The pure 3-OH PAME chemical was synthesized by Shanghai Zixia Biotechnology Co., Ltd. and the chromatographic methanol was used to dissolve it. All other chemicals were synthesized by Guangzhou Dingguo Biotechnology Co. LTD. The plants used in the following inoculation experiments (Casuarina equisetifolia cv. Wenchang, tomato cv. Provence, and peanut cv. Luhua) were purchased or self-cultivated.

The phytopathogenic bacterium used in this study was R. solanacearum NS25, which was isolated from Casuarina bacterial wilting trunk previously, and cultured in CPG medium (peptone, 10.0 g/L; glucose, 5.0 g/L; and casamino acid, 1.0 g/L; solid medium added with 15 g/L agar; Ph 7.0) at 30°C (Hendrick and Sequeira, 1984). The TTC medium was added with 5% 2,3,5-triphenyltetrazolium chloride (TTC) in a 1:1000 ratio to the melted CPG solid medium. The QQ strains Q1-7 and Q4-3 were cultivated in LB medium (NaCl, 10.0 g/L; tryptone, 10.0 g/L; and yeast extract, 5.0 g/L; solid medium added with 15 g/L agar; Ph 7.0) at 28°C (Maniatis et al., 1982).

The mineral salt medium [MSM; (NH₄)₂SO₄, 2.0 g/L; Na₂HPO₄·12H₂O, 1.5 g/L; NaH₂PO₄, 1.5 g/L; MgSO₄·7H₂O, 0.2 g/L; CaCl₂·2H₂O, 0.01 g/L; FeSO₄·7H₂O, 0.001 g/L; Ph 6.5] was used in the QQ bacterial screening and degradation tests of 3-OH PAME (Zhang et al., 2021).

The medium for swimming motility test contains 5.0 g/L bactopeptone, 5.0 g/L NaCl, and 3 g/L agar, that for swarming test contains 0.5% M8 plates supplemented with 2.0 g/L glucose and 0.5 g/L glutamate, and that for twitching test contains 10.0 g/L NaCl, 10.0 g/L tryptone, 5.0 g/L yeast extract and 15 g/L agar (Murray and Kazmierczak, 2006; Hu et al., 2021). SOBG medium for the measurement of biofilm formation contains 20.0 g/L tryptone, 5.0 g/L yeast extract, 2.4 g/L MgSO₄·7H₂O, 0.5 g/L NaCl, 0.186 g/L KCl, and 2% glycerol (Branda et al., 2005; Hu et al., 2021).

The medium for the measurement of EPS production (YEB) contains 10.0 g/L tryptone, 5.0 g/L yeast extract, 5 g/L KCl, 0.5 g/L MgSO₄·7H₂O, 5 g/L sucrose, 15 g/L agar (Romling et al., 1998). The medium for the measurement of pectate lyase (Pel) activity contains 10.0 g/L polygalacturonic acid, 10.0 g/L yeast extract, 8.0 g/L agar, 4.8448 g/L Tris-HCl, and 0.1125 g/L CaCl₂. The medium for the measurement of polygalacturonase (Peh) activity contains 5.0 g/L polygalacturonic acid, 2.0 g/L sucrose, 15.0 g/L agar, and 2.0 g/L (NH₄)₂SO₄. The medium for the measurement of cellulase (Cel) activity contains 1.0 g/L carboxymethyl cellulose, 3.8 g/L Na₃PO₄, and 8.0 g/L agar. The medium for the measurement of protease (Prt) is LB solid medium mixed with 4% non-fat milk powder at a ratio of 1:1.

A standard curve of 3-OH PAME was established using 5 different concentrations of 3-OH PAME specimens: 0.25 μM, 0.5 μM, 1 μM, 2 μM and 4 μM ( Supplementary Figure S1 ). The correlation between AA values and 3-OH PAME contents were determined. The experiment was repeated in three independent times with four replicates for each concentration each time.

The selective-enrichment culture method was used to screen QQ strains (Flagan et al., 2003). Firstly, inter-rhizosphere soil and Casuarina branches were collected from the Haitouwan Casuarina forestry in Zhanjiang, Guangdong Province. The branches were crushed into sawdust, and 10 g of which and the mixed soil sample were respectively added into 200 mL of MSM liquid medium containing 1 μM 3-OH PAME (sole carbon source), shaking at 28°C for 7 d at 200 rpm. After extracting the culture solution in three portions with 30 mL of dichloromethane, the solution was evaporated dried by rotation at 40°C, 500 Pa, eluted in two portions with 1 mL of chromatographic methanol and fixed to 1 mL, 100 μL of which was added into the injection vial containing the interpolation tube and detected by Liquid Chromatograph Mass Spectrometer (LC-MS) with methods described below. The solution with significantly reduced 3-OH PAME content, compared with the solution containing no soil or sawdust sample, was considered as a sample containing 3-OH PAME degrading bacteria.

Single colony bacteria were obtained and purified after streaking the solutions on LB plates. The degradation ability of the isolated strains was determined using the above method.

LC-MS for quantitative analysis of 3-OH PAME using Q Exactive Focus Liquid chromatography-mass spectrometry (Thermo Fisher, United States) was as follows: chromatographic column: ACQUITY UPLC HSS T3 Column (Waters, United States), flow rate of 0.3 mL/min, temperature of column: 40°C, mobile phase of methanol/water = 85/15, characteristic ion of Na⁺, molecular weight of 309, peak time of 5.5-6.5 min, injection volume of 10 µL.

To investigate the relationship between the bacterial growth and 3-OH PAME degradation, strains Q1-7 and Q4-3 were cultured in LB medium to OD₆₀₀ of 1.2 by shaking at 200 rpm at 28°C, then transferred into MSM+1 μM 3-OH PAME medium at a ratio of 1:100, shaking at 200 rpm at 28°C for 2 d. The MSM+1 μM 3-OH PAME medium added with of LB medium was served as the negative control. Samples were taken every 12 h to measure the concentrations of 3-OH PAME by LC-MS and the OD₆₀₀ values of the QQ strains by UV spectrophotometer. The experiment was repeated in three independent times.

To identify the taxonomic status of the QQ strains, Q1-7 and Q4-3 were grown in LB medium by shaking at 200 rpm at 28°C to OD₆₀₀ of 1.0. The genomic DNAs of both strains were extracted using an EasyPure Bacterial Genomic DNA Kit (Transgen, Beijing, China). DNA qualities were determined by agarose gel electrophoresis and Qubit Fluorometer Quantitation (Thermo Fisher Scientific, Waltham, MA). Genomic DNAs of Q1-7 were sent to Novogene (Tianjin, China), and Q4-3 was sent to Biomarker (Beijing, China) for sequencing. The genome sequence of strain Q1-7 was determined using the Hiseq × Ten PE150 in combination with the PacBio RS II platform. A mixed de novo assembly of PacBio long-line and Illumina short-line sequencing data was performed on the genome of Q1-7 using Unicycler (v.0.4.7). The Q1-7 genome was annotated using the Prokaryotic Genome Annotation Pipeline v.2019-05-13. build3740 (Tatusova et al., 2016). AMRFinder v.3.8.4 was used to identify antimicrobial resistance genes (Feldgarden et al., 2020); TXSScan was used to identify secretion systems (Abby and Rocha, 2017); and the tool dbCAN2 was used to identify carbohydrate-active enzymes (Zhang et al., 2018). A sequence search of the MEROPS database was used to identify proteases using the tool BlastP (v.2.9.0, NCBI) (Rawlings et al., 2010). The Signal v.4.1 program was used to detect proteins with signal peptides (Petersen et al., 2011). The genome sequence of strain Q4-3 was determined using the NovaSeq 6000 in combination with the Oxford Nanopore Technologies platform. For genome assembly, the filtered reads were assembled by Canu v1.5 software (Koren et al., 2017), and then Circlator v1.5.5 was used to cyclize genome assembly. For genome component prediction, Coding genes prediction was performed by Prodigal v2.6.3 (Hyatt et al., 2010). Transfer RNA (tRNA) genes were predicted with tRNAscan-SE v2.0 (Chan and Lowe, 2019), Ribosome RNA (rRNA) genes were predicted with Infernal v1.1.3. Repetitive sequences were predicted using RepeatMasker v4.0.5 (Tarailo Graovac and Chen, 2009).The whole genome sequences of strains Q1-7 and Q4-3 have been uploaded to NCBI under the accession numbers CP116304.1 (PRJNA922361) and CP115820.1 (PRJNA922444), respectively.

Pairwise average nucleotide identity (ANI) values between Q1-7, Q4-3, and all 327 Pseudomonas type genomes available in the NCBI RefSeq database ( Table S2 ) were calculated using fastANI v1.3. dDDH values between Q1-7, Q4-3, and their highest ANI values with Q1-7 or Q4-3 were then determined by using the GGDC server (http://ggdc.dsmz.de/). Orthologous genomes were constructed from annotated proteins in representative genomes of Q1-7, Q4-3 and 327 Pseudomonas species using OrthoFinder v2.3.2. Phylogenetic analysis of the Q1-7, Q4-3 and representative Pseudomonas genomes was inferred from the presence of single copy orthologs in most of the genomes analyzed. Single gene alignments were generated using MAFFT v7.490, and spot columns were trimmed using trimAl v1.4 and then concatenated into a supermatrix. Phylogenetic inference was performed using IQ-TREE v2.1.2 to automatically select the most appropriate evolutionary model for each gene, and the reliability of internal branching was estimated by ultra-fast bootstrap analysis with 1000 replicates.

Utilization of carbon sources of strains Q1-7 and Q4-3 were determined using Biolog GEN III microplates (bioMérieux) following the manufacturer’s protocol.

To determine the esterase activity of strains Q1-7 and Q4-3, bacteria were cultured in LB medium by shaking at 28°C at 200 rpm until OD₆₀₀ of 1.2, 1 μL of which was spotted onto the LB agar plates supplemented with 1% or 2% tributyrin and incubated in an incubator at 28°C for 2 d or 4 d (Lee et al., 2004). The diameters of the degradative halos were measured. The experiment was repeated three times.

To test the motility of the strains, Q1-7 and Q4-3 were cultured in LB medium by shaking at 200 rpm at 28°C until OD₆₀₀ of 1.2. Swimming, swarming and twitching media were prepared as described previously. Specifically, 15 mL of the medium was poured in a 90-mm diameter dish, followed by the addition of 1 μL of bacterial culture to the inside of the swimming medium, the surface of the swarming medium and the bottom of the twitching medium, and incubated at 28°C for 24, 36 and 24 h, respectively. The swimming and swarming plates can be photographed and measured directly, while the twitching plates were stained with 2% crystalline violet for 20 min, then rinsed three times with ddH₂O and then photographed and measured. The diameter of the bacterial movement was measured and collated using Image J 1. 52a. The experiment was repeated three times.

For measurement of the biofilms formed by strains Q1-7 and Q4-3, bacteria were cultured in LB medium by shaking at 200 rpm at 28°C until OD₆₀₀ of 1.2, 1 μL of which was added into 100 μL of SOBG medium in each well of polystyrene 96 well-tissue culture plate (Guangzhou Jet Bio-Filtration Co., Ltd., China). The plate was incubated for shaking at 150 rpm at 28°C for 48 h. The culture medium was discarded and 200 μL of 0.1% crystal violet (wt/vol) was added to the wells for staining. Each well of the 96-well plate was washed three times with water and dried in air, and 250 μL of 95% ethanol was added to dissolve the crystal violet from the stained bacterial biofilm cells. The amount of biofilm formation was quantified by measuring the absorption values at 595 nm on a Multifunctional Microplate Reader (MMR) (Microplate reader, BioTek, United States) (Chen et al., 2013). The experiment was repeated three times.

The antimicrobial susceptibility of the QQ strains was tested to further investigate their resistance (Jorgensen and Ferraro, 2009; Hu et al., 2021). Strains Q1-7 and Q4-3 were cultured in LB medium by shaking at 200 rpm at 28°C until OD₆₀₀ of 1.2, 1 μL of which was added into 100 μL of LB medium in each well of polystyrene 96 well-tissue culture plate (Guangzhou Jet Bio-Filtration Co., Ltd., China). The plate was incubated for shaking at 200 rpm at 28°C for 24 h. Antibiotics including ampicillin, kanamycin, streptomycin, gentamicin, tetracycline, and Polymyxin B were used in this experiment at concentrations of 5, 10, 20, 40, 80, 160, 320, and 640 μg/mL. The minimum antibiotic concentration with no visible cells was defined as MIC. Three replicates were performed for each treatment.

The Rdar phenotype, described as the red-dry-rough phenotype, can be used to detect components associated with biofilms such as EPS, cellulose and aggregated hairs (Romling et al., 1998). A final concentration of 0.005% Congo Red was added to the melted YEB medium, mixed and added to a 6-well plate with 3 mL of medium per well. The strains Q1-7 and Q4-3 were cultured in LB medium by shaking at 200 rpm at 28°C until OD₆₀₀ of 1.2. NS25 was cultured in CPG medium by shaking at 200 rpm at 30°C until OD₆₀₀ of 1.2. Every 1 µL of NS25+LB, NS25+Q1-7, NS25+Q4-3, Q1-7+CPG and Q4-3+CPG mixtures (1:1) was inoculated on the center of the well and incubated at 28°C, and photographed every 24 h. The experiment was repeated in triplicate.

In addition, NS25 was incubated with each of the QQ bacteria until OD₆₀₀ to 1.2, NS25+LB, NS25+Q1-7, NS25+Q4-3, Q1-7+CPG and Q4-3+CPG (1:1) were prepared, diluted and spread on the TTC plates and incubated upside down at 28°C. The effect of QQ bacteria on NS25 EPS yield was visualized by the extent and coverage area of EPS around NS25 colonies after 4 d. The experiment was repeated in triplicate.

Cell wall degrading enzyme activity was measured using the medium formulation described above (Zhou et al., 2016; Hu et al., 2021), and 40 mL of the medium was poured into a 130-mm diameter Petri dish, and wells in 5 mm diameter were punched. The strains Q1-7 and Q4-3 were cultured in LB medium by shaking at 200 rpm at 28°C until OD₆₀₀ of 1.2. NS25 was cultured in CPG medium by shaking at 200 rpm at 30°C until OD₆₀₀ of 1.2. Some of the NS25 bacterial culture was centrifuged at 12,000 rpm for 5 min, and the supernatant was filtered through a 0.22 μm filter. To test the cell degrading enzymatic activity of the strains, 20 μL of the bacterial culture(s)/NS25 supernatant was added to the well, and incubated at 28°C for 24 h. Pel and Peh plates were treated with 1 M HCl for 15 min. Cel plates were stained with 0.1% (w/v) Congo red for 15 min and decolored with 1 M NaCl. Prt plates were observed directly without any treatment. The diameter of the transparent halo was measured. To test the NS25 cell numbers in the wells, a concentric circle was punched around the well with an 8-mm diameter hole puncher. The cut medium was mixed thoroughly in 1 mL of CPG liquid medium, diluted in a gradient and applied to CPG plates supplemented with 50 μg/mL Polymyxin to calculate the CFU number of NS25. The experiment was repeated in triplicate.

The antagonistic activity of Q1-7 and Q4-3 against NS25 was tested by spot-on-law assay (Li et al., 2020). Firstly, bacterial cultures were grown in LB medium until OD₆₀₀ of 1.2. secondly, 300 μL of NS25 culture was added into 15 mL of 1% agarose (cooled to 50°C), mixed and poured onto the surface of 15 mL LB agar plates (9 cm × 9 cm), dried at room temperature, and then punched with a 5-mm puncher. Finally, 20 μL of overnight antagonistic candidate cultures were added to the wells. Plates were incubated at 28°C for 24 h. Antagonistic activity was assessed according to the size of the hyaline inhibition circle. The experiment was repeated in triplicate. the antimicrobial activity of NS25 strain against Q1-7 and Q4-3 was also determined using the above method.

Single colonies of strains Q1-7, Q4-3 and NS25 were grown until OD₆₀₀ of 1.2, and NS25+LB (1:1), NS25+Q1-7 (1:1) or NS25+Q4-3 (1:1) was added to 10 mL of CPG medium at a ratio of 1:50 for 7 h at 28°C for RNA extraction. RNA was extracted using the SV Total RNA Isolation System Kit (Promega, Madison, WI, USA) and further purified using the RNA Cleanup Kit (Qiagen, Hilden, Germany). DNA contamination was eliminated with DNase I. RNA purity and quality were checked by gel electrophoresis and NanoDrop 2000c (Thermo Fisher Science, Waltham, MA, USA). For RT-PCR analysis, 0.5 μg of RNA was reverse transcribed using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme Biotech Co., Nanjing, China) to generate template cDNA. Expression of the NS25 internal reference gene infB was selected to balance the cDNA concentrations in the three samples. The primers used for PCR amplification were listed in Table S4 . RT-PCR was performed using 2× Rapid Taq Master Mix (Vazyme Biotech Co., Nanjing, China) with the following cycling pattern. 1 cycle at 95°C for 3 min; followed by 40 cycles at 95°C for 15 s, 56°C for 15 s, and 72°C for 15 s; 1 cycle at 72°C for 5 min; and kept at 16°C. The expression of each gene was determined by measuring the signal intensity of the bands using Image Lab software (Bio-Rad, Hercules, CA, USA). Experiments were performed in triplicate.

Single colonies of strains NS25, NS25(ΔphcB), NS25(ΔsolI) were grown until OD₆₀₀ of 1.2, and diluted into fresh CPG medium in 1:100 ratio. Bacteria were grown with shaking at 200 rpm under 30°C, and cell density was measured at 0, 6, 12, 24, 30, 36, 42, 48, 54, and 60 h respectively. The experiment was repeated three times in triplicate.

To evaluate the efficacies of strains Q1-7 and Q4-3 on biocontrolling the bacterial wilt disease, one-year old Casuarina, one-month old tomato and peanut seedlings were used for the pathogenicity tests. Strains Q1-7 and Q4-3 were cultured in LB medium by shaking at 200 rpm at 28°C until OD₆₀₀ of 1.2. Strain NS25 was cultured in CPG medium by shaking at 200 rpm at 30°C until OD₆₀₀ of 1.2. For testing the pathogenicity on Casuarina, root-drenching method was used by pouring 10 mL of NS25+Q1-7 (1:1) or NS25+Q4-3 (1:1) into the soil close to the plant roots (Kanda et al., 2003). For testing the pathogenicity on tomato and peanut seedlings, 200 µL of NS25+Q1-7 (1:1) or NS25+Q4-3 (1:1) was injected into the center of the tomato or peanut pseudostem using 1.0 mL needleless syringes. CPG+LB broth without bacteria, and CPG+Q1-7/Q4-3 were served as the blank and negative controls, respectively, and equal volume of NS25+LB was set as the positive control. Inoculated plants were placed in a cabinet with a 14-h light/10-h dark cycle at 28°C. The experiment was repeated three times with twelve seedlings each time. All the plants were monitored for disease analysis, which was set as: survival (%) = (1 - diseased plant number/total plant number) × 100%.

For measurement the NS25 CFU in the plants, each three tomato seedlings were inoculated with 10 mL of CPG+LB, NS25+LB, NS25+Q1-7, NS25+Q4-3, Q1-7+CPG and Q4-3+CPG bacterial cultures, respectively, by irrigation into the pots with some of the roots cut off. After 7 days post inoculation (dpi), 0.5 g of plant roots or stems were cut and washed with sterile water three times, homogenized with 1 mL of ddH₂O and diluted in series. Dilutions were spread onto TTC plates. The number of NS25 CFU was counted after 3 days. Three independent tests were performed.

The statistical analysis was performed using GraphPad Prism (San Diego, CA). The results were analyzed by Student’s t-test, where * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, and **** indicates p < 0.0001.

In order to directionally cultivate QQ strains in original habitats for the control of bacterial wilt disease, C. equisetifolia branches and forest soil were collected from Casuarina planting forest in Haitouwan Forest in Zhanjiang city, Guangdong province. MSM medium with 3-OH PAME as the sole carbon source was used to enrich QQ candidate bacteria. Firstly, a standard curve of 3-OH PAME concentrations was established using a serial dilution of 3-OH PAME that generates an equation as Y = 1267805180 × X + 77323744 with a high R² as 0.9973, where Y represents the concentration of 3-OH PAME, X represents the mean AA numerical value of LC-MS ( Supplementary Figure S1 ). Secondly, crushed branches and soil samples were respectively added into the MSM+3-OH PAME medium, and cultures with significantly reduced concentration of 3-OH PAME were selected for bacterial separation. Subsequently, single bacterial colony was added into the MSM+3-OH PAME medium and the concentration of 3-OH PAME was measured. As a result, 11 strains with degrading activity of 3-OH PAME were obtained, including 4 strains (Q1-4, Q1-7, Q1-8 and Q1-9) isolated from the forest soil and the other 7 isolated from C. equisetifolia branches ( Supplementary Figure S2A ). Among these strains, strain Q1-7 and strain Q4-3 showing high 3-OH PAME degradation rates over 85% were selected for further experimental study ( Supplementary Figures S2A, B ).

To determine the 3-OH PAME degradation efficiency of Q1-7 and Q4-3 strains, bacterial cultures grown in MSM with 1 µM 3-OH PAME were sampled at different time points for measuring the residual 3-OH PAME concentrations and corresponding cell densities. At 12 hours post inoculation (hpi), the degradation rate of 3-OH PAME was 69.03% by strain Q1-7, and 88.73% by strain Q4-3; at 24 hpi, 88.78% and 95.97% of 3-OH PAME was respectively degraded by strains Q1-7 and Q4-3; at 48 hpi, 95.80% of 3-OH PAME was essentially degraded by strain Q1-7, and 100% by strain Q4-3 ( Figure 1 ). At time points of 12 hpi, 24 hpi and 48 hpi, the corresponding OD₆₀₀ of strain Q1-7 was 0.037, 0.064, and 0.080, respectively, and which of strain Q4-3 was 0.033, 0.040, and 0.060, respectively ( Figure 1 ). The results showed that the concentrations of 3-OH PAME in the solution dramatically decreased at the first 24 h with bacteria gradually grew ( Figure 1 ), suggesting that strains Q1-7 and Q4-3 could degrade 3-OH PAME by using it as a carbon source for bacterial growth.

To better understand the molecular basis of the 3-OH PAME quenching ability, the genomes of Q1-7 and Q4-3 were sequenced using both Nanopore PromethION platform and Illumina NovaSeq platform. The genomes of Q1-7 and Q4-3 were both assembled into single, circular chromosomes of 5,780,855 bp and 6,541,745 bp in sizes, with 64.77% and 66.97% in GC-content, respectively ( Figure 2 ; Supplementary Table S1 ). A total of 5,146 and 5,826 protein-coding genes, 84 and 167 RNA genes, both 12 rRNA genes, 101 and 15 transposases, and 97 and 51 pseudogenes were respectively predicted in the genomes of Q1-7 and Q4-3 ( Supplementary Table S1 ).

To determine the taxonomic classifications of Q1-7 and Q4-3, we compared their genomes with 327 Pseudomonas type genomes available in the NCBI RefSeq database, which altogether represent all the known Pseudomonas species. We first calculated the Average Nucleotide Identity (ANI) values between all genomes. The results showed that Q1-7 is most closely related to Pseudomonas lalkuanensis PE08 (GCF_008807375.1, isolated from contaminated soil collected from a paper mill yard in Lalkuan, Uttarakhand, India) with an ANI value of 88.5638% ( Supplementary Table S2 ) and a dDDH value of 33.30%, both below the commonly accepted thresholds (ANI ≥ 95%, and dDDH ≥ 70%) for species delineation. Therefore, we proposed to name the novel species represented by Q1-7 as Pseudomonas forestsoilum sp. nov. after the species name of its isolated substrate. For Q4-3, the closest relative is Pseudomonas tohonis TUM18999 (AP023189.1, isolated from the skin of a patient with burn wounds in Japan) with an ANI value of 97.8949% ( Supplementary Table S2 ) and a dDDH value of 81.90%, indicating that Q4-3 is classified as P. tohonis.

In order to visualize the evolution of Q1-7 and Q4-3 in the Pseudomonas genus, we constructed a phylogenetic tree based on 120 single-copy orthologous genes between each of them and 327 type strains of the Pseudomonas to represent their relationships. The phylogenetic tree based on whole genome sequences showed that Q1-7 and Q4-3 both belong to the Pseudomonas resinovorans group ( Figure 3 ), and Q1-7 is over 88% identical to its sister branch including P. lalkuanensis PE08 and P. resinovorans DSM 21078, suggesting that Q1-7 represents a new species. Q4-3 is closest to P. tohonis TUM18999, locating on the same branch, indicating that strain Q4-3 is P. tohonis. The evolutionary distance between the two strains is farther apart, consistent with the lower ANI (85.3783%) and dDDH values (26.50%) between them ( Supplementary Table S2 ).

In addition, biochemical characteristics of strains Q1-7 and Q4-3 were also tested, which revealed most consistent features between the two strains except that strain Q1-7 could use D-Galacturonic acid and L-Galactonic acid, while strain Q4-3 had a weak response to them; strain Q1-7 could not use D-Melibiose, α-D-Glucose, D-Fructose, L-Lactic acid, L-Alanine and L-Glutamic acid, while strain Q4-3 is weakly responsive to them; strain Q1-7 has a weak response to D-Fucose and Lincomycin, while Q4-3 could use them ( Table 1 ).

The motility of bacteria determines their ability to colonize and expand in the plant root zone. To defend the host against pathogenic bacteria, the motility of the biocontrol bacteria can also have an impact on its ability to defend itself (Piromyou et al., 2015; Liu et al., 2017). Thus, we tested the swimming, swarming and twitching motilities of strains Q1-7 and Q4-3 to reflect their colonization and expansion potential in soil rhizosphere or plants. The results showed that strain Q1-7 showed strong swimming and swarming motility, but weak twitching motility, while strain Q4-3 showed stronger swarming and twitching motilities than strain Q1-7 ( Figure 4A ). The significant motility difference between the two strains may be related with their different isolated habitats, where strain Q1-7 was isolated from the forest soil and strain Q4-3 was isolated from C. equisetifolia branch.

Biofilm is a membrane-like structure that forms under specific conditions and is formed when bacteria are adsorbed onto a solid surface and then proliferate (Branda et al., 2005). We tested the biofilm of strains Q1-7 and Q4-3 on SOBG medium. The results show that both Q1-7 and Q4-3 strains can form biofilms, with Q1-7 producing much less than Q4-3 ( Figure 4B ).

In addition, we tested the susceptibility of strains Q1-7 and Q4-3 to six different antibiotics. The results showed that they are sensitive to most tested antibiotics except ampicillin, and strain Q4-3 is also resistant to streptomycin with minimal inhibitory concentration (MIC) up to 40 µg/mL ( Supplementary Table S3 ).

The most important signaling molecule in the pathogenic regulatory networks of R. solanacearum is 3-OH PAME, the chemical nature of which is an ester compound. To confirm the ability of strains Q1-7 and Q4-3 to degrade esters, the diameters of the degradation halos on the LB plates containing tributyrin were observed to reflect their esterolytic abilities. Degradation halos were produced after the addition of strains Q1-7 and Q4-3 bacterial solution to plates containing tributyrin, indicating that strains Q1-7 and Q4-3 can produce esterase to degrade tributyrin. On the 2^nd day, both strains showed comparable esterolytic activities on the plates with 1% and 2% tributyrin, whereas, on the 4^th day, both strains had significantly greater degradation activities on the plates with 1% tributyrin than on those with 2% tributyrin ( Figure 5 ).

R. solanacearum NS25, Q1-7 and Q4-3 were all able to produce extracellular polysaccharides (EPS). On YEB plates, the EPS produced by NS25 are smooth and sticky with neat edge, while those produced by Q1-7 and Q4-3 are dry and slightly protruding with irregular edge ( Figure 6A ). After mixing NS25 with either QQ bacteria at a ratio of 1:1, the EPS formed by NS25+Q1-7 were biased towards NS25 with significantly less EPS than that of the NS25+LB group, while those formed by NS25+Q4-3 were biased towards Q4-3 ( Figure 6A ). On TTC plates, the bacterial morphology between NS25 and the two QQ strains could be visually distinguished. NS25 formed very thick and viscous EPS, while Q1-7 and Q4-3 seemed dry and flat. On both NS25+Q1-7 and NS25+Q4-3 group plates, the NS25 colonies formed significantly reduced EPS ( Figure 6B ), suggesting that both Q1-7 and Q4-3 could inhibit the yield of EPS in NS25, with Q4-3 had a stronger inhibitory capacity.

R. solanacearum NS25 produced cellulases and polygalacturonases, but not proteases or pectinases ( Supplementary Figure S3A ). Strains Q1-7 and Q4-3 did not produce plant cell wall degrading enzymes except that strain Q4-3 produced proteases ( Supplementary Figure S3A ). There was no significant difference in the cellulase degradation sizes produced by NS25 broth and supernatant, indicating that NS25 hardly grew on cellulase plates ( Supplementary Figure S3B ). Both NS25+Q1-7 and NS25+Q4-3 co-cultures produced significantly less cellulases than the NS25+LB positive control ( Figure 6C ). To test whether such difference is attributed to the lower cell density of NS25 in the co-cultures than that in the positive control, we measured the NS25 cell numbers in the cultures. The results showed that there was no significant difference between the cell numbers of NS25 in the mixed and pure cultures ( Figure 6D ), indicating that both Q1-7 and Q4-3 strains reduced the cellulase activity of NS25. We also measure the antagonistic activities between the pathogenic bacterium NS25 and quorum quenching bacteria Q1-7 and Q4-3, and found that Q1-7 was not antagonistic to the pathogen NS25, while Q4-3 showed slightly visible inhibition activity against NS25 ( Supplementary Figure S4A ), and the NS25 pathogen had no inhibitory activity against either Q1-7 or Q4-3 ( Supplementary Figure S4B ).

To investigate the consequence of the 3-OH PAME quenching caused by strains Q1-7 and Q4-3, we measured the expression of the phc QS regulated downstream genes. The results showed that co-culture of NS25 with Q1-7 or Q4-3 resulted in slightly down expression of the phc system (phcA and phcS) and dramatically down expression of the sol system (solI and solR), and significantly down expression of genes related to EPS and cellulase production ( Figure 7 ). This indicates that quenching of 3-OH-PAME caused by strains Q1-7 and Q4-3 could reduce the expression of phc regulated genes. Previously, we have obtained the deletion mutants of phcB and solI. Comparison of the growth curves between NS25, NS25(ΔphcB) and NS25(ΔsolI) revealed that before 24 h, NS25(ΔphcB) grew slightly faster than the other two strains, and no obvious difference could be observed between the strains after 24 h ( Supplementary Figure S5 ), indicating that these phc regulated genes are not closely related to the pathogen growth.

To determine the efficacy of the Q1-7 and Q4-3 strains on controlling or preventing bacterial wilt caused by R. solanacearum, three different host plants such as Casuarina, peanut, and tomato, were selected for pathogen inoculation. As shown in Figure 8A , plants inoculated with R. solanacearum NS25 showed symptomatic bacterial wilting, while that inoculated with CPG+LB medium (blank control), NS25+Q1-7, NS25+Q4-3, Q1-7+CPG and Q4-3+CPG (negative controls) kept in a good growth condition. The average survival was calculated based on the daily recorded plant incidence. The results showed that Casuarina seedlings inoculated with NS25+LB began to wilt at 15 dpi, and all died at 20 dpi, while only 10% and 20% of the seedlings inoculated with NS25+Q4-3 and NS25+Q1-7, respectively, died at 20 dpi, and all the plants in the blank and negative control groups kept healthy ( Figure 8B ). Tomato seedlings inoculated with NS25+LB began to wilt at 2 dpi and all wilted at 7 dpi, while all the seedlings kept healthy in other treatment groups ( Figure 8B ). Peanut seedlings inoculated with NS25+LB began to wilt at 6 dpi, and 75% died at 16 dpi, while only 10% of the seedlings inoculated with NS25+Q4-3 died at 10 dpi, and all the plants inoculated with NS25+Q1-7, and those in the blank and negative controls kept healthy ( Figure 8B ), suggesting that both strains are effective in prevention and control of plant bacterial wilt caused by R. solanacearum, especially for peanut and tomato bacterial wilt.

Calculation of the NS25 cell number in tomato roots and stems indicated that the CFU of NS25 slightly reduced in the roots inoculated with NS25+Q1-7, but was similar in those inoculated with NS25+Q4-3, compared with those in the positive control, and the CFU number in the stems was not obviously different in the NS25+LB, NS25-Q1-7 and NS25+Q4-3 groups ( Figure 8C ), suggesting that the quenching bacteria do not reduce the occurrence of bacterial wilt by inhibiting the growth of NS25.