Pleiotropic Effects of c-di-GMP Content in Pseudomonas syringae

The present work comprehensively analyzed the transcriptome and phenotypes that were regulated by c-di-GMP in P. syringae. Given that the majority of diguanylate cyclases and phosphodiesterases have not been characterized in P. syringae, this work provided a very useful database for the future study on regulatory mechanism (especially its relationship with T3SS) of c-di-GMP in P. syringae. In particular, we identified three promoters that were sensitive to elevated c-di-GMP levels and inserted them into luciferase-based reporters that effectively respond to intracellular levels of c-di-GMP in P. syringae, which could be used as an economic and efficient way to measure relative c-di-GMP levels in vivo in the future.

In Pseudomonas syringae pv. tomato DC3000, BifA and Chp8 have been identified as a PDE and a DGC, respectively (9,10). The overexpression of BifA reduces the c-di-GMP level in vivo and elevates the virulence of pathogens in the Pseudomonas genus, whereas a ΔbifA mutant strain shows an elevated c-di-GMP level and reduced necrosis and chlorotic symptoms during infections (10). Although BifA and Chp8 perform opposite functions, both affect virulence-related phenotypes such as motility and biofilm formation (9,10). A high intracellular c-di-GMP level drastically inhibits the biosynthesis of flagellin and bacterial motility but enhances the formation of biofilm. In contrast, a low c-di-GMP level strengthens bacterial motility and attenuates biofilm production in many bacteria (2,5,(11)(12)(13)(14). In P. aeruginosa, c-di-GMP regulates bacterial motility by controlling the expression of FliA, the key regulator of flagellar synthesis (15).
During infection in a host, c-di-GMP regulates bacterial resistance against oxidative stress, which is crucial for pathogens to survive the host immune response (25). For example, in Salmonella enterica, a low level of c-di-GMP decreases resistance against hydrogen peroxide (H 2 O 2 ) (13). In P. aeruginosa, an increased c-di-GMP level confers greater resistance to H 2 O 2 (25). Scavenger superoxide dismutases are expressed by pathogens to resist reactive oxygen species (ROS), which damage bacterial DNA, RNA, proteins, and the stability of cell membranes (26,27). For instance, the superoxide dismutases SodA and SodB convert superoxide O 2Ϫ into H 2 O 2 , which is decomposed into O 2 and H 2 O with the help of KatG, KatE, or the alkyl hydroperoxide reductase AhpC (28)(29)(30).
YedQ and YhjH, the DGC and PDE from E. coli, have been used to tune the c-di-GMP levels in Burkholderia cenocepacia (31), Comamonas testosteroni (32), Pseudomonas aeruginosa (17), Pantoea alhagi (33), and Pseudomonas putida (34) and to investigate the effects of c-di-GMP in these bacteria. Therefore, to explore the functions of c-di-GMP in P. syringae, we used plasmid pBBR1MCS to overexpress yedQ and yhjH in P. syringae. We hypothesized that c-di-GMP has a divergent effect on virulence and stress responses in P. syringae.
In this study, RNA sequencing (RNA-seq) was performed to identify the c-di-GMPdependent regulon in P. syringae. Based on the RNA-seq results, luciferase-based reporters were constructed to efficiently measure the intracellular c-di-GMP level. Phenotypic assays were further used to demonstrate that c-di-GMP regulates many important biological pathways in P. syringae, such as regulation of the type III secretion system (T3SS), motility, biofilm production, siderophore production, and oxidative stress resistance.

Overexpression of YedQ increases the intracellular c-di-GMP level in P. syringae.
To modulate the intracellular c-di-GMP level in P. syringae, OX-yedQ and OX-yhjH strains were generated via the overexpression of YedQ and YhjH, which function as c-di-GMP synthase and phosphodiesterase in P. syringae pv. syringae B728a, respectively. The c-di-GMP concentrations of wild-type, OX-yedQ, and OX-yhjH strains were quantified by liquid chromatography-mass spectrometry (LC-MS) at the stationarygrowth phase. As expected, the c-di-GMP concentration in the OX-yedQ strain (102.12 pmol/mg) was more than 20 times greater than that in the wild type (4.91 pmol/mg; P ϭ 0.00026) (Fig. 1), which demonstrates that the overexpression of the OX-yedQ strain leads to the elevated production of intracellular c-di-GMP in P. syringae. In contrast, the level of c-di-GMP in the OX-yhjH strain was below the detection limit of LC-MS ( Fig. 1), which suggests that overexpression of YhjH results in the degradation of c-di-GMP in P. syringae. Therefore, the heterogeneous expression of YedQ in P. syringae is effective in generating a high level of c-di-GMP.
c-di-GMP-dependent regulon in P. syringae. To reveal the c-di-GMP-dependent regulons in P. syringae, RNA-seq was performed to compare the transcriptomic profiles of the OX-yedQ and OX-yhjH strains; 818 differentially expressed genes (DEGs) were identified between these two strains ( Fig. 2A; see also Table S1 in the supplemental material). Of these DEGs, 352 were upregulated and 466 were downregulated in the OX-yedQ strain (P Ͻ 0.05; more than 2-fold enrichment; Fig. 2A). Enrichment analyses were performed on these DEGs via the Gene Ontology (GO) or the KEGG pathway. GO enrichment analysis highlighted the genes involved in flagellum-dependent cell motility, phosphorelay signal transduction, flagellum organization, siderophore transport, cell adhesion, and chemotaxis (Fig. 2B). Notably, the motility-related genes were significantly enriched (Fig. 2B). The genes associated with oxidoreductase activity and siderophore biosynthesis and transport were also enriched significantly (P Ͻ 0.05; Fig.  2B). According to KEGG analysis, genes associated with two-component systems were identified as the most enriched pathway (Fig. 2C). Flagellar assembly, fatty acid biosynthesis, and bacterial chemotaxis were also significantly enriched (Fig. 2C). Among the DEGs, Psyr_0610, Psyr_0685, and Psyr_5026 were selected for c-di-GMP-sensitive reporter construction. In sum, transcriptomic profiling analysis characterized the c-di-GMP-dependent regulon in P. syringae, thus indicating that c-di-GMP regulates multiple biological pathways, including those for flagellar motility, iron-binding siderophore, chemotaxis, and oxidoreductase activity. Furthermore, the RNA-seq assay was repeated by using the wild-type (WT) and OX-yedQ strains (Table S2). GO enrichment analysis highlighted the genes involved in flagellum-dependent cell motility, phosphorelay signal transduction, and bacterial chemotaxis (Table S3), which is consistent with the RNA-seq assay by using the OX-yedQ and OX-yhjH strains ( Fig. 2 and Table S1).
Luciferase-based reporters reflect the c-di-GMP level in vivo. Mass spectrometry is widely used for the accurate quantification of the intracellular c-di-GMP concentration (35,36). Here, we attempted to develop an economical and convenient method to measure the relative c-di-GMP concentrations in P. syringae using transcriptional fusion bioassays. We sought to make c-di-GMP biological reporters by fusing the promoters of c-di-GMP-induced genes to a promoterless luciferase gene. We first selected three top-induced candidates (Psyr_0610, Psyr_0685, and Psyr_5026) from our RNA-seq data sets (Fig. 3A). These three constructed plasmids were electrotransformed into the FIG 1 Quantification of high-c-di-GMP-content strain OX-yedQ, low-c-di-GMP-content strain OX-yhjH, and P. syringae pv. syringae B728a WT by mass spectroscopy analysis. nd, not determined. *, P Ͻ 0.05; ***, P Ͻ 0.001 by Student's unpaired two-tailed t test compared to P. syringae pv. syringae B728a WT with equal variance. The OX-yhjH strain was under the lowest detection value. The experiment was repeated with three independent bacterial cultures.

Effects of c-di-GMP in Pseudomonas syringae
Applied and Environmental Microbiology OX-yedQ and OX-yhjH strains to measure the luminescence value (in counts per second). All of the reporters (Psyr_0610, Psyr_0685, and Psyr_5026) showed significantly higher lux levels in the OX-yedQ strain than in the OX-yhjH strain (100-, 60-, and 20-fold, respectively; Fig. 3B), which indicates that these three promoters were very sensitive to the high levels of c-di-GMP in P. syringae pv. syringae B728a. In particular, the Psyr_0610 promoter was the most sensitive, with an induction of 100-fold, so it could be used as an efficient method to measure in vivo c-di-GMP levels (Fig. 3B). To examine whether their homogenous genes in P. syringae pv. tomato DC3000 were also sensitive to the c-di-GMP level, we constructed their counterparts in P. syringae pv. tomato DC3000 (PSPTO_0704, PSPTO_3756, and PSPTO_5471). The resulting plasmids were electrotransformed into the OX-yedQ and OX-yhjH strains and showed even better results than their counterparts in P. syringae pv. syringae B728a. These promoters were induced ϳ1,500-, 13-, and 500-fold, respectively, in the OX-yedQ strain (Fig. 3C). This luciferase-based assay also confirmed that overexpression of YedQ elevated the production of c-di-GMP in P. syringae. In sum, our newly constructed luciferase-based reporters can be used for sensitive monitoring of c-di-GMP levels in vivo.
Expression of T3SS is suppressed by c-di-GMP. Studies have shown that c-di-GMP can modulate virulence and the T3SS in some plant pathogens (4,8,36). For example, a higher level of c-di-GMP leads to repression of the T3SS in P. aeruginosa (36). To explore the effects of c-di-GMP in P. syringae T3SS and virulence, we identified 12 T3SS genes from our RNA-seq data and verified them by using reverse transcriptionquantitative PCR (RT-qPCR). As shown in Fig. 4, the expression of hrpR, hrpL, hrpA2, hrpB, hrpF, hrcC, hrcN, hrcR, avrB3, avrE1, and avrRPM1 was suppressed by 2-to 3-fold in the OX-yedQ strain in minimal medium compared with that of the OX-yhjH strain, thus suggesting an inhibitory effect of c-di-GMP on the T3SS, which is consistent with the result for P. aeruginosa (36). We also tested whether the inhibition was mediated by FIG 4 c-di-GMP negatively regulated T3SS in P. syringae. RT-qPCR of genes related to T3SS. The relative gene expression level in the OX-yhjH strain was set to 1, and the other values were adjusted accordingly. *, significantly different between OX-yedQ and OX-yhjH strains, P Ͻ 0.05; **, significantly different between OX-yedQ and OX-yhjH strains, P Ͻ 0.01; ***, significantly different between OX-yedQ and OX-yhjH strains, P Ͻ 0.001. All experiments were performed in duplicate. Error bars indicate standard deviations.

Effects of c-di-GMP in Pseudomonas syringae
Applied and Environmental Microbiology RhpR, a known T3SS repressor (37,38). However, the expression of rhpR showed no significant difference between the OX-yedQ and OX-yhjH strains ( Fig. 4), which suggests that c-di-GMP inhibits the T3SS via factors other than RhpRS. c-di-GMP negatively controls motility by regulating the expression of flhA, fliN, and fliE. In other Pseudomonas species, motility was inhibited by strains with higher intracellular c-di-GMP levels (5,12,14). Our RNA-seq results showed that seven known operons (flhAF, fliLMNPQR-flhB, fliEFJ, fliS-Psyr_3462, flgFGHIJKL, flgBCDE, and flgA) associated with flagellar synthesis were significantly downregulated in the OX-yedQ strain compared with the OX-yhjH strain (Table S1 and Fig. 5A). In addition, the transcription levels of a group of hypothetical genes, such as Psyr_3466 (encoding flagellin; 4-fold less) and Psyr_3460 (encoding flagellar sensor histidine kinase FleS), were also repressed (12-fold less) by c-di-GMP (Table S1). The phenotypic experiments showed that the swarming motility and swimming motility of the OX-yedQ strain were significantly compromised (3-fold less) compared with those of the OX-yhjH strain ( Fig. 5B and C). The elevated c-di-GMP level had significant inhibitory effects on the motility of P. syringae.
c-di-GMP is required for biofilm formation in P. syringae. In P. aeruginosa, the overexpressing YhjH strain showed less biofilm formation than did the wild type (20). In E. coli, c-di-GMP positively regulates genes associated with motile-sessile transition and biosynthesis and with the secretion of exopolysaccharides (EPS) in biofilms (3,4). Here, we speculated that c-di-GMP has similar functions in P. syringae. As expected, the expression of alg8 and alg44, which are involved in EPS biosynthesis, was increased by 3-to 4-fold in the OX-yedQ strain based on RT-qPCR (Fig. 6A). To further verify this regulation, the Congo red assay and quantitative detection of biofilms were performed in both strains. As shown in Fig. 6B, EPS production was greater in the OX-yedQ strain (more mucoid colonies) than in the OX-yhjH strain. Biofilm formation was also much higher in the OX-yedQ strain, as shown in Fig. 6C and D. These results indicate that c-di-GMP positively regulates biofilm formation in P. syringae. c-di-GMP positively regulates siderophore production. Of the DEGs in our RNA-seq data, genes catalogued as "iron ion binding" or as "siderophore transport" were significantly enriched in gene ontology analysis (Fig. 2B). The expression levels of pyoverdine transporter PvdE (32) and peptide synthase PvdP and PvsA (30,31) were upregulated 40-, 7-, and 20-fold, respectively (Fig. 7A). Given that pyoverdine was characterized by siderophore production, iron acquisition, virulence, and growth under iron-restricted conditions (24), we further tested whether a high c-di-GMP content could induce siderophore production in P. syringae. In a chrome azurol S (CAS)-based iron uptake assay, the OX-yedQ strain exhibited a larger orange halo (more than 2-fold) than the OX-yhjH strain (Fig. 7B), which is consistent with a previous study in P. syringae pv. phaseolicola 1448A (24). c-di-GMP positively regulates the biosynthesis, assembly, and transport of siderophores in P. syringae.
c-di-GMP positively regulates resistance to oxidative stress. Cytotoxic ROS, such as the superoxide radical O 2 -, H 2 O 2 , and hydroxyl radicals, can damage DNA, proteins, and lipids, resulting in a toxic effect on pathogenic bacteria (26). To protect bacteria from the ROS produced by plant cells during infection, it is important for pathogenic bacteria to inactivate ROS with their antioxidant enzymes, such as superoxide dismutase. Based on our RNA-seq data, genes that encode superoxide dismutase, such as sodA, were differentially expressed. Notably, the RNA-seq data showed that the expres- sion of sodA was induced about 40-fold in the OX-yedQ strain, while the expression of sodB and sodC was repressed (Fig. 8A). We hypothesized that c-di-GMP mediates the resistance to oxidative stress in P. syringae. To test the phenotypes of the elevated c-di-GMP level, H 2 O 2 tolerance tests were performed in liquid and solid KB media for the OX-yhjH and OX-yedQ strains. The OX-yhjH strain showed no growth in 6 mM H 2 O 2 , but the OX-yedQ strain grew well in 10 mM H 2 O 2 for 18 h (data not shown). The OX-yedQ strain grew better in 2 mM H 2 O 2 than the OX-yhjH strain (Fig. 8B). The result of the plate assay for H 2 O 2 resistance also demonstrated that the OX-yedQ strain showed stronger tolerance against H 2 O 2 than did the OX-yhjH strain (Fig. 8C). Taken together, the results suggest that c-di-GMP positively regulates resistance to oxidative stress by inducing the transcription of sodA.

DISCUSSION
The overexpression of E. coli YedQ and YhjH protein is widely used to study the function of c-di-GMP in many Pseudomonas species (34,39). The results of mass spectroscopy confirmed that the exogenous YedQ and YhjH enabled P. syringae to alter the c-di-GMP level (Fig. 9). The concentration of c-di-GMP was below the detection limit in the OX-yhjH strain. The three most sensitive c-di-GMP-responsive reporters (Psyr_0610-lux, Psyr_0685-lux, and Psyr_5026-lux) were identified and constructed and can be used to measure c-di-GMP levels in vivo. We were intrigued to find that the promoter of exogenous gene PSPTO_0704 of P. syringae pv. tomato DC3000 showed the highest induction level (1,000-fold) between the two strains. In P. aeruginosa, the c-di-GMPresponsive reporter brlR-lux was induced 100-fold by a high c-di-GMP level (40). The differences in the fold changes between PSPTO_0704-lux and brlR-lux may have been caused by the differences in the intracellular c-di-GMP levels in the two Pseudomonas strains.
Jenal et al. showed that c-di-GMP regulates bacterial lifestyles, including swimming, swarming, and biofilm formation, to alter bacterial virulence in P. aeruginosa (4). Chp8 elevates EPS production and negatively regulates swarming motility, whereas BifA negatively regulates swarming motility and positively regulates swimming motility in P. syringae pv. tomato DC3000 (9, 10). In P. aeruginosa PA14 and P. putida KT2442, BifA only affects swarming motility (43). In this study, an elevated level of c-di-GMP suppressed both swimming and swarming motility in P. syringae ( Fig. 5B and C). The differences between P. syringae pv. syringae B728a and other Pseudomonas strains may have resulted from other functions of Chp8 and BifA. Biofilm production was positively regulated by c-di-GMP in P. syringae, which is consistent with the results for other bacterial species (4,8). Furthermore, our study showed that c-di-GMP elevated the expression of lipopolysaccharide (LPS)-related genes. The expression of Psyr_0610 (encoding O-antigen ABC transporter and permease protein) and Psyr_0612 (encoding lipopolysaccharide biosynthesis protein) was 100-fold higher in the OX-yedQ strain than in the OX-yhjH strain, which suggests that c-di-GMP regulates biofilm formation by elevating the production of LPS in P. syringae.
Pyoverdine production and oxidative stress resistance are regulated by c-di-GMP in P. syringae (25,44). Pseudomonas species synthesize pyoverdine, the key virulence factor, to take up iron from the extracellular medium to rescue iron starvation (44). c-di-GMP positively controls the siderophore pyoverdine to acquire iron from ironreplete medium (16). In addition to pvdE, pvdP, and pvsA, five genes that encode the putative siderophore nonribosomal peptide synthase (Psyr_1956, Psyr_1957, Psyr_1958, Psyr_1959, and Psyr_1960) were significantly upregulated in strains that overexpressed c-di-GMP (Table S1), which suggests that c-di-GMP regulates the uptake of iron by inducing the synthesis, export, and assembly of pyoverdine in P. syringae. c-di-GMP regulates the oxidative stress resistance in many bacterial strains (13,28,45), but the mechanism remains largely unclear. These results indicate that SodA, but not SodB or SodC, plays a major role in the c-di-GMP-mediated response against oxidative stress.
Taken together, these results demonstrate that the pleiotropic molecule c-di-GMP globally regulates many important intracellular activities and behaviors in P. syringae (Fig. 9). In particular, c-di-GMP inhibits motility and T3SS and induces biofilm formation, pyoverdine production, and oxidative stress resistance in P. syringae. The newly constructed P. syringae-specific lux reporters provide an economical and effective method to detect c-di-GMP levels in vivo. We propose that tuning the c-di-GMP level offers a new strategy to protect plants from attacks by P. syringae.

MATERIALS AND METHODS
Strains, plasmids, primers, and growth conditions. The plasmids, bacterial strains, and primers used in this study are listed in Table 1. Unless otherwise indicated, Pseudomonas syringae pv. syringae B728a and its derivatives were grown in KB medium, consisting of 20 g proteose peptone, 1.5 g anhydrous K 2 HPO 4 , 15 ml glycerol, and 1.5 g MgSO 4 per liter, with or without agar at 28°C or shaking at 250 rpm. The concentration and categories of antibiotics were added as follows: for P. syringae pv. syringae B728a wild-type (WT) strain, 25 g/ml rifampin in KB agar medium; for the yedQ-overexpressing strain (OX-yedQ), 25 g/ml rifampin and an additional 60 g/ml gentamicin in KB agar medium but 30 g/ml gentamicin in KB liquid medium; for the yhjH-overexpressing strain (OX-yhjH) strain, 25 g/ml rifampin and an additional 60 g/ml tetracycline; and for the strains containing the pMS402 plasmid, 100 g/ml kanamycin was added, while in E. coli, 50 g/ml kanamycin was added. Experiments for OX-yedQ and OX-yhjH construction, RNA-seq, and liquid chromatography mass spectrometry (LC-MS) quantification of c-di-GMP were performed at Nanyang Technological University in Singapore, and other experiments were finished at the Department of Biomedical Sciences, City University of Hong Kong, Kowloon Tong, Hong Kong.

Construction of c-di-GMP reporters.
To report the c-di-GMP level in P. syringae sensitively, three c-di-GMP reporters of P. syringae pv. syringae B728a were constructed by inserting the promoters of Psyr_0610 (315 bp), Psyr_0685 (257 bp), Psyr_5026 (247 bp) to the promoterless plasmid pMS402 (46) ( Table 1). Furthermore, to examine whether their homogenous genes in P. syringae pv. tomato DC3000 are also sensitive to the c-di-GMP level in the P. syringae pv. syringae B728a strain, we constructed the corresponding homogenous promoters originating from P. syringae pv. tomato DC3000 ( Table 1). The constructed plasmids were electrotransformed into P. syringae pv. syringae B728a and its derivatives (the high-c-di-GMP-content OX-yedQ strain and low-c-di-GMP-level OX-yhjH strain) by using a Micro-Pulser (Bio-Rad) with 1.8 kV measurement during every electroporation. The OX-yedQ and OX-yhjH strains contain plasmids pBBR1MCS-5 and pBBR1MCS-3, respectively. Positive reporter strains were cultured at mid-log-growth phase (optical density at 600 nm [OD 600 ], 0.6). Luminescence values (in counts per second [cps]) of bacteria were recorded using a 96-well white opaque microplate in a BioTek microplate reader with fiber-optic type luminescence. The OD 600 of bacteria in each well was determined immediately using a 96-well transparent-bottom cell culture plate with the Fisher Scientific microplate reader.
RNA-seq analysis. To test the effect of high c-di-GMP levels in P. syringae pv. syringae B728a on the transcriptome, cells were recovered and cultured to early stationary phase (OD 600 , ϳ2, with no addition of antibiotics) in NYGB medium consisting of 8 g/liter nutrient broth, 10 g/liter glucose, and 5 g/liter yeast extract. Cells were mixed with bacterial RNAprotect reagent (Qiagen) to keep RNA intact. Total RNA was purified using the RNeasy minikit (Qiagen). The RNase-free DNase set (Qiagen) was applied to remove contaminant DNA through on-column DNase digestion. A second DNA removal step was applied using the Ambion Turbo DNA-free kit. rRNA was depleted using the Ribo-Zero rRNA removal kit (Illumina). The integrity and concentration of total RNA and DNA contamination were examined using the Agilent TapeStation system (Agilent Technologies) and Qubit 2.0 fluorometer (Invitrogen). Reverse transcription into cDNA was done using the NEBNext RNA first-and second-strand synthesis modules (NEB). Analysis of gene expression was carried out via Illumina RNA-seq. Libraries were produced using an Illumina TruSeq stranded mRNA sample prep kit. The libraries were sequenced using an Illumina HiSeq 2500 platform with a paired-end protocol and read lengths of 101 nucleotides (nt). The raw sequence data were streamlined using the Trim Galore! version 0.4.5 software (http://www.bioinformatics.babraham.ac .uk/projects/trim_galore/) to truncate or filter reads of low quality (parameters, -paired -q 20 -phred33 -illumina -length 36). High-quality reads were then aligned to the P. syringae pv. syringae B728a reference genome (GenBank accession no. NC_007005.1) and annotation file (ASM1224v1) using Tophat version 2.2.6 (47) with the parameter "-N 2 -g 1." Only the reads mapped once were considered. From the resulting alignments, SAMtools version 1.6 (48) was applied to sort the bam file. The differentially expressed genes were identified by performing Cuffdiff version 2.2.1 (49) with a P value smaller than 1eϪ5. Each sample in the RNA-seq assay was repeated three times. Reverse transcription-quantitative PCR. RT-qPCR was done using the OX-yedQ and OX-yhjH strains to validate the RNA-seq data. Briefly, 500 l fresh and overnight bacterial cells (OD 600 , 1) was harvested by centrifugation at 2,400 ϫ g for 5 min at 4°C. The extraction of total RNA followed the manufacturer's instructions for the total RNA purification kit (Sangon Biotech). cDNA was synthesized using the TIANScript reverse transcriptase (RT) kit (Tiangen). RT-qPCR used SuperReal PreMix Plus (Tiangen). The mass of cDNA is 70 ng per reaction, and 16S rRNA was selected as the internal reference. Key genes involved in T3SS and biofilm were analyzed. mRNA expression was evaluated for each sample using the cycle threshold (C T ) value. Relative gene expression was calculated as follows: ΔC T ϭ ΔC Ttarget Ϫ ΔC T16S rRNA . The fold change for the treatment was defined as the relative expression compared with the control group and was calculated by the 2 ϪΔΔCT method, where ΔΔC T ϭ C Tcontrol Ϫ ΔC T (50,51). The error bar was calculated using the standard deviation (SD) value. The experiment was repeated three times.
Quantification of c-di-GMP extracted from P. syringae pv. syringae B728a using liquid chromatography-mass spectroscopy. P. syringae pv. syringae B728a cells were recovered on a KB agar plate. Single colonies were picked and grew overnight in KB medium with respective antibiotics for strains carrying plasmid (30 g/ml tetracycline and 30 g/ml gentamicin) and no antibiotics for the wild type at 28°C and 200 rpm. Overnight cultures were diluted to an OD 600 of ϳ0.01 in KB medium and grew for 16 h to stationary phase (OD 600 ϳ3 for WT and 1.3 to ϳ1.6 for the two mutants). Biological triplicates were used for each strain. Five milliliters of cells was collected from each sample and washed with 1 mM ammonium acetate. c-di-GMP was extracted using lysis buffer consisting of ammonium acetatemethanol-water in a 40%/40%/20% ratio. Sonication using a probe sonicator was done using 40% amplitude for one-and-a-half minutes in a 10-s on/10-s off cycle to obtain better cell lysis. Cell debris was removed by centrifugation, and supernatants containing c-di-GMP were dried to remove lysis buffer using a Labconco SpeedVac concentrator. All processing steps were performed on ice or at 4°C. Dried samples were reconstituted in 200 l of LC-MS-grade water and injected into LC-MS directly with an injection volume of 5 l. A Waters BEH C 18 column (1.7 m, 2.1 by 50 mm) was used for highperformance liquid chromatography (HPLC). The gradient run was applied using water containing 10 mM ammonium formate and 0.1% formic acid as mobile phase A, and methanol containing 0.1% formic acid as mobile phase B. Mobile phase A had a gradient of 100% at 1 min, 80% at 3 min, 60% from 4 to 4.5 min, and 100% from 4.6 to 6 min, applied at a flow rate of 0.3 ml/min. A Xevo TQ-S (Waters) mass spectrometer was used to identify c-di-GMP with reference to c-di-GMP standard (Sigma) using electrospray ionization (ESI)-positive mode at 40-V collision energy. Three technical replicates were measured for each sample. c-di-GMP was detected at a retention time of 1.74 min with multiple reaction monitoring (MRM) transition to 152.0 and 135.0. Quantities of c-di-GMP were calculated and normalized against protein levels and illustrated in Fig. 1. The c-di-GMP levels in the OX-yhjH samples were undetectable.
Protein extraction was done for normalizing c-di-GMP concentrations among samples. One milliliter of bacterial culture was collected from each sample at the beginning of extraction from the same sample tubes. The cell pellet was collected and resuspended in 1 ml of 0.1 M NaOH. Cells were heated at 95°C for 10 min. The total protein content of each sample was measured using a Qubit protein assay kit (Invitrogen).
GO and KEGG pathway enrichment analysis. GO enrichment analysis and KEGG pathway enrichment analysis of DEGs were performed using DAVID 6.8 online analysis (https://david.ncifcrf.gov/ gene2gene.jsp). GO and KEGG terms with a P value of Ͻ0.05 were defined as a significantly enriched term.
Congo red assay and biofilm formation assay. A Congo red agar plate assay was carried out according to the previous study (21), with minor changes to compare the production of exopolysaccharide between the elevated-c-di-GMP-level OX-yedQ strain and the OX-yhjH strain. Germ-filtered Congo red dye (100 g/l) was added to KB medium with 1.5% agar. Five microliters of liquid overnight bacterial culture was spotted to the Congo red plate (with 25 g/ml rifampin) and cultured at 28°C. The colony staining was photographed after 3 days. The experiment was repeated with two independent bacterial cultures (3 plates for each strain).
A biofilm detection experiment was performed with minor modifications, as reported previously (52). In brief, overnight bacterial culture was transferred to a 10-ml borosilicate tube containing 2 ml KB medium (without antibiotics) at an OD 600 of 0.1 and cultured statically at 28°C for 3 to 4 days. Crystal violet (0.1%) was used to stain those biofilm cells that adhered to the tube tightly for 15 min, and other cells that bound to tube loosely were washed off with distilled deionized water (ddH 2 O). The tubes were washed gently three times with ddH 2 O, and the remaining crystal violet was fully dissolved in 250 l of 95% ethanol with constant shaking after photographs were taken. Two hundred microliters of this eluate was transferred to a clear and transparent 96-well plate to measure its optical density at 590 nm using a BioTek microplate reader. Three tubes were used for each strain, and the experiment was repeated with three independent bacterial cultures.
Measurement of H 2 O 2 resistance. For the growth assay, an overnight bacterial culture (OX-yedQ and OX-yhjH strains; OD 600 , 1.0) was diluted to OD 600 of ϳ0.01 with KB medium (without antibiotics) with or without 2 mM H 2 O 2 . Two hundred microliters of diluted culture medium was added to a sterile 96-well plate (Thermo Fisher Scientific) at 28°C. After 12 h, 24 h, 36 h, and 42 h, bacterial growth at 600 nm was recorded using a BioTek microplate reader. The experiment was repeated with two independent bacterial cultures, and 3 replicates were used for each strain. As described previously (53), in the hydrogen peroxide resistance colony assay, the bacterial culture was adjusted to an OD 600 of 1.0. KB agar plates were supplemented with 0.5 mM H 2 O 2 , and 10 l of serially diluted bacterial cultures was incubated on the plates at 28°C for 3 days. Three plates were used for two strains, and the experiment was repeated with three independent bacterial cultures.
CAS agar assay for iron uptake. A chrome azurol S (CAS) agar assay was carried out as previously described (24). To make a 100-ml CAS stock solution, 60.5 mg CAS powder (Sigma) was added to 50 ml of ddH 2 O, followed by 10 ml of 1 mM FeCl 3 . Then, 72.9 mg hexadecyltrimethylammonium bromide (HDTMA; Sigma) was fully dissolved in 40 ml ddH 2 O. At last, the entire HDTMA solution (40 ml) was slowly poured into 60 ml of CAS solution with constantly shaking to form a 100-ml CAS stock solution. A CAS agar plate was prepared by 9 parts freshly autoclaved 1.5% agar KB plate and 1 part CAS stock solution. After the agar solidified, two circular holes were dug into CAS agar by the round end of a 1-ml sterile pipette tip. About 2 l overnight bacterial culture (OX-yedQ and OX-yhjH strains; OD 600 , 1.0) of the experiment group and control group was added into one of two holes and cultivated at 28°C. The CAS plate (without antibiotics) was photographed after 3 days against a white background. Three CAS plates were used for two strains, and the experiment was repeated with three independent bacterial cultures.
Motility assay. The motility experiment was performed based on a previous study (52), with minor modifications. Swimming and swarming ability were tested on a KB plate with 0.3% and 0.4% agar (MP Biomedicals), respectively. Overnight KB cultures were inoculated and adjusted to the same bacterial density (OX-yedQ and OX-yhjH strains; OD 600 , 1.0). Two-microliter aliquots were dotted on swimming and swarming plates and cultured at 28°C. To quantify bacterial motility, the diameter of each colony was determined. Photographs were taken after 36 h for the swimming plate and 72 h for the swarming plate. Three motility plates were used for two strains, and the experiment was repeated with three independent bacterial cultures.
Statistical analysis. Data are presented as the mean Ϯ standard deviation. Data were tested for normality and analyzed using unpaired Student's t test. In Fig. 1 and 3 to 8, one asterisk (*) indicates a significant difference with a P value of Ͻ0.05. Two asterisks (**) indicate a significant difference with a P value of Ͻ0.01, and three asterisks (***) indicate a significant difference with a P value of Ͻ0.001. Each experiment was performed three times with similar results.
Data availability. The RNA-seq data sets have been deposited in the National Center for Biotechnology Information (NCBI) database with an accession number GSE120889.

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/AEM .00152-19.

ACKNOWLEDGMENTS
T.W., L.Y., and X.D. designed the study and wrote the paper. T.W. and Z.C. performed the experiments, analyzed the data, and generated the figures. L.Y. and C.H. edited the paper. W.Z. performed RNA-seq analysis. Y.X. helped construct the plasmids. Y.Z. helped measure lux luminescence. All authors reviewed the results and supported the final version of the manuscript.
This study was funded by the Health Medical Research Fund (grant 17160022 to X.D.) and the National Natural Science Foundation of China (grants 31670127 and 31870116 to X.D.).
We declare no conflicts of interest.