Hydrophobic nano-calcium superphosphate enhances leaf disease resistance via coordinated physical, biochemical, and phyllosphere responses

Abstract

Hydrophobic nano-calcium superphosphate (NPS1) is a novel nano-fertilizer that has shown great potential for enhancing plant growth and suppressing foliar diseases. However, the mechanisms underlying these effects remain unclear. We investigated the responses of grape leaves and phyllosphere microbiota to NPS1 under both open-field and rain-shelter cultivation systems. NPS1 application significantly reduced leaf wetness duration and the disease severity of downy and powdery mildew, increased leaf-associated phosphorus retention, which remained elevated at later sampling stages, while enhancing vine growth, increasing SOD and POD activities, and inhibits several grape pathogens in vitro. Integrated transcriptomic and metabolomic analyses identified substantial changes in gene expression and metabolite accumulation following NPS1 treatment, characterized by the upregulation of stress-responsive genes (WRKY41, PEROXIDASE, and EREBP) and enrichment of pathways related to phenylpropanoid and secondary metabolite biosynthesis. Exogenous validation experiments indicated that coumaric acid and lignin were associated with enhanced seedling growth and restricted pathogen growth in vitro. Furthermore, amplicon sequencing revealed significant shifts in the composition, diversity, and co-occurrence patterns of phyllosphere microbial communities, accompanied by reduced relative abundances of pathogen-associated taxa (e.g., Erysiphe and oomycetes). Notably, representative beneficial isolates (e.g., Sphingomonas spp. and Bacillus spp.) exhibited stronger pathogen inhibition when combined with NPS1 than when applied alone. Collectively, these findings indicate that NPS1 alters the leaf surface conditions through reduced leaf wetness, associated with coordinated changes in plant physiological responses, metabolite accumulation, and phyllosphere microbial communities, which may be linked to reduced disease severity and improved grapevine performance.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04828-x.

Introduction

Foliar diseases are a major constraint to grapevine productivity, particularly under humid conditions where prolonged leaf wetness facilitates pathogen infection and epidemic development. Studies have indicated that Yunnan Province is characterized by pronounced wet and dry seasons. During the wet season, heavy rainfall frequently leads to severe outbreaks of downy mildew, significantly affecting grape yield and quality [1]. At present, a substantial proportion of grape cultivation in this region (approximately 75%—80%) is conducted under open-field systems, where disease management relies heavily on repeated fungicide applications [2, 3]. In vineyard ecosystems, multiple diseases often occur simultaneously or in succession, requiring repeated fungicide applications to limit pathogen spread. In practice, fungicides are typically applied 12 to 30 times per growing season [4]. Such frequent use increases the risk of resistance development in pathogen populations [5]. Our prior research has demonstrated that rain-shelter cultivation can significantly reduce the incidence of grape downy mildew. This reduction is achieved primarily by shortening leaf wetness duration and, additionally, by facilitating the recruitment of beneficial microorganisms to the grape leaf phyllosphere, which aids the plant in resisting the disease [6, 7]. However, it remains unclear whether hydrophobic phosphate-based nanomaterials can reduce leaf wetness duration, suppress disease incidence, and enhance plant defense responses in grapevine.

Crop health is a fundamental prerequisite for sustainable agricultural production and high yields. In South China, however, crop quality is simultaneously threatened by abiotic stress from pesticide contamination and by biotic stress from phyllosphere pathogens. Although synthetic agrochemicals are routinely applied to safeguard plant health, their indiscriminate use is frequently associated with environmental and human-health hazards, including pesticide residues, diminished efficacy, resistance development, and direct toxicological risks [8, 9]. Consequently, sustainable intensification of modern agriculture requires alternative strategies; among these, nanotechnology-based approaches for mitigating abiotic stress and alleviating oxidative damage in plants have recently received considerable attention [10, 11]. Among these emerging alternatives, nanotechnology-based materials have attracted increasing interest because of their potential to improve crop performance while reducing reliance on conventional agrochemicals.

Hydrophobic nanomaterials have attracted increasing attention due to their small size and surface properties, which can modify plant–environment interactions [12]. Owing to their small size, large surface area, stability, and biocompatibility, nanotechnology-based fertilizers have been introduced into agricultural practice [13]. Recent studies have shown that nanomaterials, in comparison with conventional fertilizers, can significantly promote plant growth and increase crop yields in both greenhouse and field conditions [14–17]. Specifically, nanomaterials improve the efficiency of nutrient acquisition in plants and accelerate plant growth rates [18]. They also stimulate seed germination by enhancing water uptake and seed coat permeability, which in turn activates cellular metabolism and elevates amylase activity [19]. Notably, certain nanomaterials feature a lotus-leaf-like hierarchical structure and exhibit hydrophobic characteristics, which reduce the contact interface between pathogens and plant leaf surfaces, thus restricting pathogen adhesion and colonization [20]. Meanwhile, localized stress induced by the nanomaterial layer can suppress pathogen cell wall disruption and biofilm formation [21]. Collectively, these properties endow nanomaterials with considerable potential in alleviating various environmental stresses encountered by plants [22].

Pathogen invasion can cause plant tissue damage, suppress photosynthetic and respiratory functions, increase the generation of reactive oxygen species (ROS), damage plant cellular structures, and ultimately lead to the suppression of plant growth [23]. Foliar pathogen infection commonly constrains plant growth, disrupts physiological processes, and weakens host defense capacity, ultimately leading to substantial losses in yield and quality [23, 24]. In response to such stress, considerable attention has been directed toward strategies that not only suppress pathogen proliferation but also maintain host physiological and metabolic stability. Beyond their direct antimicrobial properties, emerging evidence suggests that nanomaterials may be associated with changes in antioxidant systems, defense-related gene expression, and secondary metabolite accumulation, which could contribute to improved plant performance under biotic stress [25–28]. These coordinated responses contribute to enhanced plant resilience under biotic stress conditions. Emerging evidence further suggests that nanomaterials may interact synergistically with plant-beneficial microorganisms, thereby improving plant performance under stress. For example, the combined application of nanomaterials and beneficial microbes has been reported to enhance photosynthate allocation in shoots and promote plant growth and development [29]. However, evidence regarding the coordinated roles of nanomaterials in disease suppression, particularly their effects on host physiological responses and plant-associated microbial communities, remains limited. Among various nanomaterials, phosphate-based nanoparticles are of particular interest due to their potential to simultaneously influence plant nutrition, redox homeostasis, and plant–microbe interactions. Nevertheless, it remains unclear whether exogenously applied superphosphate nanoparticles can alter antioxidant defenses, defense-related gene expression, metabolite accumulation, and associated microbial communities under pathogen stress, and whether these changes are associated with enhanced disease resistance. Accordingly, a mechanistic understanding of how nano-superphosphate influences host physiological processes and phyllosphere microbial communities during plant–pathogen interactions is still lacking. Integrating physiological measurements with transcriptomic, metabolomic, and microbiome analyses may provide critical insights into the coordinated processes potentially associated with its effects on plant disease suppression.

This study investigates how the nanomaterial NPS1 may contribute to the suppression of grapevine leaf diseases through coordinated changes in leaf surface properties, host defense responses, and associated microbial processes. Specifically, NPS1 is expected to enhance leaf hydrophobicity, thereby reducing leaf wetness duration and potentially limiting pathogen infection. In addition, it may be associated with the changes of antioxidant and defense- related pathways, accompanied by shifts in transcriptomic and metabolic profiles. Furthermore, NPS1 may influence the composition and interactions of phyllosphere microbial communities under different cultivation conditions. Accordingly, we tested the following hypotheses: (1) NPS1, owing to its nanoscale properties, increases leaf hydrophobicity and reduces leaf wetness duration, thereby creating less favorable conditions for pathogen survival and infection; (2) NPS1 reduces the severity of grape downy mildew and powdery mildew while maintaining plant growth; (3) NPS1 treatment is associated with changes in defense-related gene expression and metabolite accumulation linked to antioxidant and defense responses; and (4) NPS1 treatment is associated with shifts in phyllosphere microbial community composition and microbial interactions under different cultivation conditions. By integrating physiological, transcriptomic, metabolomic, and microbiome analyses, this study aimed to provide a comprehensive evaluation of the processes potentially associated with NPS1-mediated disease suppression in grapevine.

Materials and methods

Site description, materials, and experimental design

The experiment was conducted at the Yunnan Agricultural University Modern Scientific Research Base (25°57′ N, 103°25′ E) in Xundian Yi and Hui Autonomous County, Kunming City, Yunnan Province, China. The site is located at an altitude of approximately 1900 m and experiences a subtropical monsoon climate with a mean annual temperature of 16.2 °C and mean annual precipitation of 1018.8 mm. The soil is sandy loam with a pH of 7.62 and contains 25.7 g kg⁻¹ organic matter, 26.5 mg kg⁻¹ available phosphorus, 282 mg kg⁻¹ available potassium, an electrical conductivity of 240 µS cm⁻¹, and a moisture content of 34.6%. The experiment included both open-field cultivation and rain-shelter cultivation systems. Each individual greenhouse measured 3.0 m (ridge height) × 1.5 m (eave height) × 4.5 m (width) × 12.0 m (length). Seven-year-old grapevines, planted in double rows with a spacing of 2.0 m × 1.0 m (row × plant), were used.

Synthesis of Nano-Hydrophobic Superphosphate (NPS1): NPS1 was prepared by dispersing 10 g of single superphosphate (source: Kunming Renmao Agricultural Materials) in 100 mL of ultrapure water. Then, 0.1 g of stearic acid, 10 µL of 6 M hydrochloric acid, and 10 µL of hexadecyltrimethoxysilane (as a coupling agent) were added to the mixture. The reaction solution was magnetically stirred at 400 rpm and 25 °C for 2 h. Subsequently, the product was vacuum-filtered through a 0.22 µm membrane and freeze-dried to obtain the final NPS1 powder. Hydrophobicity Validation and Material Characterization: The hydrophobic property of NPS1 was verified by uniformly spraying the powder onto clean glass slides, allowing them to dry at 25 °C, and then determining the water contact angle using the sessile drop method. General material characterization was also performed.

Leaf Hydrophobicity Assay: Sterile water served as the control (CK), while NPS1 was applied at low (125 mg L⁻¹), medium (250 mg L⁻¹), and high (500 mg L⁻¹) concentrations. Healthy grape leaves of uniform age were selected, and surface impurities were removed by rinsing three times with 0.1% Tween-20 solution, followed by rinsing with deionized water and air-drying. A micro-spray system was used to simulate rainfall (intensity: 60 mm h⁻¹, duration: 10 min). After spraying, the leaves were placed in a controlled environment at 25 °C and 45% relative humidity to dry naturally, and the time required for complete drying was recorded. Each treatment included six replicates. To simulate natural leaf orientations, the leaves were tested at three inclination angles: 0°, 45°, and 90° (n = 6).

Material characterization of NPS1

The physicochemical properties of NPS1 were characterized using X-ray diffraction (XRD), Infrared Spectroscopy (IR), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), scanning electron microscopy (SEM), and water contact angle (CA) measurements, and contact angle (CA)—were conducted at the Key Laboratory of Bionic Engineering (Ministry of Education), the Electron Microscopy Center, and the State Key Laboratory of Supramolecular Structure and Materials, all at Jilin University. Sample preparation for all analyses followed standard protocols for the respective instruments [30].

Plant physiological analysis

Plant physiological analysis. Physiological analyses were conducted using four-month-old grape plants. Leaf samples were collected 24 h after NPS1 treatment for SEM and energy-dispersive X-ray spectroscopy (EDS) analyses, with three biological replicates per treatment group. Quantitative phosphorus (P) analysis was conducted at the Testing Center of Kunming Geological Exploration Institute, China Metallurgical Geology Bureau. Measurements were performed at 1, 6, and 14 d after NPS1 treatment using six sample categories: CK leaves, NPS1-treated leaves, deionized water-washed CK leaves, deionized water-washed NPS1-treated leaves, CK leaf wash eluates, and NPS1 leaf wash eluates (n = 3). For leaf washing, grape leaves were placed into sterile 50 mL centrifuge tubes containing 45 mL ultrapure water and vortexed for 20 s. The resulting wash solution was retained as the leaf wash eluate. Leaves were subsequently rinsed with deionized water for 10 min, oven-dried, ground into powder, and used for phosphorus determination. Phosphorus concentrations were determined using inductively coupled plasma optical emission spectrometry (ICP-OES; ICAP 7400, Thermo Fisher Scientific, USA) according to the National Food Safety Standard GB 5009.268–2025. Plant height, chlorophyll content, and leaf size were measured 20 d after NPS1 treatment using six independent grape seedlings (n = 6) per treatment. Chlorophyll content was measured on the third fully expanded leaf from the shoot apex of each seedling, resulting in six biological replicates per treatment (n = 6). Enzyme activities and malondialdehyde (MDA) contents were determined 24 h after NPS1 treatment using commercial assay kits (ADS-W-KY001, ADS-W-KY002, ADS-W-KY003, and ADS-W-YH002; Jiangsu ADS Biotechnology Co., Ltd., China) according to the manufacturer’s instructions (n = 3).

Disease assessment, pathogen inhibition, and microbial co-culture assays

Under open-field cultivation conditions, grape downy mildew disease index was investigated, while grape powdery mildew under rain-shelter cultivation was evaluated according to the corresponding standard method [7]. Six grapevines were used as biological replicates for each treatment, and five leaves per vine were assessed for disease severity. Disease severity was classified into six levels according to the percentage of diseased leaf area: Level 0, no visible symptoms; Level 1, lesion area accounting for 1–5% of total leaf area; Level 3, 6–25%; Level 5, 26–50%; Level 7, 51–75%; and Level 9, > 75% of total leaf area. The disease index was calculated using the following formula:

For pathogen inhibition and potential beneficial microbial co-culture assays, all microbial strains (Botrytis cinerea, Colletotrichum spp, Phytophthora spp, Sphingomonas spp, Bacillus subtilis.) were provided by the Laboratory of Crop Disease Control and Biodiversity, Yunnan Agricultural University. Potential beneficial microorganisms were selected based on core phyllosphere taxa identified in the microbiome analysis and subsequently isolated from grape leaf surfaces. For inhibition assays, a NPS1 stock suspension (1 g per 100 mL sterile water) was mixed with molten potato dextrose agar (PDA) prior to plate pouring to obtain final NPS1 concentrations of 125, 250, and 500 mg L⁻¹. Approximately 15 mL of medium was dispensed into each Petri dish. PDA plates without NPS1 served as the control (CK). Pathogen spore suspensions were adjusted to 1 × 10⁶ spores mL⁻¹, and a 10 μL aliquot was inoculated onto the center of each PDA plate after medium solidification. Plates were incubated at 25 °C, and colony diameters were measured after 4 d. Each treatment consisted of three independent replicate plates.

To evaluate the potential interactions between NPS1 and beneficial microorganisms in suppressing pathogen growth, separate co-culture assays were conducted against Colletotrichum spp. and Phytophthora spp. The NPS1 concentrations used in these assays (0.25, 0.60, and 1.00 mg L⁻¹) were selected according to previously reported minimum concentrations affecting pathogen sporulation [44], rather than the field application concentrations used in the pathogen inhibition assays. For each pathogen, seven treatments were established: CK (pathogen only), 0.25 mg L⁻¹ NPS1 + Bacillus subtilis, 0.60 mg L⁻¹ NPS1 + Bacillus subtilis, 1.00 mg L⁻¹ NPS1 + Bacillus subtilis, 0.25 mg L⁻¹ NPS1 + Sphingomonas spp., 0.60 mg L⁻¹ NPS1 + Sphingomonas spp., and 1.00 mg L⁻¹ NPS1 + Sphingomonas spp. Pathogen spore suspensions were inoculated at the center of PDA plates, while bacterial suspensions (1 × 10⁶ CFU mL⁻¹) were inoculated at four equidistant positions surrounding the pathogen colony. Plates were incubated at 25 °C, and colony diameters were measured at 0, 3, 6, 9, and 12 d. Inhibition rates were calculated as described below.

The inhibition rate was calculated according to the following equation:

where R represents the average colony radius [7].

Sample collection and pretreatment

For transcriptomic, metabolomic, and microbiome analyses, each biological replicate consisted of three healthy functional leaves collected from an individual grapevine and pooled into one composite sample. Four independent grapevines were sampled per treatment for transcriptome sequencing (n = 4), whereas six independent grapevines were sampled per treatment for metabolomic and microbiome analyses (n = 6). At 6 h after the final NPS1 treatment, functional leaves (south-facing, approximately 1.5 m above ground) were collected from uniformly growing grapevines under different cultivation conditions. This sampling time was selected to capture early molecular and microbial responses following NPS1 treatment. Damaged or malformed leaves were excluded. For transcriptomic and metabolomic analyses, samples were collected from CK and NPS1-treated grapevines, with four biological replicates used for transcriptome sequencing and six biological replicates used for metabolomic analysis per treatment group. For RT-qPCR validation, CK and NPS1 treatment groups each included 12 biological replicates. For phyllosphere microbiome analysis, four treatment groups were included: open-field CK, open-field NPS1, rain-shelter CK, and rain-shelter NPS1, with six biological replicates per treatment group. Approximately 2 g of leaf tissue per replicate was immediately frozen in liquid nitrogen and stored at − 80 °C prior to subsequent analyses.

Transcriptome library construction, sequencing, and bioinformatics analysis

Total RNA was extracted from 100 mg of grape leaf tissue using the MJZol total RNA extraction kit (Majorbio Yuhua Bio-Pharm Technology Co., Ltd., Shanghai, China) according to the manufacturer’s instructions. RNA concentration and purity were assessed using a NanoDrop spectrophotometer (Thermo Scientific, USA). RNA integrity was evaluated by agarose gel electrophoresis and an Agilent 2100 Bioanalyzer (Agilent Technologies, USA), and samples with an RNA integrity number (RIN) > 6.5 were used for subsequent analyses.​ Transcriptome sequencing was conducted using four biological replicates for each treatment group (CK and NPS1). RNA purification, library preparation, and sequencing were performed at Shanghai Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China).​ Sequencing libraries were prepared using the Illumina® Stranded mRNA Prep Kit (Illumina, San Diego, CA, USA) with 1 μg of total RNA. Poly(A) + mRNA was enriched using oligo(dT) magnetic beads and fragmented into approximately 300 bp fragments. First- and second-strand cDNA synthesis was subsequently performed, during which dUTP was incorporated into the second strand to preserve strand specificity. The resulting double-stranded cDNA underwent end repair, adenylation, adapter ligation, fragment size selection, and PCR amplification to generate the final libraries.​ Library concentration and quality were assessed using a Qubit 4.0 fluorometer (Thermo Scientific, USA). Paired-end sequencing (PE150) was performed on the Illumina NovaSeq X Plus platform.​ Raw sequencing reads were processed using Fastp to remove adapter sequences, low-quality reads, reads shorter than 20 bp, and reads containing ambiguous bases. Clean reads were aligned to the Vitis vinifera reference genome (12X assembly) using HISAT2. Transcript assembly was conducted using StringTie in a reference-guided manner.​ Gene expression abundance was estimated as transcripts per million (TPM) values using RSEM. Differential expression analysis was performed in DESeq2 using raw read counts. Genes with a false discovery rate (FDR) < 0.05 and |log₂ (fold change) |≥ 1 were considered significantly differentially expressed genes (DEGs).​ For heatmap visualization, expression values of selected DEGs were subjected to row-wise Z-score standardization according to the formula Z = (x − μ)/σ, where x represents the expression value, μ is the mean expression value of a given gene across all samples, and σ is the corresponding standard deviation. This transformation was applied to facilitate comparison of relative expression patterns among samples and genes. Functional enrichment analyses of DEGs, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses, were performed using Goatools and Python SciPy. Enrichment significance was determined using Bonferroni-corrected p-values < 0.05.

Reverse transcription quantitative PCR (RT-qPCR) analysis

Reverse transcription quantitative PCR (RT-qPCR) was performed to validate the expression patterns of selected genes in grape leaves under CK and NPS1 treatments. A total of 12 biological replicates were used for each treatment group. All samples were collected from the same field experiment described in Sect. "Sample collection and pretreatment" but were independent of those used for transcriptome sequencing. First-strand cDNA used for RT-qPCR analysis was independently synthesized from total RNA using the HiScript II Reverse Transcriptase Kit (Vazyme, China) according to the manufacturer’s instructions. The resulting cDNA was diluted five-fold prior to RT-qPCR analysis.

Primer sequences for the seven target genes and the reference gene VvActin are listed in Table S1. Each 20 μL reaction mixture contained 10 μL of 2 × ChamQ SYBR Color qPCR Master Mix, 0.4 μL each of forward and reverse primers (5 μM), 0.2 μL of 50 × ROX Reference Dye 1, 1 μL of diluted cDNA template, and nuclease-free water to a final volume of 20 μL. RT-qPCR amplification was conducted on an ABI 7500 Real-Time PCR System (Applied Biosystems, USA) using the following thermal cycling conditions: initial denaturation at 95 °C for 3 min, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. Fluorescence signals were collected during the annealing/extension step. Melting curve analysis was subsequently performed to verify amplification specificity. Relative gene expression levels were calculated using the 2^ − ΔΔCt method, with VvActin serving as the endogenous reference gene and the CK group used as the calibrator. Statistical differences between CK and NPS1 treatments were evaluated using Student’s t-test. To control the false discovery rate (FDR) associated with multiple comparisons among the seven target genes, p-values were adjusted using the Benjamini–Hochberg method, and adjusted p-values < 0.05 were considered statistically significant.

Metabolomic profiling and data processing

Leaf samples (100 mg) were homogenized in 2 mL microcentrifuge tubes containing a 6 mm grinding bead and 800 μL of extraction solvent (methanol:water, 4:1, v/v) supplemented with 0.02 mg/mL L-2-chlorophenylalanine as the internal standard. Samples were homogenized using a frozen tissue grinder for 6 min at − 10 °C and 50 Hz, followed by ultrasonic extraction at 5 °C and 40 kHz for 30 min. After incubation at − 20 °C for 30 min, the extracts were centrifuged at 13,000 × g for 15 min at 4 °C. An equal volume of supernatant from each sample was pooled to generate quality control (QC) samples. QC samples were analyzed periodically throughout the analytical sequence at an interval of one QC injection per 10 experimental samples to monitor instrument stability and analytical reproducibility. The remaining supernatants were transferred to LC–MS autosampler vials for metabolomic analysis. Chromatographic separation was performed using an ultra-high-performance liquid chromatography system coupled with Q Exactive Orbitrap Fourier transform mass spectrometry (UHPLC-Q Exactive Orbitrap MS; Thermo Fisher Scientific, USA). Metabolites were separated on a BEH C18 column (100 mm × 2.1 mm i.d., 1.7 μm; Waters, USA) with an injection volume of 3 μL. Mobile phase A consisted of water containing 0.1% formic acid and 2% acetonitrile, while mobile phase B consisted of acetonitrile containing 0.1% formic acid. The gradient elution program was as follows: 0–0.5 min, 2% B; 0.5–7.5 min, 2–35% B; 7.5–13 min, 35–95% B; 13–14.4 min, 95% B; 14.4–14.5 min, 95–2% B; and 14.5–16 min, 2% B. The flow rate was maintained at 0.40 mL/min, and the column temperature was set at 40 °C. Mass spectrometric detection was performed in both positive and negative electrospray ionization (ESI) modes with a scan range of m/z 70–1050. Spray voltages were set at 3500 V in positive mode and − 3000 V in negative mode. Sheath gas and auxiliary gas pressures were set at 50 psi and 13 psi, respectively, and the ion source temperature was maintained at 450 °C. Data-dependent MS/MS acquisition was performed using stepped normalized collision energies (NCE) of 20, 40, and 60. The resolutions for MS1 and MS2 scans were set to 70,000 and 17,500, respectively. Raw LC–MS data were processed using Progenesis QI software (Nonlinear Dynamics, Waters Corporation, Milford, MA, USA) for baseline correction, peak detection, retention time alignment, peak integration, and data normalization to generate the metabolite feature matrix. Metabolite annotation was conducted based on accurate precursor masses (MS1) and characteristic fragmentation spectra (MS2) through comparison with the METLIN database, Human Metabolome Database (HMDB), and an in-house spectral library. According to the Metabolomics Standards Initiative (MSI) guidelines, all annotated metabolites in this study were classified as Level 2 (putatively annotated compounds). Detailed metabolite identification parameters, including mass errors and fragmentation matching scores, are provided in Tables S3 and S4. The resulting data matrix was uploaded to the Majorbio Cloud Platform (Cloud.majorbio.com) for preprocessing. Metabolite features with more than 20% missing values were removed according to the 80% rule. Remaining missing values were imputed using the minimum value of the original data matrix. Peak intensities were normalized by total ion intensity, and variables with relative standard deviation (RSD) values > 30% in QC samples were excluded. The processed data were subsequently log10-transformed prior to multivariate statistical analyses. Principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) were performed using the ropls package (version 1.6.2) in R. Model robustness was evaluated using sevenfold cross-validation. Differential metabolites were identified based on variable importance in projection (VIP) values > 1 and Student’s t-test p-values < 0.05. Metabolic pathway annotation was conducted using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Pathway enrichment analysis was performed using Fisher’s exact test implemented in the Python scipy.stats package, and p-values < 0.05 were considered statistically significant.

To evaluate the potential growth-promoting effects of candidate metabolites identified through metabolomic analysis, an exogenous application experiment was conducted using four-month-old grape seedlings. Coumaric acid and lignin were selected as representative metabolites based on the metabolomic profiling results. For each metabolite, four concentration levels were established: 0 mg L⁻¹ (CK), 0.125 mg L⁻¹ (L), 0.250 mg L⁻¹ (M), and 0.500 mg L⁻¹ (H). Metabolite solutions were applied by foliar spraying, with 200 mL administered to each seedling. Twenty days after application, plant height, relative chlorophyll content (SPAD value), and leaf size (transverse and longitudinal diameters) were determined. Each concentration level consisted of six biological replicates (n = 6), with one seedling representing one biological replicate. For leaf-related measurements, three fully expanded functional leaves were selected from each plant, and the average value was used for subsequent statistical analysis. Data were analyzed using one-way analysis of variance (ANOVA). When significant treatment effects were detected (p < 0.05), means were separated using Tukey’s honestly significant difference (HSD) test for multiple comparisons.

Microbiome sequencing and analysis

For microbiome analysis, six biological replicates were used for each treatment group. Phyllosphere microbial fractions were isolated from grape leaves following standardized washing and enrichment procedures. Briefly, leaves were washed in sterile 0.1 M potassium phosphate buffer (pH 8.0) at a ratio of 10 mL buffer per gram of leaf tissue, followed by 40 kHz ultrasonication for 15 min, vortex mixing, and repeated rinsing to detach epiphytic microorganisms. The wash solutions were pooled and centrifuged at 13,000 rpm for 10 min at 4 °C to collect microbial pellets. To characterize the potentially transcriptionally active phyllosphere microbial community rather than the total microbial community, total RNA was extracted from the microbial pellets obtained from leaf surface wash solutions using the MJZol total RNA extraction kit. RNA integrity was verified by 1% agarose gel electrophoresis, while RNA concentration and purity (OD₂₆₀/₂₈₀) were assessed using a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). Complementary DNA (cDNA) was synthesized using the HiScript II Reverse Transcriptase Kit (Vazyme, China) and quantified with a Qubit fluorometer (Life Technologies, USA). The resulting cDNA served as the template for all subsequent amplicon sequencing analyses. The fungal ITS1 region (ITS1F–ITS2R), bacterial 16S rRNA V3–V4 hypervariable region (338F–806R), and oomycete ITS region (ITS1-3F/ITS1-4R) were amplified using cDNA templates on an ABI GeneAmp® 9700 PCR system. Each 20 μL PCR reaction contained 4 μL of 5 × Fast Pfu buffer, 2 μL of dNTPs (2.5 mM), 0.8 μL each of forward and reverse primers (5 μM), 0.4 μL of Fast Pfu polymerase, 10 ng of cDNA template, and nuclease-free water to a final volume of 20 μL. PCR amplification conditions were as follows: initial denaturation at 95 °C for 3 min; 35 cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s; followed by a final extension at 72 °C for 10 min and holding at 4 °C. PCR products were separated on 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen, USA). Purified amplicons were quantified using a Quantus™ Fluorometer (Promega, USA) and pooled in equimolar concentrations. Indexed sequencing libraries were subsequently constructed using the NEXTFLEX Rapid DNA-Seq Kit, including adapter ligation, magnetic bead purification to remove adapter dimers, PCR enrichment, and final library purification and quality control. Illumina adapter and index sequences were incorporated during library preparation to enable multiplex sequencing. The final libraries were sequenced on the Illumina MiSeq platform using the PE300 strategy (Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China). Because the sequencing strategy was based on RNA-derived cDNA rather than total environmental DNA, the resulting profiles preferentially reflected potentially transcriptionally active microbial fractions associated with the grape phyllosphere. Raw sequencing reads were quality-filtered using Fastp v0.19.6 to remove low-quality sequences (50 bp sliding window; minimum Phred score of 20), reads shorter than 50 bp, and reads containing ambiguous bases (N). Paired-end reads were merged using FLASH v1.2.11 with a minimum overlap length of 10 bp and a maximum mismatch ratio of 0.2. Sample demultiplexing was performed allowing zero barcode mismatches and a maximum of two primer mismatches. Quality-filtered sequences were processed using the QIIME2 pipeline with the DADA2 plugin to generate amplicon sequence variants (ASVs). Chloroplast and mitochondrial sequences were removed prior to downstream analyses. To minimize the influence of sequencing depth on alpha and beta diversity analyses, all samples were rarefied to 20,000 reads per sample, resulting in an average Good’s coverage of 99.09%. Taxonomic classification was performed using the Naïve Bayes classifier in QIIME2 against the Silva 138 database for bacterial 16S rRNA sequences and the UNITE 8.0 database for fungal and oomycete ITS sequences, with a confidence threshold of 0.7.

Statistical analysis

The numbers of biological replicates used for each experiment are indicated in the corresponding method sections and figure legends. Unless otherwise stated, all data are presented as mean ± standard deviation (SD). Statistical analyses for physiological parameters, disease incidence, disease index, pathogen inhibition assays were performed using GraphPad Prism 8.0. Differences between two groups were evaluated using Student’s t-test. Comparisons among three or more groups were analyzed using one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) test for multiple comparisons. Differences were considered statistically significant at p < 0.05. Sequence clustering of isolated bacterial strains and maximum likelihood phylogenetic tree construction were performed using MEGA 11.0. For microbiome analyses, alpha diversity indices between groups were compared using Student’s t-test. Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distances was used to visualize differences in microbial community structure, and analysis of similarities (ANOSIM) with 999 permutations was applied to assess statistical significance among treatment groups. Differential abundance analysis at the phylum level was conducted using the Wilcoxon rank-sum test with false discovery rate (FDR) correction. Differential taxa at the genus level were identified using linear discriminant analysis effect size (LEfSe), with a linear discriminant analysis (LDA) score threshold of 3.0. Co-occurrence network analysis was conducted in R 4.3.3 using the psych, Hmisc, and igraph packages. Network visualization and topological analyses were further performed using the Hiplot platform, Gephi, Lianchuan Bio Cloud Platform, and Adobe Illustrator 26.2.1. For metabolomic analyses, differential metabolites were identified based on VIP values > 1 and Student’s t-test p-values < 0.05. Metabolic pathway enrichment analysis was performed using Fisher’s exact test, and pathways with p < 0.05 were considered significantly enriched. Unless otherwise specified, bioinformatic analyses related to transcriptomic, metabolomic, and microbiome sequencing data were performed on the Majorbio Cloud Platform (Shanghai Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China). Sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) database under accession numbers PRJNA1368410, PRJNA1368411, PRJNA1368436, and PRJNA1369163.

Results

Characterization of nano-calcium superphosphate material

As shown in Fig. 1A, NPS1 exhibited a porous and agglomerated morphology characterized by a rough surface texture and well-defined particle boundaries. X-ray photoelectron spectroscopy (XPS) confirmed the presence of C, P, O, and Ca on the surface of NPS1 (Fig. 1B). Infrared spectroscopy demonstrated that the addition of ddH₂O had no significant effect on chemical structure of NPS1 (Fig. 1C). X-ray diffraction (XRD) indicated the crystallinity of the Ca(H₂PO₄)₂ phase in the material (Fig. 1D). Characteristic elemental signals corresponding to the major constituents of NPS1 were further observed in the spectra shown in Fig. 1E–G. Thermogravimetric analysis (TGA) performed under N₂ atmosphere at a heating rate of 10 °C min⁻¹ showed a total weight loss of 9.133%, attributed mainly to adsorbed water. The initial weight-loss event occurred at approximately 137.41 °C, resulting in a mass loss of 0.982%, which is consistent with the removal of physically adsorbed surface water. The second weight-loss stage, characterized by a DTG peak at 604.13 °C and a maximum weight-loss rate of 0.536% min⁻¹, is consistent with the thermal decomposition of the chemisorbed layer. These observations suggest that NPS1 maintained relatively high thermal stability under the tested conditions. The average diameter of 500 NPS1 particles was measured to be 15.46 ± 5.15 nm.

Fig. 1.

Characterisation of nano-sized superphosphate (NPS1). (A) SEM micrographs showing the morphology and surface structure of NPS1 at different magnifications. (B) XPS survey spectrum of NPS1. (C) Infrared spectra of NPS1 and NPS1 + dd H₂O. (D) XRD pattern of NPS1. (E–G) High-resolution XPS spectra of C1s, Ca 2p, and P 2p, respectively. (H) Thermogravimetric analysis (TGA) curve of NPS1 obtained under N₂ atmosphere at a heating rate of 10 °C min⁻¹. (I) Particle-size distribution of NPS1 (n = 500 particles)

NPS1 altered leaf surface hydrophobicity and phosphorus retention on grape leaves

SEM observations showed substantial deposition of NPS1 particles on grape leaf surfaces following application. During disease occurrence, pathogen-associated structures on NPS1-treated leaves frequently appeared covered by NPS1 aggregates and exhibited morphological differences compared with those observed on CK leaves (Fig. 2A). EDS analysis detected characteristic Ca and P signals on treated leaves, whereas these elements were much less evident in CK samples, indicating successful retention of NPS1-derived materials on the leaf surface (Fig. 2A). To evaluate the hydrophobic properties of NPS1, coatings were applied to glass plates, cardboard, and gauze. NPS1-coated surfaces exhibited contact angles exceeding 133° on all tested substrates, whereas water spread more readily on untreated surfaces. Surface tension measurements further supported differences in wetting behavior among treatments (Fig. 2B). These observations indicate that NPS1 possesses strong hydrophobic characteristics under the tested conditions. The effects of NPS1 on leaf wetness duration were further evaluated using grape leaves positioned at different inclination angles. Across all tested concentrations (125, 250, and 500 mg L⁻¹), NPS1-treated leaves dried more rapidly than CK leaves. Drying time generally decreased with increasing leaf inclination angle across all treatments. These results suggest that NPS1 application may contribute to reduced leaf surface wetness and altered water retention on grape leaves (Fig. 2C). To investigate phosphorus dynamics following foliar fertilization, phosphorus concentrations in unwashed leaves, deionized-water-washed leaves, and the corresponding wash eluates were determined at 1, 6, and 14 d after application. Relative to the control treatment (CK), NPS1-treated leaves maintained higher phosphorus concentrations throughout the sampling period. Phosphorus was also detected in both washed leaves and the corresponding wash eluates. These results indicate enhanced phosphorus retention following NPS1 application. The amount of water-wash-resistant leaf-associated phosphorus was greater at 1 and 6 d after treatment and decreased by 14 d (Fig. 2D).

Fig. 2.

NPS1 enhances leaf-surface hydrophobicity and is associated with increased phosphorus retention on grape leaves, Data are presented as mean ± SD. (A) SEM images of grape leaves infected by downy mildew and powdery mildew under CK and NPS1 treatments, together with EDS spectra showing elemental composition of leaf surfaces. (B) Water contact angles and surface tension on NPS1-coated glass, cardboard, and gauze substrates (n = 6). (C) Drying time of water droplets on grape leaves treated with different NPS1 concentrations or measured at different leaf inclination angles (n = 6). Different lowercase letters indicate significant differences among treatments determined by one-way ANOVA followed by Tukey’s HSD test (p < 0.05). (D) Total P, washed leaf P, and P concentration in wash eluates following NPS1 application at different sampling times (n = 3). Comparisons between CK and NPS1 treatments were performed using Student’s t-test, and different lowercase letters indicate significant differences at p < 0.05

NPS1 treatment was associated with reduced disease severity, altered physiological traits, and inhibition of pathogen growth

Field experiments were conducted to evaluate whether NPS1 treatment was associated with changes in disease development and plant physiological performance. Under open-field cultivation, the disease index of grape downy mildew remained lower in NPS1-treated plants than in CK plants across multiple assessment periods. Similar trends were observed for grape powdery mildew under rain-shelter cultivation, where disease severity was consistently lower in NPS1-treated plants throughout the observation period (Fig. 3A).​ Previous studies have found that NPS1 has good efficacy in controlling downy mildew and reducing the severity of downy mildew (Fig. S1). In the pot experiment, grape growth parameters were assessed 20 d after NPS1 application. Compared with CK, NPS1-treated plants exhibited higher chlorophyll content (SPAD value), greater plant height, and larger leaf transverse and longitudinal diameters. The medium concentration (250 mg L⁻¹) was associated with the largest values for most growth-related traits among the tested treatments (Fig. 3B). The effects of NPS1 on antioxidant-related physiological parameters were further evaluated. Compared with the CK treatment, SOD and POD activities were significantly higher in NPS1-treated leaves (p < 0.05) (Table S2). CAT activity was slightly higher, whereas MDA content was slightly lower in NPS1-treated leaves; however, these differences were not statistically significant (Fig. 3C). Following NPS1 application, physiological parameters in tomato showed trends similar to those observed in grapevine, with increases in growth-related traits and consistent changes in SOD and POD activities after NPS1 treatment (Fig. S2 A, B). The direct effects of NPS1 on pathogen growth were evaluated using in vitro culture assays. All tested concentrations inhibited the growth of Botrytis cinerea, Colletotrichum spp, and Phytophthora spp. relative to the control treatment (Fig. 3D). Correspondingly, measurable inhibition rates were observed for all three pathogens, with inhibition rates varying among NPS1 concentrations (Fig. 3E).

Fig. 3.

NPS1 suppresses grape diseases, promotes plant growth, and enhances defense-related responses. (A) Representative symptoms of grape downy mildew and powdery mildew under CK and NPS1 treatments in the field. Bar charts show disease index values at five sampling time points (mean ± SD, n = 6). (B) Plant height, chlorophyll content (SPAD), and leaf transverse and longitudinal diameters of 4-month-old grape seedlings measured 20 d after NPS1 application (mean ± SD, n = 6). (C) Activities of superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT), and malondialdehyde (MDA) content in grape leaves following NPS1 treatment (mean ± SD, n = 3). (D) Colony radius of Botrytis cinerea, Colletotrichum spp., and Phytophthora spp. grown on PDA medium supplemented with different concentrations of NPS1(n = 3). (E) Inhibition rates of the three pathogens under different NPS1 concentrations (n = 3). Different lowercase letters indicate significant differences among treatments determined by one-way ANOVA followed by Tukey’s HSD test (p < 0.05). Comparisons between CK and NPS1 in panels A and C were performed using Student’s t-test (p < 0.05)

NPS1 influences the expression of stress resistance and antioxidant-related genes in grape leaves

We further investigated the transcriptional responses associated with NPS1 treatment in grape leaves using transcriptome sequencing and compared the results with the CK group. Volcano plot analysis identified 30,661 co-annotated genes, among which 197 were significantly upregulated and 385 were downregulated in the NPS1-treated group based on the criteria of FDR < 0.05 and |log₂FC|≥ 1. Figure 4A shows seven selected differentially expressed genes associated with stress-responsive and defense-related functions. Principal component analysis (PCA) revealed distinct transcriptomic profiles between the CK and NPS1-treated groups, indicating that NPS1 treatment was associated with broad transcriptional changes in grape leaves (Fig. 4B). Several candidate genes associated with stress-responsive and antioxidant-related functions exhibited relatively higher expression levels in the NPS1-treated group (Fig. 4C). Gene Ontology (GO) annotation analysis showed that the major functional categories included catalytic activity, binding, membrane part, cell part, cellular processes, and metabolic processes (Fig. 4D). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotation identified 10 enriched pathways, with biosynthesis of other secondary metabolites, environmental adaptation, signal transduction, and folding, sorting, and degradation being the most prominent (Fig. 4E). Further GO enrichment analysis of the top 20 significant terms (Bonferroni-corrected p < 0.05) showed enrichment in biological processes such as response to chitin, transcriptional regulation, iron ion binding and localization, and response to oxidative stress, which may be associated with plant stress adaptation and defense-related responses (Fig. 4F). Similarly, KEGG enrichment analysis identified 19 significantly enriched pathways (Bonferroni-corrected p < 0.05), among which Phenylpropanoid biosynthesis and SNARE interactions in vesicular transport were the most enriched pathways (Fig. 4G). Subsequently, RT-qPCR analysis was performed on selected candidate genes identified from the sequencing dataset. Genes associated with stress-responsive functions (FTH1, EREBP, WRKY41, EREBP105, CPK) and antioxidant-related processes (PEROXIDASE, PHOSPHOLIPASE) generally exhibited relatively higher expression levels in NPS1-treated plants than in the CK group, which was consistent with the expression patterns observed in the transcriptome analysis (Fig. 4H).

Fig. 4.

Transcriptomic responses of grape leaves following NPS1 treatment. (A) Volcano plot of DEGs between NPS1 and CK. (B) PCA of transcriptome profiles. (C) Normalized expression (Z-score) of representative DEGs, identified differentially expressed genes were normalized (Z-score, FDR < 0.05 and |log₂FC|≥ 1). Expression values ​​were converted row-by-row by Z-score before visualization. (D–E) GO and KEGG classification of DEGs. (F–G) GO and KEGG enrichment analyses of DEGs. Bubble size indicates gene number and color indicates enrichment significance. (H) RT-qPCR validation of selected genes. Data are mean ± SD (RNA-seq, n = 4; RT-qPCR, n = 12)

Integrated transcriptomic and metabolomic analyses identify pathways and metabolites potentially associated with stress-related responses in NPS1-treated grape leaves

Integrated transcriptomic and metabolomic analyses were performed to explore metabolic pathways and metabolites potentially associated with NPS1 treatment in grape leaves during disease occurrence. A total of 2,018 metabolites were detected, primarily including lipids, carbohydrates and their derivatives, amino acids and their derivatives, and nucleotides and their derivatives (Fig. S3). Comparative analysis showed that 120 metabolites exhibited higher abundance and 173 metabolites exhibited lower abundance in the NPS1-treated group relative to the CK group (Fig. 5A). Orthogonal partial least squares-discriminant analysis (OPLS-DA) showed a clear separation between the NPS1 and CK groups, suggesting distinct metabolic profiles between treatments (Fig. 5B). Among the differential metabolites, ten metabolites exhibiting relatively large differences between treatments included Soyacerebroside I, P-Menthane-3,8-diol, Momordicin II, Leucocyanidin, Ellagic acid arabinoside, Caftaric acid, Testosterone, Galactosylsphingosine, Isorhamnetin 3-beta-laminaribioside, and Khellin (Fig. 5C). KEGG annotation of differential metabolites indicated that amino acids-related compounds represented a major functional category (Fig. 5D). Combined evaluation based on VIP scores and univariate statistical analysis identified 28 significantly differential metabolites among the top 30 ranked metabolites, whereas 16-OH-oleanolic ketone and Quillajasaponin did not differ significantly from the CK group (Fig. 5E). KEGG pathways enrichment analysis indicated that 16 metabolic pathways were enriched among differential metabolites. The most represented pathways included flavonoid biosynthesis, nucleotide metabolism, biosynthesis of various plant secondary metabolites, phenylalanine metabolism, arachidonic acid metabolism, and ABC transporters (Fig. 5F).

Fig. 5.

Metabolomic responses of grape leaves following NPS1 treatment (n = 6). (A) Differential metabolites identified between NPS1-treated and CK grape leaves. (B) OPLS-DA score plot of metabolite profiles. (C) Representative differentially accumulated metabolites associated with stress- and antioxidant-related pathways (FC ≥ 2, p < 0.05). (D) Classification and abundance of annotated metabolites in NPS1 and CK groups. (E) Differential metabolites ranked by variable importance in projection (VIP); significance levels are indicated as *** p < 0.001, ** p < 0.01, and * p < 0.05. (F) KEGG enrichment analysis of differential metabolites. (G) Shared pathways identified by integrated transcriptomic and metabolomic analyses. (H) Differential metabolites and associated genes involved in the biosynthesis of various plant secondary metabolites and phenylpropanoid biosynthesis pathways. Color scales represent log₂ fold changes

Joint KEGG analysis integrating transcriptomic and metabolomic datasets revealed a potential co-regulation network, with 10 pathways between differentially expressed genes and differential metabolites. Among these, Biosynthesis of various plant secondary metabolites and ABC transporters were enriched in both datasets, whereas Phenylalanine metabolism showed enrichment primarily at the transcriptomic level (Fig. 5G). Further examination of the pathways related to secondary metabolite biosynthesis and phenylalanine metabolism identified six differential metabolites, including (3S)-3-hydroxy-cyclocitral, 2-amino-3-methylsuccinic acid, coumaric acid, coumarin, coniferaldehyde, and artemidin. With the exception of 2-amino-3-methylsuccinic acid, these metabolites accumulated at relatively higher levels in NPS1-treated leaves than in the CK group. Previous studies have reported that (3S)-3-hydroxy-cyclocitral is associated with safranal biosynthesis, whereas 2-amino-3-methylsuccinic acid may participate in methionine-related metabolic processes. Coniferaldehyde has been reported as an intermediate associated with phenylpropanoid metabolism, and artemidin is considered a downstream metabolite related to isoscopoletin metabolism. Two differentially expressed genes, crocetin glucosyltransferase (UGT75L6) and peroxidase (PRDX6), were also detected within these pathways. Compared with the CK group, PRDX6 exhibited higher transcript abundance in NPS1-treated leaves and was annotated to several lignin-related biosynthetic pathways, including those involving coumaryl lignin, 5-hydroxyguaiacyl lignin, syringyl lignin, and p-hydroxyphenyl lignin biosynthesis (Fig. 5H).

Alterations in plant growth phenotypes and in vitro antimicrobial potential of candidate metabolites

To evaluate the potential functional roles of the candidate metabolites, coumaric acid and lignin were exogenously applied to four-month-old grape seedlings at varying concentrations (CK: 0 mg L⁻¹; L: 0.125 mg L⁻¹; M: 0.250 mg L⁻¹; H: 0.500 mg L⁻¹) via foliar spraying. Phenotypic observations 20 days post-application indicated a dose-dependent effect on grape growth, with both coumaric acid and lignin treatments resulting in visibly different foliage phenotypes relative to the CK group (Fig. 6A). Except for the low-concentration lignin treatment (L group), the relative chlorophyll content (SPAD values) across all other treatment groups was significantly higher than that of the CK group (p < 0.05). Compared with the CK group, plant height was significantly increased in all treatment groups (p < 0.05), with the H group reaching the maximum height. Additionally, both leaf transverse and longitudinal diameters were significantly increased in all treatment groups relative to CK (p < 0.05) (Fig. 6B). Similar trends were also observed previously in three-week-old tomato seedlings (Fig. S2 C, D). Subsequently, in vitro antagonism assays were performed against representative plant pathogens using PDA media integrated with the corresponding metabolite concentrations. The results demonstrated that media incorporated with either coumaric acid or lignin significantly restricted the growth radius of Botrytis cinerea and Phytophthora spp species, with the M and H concentrations showing superior inhibitory effects (p < 0.05) (Fig. 6C/D).

Fig. 6.

Growth-promoting and antimicrobial activities associated with coumaric acid and lignin. (A) Representative phenotypes of 4-month-old grape seedlings 20 d after foliar application of coumaric acid or lignin at different concentrations (CK, 0; L, 0.125; M, 0.250; H, 0.500 mg L⁻¹). (B) Effects of coumaric acid and lignin on chlorophyll content (SPAD), plant height, and leaf dimensions. Values represent mean ± SD (n = 6). Different letters indicate significant differences among treatments according to one-way ANOVA followed by Tukey’s HSD test (p < 0.05). F and p values are shown above each panel. (C–D) Effects of coumaric acid (C) and lignin (D) on colony growth of Botrytis cinerea and Phytophthora spp. grown on PDA medium supplemented with different metabolite concentrations. Values represent mean ± SD (n = 3). Different letters indicate significant differences among treatments according to one-way ANOVA followed by Tukey’s HSD test (p < 0.05)

Phyllosphere microbiome diversity and co-occurrence patterns associated with NPS1 treatment

To explore differences in the composition and diversity of grape phyllosphere microbiota between treatments under different cultivation systems, amplicon sequencing of 16S rRNA and ITS genes was performed (Fig. 7A). Principal coordinate analysis (PCoA) based on Bray–Curtis distances revealed distinct profiling of microbiome composition between treatments. Fungal and bacterial community structures differed significantly between the NPS1 and CK groups (p < 0.01) (Fig. 7B). For fungal communities under both cultivation systems, the first two principal coordinates explained 84.69% and 82.94% of the total variance, respectively. Clear separation of different treatment groups was observed along the first principal coordinate (70.94% and 71.16%). For bacterial communities, distinct clustering between CK and NPS1 samples was observed under rain-sheltered cultivation, accounting for 90.18% of the observed variation (Fig. 7B). Microbial diversity index analysis showed that under open-field conditions, The Shannon diversity indices of fungal and bacterial communities were higher in NPS1-treated samples than in CK samples, with a more statistically significant shift observed in bacteria (p < 0.001). Under rain-sheltered conditions, the fungal Shannon diversity index was significantly higher in the NPS1-treated group than in the CK group (p < 0.001) (Fig. 7B). LEfSe analysis at the genus level identified significantly different taxonomic groups between NPS1 and CK under different cultivation conditions. For fungi, Fifteen and ten fungal taxa were identified as biomarkers associated with NPS1 treatment under open-field and rain-sheltered conditions. Notably, the relative abundance of Erysiphe was significantly lower in NPS1-treated samples than in CK samples under both cultivation systems (p < 0.05) (Fig. 7C). For bacteria, Deinococcus and Ralstonia were identified as genera associated with NPS1 treatment (Fig. 7C).

Fig. 7.

NPS1 alters phyllosphere microbial diversity, community composition, and co-occurrence network topology under open-field and rain-shelter cultivation systems. (A) Experimental workflow. (B) PcoA and Shannon diversity analyses of fungal and bacterial communities based on ITS and 16S rRNA sequencing data (n = 6). Bars indicate mean ± SD. Statistical significance was assessed using Student’s t-test (ns, not significant; “***” p < 0.001). (C) Differential taxa identified by Lefse (LDA > 3.0). (D–E) Fungal and bacterial co-occurrence networks and robustness analyses. Nodes represent ASVs, edges represent significant correlations, and colors indicate network modules. N and L denote the numbers of nodes and edges, respectively

To evaluate differences in microbial co-occurrence patterns, co-occurrence networks were constructed. Microbial co-occurrence networks differed in their topological properties between NPS1 and CK treatments, with fewer network edges observed in the NPS1-associated fungal networks under both cultivation systems (open-field: CK—165 nodes/1894 edges, NPS1—223 nodes/1283 edges; rain-sheltered: CK—220 nodes/3235 edges, NPS1—212 nodes/1213 edges). Compared with CK samples, NPS1-treated samples exhibited a higher proportion of negative correlations (open-field: 1.06% versus 23.23%; rain-shelter: 3.15% versus 20.94%) (Table S5). Furthermore, network modularity was higher in the NPS1-treated groups than in the corresponding control groups (Table S5). Fungal co-occurrence networks in the NPS1-treated groups exhibited fewer nodes and edges than those in the CK groups under both cultivation systems, which was accompanied by lower network connectivity and complexity metrics (Fig. 7D). Robustness analysis based on the sequential ASV removal nodes showed higher robustness values in fungal networks associated with NPS1-treated samples than in those associated with CK samples (p < 0.001). A similar pattern was observed in bacterial co-occurrence networks. Under open-field conditions, the bacterial network contained 293 nodes and 3495 edges in the CK group and 115 nodes and 383 edges in the NPS1 group. Under rain-shelter cultivation, the bacterial network contained 261 nodes and 5708 edges in the CK group and 349 nodes and 4609 edges in the NPS1 group (Fig. 7E). In addition, the proportion of negative correlations increased following NPS1 treatment under both cultivation systems (open-field: 8.90% in CK and 23.76% in NPS1; rain-shelter: 2.30% in CK and 6.66% in NPS1) (Table S5). Consistent with the fungal networks, bacterial networks associated with NPS1 treatment exhibited lower connectivity and higher robustness values than the corresponding control networks.

Phyllosphere microbial responses and oomycete community assembly patterns associated with NPS1 treatment

To evaluate the compatibility between NPS1 and indigenous microorganisms phyllosphere microorganisms, random forest analysis was performed using bacterial ASVs obtained from both open-field and rain-shelter cultivation systems. Four Sphingomonas spp ASVs (ASV87, ASV213, ASV868, and ASV13) were identified as frequently detected and highly ranked taxa and were subsequently isolated from the grape phyllosphere (Fig. 8A/B). Co-culture assays involving these phyllosphere-derived Sphingomonas spp isolates and different NPS1 concentrations showed inhibitory activity against Colletotrichum spp. pathogens after 6 d of incubation. In addition, a phyllosphere-derived Bacillus spp isolate exhibited greater inhibition of anthracnose-associated pathogens in the presence of NPS1 at a concentration of 1.0 (Fig. 8G). Parallel assays against Phytophthora showed that the combination of Sphingomonas spp. and NPS1 at a concentration of 0.6 was associated with the largest inhibition zone among the tested treatments, whereas NPS1–Bacillus spp. combinations also displayed stronger inhibitory responses than the corresponding control treatments (Fig. 8H).

Fig. 8.

Associations of NPS1 with core phyllosphere microorganisms, oomycete communities, and pathogen inhibition assays. (A–B) Random Forest analysis of leaf-associated bacterial ASVs under open-field and rain-sheltered cultivation treatments. (C) Shannon and Chao indices of oomycete communities (n = 6). (D) PcoA of oomycete community composition. (E) NST analysis of oomycete community assembly. (F) Relative contributions of ecological assembly processes. (G) Effects of NPS1 combined with phyllosphere-derived Sphingomonas spp. (SPH) or Bacillus subtilis (BA) on Colletotrichum spp. growth (n = 3). (H) Effects of NPS1 combined with SPH or BA on Phytophthora spp. growth. Bars represent mean ± SD. Different lowercase letters indicate significant differences among treatments according to one-way ANOVA followed by Tukey’s HSD test (p < 0.05)

To further characterize oomycete communities associated with grape downy mildew under NPS1 and CK treatments, oomycete diversity and community assembly patterns were investigated. Both Shannon diversity and Chao indices of oomycete OTUs were significantly lower in NPS1-treated samples than in CK samples (p < 0.01) (Fig. 8C). PCoA showed clear separation between oomycete communities from NPS1-treated and CK samples, with the first principal coordinate explaining 95.12% of the observed variation (Fig. 8D). NST analysis indicated differences in inferred community assembly processes between NPS1-treated and CK samples. NST values below 50% were consistent with a greater contribution of deterministic processes, and significant differences were observed between treatments (p < 0.001) (Fig. 8E). Further analysis suggested differences in the relative contributions of inferred assembly processes between treatments. Among the inferred assembly processes, drift (including other stochastic processes), homogeneous dispersal, and homogeneous selection accounted for the largest proportions within the phyllosphere oomycete communities (Fig. 8F).

Discussion

NPS1 treatment was associated with coordinated changes in leaf surface properties, grapes physiological responses, phosphorus availability, and phyllosphere microbial communities, which collectively may contribute to grape disease suppression. In the present study, NPS1 increased leaf hydrophobicity and shortened leaf wetness duration, potentially reducing favorable conditions for pathogen survival and infection. In parallel, NPS1 treatment was associated with alterations in antioxidant enzyme activity, gene expression, metabolite accumulation, and microbial community structure, indicating the involvement of multiple stress-related processes in the observed disease resistance phenotype. In addition to the physicochemical effects of NPS1 on the leaf surface, the present study also demonstrated significantly increased P accumulation in grape leaves at 1, 6, and 14 d after NPS1 application, including after deionized water washing treatment. These observations suggest that a fraction of the applied phosphorus remained associated with the leaves after water washing, although the current experiment cannot unequivocally distinguish internalized P from strongly retained surface residues. Therefore, the physiological and molecular responses observed after NPS1 treatment may be associated not only with nanoscale surface interactions, but also with improved foliar phosphorus availability. Overall, the results support a multi-factor response model in which physical, nutritional, physiological, metabolic, and microbial processes may jointly contribute to disease suppression. However, several observations derived from transcriptomic, metabolomic, and microbiome analyses remain correlative, and direct causal relationships require further experimental validation.

Multifaceted potential of NPS1 and its link to grape disease resistance

The occurrence and epidemic of foliar diseases are major factors affecting grape yield and quality [1]. Leaf wetness represents a primary factor facilitating downy mildew development [31], while rainwater serves as the main medium for disease dispersal [32]. In this study, nano-superphosphate (NPS1), compared with conventional superphosphate fertilizers, exhibited smaller particle size, hydrophobic properties, and good stability, which may have contributed to reduced leaf wetness duration and lower disease incidence. Nano-superphosphate promoted leaf growth, increased plant height, and enhanced chlorophyll content, consistent with previous findings that nano-hydroxyapatite stimulates chlorophyll synthesis and promotes maize growth [33]. Furthermore, certain nanomaterials and nano-fertilizers have recently been reported to improve grape berry quality and yield [34], potentially through enhanced nutrient absorption and transport mediated by their modified specific surface area [35]. In the present study, dynamic analysis of leaf phosphorus (P) content showed that NPS1 treatment significantly increased leaf-associated P concentrations at 1, 6, and 14 d after application compared with the CK treatment. Moreover, elevated P levels remained detectable even after deionized-water washing, indicating that a fraction of the applied phosphorus remained associated with the leaves following washing. However, because NPS1 particles may be strongly retained on the leaf surface, the current experiment cannot unequivocally distinguish phosphorus internalized into leaf tissues from phosphorus remaining as surface-associated residues. Therefore, additional approaches such as isotope tracing, cross-sectional elemental imaging, or more rigorous surface-removal procedures will be required to determine the fate and localization of the retained phosphorus.

At the same time, NPS1 treatment increased leaf hydrophobicity and shortened leaf wetness duration. Although enhanced hydrophobicity could theoretically reduce liquid residence time on the leaf surface and thereby limit foliar nutrient uptake, the observed increase in leaf-associated P content suggests that phosphorus absorption may still occur following NPS1 application. One hypothetical explanation is that phosphorus uptake may take place during the early stage after spraying, before substantial hydrophobic modification of the leaf surface becomes fully established [36]. In addition, the nanoscale particle size and high surface activity of NPS1 might hypothetically facilitate penetration through stomatal openings or cuticular pathways, thereby allowing nutrient absorption even under relatively short liquid residence times [37]. Therefore, the enhanced hydrophobicity induced by NPS1 coincided with reduced pathogen-favorable moisture conditions, while coinciding with elevated leaf-associated phosphorus following NPS1 application. Future studies comparing nano-scale phosphorus materials with conventional phosphorus treatments under equivalent foliar P concentrations will be necessary to distinguish nutritional effects from nanomaterial-specific mechanisms. In our study, nano-superphosphate significantly reduced the disease indices of both downy mildew and powdery mildew, aligning with reports that graphene oxide (GO)-Fe₃O₄ nanocomposites effectively inhibit pathogen spore germination and prevent grape downy mildew [38]. This effect is putatively associated with its antimicrobial activity, as suggested by in vitro inhibition assays, rather than implying a mechanism identical to that of iron oxide nanoparticles [39]. The antimicrobial activity observed in plate assays supports the potential application of nano-superphosphate as a multifunctional material with both fertilizing and disease-suppressive roles in agriculture [40].

Under stress conditions, plants generate excessive reactive oxygen species (ROS). If not promptly scavenged, ROS can cause oxidative damage to host cells and potentially lead to plant death [41]. Nano-superphosphate increased SOD activity, enhanced POD enzyme levels, and showed a trend toward reduced MDA content, a phenomenon consistent with observations of nano-phosphate effects in common bean [42]. Previous research has shown that nano-selenium enhances antioxidant capacity and improves melon resistance to Verticillium wilt [43]. Because phosphorus is an essential macronutrient involved in energy metabolism, membrane synthesis, and stress adaptation, the observed antioxidant responses may reflect both improved nutritional status and physicochemical effects associated with NPS1 treatment. These results collectively indicate that NPS1 treatment is concurrently associated with reduced leaf wetness, increased phosphorus retention on leaves, and physiological responses linked to antioxidant defense systems; however, whether these multifaceted changes operate via a single integrated mechanism or act as parallel, independent responses remains to be fully determined.

NPS1 is associated with transcriptional and metabolic responses related to stress resistance and antioxidant processes in grape leaves

Nanomaterial-mediated regulation of plant stress responses is often accompanied by transcriptional and metabolic reprogramming [44, 45]. In the present study, NPS1 treatment significantly altered gene expression and metabolite accumulation patterns in grape leaves, suggesting that NPS1 may influence stress-related physiological pathways. However, the direct causal relationships between these molecular responses and enhanced disease resistance remain to be fully validated. GO enrichment analysis revealed that differentially expressed genes were mainly associated with catalytic activity, binding, and cellular metabolic processes, indicating broad metabolic adjustments and signal transduction responses following NPS1 treatment [46]. In addition, enrichment of oxidative stress-related pathways, together with the increased expression of PEROXIDASE validated by RT-qPCR, suggested activation of antioxidant-related processes. Peroxidases are known to participate in H₂O₂ scavenging and mitigation of oxidative damage in plants [26]. Similar activation of antioxidant systems has also been reported in plants treated with other nanomaterials, including nano-iron formulations [47]. KEGG enrichment analysis further indicated that pathways associated with secondary metabolite biosynthesis, environmental adaptation, and signal transduction were significantly affected by NPS1 treatment. In particular, enrichment of phenylpropanoid biosynthesis pathways suggested enhanced synthesis of stress-associated secondary metabolites, including lignin- and flavonoid-related compounds [48]. In addition, enrichment of chitin-responsive pathways and the upregulation of transcription factors such as WRKY41 and EREBP may indicate activation of defense-related signaling pathways. Previous studies have shown that WRKY transcription factors participate in disease resistance through regulation of defense-related genes [49], whereas EREBP family members are involved in responses to both biotic and abiotic stresses [50]. Increased expression of CPK also suggested possible activation of calcium-mediated signaling pathways involved in stress signal transduction [51]. Untargeted metabolomics analysis revealed substantial metabolic alterations following NPS1 treatment. Metabolite annotation was primarily based on MS/MS spectral matching and database comparison according to the Metabolomics Standards Initiative (MSI) criteria, and most metabolites identified in this study were classified as MSI Level 2 (putatively annotated compounds). Therefore, interpretations regarding the biological functions of these metabolites should be considered tentative and require further validation using authentic standards. Among the metabolites showing increased accumulation after NPS1 treatment, compounds putatively identified as Leucocyanidin, Ellagic acid arabinoside, and Caftaric acid have previously been associated with antioxidant activity and stress adaptation in plants. These metabolites may contribute to reactive oxygen species scavenging and defense-related metabolic regulation. The enrichment of flavonoid biosynthesis and phenylalanine metabolism pathways corresponded well with transcriptomic enrichment of phenylpropanoid biosynthesis pathways, suggesting coordinated regulation between transcriptional and metabolic responses [52]. Integrated transcriptomic and metabolomic analyses further indicated co-enrichment of secondary metabolite biosynthesis pathways, implying that NPS1 treatment may be associated with coordinated metabolic adjustments linked to stress adaptation. Several differentially expressed genes, including UGT75L6 and PRDX6, were also associated with altered metabolite accumulation patterns. UGT family genes are generally involved in glycosylation-related processes in secondary metabolism, whereas PRDX family proteins are commonly associated with redox homeostasis and antioxidant regulation. However, the specific biological functions and regulatory mechanisms of PRDX6 in grape remain poorly understood and require further investigation. Importantly, these observations should currently be interpreted as associations rather than direct mechanistic evidence. In addition, enrichment of ABC transporter pathways and SNARE-mediated vesicular transport pathways suggested that NPS1 treatment may influence intracellular metabolite transport and stress-response signaling processes [53]. However, the functional significance of these pathways under NPS1 treatment has not yet been experimentally validated. Although NPS1 treatment was associated with substantial transcriptomic, metabolic, and physiological changes, the primary mechanisms underlying these responses remain unclear. The observed effects may involve multiple contributing factors, including altered leaf surface hydrophobicity, changes in nutrient availability, and potential nano-scale interactions with plant tissues. In particular, increased leaf hydrophobicity may reduce leaf wetness duration while simultaneously influencing foliar nutrient absorption efficiency. Therefore, the relative contributions of physicochemical properties and nutrient-related effects to the observed biological responses require further investigation. Overall, the present results suggest that NPS1 treatment is associated with coordinated transcriptomic and metabolomic changes linked to antioxidant activity and stress adaptation in grape leaves. Nevertheless, further studies involving targeted metabolite quantification, functional validation of key genes, and mechanistic investigations will be necessary to clarify the specific contributions of these pathways to grape disease resistance and to support the rational application of nanomaterials in sustainable plant protection.

NPS1 reshapes the grape phyllosphere microbiome and exhibits parallel associations with specific microbial taxa

The phyllosphere microbiome serves as the plant’s first line of biological defense, with its community structure balance and functional homeostasis being closely linked to host disease resistance [54]. This study indicates that NPS1 treatment coincided with enhanced disease suppression, accompanied by substantial shifts in the grape phyllosphere microbiome composition, co-occurrence network topology, and community assembly processes. While NPS1 treatment was associated with microbial taxa previously linked to plant health, these observed variations provide ecological clues regarding nanomaterial-mediated microenvironmental dynamics rather than direct functional evidence of targeted ecological regulation. NPS1 treatment was associated with a restructuring of the phyllosphere microbial community, with effects varying by cultivation mode. Under open-field conditions, it enhanced both fungal and bacterial diversity (particularly the bacterial Shannon index), while showing opposite trends under rain-sheltered conditions. This discrepancy may relate to differences in leaf microhabitats (humidity and nutrient exudates) across cultivation environments [7], which might be influenced by altered leaf surface properties following NPS1 treatment. LEfSe analysis showed that NPS1 significantly reduced the relative abundance of the powdery mildew pathogen Erysiphe under both cultivation modes while enriching taxa such as Deinococcus and Ralstonia. Previous studies have suggested that the radiation resistance and antioxidant characteristics of Deinococcus may contribute to microbial stress tolerance and community adaptation [55], potentially reflecting microbial community shifts associated with pathogen suppression. However, whether these taxa are actively recruited by the host or simply respond passively to altered microenvironments remains speculative and requires future functional validation.

Changes in microbial network topology revealed shifts in community architecture that may relate to altered ecological organization. Following NPS1 treatment, both fungal and bacterial networks showed reduced connectivity, increased proportions of negative correlations, and enhanced modularity under both cultivation modes. Higher proportions of negative correlations are frequently interpreted in the literature as mathematical indicators potentially associated with competitive or exclusive microbial relationships that may limit pathogen proliferation [56], whereas increased modularity may reflect altered ecological partitioning within complex microbial communities [57]. However, it is important to emphasize that amplicon-based co-occurrence networks cannot directly demonstrate actual physical interactions, biological competition, or antagonistic relationships among microorganisms [59]. Robustness analysis indicated that NPS1 treatment was associated with higher inferred network stability metrics, suggesting that NPS1 may influence microbial taxa co-varying with network topology patterns [58]. Nevertheless, these topological changes should be interpreted cautiously because network analysis alone cannot distinguish whether niche overlap or exclusion results from stochastic processes, habitat filtering, or true biological interactions among microorganisms [59].

The concurrent associations between NPS1 treatment and specific phyllosphere bacterial taxa suggested a potential relationship with pathogen suppression. Random forest analysis identified Sphingomonas spp as a core bacterial taxon statistically associated with reduced abundance of Colletotrichum spp and Phytophthora spp. We speculate that NPS1 may potentially influence the persistence of antimicrobial compounds (such as chitinase from Sphingomonas spp or lipopeptides from Bacillus spp) [60, 61], while simultaneously affecting the colonization microenvironment of phyllosphere microorganisms through plant metabolic regulation (e.g., phenolic compound accumulation). However, these mechanistic pathways remain hypothetical and have not been validated through in-planta functional assays or microbial reinoculation experiments. Furthermore, NPS1 significantly reduced oomycete diversity and shifted community assembly toward more deterministic processes, suggesting a potential ecological filtering effect that might be indirectly mediated through host physiological responses, immune signaling, or microhabitat specificity [62].

In summary, NPS1 treatment was associated with grape disease resistance alongside three potentially parallel microbial phenomena: reshaped microbial community structures associated with reduced pathogen abundance, altered network topology patterns, and concurrent associations with specific phyllosphere bacterial taxa. Because direct causal relationships remain unverified, future time-series analyses and functional characterization of key microbial taxa are necessary to determine whether these microbial shifts actively contribute to disease resistance or instead represent secondary responses to host physiological changes.

Integrated multi-pathway associations of NPS1 in grape leaf disease suppression

The phyllosphere serves as a critical interface for plant–microbe–environment interactions [54], where the microbiome, metabolome, and transcriptome are widely considered key components reflecting plant stress responses, antioxidant regulation, and disease resistance strategies [63, 64]. Correlation analyses among differentially expressed genes, metabolites, and dominant microorganisms in this study revealed extensive interconnections. Microbial taxa previously associated with grape stress resistance and disease defense (Deinococcus, Massilia, Ralstonia), upregulated stress-related genes (WRKY41, EREBP, PEROXIDASE), and key enriched metabolites were identified; these molecular and microbial components are known to participate in pathways associated with reactive oxygen species scavenging, cell wall reinforcement, and pathogen suppression [65–67]. Previous studies have suggested that nanomaterials can influence metabolite synthesis concurrently with changes in plant gene expression, thereby potentially modifying microbial colonization microenvironments [68]. Our findings indicate that under NPS1 treatment, the mathematical gene–metabolite–microbe correlation network appeared to become more interconnected. This statistical correlation pattern was observed concurrently with reductions in foliar pathogens such as Erysiphe, enrichment of specific microbial taxa, and activation of secondary metabolic pathways, while potentially aligning with concurrent associations involving phyllosphere-isolated Sphingomonas spp and Bacillus strains. However, because these multi-omics insights are entirely derived from correlative network analyses, a definitive causal chain demonstrating that NPS1 directly drives transcriptomic reprogramming, which subsequently orchestrates downstream metabolic shifts and microbial recruitment, has not yet been experimentally established.

In conclusion, our results suggest that NPS1 treatment is associated with a multi-layered response network in the grape phyllosphere involving phosphorus retention, leaf surface physicochemical properties, host physiological responses, defensive metabolites, and core microbial taxa, together with potential associations involving phyllosphere bacterial taxa previously linked to plant health. This research provides valuable insights into the multifaceted changes through which nanomaterials may correlate with plant stress responses and disease immunity, while highlighting close associations among phyllosphere multi-omics components and plant health. The observed associations among genes, metabolites, and microbial taxa offer conceptual framework for exploring sustainable agricultural strategies utilizing nanotechnology, although explicit verification of the underlying mechanisms remain to be experimentally validation.

Conclusions

In conclusion, NPS1 treatment was associated with coordinated shifts in leaf surface characteristics, host physiological and metabolic responses, and phyllosphere microbial communities. These changes were accompanied by reduced foliar disease severity and improved grapevine performance under both cultivation systems. While the causal relationships among these responses require further investigation, the present findings suggest that NPS1 may contribute to disease suppression through the combined modulation of plant traits and phyllosphere microbiota. This study provides a foundation for the development of hydrophobic nano-fertilizers as sustainable tools for integrated disease management in grape production.

Author contribution

Tao He †: Conceptualization, Sampling, Investigation, Formal analysis, Visualization, Writing – original draft. Mingxin Fu †: Conceptualization, Sampling, Investigation, Formal analysis. Meijing Wang: Sampling, Investigation. Meng Yang: Sampling, Investigation. Mei Li: Sampling, Investigation. Ronghui Du: Sampling, Investigation. Rongqi Lin: Formal analysis. Weiping Deng: Conceptualization, Formal analysis. Min Yang: Formal analysis. Airong Li: Writing – review & editing, Resources. Shusheng Zhu: Conceptualization, Resources. Fei Du: Conceptualization, Resources, Writing – review & editing. Tao He and Mingxin Fu contributed equally to this work. The author’s order was determined on the basis of seniority.

Funding

National Natural Science Foundation of China,32360712,Yunnan Province Agricultural Joint Special Project,202101BD070001-125,The Expert Workstation Project in Yunnan Province,202305AF150129.

Data availability

The datasets generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under accession numbers PRJNA1368410, PRJNA1368411, PRJNA1368436, and PRJNA1369163. https://www.ncbi.nlm.nih.gov/sra/.

Declarations

Competing interests

The authors declare that the research was conducted in the absence of any commercial or financiall relationships that could be construed as a potential conflict of interest.