The pine wood nematode (PWN) Bursaphelenchus xylophilus (Steiner and Buhrer, 1934) Nickle 1970 isolate was previously isolated and stored in our laboratory, Guangdong Province Key Laboratory of Microbial Signals and Disease Control, South China Agricultural University, China (Kamaruzzaman et al., 2024). PWN was grown on the mycelium of Botrytis cinerea (7-days old culture) cultured at 25°C for one week in the dark on potato dextrose agar. The Baermann funnel method was employed to harvest the PWN from the culture medium (Mannaa and Seo, 2023). The harvested nematodes were thoroughly washed with sterilized distilled water (sdH₂O) to remove most contaminants from the cuticle and stored at 4°C for future use.

Soil samples were collected from roots of the of healthy Pinus massoniana free from PWN infestation, on the campus of South China Agricultural University, Guangzhou, China. Three soil samples were collected from P. massoniana and mixed together to create a composite sample. Bacterial strains were isolated from the collected soil sample using a serial dilution method described by Ahmed (Ahmed et al., 2022c). For the isolation of bacterial strains, one gram of soil was thoroughly mixed in 9 mL of sdH₂O and incubated at 28°C and 180 rpm for 30 minutes. The mixture was subsequently serially diluted (10-folds), and 100 μL of the diluted solution was spread on nutrient agar (NA) media supplemented with 50 mg/L Nystatin (to inhibit fungi growth) with a sterile spreader and incubated at 28°C for two days. On the basis of colony morphology (shape and color), a single colony was picked and streaked on NA medium for purification and 63 rhizobacterial strains were collected. After five rounds of streaking and purification, the purified strains were stored in 50% glycerol stock solution (v/v) at −20°C for future use (strains identification and nematode mortality assays).

The isolated bacterial strains were reactivated on NA media at 28°C for 48 hours to determine their nematocidal potential against PWN. The nematicidal activity of the isolated strains was assessed against PWN via the preparation of fermentation filtrates (FFs) according to the methodology of Puttawong et al. (2024). The seeds solutions of the isolated bacterial strains were prepared by adding a single colony from each strain into 50 mL of nutrient broth (NB) and incubating at 28°C for 15 hours. For the preparation of FFs, 100 mL of NB was inoculated with 1% seed solution and incubated on a shaker for 72 hours at 28°C and 180 rpm. The fermentation mixture was subsequently centrifuged at 4°C and 10,000 rpm for 10 minutes, which was repeated twice. The supernatant was collected and subsequently filtered using a 0.22 μm bacterial filter to obtain FFs and then stored at 4°C for later use. We used the immersion method to evaluate the nematicidal activity of FFs of each bacterial strain against PWN (Moo-Koh et al., 2022). A mixture of 10 μL of nematodes suspension (∼100 PWN) and 90 μL of FFs was added to a 48-well cell culture plate (each well) and incubated in the dark at 28°C for 48 hours, whereas sdH₂O was used as a control. Nematodes morphology was observed every 24 hours using a stereomicroscope (Nikon SMZ 745; Nikon, Tokyo, Japan) to calculate the corrected mortality rate (C-M-R). Nematodes were determined to be dead when their bodies were stiff and unresponsive to stimulation by a dissecting needle (Ali et al., 2023). The C-M-R was calculated via the following formula (Ayaz et al., 2021): CM (%) = (T− CK)/(1 − CK) × 100. Where T represents the mortality rate in the treatment group and CK represents the mortality rate in the control group. The assay was repeated three times, with minimum three replicates per treatment. From the isolated bacterial strains, strain DP2-30 exhibited strong nematicidal activity against PWN and was selected for subsequent analyses.

The potential biocontrol strain DP2-30 was identified on the basis of morphological (colony shape, color, cell size, and Gram staining) and molecular (PCR amplification and phylogenetic analysis) characteristics. Morphological identification of strain DP2-30 was performed by growing it on NA media at 28°C for 48 hours. Gram staining of strain DP2-30 was performed using a Gram Stain Kit, Solarbio Life Sciences & Technology Co., Ltd. (Beijing, China). The cell morphology and size were observed by Scanning Electron Microscope (SEM; (LEO 1450 VP SEM, Germany) (Taheri et al., 2023). Molecular identification of strain DP2-30 was performed by PCR amplification of 16S rRNA gene via primer pair 27F-(5-AGAGTTTGATCCTGGCTCAG-3) and 1492R (5-GGTTACCTTGTTACGACTT-3) (Zhang et al., 2022b). The PCR products were observed using a 1% agarose gel electrophoresis in a gel documentation system (Tanon 2500) and sent to Guangzhou Aiji Biotechnology Co., Ltd. (Guangdong, China) for sequencing. The obtained sequencing results were analyzed by the BLASTN tool in the NCBI database. Mega 11 was used to edit, build, Clustal W, and alignment of sequences, and a phylogenetic tree was constructed with the neighbor-joining method (bootstrap analysis with 1000 replicates) by taking Bacillus sp. as an outgroup (Tamura et al., 2021).

The effects of different culture medium conditions, such as incubation time (days), temperature (°C), shaking speed (rpm/min), and culture medium pH on the nematicidal activity of the L. pinisoli DP2-30 against PWN were assessed to optimize the fermentation conditions to maximize the nematicidal activity (Kamaruzzaman et al., 2024). The seed solution of strain DP2-30 was prepared as described above (Section 2.3), and 1% of the seed solution was used for subsequent analyses. The impact of incubation time on the nematicidal activity of strain DP2-30 was examined by growing strain DP2-30 in 100 mL of NB for 1, 2, 3, 4, 5, 6, and 7 days and the nematicidal activity was checked every 24 hours. To study how the pH of the culture medium influences the nematicidal activity of DP2-30 against PWN, the pH of the culture medium was adjusted to 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, and 10.0. Subsequently, 100 mL of culture mediums (NB) with different pH values were inoculated with strain DP2-30 seed solution and incubated at 180 rpm/min and 28°C for 72 hours. To assess the effect of incubation temperature (°C), 100 mL of culture medium (NB) containing DP2-30 seed solution was cultured at 180 rpm/min and different temperatures 22, 25, 28, 31, 34, and 37°C for 72 hours. Finally, the impact of shaking speed (rpm/min) on the nematicidal activity of strain DP2-30 was determined by culturing 100 mL of NB at 120, 140, 160, 180, 200 and 220 rpm/min and 28°C for 72 hours. The FFs of strain DP2-30 for each assay were collected via centrifugation as mentioned in section 2.3, and their nematicidal activity was assessed as described above (Section 2.3). All the assays were performed three times, with at least three replications per treatment for each assay.

The effects of different conditions (FFs, fermentation broth (FB), and bacterial suspension) of strain DP2-30 on PWN eggs hatching were studied as described by Sun et al. (2022). The PWN eggs were collected following the methodology of Liu and colleagues (Liu et al., 2019a) and adjusted to a concentration of 100 eggs per 10 μL via a microscope (Nikon SMZ 745; Nikon, Tokyo, Japan). A 10 μL of egg suspension (contained ~100 PWN eggs) was added to a 48-well cell culture plate containing 90 μL of FFs, FB, and bacterial suspension (OD₆₀₀ = 0.5) and incubated in the dark at 28°C for 36 h, while sdH₂O was used as a control. The PWN (L2 stage) number was counted under a stereomicroscope (Nikon SMZ 745; Nikon, Tokyo, Japan). The PWN egg hatching rate (HR: %) and relative inhibition rate (IR: %) were calculated using the following formulas: HR (%) = L2 × 100/eggs + L2 and IR (%) = (HR of control – HR of treatment) × 100/HR of control. The experiment was performed three times, with three replication per treatment.

The growth-promoting potential of strain DP2-30 was assessed on a model plant (tobacco; Nicotiana benthamiana). For the growth promotion assay, the FB and FFs of strain DP2-30 were obtained by growing it in NB media at 28°C and 180 rpm/min for 72 hours. Tobacco seeds were first surface sterilized with 10% sodium hypochlorite solution for 5 min, followed by 4 times washing with sdH₂O (Shoyeb et al., 2020). After surface sterilization, the tobacco seeds (30) were soaked in 2 mL of FFs and FB for 12 hours, while NB and sdH₂O were used as controls. The seeds (30 seeds per dish, 3 dishes per treatment) were then placed in Petri dishes containing MS media (0.8% agar, 1.5% sucrose, and pH 5.7) and placed in a light incubator at 25°C (12 h light and 12 h dark) under controlled conditions (Sheikh et al., 2023). The assay was performed three times with three dishes per treatment as replicates and each dish contained ten seeds. After 14 days, the seed germination rate, fresh weight, root length, and stem length were recorded and calculated (Sheikh et al., 2023).

The biocontrol potential of strain DP2-30 against PWD was investigated in a greenhouse pot experiment on 2-year-old pine seedlings (P. massoniana) artificially inoculated with PWN. Before the experiment, strain DP2-30 was grown in NB medium at 28°C and 180 rpm/min for 72 hours to prepare FB. The experiment was executed following a completely randomized design under three conditions: T1 (inoculation of ∼3000 PWN per mL/plant), T2 (combined application of DP2-30 FB and ∼3000 PWN per mL/plant), and CK (application of sdH₂O as blank control). We used sdH₂O as control to establish a baseline for comparison, which allows us to assess the effects of DP2-30 without any confounding variables that might arise from other treatments. Strain DP2-30 FB (30 mL/plant) was inoculated via the root drenching method around the base of pine plant by disturbing the above 1 cm layer, and the PWN suspension was artificially inoculated into the main stem 5 cm below the top with superficial longitudinal incisions in a well-sealed premade parafilm funnel and kept moist (Rodrigues et al., 2021). Disease development was observed at 9 days of post-inoculation with PWN to 45 days with an interval of 9 days. The disease severity index (%) was calculated for each treatment using a 5-point disease grading scale (0: asymptomatic plants, 1: 1−25% wilting of pine needles, 2: 26−50% wilting of pine needles, 3: 51−75% wilting of pine needles, and 4: 76−100% wilting of pine needles or death of plant). The disease severity index (%) and control effect (%) were calculated using formulas as previously described by Kamaruzzaman et al. (2024). The experiment was repeated three times and 8 pine seedlings were inoculated per treatment as replicates for each individual assay. To study, the impact of strain DP2-30 on the host microbiome, particularly in the context of its potential role in managing PWN infestations through modulation of host microbiome. Root and stem samples were collected from the pine seedlings under different treatments from the in-planta biocontrol assay. A total of 36 samples (18 roots and 18 stems; 6 samples per treatment) were collected under different treatments (CK, T1, and T2), placed in liquid nitrogen, and stored at −80°C for host-microbiome analysis. The samples collected from each treatment were labelled according to sample type: CK-R, T1-R, and T2-R (roots) and CK-S, T1-S, and T2-S (stem).

The total genomic DNA from the plant samples were extracted as previously described in our study (Ahmed et al., 2022a), using a Plant DNA Extraction Kit (Zymo Research, USA). The quality of the extracted DNA was assessed at an OD_A260/A280 ratio of≥1.8 using an ND2000 nanodrop spectrophotometer (Thermo Scientific, USA). For microbiome analysis, the V5-V7 region of bacterial 16S rRNA gene and fungal ITS1 region were amplified with the primer sets 799F (5-AACMGGATTAGATACCCKG-3)/1193R (5-ACGTCATCCCCACCTTCC-3) (Li et al., 2023b) and ITS1-F_KYO2 (5-TAGAGGAAGTAAAAGTCGTAA-3)/ITS-86R (5-TTCAAAGATTCGATGATTCAC-3) (Toju et al., 2012), respectively. The obtained PCR samples were subjected to paired-end sequencing on an Illumina MiSeq platform at Gene Denovo Biotechnology Co., Ltd. (Guangzhou, China). The raw sequences of bacteria and fungi were analyzed and processed using the UPARSE pipeline according to standard protocols and quality controlled by USEARCH software to generate clean reads (Edgar, 2013). Chimeras were eliminated via UCHIME, and subsequently, the remaining clean sequences were allocated to operational taxonomic units (OTUs) on the basis of a similarity standard of 97% (Edgar et al., 2011). Ultimately, the sequence of each OTU was compared to the RDP database of bacteria and UNITE database of fungi according to the Naïve Bayesian algorithm of the RDP Classifier for taxonomic annotation at a confidence threshold level of 0.8~1 (Wang et al., 2007; Nilsson et al., 2019). The raw data from the greenhouse experiments are available in the NCBI Sequence Read Archive (SRA) under BioProject No. PRJNA1132778 with the sample names of CK-R, T1-R, and T2-R (R; root) and CK-S, T1-S, and T2-S (S; stem).

The alpha diversities of diversities (Shannon, Chao, PD-whole-tree, and Chao-1) bacterial and fungal communities alpha were calculated using QIIME 2 (Bolyen et al., 2019). We used principal coordinate analysis (PCoA) based on the Bray–Curtis distance metric to assess the variations in the structure of the bacterial and fungal community compositions. PERMANOVA with Adonis function was employed to determine the overall dissimilarity in the structure of the microbial communities using “vegan” package in R (version 4.0.1) (Oksanen et al., 2013). The relative abundance bar, box, and bubble plots for the bacterial and fungal communities were generated using the “ggplot” package in R (version 4.0.1). The intra-kingdom microbial network analysis was performed at the OTU level by excluding those with a relative abundance < 0.01, p < 0.05, and Pearson correlation coefficient > 0.7 using R sparcc package and visualized with Gephi 0.9.2 (Zhang et al., 2024). We employed one-way ANOVA in SPSS v20.0 (SPSS Inc. USA) to evaluate the impact of different treatments on the disease severity index, nematicidal activity, alpha diversity, and microbial community composition and were considered significant if p < 0.05 according to Tukey-HSD and Duncan’s multiple range tests. The figures were edited, merged, and combined using Adobe Illustrator 2022.

A total of 63 bacterial strains were preliminarily isolated and purified from the rhizosphere soil of healthy Pinus massoniana plants. The nematicidal activity of isolated bacterial strains was investigated against PWN using fermentation filtrates (FFs) by the immersion method, among which 10 bacterial strains showed various degree of nematicidal activity (20.70% to 92.08%) against PWN after 48 h of exposure. The strain DP2-30 displayed the highest nematicidal activity with a corrected mortality rate (C-M-R) of 92.08% against PWN ( Figures 1A, B ), followed by strains DP1-4 and DP2-15 with C-M-R of 75.01% and 73.94% (moderate nematicidal activity), respectively. Several strains, including DP2-13, DP2-27, DP1-2, DP1-18, Dp2-2, DP2-24, and DP2-41 exhibited low nematicidal activity against PWN, with C-M-R values less than 40% ( Table 1 ). In light of these findings, we therefore, chose DP2-30 as a possible biocontrol strain for further investigation. Morphological and biochemical analyses were performed for strain DP2-30. On NA media strain DP2-30 produced smooth, opaque, and round to convex yellow color colonies ( Figures 1C, D ). Scanning electron microscopy revealed that the bacterial cells were rod-shaped with 1.19 ± 0.20 μm in length and 0.37 ± 0.03 μm in diameter, and Gram staining was negative ( Figures 1E, F ). Molecular identification of strain DP2-30 was performed via PCR amplification of the 16S rDNA gene using primer set 27F/1492R. A 938 bp PCR product was obtained, and NCBI BLASTN analysis showed 98.80% similarity with Luteibacter pinisoli MAH-14 (NR_179163.1). Further phylogenetic tree analysis via neighbor-joining method revealed that strain DP2-30 was in the same clade as L. pinisoli and was different from other Luteibacter species ( Figure 1G ). Thus, on the basis of morphological, biochemical, and molecular analyses, the isolated strain DP2-30 was determined to be L. pinisoli and was named as L. pinisoli P2-30.

The effects of various fermentation conditions, including culture medium time, pH, temperature, and shaker speed on the nematicidal activity of L. pinisoli DP2-30 were assessed against pine wood nematode (PWN) using fermentation filtrates (FFs). Results revealed that the nematicidal activity of strain DP2-30 FFs first increased with the increasing in incubation time, temperature, speed, and fermentation medium pH and then gradually decreased ( Figures 2A–D ). The nematocidal activity of strain DP2-30 FFs was the highest against PWN at 3 d of incubation with a C-M-R of 95.21% ( Figure 2A ). At pH 7, the nematicidal activity of strain DP2-30 FFs was maximum, with a C-M-R of 94.10% ( Figure 2B ). When the incubation temperature was 28°C, strain DP2-30 FFs presented the highest killing rate against PWN with a C-M-R of 94.60% ( Figure 2C ). The FFs of strain DP2-30 exhibited strong nematicidal activity against PWN when cultured at 180 rpm/min and 200 rpm/min, having a C-M-R values of 92.43% and 91.71%, respectively ( Figure 2D ). On the basis of these findings, the optimal fermentation conditions for maximizing the nematicidal activity of L. pinisoli DP2-30 against PWN appear to be 3 days of incubation at 28°C, pH 7, and a shaking speed of 180 rpm/min.

The effects of strain DP2-30 fermentation broth (FB), FFs, and bacterial suspension on PWN egg hatching were calculated after 36 hours of incubation at 28°C. The FB and FFs significantly reduced PWN egg hatching compared to the control group. The hatchability rates were recorded at 41.17% and 46.33% for the FB and FFS, respectively, which were substantially lower than the control. In contrast, the bacterial suspension of strain DP2-30 itself had only a slight inhibitory effect on egg hatching, with a hatchability rate slightly lower than that of the control group. The FB and FFs demonstrated a relatively high inhibition rates of 49.38% and 43.055%, respectively, on PWN egg hatching compared to the control ( Table 2 ). These results indicate that the metabolites produced during fermentation by strain DP2-30, rather than the bacteria themselves, are responsible for the significant inhibition of PWN egg hatching.

The effects of strain DP2-30 FB and FFs were evaluated on tobacco seed germination and seedling growth ( Supplementary Table S1 ; Supplementary Figure S1 ). After the tobacco seeds were soaked in strain DP2-30 FB, FFs, nutrient broth (NB medium) and sdH₂O for 12 h, they germinated normally and no significant effects were observed among treatments. The germination rate reached more than 90.00% in all treatments, and no significant difference was observed among treatments, indicating that strain DP2-30 did not affect the tobacco seed germination rate. After 14 days of tobacco seed growth in the dish, the tobacco seeds in the treatment group had stronger buds, thicker green leaves, longer roots, and greater fresh weight than the control group ( Supplementary Figure S1 ). The root length of tobacco seedlings treated with FB was significantly increased by 75.05% and 74.30% compared with control groups, while the fresh weight was increased by 58.6% and 61.4%, respectively. Similarly, the root length of tobacco seedlings treated with FFs significantly increased by 60.4% and 59.8% compared with control groups, whereas the fresh weight increased by 49.1% and 51.8%, respectively ( Supplementary Table S1 ). The results showed that strain DP2-30 significantly promote the growth of the tobacco seedlings, and could be used as a potential plant growth promoting agent.

The biocontrol efficacy of L. pinisoli DP2-30 was assessed in mitigating PWD in pine seedlings treated with PWN and strain DP2-30 FB. The disease severity index (%) was calculated from the 9^th day after inoculation with PWN till the 45^th days, with the interval of 9 days, whereas the control effect (%) was calculated at the end of the experiment ( Table 3 and Figure 3 ). The results indicated that symptom of PWD (fade, yellowing, and wilting of needles) were first observed on the 9^th day post-inoculation with PWN in treatment T1 (pine seedlings solely inoculated with PWN). In contrast, the application of DP2-30 FB (T2; combined treatment with DP2-30 FB and PWN) resulted in delayed symptom development in pine seedlings, and PWD symptoms were first observed on the 27^th after inoculation with PWN. The disease incidence of PWD gradually increased with the increasing time in treatments T1 and T2, whereas the control plants (CK) treated with sdH2O exhibited normal growth, were healthy, and remained lush green ( Figure 3 ). At 45 days after inoculation with PWN, the highest disease severity index of 100% was observed in treatment T1. However, pine seedlings treated with DP2-30 FB + PWN (T2) displayed a minimum disease severity index of 37.5% and maximum control effect of 62.50% ( Table 3 ). These results showed that the application of DP2-30 FB (T2) significantly delayed the occurrence of PWD in pine seedlings.

Thirty-six samples (18 roots and 18 stems) of pine plants from different treatments (CK, T1, and T2) were subjected on the Illumina platform for analysis of the host endophytic microbiome (bacteria and fungi). The 16S rRNA and ITS amplicon sequencing of V5-V7 and ITS1 variable regions of bacteria and fungi resulted in a total of 4.547 and 4.523 million raw reads, respectively. After quality control and chimera removal, 3.285 and 4.228 million effective tags of bacteria and fungi were obtained, with an average of 0.091 and 0.117 million effective tags per sample of bacteria and fungi, respectively. These effective tags were then clustered into 29909 bacterial and 16952 fungal OTUs for taxonomic annotation, with an average of 830 bacterial and 470 fungal OTUs per sample ( Supplementary Tables S2, S3 ). The obtained OTUs were used to calculate the alpha diversity indices (Chao-1, Shannon, Simpson, and PD-whole-tree indices) and the beta diversity (changes in the structure) of the bacterial and fungal communities in different samples. A significant difference was observed in the alpha diversity indices of bacterial and fungal communities among the samples under different treatments ( Figures 4A, B ; Tukey-HSD, p < 0.05). The values of alpha diversity indices for bacterial communities under treatment T2 (T2-S) were significantly decreased as compared to other treatments ( Figure 4A ; Tukey-HSD, p < 0.05), while for fungal community values of alpha diversity indices were increased dramatically under treatment T2 (T2-S) than other treatments ( Figure 4B ; Tukey-HSD, p < 0.05). PCoA based on Bray–Curti’s dissimilarity matrix and PERMANOVA (Adonis) demonstrated a significant difference in the structure of bacterial and fungal communities under different treatments. The first two components of PCoA explained 77.10% (PERMANOVA, p = 0.001 and R² = 0.8259) and 60.74% (PERMANOVA, p = 0.001 and R² = 0.7257) of the variation in the structure of the bacterial and fungal communities, respectively under different treatments ( Figures 4C, D ; Supplementary Table S4 ). First, we generated Upset plots to show the shared and unique bacterial and fungal OTUs among samples (roots and stems) under different treatments ( Supplementary Figures S2A–C ). The flower Venn diagram also revealed 303 bacterial and 92 fungal shared OTUs among the samples (roots and stems) under the different treatments. Further OTUs analysis revealed that 397, 51, 73, 130, 88, and 13 bacterial and 68, 22, 66, 150, 20, and 158 fungal OTUs were presented as unique OTUs in CK-R, T1-R, T2-R, CK-S, T1-S, and T2-S, respectively ( Figures 4E, F ). These results suggested that the variations in the diversity and structure of the bacterial and fungal communities might be due to these common and shared OTUs.

Analysis of the microbial community composition across different treatments revealed significant variations in the relative abundances of major bacterial and fungal communities at the phylum, class, and family levels in the samples (root and stem) (ANOVA, p-adjusted < 0.0001; Figures 5 , 6 ). We detected the top 10 bacterial and fungal phyla in the samples (root and stem) under different treatments and their average relative abundance (RA) in each sample is shown by taxon bar plots ( Figures 5A, B ). Bacterial phylum, such as Proteobacteria, Actinobacteriota, Chloroflexi, and Acidobacteriota and fungal phylum, including Ascomycota and Basidiomycota were the dominant microbial phyla in all the samples with an average RA > 1% across the treatments. Phylum Proteobacteria and Actinobacteriota accounted for 81.63% and 11.45% of the host total bacteriome, respectively ( Figure 5A ), and the phylum Ascomycota and Basidiomycota described 86.38% and 12.82% of host total fungal communities, respectively ( Figure 5B ). Further box plots were constructed to observe the overall variations in microbial community compositions of the most abundant bacterial and fungal phyla among the treatments. Phylum Proteobacteria was present in significantly high RA in T1-R, T2-R, and T2-S compared to CK-R, CK-S, and T1-S (Tukey-HSD, p < 0.05). The RA of phyla Actinobacteriota, Chloroflexi, and Armatimonadota was significantly higher in CK-S, and Acidobacteriota was the dominated bacterial phyla in CK-R and T1-S than other treatments (Tukey-HSD, p < 0.05; Figure 5C ). In the fungal communities, the RA of phylum Ascomycota varied significantly among treatments (p < 0.05) and was present in significantly high RA in T1-R, T1-S, and T2-S samples compared to others CK-R, T2-R, and CK-S. The RA of phylum Basidiomycota was substantially higher in T2-R than others, and phylum Mortierellomycota was present in significantly higher RA in CK-R than others (Tukey-HSD, p < 0.05; Figure 5D ).

At the class and family levels, the taxonomic distribution patterns of bacterial and fungal communities became more apparent (Tukey-HSD, p < 0.05; Figure 6 ). The RA of the top 10 most abundant bacterial and fungal communities at class and family levels are shown by bubble plots ( Figures 6A–D ). Among the bacterial communities, Gammaproteobacteria and Alphaproteobacteria were the highly abundant classes, and Rhodanobacteraceae and Burkholderiaceae were the most abundant families in all samples across the treatments. The class Gammaproteobacteria was present in significantly high RA in T1-R, T2-R, and T2-S as compared to other samples and class Alphaproteobacteria was significantly dominated in CK-S than in the other samples (Tukey-HSD, p < 0.05; Figure 6E ). The RA of the family Rhodanobacteraceae increased considerably in T2-S followed by T2-R compared with CK-R, CK-S, T1-R, and T1-S, and the family Burkholderiaceae was present at a high RA in T1-R than in the other samples (Tukey-HSD, p < 0.05; Figure 6F ). Among the fungal community’s classes, Sordariomycetes and Dothideomycetes and the families Nectriaceae and Diaporthaceae were present at high RA in all samples among the treatments. The class Sordariomycetes had significantly higher RA in T1-S than in the other samples, and Sordariomycetes exhibited in significantly greater RA in CK-S and T2-S than in the other samples under different treatments. The family Nectriaceae was more dominant in the T1-R sample than other samples, and the family Diaporthaceae was more abundant in the T1-S sample than in the other samples across the treatments (Tukey-HSD, p < 0.05; Figures 6G, H ). Analysis of the bacterial community composition at the family level revealed that the application of biocontrol agent L. pinisoli DP2-30 significantly increased the RA of the family Rhodanobacteraceae in the roots and stems (T2-R and T2-S) of pine seedlings, as L. pinisoli belongs to the family Rhodanobacteraceae.

To further explore the impact of different treatments on the host microbiome, we assessed the intra-kingdom co‐occurrence patterns of the bacterial and fungal communities in the different components (root and stem) of the host plants ( Figures 7 , 8 ). For bacterial communities, the control group (CK) exhibited a complex co‐occurrence network; however, the complexity of the bacterial co‐occurrence network was reduced significantly from the root to stem in CK and T2 than in T1, in which the network complexity increased ( Figure 7 ). For fungal communities treatment T2 showed a more complex co‐occurrence network as compared to CK and T1, and network complexity increased from the root to stem ( Figure 8 ). The modularity of all networks was calculated to be >0.4, which suggests a modular network structure of bacterial and fungal communities ( Figures 7 , 8 ). The bacterial co‐occurrence network of CK-R and T1-S was divided into 3 modules, which showed a more complex network with a maximum number of nodes (CK-R; 541, T1-S; 307), edges (CK-R; 9444, T1-S; 3557), and average degree (CK-R; 34.913, T1-S; 23.173). In contrast, the networks of T1-R (nodes; 264, edges; 2719, and average degree; 20.598), T2-R (nodes; 297, edges; 3401, and average degree; 22.902), CK-S (nodes; 331, edges; 3453, and average degree; 20.864), and T2-S (nodes; 181, edges; 1081, and average degree; 11.945) were divided into 4, 4, 5, and 6 modules, respectively which demonstrated a decrease in the network complexity. The number of nodes decreased in CK and T2 from the root to stem but increased in T1 from the root to stem. A higher number of positive edges were observed in T1-S (2670) than in CK-S (1802) and T2-S (652), which increased in the order T1-S > CK-S > T2-S ( Figure 7 ). In contrast, the complexity of the fungal co‐occurrence network was increased from root to stem under different treatments. A highly complex microbial co-occurrence network was observed for fungal communities in T2-S S (nodes; 260, edges; 2163, average degree; 16.431, and module; 4) and CK-S (nodes; 243, edges; 1821, average degree; 14.987, and module; 4). In comparison, the networks of CK-R (nodes; 165, edges; 855, average degree; 10.363), T1-R (nodes; 148, edges; 794, average degree; 10.730), T2-R (nodes; 162, edges; 880, average degree; 10.864), and T1-S (nodes; 164, edges; 873, average degree; 10.646) were divided into 5 modules, which showed less complexity in co-occurrence network. In the stem, the number of positive edges decreased from T2-S (1104) to CK-S (1034) to T1-R (525), while in root, the number of positive edges decreased in order CK-R (455) < T1-R (562) < T2-R (578) ( Figure 8 ). These results suggest that the application of L. pinisoli DP2-30 significantly reduced the positive correlation between the PWN and bacterial communities, which resulted in low disease development. The decrease in complexity of the T2 bacterial co‐occurrence network might be due to the strong colonization L. pinisoli DP2-30 ability in the host plant, as it significantly enhances the RA of Rhodanobacteraceae family in the host plant. Moreover, L. pinisoli DP2-30 make the fungal network more complex by recruiting more fungal communities that plays an important role in disease suppression.