In this study, a biocontrol strain MY-J3 with strong antagonistic activity against X. fragariae was isolated from healthy strawberry plants in Mayi Village, Daibu Town, Qujing City, Yunnan Province, and characterized through morphological observation, physiological and biochemical profiling, and molecular identification. Colony morphology was examined on Nutrient Agar (NA), and cellular features were assessed via Gram staining and spore staining under a light microscope. Physiological profiling was performed using the Biolog GEN III system. For molecular identification, a single colony was inoculated into Nutrient Broth (NB) and cultured for 48 hours. Genomic DNA was extracted and used as a template for PCR amplification of the 16S rRNA and gyrA genes. The amplicons were sequenced by Beijing Tsingke Biotech Co., Ltd. Kunming Branch. (http://Tsingke.com.cn), and resulting sequences were aligned against the NCBI GenBank database using BLAST. Phylogenetic analysis was conducted with MEGA 11 (Tamura et al., 2021).

A compatibility assay was performed between P. polymyxa MY-J3 and three pre-selected antagonistic strains—Lysobacter antibioticus HY (isolated from konjac), Pantoea ananatis XP-1 (isolated from rice), and Pseudomonas mediterranea YX5-4 (isolated from strawberry)—all of which had demonstrated in vitro efficacy against X. fragariae YM2 and proven effective in controlling ALS under greenhouse conditions. The four strains were individually cultured in King’s B broth at 28 °C for 48 hours. Subsequently, 100 µL of each suspension was spread on a KB agar plate, while 200 µL aliquots from each of the other three strains were added into separate Oxford cups placed on the same plate. Following 48 hours of incubation at 28 °C, presence or absence of inhibition zones between strains was assessed.

The susceptible strawberry cultivar ‘Monterey’ was used in greenhouse trials. The pot experiment was conducted in a controlled greenhouse environment with regulated temperature and irrigation. Ambient conditions were maintained at 80–90% relative humidity and 20–30 °C. X. fragariae was inoculated into WBN medium and cultured at 28 °C with shaking at 160 rpm for 48 hours to obtain the bacterial suspension. The suspension concentration was adjusted to an OD₆₀₀ value of 0.5, followed by spray inoculation onto the abaxial surface of strawberry leaves until runoff occurred. Nine treatments were applied in triplicate, including a conventional chemical control (bromothalonil · bromonitrol, 1000× dilution). Each biocontrol bacterium was cultured separately in shake flasks at 28 °C and 160 rpm for 3 days to obtain fermentation broth. The concentration of all fermentation broths was uniformly adjusted to an OD₆₀₀ value of 2.0 prior to dilution. For single-strain treatments, the fermentation broth was diluted 20-fold. For synthetic consortium treatments, strains were first mixed at a 1:1 (v/v) ratio and then diluted 20-fold. Treatment details are provided in Table 1. All treatments were applied by spraying on the second day after pathogen inoculation, followed by subsequent applications at two-day intervals for a total of three sprays. The control group was sprayed with an equivalent amount of sterile water. ALS incidence was assessed 15 days after the final pathogen inoculation. Disease severity was graded according to Supplementary Table 1. The disease index (DI) and relative control effect (RCE) were calculated as follows:Disease Index (DI)= 100 × ∑ (Number of diseased leaves per grade × Grade value)/(Total leaves surveyed × Maximum grade value); Relative Control Effect (RCE, %) = (DI – control – DI – treatment)/DI – control × 100.

Field trials were conducted at Lujiajing, Daibu Town, Huize County, Qujing City, Yunnan Province, China (geographical coordinates: 26.223°N, 103.436°E, WGS84 datum). The trials were carried out in a greenhouse cultivation area under drip irrigation during the summer of 2024 (June to August). The experimental period was warm, humid, and rainy, with average temperatures ranging from 17 °C to 20 °C and relative humidity between 75% and 85%. The field trial was conducted using the strawberry cultivar ‘Monterey’ in plots with uniformly distributed incidence of ALS. Five treatments were evaluated, each with three replicates, totaling 15 plots. Each raised bed served as a replicate, containing 4 rows of strawberry plants, with each row comprising no fewer than 200 plants. Constructed bacterial consortia were applied as the main treatments, with a conventional chemical control (bromothalonil · bromonitrol, 1000× dilution) included for comparison (Table 2). The concentration of all fermentation broths was uniformly adjusted to an OD₆₀₀ value of 2.0 prior to dilution (28 °C, 160 rpm, 3 days). The compound microbial consortia were mixed in a 1:1 (v/v) ratio and diluted 50-fold. All solutions were sprayed onto the abaxial surfaces of strawberry leaves at three-day intervals for a total of three applications, with an equivalent amount of water sprayed as the control. Disease incidence and control efficacy were assessed according to the methods described in Section 2.3.1.

A standard curve for qPCR was constructed using ten-fold serial dilutions of genomic DNA from X fragariae YM2. The logarithm of DNA concentration (log₁₀) was plotted against Ct values to generate a regression equation for pathogen quantification. Method validity was evaluated based on the correlation coefficient (R²), slope, and amplification efficiency (E), calculated as: E = 10 ^(−1/slope) − 1.

Leaf samples were collected 10 days post-treatment in the greenhouse. Total DNA was extracted using the SteadyPure Plant Genomic DNA Extraction Kit. Pathogen-specific primers XopAF-F/R were used for qPCR analysis, and bacterial abundance in leaf tissues was quantified by interpolating Ct values into the standard curve established in Section 2.4.1.

Total RNA was extracted from liquid liquid nitrogen-snap-frozen leaves collected 10 days after consortia treatment and reverse-transcribed into cDNA. Using FaGAPDH2 as a reference, expression of defense-related genes (FaSnRK2, AOS, FaBG2-3, FaPR1, FaPAL, FaChi3, FaJAR1, FaEDS1) was quantified via qPCR (Xu, 2019). Primer sequences are listed in Supplementary Table 2. Reactions were performed on a Roche LightCycler^® 96 using a two-step protocol, with relative expression calculated by the 2^(–ΔΔCt) method.

Total RNA was extracted from liquid liquid nitrogen-snap-frozen leaves collected 10 days after consortia treatment and reverse-transcribed into cDNA. Expression of virulence genes (xopR, Hpa2, flgG, fliA, rpfE, gumG, rtxD, rtxE) was analyzed using the X. fragariae housekeeping gene gyrB as an internal control. Primer sequences are provided in Supplementary Table 3. Methods followed those described in Section 2.5.1.

Biofilm formation was assessed using the crystal violet staining method (Jiu et al., 2020). Bacterial strains were cultured individually in KB medium for 48 h, then mixed in equal proportions according to the consortium treatment groups and adjusted to OD₆₀₀ ≈ 0.5. A 100 μL aliquot of each bacterial suspension was added to 96-well plates containing 100 μL of fresh KB medium per well. Uninoculated KB medium served as the blank control. All treatments were performed in triplicate and incubated at 28 °C for 48 hours.

Siderophore production was evaluated with the chrome azurol S (CAS) assay (Sun et al., 2011). A 1% (v/v) inoculum of each bacterial consortium was cultured in MKB liquid medium for 72 hours. The supernatant was collected by centrifugation at 8000 ×g for 10 min and mixed with an equal volume of CAS detection solution. After 1 hours of reaction in the dark, the absorbance at 630 nm (A_s) was measured. The reference (A_r) was prepared using uninoculated MKB medium treated identically. Siderophore activity units (SU) were calculated as:SU = (A_r − A_s)/A_r × 100%.

Cellulose decomposition ability was qualitatively determined using Congo red staining (Liao et al., 2014). After centrifugation and concentration, the bacterial consortium was inoculated onto carboxymethyl cellulose sodium (CMC-Na) plates and incubated at 28 °C for 48 hours. Plates were stained with 1 mg/mL Congo red for 10–15 min, then destained with 1 mol/L NaCl. The formation of a hydrolysis zone indicated cellulose degradation. The hydrolysis capacity (H) was evaluated as the ratio of the hydrolysis zone diameter (D) to the colony diameter (d): H = D/d.

P. polymyxa MY-J3 (M) and L. antibioticus HY (H) were inoculated into their respective optimal media (KB medium for MY-J3 and YH medium for HY) and cultured under shaking conditions for 3 days. Using an equal volume of uninoculated medium as control, the cultures were mixed at a 1:1 volume ratio and extracted three times with an equal volume of ethyl acetate by vigorous shaking. The upper ethyl acetate phase was collected via a separatory funnel and concentrated using a rotary evaporator, yielding a brown oily crude extract. After weighing, the extract was dissolved in a defined volume of dimethyl sulfoxide (DMSO), filtered through a 0.22 μm membrane, and stored in amber vials for subsequent use. Since no extract was obtained from the uninoculated medium control, an equal volume of DMSO was used as the solvent control in subsequent bioassays to exclude any effects of the extraction solvent.

The crude metabolites (M+H) were added to 50 mL suspensions of X. fragariae YM2 to achieve final concentrations of 5, 10, 20, 40, 80, 120, and 160 μg/mL. A control was prepared using an equal volume of 1% DMSO. Bacterial growth was monitored by measuring OD₆₀₀ at 0, 12, 24, 36, 48, 60, and 72 h. The minimum inhibitory concentration (MIC) was determined as the lowest concentration showing no visible growth. All treatments were performed in triplicate.

Pathogenicity was assessed by calculating the disease index following treatment with the crude metabolites. X. fragariae YM2 was cultured in WBN medium for 24 hours, then treated with crude metabolites at 2×MIC, MIC, and 1/2MIC concentrations. A control received 1% DMSO. After incubation (28 °C, 160 rpm, 24 h), cells were harvested, washed with PBS, and resuspended to OD₆₀₀ ≈ 0.5. Strawberry leaves were sprayed with the suspensions, and the disease index was recorded 10 days post-inoculation. Each treatment included three replicates with 10 leaves per replicate.

Biofilm biomass was quantified using crystal violet staining (Wang et al., 2021). X. fragariae YM2 suspensions were treated with crude metabolites at concentrations of 5–160 μg/mL (with 1% DMSO control) in 96-well plates and incubated at 28 °C for 48 hours. Cell surface hydrophobicity (CSH) was determined using the microbial adhesion to hydrocarbons (MATH) method (Danchik and Casadevall, 2020). Bacteria were treated with crude metabolites at 1/2MIC, MIC, and 2×MIC, incubated for 12–48 h, washed, and resuspended in PBS. A 5 mL aliquot was mixed with n-hexane, vortexed, and allowed to separate. The aqueous phase absorbance (OD₆₀₀) was measured. CSH was calculated as:CSH (%) = [1 − (OD₆₀₀ − treatment/OD₆₀₀ − control)] × 100.

EPS production was measured using ethanol precipitation (Zhang, 2011). X. fragariae YM2 was treated with crude extract at 1/2MIC, MIC, and 2×MIC (1% DMSO control) and cultured for 5 days (28 °C, 160 rpm). A 10 mL sample was mixed with 20 mL absolute ethanol, stored at −20 °C overnight, and centrifuged at 4,000 ×g. The precipitate was dried and weighed to determine crude EPS yield.

Through systematic isolation and screening, a potent antagonistic strain MY-J3 was obtained from healthy strawberry tissues, demonstrating remarkable inhibitory activity against ALS pathogen X. fragariae YM2. The inhibition zone diameter produced by MY-J3 reached 2.70 cm, significantly surpassing other isolates. The strain was identified as a Gram-positive, spore-forming rod-shaped bacterium (Figures 1A–C). Based on 16S rRNA and gyrA gene sequence analysis, it was classified as P. polymyxa with a bootstrap value of 99% in the phylogenetic tree (Figures 1D, E). Physiological profiling using the Biolog GEN III system showed that MY-J3 could utilize 64 carbon sources (Supplementary Table 4) and exhibited strong metabolic reducing capacity. The 16S rRNA and gyrA gene sequences of strain MY-J3 have been submitted to GenBank with accession numbers PP732390.1 and PX676550, respectively.

To ensure consortium stability, we systematically evaluated the compatibility between MY-J3 and three preserved antagonistic strains (L. antibioticus HY (H), P. mediterranea YX5-4 (Y), P. ananatis XP-1 (X)). Dual verification through cross-streaking and oxford cup assays demonstrated complete absence of inhibition zones in all strain combinations, with no inhibitory phenomena observed at colony junctions (Supplementary Figure 1), indicating excellent ecological niche compatibility. Comparison of two consortium construction strategies revealed that the “mix-then-ferment” approach (Figure 2A) consistently outperformed the “ferment-then-mix” method (Figure 2B). The M+H consortium (MY-J3 + HY) constructed using the “mix-then-ferment” strategy produced an inhibition zone of 3.53 cm, significantly larger than those of individual strains MY-J3 (2.47 cm) and HY (2.53 cm), demonstrating clear synergistic effects. Among other combinations, M+Y (5.20 cm) and M+H+Y+X (5.30 cm) showed prominent antibacterial activity, while the M+X combination exhibited relatively weaker effects. Although the M+Y and M+H+Y+X consortia produced larger inhibition zones in vitro, the M+H consortium was selected for subsequent mechanistic studies because it showed superior and more consistent disease control efficacy in field trials, coupled with the lowest pathogen load in planta as quantified by qPCR. These systematic in vitro evaluations provide experimental evidence for selecting the M+H consortium for further mechanistic investigation.

Following confirmation of in vitro synergy, we evaluated the practical efficacy of the M+H consortium against ALS under both greenhouse and field environments. Comparison of nine treatment schemes revealed that all treatments significantly reduced disease incidence (Supplementary Figure 2). Among them, the M+H consortium exhibited the optimal control efficacy in greenhouse trials, achieving a relative control effect of 76.15%, significantly higher than individual strain treatments M (58.33%) and H (60.00%) (Table 3) (P<0.05). The four-strain combination M+H+Y+X provided 77.69% control efficacy, showing no significant difference from M+H. Notably, the efficacy of M+H treatment significantly surpassed that of conventional chemical agent bromonitrol (60.66%). In field validation trials, the M+H treatment similarly exhibited excellent disease control performance, with a relative control efficacy of 74.26%. Under consistent conditions with biocontrol application as the only variable, the M+H consortium significantly outperformed the chemical agent bromonitrol NY, showing a higher yield increase (78.40% vs. 50.94%) and control efficacy (P < 0.05) (Table 4). It is therefore hypothesized that the yield improvement stems from the plant growth-promoting properties of the biocontrol agents, and that disease control further enhances strawberry quality. These results fully demonstrate that the M+H consortium possesses stable and reliable disease control capacity under practical application conditions.

To elucidate the physiological basis of disease control, we quantified pathogen load in plant tissues using a specific qPCR assay. A standard curve was established with the regression equation Ct = -3.2193 log_[DNA] + 35.349 (R² = 0.9993) and amplification efficiency of 104%, ensuring accurate quantification (Supplementary Figure 3). Results showed that 10 days post-inoculation, X. fragariae YM2 load in control plants reached 1.58×10⁵ fg/μL. In contrast, the M+H treatment group exhibited only 9.12×10¹ fg/μL, representing a 99.99% reduction compared to the control and demonstrating the most significant suppression (Figure 3) (P<0.05). This molecular evidence confirms that the M+H consortium effectively inhibits pathogen colonization and spread in strawberry plants. The significant positive correlation between pathogen load and disease severity further demonstrates that the consortium controls disease through direct suppression of pathogen growth.

qPCR results demonstrated that compared to the control and individual strain treatments, the M+H treatment significantly upregulated the transcriptional levels of a series of core defense-related genes (Figure 4) (P<0.05). Specifically, the expression of FaBG2-3, a β-1,3-glucanase gene associated with cell wall reinforcement, was upregulated by 2.40-fold, the pathogenesis-related protein gene FaPR1 was upregulated by 2.32-fold, phenylalanine ammonia-lyase gene FaPAL by 2.31-fold, and chitinase gene FaChi by 3.63-fold, all significantly exceeding the levels observed in individual strain treatments. Notably, in the four-strain combination M+H+Y+X treatment, the upregulation of FaPR1, FaPAL, and FaChi3 expression reached maximum levels of 13.27-fold, 29.07-fold, and 10.41-fold compared to the control group, respectively, indicating that the multi-strain combination has a synergistic enhancing effect on inducing plant defense gene expression. The coordinated upregulation of these PR proteins demonstrates that the M+H consortium successfully triggers the plant’s basal immune response, establishing a systemic disease-resistant state.

To understand the initiation mechanisms of immune signaling, we systematically analyzed the expression patterns of marker genes from different hormone pathways (Figure 4) (P<0.05). The study found that AOS, a key gene in the jasmonic acid pathway, was significantly induced in all consortium treatments, with the M+H treatment group showing the induction amplitude of 5.08-fold compared to the control. Particularly noteworthy, in the M+H treatment, the key ABA signaling receptor gene FaSnRK2 showed specific upregulation, with expression levels 1.42-fold higher than the control, while in other treatment groups, the expression of this gene was lower than the control. In contrast, the expression pattern of FaEDS1, an important component of the salicylic acid pathway, varied among treatments: upregulated by 1.38-fold in the M treatment, 1.64-fold in the M+Y treatment, but downregulated in the M+H treatments. Another key jasmonic acid pathway gene, FaJAR1, showed significant down-regulation in the M+H treatment. Our results suggest that the M+H consortium fine-tunes the JA pathway, as indicated by the concurrent strong upregulation of the biosynthetic gene AOS and downregulation of the signaling gene FaJAR1.

We first evaluated key functional traits of the consortium itself. The M + H consortium demonstrated the strongest biofilm production, with an OD₅₉₀ value of 0.7091, representing 2.18-fold and 1.41-fold increases over the H and M monocultures, respectively (Figure 5A) (P<0.05), indicating a significantly enhanced ability to colonize plant surfaces and form stable microenvironments. Meanwhile, Siderophore content in the fermentation broth was quantified using the CAS assay. The M+H consortium yielded significantly higher siderophore levels than the H or M monocultures, exceeding them by 3.66-fold and 1.47-fold, respectively (Figures 5B, D) (P<0.05), revealing a substantially improved capacity to compete with the pathogen for iron. Furthermore, assessment of cellulose degradation ability using Congo red staining revealed that the M+H consortium formed a hydrolysis zone-to-colony diameter ratio (H/d) of 3.29 on CMC-Na plates, significantly higher than those of the individual strains M (2.81) and H (2.77) (Figures 5C, E) (P<0.05), demonstrating enhanced substrate degradation capability and environmental adaptability. This cellulolytic activity was only detectable under in vitro conditions with highly concentrated cells. Crucially, no leaf damage was observed on strawberry plants treated with the 20-fold diluted fermentation broth, indicating minimal risk under practical application. These enhanced physiological traits collectively improve the consortium’s competitive survival on plant surfaces. This may directly or indirectly inhibit pathogen growth and colonization, and could also contribute to the activation of the plant’s immune system.

We further extracted secondary metabolites from the M+H fermentation broth. Growth curve assays demonstrated that the crude metabolites exhibited significant concentration-dependent inhibition against X. fragariae, with complete suppression of pathogen growth at 80 μg/mL, determined as the minimum inhibitory concentration (MIC) (Figure 6A). At concentrations of 120 μg/mL and 160 μg/mL, the pathogen failed to grow throughout the 72-hour incubation period. More importantly, in vivo pathogenicity assays revealed that when the pathogen was treated with MIC (80 μg/mL), and 2×MIC (160 μg/mL) concentrations of the crude metabolites, the resulting d X. fragariae isease indices after inoculation onto strawberry plants significantly decreased to 16.67 and 9.45, respectively (control: 43.89) (Figures 6B, C) (P<0.05), demonstrating that the crude metabolites effectively impair the pathogen’s pathogenic capability with a clear dose-response relationship.

The crude metabolites markedly inhibited biofilm formation and altered cell surface hydrophobicity (CSH) of X. fragariae. Biofilm biomass decreased with increasing metabolite concentration, with an OD₅₉₀ of 0.70 at MIC compared to 1.89 in the control (Figure 6D) (P<0.05). CSH values were 50.0%, 16.7%, and 14.3% after treatment with 1/2MIC, MIC, and 2×MIC, respectively (Figure 6E) (P<0.05), indicating reduced surface hydrophobicity and potential impairment of host colonization. Furthermore, treatment with the crude metabolites significantly reduced the production of extracellular polysaccharides (EPS), EPS yields decreased to 75% and 68% of the control level at 1/2MIC and MIC, respectively. The 2×MIC treatment resulted in the most substantial reduction, confirming the dose-dependent inhibitory effect on EPS synthesis (Figure 6F) (P<0.05).

After 10 days of treatment, the expression of virulence-related genes in X. fragariae was analyzed. The results demonstrated that most consortia treatments downregulated virulence gene expression.In single-strain treatments, upregulation was observed only in fliA and flgG (flagellar motility genes) after YX5–4 treatment, and in rpfE (quorum sensing), gumG (exopolysaccharide synthesis), and Hpa2 (T3SS) after HY treatment, compared to the CK group. All other treatments, particularly the synthetic consortia, resulted in downregulation of these genes (Figure 7) (P<0.05).These findings suggest that biocontrol agents can impair motility, expansibility, and pathogenicity of X. fragariae in plant tissues. Overall, consortia treatments consistently suppressed the expression of pathogenicity-related genes, with single-agent treatments of XP-1 and MY-J3 also showing significant downregulatory effects.

It is noteworthy that these molecular-level changes are highly consistent with the phenotypic results we observed—the inhibition of virulence gene expression directly explains the reduced biofilm formation, decreased cell surface hydrophobicity, and diminished EPS production. This coordinated regulation of phenotypes and gene expression indicates that the M+H consortium comprehensively impairs the virulence of X. fragariae by simultaneously interfering with multiple pathways, including attachment capacity, motility, production of pathogenicity factors, and cell-cell communication. This multi-target mechanism of action may represent an important molecular basis for the exceptional disease control efficacy demonstrated by the M+H consortium in both greenhouse and field conditions.

The 16S rRNA gene amplicon sequencing of nine samples from three treatment groups was performed on the Illumina MiSeq platform. After filtering raw reads and removing chimeras, the average proportion of high-quality Tags reached 92.24%, with low variability among triplicate samples per group, indicating that the sequencing data were suitable for downstream analysis (Table 5). Rarefaction curves were used to assess the adequacy of sequencing depth and indirectly reflect species richness across samples. The curves for all three treatment groups rose steeply initially and then plateaued, demonstrating sufficient sequencing coverage and confirming that the data met analytical requirements (Figure 8).

High-throughput sequencing of the 16S rRNA gene V3–V4 region from field-grown strawberry leaves demonstrated that the M+H treatment significantly altered the structure of the phyllosphere bacterial community. Principal coordinates analysis (PCoA) revealed clear separation between the treatment and control groups (Figure 9A), indicating substantial structural shifts in community composition. Alpha diversity analysis further confirmed that the M+H-treated group exhibited significantly higher Chao1 (1250.43), Sobs (1146.00), and Shannon (1.82) indices, along with a significantly lower Simpson index, compared to the control (Figure 10) (P<0.05). These results demonstrate that the application of M+H significantly increased both species richness and evenness of the phyllosphere bacterial community. Venn diagram analysis showed that the M+H treatment group contained 1224 operational taxonomic units (OTUs), including 736 unique OTUs, while the control group had only 599 OTUs with 318 unique OTUs (Figure 9B), further supporting that the introduced consortium substantially enriched the indigenous bacteria microbial resources.

After quality control, the obtained sequences met the requirements for downstream analysis. Comparative analysis of the top 12 most abundant bacterial phyla across samples revealed similar taxonomic profiles among treatments, though with variations in relative abundance.In treatments T1(M+H) and T2 (M+H+Y+X), Cyanobacteria (which may include chloroplast sequences from the host plant) and Proteobacteria were the dominant phyla, with relative abundances of 70.62% and 27.20% in T1, and 64.57% and 30.86% in T2, respectively. Compared to the CK group, T1 and T2 showed a significant increase in Cyanobacteria and a decrease in Proteobacteria. Minor phyla (<1% abundance) were slightly enriched in both treatments relative to CK (Figure 11A). The increase in Cyanobacteria relative abundance in T1 and T2 suggests an enrichment of potentially beneficial taxa, potentially enhancing plant growth and disease resistance. At the genus level, Methylobacterium, Prevotella, Cossackia, and Halomonas showed increased abundance. The consortium treatments may have enriched genera such as Sphingomonas, Cossackia, and Halomonas, which are associated with toxin degradation, pathogen cell wall lysis, and nitrogen fixation, thereby improving plant stress resistance. Notably, the relative abundance of Xanthomonas was significantly reduced in T1 and T2, accounting for only 3.20% and 0.28% of that in the CK group, respectively (Figure 11B). These findings indicate that the M+H consortium did not disrupt the native ecology but instead steered the phyllosphere microbiome toward a healthier, more stable, and pathogen-suppressive state.

Notably, compared to the control group (CK), the overall abundance of Proteobacteria in the treatment groups (T1, T2) showed a decreasing trend, and the exogenously applied Firmicutes biocontrol agent was also not significantly detected at the phylum level. This pattern reflects the distinct ecological dynamics at the 21-day sampling point. In the control, explosive growth of the pathogen X. fragariae (Proteobacteria) dominated the community, leading to high Proteobacteria abundance (Figure 11A). In the treatment groups, after achieving early pathogen suppression, the populations of the exogenous biocontrol agents (including MY-J3 from Firmicutes and HY from Proteobacteria) naturally declined to low levels due to ecological competition. Simultaneously, through ecological regulation, they reshaped a microbial network centered on indigenous beneficial microbiota. This network continuously suppressed pathogen resurgence, resulting in a reduction in pathogen biomass that far exceeded the transient colonization contribution of the exogenous agents. Consequently, this led to a net decrease in Proteobacteria abundance and an overall optimization of the community structure in the treatment groups. These findings reveal the ecological principle of biocontrol treatment achieving sustained disease suppression through a “priming-remodeling” mechanism.