKY1 was significantly elevated. Previous research showed that HcWRKY1 directly binds to the promoter of HcTPS1 to regulate terpenoid synthesis [67]. Hence, we detected HcTPS1 expression level from the VIGS experiment samples to test our hypothesis that HcPUB30 may indirectly regulates HcTPS1, thereby influencing floral scent emission. The results showed that the expression of HcTPS1 was significantly up-regulated in pCaBS-γ-HcPUB30 flowers compared to control, indicating that HcPUB30 negatively regulates HcTPS1 expression (Fig. 9C). Together, these results strongly suggest that HcPUB30 negatively regulates the expression of HcWRKY1, which further diminishes HcTPS1 transcript level, ultimately influencing the production of monoterpenoids in ginger lily petals. Dissusion Phylogenetic analysis of PUB proteins in Hedychium coronarium Plant PUB proteins represent a class of U-box type E3 ubiquitin ligases. Genomic analyses in horticultural crops have demonstrated that PUB gene family members play significant roles in plant growth, development, and environmental signal responses. Generally, the PUB gene family can be categorized into distinct groups based on their domain architecture, with each group possessing conserved protein domains. For example, 155 FaPUB family members were clustered into four groups according to their conserved domains: Group I primarily contained U-box, Arm, and PK_Tyr_Ser_Thr domains; Group II mainly included the U-box domain alongside others; Group III typically contained only the U-box domain; while Group IV predominantly featured U-box and Arm domains, involved in abiotic stress, hormone responses, and fruit ripening [44]. Similarly, in tea plant (Camellia sinensis), 89 CsPUBs were classified into 10 classes based on the presence of additional domains/motifs (such as UFD2, U-box, ARM, WD40, and PRR), with 34 CsPUB genes responding to stress induction and 8 CsPUBs regulating anthracnose pathogen and drought stress [19]. Furthermore, studies on BoPUBs, StPUBs, IbPUBs (sweet potato), and CsPUBs (Cucumis sativus) have all reported classification systems where members are grouped according to conserved gene structures. In Brassica oleracea, 99 PUB family genes are unevenly distributed across nine chromosomes and classified into seven groups, potentially involved in abiotic stress responses and pollen tissue-specific regulation [69]. In cabbage, 65 BoPUB genes were categorized into six subfamilies according to conserved gene structures, and promoter sequence analysis indicated their involvement not only in light signaling, hormone responses, stress adaptation, and growth but also in cuticular wax synthesis [28]. Potato has 74 StPUBs located on 10 chromosomes, including three associated with drought stress [22]. Sweet potato identified 81 IbPUB genes distributed across 15 chromosomes, participating in biological processes and drought stress responses, and they were categorized into eight distinct groups according to domain architecture [23]. Similarly, our study identified 96 HcPUB genes in Hedychium coronarium, unevenly distributed across 16 chromosomes (Fig. 1). The cis-acting element prediction suggested their participation in environmental signaling, stress responses, protein binding, mitosis, and cell differentiation (Fig. 3). Phylogenetic analysis with 58 Arabidopsis PUB genes indicated that the 96 HcPUBs have undergone significant gene family expansion through duplication events, followed by subfunctionalization, neofunctionalization, or pseudogenization. This suggests that HcPUBs may exhibit greater functional diversity compared to AtPUBs. The HcPUBs were classified into 10 groups. Group 1 members primarily contain RING-U-box and U-box domains; Group 4 possesses RING-U-box and Armadillo (Arm) conserved domains; whereas Group 9 members have distinct USP, PKc_like superfamily, and RING-U-box domains (Figs. 2 and 4). The conserved domains within each group may indicate functional similarity among members; however, inferences regarding functional conservation require experimental validation. Moreover, proteins function only when folded into a well-defined 3D structure. A notable example of ortholog functional divergence is AtCHIP, the sole TPR-repeat PUB, named for its sequence conservation with animal CHIP proteins. Unlike animal CHIP overexpression, which confers stress protection [70], AtCHIP overexpression in Arabidopsis increases sensitivity to temperature stress and the hormone ABA [71]. Furthermore, genes can acquire new functions through changes in regulatory regions, altering expression patterns. While phylogenetic analysis and conserved domains provide clues for potential functional conservation, further analysis of gene expression patterns, phenotypic characterization, and functional validation is required. In our study, the clustering of HcPUBs and AtPUBs within the same subclades, such as HcPUB30 with AtPUB10/11 and HcPUB13 with AtPUB24 (Fig. 4), suggests potential functional conservation, which warrants further investigation. Although phylogenetic closely related PUBs may indicate conserved domains and similar functions, gene function is often exerted through interactions within networks. A change in the expression of one gene can trigger coordinated changes in its interacting partners to maintain network integrity. Following polyploidization, duplicated gene pairs within an interaction network may diverge in expression in a correlated manner, potentially resulting in two parallel network versions expressed in distinct cell types, developmental stages, or environmental conditions. Furthermore, the adaptive conflict model suggests that functional specialization of gene copies can arise from expression pattern differentiation [72]. We infer that the HcPUB family has adaptively evolved following a tetraploid event in Hedychium coronarium, as evidenced by their distribution across various tissues and diverse expression patterns during development (Fig. 5). Closely related HcPUBs may exhibit similar or divergent expression patterns, depending on the adaptive mechanism. Duplicate gene pairs with multiple regulatory regions can undergo subfunctionalization, neofunctionalization, or pseudogenization. Conserved paralogs often retain similar expression patterns when functionally redundant, whereas divergent paralogs evolve distinct patterns. For example, AtPUB22 and AtPUB23 function as negative regulators of water stress responses. They are rapidly and coordinately induced by abiotic stresses and exhibit similar expression patterns. Transgenic plants overexpressing PUB22 or PUB23 are hypersensitive to drought, whereas pub22 pub23 double knockouts show increased drought tolerance [73]. Conversely, functional divergence can occur via neofunctionalization (one copy acquires a new function) or subfunctionalization (copies partition ancestral functions) [74]. In citrus, PUB21 encodes an E3 ubiquitin ligase that suppresses JA-dependent defense and enhances susceptibility to CLas. However, the dominant-negative allele PUB21DN, a syntenic paralog with a C39A substitution in the U-box domain, inhibits PUB21’s E3 activity, and its expression positively correlates with HLB resistance in Rutaceae species [75]. Functional diversity of HcPUBs is involved in floral scent regulation in Hedychium coronarium The PUB family is functionally diverse, with substantial membership across plant species, primarily regulating growth, development, environmental responses, and stress tolerance. Emerging evidence suggests PUB family participation in plant secondary metabolite synthesis. Grape VvPUB26 ubiquitinates VvWRKY24 to influence proanthocyanidin content [10]. Strawberry FaU-box127 promotes fruit ripening and significantly increases organic acid and soluble sugar content [44]. Wheat TaPUB4 primarily affects the sucrose-starch pathway in anthers [38]. Rice OsPUB41 regulates lignin biosynthesis [14]. However, PUB regulation of secondary metabolism in ornamental plants remains limited. BSMV-VIGS is an efficient gene silencing technique using single-stranded RNA virus with three genome components designated alpha, beta, and gamma, and it has been successfully used in culinary ginger (Zingiber officinale) [76], barley, wheat, Brachypodium distachyon [77] and Hedychium coronarium [66, 78–80] to validate gene functions. Our study demonstrates that HcPUB30 involve in biological processes during late petal development (Fig. 8D). VIGS of HcPUB30 in petals confirms its negative regulation of linalool, ocimene, and eucalyptol (Fig. 9). This regulatory mechanism may also participate in environmental signal responses and hormone induction, requiring further investigation in future studies. HcPUB30 negatively regulates monoterpenoid emission by mediating the polyubiquitination and degradation of HcWRKY1. In contrast, silencing HcPUB13 decreased monoterpenoid content, suggesting it may positively regulate terpene biosynthesis. This functional divergence between HcPUB paralogs likely explains their distinct roles in floral scent regulation. Notably, HcPUB30 interacted with HcWRKY1, whereas HcPUB13 did not (Fig. 6). We speculate that these phenotypic differences arise because they ubiquitinate different substrates. Although both HcPUB30 and HcPUB13 expression increased during flower senescence, HcPUB30 may function in plant development and cell death or hypersensitive responses by ubiquitinating specific substrates, while HcPUB13 may target substrates involved in monoterpenoid biosynthesis to positively regulate abiotic stress responses, as increased terpene levels can enhance stress tolerance. AtPUB24, the closest sequence homolog of HcPUB13, has been shown to negatively regulate plant hypoxia and submergence tolerance by ubiquitinating STOP1 [81]. However, the specific function of HcPUB13 requires further investigation. In this study, HcPUB13 served as a paralogous control to highlight the unique functional mechanism of HcPUB30. Furthermore, PUB proteins can mediate various ubiquitin modifications, including polyubiquitination leading to substrate degradation and/or monoubiquitination altering protein activity or localization [82]. Key questions remain as follow. Which substrates interact with HcPUB13? Does HcPUB13 function through diverse ubiquitin modifications? Could HcPUB13 act as a molecular chaperone or co-chaperone in addition to its E3 ligase activity? These questions warrant further study. HcPUB30-mediated ubiquitination and degradation of HcWRKY1 to regulate floral scent PUB family members bind downstream proteins, mediate ubiquitination, and facilitate 26S proteasomal degradation. Previous researches demonstrate that rice OsPUB75, an ARM/U-box E3 ubiquitin ligase, negatively responses to abiotic stresses (drought and salinity) by ubiquitinating OsSK41. Its transcript levels decrease in response to salt and drought treatment [25]. OsPUB41 was significantly induced by Magnaporthe oryzae, showing more than a 10-fold increase at 1 h post-inoculation. It promotes OsPAL1 degradation, negatively regulating rice blast resistance [14]. In tomato, both in vitro and in vivo experiments confirm that SlPUB13 ubiquitinates FLS2, modulating plant immunity [15]. AtPUB22 and AtPUB23 function as negative regulators of abiotic stress responses. Their expression is rapidly induced by cold, drought, and high salinity. Transgenic plants overexpressing PUB22 or PUB23 are hypersensitive to drought, whereas pub22 pub23 double knockouts exhibit increased drought tolerance. Mechanistically, PUB22 and PUB23 coordinately regulate drought signaling by ubiquitinating cytosolic RPN12a in Arabidopsis [73]. Similarly, ABA and drought treatments promote PUB11 accumulation. PUB11 negatively regulates ABA-mediated drought tolerance by ubiquitinating two receptor-like protein kinases, LEUCINE RICH REPEAT PROTEIN 1 (LRR1) and KINASE 7 (KIN7), and mediating their degradation both in vitro and in vivo [83]. In sweet orange (Citrus sinensis), PUB21 is significantly upregulated in susceptible varieties during CLas infection. PUB21-mediated ubiquitination targets MYC2 for proteasomal degradation, thereby suppressing JA-dependent defenses and enhancing susceptibility to CLas [75]. Collectively, these PUB proteins act as negative regulators of both abiotic stress and defense responses. Interestingly, our study also demonstrated that HcPUB30 negatively regulates monoterpenoid production (Fig. 9) by specifically ubiquitinating HcWRKY1 and facilitating its degradation via the 26S proteasome (Fig. 7). Given that terpenes are important for tolerance to various stresses, we speculate that this mechanism may be involved in developmental cell death and/or susceptibility to abiotic stresses, which warrants further investigation. In contrast, overexpression of VvPUB26 in grape enhances proanthocyanidin (PA) content and resistance to powdery mildew. VvPUB26 positively regulates PA accumulation and pathogen tolerance by ubiquitinating VvWRKY24, a negative regulator of PA biosynthesis [10]. As one of the closest sequence homologs of HcPUB30, AtPUB10 functions as an E3 ubiquitin ligase that ubiquitinates and degrades MYC2, with its ARM repeats identified as the functional interaction domains [37]. Another key homolog is AtPUB11. Both the U-box and ARM repeats of AtPUB11 interact with AtHPPD, with the ARM repeats exhibiting stronger binding affinity. This interaction is enhanced under oxidative stress, promoting the ubiquitination and 26S proteasome-mediated degradation of AtHPPD to maintain reactive oxygen species (ROS) homeostasis [84]. The U-box and ARM domains of PUB13 are both involved in binding BRI1, with the ARM repeats stabilizing the interaction [36]. HcPUB30 contains a RING-U-box domain and six ARM repeats. The ARM repeats are putative peptide-binding sites; however, further research is needed to validate the specific polyubiquitination interaction sites. Additionally, numerous studies indicate that PUB genes respond to hormone signaling pathways (ABA, BR, JA, SA) and participate in various biotic and abiotic stresses [24, 35, 38, 43]. Promoter analysis of the HcPUB family also revealed ABA-responsive and drought-responsive elements, warranting further investigation into PUB functions in hormone and stress responses. Hormones can synergistically or antagonistically regulate monoterpenoid biosynthesis through the MEP and MVA pathways. In the auxin pathway, silencing HcIAA2 significantly increased the content of ocimene, linalool, and methyl benzoate [79]. MeJA regulates terpenoid synthesis via MYB and WRKY transcription factor families. HcMYB132 and HcMYB1 are key regulators of terpenes. Their transcript levels peak at the full-bloom stage and decrease during senescence [66, 78]. HcMYB132 expression is substantially induced by IAA treatment and is specifically expressed in flowers. HcMYB132 positively regulates floral aroma production; its suppression significantly reduces eucalyptol emission, although ocimene, linalool, and methyl benzoate contents remain unchanged. Furthermore, HcTPS1 transcript level is significantly decreased in HcMYB132-silenced flowers [78]. Similarly, silencing HcMYB1 significantly reduces the contents of methyl benzoate and linalool [66]. Our previous study indicated that HcWRKY1, a MeJA-responsive transcription factor, positively regulates terpene biosynthesis [67]. We speculate that the divergent VIGS results for HcIAA2, HcMYBs, and HcWRKY1 suggest that their upstream regulators may employ distinct modification mechanisms to exert antagonistic or synergistic effects on gene expression, thereby modulating floral scent emission. However, this study specifically focuses on the post-transcriptional regulation of HcWRKY1. The potential crosstalk between HcPUBs and other transcription factors (e.g., HcMYBs and HcIAAs) in mediating phytohormone signaling during different developmental stages requires further investigation. Our study demonstrates that HcPUB30 shows sharply elevated expression in petals during late developmental stages compared to blooming stage, indicating its involvement in biological processes during petal late development (Fig. 8D). Moreover, HcWRKY1 and HcTPS1 exhibited the highest expression levels in floral tissues. Their expression patterns showed opposite trends to that of HcPUB30 from the F4 to F6 stages, with both genes peaking at the full-bloom stage (Fig. S6) [67, 68]. The expression profile of HcTPS1 in ginger lily is consistent with previous findings in transgenic Nicotiana tabacum [85]. The function of the HcTPS1 promoter was validated by GUS activity assays in transgenic tobacco, confirming that HcTPS1 is specifically expressed in flowers and developmentally regulated. Promoter activity increased dramatically during flower opening, peaking at full bloom, thereby paralleling floral development [85]. Collectively, HcPUB30 may negatively regulate the expression of HcWRKY1 and HcTPS1 in Hedychium coronarium flowers, as evidenced by the significantly increased expression levels of HcWRKY1 and HcTPS1 in HcPUB30-silenced petals (Fig. 9C). Furthermore, VIGS-mediated silencing of HcPUB30 in petals confirmed its negative regulatory role in the production of monoterpenoids (linalool, ocimene, and eucalyptol) (Fig. 9B). These findings are consistent with previous studies showing that HcTPS1 encodes a monoterpene synthase involved in floral scent formation [60]. Based on these results, we propose a floral scent regulation model during flower development. The regulatory mechanism of the HcPUB30-HcWRKY1-HcTPS1-monoterpenoid pathway is dynamic, reflecting the precise expression patterns of these genes across developmental stages. During the blooming stage, high expression of HcWRKY1 facilitates its binding to the HcTPS1 promoter, promoting monoterpene synthase activity and monoterpenoid accumulation. Conversely, during the senescence stage, the expression pattern shifts. HcPUB30, an E3 ubiquitin ligase that functions predominantly in late development, is highly expressed and targets HcWRKY1 for ubiquitination. The subsequent degradation of HcWRKY1 via the 26S proteasome inhibits HcTPS1 expression, leading to reduced monoterpenoid content (Fig. 10). Fig. 10Proposed mechanism of HcPUB30 target HcWRKY1 during petal development. Red arrows indicate increasing of expression or volatile content, while dark blue indicate decreasing of expression or volatile content Proposed mechanism of HcPUB30 target HcWRKY1 during petal development. Red arrows indicate increasing of expression or volatile content, while dark blue indicate decreasing of expression or volatile content Conclusion Based on our findings, we propose a molecular mechanism by which HcPUB30 targets HcWRKY1 during flower development (Fig. 10). At the full-bloom stage, HcPUB30 expression is low, allowing the highly expressed HcWRKY1 to bind the HcTPS1 promoter and positively regulate the emission of monoterpenoids (linalool, ocimene, and eucalyptol). During the late senescence stage, HcPUB30 expression increases, and the accumulated HcPUB30 protein ubiquitinates HcWRKY1. The ubiquitinated HcWRKY1 is subsequently recognized and degraded by the 26S proteasome. Consequently, HcTPS1 expression is diminished, leading to reduced monoterpenoid content. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3.