|Home | About | Journals | Submit | Contact Us | Français|
Bacteria in biofilms often undergo active dispersal events and revert to a free-swimming, planktonic state to complete the biofilm life cycle. The signaling molecule nitric oxide (NO) was previously found to trigger biofilm dispersal in the opportunistic pathogen Pseudomonas aeruginosa at low, nontoxic concentrations (N. Barraud, D. J. Hassett, S. H. Hwang, S. A. Rice, S. Kjelleberg, and J. S. Webb, J. Bacteriol. 188:7344-7353, 2006). NO was further shown to increase cell motility and susceptibility to antimicrobials. Recently, numerous studies revealed that increased degradation of the secondary messenger cyclic di-GMP (c-di-GMP) by specific phosphodiesterases (PDEs) triggers a planktonic mode of growth in eubacteria. In this study, the potential link between NO and c-di-GMP signaling was investigated by performing (i) PDE inhibitor studies, (ii) enzymatic assays to measure PDE activity, and (iii) direct quantification of intracellular c-di-GMP levels. The results suggest a role for c-di-GMP signaling in triggering the biofilm dispersal event induced by NO, as dispersal requires PDE activity and addition of NO stimulates PDE and induces the concomitant decrease in intracellular c-di-GMP levels in P. aeruginosa. Furthermore, gene expression studies indicated global responses to low, nontoxic levels of NO in P. aeruginosa biofilms, including upregulation of genes involved in motility and energy metabolism and downregulation of adhesins and virulence factors. Finally, site-directed mutagenesis of candidate genes and physiological characterization of the corresponding mutant strains uncovered that the chemotaxis transducer BdlA is involved in the biofilm dispersal response induced by NO.
Bacterial biofilms are highly dynamic communities which display a range of differentiated phenotypes during the course of development. By exchange of cell-cell signals, subpopulations of cells can coordinate their activity and undertake particular metabolic tasks or defense strategies (56). At times, the bacterial community releases single cells that escape from the biofilm and revert to a free-swimming, planktonic mode of growth, leaving behind hollow voids in the biofilm architecture (5, 37, 57). This process, referred to as dispersal, completes the biofilm life cycle and is thought to be important for successful colonization of new surfaces. Although the mechanisms underlying these events remain to be fully elucidated, previous studies of various species, including the opportunistic pathogen Pseudomonas aeruginosa, have revealed that dispersal events correlate with the induction of a specific phenotype that involves cellular motility (37, 42).
In P. aeruginosa, biofilm dispersal can be triggered by environmental factors, including nutrient (42, 45) and iron (4, 36) availability, and has recently been linked to the intracellular second messenger cyclic di-GMP (c-di-GMP) (45, 47). Numerous studies revealed that decreased c-di-GMP levels are related to a motile mode of growth and to cell dispersal in eubacteria. In this second messenger system, diguanylate cyclases (DGCs) and specific phosphodiesterases (PDEs) are responsible for the biosynthesis and the degradation of c-di-GMP, respectively. DGCs and PDEs contribute to a genetic network that responds to a broad range of environmental cues and/or cell-cell signals and modulate intracellular levels of c-di-GMP, which has been shown to regulate various cellular functions, including biofilm formation, virulence, and dispersal, in many bacterial species (47, 51-53). Recently, we identified the gas nitric oxide (NO) as an important factor in the regulation of dispersal in P. aeruginosa biofilms (5). Exogenous addition of nontoxic concentrations of NO, typically in the low nanomolar range, was found to stimulate motility and biofilm dispersal in P. aeruginosa. A role for anaerobic metabolism and NO in biofilm dispersal and survival was further supported by other studies of P. aeruginosa (54, 61), Staphylococcus aureus (44), and various single and multispecies biofilms (6).
NO is a water-soluble, hydrophobic free radical that can freely diffuse in biological systems. At high concentrations (micromolar to millimolar range), NO and downstream reactive nitrogen species (e.g., peroxynitrite [ONOO−]) can have severe deleterious effects on cells. In recent years, it has become apparent that multiple regulatory systems in bacteria can mediate responses to NO exposure and activate detoxification mechanisms, thereby preventing damage from nitrosative stress (38, 49). Nevertheless, bacteria do not fully eliminate this radical, and lower concentrations of NO (nanomolar range) are commonly present in healthy bacteria (62, 63). For example, in denitrifying bacteria, NO is endogenously produced as an obligate intermediate of the anaerobic respiratory pathway, mainly through the activity of the nitrite reductase (Nir). To limit toxicity, its concentration is maintained at low levels within the bacterial cells by tight transcriptional control of nirS and the norCB genes, encoding the nitrite reductase precursor and two subunits of the nitric oxide reductase, respectively. In P. aeruginosa, NO itself may exert feedback control by inducing or repressing expression of the entire denitrification pathway, including nirS and norCB genes, via direct interactions with the regulators DNR (induction [2, 15]) and ANR (repression ). However, very little is known about the cellular and genetic responses to such low, physiological concentrations of NO. In biological systems, NO can react with high affinity to a broad variety of molecules, giving it a very short life span, in the order of seconds. Autooxidation with molecular oxygen (O2) is the most abundant reaction, the most rapid of which occurs with superoxide (O2−) (11). Molecular targets of NO encompass heme/nonheme iron cofactors, iron-sulfur clusters, and other redox metal sites, all forming metal-nitrosyl complexes as well as cysteine thiols (S nitrosylation), amines, and others (50). Reactions with iron moieties appear predominant, and interestingly, in several bacterial species, NO can derepress iron acquisition mechanisms through inactivation of the ferric uptake regulator (Fur) (12, 34). The amount of NO required for inactivation of Fur was found to be in the micromolar range (12, 34), and thus, it remains unclear whether endogenous NO could exert such regulatory effect.
Sensory receptors encoded by conserved PAS (Per-Arnt-Sim) or H-NOX (heme-nitric oxide/oxygen binding) domains that can potentially bind and respond to NO are commonly present in both eukaryotes and prokaryotes (9, 13, 20). Interestingly, genome sequence analyses revealed that P. aeruginosa contains multiple genes with such redox sensor domains associated with conserved GGDEF and EAL motifs with known or putative DGC and PDE activities, respectively (30). This strongly suggests a potential link between NO-sensing and modulation of c-di-GMP levels in P. aeruginosa.
The goal of this study was to test this hypothesis by performing PDE inhibitor studies, PDE enzyme activity measurements, and quantification of c-di-GMP levels in cells by liquid chromatography-tandem mass spectrometry (LC-MS-MS) analysis in the presence or absence of NO. Results from this study demonstrated that NO-mediated dispersal requires the chemotaxis transducer BdlA (biofilm dispersion locus) and involves increased PDE activity, resulting in an overall decrease in intracellular c-di-GMP levels in P. aeruginosa.
The P. aeruginosa strains used in this study are listed in Table Table1.1. Overnight cultures were routinely grown in Luria-Bertani (LB) medium with 10 g liter−1 NaCl with shaking at 37°C and supplemented with antibiotics where appropriate. Biofilms were grown in modified M9 minimal medium (57) containing 48 mM Na2HPO4, 22 mM KH2PO4, 9 mM NaCl, 19 mM NH4Cl, 2 mM MgSO4, 100 μM CaCl2 (pH 7.0), and glucose at 5 mM for continuous-culture-flow-reactor experiments and 20 mM for batch culture petri dish experiments. Fresh solutions of the NO donor sodium nitroprusside (SNP) (Sigma) or the NO scavenger 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (PTIO) (Sigma) were made daily in culture medium and protected from light. Although the exact amount and location of NO liberated in vivo within biofilms from SNP have not yet been elucidated, a 1,000-fold linear relationship between SNP concentrations and NO concentrations was measured in vitro (R2 ≥ 0.95) by using a NO analyzer (Apollo 4000 with ISO-NOP electrode; World Precision Instruments, Sarasota, FL). NO concentrations (100 nM to 1 μM) were determined for a range of SNP solutions (100 μM to 1 mM) in M9 medium. In these experiments, NO levels reached steady state after 10 to 15 min and were maintained over several hours. The sensitivity limit for the electrode system was approximately 20 nM NO. c-di-GMP (Biolog, San Diego, CA) and GTP (Sigma) stock solutions of 1 mM and 200 mM, respectively, were made in Milli-Q H2O and stored at −20°C. bis(p-nitrophenyl) phosphate (bis-pNPP) (Sigma) was directly dissolved in buffer prior to the enzymatic assay.
P. aeruginosa biofilms were cultivated on the interior surfaces of tubing of a once-through continuous-flow-reactor system at room temperature with a flow rate of 10 ml h−1. Earlier studies showed that P. aeruginosa biofilms grown on the inner walls of silicone tubing exhibited morphological similarities with biofilms grown on glass substrata in flow cells (17). The reactor was composed of Tygoprene XL-60 silicone tubing (Saint-Gobain Performance Plastics) (internal diameter, 3.2 mm; wall thickness, 1.6 mm; length, 200 mm; internal volume, 1.6 ml). Biofilms were cultivated for 5 days in M9 medium in triplicate. Then, biofilm dispersal was induced by adding SNP at a final concentration of 5 μM to the growth medium. This concentration of SNP was determined to generate steady-state NO concentrations of 5 nM in M9 medium in a static glass vial by using the Apollo 4000 NO analyzer (data not shown). Because the type of tubing had an impact on biofilm biomass, we selected Tygoprene tubing and 5 μM SNP for our experiments, as these conditions allowed sufficient biofilm biomass remaining after induction of dispersal to be harvested for subsequent extractions and analysis. The total number of viable bacteria in the effluent runoff of biofilms was measured by performing serial dilutions and CFU counts on LB agar at 37°C. Data are presented as changes in effluent runoff CFU numbers for SNP-treated biofilms in comparison to the level for untreated biofilms.
Biofilms of P. aeruginosa wild-type and mutant strains were grown in batch cultures in petri dishes (90 mm in diameter) containing glass microscope slides, as previously described (5). Biofilms were first allowed to develop for 24 h in M9 medium in the absence of NO. After 24 h, planktonic bacteria were removed and replaced with fresh M9 medium with or without the NO donor SNP at a final concentration of 500 nM to induce biofilm dispersal (5) and with or without 250 μM or 1 mM GTP. Inhibition of PDE activity upon exogenous addition of GTP was validated by using a PDE enzymatic assay as described below. The system was then incubated for an additional 24 h with constant incandescent lighting. Slides were rinsed gently in sterile phosphate-buffered saline (PBS). The biofilm on the slides was then stained with 250 μl of SYTO 9 (3 μl ml−1) (Molecular Probes, Eugene, OR) for 20 min in a humidified chamber. Fifteen selected fields of view per slide were imaged in the x-y plane by using an epifluorescence microscope (Leica model DMR) at regular intervals and across the entire slide. Image analysis (ImageJ software; NIH) was performed to determine the percentage of the glass surface covered with biofilm. Experiments were performed in triplicate, and a statistical comparison of the percentages of surface covered by biofilms was performed using one-way analysis of variance and Tukey's multiple comparison tests.
The effect of NO on PDE activity in P. aeruginosa was assessed in vitro by using the PDE-specific synthetic substrate bis-pNPP, as previously described (8, 29). Three milliliters of late-log-phase P. aeruginosa cells grown in LB medium and adjusted to an optical density at 600 nm (OD600) of 2.0 were harvested by centrifugation at 6,000 × g at 4°C and resuspended in buffer containing 50 mM NaCl, 50 mM Tris base (pH 8.1), 1 mM MnCl2, and bis-pNPP at a final concentration of 5 mM. Cells were either lysed with CelLytic Express (Sigma) or kept intact. Then, SNP was added to the mixtures in 50-ml Falcon tubes (BD), and the tubes were incubated vertically in the dark at 37°C with shaking at 60 rpm in triplicate. In these experiments, degradation of bis-pNPP and the release of p-nitrophenol was linear and quantified with a 200-μl aliquot by measuring OD405 with a spectrophotometer (Wallac Victor2; PerkinElmer, Waltham, MA) after 2 h of incubation, at a time point when the highest differences in p-nitrophenol production could be observed as identified in preliminary experiments (not shown). The controls consisted of medium without cells with or without SNP or CelLytic reagent in medium without cells.
For c-di-GMP extraction from planktonic bacteria, overnight cultures were diluted to an OD600 of 0.2 in fresh LB or M9 medium and grown with shaking at 37°C to an OD600 of 0.4. SNP was applied to 5 ml aliquots of the culture, and the solutions were incubated for 2 h. After treatment, heat inactivation of cells and ethanol extraction of c-di-GMP were performed as previously described (47, 51). Specifically, cell pellets were heated at 100°C for 3 min, cooled on ice, resuspended in 1 ml of ice-cold 65% ethanol, and homogenized by passing five times through a 20-gauge needle. Then, samples were incubated 20 min on ice, followed by centrifugation for 10 min at 4°C, and the supernatants retained. The extraction was repeated, and the supernatants were combined. The remaining cell debris was used for subsequent protein quantification by using a Lowry-based total protein assay (25). Nucleotide extracts were frozen at −80°C and lyophilized. Samples were then dissolved in 500 μl of Milli-Q H2O, and chloroform (1 ml) was added to the suspensions, vortexed, and separated by centrifugation for collection of 400 μl. For c-di-GMP extraction from biofilm bacteria, the biofilm on the inner walls of the tubing reactor were washed by a gentle bath of 3 ml ice-cold PBS and the tubing containing the biofilm was immersed in boiling water for 5 min. After cooling, biofilm cells were resuspended in 3 ml ice-cold 65% ethanol and harvested by compressing the exterior of the tubing by hand. Then, the extraction procedure was continued as described above. Nucleotides were extracted in at least three independent experiments.
For detection of c-di-GMP in cell extracts, reversed-phase high-pressure LC on a Nucleosil C18 100-5 column (3 by 125 mm; particle size, 5 μm) (Macherey-Nagel, Düren, Germany) coupled with the use of an electrospray ionization triple-quadrupole mass spectrometer in positive-ion mode was performed using the TSQ Quantum Access Max system (Thermo Scientific, Waltham, MA). The mobile phases were 50 mM ammonium acetate in water (pH 4.0) (phase A) and 100% acetonitrile (phase B). The gradient program was initiated with 100% phase A at a flow rate of 500 μl min−1, and after 0.5 min, the portion of phase A was decreased linearly to 10% in 4 min and then restored to 100% in 1 min; then, the portion of phase A was again decreased to 10% in 1 min and restored to 100% in 0.5 min to eliminate carryover between injections. The temperature was set at 30°C, and the cycle time was 10 min. Authentic c-di-GMP (Biolog) was used to identify c-di-GMP in cell extracts on the basis of (i) identical elutions (i.e., cochromatography) at 3.6 min, (ii) identical masses of the protonated molecular ion ([M+H]+) (m/z = 691), and (iii) identical MS-MS fragmentation patterns of isolated precursor ions for major components with m/z values of 152 and 540 (loss of guanine) (47).
For quantification of c-di-GMP in cell extracts, since no isotopically labeled c-di-GMP was available, a standard curve, whereby defined concentrations of authentic c-di-GMP were spiked to pooled cell extracts and analyzed against unspiked, pooled cell extracts, was established. The areas of the major fragments of c-di-GMP (m/z = 152 and 540) were used to estimate the amount of c-di-GMP in a sample, referred to in mg of cellular protein. One-way analysis of variance and Tukey's multiple comparison tests were used for statistical analysis of these experiments.
To determine gene expression patterns associated with NO-mediated biofilm dispersal in P. aeruginosa, the relative abundances of mRNA were compared for biofilm cells grown in continuous-flow reactors, with and without exposure to NO in two settings. First, P. aeruginosa biofilms were grown for 5 days in the absence of SNP, before a mature biofilm was exposed to 5 μM SNP to initiate dispersal and another biofilm was left untreated as a control (SNP versus untreated control experiment). After 1 h, RNA was extracted from both biofilm populations. Three independent replicate experiments were performed. Second, wild type and ΔnirS mutant (unable to produce endogenous NO) biofilms were grown for 5 days, and RNA was extracted from both biofilm populations (wild type versus ΔnirS mutant experiment). This experiment was performed in duplicate. For RNA extraction, the biofilm remaining on the walls of the tubing was washed by a gentle bath of 3 ml ice-cold PBS to remove planktonic bacteria. Biofilm cells were immediately resuspended in 3 ml ice-cold RNAlater RNA stabilization reagent (Qiagen, Hilden, Germany) and harvested from the interior surface of the tubing. Total RNA was extracted by using RNeasy RNA isolation kits (Qiagen) in accordance with the manufacturer's instructions and precipitated with 0.1 volume of 3 M sodium acetate, pH 5.2, and 3 volumes of ethanol. The RNA concentration and purity were determined by OD260 and OD280 measurements (OD260 = 1.0 = 40 μg RNA ml−1, and 2.0 < OD260/OD280 < 2.1, respectively). Biotinylated cDNA samples were synthesized and hybridized to PaeG1a oligonucleotide arrays by using Affymetrix GeneChip microarrays (Affymetrix, Santa Clara, CA) in accordance with the manufacturer's manual at the Ramaciotti Centre for Gene Function Analysis, The University of New South Wales.
Gene name, number, and annotation are from the Pseudomonas Genome Database (60). ArrayAssist version 5.0 (Stratagene, La Jolla, CA) was used for hybridization data processing, filtering, and statistical analysis. The probe array images were inspected for any image artifact. Polyadenosine RNA control intensities were used to monitor the labeling process. The overall fluorescence intensity of each microarray was normalized using the probe logarithmic intensity error estimate method to remove intensity-dependent effects of the arrays. Then, statistical analysis was done by unpaired t testing, and P values were obtained from calculated t statistics by using asymptotic analysis under the assumptions of normal distributions of expression values and equal variances. Genes were considered to be differentially expressed if the differences in change over three independent experiments (SNP versus untreated control) or two independent experiments (wild type versus ΔnirS mutant) had a P value of ≤0.05 and the absolute value of the average was equal to or greater than the standard deviation (≥1.3-fold) for the whole-microarray experiments (53).
To validate changes in gene expression observed in the microarray experiments, the mRNA levels of six selected genes and one housekeeping gene (proC) (43) from biofilm cells treated with or without 5 μM SNP by using the same experimental conditions as those for microarray analysis were examined by real-time quantitative PCR (qPCR) in triplicate. cDNA strands were synthesized by a SuperScript III reverse transcriptase reaction mixture (Invitrogen, Carlsbad, CA), and qPCR was carried out with a Rotor-Gene RG-3000A qPCR detection system (Corbett Research, New South Wales, Australia) by using Platinum SYBR green qPCR SuperMix-UDG (Invitrogen) with 100 ng of cDNA and a 500 nM concentration of each primer. Controls consisted of samples to which no cDNA template had been added and to which unmodified mRNA was added. Similar levels of mRNA for the housekeeping gene proC were found in treated and untreated biofilm cells (change, 1.0-fold). Primers for each selected gene were designed using PrimerSelect software (DNASTAR, Madison, WI) (see Table S1 in the supplemental material). Following PCR, relative expression ratios were calculated using 2ΔCT, where ΔCT represents the difference in cycling times for the two samples being compared. Unpaired t testing was used for statistical analysis.
To confirm upregulation of the nirS gene upon exposure to low doses of NO, a P. aeruginosa transcriptional fusion reporter strain, NSGFP, which expresses green fluorescent protein (GFP) when the anaerobic gene nirS is expressed (5), was used. Overnight cultures of NSGFP were diluted to an OD600 of 0.2 in fresh LB medium and grown with shaking at 37°C to an OD600 of 0.4. SNP treatments were added to 10-ml aliquots of the cultures in 50-ml Falcon tubes (BD) in triplicate, and the bacteria were incubated in the dark for a further 2 h. After exposure, cells were washed once in ice-cold PBS and resuspended in 1 ml ice-cold PBS. Two-hundred-microliter aliquots were transferred to a microtiter plate for fluorescence measurements (excitation, 485 nm; emission, 535 nm) (Wallac Victor2).
To assess the effect of low doses of NO on the production of the virulence factor pyoverdine in P. aeruginosa, overnight cultures of wild-type bacteria were washed once with PBS and diluted 50 times in 5 ml King's medium B (26) in 50 ml Falcon tubes (BD, Franklin Lakes, NJ) in the presence or absence of the NO donor SNP in triplicate. SNP was added to the tubes to give final concentrations of 500 nM or 5 μM or was not added, and the cells were incubated in the dark at 37°C at 100 rpm for 4 h to allow for sufficient release of pyoverdine in untreated control cultures. The relative concentrations of pyoverdine siderophore in the supernatants were quantified from 200-μl aliquots by using a microtiter plate fluorometer (excitation, 355 nm; emission, 460 nm) (Wallac Victor2).
To construct insertional mutants of the genes bdlA, fimX, morA, and rocS1, internal fragments of these genes were amplified by PCR, using P. aeruginosa PAO1 genomic DNA as a template, with appropriate primer pairs (see Table S1 in the supplemental material). The purified PCR products of bdlA, fimX, morA, and rocS1 were digested with SalI, SpeI-PstI, BamHI-HincII, and BamHI-HincII, respectively, and the correct-sized fragments were gel purified and cloned into the respective restriction sites of the suicide vector pKnockout-G (59) before transformation into Escherichia coli XL1-Blue (Stratagene). The resulting suicide vectors pKO[bdlA], pKO[fimX], pKO[morA], and pKO[rocS1] were transferred in strain PAO1 by biparental mating, with E. coli strain S17-1 as a donor, as described previously (27). Correct chromosomal insertion of the vectors was confirmed by PCR with appropriate primer pairs (see Table S1 in the supplemental material).
Previously, NO was found to induce a transition from the biofilm to the planktonic mode of growth and to increase swimming and swarming motilities in P. aeruginosa (5). Exposure to the NO donor SNP also induces dispersal events in biofilms cultivated under continuous-flow laboratory conditions. After 1 h of treatment with 5 μM SNP, biofilms exhibited a 2.4-fold increase in CFU released in their effluent runoff compared to untreated biofilms (Fig. (Fig.1A).1A). In this assay, SNP was added at 5 μM rather than 500 nM to compensate for the higher O2 levels in the biofilm medium from diffusion through the silicone tubing than in the batch culture biofilm assay, which has previously demonstrated the NO-induced dispersal event (5). Because the effects of NO on P. aeruginosa physiology closely resemble those regulated by the intracellular secondary messenger c-di-GMP (47), and because genes putatively involved in the turnover of c-di-GMP are often found to be associated with NO sensors (9, 20), the potential role of this signaling dinucleotide in the response to low doses of NO was investigated.
To assess the potential involvement of PDE activity during NO-mediated dispersal in P. aeruginosa biofilms, GTP, the precursor of c-di-GMP, which is known to inhibit some PDE activity and the degradation of c-di-GMP (40), was used. First, the inhibition of intracellular PDE activity by GTP in P. aeruginosa was confirmed when intact cells grown in suspension cultures were incubated with the PDE-specific substrate bis-pNPP, as previously described (8, 29). Addition of 250 μM and 1 mM GTP to intact cells of the P. aeruginosa wild type in buffer with bis-pNPP resulted in dose-dependent inhibition of PDE activity as determined by p-nitrophenol production (Fig. (Fig.1B).1B). Second, with the use of the previously described batch culture biofilm dispersal assay (5), which focuses on NO induction effects during the initiation of dispersal, preestablished P. aeruginosa biofilms were exposed to 500 nM SNP to trigger dispersal in the presence or absence of various concentrations of GTP. SNP treatment alone induced a 68% reduction in biofilm surface coverage compared to the level for untreated control biofilms, indicative of dispersal events (Fig. (Fig.1C).1C). NO-mediated dispersal was significantly abrogated in the presence of 250 μM and 1 mM GTP in a dose-dependent manner (Fig. (Fig.1C)1C) (P < 0.001). Treatment with 500 nM SNP in the presence of 250 μM GTP and 1 mM GTP resulted in only 43% and 10% reduction in biofilm surface coverage, respectively, compared to the level for untreated biofilms. The addition of GTP alone was not found to induce any increase in surface coverage when assessed for biofilms without NO addition (Fig. (Fig.1C).1C). Thus, the results revealed that NO induction of biofilm dispersal was significantly impaired in the presence of GTP and linked to altered PDE activity within the cells.
To determine whether NO stimulates the activity of PDEs present in P. aeruginosa, the overall PDE activity in cell extracts of P. aeruginosa grown in suspension was determined by using bis-pNPP as a substrate in the presence or absence of SNP. Addition of SNP at final concentrations of 500 nM, 5 μM, and 50 μM increased the overall PDE activities by 20%, 45%, and 31%, respectively, compared to the level for untreated cell extracts (Fig. (Fig.1D).1D). In contrast, higher concentrations of SNP, 5 mM and 50 mM (releasing 5 μM and 50 μM NO, respectively), caused inhibitions of PDE activity of 66% and 75%, respectively, in cell extracts. Moreover, addition of the NO scavenger PTIO (100 μM), which was previously found to be effective at preventing exogenous NO from inducing biofilm dispersal (5), decreased the apparent PDE stimulation in cell extracts after SNP addition. Interestingly, PTIO treatment also decreased the overall PDE activity in cell extracts not treated with SNP (Fig. (Fig.1D).1D). This observation suggests that low, basal levels of NO may have been present in the cell extracts and contributed to the basal PDE activity, even without exogenous addition of NO via SNP. Addition of PTIO was also tested with lysed cell solutions in the presence of p-nitrophenyl phosphate, a substrate specific for phosphomonoesterase. No significant difference from solutions without PTIO could be detected, suggesting that PTIO itself does not have any unspecific inhibitory effect on phosphomonoesterase activity (not shown). Finally, similar stimulatory effects on PDE activity by low doses of NO were also observed by using intact P. aeruginosa cells or cells from 5-day-old biofilms grown in continuous-flow reactors (data not shown).
The stimulatory effect of NO on PDE activity described above implied an increased degradation of c-di-GMP in P. aeruginosa during NO-mediated dispersal. In order to test this, a method for quantification of intracellular c-di-GMP by use of LC-MS-MS analysis was adapted (see Materials and Methods). With this method, the average c-di-GMP concentration was determined to be 101 pmol per mg cellular protein (pmol mg−1) in cells of P. aeruginosa taken from 5-day-old biofilms grown in continuous-flow cultures (Fig. (Fig.1E).1E). In contrast, in planktonically grown cells of P. aeruginosa, the concentration of c-di-GMP was 30 pmol mg−1, representing a 3.4-fold-decreased c-di-GMP content in planktonic cells compared to the level for biofilm-grown cells. We applied this method to quantify c-di-GMP in biofilm and planktonic bacteria that were exposed to low levels of the NO donor SNP. Mature biofilms of the parent strain and a ΔnirS mutant, which is unable to produce endogenous NO, grown for 5 days in the absence of SNP were exposed to 5 μM SNP to induce dispersal (Fig. (Fig.1A).1A). After 1 h of treatment, at a time point when dispersal had been initiated but sufficient amounts of biofilm could still be harvested, c-di-GMP was extracted from the biofilm cells. Treatment with SNP resulted in 47% (54 pmol mg−1) and 45% (63 pmol mg−1) (P < 0.01) decreases of intracellular c-di-GMP levels in cells of the parent strain and the ΔnirS mutant, respectively, compared to the level for untreated controls (Fig. (Fig.1E).1E). Further, cells from planktonic cultures of P. aeruginosa exposed to 5 μM and 50 μM SNP for 2 h showed 44% (17 pmol mg−1) and 40% (18 pmol mg−1) decreases in c-di-GMP levels, respectively, compared to the level for untreated planktonic bacteria (Fig. (Fig.1E)1E) (P < 0.01). Overall, these results strongly suggest that low levels of NO stimulate PDE activity in P. aeruginosa, leading to a decrease of intracellular c-di-GMP levels, which results in enhanced dispersal of biofilm bacteria.
To further investigate the regulatory mechanisms involved in NO-mediated dispersal in P. aeruginosa, we examined changes in gene expression in biofilm cells during induction of dispersal by NO. For each microarray analysis, we focused on genes with expression levels that were consistently affected across replicate experiments, with a P value of ≤0.05 (unpaired t test), and examined those that were the most up- or downregulated, with an absolute level of change higher than the standard deviation for the whole-microarray experiments. The observed changes in expression were supported by real-time qPCR analysis for six genes from 5-day-old biofilms after exposure to 5 μM SNP, which confirmed the transcriptomic pattern of the corresponding microarray analysis (Table (Table22).
The results indicated that the expression of genes involved in anaerobic metabolism, motility, attachment, cell envelope biogenesis, adaptation, protection virulence, and c-di-GMP signaling as well as bacteriophage genes were differentially expressed (Table (Table2;2; see also Table S2 in the supplemental material). The pilA gene, involved in twitching and swarming motilities (28), was found to be upregulated in the presence of NO, which correlates with increased motility behavior as previously observed for P. aeruginosa after NO exposure (5). In contrast, expression of adhesion genes, such as cupB and cupC, was reduced (Table (Table2;2; see also Table S2 in the supplemental material). The cupB and cupC genes, which are involved in interconnection of cells and biofilm formation, are known to be regulated by the Roc two-component system in which RocS1 is a membrane sensor with a PAS domain and one of the corresponding response regulators has an EAL domain (31). Further, genes involved in the production of virulence factors, such as pvd, essential for pyoverdine biosynthesis, and genes involved in defense mechanisms, such as fhp, implied in resistance to nitrosative stress (14), were found to be downregulated in biofilms exposed to 5 μM SNP versus untreated biofilms (Table (Table2).2). To confirm these NO-induced changes in gene expression, production of pyoverdine was directly assessed in the presence or absence of NO in P. aeruginosa grown in King's B medium (26). Addition of 500 nM and 5 μM SNP showed dose-dependent responses, resulting in 2.6- and 8.8-fold decreases in pyoverdine production, respectively, compared to the level for untreated cells (Fig. (Fig.2A2A).
Genes involved in energy metabolism, in particular the denitrification pathway, including nirS and norC, were found to be upregulated in biofilms treated with SNP (Table (Table2;2; see also Table S2 in the supplemental material). To further investigate this transcriptional activation, we used a P. aeruginosa reporter strain, NSGFP (5), which expresses GFP under the control of the nirS promoter, to monitor nirS expression levels in response to low doses of NO. SNP addition to NSGFP cells at final concentrations of 5 μM, 50 μM, and 250 μM induced 1.6-, 2.1-, and 3.1-fold increases in nirS expression, respectively, compared to the level for untreated cells (Fig. (Fig.2B),2B), verifying the microarray data described above.
DNA microarray results comparing the wild type versus ΔnirS also indicated that NO altered the expression of multiple genes suggested to be involved in the turnover or regulation of the secondary messenger c-di-GMP (see Table S2 in the supplemental material), including morA (downregulated) and bdlA (upregulated). MorA has been linked to the expression of adhesive surface structures essential for biofilm formation, thus suggesting that MorA may exhibit DGC activity (33), while BdlA is known to regulate c-di-GMP levels in P. aeruginosa (35).
The results presented thus far strongly suggest that low levels of NO trigger dispersal events in biofilm cells via c-di-GMP signaling by altering gene expression patterns that favor a planktonic lifestyle, e.g., increased motility and decreased surface attachment, and by changes in the expression of genes which are involved in c-di-GMP signaling. On the basis of these results, we selected genes which might be involved in the NO-induced dispersal by applying the following criteria: (i) the gene was found to be differentially expressed in the microarray data or is involved in the regulation of a gene which was identified in the microarray, (ii) the gene product is known or predicted to be involved in the turnover of c-di-GMP (EAL/GGDEF domains ) or known to influence intracellular c-di-GMP levels, (iii) the gene product is predicted to harbor a PAS domain for potential NO sensing (13, 16), and (iv) the gene product is known to regulate or alter the expression of surface structures involved in biofilm development or motility. Finally, genes which met at least three of these criteria were selected for further analysis using a site-directed mutagenesis approach with P. aeruginosa. This screening process identified seven genes, namely, bdlA (i to iii) (35), fimX (ii to iv) (24), morA (i to iv) (33), rocS1 (i to iv) (31), PA0575 (i to iii) (30), PA1181 (i to iv) (30), and PA2072 (i to iii) (30). Insertional knockout mutants of these genes were then tested for their ability to disperse from batch culture biofilms in response to low doses of NO. Preestablished biofilms grown on glass slides were exposed to 500 nM SNP to induce biofilm dispersal as previously reported (5). Addition of NO to all strains resulted in 40 to 66% reductions (statistically significant [P < 0.05]) in biofilm surface coverage in comparison to the level for the untreated control, except for the BdlA mutant strain KO[bdlA], which showed only a 23% reduction and was not significantly different from the untreated control (Fig. (Fig.33).
To confirm the involvement of BdlA in the NO-mediated dispersal, biofilms were allowed to grow under continuous-flow conditions for 5 days without SNP, and then 5 μM SNP was added to the growth medium. After 1 h of exposure and initiation of biofilm dispersal (Fig. (Fig.1A),1A), dispersing bacteria in the biofilm effluent runoff were enumerated by CFU counts on agar plates and compared to the corresponding untreated control. In the untreated controls, no significant difference was observed for the biofilm effluent runoff of strain KO[bdlA] in comparison to the P. aeruginosa wild-type strain (3.0 and 2.8 CFU cm−2 min−1, respectively). Exposure to 5 μM SNP increased the number of cells that detached from the wild-type biofilm by 2.7-fold (P < 0.05), whereas the number of cells detached from the KO[bdlA] strain biofilm remained unchanged (Fig. (Fig.4A4A).
To further investigate the role of BdlA in the response to NO, levels of c-di-GMP in planktonically grown cells of strain KO[bdlA] were determined and compared to the level for the parent strain in the presence or absence of SNP. When strain KO[bdlA] was cultured in the presence or absence of NO treatment, no significant difference in intracellular concentration of c-di-GMP (31, 27, and 26 pmol mg−1 with 0, 5, and 50 μM SNP, respectively) (Fig. (Fig.4B)4B) was observed in comparison to the level for the parent strain, which showed significant reduction of c-di-GMP levels upon exposure to NO (30, 17, and 18 pmol mg−1 with 0, 5, and 50 nm SNP, respectively) (Fig. (Fig.4B4B).
NO was previously found to induce a transition from a biofilm to a planktonic lifestyle in P. aeruginosa. In this study, we show that inhibition of PDE activity impaired NO-mediated biofilm dispersal and that low levels of NO stimulated PDE activity in P. aeruginosa and induced an overall decrease in intracellular c-di-GMP levels. In addition, microarray analysis revealed that several responses are affected upon addition of low, physiological doses of NO, mostly corresponding to c-di-GMP-regulated phenotypes and a transition from a biofilm to a free-living mode of growth. Hence, the data suggest that NO signaling is part of a global regulatory network that controls the switch between biofilm and planktonic phenotypes and involves the secondary messenger c-di-GMP. Further, analysis of several mutant strains showed that NO-mediated dispersal in P. aeruginosa requires the chemotaxis transducer BdlA.
Comparison of c-di-GMP levels in wild-type P. aeruginosa biofilm and planktonic cells demonstrated lower levels of c-di-GMP present in cell extracts of planktonic bacteria than for biofilm-grown cells. This is in agreement with previous reports, which demonstrated that reduced c-di-GMP is associated with dispersal or a planktonic phenotype in various bacterial species, including P. aeruginosa (30, 41, 47). Furthermore, our results show that exogenous NO addition reduced the overall intracellular c-di-GMP content in both biofilm and planktonic cells. A recent study demonstrated that cells of P. aeruginosa exist mostly in small aggregates during growth in suspension cultures and that c-di-GMP levels are significantly higher in these cell cultures than in those consisting of freely suspended cells (45). Therefore, the decrease in c-di-GMP levels in planktonically grown cells after exposure to SNP observed here is in agreement with the current hypothesis of c-di-GMP-regulated phenotypes, as it indicates a dispersal event from mostly aggregated cells toward freely suspended cells in the culture.
DGC and PDE enzymes that regulate the turnover of c-di-GMP signals are predicted to respond to environmental stimuli on the basis of their association with putative sensory receptors or of results from in vitro experiments. In this study, differences in intracellular levels of c-di-GMP in response to exogenous stimuli in wild-type bacteria were observed. Our data suggest that c-di-GMP signaling involves continuous changes in the intracellular content of this secondary messenger and does not imply an on/off mechanism. These observations correlate with the hypothesis that several, most likely localized, c-di-GMP pools are present in the cell and are representative of several, parallel c-di-GMP signal transduction cascades for various environmental signals (23, 58). Thus, changes of individual environmental signals might only induce alterations of discrete c-di-GMP pools, which can result in a partial decrease of the total cellular c-di-GMP levels. The changes in the expression levels of several genes encoding potential PDEs and/or DGCs observed in this study support the involvement of multiple pathways in the response to NO. All of those genes identified in this study encode proteins harboring both a GGDEF and an EAL domain, respectively, with the exception of the putative DGC gene PA0290 and the known PDE gene arr (18). However, the function of these composite proteins is not entirely understood, as some of the domains might have regulatory rather than catalytic properties (22).
Signal transduction mechanisms involving c-di-GMP were previously suggested to imply multilayer control at the transcriptional, translational, and posttranslational levels (22). For example, it was found that c-di-GMP binds to PelD to regulate production of the PEL polysaccharide (32). Moreover, the PDE stimulation by NO in cell extracts observed in this study suggests a posttranslational activation of PDE activity rather than a transcriptional control by NO. Such posttranscriptional regulatory effects could translate the relatively small changes in transcriptomic response into significant phenotypic effects, such as increased pyoverdine production, motility, and biofilm dispersal, which were observed here.
Finally, the chemotaxis transducer BdlA, which harbors two PAS domains and can indirectly modulate c-di-GMP levels (35), was found to be required for the dispersal response to low doses of NO in P. aeruginosa biofilms. Disruption of the corresponding single-gene operon encoding BdlA caused the loss of the dispersal phenotype in response to NO exposure in strain KO[bdlA]. We further demonstrated that BdlA is able to influence c-di-GMP levels, as cells of strain KO[bdlA] did not respond with decreased c-di-GMP concentrations upon exposure to NO in comparison to the level for the parent strain. These results strongly support our hypothesis that NO-induced dispersal is regulated via a c-di-GMP-dependent regulatory network. Since BdlA is not predicted to have any PDE activity itself, the effect of BdlA on c-di-GMP levels is most likely based on a regulatory mechanism, although the target is currently unknown. Further studies will be carried out to establish whether BdlA can directly bind NO signals at concentrations found to induce dispersal.
The nontoxic effect of low concentrations of NO in our experiments included the observed reduction in transcription of the flavohemoprotein-encoding gene, fhp, involved in aerobic NO scavenging (Table (Table2).2). In contrast, expression of this gene was previously reported by Firoved et al. (14) to be highly upregulated as a result of nitrosative stress, i.e., in the presence of damaging levels of NO generated from 5 mM S-nitroso-l-glutathione (GSNO). Here, we exposed bacteria to 5 μM SNP, which generates low, physiological concentrations of NO, kinetically lower on a molar basis by at least 1,000-fold. Moreover, NO-mediated dispersal was unchanged in a P. aeruginosa knockout mutant of fhpR (data not shown), encoding a NO-responsive regulator (FhpR, ortholog to NorR) that activates transcription of fhp under nitrosative stress (1). Thus, we suggest that FhpR is not involved in the dispersal response induced by NO in P. aeruginosa biofilms but is primarily responsible for inducing fhp in the presence of damaging concentrations of NO.
Many genes involved in the anaerobic respiratory pathway were positively affected in biofilms that were exposed to exogenous NO. The induction of nirS and norC transcription by NO was shown previously and found to be controlled by the NO-responsive activator DNR and the regulatory protein NirQ in P. aeruginosa (2). In Pseudomonas stutzeri, similar activations were observed in the presence of 5 to 50 nM NO but not in the presence of 500 nM NO under anoxic conditions (55). Bacteria in biofilms usually exhibit reduced metabolic activity, and by enhancing anaerobic metabolism, NO may favor a transition to a more active state allowing dispersal from the biofilm. Interestingly, recent studies have shown that the addition of nitrate (NO3−), nitrite (NO2−), or arginine to P. aeruginosa biofilms resulted in an increase in the sensitivity of the biofilm cells toward antibiotics (10) similar to that observed with NO (5). A NO3− sensor-response regulator system was also found to be involved in biofilm formation and motility in P. aeruginosa (54). Therefore, it is possible that NO may mediate these effects in P. aeruginosa biofilms.
In this study, we provide evidence for a direct link between NO and c-di-GMP in P. aeruginosa. This regulatory pathway may be conserved across microbial species. Indeed, NO, which is ubiquitous in nature, was found to induce dispersal in biofilms of several gram-positive and gram-negative species and Candida albicans as well as in mixed-species biofilms (6). Further, GGDEF and EAL domains are widely distributed among bacteria and are often associated with redox sensors capable of sensing NO (3, 39). These observations support the hypothesis that NO is an ancient and highly conserved regulator of dispersal and life histories (7).
The molecular machinery underlying NO-induced dispersal in bacterial biofilms shares striking similarities with the well-characterized NO/cGMP signaling pathway in eukaryotes, where NO sensing domains are commonly found associated with guanylate cyclases and/or PDEs (46). In animals, the NO/cGMP regulatory system is involved in sophisticated multicellular processes, including vasodilation and neurotransmission. Thus, multiple components of NO signaling mechanisms in biofilms appear to be represented in higher multicellular organisms, suggesting an intriguing relationship between biofilm dispersal signaling and the evolution of eukaryotic regulatory pathways.
We thank A. L. Vivant for help with the BdlA mutant biofilm experiments and our colleagues at the University of New South Wales (UNSW) for their support. The microarray experiments were performed at the Clive and Vera Ramaciotti Centre for Gene Function Analysis (UNSW). Mass spectrometric results were obtained at the Bioanalytical Mass Spectrometry Facilities within the Analytical Centre (UNSW). Subsidized access to this facility is gratefully acknowledged.
This work was partially supported by grants from the Australian Research Council, the Centre for Marine Bio-Innovation, and the Environmental Biotechnology Cooperative Research Centre. D.S. was supported by UNSW Vice-Chancellor's Postdoctoral Fellowship. D.J.H. was supported by the National Institutes of Health (GM-69845), the Cystic Fibrosis Foundation, and Cure Finders, Inc. (Sevierville, TN).
Published ahead of print on 2 October 2009.
†Supplemental material for this article may be found at http://jb.asm.org/.