PmWRKY‐10 Enhances Pine Wilt Disease Resistance by Promoting Pinosylvin Monomethyl Ether Biosynthesis Through Transcriptional Activation of PmPMT2‐1

ABSTRACT

Pine wilt disease (PWD), caused by the pine wood nematode (PWN, Bursaphelenchus xylophilus), poses a severe threat to global pine forests. Plants can enhance their resistance to external stresses by eliciting secondary metabolites; however, the transcriptional regulatory mechanisms governing this defence response in pines remain largely unexplored. Here, we report a novel defence regulatory module in Pinus massoniana . PmWRKY‐10 may participate in PWN resistance by regulating the biosynthesis of the nematicidal metabolite pinosylvin monomethyl ether (PME). We first demonstrated that PME accumulation is closely linked to resistance in P. massoniana and provided evidence that PME exhibits direct and potent nematicidal activity against PWN in vitro. By integrating multi‐omics profiling with weighted gene co‐expression network analysis (WGCNA), we identified PmPMT2‐1 as a key candidate associated with PME biosynthesis and accumulation, while recognizing that it is unlikely to function as the sole rate‐limiting determinant. Functional analyses demonstrated that PmPMT2‐1 overexpression promoted PME accumulation and enhanced resistance, whereas silencing the gene impaired both processes. Mechanistically, we identified PmWRKY‐10 as an upstream transcriptional activator that directly binds to the W‐box element in the PmPMT2‐1 promoter, inducing its transcription. Our work delineates a complete PmWRKY‐10→PmPMT2‐1→PME transcriptional‐metabolic pathway that is essential for stilbene‐based resistance in pine. This discovery not only provides fundamental insights into the inducible defence mechanisms of conifers but also identifies key genetic targets for breeding resistant varieties and developing sustainable forest management strategies.

We identified a PmWRKY‐10→PmPMT2‐1→PME defence pathway in Pinus massoniana against pine wilt disease. PmWRKY‐10 activates PmPMT2‐1 to promote the accumulation of the nematicidal metabolite PME. This discovery reveals the mechanism of stilbene‐mediated resistance and provides key genetic targets for forest management.

1. Introduction

Pine wilt disease (PWD) is caused by the pine wood nematode (PWN) Bursaphelenchus xylophilus. It poses a devastating global threat to conifer ecosystems, resulting in widespread tree mortality and significant economic and ecological losses (Futai 2008; Li et al. 2020). Since its introduction to China in 1982 (Li et al. 2020), the pathogen has spread rapidly across the country; as of 2024, official reports confirm its presence in 18 provinces (https://www.forestry.gov.cn/c/www/gkgjlyjgb/547477.jhtml), highlighting a severe and escalating challenge to forest health and management. This relentless expansion underscores an urgent need to shift management strategies from temporary, high‐cost containment toward a sustainable control paradigm (Chen and Li 2024). In this context, harnessing the plant's innate defence system, particularly the inducible biosynthesis of antimicrobial secondary metabolites, has become a promising frontier for developing durable resistance (Mithöfer and Boland 2012).

In conifers, stilbenoids such as pinosylvin and its methylated derivative pinosylvin monomethyl ether (PME) function as crucial phytoalexins against biotic stressors (Hwang et al. 2021; Koo et al. 2022). Notably, resistant pine species accumulate significant levels of PME upon nematode challenge, whereas susceptible ones show an attenuated response, implicating PME in disease resistance (Hwang et al. 2021). The biosynthesis of PME from pinosylvin is catalysed by pinosylvin O‐methyltransferase (PMT). In Pinus sylvestris , PMT2 has been identified as the key enzyme responsible for this specific methylation step (Chiron et al. 2000; Paasela et al. 2017). Transcriptomic studies further suggest the coordinated induction of PMT and stilbene pathway genes during defence, hinting at a regulated transcriptional programme (Paasela et al. 2017).

Transcription factors of the WRKY family are master regulators of plant defence, often governing the expression of pathway genes for specialized metabolites (Javed and Gao 2023). For instance, in Arabidopsis, AtWRKY33 directly activates genes for camalexin synthesis (Birkenbihl et al. 2012), and in grapevine, several WRKYs co‐regulate stilbene synthase (STS) genes (Vannozzi et al. 2018). In pines, the genes of the stilbene pathway, and WRKY, MYB and bHLH transcription factors are upregulated during nematode infection, suggesting their potential role in mobilizing defences (Hwang et al. 2021; Modesto et al. 2022). Our previous study also found that WRKY transcription factors exhibit coordinated changes with disease‐resistance‐related metabolites.

Collectively, these findings suggest that PME may serve as a key effector metabolite in the defence response of Pinus massoniana , and that WRKY transcription factors may play a central role in this process. bHLH and MYB transcription factors may also be involved in regulating its biosynthesis. Our previous work demonstrated that PME possesses direct nematicidal activity against PWN. Among the candidate transcription factors associated with PME biosynthesis, the WRKY family showed the strongest association with PME accumulation, whereas MYB and bHLH factors may function in parallel or downstream regulatory networks. Nevertheless, the molecular mechanism underlying PME biosynthesis and its transcriptional regulation in P. massoniana remains largely unclear. Therefore, elucidating the PME regulatory pathway is essential for understanding inducible chemical defence in conifers and for identifying potential targets for molecular breeding against PWD.

To elucidate this defence mechanism in depth, this study aims to integrate multi‐omics analyses with functional genomics approaches to systematically compare the responses of resistant and susceptible pine genotypes to PWN infection. The study seeks to identify key metabolites, biosynthetic enzymes and upstream transcriptional regulators involved in the accumulation of resin‐related defensive compounds and nematode resistance, thereby clarifying the molecular basis and regulatory networks that underlie inducible chemical defences in conifers and providing important candidate targets for resistance‐oriented molecular breeding.

2. Results

2.1. The Phenotypic and Histological Responses of Resistant and Susceptible P. massoniana Clones to B. xylophilus Infection

We compared symptom development and early cellular changes in a resistant clone (R: 1144‐2, three individual trees) and a susceptible clone (S: He9‐5, three individual trees) following inoculation with B. xylophilus. External phenotypes of 1‐year‐old stem segments revealed clear divergence between the two clones over time (Figure 1A–D). The resistant clone showed only mild symptoms from 10 to 30 days post‐inoculation (dpi) (Figure 1A,B). In contrast, the susceptible clone displayed severe discolouration and tissue necrosis, which progressed markedly from 10 to 30 dpi (Figure 1C,D).

FIGURE 1.

Tissue and cellular phenotypes of Pinus massoniana clones following inoculation with Bursaphelenchus xylophilus (pine wood nematode, PWN). Phenotypic symptoms of 1‐year‐old stem segments of Pinus massoniana resistant (R) and susceptible (S) clones at 10 and 30 days post‐inoculation (dpi). (A) R at 10 dpi; (B) R at 30 dpi; (C) S at 10 dpi; (D) S at 30 dpi. Scale bar = 1 cm. (E–G) Histological staining of paraffin sections from control and PWN‐inoculated stem segments. (E) Control stem segment; (F) R clone at 5 dpi; (G) S clone at 5 dpi. C, cortex; Ca, cambium; P, periderm; Ph, phloem; RD, resin duct; X, xylem. Scale bar = 100 μm.

Because visible symptom development showed a clear time lag. We performed paraffin sectioning and staining of resistant and susceptible stem segments 5 dpi to examine the early histological responses. The two clones exhibited distinct tissue staining patterns at this early stage (Figure 1E,F), indicating that cellular changes occurred before obvious external symptoms became apparent. These staining differences may reflect lignification‐associated responses and are consistent with previous anatomical studies of resistant and susceptible P. massoniana individuals infected with PWN (Li et al. 2023).

2.2. Elucidation of the PME Biosynthetic Pathway and Its Nematicidal Activity Against B. xylophilus

Subsequently, we performed integrated transcriptomic and metabolomic analyses to identify genes differentially expressed between the R and S clones at each time point after inoculation. This analysis identified 7077 differentially expressed genes (DEGs) and 577 differentially accumulated metabolites (DAMs). KEGG enrichment analysis revealed that the DEGs and DAMs were significantly co‐enriched in several biosynthetic pathways, including stilbenoid, diarylheptanoid and gingerol biosynthesis, as well as flavonoid biosynthesis (Figure 2A).

FIGURE 2.

Screening and validation of pinosylvin monomethyl ether (PME) in resistance to pine wood nematodes. (A) KEGG enrichment analysis of integrated transcriptomic and metabolomic data. (B) WGCNA analysis showing module‐metabolite correlations for candidate metabolites and differentially expressed genes (DEGs). (C) The putative biosynthetic pathway of PME and a heatmap of related DEGs. (D) Heatmap of differences of PME between highly resistant and susceptible individuals of Pinus massoniana (Ponderosa pine) at different time points. (E) Quantification of PME content in stem segments at different time points using LC–MS. Data are presented as mean ± SD; *p < 0.05, **p < 0.01. (F) Mortality rates of Bursaphelenchus xylophilus treated with a gradient of PME concentrations in vitro. Nematode survival was assessed at 1/3/5 days post‐treatment.

These enriched pathways comprised nine metabolites, which were subsequently used as phenotypic traits for weighted gene co‐expression network analysis (WGCNA) to identify gene modules highly correlated with their accumulation patterns (Figure 2B). Comparative analysis of these nine metabolites between the R and S clones showed that, at 7 dpi, PME, a stilbenoid derivative, accumulated at significantly higher levels in the R than in the S clones (Figure 2D). Moreover, multiple genes involved in the PME biosynthetic pathway exhibited changes coordinated with PME accumulation (Figure 2C). The specific genes associated with this pathway are listed in Table S1.

In order to verify the reliability of the above results. We quantified PME levels of inoculated stem segments by LC–MS. The results showed that at 7 dpi, the PME content in the resistant stem segments reached as high as 2957 μg kg⁻¹, which was more than 5 times higher than that in the susceptible stem segments (Figure 2E). The specific data obtained from the LC–MS are presented in Table S2. Based on the actual content as the standard, we conducted an in vitro nematode treatment experiment. The results demonstrated a strong dose‐ and time‐dependent nematicidal effect of PME against PWN. After 1 day of exposure, the number of surviving nematodes declined sharply at PME concentrations above 500 μg kg⁻¹, and the survival rate decreased to below 50% at 1000 μg kg⁻¹. Consistently, the LC₅₀ values varied with prolonged exposure time (Table S3). After 3 days of treatment, PME at 500 μg kg⁻¹ induced more than 50% nematode mortality. By 5 days, even 200 μg kg⁻¹ PME resulted in over 50% mortality, whereas 2000 μg kg⁻¹ PME caused complete mortality. Together, these findings demonstrate that PME has potent nematicidal activity against B. xylophilus (Figure 2F). This provides direct evidence that PME is a key defensive metabolite with anti‐nematode activity in P. massoniana .

2.3. Phylogenetic Analysis, Expression Profiling, and Subcellular Localization of PmPMT2‐1

To identify the key gene of PME biosynthesis, we focused on the pinosylvin O‐methyltransferase (PMT) gene family, as WGCNA indicated its strong correlation with PME accumulation. We generated a phylogenetic analysis of PMT homologues from several pine species and found that PmPMT2‐1 and PsPMT2 were clustered in the same branch. They may have similar functions (Figure 3A). Expression profiling showed that PmPMT2‐1 transcript levels were rapidly and strongly induced upon B. xylophilus inoculation, peaking at 5 dpi in the resistant clone—a pattern mirroring PME accumulation (Figure 3B,C). Subcellular localization confirmed that the PmPMT2‐1‐GFP fusion protein was localized to the cytoplasm and nucleus (Figure 3D), consistent with its role as a metabolic enzyme.

FIGURE 3.

PmPMT2‐1 is closely related to the synthesis of pinosylvin monomethyl ether (PME). (A) Phylogenetic tree of the PMT gene family of Pinus strobus , Pinus massoniana and Pinus sylvestris . (B) Heatmap showing the relative expression levels of PMT family genes between the resistant (R) and susceptible (S) clones. Red indicates high expression in R clones, blue indicates high expression in S clones, and the size of the circle represents the fold difference in expression compared to susceptibility. (C) Reverse transcription‐quantitative PCR analysis of PmPMT2‐1 expression in R and S clones; **p < 0.01. (D) Subcellular localization of PmPMT2‐1. From left to right: mCherry‐labelled nuclear marker, bright field, GFP signal and merged field. Scale bar = 20 μm.

2.4. Overexpression of PmPMT2‐1 Enhances PME Accumulation in P. massoniana

We validated the function of PmPMT2‐1 through overexpression and gene‐silencing assays. Transient overexpression in pine stem segments increased PmPMT2‐1 transcript levels by 194‐fold, whereas PME content increased by only 1.8‐fold (Figure 4B,C). This discrepancy may be attributable to limited substrate availability or to the absence of specific inducing conditions required for PME accumulation, as PME is an inducible metabolite. Virus‐induced gene silencing (VIGS) of PmPMT2‐1 reduced PME accumulation by approximately 50% (Figure 4D,E). In addition, we generated stable transgenic calli (Figure 4F–I) and hairy roots (Figure 4J) overexpressing PmPMT2‐1. Compared with the wild‐type (WT) and empty‐vector controls, PmPMT2‐1 transcript levels were significantly elevated in the transgenic calli and hairy roots (Figure 4K,M). PME production was detected in aged transgenic calli (Figure 4N); however, the PME content was lower than the concentration required to induce complete nematode mortality under in vitro conditions, which may likewise be associated with the inducible nature of PME. Moreover, PME‐mediated resistance may not depend solely on direct nematode killing but may also involve suppression of nematode activity and reproduction, as well as synergistic interactions with other defence responses. Furthermore, we confirmed the enzymatic activity of the PmPMT2‐1 protein through an in vitro enzyme assay, demonstrating that it catalyses the conversion of pinosylvin to PME (Figure 4O).

FIGURE 4.

Validation of PmPMT2‐1 overexpression. (A) Tissue‐cultured seedling of Pinus massoniana . (B) Reverse transcription‐quantitative PCR (RT‐qPCR) analysis of transient overexpression of PmPMT2‐1. Control represents the same tissue‐cultured seedling transformed with an empty vector. (C) Content determination of pinosylvin monomethyl ether (PME) in seedlings transiently overexpressing PmPMT2‐1. Control represents the PME content in seedlings transformed with an empty vector. (D) RT‐qPCR analysis of PmPMT2‐1 silencing via virus‐induced gene silencing (VIGS). Control represents the same tissue‐cultured seedling transformed with an empty vector. (E) PME content determination in PmPMT2‐1‐silenced seedlings via VIGS. Control represents the PME content in seedlings transformed with an empty vector. (F) Phenotype of PmPMT2‐1‐overexpressing calli. WT, wild type, untransformed calli. (G) Cell morphology of PmPMT2‐1‐overexpressing calli. (H) Morphology of PmPMT2‐1‐overexpressing embryoids. (I) Morphology comparison of WT and PmPMT2‐1‐overexpressing calli after 3 months of continuous culture. (J) Phenotype of PmPMT2‐1‐overexpressing hairy roots. (K) RT‐qPCR analysis of PmPMT2‐1 expression in WT and overexpressing calli. (L) RT‐qPCR analysis of PmPMT2‐1 expression in WT and overexpressing calli after 3 months of continuous culture. (M) RT‐qPCR analysis of PmPMT2‐1 expression in overexpressing hairy roots. (N) PME content determination in PmPMT2‐1‐overexpressing calli after 3 months of culture. (O) Enzymatic activity assay of PmPMT2‐1, **p < 0.01.

To further elucidate the role of PmPMT2‐1 in resistance to B. xylophilus infection, we performed aseptic B. xylophilus inoculation assays using overexpressing calli. Five days after inoculation with aseptic B. xylophilus, wild‐type P. massoniana calli exhibited a pronounced water‐soaked phenotype. Although the empty‐vector control did not show hyperhydricity, it displayed marked browning and ceased proliferation; in contrast, the overexpressing calli continued to grow normally (Figure 5). Collectively, these results indicate that PmPMT2‐1 is an essential methyltransferase gene required for PME biosynthesis and may represent one of the key genetic factors underlying resistance to PWD.

FIGURE 5.

Phenotypes of PmPMT2‐1‐overexpressing calli after inoculation with Bursaphelenchus xylophilus. Wild type (WT) represents untransformed calli, Control represents calli transformed with an empty vector, and PmPMT2‐1 represents overexpressing calli.

2.5. PmWRKY‐10 Binds Specifically to the PmPMT2‐1 Promoter and Functions as a Transcriptional Activator

To elucidate the upstream transcriptional regulatory mechanism of PmPMT2‐1, we constructed a yeast library using stem tissues of P. massoniana collected at different time points after inoculation with B. xylophilus. We then cloned the PmPMT2‐1 promoter and used its core region, containing a W‐box motif, to generate the pAbAi‐PmPMT2‐1 bait vector for yeast one‐hybrid (Y1H) library screening. This screen identified PmWRKY‐10 as a candidate transcription factor. Subsequent one‐to‐one Y1H assays confirmed that PmWRKY‐10 specifically bound to the PmPMT2‐1 promoter region (Figure 6D).

FIGURE 6.

PmWRKY‐10 promotes the expression of PmPMT2‐1 by binding to its promoter. (A) The imaging of ProPmPMT2‐1‐Luc and PmWRKY‐10‐62sk co‐expression in Nicotiana benthamiana leaves. The combinations of Luc and 62sk, Luc and PmWRKY‐10‐62sk and ProPmPMT2‐1‐Luc and 62sk were co‐injected as the controls. (B) The LUC/REN ratio of ProPmPMT2‐1‐Luc+PmWRKY‐10‐62sk, ProPmPMT2‐1‐Luc+62sk. *p < 0.05. (C) Electrophoretic mobility shift assay (EMSA) revealing PmWRKY‐10 binding to W‐box in the PmPMT2‐1 promoter. The PmWRKY‐10‐His protein was incubated with the biotin‐labelled hot probe or mutant probe, along with or without 10‐fold and 50‐fold excess of unlabelled cold probe as a competitor. The shifted probes and free DNA probes are indicated by black arrows. +, presence. (D) Yeast one‐hybrid (Y1H) assays showing that PmWRKY‐10 binds to PmPMT2‐1 promoter. p53‐pHis2+pGADT7‐p53 was the positive control, pHis2+pGADT7 was the negative control.

To verify the regulatory effect of PmWRKY‐10 on PmPMT2‐1 promoter activity, we performed a dual‐luciferase reporter assay in Nicotiana benthamiana leaves. Co‐expression of PmWRKY‐10 with a LUC reporter driven by the PmPMT2‐1 promoter resulted in a markedly enhanced luminescence signal, whereas the empty effector vector control showed only weak background activity (Figure 6A). Quantitative analysis further showed that PmWRKY‐10 increased PmPMT2‐1 promoter activity by approximately threefold compared with the control (Figure 6B), indicating that PmWRKY‐10 significantly activates the transcriptional activity of the PmPMT2‐1 promoter.

To further determine whether PmWRKY‐10 directly recognizes and binds the core cis‐acting element within the PmPMT2‐1 promoter, we performed an electrophoretic mobility shift assay (EMSA) (Figure 6C). Incubation of purified PmWRKY‐10 protein with the labelled probe corresponding to the promoter core region produced a clear shifted band, indicative of DNA–protein complex formation. Addition of 10‐ fold or 50‐fold molar excesses of unlabelled competitor probe progressively reduced the shifted signal, demonstrating the competitive and specific nature of this interaction. In contrast, the shifted band completely disappeared when a mutated probe was used, further confirming that recognition of the promoter core motif by PmWRKY‐10 depends on the specific cis‐element sequence.

Collectively, these results demonstrate that PmWRKY‐10 directly binds the core region of the PmPMT2‐1 promoter and activates its transcription. This finding reveals a key transcriptional regulatory module, PmWRKY‐10–PmPMT2‐1, which may play a crucial role in the defence response of P. massoniana induced by B. xylophilus infection.

2.6. Functional Characterization of PmWRKY‐10 and Its Regulatory Role in PME Biosynthesis

To elucidate the biological function of PmWRKY‐10, we first performed a phylogenetic analysis of WRKY transcription factor family members in the Pinus genome (Figure 7A). The results showed that the WRKY family members could be classified into eight subgroups. Among them, PmWRKY‐10 clustered within subgroup IIc and grouped with several conifer WRKY proteins previously implicated in defence responses, suggesting that PmWRKY‐10 may be involved in regulating stress responses and defence‐related metabolic processes.

FIGURE 7.

Functional verification of PmWRKY‐10. (A) Phylogenetic tree analysis of WRKY transcription factors, with different colours representing distinct subgroups. (B) Reverse transcription‐quantitative PCR (RT‐qPCR) analysis of PmWRKYs, Resistant: 1144‐2, Susceptible: He 9‐5; *p < 0.05, **p < 0.01. (C) Phenotype of Pinus massoniana calli overexpressing PmWRKY‐10. (D) Phenotype of P. massoniana calli overexpressing PmWRKY‐10 after continuous culture for 3 months. (E) The subcellular localization of PmWRKY‐10. mCherry is the nuclear marker, the second one is the bright field, the third one is the GFP signal, and the last one is the superimposed field, scale bar = 20 μm. (F) RT‐qPCR identification of PmWRKY‐10‐overexpressing calli, **p < 0.01. (G) RT‐qPCR identification of PmWRKY‐10‐overexpressing calli after continuous culture for 3 months. **p < 0.01. (H) RT‐qPCR identification of PmWRKY‐10 transient overexpression. Control refers to transient transformation with empty vector in the same tissue‐cultured seedling. *p < 0.05. (I) Content determination of pinosylvin monomethyl ether (PME) in PmWRKY‐10 transient overexpression; Control refers to the content of PME after transient transformation with empty vector in the same tissue‐cultured seedling. *p < 0.05. (J) RT‐qPCR identification of PmWRKY‐10 virus‐induced gene silencing (VIGS). Control refers to transformation with empty vector in the same tissue‐cultured seedling. **p < 0.01. (K) Content determination of PME after PmWRKY‐10 VIGS silencing. Control refers to the content of PME after transformation with empty vector in the same tissue‐cultured seedling. **p < 0.01. (L) Phenotypic changes of P. massoniana calli overexpressing PmWRKY‐10 after inoculation with Bursaphelenchus xylophilus nematodes; WT represents wild type, Control represents empty vector control, scale bar = 1 cm. (M) Expression changes of PmPMT2‐1 in PmWRKY‐10 transgenic calli. **p < 0.01.

We next examined the expression dynamics of PmWRKY‐10 during infection by B. xylophilus. Reverse transcription‐quantitative PCR (RT‐qPCR) analysis showed that PmWRKY‐10 transcript levels were significantly induced after B. xylophilus inoculation, reaching a peak at 5 dpi and subsequently remaining at relatively high levels (Figure 7B). This inducible expression pattern was highly consistent with that of PmPMT2‐1, further supporting the possibility that these two genes function in the same defence‐related regulatory pathway. Subcellular localization analysis revealed that the PmWRKY‐10 fluorescence signal was predominantly localized in the nucleus (Figure 7E), consistent with its putative role as a transcription factor.

To further validate the regulatory function of PmWRKY‐10, we performed stable genetic transformation, transient overexpression, and VIGS‐mediated gene silencing assays using P. massoniana calli and young stem segments. Transient overexpression of PmWRKY‐10 increased its transcript level by approximately 3.4‐fold compared with the control (Figure 7H), accompanied by a 1.8‐fold increase in the content of its downstream metabolite, PME (Figure 7I). Conversely, VIGS‐mediated silencing of PmWRKY‐10 significantly reduced its expression (Figure 7G) and resulted in an approximately 1.8‐fold decrease in PME accumulation (Figure 7K). These results indicate that PmWRKY‐10 positively regulates PME biosynthesis. In stably transformed transgenic calli, PmWRKY‐10 expression was significantly higher than that in the control (Figure 7C,D,F,G), confirming the successful generation of transgenic materials.

Notably, PmPMT2‐1 expression was also significantly upregulated in PmWRKY‐10‐overexpressing calli (Figure 7M), further demonstrating that PmWRKY‐10 can activate PmPMT2‐1 transcription in vivo. To evaluate its role in resistance to B. xylophilus infection, we inoculated transgenic calli with B. xylophilus. Compared with the control, PmWRKY‐10‐overexpressing transgenic calli exhibited a stronger resistance phenotype (Figure 7L), suggesting that PmWRKY‐10 may also participate in the defence response of P. massoniana against B. xylophilus infection.

Collectively, these results demonstrate that PmWRKY‐10 acts as a positive transcriptional activator of PmPMT2‐1. Upon induction by B. xylophilus infection, PmWRKY‐10 is strongly activated and promotes PME accumulation by directly upregulating PmPMT2‐1 expression. These findings indicate that the PmWRKY‐10–PmPMT2‐1–PME regulatory module contributes to the defence of P. massoniana against B. xylophilus.

3. Discussion

Previous studies have shown that resistant pine genotypes can effectively restrict the migration, colonization and reproduction of B. xylophilus within host tissues, thereby reducing nematode population size and delaying symptom development (Carrasquinho et al. 2018; Menéndez‐Gutiérrez et al. 2018; Gaspar et al. 2020; Modesto et al. 2021). In contrast, susceptible genotypes generally fail to establish such restrictions. In addition, multiple classes of defence‐related compounds induced in pine under stress, including resin acids, terpenoids and phenolic compounds, may contribute collectively to host resistance by inhibiting nematode movement, interfering with feeding and reproduction, or increasing nematode mortality (Kusumoto et al. 2014; Celedon and Bohlmann 2019; Faria et al. 2021). Therefore, resistance of P. massoniana to B. xylophilus infection is unlikely to depend on a single defence factor alone (Kusumoto et al. 2014).

In this study, although we successfully obtained transgenic calli and hairy root materials and used these systems for candidate gene functional validation, these materials cannot fully recapitulate the complex physiological processes occurring in intact plants, including the vascular system, resin duct structure, xylem microenvironment and systemic defence responses. Therefore, their application remains constrained by tissue type and functional specialization (Ikeuchi et al. 2016, 2019). Further optimization of the genetic transformation and regeneration system in P. massoniana is still required to establish a stable and efficient platform for generating transgenic plants, which would provide more suitable materials for validating candidate gene functions at the whole‐plant level (Gutierrez‐Valdes et al. 2020). In addition, direct quantification of nematode population size, migration distance and reproductive dynamics in transgenic materials or intact plants will be an important direction for future studies aimed at clarifying the resistance functions of candidate genes.

In conifers, stilbenes such as pinosylvin and its derivatives have been widely recognized as important chemical defence compounds against fungi, insects and nematodes (Koo et al. 2022). Previous studies have shown that stilbenes such as PME play defence‐related roles in several Pinus species (Bryant et al. 1983; Chiron et al. 2000; Hwang et al. 2021; Proença et al. 2017; Valletta et al. 2021).

By integrating resistance phenotyping, biochemical assays, transcriptional regulatory validation and functional genomics, this study identified a previously uncharacterized defence regulatory pathway, PmWRKY‐10→PmPMT2‐1→PME. Our results provide the first evidence that PME participates in the resistance of P. massoniana to B. xylophilus and further clarify its upstream transcriptional regulatory mechanism. The accumulation of PME may, in turn, contribute to disease resistance in P. massoniana .

However, in this study, the PME concentration detected in transgenic calli differed from the lethal concentration observed in vitro. This discrepancy may arise because plant defence responses in vivo are influenced by multiple factors (Jeandet et al. 2013; Celedon and Bohlmann 2019; Desmedt et al. 2020). Thus, although in vitro assays demonstrated that PME has direct nematicidal activity, the PME content detected in transgenic cell masses or tissues should not be directly interpreted as equivalent to the exogenous treatment concentration required to kill nematodes. For example, the PME concentration of 2000 μg kg⁻¹ in the in vitro assay represents the level required to achieve 100% mortality of B. xylophilus under specific artificial conditions, whereas the PME content measured in plant tissues reflects the overall endogenous abundance of this compound within biological materials. These two measurements differ substantially in exposure mode, biological context and bioavailability (Desmedt et al. 2020; Faria et al. 2021).

In planta, resistance may depend more strongly on local PME enrichment at infection or feeding sites, prolonged exposure of nematodes to PME‐containing tissues and combined effects between PME and other pine‐derived defence metabolites. Therefore, even when the overall accumulation level of PME is below the concentration required for complete lethality in vitro, it may still contribute to resistance by reducing nematode motility, disrupting feeding behaviour, suppressing reproduction or delaying population expansion, rather than causing direct and rapid complete mortality. Accordingly, the enhanced resistance observed in transgenic calli likely reflects the combined effects of increased PME accumulation and associated defence responses.

3.1. Elucidating the PME Biosynthetic Regulatory Pathway

Phylogenetic analysis showed that PmPMT2‐1 clusters with PMT2‐like proteins from P. sylvestris , suggesting that these enzymes may share similar catalytic functions. Previous studies in P. sylvestris demonstrated that PMT2 is a key O‐methyltransferase responsible for catalysing the methylation of pinosylvin to PME, and that PMT2 exhibits markedly greater substrate specificity for pinosylvin than PMT1, which has a broader catalytic spectrum (Chiron et al. 2000). Moreover, the expression of PMT2 in P. sylvestris is coordinated with STS and upstream phenylpropanoid biosynthetic genes, suggesting that PMT2 may be precisely regulated at the transcriptional level. However, the transcription factor directly controlling PMT2 promoter activity had not previously been identified (Paasela et al. 2017). Against this background, our study provides functional evidence, using complementary overexpression and gene‐silencing approaches, that the expression level of PmPMT2‐1 affects PME accumulation and thereby contributes to the resistance of P. massoniana to B. xylophilus. This finding is consistent with observations in P. strobus , where PMT genes are strongly induced after nematode infection and their expression is closely associated with PME content.

Notably, transient overexpression of PmPMT2‐1 in our experimental system did not lead to a significant increase in PME content. In addition, the basal PME concentration detected in transgenic calli was substantially lower than the lethal concentration observed in the in vitro toxicity assay. These results suggest that PME accumulation is not determined solely by PmPMT2‐1 transcript abundance but may depend on more complex factors, including elicitation status, substrate availability and the overall metabolic network context (Zaynab et al. 2018; Isah 2019; Desmedt et al. 2020; Erb and Kliebenstein 2020). As a typical inducible defence metabolite, PME may require pathogen infection, wounding signals or specific elicitors for robust accumulation, together with sufficient availability of upstream substrates and methyl donors (Celedon and Bohlmann 2019; Isah 2019; Dong and Lin 2021). Therefore, under the present experimental conditions, upregulation of a single biosynthetic gene may be insufficient to drive substantial accumulation of the final product. It should also be emphasized that callus tissue differs markedly from intact plant tissues in cellular composition, tissue organization, metabolic compartmentalization and defence‐related physiological status (Ikeuchi et al. 2019; Modesto et al. 2022). Thus, its metabolic capacity and defence responsiveness may not fully recapitulate the in planta resistance process. Under natural infection conditions, PME may act synergistically with other inducible defence compounds and host responses, such as cell wall reinforcement, reactive oxygen species (ROS) burst and localized programmed cell death (Desmedt et al. 2020; Erb and Kliebenstein 2020; Modesto et al. 2022). Moreover, the local PME concentration at infection sites may be substantially higher than the average concentration measured at the whole‐tissue level. Accordingly, PME need not reach the concentration required for 100% nematode mortality in vitro to contribute to resistance in vivo (Faria et al. 2021).

The non‐linear relationship between PME accumulation and the expression of related biosynthetic genes also reflects a common feature of plant secondary metabolism. Although key genes in the PME biosynthetic pathway may be transcriptionally induced, metabolite accumulation is often constrained by multiple regulatory and metabolic factors (Fernie and Tohge 2017; Sweetlove et al. 2017; Dong and Lin 2021). One important limitation may be the availability of the upstream substrate pinosylvin and the methyl donor S‐adenosyl‐L‐methionine (SAM). In plant secondary metabolic pathways, precursor availability is frequently a central determinant of metabolic flux (Deng and Lu 2017; Dong and Lin 2021). For example, malonyl‐CoA and cinnamic acid are not only required for stilbene biosynthesis but also serve as shared substrates for multiple phenylpropanoid‐derived pathways; their availability, therefore, directly influences the final accumulation of target metabolites (Jeandet et al. 2013; Deng and Lu 2017; Dong and Lin 2021). Thus, even when stilbenoid biosynthetic genes are strongly activated under stress conditions, insufficient upstream precursor supply may still restrict PME biosynthesis (Sweetlove et al. 2017).

In addition to substrate availability, flux bottlenecks at any enzymatic step in the PME biosynthetic pathway may limit final product accumulation (Fernie and Tohge 2017; Sweetlove et al. 2017; Fu et al. 2018). PME formation depends on the coordinated action of multiple enzymes in the phenylpropanoid and stilbenoid pathways, including phenylalanine ammonia‐lyase (PAL), 4‐coumarate:CoA ligase (4CL), stilbene synthase (STS) and O‐methyltransferases (Jeandet et al. 2013; Deng and Lu 2017; Dong and Lin 2021). A rate‐limiting step at any point in this multienzyme process may lead to intermediate accumulation or insufficient downstream product formation, thereby restricting the overall metabolic flux of the pathway (Sweetlove et al. 2017; Fu et al. 2018). Such flux constraints are commonly observed in both natural plant defence metabolism and engineered metabolic systems, and may explain why upregulation of a single key enzyme gene does not always translate linearly into increased accumulation of the target metabolite (Fernie and Tohge 2017).

Furthermore, transcript abundance does not necessarily reflect the abundance or activity of the corresponding enzyme (Liu et al. 2016; Buccitelli and Selbach 2020; Mergner et al. 2020). Post‐transcriptional regulation, translational efficiency, protein stability, post‐translational modification and subcellular localization can all influence enzymatic output (Vu et al. 2018). For instance, phosphorylation, acylation or other protein modifications may alter enzyme activity, stability or access to substrates, thereby causing discrepancies between gene expression and metabolite accumulation (Vu et al. 2018; Buccitelli and Selbach 2020; Mergner et al. 2020). Such non‐linear relationships among transcript abundance, protein activity and metabolic output have been documented in a wide range of plant secondary metabolic pathways (Liu et al. 2016; Fernie and Tohge 2017; Sweetlove et al. 2017). Therefore, transcriptomic or gene expression data alone are insufficient to fully resolve the regulatory logic of PME biosynthesis. Future studies integrating proteomics, enzyme activity assays, metabolic flux analysis and subcellular localization will be required to more precisely define the metabolic regulation mediated by PmPMT2‐1.

Competition among parallel branches of the metabolic network may also affect PME biosynthesis (Deng and Lu 2017; Fu et al. 2018; Dong and Lin 2021). Shared precursors such as cinnamic acid and malonyl‐CoA can be diverted into the biosynthesis of lignin, flavonoids, other stilbenes and diverse phenolic compounds (Erb and Kliebenstein 2020). During pathogen infection, multiple defence‐related branches may be activated simultaneously, resulting in the redistribution of limited precursor pools among competing metabolic routes (Deng and Lu 2017; Liu et al. 2016). If more flux is directed toward lignin deposition, cell wall reinforcement or the synthesis of other defensive phenolics, the flux toward pinosylvin and PME may be correspondingly reduced (Erb and Kliebenstein 2020; Modesto et al. 2021). Such competition and redirection of shared intermediates are important features of plant metabolic network regulation and can substantially influence the accumulation of target defence metabolites (Deng and Lu 2017).

Temporal uncoupling between transcriptional activation and metabolite accumulation may further contribute to the observed discrepancy between PmPMT2‐1 expression and PME content (Isah 2019). In plant defence responses, transcriptional induction often occurs rapidly, whereas metabolite accumulation requires subsequent processes such as substrate biosynthesis, enzyme translation and maturation, metabolic transport and product stabilization (Fernie and Tohge 2017; Sweetlove et al. 2017). Consequently, metabolite accumulation often lags behind gene expression. In long‐lived woody plants such as pines, secondary metabolic regulation may be further influenced by tissue differentiation status, resin duct development and the formation of localized defence structures (Celedon and Bohlmann 2019). Therefore, transcriptional upregulation detected at a single time point may not coincide with the peak of PME accumulation. This temporal decoupling further supports the view that PME biosynthesis is governed by a multilayered regulatory network rather than by a simple linear relationship between gene expression and metabolite accumulation (Vu et al. 2018; Isah 2019).

Defence metabolite profiles induced by B. xylophilus infection may differ substantially among pine species. For example, P. densiflora , P. koraiensis and P. strobus display distinct patterns of stilbenoid accumulation following nematode challenge (Hwang et al. 2021). Such interspecific variation in metabolic responses may represent an important basis for differences in disease resistance among pine species and suggests that different pines may employ distinct chemical defence strategies to restrict pathogen progression.

WRKY transcription factors are major regulators of plant responses to biotic and abiotic stresses and play central roles in pathogen perception, signal transduction and activation of defence metabolism (Javed and Gao 2023). In P. strobus , WRKY, MYB and bHLH transcription factors are significantly upregulated after B. xylophilus infection, and MYB4 has been proposed as a potential regulator of PME biosynthesis (Kim et al. 2024). These findings suggest that transcriptional regulators may contribute to the establishment of pine defence responses (Hwang et al. 2021). Similarly, in P. sylvestris , the coordinated expression of PMT2/STS with specific MYB or NAC transcription factors during heartwood formation and wound responses indicates that stilbenoid‐ and PME‐related metabolic pathways are likely governed by complex transcriptional networks (Paasela et al. 2017). In P. massoniana , members of the WRKY family have also been implicated in multiple stress responses (Sun et al. 2022). Our study further establishes a direct mechanistic link between WRKY‐mediated transcriptional regulation and PME biosynthesis. Using promoter–reporter assays, EMSA and Y1H assays, we demonstrated that PmWRKY‐10 specifically binds to the W‐box cis‐element in the PmPMT2‐1 promoter and activates its transcription. These results provide direct mechanistic evidence for the existence of a WRKY–PMT regulatory module in conifers and fill an important knowledge gap regarding the upstream transcriptional regulation of PME biosynthesis.

Collectively, our phylogenetic analysis, expression profiling, functional validation, promoter‐binding assays, metabolite measurements and resistance phenotyping support a model in which the PmWRKY‐10–PmPMT2‐1 module plays a central role in regulating PME biosynthesis. This module may be activated during B. xylophilus infection, promoting the accumulation of PME and related defence metabolites and thereby enhancing resistance in P. massoniana . Nevertheless, although transcriptional activation assays and expression analyses support the positive regulation of PmPMT2‐1 by PmWRKY‐10, the genetic dependency between these two genes remains to be further validated. In this study, we attempted to combine PmWRKY‐10 overexpression with PmPMT2‐1 silencing to determine whether PmWRKY‐10‐mediated resistance enhancement depends on PmPMT2‐1. However, simultaneous manipulation of the two genes showed low efficiency and poor reproducibility in the current pine transformation system. In addition, the stable generation of PmWRKY‐10‐overexpressing hairy roots remains technically challenging. Future optimization of P. massoniana hairy‐root and stable transformation systems, together with CRISPR/Cas‐mediated genome editing, dual‐gene genetic interaction analyses, and in situ metabolite detection, will enable a more direct dissection of the function of the PmWRKY‐10–PmPMT2‐1 regulatory axis in resistance to B. xylophilus.

It is also noteworthy that, in Arabidopsis, MAPK cascades can activate WRKY transcription factors through phosphorylation, thereby inducing phytoalexin biosynthesis (Mao et al. 2011). Therefore, an important future direction will be to determine whether a similar MAPK–WRKY signalling module operates in P. massoniana , and whether nematode infection, mechanical injury or pathogen‐associated molecular patterns activate PmWRKY‐10 through upstream kinase cascades. Elucidating this signalling pathway will help establish an integrated mechanistic framework linking pathogen perception, signal transduction, transcriptional regulation, defence metabolism and resistance formation in P. massoniana against B. xylophilus.

3.2. Biological Significance and Application Prospects of PME

As a constitutive and inducible defence compound, PME likely confers broad‐spectrum resistance, potentially acting against fungal pathogens and decay organisms in addition to nematodes (Jeandet et al. 2002; Steiner and Buhrer 1934). This aligns with studies linking heartwood stilbene content in P. sylvestris to decay resistance, suggesting that the induction of such metabolites is an evolutionary adaptation to multifactorial biotic stress.

Our identification of a direct WRKY‐stilbene pathway link in pine complements the previously reported WRKY‐STS regulatory module in grapevine (Vannozzi et al. 2018), indicating a potentially conserved regulatory mechanism across woody plants. From a translational perspective, PmWRKY‐10 and PmPMT2‐1 serve as valuable molecular markers or potential targets for genome editing to enhance resistant germplasm. Moreover, the demonstrated nematicidal activity of PME suggests its potential, or that of its derivatives, for development into topical biopesticides for sustainable forest management.

Future research should focus on elucidating the regulatory network surrounding PmWRKY‐10, including its potential interactions with other transcription factors (e.g., MYB, bHLH and NAC) and its activation by upstream signalling cascades. Such work will further unravel the complexities of conifer defence and inform strategies to enhance tree resilience.

In conclusion, this study systematically deciphers the molecular basis of P. massoniana resistance to B. xylophilus infection, culminating in the identification of a novel defence regulatory pathway: PmWRKY‐10→PmPMT2‐1→PME (Figure 8). By integrating phenotypic, cellular, metabolic and molecular evidence, we demonstrate that PmWRKY‐10 functions as an upstream transcriptional activator, PmPMT2‐1 serves as the key biosynthetic enzyme, and PME acts as the terminal defensive metabolite with direct nematicidal activity.

FIGURE 8.

The disease resistance control model of Pinus massoniana .

Our findings provide a coherent mechanistic framework for understanding induced stilbenoid‐based defence in conifers. The elucidation of this pathway not only advances fundamental knowledge of conifer resistance but also identifies PmWRKY‐10 and PmPMT2‐1 as promising targets for molecular breeding and the development of sustainable strategies for forest disease management.

4. Experimental Procedures

4.1. Plant and B. xylophilus Materials

We used 7‐year‐old clonal progenies of P. massoniana , consisting of a genetically resistant genotype (R: 1144‐2) and a susceptible genotype (S: He 9‐5). The resistance status of these genotypes was rigorously established through three consecutive years of controlled artificial inoculation with B. xylophilus. Each year, we inoculated individual plants with 10,000 nematodes and included 30 biological replicates per genotype. We classified a genotype as highly resistant if its average disease incidence remained below 10%, and as susceptible if the incidence exceeded 80% across all trials. All experimental pines were grown at the Linhai Forest Farm, Zhejiang Province, China (28°52′49″ N, 121°00′06″ E). For subcellular localization assays, we employed N. benthamiana plants grown under standard greenhouse conditions. Furthermore, we obtained P. massoniana callus cultures from the Jiangxi Academy of Forestry. The wild‐type B. xylophilus isolate used in this study was freshly extracted from diseased pine wood collected in the current year.

4.2. Nematode Culture and Plant Inoculation

We maintained and propagated B. xylophilus on cultures of Botrytis cinerea. We first cultured B. cinerea on a wheat‐based medium (containing 20 g dehulled wheat and 15 mL of a 2.5% sucrose solution, autoclaved) at 23°C for 5 days. We then transferred the nematodes to these fungal cultures and incubated them at 26°C for 5 days. We harvested nematodes using the Baermann funnel technique.

For inoculations, we chose 1‐year‐old lateral branches from both the 1144‐2 (R) and He 9‐5 (S) genotypes. Three individual trees from each clone were selected as biological replicates. We inoculated each branch with approximately 10,000 B. xylophilus individuals in aqueous suspension. Monitored and recorded disease symptom progression over time. Then harvested inoculated branch tissues at specified time points (0, 1, 3, 5, 7, 10, 15 and 30 dpi), immediately froze them in liquid nitrogen, and stored them at −80°C for subsequent analyses.

4.3. Phenotypic Assessment and Histological Analysis of PWN‐Inoculated Stem Tissues

We monitored external disease symptoms at 0, 1, 3, 5, 7, 10, 15 and 30 dpi. We assessed disease progression according to visible changes in stem tissues, including discolouration, browning and necrosis. At each observation point, we photographed representative stem segments to compare symptom development between the resistant and susceptible clones.

For histological analysis, we collected stem tissues adjacent to the inoculation site at 5 dpi. We immediately fixed the samples in FAA (formaldehyde/ethanol/acetic acid) solution, dehydrated them through a graded ethanol series, cleared them in xylene and embedded them in paraffin. We prepared 8‐μm‐thick transverse paraffin sections using a rotary microtome, then deparaffinized, rehydrated, and stained the sections with Safranin O–Fast Green. Three individual plants were sampled from both resistant and susceptible P. massoniana . Ten sections from each plant were stained, and representative sections were selected for imaging. We examined the stained sections under a light microscope and acquired images using a digital imaging system and annotated major anatomical regions, including the periderm, cortex, resin ducts, phloem, cambium and xylem. Finally, we evaluated histological differences between the R and S clones based on tissue organization and staining patterns.

4.4. Integrated Multi‐Omics Analyses

4.4.1. Transcriptome Sequencing

We extracted total RNA from frozen pine tissues using a TRIzol‐based method. To ensure high‐quality input material, we assessed RNA integrity using an Agilent 2100 Bioanalyzer, confirming an RNA Integrity Number (RIN) greater than 8.0 for all samples. We constructed sequencing libraries from 1 μg of total RNA per sample using the NEBNext Ultra II RNA Library Prep Kit. After quantifying the final libraries by Qubit fluorometry and verifying their size distribution on an Agilent 4200 TapeStation, we pooled them in equimolar ratios. We performed paired‐end sequencing (2 × 150 bp) on an Illumina NovaSeq 6000 platform, generating about 40 million read pairs per sample.

For full‐length transcriptome analysis, we processed raw data to generate Circular Consensus Sequences (CCS), which we subsequently corrected using LoRDEC with default parameters. We identified and clustered full‐length, non‐chimeric reads into transcript isoforms using the ICE algorithm in SMRT Link v. 11.1, resulting in a high‐confidence consolidated transcriptome for downstream analysis.

4.4.2. Metabolomic Profiling

We conducted untargeted metabolomics at Wuhan Meview Metabolomics Biotechnology Co. Ltd. We homogenized 100 mg of frozen plant powder in 1 mL of ice‐cold extraction solvent (80% methanol, 0.1% formic acid and 10 ng μL⁻¹ chlorophenylalanine as an internal standard). After vortexing and sonication for 30 min, we centrifuged the extracts at 12,000 g for 15 min. We analysed the supernatants using a UPLC‐MS/MS system (ExionLC AD coupled to a QTRAP 4500) operating in both positive and negative ESI modes.

Chromatographic separation was achieved on a Waters HSS T3 column (2.1 × 100 mm, 1.8 μm) with a 15‐min linear gradient from 5% to 95% acetonitrile (containing 0.1% formic acid) at 0.3 mL/min, with the column maintained at 40°C. MS parameters were ion source temperature 550°C, curtain gas 35 psi, and ion spray voltage ±4500 V. We identified metabolites by matching their accurate mass (±5 ppm) and MS/MS spectra against commercial standards (pinosylvin and PME; Sigma‐Aldrich) and the HMDB database. We quantified metabolites using external calibration curves with at least six concentration points.

4.4.3. Transcriptome–Metabolome Integration and WGCNA Analysis

We identified DEGs between resistant (R) and susceptible (S) samples at each sampling time point and also compared gene expression at each time point with that at Day 0. To eliminate transcriptional changes caused by mechanical wounding or other non‐specific responses, we excluded genes that showed significant differential expression following water inoculation.

We analysed transcriptome data using the DESeq2 package implemented in TBtools (v. 1.098722). We defined DEGs as genes with a log₂FC ≥ 1 and a false discovery rate (FDR) < 0.05. We visualized DEG expression patterns by hierarchical clustering and heatmap analysis (Huang et al. 2022).

For metabolomic analysis, we identified DAMs using the criteria log₂FC ≥ 1, FDR < 0.05 and VIP ≥ 1.0. We then performed KEGG pathway enrichment analyses separately for DEGs and DAMs and compared the results to identify pathways enriched in both datasets.

Next, we selected nine metabolites from the co‐enriched pathways for integrative network analysis. We used the abundance values of these metabolites as phenotypic traits and conducted WGCNA in R to identify gene modules associated with metabolite accumulation patterns. The WGCNA scripts used in this study are provided in File S2.

Before network construction, we filtered out low‐expression genes (FPKM < 1) to reduce background noise and improve network stability. We then selected the soft‐thresholding power based on the criterion that the resulting network achieved a scale‐free topology fit index (R ²) greater than 0.85. We constructed the co‐expression network using the normalized gene expression matrix and identified modules with a minimum module size of 35 genes. Subsequently, we merged highly similar modules using a merge cut height of 0.25 to improve module robustness and biological relevance (Langfelder and Horvath 2008).

Finally, we calculated correlations between module eigengenes and metabolite abundance traits and identified modules significantly associated with the selected metabolites. We prioritized candidate resistance‐related genes based on their membership in trait‐associated modules, their differential expression patterns and their connectivity within the co‐expression network.

4.5. Quantification of PME and In Vitro Treatment of B. xylophilus

Based on the expression patterns of the target genes in the transcriptome data, stem segments from resistant and susceptible clones were collected at 0 and 7 dpi. The nine selected metabolites were quantified by LC–MS (Thermo Fisher Scientific) to determine their actual abundance in pine tissues. Three experimental replicates were performed. Detailed information on the selected metabolites is provided in Table S6. For sample preparation, we ground frozen tissues to a fine powder, weighed 0.1 g aliquots, and added 5 mL of methanol. Following 1 h of ultrasonic extraction, we centrifuged 2 mL of the supernatant at 12,000 g for 5 min and filtered it through a 0.22 μm organic membrane prior to LC–MS analysis.

LC–MS conditions were as follows: column: Waters HSS T3 (2.1 × 100 mm, 1.8 μm); mobile phase: 0.01% formic acid in water and acetonitrile; injection volume: 1 μL; column temperature: 35°C; ion source: ESI; polarity: negative ion mode; ion source temperature: 120°C; capillary voltage: 2.5 kV; detection: Multiple Reaction Monitoring (MRM). In this section, a single‐sample standard‐based quantitative assay was used, whereas the metabolomic analysis was performed using a widely targeted metabolomics approach. These two methods differ in their analytical objectives, sample preparation procedures and data acquisition and analysis strategies; therefore, the experimental methodologies employed are not the same.

Integrating the metabolomic data with the changes in the actual abundance of the nine metabolites in pine tissues, we selected three metabolites based on their accumulation patterns and designed concentration‐gradient treatment assays. For each treatment, 1000 nematodes were incubated in solutions containing different concentrations of the selected metabolites. Nematode survival was recorded by direct counting under an optical microscope (Axiocam ERc 5 s; Zeiss). Three experimental replicates were performed for each treatment.

Mortality at each concentration was calculated based on the observed survival data, and dose–response curves were generated according to the relationship between metabolite concentration and nematode mortality. The median lethal concentration (LC₅₀) was estimated using both probit and logit analyses.

The mortality rate was calculated using the following formula:

$$\text{Mortality}\left(\%\right)=\frac{N_{\text{dead}}}{N_{\text{total}}}\times 100$$

where $N_{\text{dead}}$ represents the number of dead nematodes and $N_{\text{total}}$ represents the total number of nematodes examined.

Probit analysis:

The mortality rate was transformed into a probit value using the following equation:

$$\text{Probit}\left(p\right)={\mathrm{\Phi }}^{-1}\left(p\right)$$

where ${\mathrm{\Phi }}^{-1}\left(p\right)$ denotes the inverse cumulative distribution function of the standard normal distribution, and $p$ represents the mortality proportion (e.g., a mortality rate of 50% corresponds to $p=0.5$).

4.6. Gene Cloning, Sequence Analysis and Subcellular Localization

Based on the in vitro treatment assays, WGCNA analysis and RT‐qPCR results, genes whose expression patterns were consistent with the disease phenotype and PME accumulation trends were selected as candidate genes for further validation. We isolated total RNA from inoculated pine stems using the EASY38spinPlus Plant RNA Kit (Aidlab) and synthesized first‐strand cDNA using a commercial reverse transcription kit (Takara). We designed gene‐specific primers (Table S5) to amplify the full‐length coding sequence of PmPMT2‐1 via PCR. We confirmed the sequence and deposited it in GenBank under accession number PQ217824.

To determine the subcellular localization of PmPMT2‐1, we constructed C‐terminal GFP and RUBYRed fusion vectors (pCambia1300‐GFP‐PmPMT2‐1 and pCAMBIA2300‐RUBYRed‐PmPMT2‐1) using seamless cloning (vector schematics in Figure S1). The RUBY reporter system was used for visual identification of transformed tissues. RUBY is a convenient, equipment‐free reporter that enables direct detection of transgene expression based on the accumulation of the red pigment betalain in transformed plant tissues. The betalains are tyrosine‐derived secondary metabolites whose biosynthesis requires the coordinated activity of three enzymes: CYP76AD1, DODA and a glucosyltransferase (GT). In conventional eukaryotic expression systems, simultaneous expression of multiple proteins typically requires separate promoter–terminator cassettes for each coding sequence. In contrast, the RUBY system employs a single promoter to drive the transcription of a polycistronic expression cassette, in which the coding sequences of CYP76AD1, DODA and GT are linked by self‐cleaving 2A peptide sequences. This design enables coordinated expression of the three betalain‐biosynthetic enzymes from a single transcript, allowing transformed tissues to be readily distinguished by visible red pigmentation without the need for specialized instruments or staining procedures. We transformed these constructs into Agrobacterium tumefaciens GV3101 and verified them by PCR.

Subcellular localization: for transient expression in N. benthamiana, we grew the transformed agrobacteria (GFP‐PmPMT2‐1 and mCherry nuclear marker) to an OD₆₀₀ of 1.0, harvested the cells by centrifugation, and resuspended them in infiltration buffer (10 mM MgCl₂, 10 mM MES, pH 5.6, 200 μM acetosyringone). After incubating the suspension for 1 h, we infiltrated it into the abaxial side of young leaves. We kept the plants in the dark for 16 h post‐infiltration and then visualized fluorescence signals using a Zeiss LSM 700 confocal microscope (excitation 488 nm).

4.7. Recombinant Protein Expression and Enzyme Assay

We cloned the PmPMT2‐1 CDS into the pET‐28a vector for expression as an N‐terminal His‐tag fusion in Escherichia coli . We induced protein expression with 0.5 mM IPTG and purified the recombinant protein using Ni‐NTA affinity chromatography.

We assayed the methyltransferase activity of purified PmPMT2‐1 in a 1 mL reaction containing: 100 mM Tris–HCl (pH 7.5), 10 μM S‐adenosylmethionine (SAM), 10 μM pinosylvin and 5 μg of purified enzyme. We incubated the reaction at 30°C for 1 h (n = 3) and terminated it by adding 5 mL of methanol. We then processed the samples as described in section 4.4.2 and quantified PME production via LC–MS.

4.8. Functional Verification of PmPMT2‐1 in P. massoniana

4.8.1. Transient Overexpression

We transiently expressed PmPMT2‐1 in pine stem segments via Agrobacterium‐mediated transformation. The procedure was performed using vacuum infiltration, and the Agrobacterium suspension was adjusted to an OD₆₀₀ of 1.0. All other procedures were carried out as described for the subcellular localization assay. We vacuum‐infiltrated stem segments from tissue‐cultured seedlings with a suspension of A. tumefaciens GV3101 harbouring the pCambia1300‐PmPMT2‐1 overexpression vector. After co‐cultivation on Murashige and Skoog (MS) medium in the dark for 2–4 days, we harvested the tissues for RT‐qPCR and PME quantification. The experiment was performed with three biological replicates and three technical replicates.

4.8.2. Virus‐Induced Gene Silencing

To knock down PmPMT2‐1 expression, we constructed a VIGS vector (pTRV2‐pmt) containing a fragment of the target gene. A 300‐bp sequence fragment corresponding to the first 300 bp of PmPMT2‐1 was selected. The sequence alignment results are shown in Figure S2. We infiltrated pine stems with the VIGS construct and assessed silencing efficiency by RT‐qPCR, correlating it with PME content. The experiment was performed with three biological replicates and three technical replicates. The detailed procedure was the same as that used for transient transformation.

4.8.3. Stable Transformation of Callus

We vacuum‐infiltrated calli with the bacterial suspension for 10 min, co‐cultured them for 2 days, and then transferred them to recovery medium for 10 days. Finally, we selected transformed calli on medium containing 100 mg/L kanamycin and 400 mg/L timentin. The specific methods and medium formulations are as follows:

For callus transformation, we prepared the infiltration buffer (H + 200 μM AS) as per the method used for transient transformation. We placed the calli in the infection solution and subjected them to vacuum infiltration for 10 min. Subsequently, we blotted them dry on filter paper and transferred them to co‐culture medium (H + 200 μM AS+4.5 g/L gellan gum). Following a 2‐day co‐culture period, we washed three times with sterile water containing Tim, followed by two rinses with sterile water. We then transferred the calli to recovery medium for 10 days before moving them to selection medium (H + 4.5 g/L gellan gum+100 mg/L Kan + 400 mg/L Tim). The specific medium recipe can be found in File S1. Positive transgenic calli were identified based on colour changes and PCR analysis (Figure S3).

4.8.4. Agrobacterium rhizogenes ‐Mediated Hairy Root Induction in P. massoniana Seedlings

4.8.4.1. Cultivation of P. massoniana Seedlings

Vermiculite, peat moss and perlite were thoroughly mixed at a ratio of 1:3:1 (v/v/v) and used as the substrate for seed germination and seedling growth of P. massoniana . The seeds were evenly sown on the surface of the substrate and covered with a thin layer of soil. After thorough watering, the containers were covered to maintain adequate humidity for germination. Seeds typically germinated approximately 10 days after sowing. Seedlings grown for 1–2 months were used for subsequent hairy root induction experiments. We inoculated 200 seedlings for each gene.

4.8.4.2. Construction of the Transgenic Vector and Transformation of A. rhizogenes

Plant expression vectors were constructed according to the experimental objectives and introduced into A. rhizogenes . In this study, A. rhizogenes K599 was generally used for hairy root induction in P. massoniana . Positive transformants were selected on antibiotic‐containing medium and verified by PCR before being used for subsequent infection experiments.

4.8.4.3. Culture and Infection With A. rhizogenes

A PCR‐verified positive A. rhizogenes colony was inoculated into 5 mL of TY liquid medium containing the appropriate antibiotics and cultured overnight with shaking under suitable conditions. The antibiotics typically included 50 mg/L streptomycin and 50 mg/L kanamycin. The overnight culture was then transferred into 200 mL of fresh TY liquid medium at a ratio of 1:100 and cultured with shaking for approximately 8 h until the OD₆₀₀ reached approximately 1.0.

After preparation of the bacterial suspension, infection was performed as follows. Healthy and uniformly developed P. massoniana seedlings were selected, and their original root systems were excised. The cut sites were immersed in the prepared A. rhizogenes suspension for 20–30 min. After infection, the seedlings were transplanted into sterilized vermiculite, and approximately 3 mL of bacterial suspension was applied to the rhizosphere of each seedling. The plants were then maintained in a growth room under appropriate temperature and humidity conditions to promote hairy root induction. After new roots had formed, root samples were collected for transgenic verification.

It should be noted that not all newly formed roots are transgenic hairy roots; therefore, positive roots must be further identified by molecular detection and/or reporter‐gene phenotyping. In this system, the pCAMBIA2300‐RUBYRed‐SEQ vector can be used for transformation, allowing preliminary screening of putative transgenic roots using the RUBY reporter system, followed by further confirmation by PCR or other molecular methods.

4.9. Promoter Cloning and cis‐Element Analysis

We isolated genomic DNA from pine stems using the Genomic DNA Kit (Tiangen) and amplified the PmPMT2‐1 promoter region by PCR. The detailed procedure was the same as that used for gene cloning. We analysed the promoter sequence for putative cis‐regulatory elements using the PlantCARE database (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Table S4).

4.10. Y1H Screening

We constructed a yeast library using stem segments of P. massoniana collected at different time points after nematode inoculation. The library capacity and quality assessment are shown in Figure S4. We cloned a core fragment of the PmPMT2‐1 promoter into the pAbAi vector to create the bait construct. The pAbAi‐pmt vector was linearized with BstBI and co‐transformed with the pGADT7 library vector into the yeast strain Y1HGold. Autoactivation tests were performed to determine the optimal screening conditions (abscisic acid, ABA, concentration).

Yeast library screening: After determining the optimal ABA concentration, the pAbAi‐pmt bait vector and the library were co‐transformed into yeast cells for the screening experiment.

Candidate transcription factor identification: Following the Y2H‐seq method (Chen et al. 2021), colonies obtained from the screen were subjected to colony PCR. The products were then analysed by Next‐Generation Sequencing (NGS). Sequences were assembled and aligned, and transcripts per million (TPM) values were calculated. Candidate transcription factors were selected based on high TPM values. The full‐length sequences of candidate transcription factors were cloned and verified through one‐to‐one interaction assays.

4.11. Dual‐Luciferase Reporter Assay

We cloned the candidate transcription factor genes into the pGreenII 62‐SK effector vector and the PmPMT2‐1 promoter into the pGreenII 0800‐LUC reporter vector. We co‐infiltrated N. benthamiana leaves with agrobacteria carrying the effector and reporter constructs. The bacterial culture preparation and infiltration procedure were performed as described for the subcellular localization assay. OD₆₀₀ = 1.0. For co‐infiltration, the bacterial suspensions were mixed at a 1:1 ratio prior to injection. After 2–3 days, we used the in vivo imaging system to observed the fluorescence intensity in N. benthamiana leaves. We quantified the relative luciferase activity using a Dual‐Luciferase Reporter Assay Kit (Yeasen) to measure transcriptional activation according to the manufacturer's recommendations.

The experiment was performed with six biological replicates and three technical replicates.

4.12. EMSA

To study the interaction between transcription factors and DNA, we inserted the transcription factor gene into the pET‐22b vector with a His tag. The binding element sequence of the core promoter region from PmPMT2‐1 was selected for probe design, and a biotin‐labelled double‐stranded DNA probe corresponding to the wild‐type promoter sequence was synthesized. Simultaneously, an unlabelled competitive probe with the presumed binding site mutation was prepared. The selected probe sequence was 5′‐AACATAGTCAACCACA‐3′, and the corresponding mutant probe sequence was 5′‐AAAAAAAAAAAAAAA‐3′. The binding affinity and specificity of the purified transcription factors were EMSA. The labelled probes were incubated with the recombinant protein, and the complexes were separated by native PAGE. To confirm the binding specificity, competitive experiments were conducted by adding excess unlabelled wild‐type or mutant probes.

Author Contributions

Ziyan Nie: writing – original draft, formal analysis, investigation, validation, visualization, writing – review and editing, software, data curation. Shan Hu: resources. Bin Liu: supervision, investigation. Qinghua Liu: funding acquisition, supervision. Hengfu Yin: conceptualization, methodology, software, writing – review and editing, resources. Zhichun Zhou: funding acquisition, supervision. Kai Gao: writing – original draft, supervision, resources, project administration.

Funding

This work was supported by National Natural Science Foundation of China (32501677), Biological Breeding‐National Science and Technology Major Project (2022ZD04016), National Key Research and Development Program of China (2022YFD2200202), and Anhui Provincial Key Laboratory of Forest Resources and Silviculture (LMZYPY‐03).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

File S1: The detailed recipe of the culture medium used in the article.

File S2: The WGCNA scripts used in this study.

Figure S1: The vector schematics of pCambia1300‐GFP‐PmPMT2‐1 and pCAMBIA2300‐RUBYRed‐PmPMT2‐1.

Figure S2: Multiple amino acid sequence alignment of all PMT family members in Pinus massoniana .

Figure S3: Identification of positive transgenic calli.

Figure S4: Insert fragment size and library capacity assessment of the yeast library.

Table S1: Genes involved in the pinosylvin monomethyl ether (PME) biosynthetic pathway.

Table S2: The specific data obtained from the LC–MS.

Table S3: Estimated LC₅₀ values and 95% confidence intervals of pinosylvin monomethyl ether (PME) against nematodes after 1, 3 and 5 days of treatment.

Table S4: The promoter cis‐element analysis.

Table S5: The gene‐specific primers.

Table S6: Profiles and extreme levels of eight metabolites in resistant and susceptible clones at 0 and 7 days post‐inoculation (dpi) with pine wood nematode.

Data Availability Statement

The raw transcriptome data are available from the NCBI database under accession number PRJNA1458328.