Inhibition of Citrus Huanglongbing Disease by Paenibacillus polymyx KN-03 and Analysis with Transcriptome and Microflora

Abstract

Soil drench treatment using Paenibacillus polymyxa strain KN-03 was applied to citrus plants infected with Candidatus Liberibacter asiaticus (CLas). The infection status was assessed using PCR and a real-time quantitative PCR detection system (qPCR). The application of KN-03 resulted in a notable reduction in CLas levels in citrus plants. Specifically, by the 257th day post treatment commencement, following 24 KN-03 applications, the negative rates of CLas in the vein, root tip, and shoot tip were 50%, 0%, and 50%, respectively. After 24 cycles, KN-03 application significantly enhanced plant growth and stimulated reactive oxygen production in citrus leaves compared to control plants. Transcriptome analysis identified specific upregulated pathways. Furthermore, flora analysis revealed an increased abundance of microorganisms possessing potential utilization value, including Burkholderia-Caballeronia-Paraburkholderia, uncultured_bacterium_o_Acidobacteriales, uncultured_bacterium_f_Gemmatimonadaceae, and Rhodanobacter, in the root zone. Moreover, the BugBase analysis indicated that KN-03 treatment increased the abundance of beneficial rhizosphere bacteria associated with biofilm formation, element mobilization, and stress tolerance. These findings support the utility of Paenibacillus polymyxa KN-03 as an effective plant-growth-promoting bacterium for CLas management, with additional benefits for plant growth and soil health, specifically offering detoxification resources for shoot tip grafting.

1. Introduction

Huanglongbing (HLB), also referred to as citrus greening, is a formidable and highly contagious ailment that affects citrus crops. This malady has substantial ramifications for the global citrus production arena, affecting more than 50 countries and regions [1]. The causal agent of HLB is Candidatus Liberibacter asiaticus (CLas) [2], an uncultivable bacterium that demonstrates varying degrees of infection within primary cultivars. In citrus plants afflicted by CLas, observable symptoms include yellowing of upper branches, yellow and mottled leaves, and the emergence of “red nose fruit”. This ailment disrupts metabolic processes related to nutrient assimilation, resulting in the accumulation of starch in the phloem, impeding the transportation of essential nutrients, and culminating in eventual desiccation and demise [3,4]. At present, efficacious strategies for managing HLB are limited to the removal of CLas-infected trees; application of insecticides to eradicate citrus psyllids, which serve as vectors for CLas; and cultivation of virus-free plants in China.

1.1. Paenibacillus polymyxa Review

In the context of addressing HLB, numerous strains of Paenibacillus polymyxa (P. polymyxa), commonly referred to as plant-growth-promoting bacteria, have exhibited promise as potent promoters of plant growth. These bacteria are typically endophytically encountered within the root tissues of plants [5]. Genomic analysis of P. polymyxa strains revealed that approximately 12% of their genetic makeup comprises gene clusters responsible for synthesizing biologically active metabolites. These metabolites include antagonistic proteins, polypeptide antibiotics, and plant growth regulators [6,7]. Paenibacillus species influence crop growth and vitality through various mechanisms, including nitrogen fixation, phosphate solubilization, production of plant hormones such as indole-3-acetic acid auxin and isopentenyl adenine cytokinin, siderophore-mediated iron chelation, and the synthesis of antibacterial lipopeptides such as fusaricidin and polymycin. Additionally, they can induce a plant defense response known as induced systemic resistance (ISR). Other beneficial metabolites produced by these bacteria include exopolysaccharides and enzymes [8].

Further examination has indicated that certain strains of P. polymyxa confer benefits to soil microbiomes. The introduction of P. polymyxa into the soil results in an augmentation of bacterial populations while concurrently reducing the number and diversity of fungi, including potentially pathogenic species such as Rhizoctonia sp. This is accompanied by an increase in the poplar yield [8]. P. polymyxa strain E681 is capable of emitting bacterial volatile compounds, suppressing plant pathogens, and bolstering the immune system of plants [9]. Moreover, it provides assistance in mitigating abiotic stressors such as drought, salinity, and cold [10]. The NCBI reference sequence of P. polymyx KN-03: NZ_CP036496.1 has been studied clearly. In disease resistance research, the control effect of P. polymyxa KN-03 on both deciduous and non-deciduous strains of Verticillium dahliae in Xinjiang cotton was higher than 40% under 100 times solution [11]. After optimizing the culture conditions, the lipopeptide content produced by KN03 reached 60.8 mgL⁻¹. Compared with the supernatant of the original culture, the supernatant of the P. polymyxa KN-03 culture from the optimized medium and feeding fermentation showed an enhanced antibacterial effect, with increases in the inhibition zone of 59%, 45%, and 26% against Ralstonia solanacearum, Erwinia carotovora, and Xanthomonas campestris, respectively [12]. These attributes position them as potential agents for biocontrol, biostimulation, and biofertilization, ensuring sustainable crop productivity and environmental preservation.

1.2. Research Idea on the Prevention and Control of Citrus HLB with P. polymyxa Plant-Growth-Promoting Bacterium

Currently, a specific medication for the complete eradication of citrus HLB remains elusive. Several strategies have exhibited some efficacy in agricultural practice. These include tree removal, containment of psyllids (carriers of the disease), and cultivation of pathogen-free plants. Natural adversaries such as parasitic wasps and ladybugs have been employed in the domain of biological control [13]. Physical control measures commonly used in production involve the use of insect lamps to eliminate citrus psyllids. Chemical control, on the other hand, relies on the widespread application of insecticides such as organophosphorus compounds and mineral oils in production systems. However, it is crucial to note that this approach has substantial environmental implications.

Notably, broad-spectrum antibiotics have been shown to be effective against HLB. Studies have reported that the injection of tetracycline or penicillin can partially enhance the health of trees. Nonetheless, it does not offer a complete cure for HLB disease [14,15]. Simultaneously, significant strides have been made to explore the effects of specific small-molecule antibacterial substances. In particular, benzbromarone and phloderin have proven to be effective inhibitors of LdtR activity, a pivotal transcription factor in HLB [16]. Additionally, antimicrobial peptides are promising for the prevention and treatment of HLB [17].

Polymerase chain reaction (PCR) technology has been extensively used for the diagnosis of HLB [2]. In 2009, a research team led by Duan Yongping obtained the complete genome sequence of the citrus HLB bacterium Candidatus Liberibacter asiaticus str. pys62 using metagenomics [18]. This laid the foundation for further improving detection sensitivity. Li et al. (2006) used qPCR to detect pathogen CLas, the cause of HLB disease. Their observations indicated relatively high pathogen concentrations in the leaf veins and petioles of citrus plants [19]. Zhang et al. [20] further increased the detection sensitivity by designing the primer RNRf/RNRr with characteristics based on the five-copy number traits of the CLas bacterial nrdB gene sequence. In HLB disease research, qPCR has progressively emerged as the prevailing molecular detection method.

Transcriptomics, the study of gene expression at the RNA level, considers temporal and spatial constraints. Gene expression in the same cells can vary under different growth conditions and environments. BugBase, a method for predicting functional pathway coverage within complex microbiomes and explaining bio-related phenotypes, can significantly contribute to a comprehensive analysis of KN-30 function in citrus plants. In this study, we focused on the control of HLB disease and promotion of plant and root growth using a commercial suspension formulation of Paenibacillus polymyx KN-03. This formulation includes antibacterial lipopeptides and phytohormones, thus enhancing its potential to address these critical aspects of citrus cultivation. The purpose of this study was to explore the application of Paenibacillus polymyx KN-03 as an effective plant growth promoter in the management of CLas and to explore its role in citrus growth and soil health, especially if it can provide non CLas-infected materials for shoot-tip grafting and propagation of excellent varieties.

2. Materials and Methods

2.1. Plant Materials and Bacteriotics

The plant materials used in this experiment were 2-year-old Citrus unshiu var. Youliang plants grafted onto Poncirus trifoliata. These plants were individually nurtured in the controlled environment of the Microbiology Engineering Center at the Huazhong Agricultural University. The bacteriological agent used in this investigation was P. polymyx KN-03, obtained from Wuhan Kenuo Biotechnology Co., Ltd. (Wuhan, China), at a concentration of 500 million CFU/g.

2.2. Treatment Method of P. polymyx KN-03

The experiment spanned July 2021 to March 2022. Initially, CLas was detected by PCR and qPCR in the leaf veins of “Yura” nursery plants, and only those showing positive results were selected for further investigation. Twelve positive plants were transplanted into black plastic pots with a substrate comprised of equal parts of vermiculite, peat, and garden soil. To ensure uniform growth conditions, all plants received regular application of compound fertilizer and decomposed chicken manure, with each application amounting to 10 g per plant every two months.

The active P. polymyx KN-03 aqueous suspension (AS) used in this study had a concentration of 500 million CFU/g. The experimental groups included the P. polymyx KN-03 treatment group, heat-inactivated P. polymyx KN-03 treatment group, and the control group, with each group consisting of four biological replicates. For both live KN-03 bacteria and inactivated treatments, root irrigation was conducted every ten days using a diluted primary solution containing 6 g/L water administered at a dosage of 500 mL per plant. In the control group, an equivalent volume of water was applied to each plant every 10 days. The interval between the initial and 24th treatments was 257 days.

2.3. Detection of CLas before and after Treatment with KN-03

For molecular detection of HLB disease, the Corbett RG-6000 (N15128) Designed and manufactured in Sydney, Australia and the Bio-Rad PCR instrument from the Hercules, CA, USA were employed. The primers used for HLB disease detection are listed in

Table 1. Primers used for detection of HLB disease by PCR and qPCR.

Primers	Nucleic Acid Sequence (5′–3′)	Fragment Size (bp)	GenBank Number	Sources
COX+	GTATGCCACGTCGCATTCCAGA	68	CX297817 72-93	Li et al., 2006 [19]
COX−	GCCAAAACTGCTAAGGGCATTC		CX297817 118-139
RNRf	CATGCTCCATGAAGCTACCC	80	CP010804.1 5434-5455	Zheng et al., 2016 [20]
RNRr	GGAGCATTTAACCCCACGAA		CP010804.1 5493-5514
OI1	GCGCGTATGCAATACGAGCGGCA	1167	CP001677.5 786316-786339	Jagoueix et al., 1994 [2]
OI2c	GCCTCGCGACTTCGCAACCCAT		CP001677.5 787462-787483

. DNA was extracted from each tissue sample using the CTAB method [21]. Leaf veins were specifically gathered from the third to fifth leaves at the branch tip and collected from four distinct locations on the plants. Data analysis was performed using Excel 2016 and SPSS 25.0, which included variance analysis and significance-level testing to extract meaningful insights from the results.

Genomic DNA extracted from citrus leaves was amplified using OI1/OI2c primers targeting the 16S rDNA locus. The amplification protocol included an initial pre-degeneration step at 94 °C for 5 min, followed by denaturation at 94 °C for 30 s, annealing at 61 °C for 30 s, and extension at 72 °C for 45 s, with a total of 34 cycles. The procedure was performed as a final extension step at 72 °C for 5 min. Sterilized distilled water was used as the negative control, whereas a plasmid vector containing the HLB target fragment served as the positive control.

For qPCR detection, RNR+/RNR− primers with heightened sensitivity were carefully chosen, with each sample undergoing three biological replicates. As a negative control, DNA extracted from healthy “Newhall” calli was utilized, with its 2^−∆∆CT = 1. Positive controls comprised DNA extracted from the leaf veins of sugar oranges, which had previously tested positive in our laboratory. Data analysis was performed using the 2^−∆∆CT method, and 2^−∆∆CT >1 indicated a positive result for plant samples, in accordance with [22].

2.4. Determination Method of Morphology and Biochemistry

2.4.1. Determination of the Morphological Index

Plant height, crown diameter, and stem diameter were measured in this study. After subjecting the plants to 24 treatments with KN-03, the plant height was gauged from the collar to the apex of the tree. The crown width was calculated as the average of the southeast and northwest dimensions, while the stem diameter at the top 10 cm of the grafting site was assessed using a Vernier caliper. Each experimental group included four biological replicates.

2.4.2. Detection of ROS

To detect the citrus active oxygen content, DAB (3,3-diaminodiamindiamine) of H₂O₂ was utilized for histochemical staining [23]. Randomly selected citrus leaves were cleaned and divided into treatment and control groups. The treatment group was soaked in KN-03 solution for 4 h, whereas the control group was soaked in distilled water for the same duration. Subsequently, the samples were immersed in 1 mg/mL DAB solution (pH = 3.8) and left to stain in a dark environment for 1 h until visible staining occurred on the leaves. Anhydrous ethanol was used until the green color faded, and finally, the samples were cleaned with distilled water before being pressed and photographed [24].

For NBT histochemical staining of superoxide anions, citrus leaves were collected, cleaned, and randomly divided into treatment and control groups. The treatment group was immersed in a KN-03 solution for 4 h, whereas the control group was soaked in distilled water for the same duration. NBT was dissolved in PBS buffer at pH=7.8, and the samples were soaked in a 1 mg/mL NBT solution and stained for 1 h under natural light until a blue phenotype was visible on the leaves. Anhydrous ethanol was then employed until the green color faded, followed by washing the leaves with distilled water and capturing the photographs [25].

2.4.3. Assays of Malondialdehyde (MDA) Content and Defense Enzyme Activity

The relative content of malondialdehyde (MDA) in the plants was determined using an MDA content kit (Suzhou Gerith Biotechnology Co., Ltd., Suzhou, China) after subjecting the plants to 10 rounds of KN-30 treatment. Each group consisted of four biological replicates, and the operating procedure was performed according to the manufacturer’s instructions. Mature leaves were collected from the branches of each treatment group at 0 h prior to as well as 24 h post inoculation with KN-03. Four biological replicates were used for each experimental group. The activities of resistance-related enzymes (CAT, POD, and PPO) were determined according to the guidelines outlined in [25,26]. The variance and significance level tests were performed using Excel 2016 and SPSS 25.0.

2.4.4. Detection of Phenol Oxidase Activity

Soil phenol oxidase activity was measured by taking 0.1 g of air-dried soil (< 2 mm) in a 2 mL centrifuge tube. Next, 0.75 mL of acetic acid buffer (pH 5.5) and 0.75 mL of dopamine substrate solution were added to the experimental group. In contrast, the control group involved measuring 0.1 g of air-dried soil in a 2 mL centrifuge tube with 1.5 mL of buffer. The mixtures were shaken to ensure thorough mixing, capped, and incubated for 1 h at 37 °C under flat shaking conditions (100 rpm). Subsequently, centrifugation (5000 rpm, 5 min) and filtration were performed, followed by colorimetric determination of the filtrate at 460 nm. Enzyme activity was calculated as nmol diqc h⁻¹g⁻¹dry soil [27].

Asb_sample, Asb_CK, absorbance values of sample and control; V, volume (L); M, mass (g); T, time (h); K = 0.834 × 10³.

2.4.5. Detection of Peroxidase (POD)

Soil peroxidase (POD) was quantified using 0.1 g of air-dried soil (<2 mm) in a 2 mL centrifuge tube. In the experimental group, 0.6 mL of acetic acid buffer (pH 5.5), 0.6 mL of dopamine substrate solution, and 0.6 mL of 0.3% H₂O₂ were introduced. In contrast, the control group entailed measuring 0.1 g of air-dried soil in a 2 mL centrifuge tube, to which 1.2 mL of buffer and 0.6 mL of 0.3% H₂O₂ were added. The mixtures were then thoroughly combined and capped. They were then cultured for 1 h in a 37 °C incubator with gentle agitation (100 rpm) and subsequently subjected to centrifugation (5000 rpm, 5 min) for filtration. The filtrate was then subjected to colorimetric determination at 460 nm [28]. Enzyme activity was calculated according to the following formula: units, nmol diqc h⁻¹g⁻¹dry soil.

2.5. Transcriptome Analysis Methods

2.5.1. Sample Preparation, Library Construction, and Sequencing

Citrus leaf samples were collected for analysis. Total RNA extraction from these citrus leaves was performed using TRIzol^® reagent (Beijing Norbolide Technology Co., Ltd., Beijing, China). Subsequently, genomic DNA was removed using DNase I. The concentration and quality of the RNA samples were rigorously evaluated using a 2100 Bioanalyzer (Agilent Technology (China) Co., Ltd., Shanghai, China) and ND-2000 (America Thermo, Waltham, MA USA), respectively. Following this assessment, high-quality RNA samples were used to construct the sequencing libraries.

2.5.2. Library Preparation and Illumina NovaSeq 6000 Sequencing

RNA purification, reverse transcription, library construction, and sequencing procedures were performed in accordance with the manufacturer’s instructions provided by Illumina, San Diego, CA, USA. To prepare the transcriptome library for RNA-SEQ analysis, 1 μg of total RNA was used using the TrSeqTM RNA sample preparation kit from Illumina (San Diego, CA, USA).

Initially, mRNA molecules were selectively isolated by polyA selection using magnetic beads. Subsequently, the fragmented mRNA was subjected to double-stranded cDNA synthesis using the SuperScript double-stranded cDNA synthesis kit (Invitrogen, CA, USA) and random hexameric primers provided by Illumina. The resulting cDNA fragments underwent processes such as end repair, phosphorylation, and the addition of “A” bases, all in accordance with Illumina’s library construction protocol. Library size selection targeted cDNA fragments ranging from 200 to 300 bp and was accomplished through electrophoresis on a 2% low-range agarose gel. Following this, PCR amplification was performed using Phusion DNA polymerase for 15 cycles. After quantification using the TBS380 assay system, paired-end RNA-seq sequencing libraries were sequenced on an Illumina NovaSeq 6000 platform, with a read length of 2 × 150 bp.

2.5.3. Bioinformatics Analysis

The data underwent a rigorous filtering process to obtain clean data, which were subsequently aligned with Citrus sinensis s. hqb _v1. genome.fa, utilizing the highly efficient comparison system HISAT2 [29]. The reads that resulted from this alignment were then assembled using StringTie [30] (Pertea et al., 2015), and the percentage of mapped reads in clean reads was calculated. Fragments per kilobase of transcript per million fragments mapped (FPKM) served as a metric for assessing transcript or gene expression levels [31].

Next, differential expression analysis was carried out using DESeq2 [32], with a fold change ≥ 2 and FDR < 0.01 serving as the screening criteria. The fold change represents the ratio of expression between the two groups, highlighting significant differences in expression levels. Subsequently, comprehensive functional annotation and enrichment analyses of differentially expressed genes (DEGs) were conducted. Additionally, gene products were classified based on lineology using the Clusters of Orthologous Groups of Proteins (COG) database. Pathway enrichment analysis of DEGs was performed to identify any significant differences in specific pathways.

2.6. Prediction and Analysis of BugBase Phenotypes

Microbial diversity analysis was performed using the Illumina NovaSeq sequencing platform. A library was constructed and sequenced using the paired-ends method. This approach facilitated the identification of species, abundance analysis, and determination of species composition within the samples. These analyses were performed using various techniques, including filtering, clustering, and denoising.

2.6.1. Library Construction and Sequencing

The process began with the separate extraction of total DNA from both root and rhizosphere soil samples. Primers were designed based on conserved regions. The sequencing adapters were fixed to the primers. PCR was used to amplify the target sequences, and the resulting products were purified, quantified, and homogenized to generate a sequencing library. Subsequently, library quality control was performed to ensure that the libraries met the required standards.

Once qualified, the libraries were sequenced using the Illumina NovaSeq 6000 platform. The original image data files generated from high-throughput sequencing were subjected to base-calling analysis to yield the sequenced reads. These results were stored in FASTQ format files, which included sequence information for the reads as well as their corresponding sequencing quality information.

2.6.2. BugBase Phenotype Prediction

BugBase [33] is an algorithm designed to predict functional pathway coverage within complex microbiomes and provides insights into bio-explanatory phenotypes, such as oxygen tolerance, gram staining, and pathogenic potential. BugBase initially normalizes operational taxonomic units by considering the predicted 16S copy numbers. It then utilizes the precomputed documentation provided by BugBase to predict the microbial phenotypes.

For each sample within the biological dataset, BugBase estimates the relative abundance of traits across a threshold range from 0 to 1, with increments of 0.01. Subsequently, BugBase identifies the highest coverage threshold for each feature among all samples. Once this threshold value is determined, BugBase generates an organism-level trait-prediction table that contains the relative abundance of the predicted traits for each sample.

BugBase also offers the capability of selective automatic hypothesis testing and visualization of differential traits based on user-specified metadata. In addition, it can generate taxonomic contribution maps that describe the relative abundance of taxonomic groups with characteristic traits. Outputs from BugBase include statistical summary files for nonparametric differential tests, such as the Mann–Whitney U or Kruskal–Wallis tests.

3. Results

3.1. Effects of P. polymyx KN-03 on Plant Morphological Indices

Following 24 treatments in the experimental group, the results revealed that in the KN-03 treatment group, the plant height, crown width, stem diameter, and root crown cross-sectional area were 84.37 ± 10.84 cm, 57.25 ± 10.37 cm, 11.21 ± 1.19 mm, and 84.16 ± 8.41 cm², respectively. In the inactivated group, these measurements were 66.25 ± 4.78 cm, 41.24 ± 11.08 cm, 9.37 ± 0.91 mm, and 62.61 ± 11.37 cm², respectively. Meanwhile, in the control group, the growth indices were recorded at 77.25 ± 11.38 cm, 51 ± 2.94 cm, 9.712 ± 0.91 mm, and 64.56 ± 16.13 cm² (Figure 1A–D).

It is noteworthy that plants treated with active KN-03 AS bacteria exhibited significantly greater plant height, crown width, stem diameter, and root crown cross-sectional area than those treated with heat-inactivated KN-03 (p < 0.05). These findings suggest that inactivation of KN-03 bacteria has a detrimental effect on plant growth, whereas treatment with active KN-03 promotes new root development and substantially enhances the growth of CLas-infected plants.

3.2. Changes in HLB Disease before and after Treatment with KN-03

The comprehensive evaluation of Huanglongbing disease is grounded in the results obtained from conventional PCR and qPCR detection methods. If either of these tests yields a positive result, the sample is considered positive. According to routine PCR detection for HLB disease (Supplementary Figure S2A,B), after 12 treatments, two out of four samples in both the live bacteria treatment group and the inactivated treatment group tested negative post treatment. However, after 12 treatments, qPCR test results indicated that all treatment groups showed no conversion to negative status; nonetheless, they did demonstrate a decrease in the relative concentration of leaf vein pathogens compared to pretreatment levels. In the inactivated group, three out of four plants exhibited a reduction in the relative concentration of leaf vein pathogens. No significant change in pathogen concentration within the leaf veins of plants treated with water was observed (

Table 2. Detection of plant pathogen of Huanglongbing in leaf vein before and after treatment.

	Before Treatment 1					After 12 Times Treatment 2				After 24 Times Treatment 3
	Sample Number	Results of qPCR*	PCR Result	Evaluation	Negative Rate %	Results of qPCR	PCR Result	Evaluation	Negative Rate %	Results of qPCR	PCR Result	Evaluation	Negative Rate %
KN-03 viable bacteria	YL48	5.03 ± 0.04 (+)	+	+	0	8.69 ± 1.96 (+)	+	+	0	0.24 ± 0.01 (−)	−	−	50
	YL39	15.61 ± 0.13 (+)	+	+		8.55 ± 0.02 (+)	−	+		0.99 ± 0.01 (−)	−	−
	YL20	127.15 ± 3.99 (+)	+	+		72.03 ± 3.22 (+)	+	+		59.14 ± 0.53 (+)	+	+
	YL32	35.03 ± 1.55 (+)	+	+		4.68 ± 0.07 (+)	−	+		8.16 ± 0.84 (+)	−	+
Inactivated bacteria	YL35	29.75 ± 0.47 (+)	+	+	0	21.71 ± 1.09 (+)	+	+	0	3.27 ± 0.17 (+)	−	+	0
	YL16	33.94 ± 4.18 (+)	+	+		55.30 ± 4.06 (+)	−	+		9.62 ± 1.64 (+)	−	+
	YL15	756.04 ± 1.76 (+)	+	+		4.83 ± 2.76 (+)	+	+		15.37 ± 1.12 (+)	+	+
	YL14	26.42 ± 0.19 (+)	+	+		3.36 ± 0.02 (+)	−	+		2.19 ± 0.21 (+)	+	+
CK	YL18	54.61 ± 2.94 (+)	+	+	0	2.73 ± 0.43 (+)	+	+	0	33.46 ± 1.25 (+)	+	+	0
	YL10	2.22 ± 0.27 (+)	+	+		114.46 ± 5.85 (+)	+	+		29.42 ± 3.62 (+)	+	+
	YL17	5.03 ± 0.92 (+)	+	+		11.67 ± 3.57 (+)	−	+		6.43 ± 0.61 (+)	−	+
	YL45	36.76 ± 1.13 (+)	+	+		16.64 ± 0.63 (+)	+	+		7.71 ± 0.56 (+)	+	+

Note: Negative control and 2^−∆∆CT = 1. If 2^−∆∆CT > 1, it is considered positive, and if 2^−∆∆CT < 1, it is considered negative. Sampling time: ¹ 6 July 2021; ² 22 September 2021; ³ 24 March 2022.

).

Following 24 KN-03 treatments, PCR testing revealed that three-quarters of plants treated with live bacteria tested negative (refer to Supplementary Figure S2C). Similarly, half of the plants treated with inactivated bacteria yielded negative results, while only one out of four control plants (25%) tested negative. QPCR analysis demonstrated that after 24 treatments, half of the samples from the live bacteria treatment turned negative, whereas both samples from the inactivated bacteria treatment and the control remained positive (Table 2).

After 24 treatments, the root test results from both methods indicated HLB positivity, with only one strain in the control group testing negative. PCR and qPCR tests were conducted on the apical buds, revealing that two out of the four plants tested negative after live bacterial treatment, while one out of the four plants tested negative after inactivated bacterial treatment. The control group remained positive (Supplementary Figure S1D,E;

Table 3. Detection of plant pathogens of Huanglongbing before and after treatment in the apical meristem and root.

Treatment	Pathogen Content of Roots 1					Pathogen Content of Buds 2
	Sample Number	Results of qPCR	PCR Result	Evaluation	Negative Rate %	Results of qPCR*	PCR Result	Evaluation	Negative Rate %
KN-03 viable bacteria	YL48	4.59 ± 0.94 (+)	+	+	0	1.27 ± 0.75 (+)	−	+	50
	YL39	4.11 ± 0.81 (+)	−	+		1.84 ± 0.17 (+)	+	+
	YL20	12.18 ± 1.59 (+)	+	+		0.85 ± 0.02 (−)	−	−
	YL32	9.33 ± 0.15 (+)	−	+		0.27 ± 0.01 (−)	−	−
Inactivated bacteria	YL35	4.31 ± 0.37 (+)	−	+	0	7.11 ± 1.37 (+)	−	+	25
	YL16	6.03 ± 0.48 (+)	−	+		2.11 ± 0.42 (+)	+	+
	YL15	2.45 ± 0.02 (+)	+	+		0.05 ± 0.01 (−)	−	−
	YL14	60.46 ± 6.56 (+)	+	+		13.71 ± 5.25 (+)	−	+
CK	YL18	3.59 ± 0.26 (+)	+	+	25	345.07 ± 11.40 (+)	+	+	0
	YL10	0.66 ± 0.17 (−)	−	−		5.05 ± 0.91 (+)	+	+
	YL17	3.60 ± 0.26 (+)	+	+		4.84 ± 0.16 (+)	−	+
	YL45	1.62 ± 0.23 (+)	−	+		0.67 ± 0.32 (−)	+	+

Note: The result of qPCR* is 2^−∆∆CT. Negative control, and 2^−∆∆CT = 1. If 2^−∆∆CT > 1, it is positive; 2^−∆∆CT < 1 is considered negative. Note: The sampling time, ¹: 26 March 2021, ²: 26 March 2022.

).

In conclusion, the results demonstrated that the application of KN-03 effectively reduced the CLas content in nursery citrus plants. After the 24 AS treatments, the negative rates of CLas in the vein, root tip, and shoot tip were 50%, 0%, and 50%, respectively. In comparison with heat-inactivated KN-03 AS or the water control, the negative rates of CLas in the vein, root tip, and shoot tip were 0%, 0%, and 25% or 0%, 25%, and 0%, respectively. This provides high-probability CLas-free shoot-tip meristems as materials and enhances detoxification efficiency for healthy citrus plant production.

3.3. Effects of P. polymyxomyces KN-03 on Biochemical Indices

Catalase (CAT) is a pivotal hemoglobin protease in the in vivo environment and plays a crucial role in plant resistance because of its capacity to catalyze the decomposition of excess H₂O₂ into H₂O and O₂. When plants confront disease-induced stress, polyphenol oxidase (PPO) fosters the development of lignin and various immune compounds, thereby bolstering disease resistance. Furthermore, POD engages in metabolic processes during plant stress conditions by detoxifying harmful substances and preserving the internal tissue equilibrium. The findings revealed that CAT enzyme activity significantly increased after 24 h of treatment with inactivated bacteria compared to both the control group and activated bacterial treatment. POD enzyme activity was higher after 12 h of treatment with inactivated bacteria in comparison to live bacteria treatment, and it was also markedly greater than that observed in the control group after 24 h. Additionally, the activity of the PPO enzyme exhibited a notable increase after 12 h of treatment with inactivated bacteria when compared with the live bacteria treatment and the control group (Figure 2A–C).

MDA production occurs when plants endure various stressors, such as high temperature, drought, and injury, and is accompanied by related metabolic reactions in tree tissues or organs. The MDA content is closely associated with plant senescence and stress-related injuries. As indicated in Figure 2D, the average MDA content in the active, inactivated, and control groups was 42.00 nmol/g, 44.26 nmol/g, and 36.48 nmol/g, respectively. The MDA content in the active bacterial treatment group and inactivated treatment group was significantly higher than that in the clean water control group (p < 0.05), showing increases of 15.12% and 21.31%, respectively. These results highlight an elevation in the MDA content in plants due to both active and inactivated treatments (Figure 2D).

Soil PPO plays a pivotal role in soil organic matter formation by facilitating the conversion of aromatic compounds into soil organic constituents. This significantly contributes to augmenting the soil organic matter content and improving soil fertility. The alterations in soil PPO activity are presented in Figure 2E: Both the active bacteria treatment group (36.29 U/g∙min) and the inactivated treatment group (34.89 U/g∙min) displayed significantly higher enzyme activity in comparison to the control group (24.08 U/g∙min) (p < 0.05). The active bacterial treatment group demonstrated a 50.70% increase, while the inactivated treatment group exhibited a 44.89% increase compared to the control group (Figure 2E).

Serving as an indispensable REDOX reductase, soil CAT participates in biological respiratory metabolism by decomposing hydrogen peroxide generated during respiration, thereby mitigating its toxic effects on organisms. The variations in CAT activity are depicted in Figure 2F: Both the active bacteria treatment group (45.37 U/g∙min) and the inactivated treatment group (46.07 U/g∙min) exhibited significantly higher enzyme activity than the control group (38.8 U/g∙min) (p < 0.05). Compared to the control group, the active bacteria treatment group displayed a 19.14% increase, while the heat-inactivated treatment group showed a 20.98% increase.

DAB and NBT staining were used for in vitro staining of citrus leaves. As depicted in Figure 2G,H, leaves treated with KN-03 bacterial solution exhibited conspicuous staining, indicating the stimulation of reactive oxygen species production in leaves through KN-03 bacterial treatment.

3.4. Transcriptomic Analysis of RNA-Seq

3.4.1. Transcriptome Data Meet Quality Standards

Transcriptome analysis included 12 samples derived from three distinct groups: KN-03 treatment, inactivated bacterial treatment, and a control group. A total of 125.78 gigabases (Gb) of clean data was acquired, with each sample exhibiting a Q30 base percentage exceeding 93.13% (Table S1). The comparative results revealed that the efficacy of comparing the reads from each sample to the reference genome ranged from 85.08% to 95.50% (Table S2). Principal component analysis was conducted on both the KN-03 treatment and control groups. In this analysis, PC1 accounted for 69.6% of the contribution rate, whereas PC2 accounted for 17.5% (Figure 3A). In the KN-03 treatment group and the control group, PC1 explained 83.3% of the contribution rate, with PC2 contributing 12.3%. These results indicated notable distinctions between groups and robust repeatability within groups, meeting the criteria for subsequent statistical analysis (Figure 3B).

3.4.2. Differential Expression Gene and KEGG Cluster Analysis

Cluster heatmap analysis revealed distinct gene expression patterns among the examined groups. Compared to the control group, the KN-03 treatment group exhibited upregulation of 36 genes and downregulation of 51 genes (Figure 4A). Conversely, the inactivated bacterial treatment group showed upregulation of 23 genes and downregulation of 12 genes (Figure 4B). Moreover, the KN-03 treatment group displayed an additional upregulation of 13 genes and a further downregulation of 39 genes compared to the inactivated bacteria treatment group (Table S3).

KEGG enrichment analysis identified 17 pathways with significantly upregulated expression levels. All these pathways were upregulated with DEGs, FDR < 0.01, and log₂FC between 1 and 2.45 (Figure 5A,B).

The top three upregulated pathways following KN-03 treatment were as follows: (1) protein processing pathways in the endoplasmic reticulum (16/850), involved in managing extracellular proteins such as antibodies and hormones; (2) endocytosis (6/579), which refers to both phagocytosis of harmful substances by phagocytes and the uptake of extracellular nutrients mediated by cell plasma membrane receptors to maintain normal metabolic activities; and (3) spliceosome (4/1002), a multicomponent complex crucial for RNA splicing that plays a significant role in genetic regulation. Additionally, one differential gene was identified in each of the remaining seventeen pathways, all of which exhibited up-regulation patterns (Figure 5A, Table S4).

After treatment with inactivated KN-03, the enriched KEGG pathways included photosynthesis-antenna protein (1/37), MAPK signaling pathway–plant (1/1014), spliceosome (2/1002), plant–pathogen interaction (2/2521), endocytosis (3/579), and protein processing in the endoplasmic reticulum (14/850) (Figure 5B, Table S5). Notably, KEGG enrichment pathways following KN-03 treatment differed from those observed after treatment with heat-inactivated bacteria. Apart from protein processing in the endoplasmic reticulum, spliceosome, and endocytosis, there were 17 unique KEGG upregulation enrichment pathways specific to KN-03 treatment. Photosynthesis-antenna protein, MAPK signaling pathway–plant, and plant–pathogen interaction represent three distinct enrichment pathways following treatment with inactivated bacteria.

The COG database was constructed based on phylogenetic relationships among bacteria, algae, and eukaryotes. This database was used to categorize the gene products in treatment experiments involving KN-03 and inactivated bacteria. The classification of differentially expressed COG genes includes various categories (Figure 6A, Table S6): The three most frequent functional classification sequences are posttranslational modification, protein turnover, and chaperones (17); carbohydrate transport and metabolism (11); and secondary metabolite biosynthesis, transport, and catabolism (5). The numbers in parentheses represent the number of genes within each category. The proportion of genes in different functional classes signifies metabolic or physiological bias resulting from treatment with live bacteria.

In contrast, the COG DEGs following inactivated bacterial treatment were classified, and the three most frequent functional classification sequences were as follows (Figure 6B, Table S7): posttranslational modification, protein turnover, and chaperones (15); cell wall/membrane/envelope biogenesis (2); and carbohydrate transport and metabolism (2). These DEGs indicated distinctions in treatment functions between live and inactivated bacteria, with live bacteria exhibiting a broader range of functions in promoting seedling growth and enhancing disease resistance.

3.5. Microflora Analysis and BugBase Phenotypic Prediction

Based on the results obtained from the root sample sequencing data, “Clean Reads” refers to the number of high-quality reads obtained after the initial sequence quality control, with Q30% ranging from 95.09% to 97.1% (Table S8). Regarding rhizosphere sample sequencing data processing, the Q30% ranged from 95.03% to 95.45% (Table S9). This assessment of sequencing data quality confirmed that the samples adhered to the required standards.

Following KN-03 treatment, the relative abundance of Burkholderia-Caballeronia-Paraburkholderia, uncultured_bacterium_f_Xanthobacteraceae, Rhodanobacter sp., Allorhizobium neorhizobium-Pararhizobium-Rhizobium, and uncultured_bacterium_f_Burkholderiaceae in the roots increased by factors of 1.23, 1.06, 1.25, 2.21, and 1.41, respectively, compared with the control group. After heat-inactivated KN-03 treatment, the relative abundance of Burkholderia-Caballeronia-Paraburkholderia in the roots was 1.62 times higher than that of the control group. In comparison to inactivated KN-03 treatment, after KN-03 treatment, the abundances of Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, uncultured_bacterium_f_Enterobacteriaceae, and uncultured_bacterium_f_Burkholderiaceae were 2 times, 3.53 times, and 1.6 times higher than those in the control group, respectively (Table S10; Figure 7A–C).

After KN-03 treatment, the abundance of Burkholderia-Caballeronia-Paraburkholderia, uncultured_bacterium_o_Acidobacteriales, uncultured_bacterium_f_Gemmatimonadaceae, and Rhodanobacter in the rhizosphere increased by factors of 1.08, 1.25, 1.11, and 1.11, respectively, compared with the control group (Table S10, Figure 7D). Following treatment with inactivated bacteria, the abundance of Burkholderia-Caballeronia-Paraburkholderia and uncultured_bacterium_o_Betaproteobacteriales in the rhizosphere was 1.12 and 3.19 times higher than that of the control group, respectively. Compared with the inactivated bacteria treatment, after KN-03 treatment, there was a significant increase in the abundance of uncultured_bacterium_o_Acidobacteriales (by 1.19 times), uncultured_bacterium_f_Xanthobacteraceae (by 1.16 times), Rhodanobacter (by 1.30 times), and uncultured_bacteria_f_JG30-KF-AS9 (by 1.23 times) in the rhizosphere (Table S10; Figure 7E,F). KN-03 treatment resulted in a higher abundance of these probiotics within the rhizosphere.

The predicted phenotypes in BugBase typically include the following nine categories: aerobic, anaerobic, facultative anaerobic, Gram-positive, Gram-negative, biofilm-forming, potential pathogenicity, mobile element containing, and stress tolerance. In this experiment, treatment with KN-03 resulted in a higher relative abundance of root flora exhibiting the phenotypes of aerobicity, mobile element containment, biofilm formation capability, Gram-negativity, potential pathogenicity, and stress tolerance compared to the control group (Supplementary Figure S2A–I). Additionally, the number of genera displaying aerobicity, mobile element containment, biofilm formation capability, Gram-negativity, potential pathogenicity, and stress tolerance increased by 50 in the roots (Table S11,

Table 4. Comparison of the number of genera of flora at the root and rhizosphere in KN-03 treatment tests.

Materials	Phenotype
Number of Treatments and Genera	Aerobic	Anaerobic	Contains_Mobile_Elements	Facultatively_Anaerobic	Forms_Biofilms	Gram_Negative	Gram_Positive	Potentially_Pathogenic	Stress_Tolerant
SA    +5	9D	3	29U	7	16U	16U	5D	14U	13U
SACK	10	3	27	7	15	15	8	12	10
SB     −1	9D	4	29U	7	14D	15	6D	15U	9D
SBCK	10	4	27	7	15	15	8	12	11
RA   +50	20U	8	20U	6D	30U	30U	7D	21U	19U
RACK	11	15	11	7	20	21	14	11	11
RB    −3	14U	15D	6U	8	15D	15D	13D	10D	10D
RBCK	13	16	4	8	16	16	16	11	11

Note: R, root; S, rhizosphere soil; RA, activated KN-03 treatment; RB, heat-inactivated KN-03 treatment, U, an increase in the number of genera compared to the control; D, a reduction in the number of genera compared to the control.

). Meanwhile, inactivated bacterial treatment led to an increase in the relative abundance of root flora compared to the control group, focusing on phenotypes such as aerobicity, biofilm formation, Gram-negativity, potential pathogenicity, and stress tolerance (Supplementary Figure S3A–I).

After treatment with KN-03, the relative abundance of rhizosphere microflora was significantly higher than that in the control group. These changes were characterized by phenotypes such as anaerobicity, presence of mobile elements, facultative anaerobes, biofilm formation, Gram-negative bacteria, potential pathogenicity, and stress tolerance (Supplementary Figure S4A–I). Furthermore, there was an increase in the number of genera exhibiting traits including mobile element carriage, biofilm formation ability, Gram-negative characteristics, potential pathogenicity, and stress tolerance; a total of five genera showed an increase (Table S11, Table 4). After treatment with inactivated bacteria, the relative abundance of flora in the rhizosphere, representing anaerobic, facultatively anaerobic, Gram-negative, and potentially pathogenic bacteria, was found to be higher than that in the control group (Supplementary Figure S5A–I). The number of genera associated with phenotypes including mobile elements and potential pathogenicity increased, while a total of one genus showed a reduction (Table S11, Table 4).

4. Discussion

The current study explored the effect of KN-03 on HLB management. KN-03 played a multifaceted role in stimulating plant growth, bolstering disease resistance, modulating biodiversity, and increasing the abundance of beneficial microorganisms. The findings revealed that after 24 applications of KN-03, there was a 50% reduction in infection rates in the leaf veins and shoot tips of the positive plants. Detoxification of shoot-tip CLas disease holds practical significance for safeguarding citrus plants from HLB-related damage.

4.1. P. polymyxoides KN-30 May Trigger Citrus Autoimmunity

Rhizosphere microorganisms play a crucial role in plant growth and development, with beneficial rhizosphere microorganisms establishing mutualistic symbiosis with plant roots in the rhizosphere environment [34,35]. P. polymyxa, known for its excessive production of exopolysaccharides, is easily colonizable and cultivable in the rhizosphere environment. It is capable of ISR in plant defense, thereby enhancing plant adaptability to stressful environments [36].

This study demonstrated that KN-30 stimulated an increase in CAT, POD, and PPO activity; MDA content; and activated reactive oxygen species production. These findings indicate that the live bacteria KN-30, diluted at a ratio of 500:1, has the capacity to elicit a sensitive response in citrus plants, thereby enhancing their autoimmunity. This effect may be attributed to lipopeptides or other signaling molecules produced by P. polymyxa [37,38]. Evaluation of resistance to polymyxin-inoculant treatment revealed that both live KN-03 AS treatment and heat-inactivated treatment induced ISR to some extent after administration. Notably, the heat-inactivated treatment significantly and rapidly improved the overall plant CAT, POD, PPO, and MDA indices. This result suggests that during the polymyxin agent-deactivation process, immune activators are rapidly released from polymyxin, directly contacting the plant roots during treatment procedures, which leads to a more sensitive and rapid induction of plant defense responses.

4.2. Influence of KN-30 Treatment on HLB Disease

Several strains of P. polymyxa exhibit broad-spectrum antimicrobial activity and are effective in controlling various fungi, bacteria, and nematodes [39,40,41,42]. Numerous reports have highlighted the effectiveness of Paenibacillus spp. in the disease management of various cultivated crops. For instance, HKA-15 has been effective in controlling citrus canker [43], GBR-1 has been successful in inhibiting tomato root knot formation [44], A21 efficiently manages gray mold in tomatoes, and SQR-21 has demonstrated activity against Fusarium oxysporum [45].

In this study, a comparison of CLas detection results in leaf veins before and after treatment indicated that the CLas content was reduced with the KN-30 AS formulation in the soil drench treatment. This suggests that the proliferation of CLas cells was partially inhibited by KN-30, with a lesser effect observed with heat treatment. These findings align with previous studies on the application of Bacillus amylolyticus GJ1 to control citrus HLB disease [46] and the use of endophytic Bacillus subtilis L1-21, which showed potential for Asian citrus psyllid and CLas control, as reported by Li et al. (2022) [47]. Unexpectedly, one citrus plant in the control group exhibited a 25% reduction in CLas content. Further investigations into disease mobility should be conducted.

4.3. Transcriptome Analysis Revealed a Significant KEGG Upregulation Pathway Induced by KN-30

Following treatment with KN-30 AS formulations, we observed upregulation of pathways involved in cellular processes, genetic information processes, metabolism, and biological systems. Specifically, the KEGG upregulated pathways of DEGs showed significant enrichment in pathways, such as protein processing in the endoplasmic reticulum, endocytosis, and spliceosome. Protein processing in the endoplasmic reticulum is responsible for the secretion of antibodies and hormones [48]. Another important pathway is endocytosis, which involves the formation of vesicles through invagination of the cytoplasmic membrane to internalize external substances into cells [49]. The third pathway, the spliceosome, is a multicomponent complex composed mainly of small nuclear RNA and proteins that play crucial roles in regulating genetic activity [50].

Additionally, among the 17 enriched upregulated KEGG pathways, phenylpropanoid biosynthesis, flavonoid biosynthesis, diarylheptane and gingerol biosynthesis, and zein biosynthesis were identified. These pathways have antibacterial, antiviral, and antioxidant effects and are closely related to disease resistance [51]. Previous studies have reported that the expression of flavonoid pathway genes in poplar can enhance its resistance to pathogens [52]. Anthrax (Colletotrichum sublineolum) produces a flavonoid that exerts antagonistic effects after infecting sorghum [53]. Zein biosynthesis is directly involved in cytokinin synthesis and is related to the stimulation of cell division. The results in the above-mentioned literature are in line with the identified upregulated pathways and functions.

4.4. KN-03 Treatment Increased Microbial Diversity and the Abundance of Beneficial Microorganisms

Within the rhizosphere, certain Paenibacillus species engage in respiration, releasing carbon dioxide or metabolizing acids to facilitate the dissolution of insoluble minerals, thus enhancing plant absorption of essential mineral elements such as phosphorus. Bacillus species have been found to directly influence plant growth through the production of indole-3-acetic acid and other auxin-like hormones. Some Bacillus species also possess nitrogen-fixing capabilities from atmospheric sources [54]. Additionally, some rhizosphere microorganisms produce antibiotics that inhibit the proliferation of pathogenic microbes that are harmful to plants [55]. In this study, the application of KN-03 indicated its growth-promoting function. Notably, the rhizospheres of plants treated with KN-03 exhibited a higher abundance of Burkholderia-Caballeronia-Paraburkholderia, uncultured_bacterium_o_Acidobacteriales, uncultured_bacterium_f_Gemmatimonadaceae, and Rhodanobacter than that of the control group.

From the published literature, we know that these bacteria have important functions. For example, Burkholderia-Caballeronia-Paraburkholderia is a dominant genus in corn–peanut intercropping systems and is closely associated with alleviating the autotoxic effects caused by continuous cropping [55,56]. The dominant strains of ZQ01 were mainly derived from Sinomonas, Pandoraea, and Burkholderia-Caballeronia-Paraburkholderia, which were able to rapidly colonize and grow during in situ bioreaction in contaminated soil and water [57]. Matsumoto et al. (2021) [58] found significantly higher proportions of Burkholderia bacteria in rice seeds. Additionally, Wilhelm et al. (2021) [59] demonstrated that Pandoraea-Burkholderia-Caballeronia-Paraburkholderia exhibited enhanced pollution-coping capabilities. Furthermore, following biochar application, Gemmatimonadaceae was enriched by a ¹³C stable isotope label; Acidobacteria represented a dominant bacterial group in paddy soils, playing a crucial role in regulating soil ecological balance, while hydroxy-ferric oxide was shown to alter microbial community structures and enrich denitrifying bacteria [60]. Rhodanobacter, Chryseobacterium, and Rhizomicrobium are the dominant endophytic bacteria endemic to sugarcane during sugarcane–mung bean intercropping treatment [61].

In this study, based on the predicted phenotypes from BugBase, the abundance of related bacteria in five phenotypes (aerobic, mobile elements, biofilm formation, Gram-negative bacteria, and stress tolerance) was found to be higher in the roots after KN-03 treatment than in the control group. Similarly, in the rhizosphere, there was a higher abundance of microflora associated with five phenotypes (anaerobic, amphoanaerobic, Gram-negative, biofilm formation, and stress tolerance) following KN-03 treatment. These changes indicated the positive effects of KN-03. However, it should be noted that potentially pathogenic bacteria show increased abundance, and the potential effect of these changes on citrus health remains unknown and requires further investigation and analysis. In the future, we should further explore the disease resistance mechanism and explore beneficial microorganisms with stronger anti-Huanglongbing disease effects so that they may play a better role in the field of Huanglongbing disease resistance.

Our results suggest that an ecologically balanced approach to controlling HLB is better than a single-factor approach, especially with the help of beneficial microorganisms. Therefore, exploring new methods from an ecological balance perspective is necessary. P. polymyxoides KN-03 has been shown to promote growth and disease resistance while regulating biodiversity and increasing probiotic abundance. Thus, creating a favorable ecological environment for citrus trees with rich microbial diversity and continuous accumulation of probiotics is crucial for the effective management of HLB disease through cooperative interactions between various microorganisms alongside Clas.

5. Conclusions

Bacillus polymyxis KN-03 effectively reduces the concentration of the Huanglongbing pathogen in the citrus apex meristem and provides a favorable substrate for the detoxification of citrus shoot tips. Transcriptome analysis revealed that KN-03 treatment significantly up-regulated 17 KEGG pathways associated with vital functions such as protein processing in the endoplasmic reticulum, endocytosis, spliceosome, phenylalanine synthesis, flavonoid biosynthesis, and gingerol biosynthesis. These findings are closely linked to growth promotion and disease resistance. Comparison of the top 10 species based on genus-level abundance showed that after KN-03 treatment, potential Burkholderia-Caballeronia-Paraburkholderia, uncultured_bacterium_f_Xanthobacteraceae and Rhodanobacter Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium as well as uncultured_bacterium_f_Burkholderiaceae were found in higher abundance in the roots compared to the control group. Following KN-03 treatment, there was an increase in aerobic bacteria, biofilm-forming bacteria, Gram-negative bacteria and stress-tolerant phenotypes among rhizosphere flora at the genus level when compared to the control group. Bacillus polymyxoides KN-03 exhibited multifaceted effects including growth promotion, disease resistance enhancement, regulation of biodiversity, and increased abundance of potential utilization-value strains.