Bacterial and fungal strains, plasmids and growth conditions. Phytopathogenic Xanthomonas spp. were cultivated in nutrient agar (NA) at 28˚C59, and E. coli strains DH5α and BL21(DE3) were cultured in Luria Bertani (LB)60 medium at 37˚C. Tryptone soy broth (TSB; 30 g/L) was used for culturing Pseudomonas strains at 30˚C. The fungus M. oryzae R01-1 was grown in oatmeal (OAM)61 medium at 25˚C. The strains, plasmids and primers used in this study are listed in Supplementary Tables 1 and 2. When required, antibiotics were added at the following final concentrations (µg mL− 1): kanamycin (Km), 25; rifampicin (Rif), 75; gentamicin (Gm), 20; and spectinomycin (Sp), 25. Chemical reagents including petroleum ether, ethyl acetate, n-butyl alcohol and methanol were purchased from Macklin (Shanghai, China). M. oryzae R01-1 was obtained from Dr. Jiangbo Fan, School of Agriculture and Biology, Shang Hai Jiao Tong University. Phenazine-1-carboxylic acid (PCA) was provided by Dr. Yawen He, School of Life Sciences and Biotechnology, Shang Hai Jiao Tong University. Restriction enzymes were purchased from Takara Bio (Europe AB).
Antibacterial tests. To isolate bacteria with antimicrobial activity for Xoc RS105, 248 soil samples were collected from the rhizosphere as described previously5. Antagonism of soil bacteria towards Xoc, Xoo and other pathogens was assessed by the Kirby-Bauer (KB)62 method. All strains were tested in triplicate and their inhibition zones were measured after 2–3 d cultivation at 28°C.
Genomic analysis and phylogenomics. The 16S rRNA of strain 923 was amplified using universal primers 27F and 1492R (Supplementary Table 2) using established methods5. The complete gene was purified and sequenced by BioSune Biotech (Shanghai, China) and used for Blast searches in the National Center for Biotechnology Information (NCBI) database. The 16S rRNA sequences for 14 representative Pseudomonas spp. were obtained from the NCBI database, and a phylogenetic tree based on 16S rRNA was constructed using neighbor-joining analyses with MEGA 7.063. A bootstrap test with 1000 replicates was used to evaluate the confidence level, and the internodes of branches represented percentage of confidence > 50%. The whole genome of 923 was sequenced using the Illumina Miseq PacBio sequencing platform at Personalbio (Shanghai, China). Average nucleotide identity (ANI) values were calculated using the J Species WS online service64. To determine the precise phylogenetic position of strain 923, all publicly available Pseudomonas genome sequences and related strains were retrieved from the NCBI GenBank database and included in the phylogenetic analyses according to the Type Genome Server (TYGS) (https://tygs.dsmz.de). The whole-genome sequence of strain 923 was deposited in the NCBI BioProject Database (BioProject ID: PRJNA826312).
Isolation and purification of antibacterial metabolites from P. mosselii 923. Strain 923 was cultured in 50 mL LB broth at 30°C with aeration at 220 rpm for 12 h, then diluted 1:100 into 2 L flasks containing 300 mL LB broth, and incubated for 24 h at 30°C, 220 rpm. The fermentation was extracted three times with an equal volume of petroleum ether, ethyl acetate and n-butyl alcohol at room temperature, respectively. The organic and aqueous phases were concentrated with a rotary evaporator, freeze-dried, and resuspended in 1 mL methanol and sterile water, respectively. Fifty microliters of the above four samples were removed and assessed for antibacterial activity using Xoo PX099A as previously described5.
The ethyl acetate extract exhibited the greatest antagonistic activity against Xoo PX099A. The large-scale fermentation was carried out and a 10 L batch of the fermentation was further extracted by ethyl acetate. The gelatinous crude extract was then loaded onto a C18 reversed-phase silica gel column (MeOH/H2O, 3:7 to 9:1). Twelve fractions were obtained from the column and further purified by high performance liquid chromatography (HPLC) with a C18 column (Agilent C18 column, 250 × 20 mm, 5 µm, USA) (method A). Using this approach, seven compounds were isolated and designated PM-1, PM-2, PM-3, PM-4, PM-5, PM-6 and PM-7. The antagonistic activity of each compound was tested against PX099A after concentration with methanol as a control.
HPLC preparation and analysis method: Method A (preparation): The initial concentration of 15% methanol was maintained for 2 min and elution gradients proceeded as follows: linear gradient of methanol from 15–100%, 2–18 min; 100% methanol, 18–22 min; 100–15% methanol, 22–26 min; and 15% methanol, 26–30 min. The flow rate was 5 mL/min with a UV detector (wavelength: 210 and 500 nm). Method B (analysis): HPLC using solvents A (water) and B (methanol) with a flow rate of 0.4 mL/min (Agilent C18 column, 150 × 4.6 mm 5 µm, USA, 210 and 500 nm). The initial concentration of 5% methanol was maintained for 4 min, followed by the following gradients: 5–20% methanol, 4–25 min; 20–100% methanol, 25–26 min; 100% methanol, 26–32 min, 100% − 5% methanol, 32–33 min, and 5% methanol, 33–40 min.
Structural elucidation. The molecular formula of PM-3 was determined to be C6H7N5O by UPLC/Q-TOF-MS (m/z 166.0722 [M + H]+, calculated for C6H8N5O, 166.0729) with six IHDs (index of hydrogen deficiency). The 1H and 13C NMR data (Supplementary Fig. S4b-c) showed the presence of two methyls (δH 4.27, 4.44; δC 43.3, 57.7) and one aromatic proton (δH 9.14). Because there are too many quaternary carbon and nitrogen atoms in the molecule, it was difficult to further determine the planar structure of PM-3. Fortunately, a tiny crystal was generated in the methanol solvent by the solvent evaporation method after many attempts. Finally, the complete structure of PM-3 was determined by X-ray diffraction (Fig. 2c) and it was identical to pseudoiodinine.
Biocontrol assays. Twenty seeds of rice cultivar Yuanfengzao were sown in plastic pots (16 × 12 cm), which were maintained in the greenhouse at 14 h light (30˚C)/10 h dark (28˚C) and 65% RH. Rice seedlings were used for biocontrol assays after 21 d of cultivation (3–6 leaf stage). Cell suspensions of Xoc RS105 and Xoo PXO99A (OD600 = 0.6) and P. mosselii 923 (OD600 = 0.5) were prepared in advance. In brief, these strains were cultivated to mid-exponential phase, centrifuged, and washed twice with sterile distilled water to an OD600 of 0.5–0.6. The supernatant of the 923 culture (SUP) was prepared by centrifugation at 4°C, 12,000 rpm for 10 min. Pseudoiodinine stock solutions with MIC values of 4 µg/mL for Xoc RS105 and 0.5 µg/mL for PXO99A were prepared. Sterile distilled water (MOCK) was used as a control.
Biocontrol assays on rice seedlings of three-leaf stage were inoculated by needleless syringe as described previously5 and were maintained in a greenhouse as described above. Disease lesion lengths were measured at 7 d. Yuanfengzhao rice plants at the six-leaf stage were sprayed with bacterial cell suspensions (2 mL of 1.0×108 cfu mL− 1). Treatments were as follows: (1) rice plants inoculated with strains RS105 or PXO99A; (2) rice inoculated with pathogens RS105 or PXO99A and then treated with strain 923 (2 mL suspension) or pseudoiodinine (psd-Tre); (3) rice inoculated with strain 923 (2 mL) or pseudoiodinine 12 h prior to spray-inoculation with RS105 or PXO99A (psd-Pre); and (4) sterile distilled water-inoculated controls (MOCK). Disease lesion areas were measured on leaves (n = 20) at 15 d and bacterial growth were at 7, 14, and 21 d after inoculation. Briefly, three 4-cm pieces from the leaf tip were excised with sterile surgical scissors, and treated with 75% ethanol (30 s), 3% sodium hypochlorite (3 min) and sterile-distilled water (1 min). Leaf material was then macerated with a tissue grinder (JX-FSTPRP) at 55 Hz (twice for 1 min) and then maintained at room temperature for 30 min. Serial dilutions were prepared and incubated on NA supplemented with the appropriate antibiotics at 28°C for 3–4 d until single colonies could be counted (> 100/plate). The experiment was repeated three times with similar results.
For biocontrol assays in the field, rice cultivar Yuanfengzao was sown in paddies (1 × 2 m) and 50 plants were inoculated by spraying as described above. BLS and BLB disease severity was investigated on leaves (n = 15) at 15 d after inoculation. Field experiments were repeated three times independently with similar results.
β-glucuronidase (GUS) assays. Our previous research demonstrated that pHM1-derived vectors systems including pHG1 were efficient, stable and suitable for gene expression analysis in X. oryzae strains and for tracking infections in rice leaves40. Prior research with pHM1 vectors indicated that the intensity of GUS staining in rice leaves correlated with bacterial multiplication; thus, the GUS activity of Xoc RS105-Gus strains including staining and quantitative assays were conducted following our previous protocol65. Each treatment was replicated on three leaves, and three independent experiments were performed with the similar results.
GUS activity in P. mosselii strains were performed using a modified protocol. Briefly, P. mosselii 923 and mutants were cultured in LB containing spectinomycin at 30°C overnight. Bacterial cells were collected in 2 mL microcentrifuge tubes (three replicates), washed twice with 1 mL sterile water, and adjusted to an OD600 of 0.5. Qualitative and quantitative GUS analyses of P. mosselii 923 and mutants were conducted as described previously65.
Construction of promoter-probe vectors. The multiple cloning sites (MCS) in the promoter probe vector pNG1 were positioned upstream of uidA, and the NdeI site overlapped with the translational start site (ATG). Thus we created uidA promoter-probe fusions at the transcriptional and post-transcriptional level. The psdA promoter region was cloned into pNG1 as an EcoRI/KpnI fragment to create pNG1-P2064 with the psdA promoter-uidA fusion (Supplementary Table 1). Using a similar approach, the psdA promoter was cloned into pNG1 as an EcoRI/NdeI fragment, resulting in a translational fusion with uidA; the resulting construct was designated pNG1-P2064-post (Supplementary Table 1).
EZ-Tn5 mutagenesis screening and identification. Mutagenesis of P. mosselii strain 923 with EZ-Tn5 and characterization of insertion sites was conducted using established protocols5.
Generation of gacA deletion, complementation and overexpression strains. A gacA deletion mutant was generated by sacB-mediated double homologous recombination. Sequences flanking gacA at 716 bp upstream and 487 bp downstream were amplified by PCR using primers gacA-up-F1/gacA-up-R1 and gacA-down-F2/gacA-down-R2 (Supplementary Table 2), respectively. The upstream and downstream amplified products were gel-purified, digested with NdeI/KpnI and XbaI/NdeI, respectively, and cloned into p2P24Km66 to obtain construct p24-gacA (Supplementary Table 1). After ligation, this construct was introduced into E. coli DH5α competent cells and cultured on LB containing Km. Colonies containing p24-gacA were confirmed by PCR with M13-F and M13-R primers (Supplementary Table 2) and sequenced. Plasmid p24-gacA was introduced into P. mosselii 923 by electroporation, and colonies with kanamycin resistance and sucrose sensitivity were selected on NAN (NA without sucrose) containing Km (NAN + Km) and NAS (NA containing 10% sucrose) in succession. NAS-surviving colonies were individually cultured on NA and NA + Km, and colonies that grew on NA but not NA + Km were confirmed by PCR with primers in Supplementary Table 2. Deletion of gacA was confirmed by sequencing analysis and further verified by testing for bacteriostasis and pseudoiodinine production as described above.
A 1074-bp fragment containing gacA and the upstream promoter region was amplified by PCR with primers gacA-com-F and gacA-com-R (Supplementary Table 2). The PCR product was cloned into pUFR034 as an EcoRI/KpnI fragment to yield the recombinant plasmid pUFR-gacA; this construct was confirmed by PCR with M13 primers, sequenced, and then introduced into P. mosselii 923 and the ΔgacA mutant by electroporation. Colonies were selected on LB containing Km, identified by PCR and further confirmed by testing for bacteriostasis and pseudoiodinine production. The complemented mutant and gacA overexpressing strains were designated CΔgacA and 923 pUFR-gacA, respectively.
RNA-seq and transcriptome analyses. Wild-type P. mosselii 923 (WT) and the knockout mutant ΔgacA (KO) were cultured in LB for 24 h. Three biological replicates were prepared of each strain. Bacterial cells (1 mL) were collected and centrifuged for 2 min at 10,000 rpm at 4°C; this was repeated with another 1 mL aliquot of cells. Supernatants were removed, frozen in liquid nitrogen for 15 min, and stored at -80°C until needed.
RNA-seq was performed by Shanghai Personal Biotechnology Co., Ltd. with the Illumina HiSeq system. DESeq v. 1.18.0 was used to analyze differentially expressed genes (DEGs) of KO compared to WT and a corrected P value (q value) < 0.005 and a log2 fold-change > 1 were used to establish significance. Volcano plots were created using the R language ggplots2 package and plot_volcano from soothsayer (https://github.com/jolespin/soothsayer) in Python v. 3.6.6. Heatmaps were produced by R language and the Pheatmap software package (https://rdrr.io/cran/pheatmap/). Euclidean and complete linkage methods were used to calculate distance and clustering, respectively.
Quantitative RT-PCR. To verify RNA-seq data, three independent real-time quantitative PCR (RT-qPCR) analyses were performed on WT and KO samples using the same treatments as used for RNA-seq. Fifteen DEGs were randomly selected for secondary confirmation by RT-qPCR analysis (Supplementary Fig. S11). Total RNA of WT and KO samples were extracted using the EasyPure RNA Kit (Transgen Biotech). cDNA was prepared with the cDNA Synthesis Super Mix Kit (Transgen Biotech). SYBR green-labelled PCR fragments were amplified, and RT-qPCR was performed with an ABI7500 Real-Time PCR System (Applied Biosystems, USA). rpoD was used as an internal control and reference gene, and the 2−ΔΔCt method was used for relative quantification67. All RT-qPCR reactions were performed three or more times using primers listed in Supplementary Table 2.
Generation of psd deletion mutants and complementation analyses. Deletion of individual psd genes was accomplished by sacB-mediated double homologous recombination as described above. The primers used to amplify sequences flanking each gene are listed in Supplementary Table 2. Deletion mutants were identified by PCR with corresponding primers (Supplementary Table 2) and further evaluated for bacteriostasis and pseudoiodinine production. The other deletion mutants described in this study were also generated using sacB-mediated double homologous recombination.
For complementation analyses, psdA, psdB, psdC, psdD, psdE, psdF, and psdG and the psdA promoter were amplified and cloned in pBSPPc to obtain plasmids pBS-psdA, pBS-psdB, pBS-psdC, pBS-psdD, pBS-psdE, pBS-psdF and pBS-psdG, respectively. (Supplementary Table 1). Constructs then were introduced into the corresponding deletion mutants and tested for complementation. The rsmY and rsmZ mutants were complemented through cloning in pUFR034 using the same protocol.
Plasmid construction with One-Step ClonExpress technology. Plasmids pBSPPc-Pseu-ORF and pNG1-P2064 were constructed as described above. The primers used are listed in Supplementary Table 2 and were synthesized by Shanghai Generay Biotech Co., Ltd. Linearized pBSPPc-Pseu-ORF was obtained by reverse PCR with primers pBS-Pseu-F2 and pBS-Pseu-R2 and mixed with linearized pNG1-P2064 (Supplementary Fig. S10d); recombination proceeded using protocols supplied with the ClonExpress II One Step Cloning Kit (Vazyme, Nanjing, China). The recombination product, pBSPPc-P2064-RBS-Pseu-ORF, was used to transform recipient cells.
Heterologous expression of pseudoiodinine in P. putida KT2440. The seven gene pseudoiodinine operon (psdABCDEFG) was cloned with its native promoter in pBSPPc. The 7.2-kb DNA fragment was amplified with primers pBS-Pseu-F and pBS-Pseu-R (Supplementary Table 2) and purified using a gel extraction kit. The 7.2-kb fragment was inserted into XbaI/BamHI-digested pBSPPc and cloned using directions provided with the ClonExpressII One Step Cloning Kit. The resulting construct was designated pBSPPc-ABCDEFG; this was introduced into P. putida KT2440 through electroporation and verified by PCR.
The engineered P. putida strain was cultivated in 20 mL TSB supplemented with 20 µg/mL gentamicin, and cells were grown at 30°C, 220 rpm, for 24 h. The fermentation broth was centrifuged, and the supernatant was extracted three times using equal volumes of ethyl acetate. The combined organic phases were concentrated and redissolved in methanol for HPLC analysis (Method B).
Generation of psd operon overexpressing strains. For overexpression experiments, the plasmid pBSPPc-ABCDEFG (Supplementary Table 1) contained the psdABCDEFG operon and its native promoter was introduced into P. mosselii ΔcsrA1A2A3 as described above and analyzed for pseudoiodinine production.
Heterologous complementation. The pseudoiodinine homologous gene cluster (orf2184-2190) from P. mosselii DSM17497 and its native promoter was cloned as a XbaI/BamHI fragment in pBSPPc, resulting in pBS-DSM17497-Pseu (Supplementary Table 1). This clone was introduced into the P. mosselii 923 ∆psdB, ∆psdC and ∆psdG mutants to determine if the DSM17497 psd operon could complement the 923 mutant strains. Two strains of each mutant (PsdB-comp.1 and PsdB-comp.2; PsdC-comp.1 and PsdC-comp.2, and PsdG-comp.1 and PsdG-comp.2) were generated by introducing plasmid pBS-DSM17497-Pseu into the corresponding mutants by electroporation (Supplementary Table 1). Complemented mutants were evaluated for bacteriostasis and pseudoiodinine production as described above.
RT–PCR analysis in P. mosselii DSM17497. DSM17497 contains seven genes homologous to the pseudoiodinine operon in P. mosselii 923, namely orf2184 (psdA), orf2185 (psdB), orf2186 (psdC), orf2187 (psdD), orf2188 (psdE), orf2189 (psdF) and orf2190 (psdG). RT-PCR was used to determine whether the seven putative pseudoiodinine genes were expressed in P. mosselii DSM17497. Total RNA of P. mosselii DSM17497 was extracted using the EasyPure RNA Kit (Transgen Biotech), and cDNA was prepared with the cDNA Synthesis Super Mix Kit (Transgen Biotech). The 16S rRNA gene was used for normalization. The primers used for RT-PCR are listed in Supplementary Table 2, and the experiment was repeated three times with similar results.
Expression and purification of GacA. The gacA ORF was amplified by PCR with primers gacA-F and gacA-R (Supplementary Table 2) and cloned into pET28a as a NdeI/XhoI fragment. The resulting construct, pET28-gacA, was transformed into E. coli BL21 (DE3). For overexpression, a single colony E. coli BL21 colony was inoculated into 20 mL LB with Km and cultivated overnight at 37℃. Cells (5 mL) were then transferred to 500 mL LB and grown at 37°C, 220 rpm to OD600 = 0.6–0.8; cells were then induced with 0.5 mM IPTG and cultivated overnight at 16°C. Pellets were harvested by centrifugation at 4°C, 6000 rpm for 10 min, washed twice in PBS, and suspended in buffer A (50 mM Tris-HCl, 300 mM NaCl, pH 7.5) containing a final concentration of 1 mM PMSF (phenylmethanesulfonyl fluoride). Cells were disrupted by sonication, and cell free extracts were obtained by centrifugation at 4°C, 12,000 rpm for 30 min. Supernatants were applied to His Sep Ni-NTA Agarose Resin (Yeasen Biotechnology), which was equilibrated with buffer A prior to use. The Ni-NTA column was washed three times with buffer A, and the column was eluted with a gradient of 20–250 mM imidazole in buffer A. Fractions from 20 to 250 mM were pooled and separated by SDS-PAGE. The fraction containing GacA was concentrated and exchanged with storage buffer (50 mM Tris-HCl, 150mM NaCl, 5% glycerol, pH 7.5). The protein concentration was estimated using a Nano-300 MicroSpectrophotometer (Hangzhou Allsheng Instruments Co.), and the preparation was stored at -80°C until needed.
Electrophoretic mobility shift assays (EMSA). Cy5-labeled promoter probes containing the UAS motif (TGTAAG-N6-CTTACA) in rsmY and rsmZ were synthesized commercially (Shanghai DNA Bioscience Co. Ltd). EMSA was performed using established protocols68. Briefly, the purified His-GacA was mixed with Cy5-labeled rsmY and rsmZ promoter fragments, respectively, then loaded on a 4.5% nondenaturing polyacrylamide gel for electrophoresis. The Cy5 fluorophore was detected using an Amersham Typhoon RGB Biomolecular Imager (Cytiva, Sweden). EMSA experiment was repeated three times with similar results.
MIC and EC 50 determination. The MICs of pseudoiodinine and PCA for Xanthomonas spp. were determined by serial dilution. NA was prepared containing pseudoiodinine at 0–64 µg/mL or PCA at 0-256 µg/mL. Two microliters of bacterial suspension (OD600 = 1.0) were diluted threefold and spotted to NA plates. After a 48-h incubation at 28°C, MICs were defined as the lowest concentration at which no growth was visible.
EC50 values of pseudoiodinine and PCA were determined according to growth inhibition. Briefly, 10 µL of bacterial suspension was added to 5 mL NB containing diluted concentrations of pseudoiodinine and PCA. OD600 values of the tested suspensions were measured when the control suspensions increased to OD600 = 1.0. The log of percentage inhibition based on OD600 values were regressed on the log of compound concentrations, and EC50 values were calculated. This experiment was performed three or more times with similar results.
The EC50 of M. oryzae R01-1 was determined in vitro by transferring plugs (0.5 cm2 diameter) of mycelium from an actively growing fungal colony to a series of OAM plates containing pseudoiodinine at 10, 15, 20, 25, 30, 35, 40, 45 and 50 µM. Fungal colony diameters were measured after a five-day incubation at 25oC in darkness, and inhibition was calculated as percent of the control growth. EC50 values were calculated based on linear regression of colony diameter on log-transformed pseudoiodinine concentrations. Experiments were conducted three times with similar results.
Statistical analyses and reproducibility. The statistical significance was calculated using the least significant difference (LSD) test method and one-way ANOVA analysis through SPSS version 22.0 programme. A value of P < 0.05 was regarded as statistically significant. All values are presented as mean ± s.d. The statistics data were performed using GraphPad Prism version 8.00. Lesions areas were calculated from infected leaves using Adobe Photoshop CS5. Experiments were repeated at least three times to confirm reproducibility.