The bacterial strains and plasmids used in this study are listed in Supplementary Table S1. P. aeruginosa strain PAK was used as the parental strain. Escherichia coli strain DH5α was used as a routine cloning host for plasmid construction. P. aeruginosa and E. coli strains were grown in Lysogeny broth (LB) medium (10 g of tryptone, 5 g of yeast extract, and 5 g of NaCl per liter, pH 7.0) with shaking or on LB Agar (LB medium containing 1.5% [w/v] agar) at 37°C unless otherwise noted. Antibiotics were used at the following concentrations where appropriate: 100 μg/ml gentamicin, 150 μg/ml carbenicillin and 50 μg/ml tetracycline for P. aeruginosa and 100 μg/ml ampicillin, 50 μg/ml kanamycin, 10 μg/ml gentamicin and 10 μg/ml tetracycline for E. coli.

In-frame deletions of suhB, gacA and the double deletion rsmY rsmZ were previously constructed (Li et al., 2013), by allelic exchange employing the sucrose counter-selection system with the gene replacement vector pEX18Tc (Hoang et al., 1998). For engineering the gcbA (PA4843) deletion mutant, similar procedures were used except the different primers (Supplementary Table S2). Complementation and overexpression were accomplished by placing the respective genes under the control of the lac promoter in the pUCP20 vector. The plasmid pDN19lacZW carrying a promoter-less lacZ reporter gene was used to construct promoter-lacZ reporter fusions of the cdrA and other genes as previously described (Sun et al., 2014).

Swimming motility was assessed on tryptone plates (10 g of tryptone, 5 g of NaCl per liter containing 0.3% [w/v] agar, pH 7.0) as previously described (Rashid and Kornberg, 2000; Amiel et al., 2010). Briefly, plates were inoculated with bacteria from overnight LB cultures by puncturing inoculates halfway through the depth of the agar with a sterile toothpick and incubated at 37°C for 18 h. The diameter of the circular turbid zone as an indication of swimming motility was measured in millimeters. Plates were subsequently photographed with a digital camera. The experiments were performed three times with three replicates each time.

Swimming motility was also monitored in 3% (w/v) Ficoll (low-viscosity mimicking swimming conditions) as described elsewhere (Toutain et al., 2005). Bacterial strains expressing ZsGreen1 green fluorescent protein (GFP) were visualized via fluorescence microscope (Eclipse Ni-U, Nikon) equipped with a 100×/1.45 oil objective. The GFP-expressing bacteria were generated by transformation of the indicated strains with a multicopy plasmid (pUCP-zsGreen1, GFP⁺) that constitutively expresses ZsGreen1 under the control of a P_lac promoter (West et al., 1994). Approximately 100 cells were visualized for each strain.

Bacterial cultures were gently suspended in 2.5% (v/v) glutaraldehyde in 0.1 M PBS (pH 7.2) and the fixation process started at the same time. After 15 min, carbon-coated copper grids were dipped into a drop of the cell suspension for 30 s to allow bacterial adhesion. Excess liquid was eliminated with filter paper and grids were rinsed twice with distilled water and excess liquid was eliminated again. The grids were stained for 2 s with 2% (w/v) uranyl acetate, washed twice for 10 s in a drop of water and air-dried. The negatively stained cells were then visualized with a Hitachi H-7500 transmission electron microscope to confirm cell morphology and the presence of flagella.

Swim reversal rate measures the frequency at which a swimming cell changes its direction as previously described (Toutain et al., 2005; Caiazza et al., 2007; Petrova et al., 2014). Briefly, ZsGreen1-expressing P. aeruginosa strains were grown to exponential (OD₆₀₀ = 0.3) and stationary phase (OD₆₀₀ = 2.5), diluted 1:10–1:100 in 0.9% (w/v) NaCl containing 3% (w/v) Ficoll. Cells were observed via fluorescence on a Nikon Eclipse Ni-U microscopy with a 100×/1.45 oil objective, and real-time videos were captured using a Nikon DS-Fi1C camera, NIS-Elements F Ver4.00.00 and Camtasia Studio V7.5 software package. The videos were subsequently analyzed to monitor individual cells within the field of view for the number of changes in motility direction. Only cells remaining within the field of view for the duration of the video were considered. Approximately 50 cells were counted for each strain and reversal rates are reported as reversals per cell per minute.

Quantification of c-di-GMP was adapted from procedures previously described (Rybtke et al., 2012; Moscoso et al., 2014;Lo et al., 2016), with the following modifications. Briefly, a plasmid-based c-di-GMP reporter was created by transcriptionally fusing the c-di-GMP-responsive cdrA promoter to the lacZ gene encoding the enzyme β-galactosidase. The cdrA (PA4625) promoter from P. aeruginosa PAK was amplified using the primer pair cdrA-LacZ-F and cdrA-LacZ-R (Supplementary Table S2). The amplified promoter was inserted between the EcoRI and BamHI sites of the promoterless lacZ expression vector pDN19lacZW. The intracellular levels of c-di-GMP were gauged by measuring the β-galactosidase activity, and data are presented as relative c-di-GMP levels which were normalized to the levels of the reference strains.

For β-galactosidase measurements, overnight cultures were diluted 1:100 into LB and the cultivation was continued for additional 3.5 h (Chen et al., 2016). The β-galactosidase activity was determined by the Miller method (Sambrook et al., 1989). All assays were carried out in triplicates and each assay was performed at least twice.

Biofilm assay was performed as described (O’Toole and Kolter, 1998) with minor modifications. Briefly, overnight cultures were inoculated at a final OD₆₀₀ of 0.0025 into LB medium in glass tubes or 24-wells polystyrene microtitre dishes and grown for 24 h. Initial attachment to a polystyrene surface was measured following 6 h of growth as previously described (Petrova et al., 2014). Then planktonic cells were removed and tubes were washed gently with water for three times. Biofilms were stained with 0.1% (w/v) crystal violet for 15 min and tubes were rinsed three times with distilled water. The tubes can be photographed when dry and for quantification of the biofilm crystal violet stain was solubilized in 95% (v/v) ethanol before measuring the absorbance at 600 nm.

For visualization of biofilm architecture, biofilms were grown on glass coupons in 24-well microtitre plates for 48 h under conditions described above. Then sample coupons were rinsed three times with PBS and stained with propidium iodide (60 μM) for 30 min at room temperature, protected from light. Excess stain was removed by rinsing the sample gently with sterilized water. Biofilm architecture was visualized via confocal laser scanning microscopy (CLSM) by using a Nikon A1 confocal microscope. The CLSM images acquisition and 3D reconstructions were done with the Nikon’s NIS-Elements AR software.

Overnight bacterial cultures were subcultured in LB medium and grown to an OD₆₀₀ of 1.0. Total RNA was extracted using the RNAprep pure Cell/Bacteria Kit according to the manufacturer’s protocol with on-column DNase I digestion (TIANGEN Biotech, Cat #DP430). RNA samples were further subjected to DNase I treatment and purified using the RNAclean Kit (TIANGEN Biotech, Cat #DP412). RNA concentration and purity were analyzed with an NanoDrop 2000 spectrophotometer (Thermo Scientific) and integrity was verified in denaturing agarose gels. cDNA was generated using the PrimeScript RT Reagent Kit (Perfect Real Time) (Takara, Cat #RR037A) according to the manufacturer’s instructions with random hexamers. Quantitative real-time reverse-transcription PCR (qRT-PCR) was performed using the SYBR^® Premix Ex Taq^TM (Tli RNaseH Plus) (Takara, Cat #RR420A) with the primers listed in Supplementary Table S2. The level of 30S rRNA rpsL was used as a control (Wu and Jin, 2005).

Statistical analysis was carried out using SPSS program and results were expressed as the mean values ± standard deviations. As indicated, one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test or Student’s t-test analysis was performed to assess statistical significance of the data.

It has previously been shown that SuhB is required for both swimming motility and biofilm formation (Li et al., 2013), suggesting this protein might be involved in coregulating these processes. We hypothesized that suhB may serve as a genetic link between these phenomena and attempted to determine the molecular basis of this link in the current studies. We firstly pursued the flagella-mediated swimming motility using tryptone soft agar motility assay. In agreement with previous results on 0.3% LB agar (Li et al., 2013), the suhB mutant is defective in swimming on the tryptone soft agar plates, and complementation of the mutant with a suhB-expressing plasmid (pUCP-suhB) restored normal swimming (Figures 1A,B).

To characterize the swimming motility defect in more detail, cells constitutively expressing GFP were generated and observed by light microscopy in media with viscosity mimicking swimming (3% Ficoll) conditions (Toutain et al., 2005). In general, the mono-flagellated P. aeruginosa swims in a straight forward-backup/reversal-straight forward mode (Pratt and Kolter, 1998; Ping et al., 2013). The results showed that during exponential phase (OD₆₀₀ = 0.3) about 35% of wild-type cells were motile and exhibited the run/back-up/run mode of swimming, whereas approximately 19% of the suhB mutant cells were also motile, but tended to swim straight and exhibited less frequent reorientations as compared with the wild-type. Interestingly, during stationary phase (OD₆₀₀ = 2.5) the percentage of motile cells in the suhB mutant was comparable to that of the wild-type strain (75% for wild-type strain, and 70% for suhB mutant). Transmission electron microscopy (TEM) analysis of cells from stationary phase showed that the suhB mutant possessed apparently normal flagellum and the cell morphology was comparable to the wild-type PAK (Figure 1C). However, when the flagellar reversal rates (the frequency at which a motile cell changes its movement direction) was tested at stationary phase, inactivation of suhB resulted in a significantly reduced frequency of changes in cellular movement direction compared to those of the wild-type strain (Figure 1D), suggesting an abnormal function of flagella in the suhB mutant.

We have previously shown that suhB mutation leads to the upregulation of GacA and its downstream small RNAs, RsmY, and RsmZ (Li et al., 2013). Furthermore, we have shown that SuhB regulates the expression of RsmY/RsmZ through GacA (Li et al., 2013). Since the Gac system negatively regulates motility in pseudomonads (Heurlier et al., 2004; Navazo et al., 2009), we hypothesized that the Gac system could be implicated in the repression of swimming motility in the suhB mutant. Consequently, the effect of gacA mutation on swimming motility was tested. As shown in Figures 2A,B, inactivation of gacA in wild-type strain PAK resulted in small, but significant increases in swimming motility, while mutation of gacA in the suhB mutant restored swimming motility to nearly wild-levels. Similar results were obtained when both rsmY and rsmZ were simultaneous deleted in the suhB mutant (Supplementary Figures S1A,B). These results suggest that GacA and RsmY/Z are involved in the SuhB-mediated regulation of swimming motility.

Since the major role of RsmY and RsmZ is to antagonize the RNA binding regulator RsmA, we therefore asked whether the observed motility defect of suhB mutant is linked to reduced RsmA levels and can be rescued by elevating levels of this protein. As expected, while plasmid overexpression of RsmA under the control of the lac promoter (pUCP-rsmA) in wild-type PAK resulted in enhanced motility, multicopy expression of rsmA partly rescued the swimming motility of the ΔsuhB mutant (Figures 2C,D), indicating that RsmA is responsible for the suhB mutant phenotype. Together, these findings indicated that the regulation of motility by SuhB acted, at least partially, through the GacA-RsmY/Z-RsmA regulatory pathway.

In P. aeruginosa, mutation in either suhB or retS results in similar phenotypes from motility to biofilm formation, and SuhB-mediated regulatory pathways have been demonstrated to partially overlap those under the control of RetS (Goodman et al., 2004; Coggan and Wolfgang, 2012; Li et al., 2013). The retS mutant displays elevated levels of c-di-GMP, and RetS has been demonstrated to switch the T3SS/T6SS via c-di-GMP signaling (Moscoso et al., 2011). Based on these studies, we hypothesized that loss of suhB might also result in an increase in intracellular c-di-GMP levels. To address this, we compared the levels of c-di-GMP between wild-type PAK and the suhB mutant by using a transcriptional PcdrA-lacZ reporter system that responds to c-di-GMP (Rybtke et al., 2012). As shown in Figure 3, the suhB mutant displays a ∼3-fold increase in c-di-GMP levels as compared with the wild-type strain. When the suhB mutant was complemented with the suhB gene, the intracellular c-di-GMP was restored to wild-type levels, indicating that SuhB modulates intracellular c-di-GMP concentrations.

Since c-di-GMP signaling has been implicated in regulating the transition between motile and sessile lifestyles, and mutation of suhB results in increased levels of c-di-GMP, we questioned whether the phenotypes of a suhB mutant were correlated with a variation in intracellular c-di-GMP levels. If the elevated level of c-di-GMP was responsible for the suhB mutant phenotypes, then reducing intracellular c-di-GMP should reverse these phenotypes. To evaluate this possibility, a plasmid expressing a phosphodiesterase (PDE) PA2133 (pUCP-2133) or the empty vector pUCP20 was introduced into the suhB mutant. As shown in Figures 2A,B, overexpression of PA2133 in the suhB mutant partially restored swimming motility to wild-type levels. Moreover, we found that both PAK and ΔsuhB carrying the PA2133 expression vector exhibited very low to undetectable levels of c-di-GMP (Figure 3), suggesting that the motility defect of the ΔsuhB mutant can be partly attributed to the increases in intracellular c-di-GMP concentrations.

Given that both Gac-Rsm pathway and c-di-GMP were responsible for the motility inhibition in suhB mutant, we asked whether the two signaling networks act in parallel or there was a link between them. Whereas deletion of gacA or overexpressing PA2133 in ΔsuhB resulted in significant increases in swimming motility, the effects were not additive when PA2133 was overexpressed in the suhB gacA double mutant cells (Figures 2A,B). ΔsuhBΔgacA mutant overexpressing PA2133 had swimming phenotypes that were comparable to those carrying the vector control (Figures 2A,B). The same phenomenon was observed in the ΔsuhBΔrsmYΔrsmZ mutant strains (Supplementary Figures S1A,B). These findings indicated that the presence of GacA and sRNAs might be required for the c-di-GMP-mediated regulation of swimming motility in the suhB mutant.

Further, we extended our observation to biofilm formation. The biofilm phenotype was tested in glass tubes and quantified using the crystal violet assay as previously described (O’Toole and Kolter, 1998). Consistent with elevated c-di-GMP levels, the ΔsuhB mutant showed a hyperbiofilm phenotype compared with the wild-type (Figure 4A). This enhanced biofilm formation was dramatically abolished upon overexpression of PA2133 (Figure 4A), indicating the hyperbiofilm phenotype of the suhB mutant is dependent on elevated c-di-GMP. Additionally, CLSM analysis demonstrated that the suhB mutant formed thicker biofilms with more biomass than the wild-type strain PAK, while overexpressing PA2133 significantly reduced biofilm biomass and thickness in the suhB mutant (Figure 4B). Taken together, these results demonstrate that c-di-GMP dysregulation plays an essential role in the impaired swimming motility and enhanced biofilm formation in suhB mutants.

Cyclic di-GMP is produced by diguanylate cyclases (DGC) and degraded by phosphodiesterases (PDE). We hypothesized that the increased levels of c-di-GMP in suhB mutant could be a consequence of either activation of gene(s) encoding a DGC or repression of gene(s) encoding a PDE. To define the mechanistic basis for c-di-GMP accumulation and to identify potential target c-di-GMP-metabolizing enzymes in the suhB mutant, we compared their transcriptional profiles between wild-type PAK and the ΔsuhB mutant by real-time PCR. Of the 41 genes encoding GGDEF/EAL/HD-GYP domain proteins, we identified 12 genes exhibiting moderate differences in expression levels (Table 1). The DGC genes PA4843 (gcbA), PA1107 (roeA) showed higher levels, and the PDE genes PA5017 (dipA), PA4367 (bifA) showed lower levels of transcripts in the suhB mutant compared to wild-type PAK. It is worth noting that the transcript level of the PDE gene PA2133 was not changed in the ΔsuhB strain.

We next investigated the c-di-GMP-metabolizing genes responsible for the SuhB-mediated regulation of c-di-GMP concentration and hence the motility and biofilm phenotype. In light of the PA4843 (gcbA, also identified as adcA) upregulation, and the fact that GcbA was recently characterized to facilitate the transition from planktonic to surface-associated growth by acting as a functional DGC (Jones et al., 2014; Petrova et al., 2014), we engineered a suhB gcbA mutant and analyzed the level of c-di-GMP in this strain. As a control, a gcbA single mutant was also included. As shown in Figure 5A, while inactivation of gcbA resulted in a ∼2-fold decrease in c-di-GMP levels of the wild-type, deletion of gcbA in the suhB background significantly reduced the c-di-GMP concentration of the suhB mutant (Figure 5A), suggesting that the GcbA diguanylate cyclase is active in the suhB mutant and is responsible for the elevated levels of c-di-GMP. Moreover, the ΔsuhB ΔgcbA strain demonstrated significantly elevated swimming motility relative to the suhB mutant (Figures 5B,C), corresponding to its intermediate level of intracellular c-di-GMP (Figure 5A). Finally, inactivation of gcbA in ΔsuhB resulted in a significant reduction in attachment to surfaces relative to the parental suhB mutant strain (Figure 5D). Altogether, these data suggested GcbA is involved in the elevated levels of c-di-GMP in ΔsuhB that leads to the transition from planktonic to surface-associated growth.