The Nitrite Transporter Facilitates Biofilm Formation via Suppression of Nitrite Reductase and Is a New Antibiofilm Target in Pseudomonas aeruginosa

Bacterial biofilms play roles in infections and avoidance of host defense mechanisms of medically important pathogens and increase the antibiotic resistance of the bacteria. Nitric oxide (NO) is reported to be involved in both biofilm formation and dispersal, which are conflicting processes. The mechanism by which NO regulates biofilm dispersal is relatively understood, but there are no reports about how NO is involved in biofilm formation. Here, by investigating the mechanism by which complestatin inhibits biofilm formation, we describe a novel mechanism for governing biofilm formation in Escherichia coli and Pseudomonas aeruginosa. Nitrite transporter is required for biofilm formation via regulation of NO levels and subsequent c-di-GMP production. Additionally, the nitrite transporter contributes more to P. aeruginosa virulence than quorum sensing. Thus, this study identifies nitrite transporters as new antibiofilm targets for future practical and therapeutic agent development.

biofilms include, but are not restricted to, lung infections of patients with cystic fibrosis (CF), ear infections, burn wound infections, catheter infections, chronic wound infections, tooth decay, and bacterial endocarditis (7). Biofilm-mediated infections prolong hospital stays, increase fatality, and consequently place a substantial financial burden on health care systems worldwide (8,9). Thus, novel therapeutic strategies are needed to fight biofilm-mediated infections (10).
An example of a biofilm-forming multidrug-resistant bacterium is the opportunistic pathogen Pseudomonas aeruginosa, which was recently listed among the 12 antibioticresistant "priority pathogens" by the WHO (5,7,11). P. aeruginosa is fatal to CF patients, forming mucoid masses in lung tissue that lead to pneumonia; in addition, it causes severe infections in immunocompromised patients and is liable for the majority of nosocomial infections (12). Biofilm-grown P. aeruginosa persists despite frequent antibiotic treatment, causes reduced activation of complement, and shows less susceptibility to phagocytosis, indicating the role of biofilm formation on P. aeruginosa antibiotic resistance and virulence (4,5).
P. aeruginosa biofilm formation is regulated by an intercellular chemical communication system, quorum sensing (QS) (13,14). P. aeruginosa has three main QS systems: las, rhl, and pqs. Each QS system consists of autoinducer synthesis genes lasI, rhlI, and pqsABCD, as well as cognate regulatory genes lasR, rhlR, and pqsR (13). Bacteria within a biofilm display phenotypes distinct from those of planktonic cells, specifically those relating to growth and gene expression (4). QS systems are responsible for the expression of biofilm matrix genes and subsequent development of biofilm architecture (15). The intracellular second messenger cyclic di-GMP (c-di-GMP), synthesized by diguanylate cyclases (DGCs) and degraded by phosphodiesterases (PDEs) (16), can be regulated by QS systems because the activities of DGC-and c-di-GMP-specific PDE are reduced in the lasR and rhlR mutants, respectively (17). In addition, biofilm formation in P. aeruginosa is positively regulated by c-di-GMP (18,19). c-di-GMP induces the transition from planktonic to biofilm lifestyles by downregulating motility-associated genes and upregulating exopolysaccharide-and biofilm maturation-associated genes (20). Conversely, low c-di-GMP levels induce biofilm dispersal by activating expression of motility mechanisms such as flagella and pili (21).
Recently, nitric oxide (NO) has also been reported to regulate biofilm dynamics in a wide variety of bacteria, including P. aeruginosa (22)(23)(24). P. aeruginosa produces NO via the denitrification pathway that reduces nitrate (NO 3 Ϫ ) to dinitrogen (N 2 ) via nitrite, nitric oxide (NO), and nitrous oxide (N 2 O) (25). Each step of the pathway is catalyzed by individual enzymes: nitrate reductase (NAR), nitrite reductase (NIR), nitric oxide reductase (NOR), and nitrous oxide reductase (NOS), respectively. Under anaerobic conditions, such as the ones encountered in the CF airway mucus and inside biofilms, P. aeruginosa can obtain sufficient energy through a denitrification pathway using nitrate (NO 3 Ϫ ) and nitrite (NO 2 Ϫ ) as final electron acceptors (26,27). P. aeruginosa can obtain nitrate and nitrite from the host using nitrate and nitrite transporters (28,29). The denitrification pathway is also active under aerobic conditions (23). Exogenous addition of nontoxic concentrations of NO (approximately nanomolar to micromolar) stimulates biofilm dispersal or inhibits biofilm formation in P. aeruginosa (30). The exogenous NO stimulates motility and biofilm dispersal in P. aeruginosa by enhancing PDE activity and subsequently decreasing c-di-GMP levels (23,30,31). Accordingly, P. aeruginosa biofilm dispersal requires expression of the nitrite reductase (NIR), an NO-generating enzyme in bacteria (30,31). However, NIR and NO production are also reported to be important for biofilm formation (28,32), which seemingly conflicts with their role in biofilm dispersal (30,31). Although the regulation of biofilm dispersal by NIR-derived NO is relatively well understood (22,33), how NIR is involved in biofilm formation is not yet understood.
The discovery and development of agents that can restrict biofilm formation or even eradicate established biofilms by targeting the above-mentioned pathways are urgently needed. Such therapeutic entities would have profound effects on antibiotic resistance and pathogenesis management (4,29).
Here, while screening a microbial metabolite library for a P. aeruginosa biofilm formation inhibitor, we identified complestatin (Fig. 1A), a structural analog of vancomycin, derived from a fermentation culture of Streptomyces chartreusis AN1542 (34). We then investigated the in vitro and in vivo effects of complestatin on inhibition of P. aeruginosa biofilm formation and the molecular mechanisms underlying this inhibition. Using P. aeruginosa and Escherichia coli in vitro biofilm formation systems and an E. coli mutant library, we found that complestatin enhances endogenous NO production to high levels via overactivation of NIR expression. Specifically, we identified the NasA nitrite transporter in P. aeruginosa as the complestatin target. We then demonstrated that the nitrite transporter partially suppressed NIR expression to produce low levels of NO necessary for biofilm formation. This can explain how NIR is involved in biofilm formation. Our findings also demonstrate that nitrite transporters are a new antibiofilm target.

RESULTS
Complestatin inhibits P. aeruginosa biofilm formation but not planktonic growth. Screening of 5,839 microbial fermentation extracts for a P. aeruginosa PA14 biofilm formation inhibitor led to the selection of S. chartreusis AN1542. Bioactivityguided fractionation of the mycelium of this strain identified complestatin that was previously isolated by our group (Fig. 1A) (35). Biofilm formation was significantly reduced by treatment with 1, 3, and 10 M complestatin in a concentration-dependent manner in P. aeruginosa strains PAO1 and PA14 (Fig. 1B), whereas planktonic cell growth was not affected. The inhibitory effect on biofilm formation was not a result of a reduction in total cell numbers or cell viability as confirmed by optical density and viable cell growth assays (see Fig. S1A and B in the supplemental material). Conversely, vancomycin, a structural analog of complestatin, significantly induced biofilm formation at high concentrations (10 to 30 M), whereas it had no effect on biofilm formation at lower concentrations (Fig. S1C). In addition, vancomycin inhibited planktonic cell growth at high concentrations (10 to 30 M) (Fig. S1C).
The effect of complestatin treatment (10 M) on biofilm formation was further explored by confocal laser scanning microscopy (CLSM) using a BacLight Live/Dead Viability kit to stain the cells (Fig. S2A). The depth of complestatin-treated biofilms was approximately 40% less than that of untreated biofilms (Fig. S2B). The cell viable fluorescence intensity of complestatin treated P. aeruginosa PA14 biofilms was also reduced by 89.2% compared to that of untreated biofilms (Fig. S2C). Green fluorescence was not observed from dead bacteria.
To further confirm the biofilm inhibitory activity of complestatin, we quantified extracellular polymer substance (EPS) components in P. aeruginosa PA14 biofilms treated with or without complestatin. Complestatin (1 to 10 M) significantly reduced the amounts of extracellular polysaccharides, proteins, and eDNA in P. aeruginosa biofilms compared to those in an untreated control ( Fig. S2D to F). Overall, these results indicate that in contrast to vancomycin, complestatin inhibits P. aeruginosa biofilm formation without suppressing cell growth.
Complestatin treatment decreases c-di-GMP levels. To identify the mechanism by which complestatin inhibits biofilm formation, we measured the production of QS molecules [N-(3-oxododecanoyl)-L-homoserine lactone, N-butanoyl homoserine lactone, and 2-heptyl-3-hydroxy-4(1H) quinolone] and c-di-GMP in complestatintreated P. aeruginosa PA14 biofilm cells. Treatment with furanone C-30 (FC), a well-known QS inhibitor, resulted in reductions in the levels of QS molecules and pyocyanin, a QS-related virulence factor, whereas treatment with complestatin (10 M) had no effect ( Fig. S3A to D). In contrast, c-di-GMP levels were significantly lower in PAO1 and PA14 biofilm cells treated with 1 to 10 M complestatin than in untreated control cells (Fig. S3E). The effect was observed at lower complestatin concentrations in planktonic cells (0.1 and 0.3 M) than that in biofilm cells (Ն1 M) (Fig. 1C). These results suggested that complestatin inhibits biofilm formation in P. aeruginosa by interfering with c-di-GMP formation.
Next, to confirm whether complestatin exerted its antibiofilm activity via inhibition of c-di-GMP production, P. aeruginosa PA14 biofilm formation inhibition by complestatin treatment was assessed in a c-di-GMP-overproducing mutant strain that displays enhanced biofilm formation (ΔwspF) (36). In the wild-type strain P. aeruginosa PA14, 3 M complestatin inhibited biofilm formation by approximately 55.6%, and 10 M complestatin inhibited biofilm formation by approximately 61% compared to that of an untreated control (Fig. 1D). In contrast, in the wspF mutant, inhibition of biofilm formation was only observed at the highest complestatin concentration tested (10 M), and it was much lower than in the wild-type strain (9.3%) (Fig. 1D). In contrast, as expected, no differences in biofilm inhibition between the ΔwspF and wild-type strains were observed when using FC (data not shown) (17).
Overall, these results indicated that complestatin inhibits biofilm formation by decreasing cellular c-di-GMP levels.
Complestatin enhances PDE activity but not DGC activity. As DGCs and PDEs are responsible for c-di-GMP biosynthesis and degradation, respectively (16,37), we measured PDE and DGC activities in P. aeruginosa PA14 cells treated with and without complestatin to understand the mechanism by which complestatin decreased intracellular c-di-GMP levels. We observed significantly higher PDE activity in complestatintreated cells (3 to 10 M) than in untreated cells, whereas DGC activity was not changed with 0.1 to 10 M (Fig. 1E). These results suggest that complestatin decreases c-di-GMP levels by stimulating PDE activity and, hence, c-di-GMP degradation.
We found that C-PTIO completely reversed the inhibition of biofilm formation, the decrease in c-di-GMP levels, and the stimulation of PDE activity caused by complestatin in P. aeruginosa, whereas it had no effect on DGC activity ( Fig. 2A). In contrast, the NO scavenger did not reverse the effects of FC treatment on biofilm formation, c-di-GMP levels, or PDE activity ( Fig. 2A).
To further confirm the role of NO in complestatin inhibition of biofilm formation, we next measured NO levels in biofilm cells of P. aeruginosa after complestatin treatment using CLSM intrabiofilm and the fluorescent NO probe 4,5-diaminofluorescein diacetate (DAF-2DA). NO levels were significantly higher in complestatin-treated P. aeruginosa cells than in untreated cells ( Fig. 2B and C). Similarly, treatment with sodium nitroprusside (SNP; 5 M), an NO donor that served as a positive control, resulted in higher NO levels and biofilm formation inhibition, whereas treatment with FC did not affect NO production but inhibited biofilm formation as expected ( Fig. 2B and C).
Since high levels of exogenous NO (approximately nanomolar to micromolar) inhibits biofilm formation as well as disperses biofilms in P. aeruginosa (30), we tested the effect of complestatin on preestablished biofilms. Interestingly, complestatin did not affect dispersal and c-di-GMP levels of preestablished biofilms, whereas 5 M SNP as a positive control did, as reported (30) (Fig. 2D).
Overall, these results indicated that complestatin elevates NO production in P. aeruginosa, followed by stimulation of PDE activity, a decrease in c-di-GMP levels, and biofilm formation inhibition.
Complestatin inhibits biofilm formation by targeting the nitrite transporter NasA in P. aeruginosa. We then used the knowledge of the mechanism involved in complestatin-mediated biofilm formation inhibition to identify the specific target protein of this compound. To this end, we used the extensive Keio E. coli gene knockout library, which consists of in-frame, single-gene deletion mutants for all nonessential genes in E. coli BW25113 (38). A knockout library with the same characteristics is not available in P. aeruginosa, and since these two bacteria are closely related, we decided to perform the screening in E. coli. To validate the use of E. coli as a surrogate of P. aeruginosa, we first checked that the mechanism of biofilm formation inhibition of complestatin in E. coli was the same as in P. aeruginosa. Complestatin inhibited biofilm formation, decreased c-di-GMP levels, and stimulated PDE activity in E. coli BW25113, the E. coli Keio knockout parent strain, and these effects were abrogated by treatment with C-PTIO (see Fig. S4). To identify the target protein of complestatin, we screened the E. coli Keio collection library for E. coli mutants that exhibited the same phenotype as complestatin-treated cells. Among the 3,801 E. coli Keio mutants tested, 11 showed 30% less biofilm formation than the wild-type BW25113, without cell growth being affected (see Fig. S5A). In 6 of the 11 mutants, biofilm formation was restored to wild-type levels in the presence of the NO scavenger C-PTIO (Fig. S5A), and these mutants were further screened for reduced c-di-GMP levels, increased PDE activity, and unchanged DGC activity compared to those of the wild type, which led to the selection of four mutants. The genes mutated in these four mutants (ΔyhfS, ΔnirC, ΔybfA, and ΔprfC) were separately overexpressed in E. coli BW25113 using the plasmid pBAD with the arabinose-inducible promoter (see Table S1), and the effects of complestatin on biofilm formation and c-di-GMP levels were tested (Fig. S5B). In the presence of arabinose, complestatin failed to inhibit biofilm formation and c-di-GMP production and to enhance PDE activity, whereas it stimulated DGC activity only in a nitrite transporter (nirC)-overexpressing E. coli strain (Fig. S5B and Fig. 3A to D). In contrast, the QS inhibitor FC inhibited biofilm formation in the nirC-overexpressing E. coli strain regardless of the presence of arabinose, as expected (Fig. 3A).
To further confirm the role of the nitrite transporter in complestatin-mediated inhibition of biofilm formation, we tested nitrite transporter mutants in E. coli BW25113 and P. aeruginosa PA14 (ΔnirC and ΔnasA, respectively). As expected, these mutants presented reduced biofilm formation, decreased c-di-GMP levels, and enhanced PDE activity even in the presence of complestatin, and these effects were reversed by C-PTIO treatment, whereas DGC activity was not affected under any of the conditions tested ( Fig. 4A to D). Additionally, elevated NO production in the nirC mutant was confirmed by NO detection using CLSM (Fig. 4E). These results indicated that complestatin inhibits biofilm formation via NO production by targeting the nitrite transporters NirC and NasA in E. coli and P. aeruginosa, respectively.
Nitrite transporters repress nitrite reductase transcription. Next, to elucidate how complestatin ultimately inhibits biofilm formation, we investigated how the nitrite transporter inhibition by complestatin caused increased NO production. In bacteria, including E. coli and P. aeruginosa, NO is synthesized from nitrite by NIR. Thus, we hypothesized that the nitrite transporter represses transcription of the gene encoding NIR. To test this hypothesis, we analyzed the mRNA levels of the genes encoding NIRs in the E. coli nitrite transporter mutant ΔnirC, the ΔnirC mutant complemented with nirC, and a nirC-overexpressing strain by real-time quantitative PCR (RT-qPCR). In E. coli, there are two NIRs, a cytoplasmic NADH-dependent NIR (encoded by nirBD) and a periplasmic cytochrome c-dependent NIR (encoded by nrfABCDEFG), both of which can reduce nitrite to NO (39,40). We analyzed the mRNA levels of the first gene in each NIR operon. mRNA levels of the cytoplasmic NADH-dependent NIR (nirB) dramatically increased in the nirC mutant compared to those in the wild-type BW25113, whereas those of the periplasmic formate-dependent NIR (nrfA) were not changed (Fig. 5A). The increased nirB levels in the nirC mutant returned to normal in the nirC complemented strain (Fig. 5A). The decrease in nirB transcription in response to nirC overexpression confirmed that NirC suppressed nirB transcription and, importantly, also indicated that NirC partially suppressed nirB transcription in the wild-type E. coli (Fig. 5A), which is consistent with NO detection in wild-type E. coli BW25113 biofilm cells (Fig. 4E). Additionally, NIR suppression by the nitrite transporter was demonstrated in P. aeruginosa. P. aeruginosa has only one type of NIR (encoded by nirS), which is periplasmic. The mRNA levels of nirS were increased 1.5-fold in the nitrite transporter ΔnasA mutant (see Fig. S6A). These results indicated that the nitrite transporter partially suppressed the transcription of NIRs in both E. coli and P. aeruginosa.
Next, we tested whether complestatin induced the transcription of NIR via inhibition of a nitrite transporter. In E. coli, complestatin treatment resulted in a dose-dependent increase in nirB mRNA levels in the wild-type BW25113 but not in the ΔnirC mutant, whereas the effect of complestatin on nirB mRNA levels was restored in the ΔnirC strain complemented with nirC (Fig. 5B). Moreover, an increase in nirS transcription was observed in both PAO1 and PA14 treated with complestatin but not in the nitrite transporter mutant (ΔnasA) (Fig. S6B to D). These results indicated that complestatin restored the transcription of NIR via nitrite transporter inhibition in both E. coli and P. aeruginosa.
Nitrite reductase, partially suppressed by nitrite transporter, produces NO that activates DGC to produce c-di-GMP. Because the nitrite transporter partially sup- pressed the transcription of NIRs in the wild type and, as shown in Fig. 3A and B, nitrite transporter overexpression enhanced biofilm formation and c-di-GMP production, we hypothesized that NO, produced by the partially suppressed NIR, stimulates c-di-GMP production and biofilm formation. Indeed, in E. coli, nirC overexpression enhanced c-di-GMP production and biofilm formation, and these effects were revered by C-PTIO treatment (Fig. 6A and B). These results indicated that NIR, partially suppressed by NirC, produced endogenous NO that caused c-di-GMP production and subsequent biofilm formation.
Next, given that at low concentrations, endothelial NO synthase-derived NO binds to soluble guanylyl cyclase (sGC) to produce cyclic-GMP (c-GMP) in mammals (41), we hypothesized that a limited amount of NO produced by a partially suppressed NIR activates DGCs to produce c-di-GMP for biofilm formation. Indeed, nirC overexpression enhanced DGC activity but did not affect PDE activity, which was prevented by C-PTIO treatment ( Fig. 6C and D). This result indicated that NIR, partially suppressed by NirC, produced endogenous NO that in turn activated DGCs.
Thus, taken together, these results indicate that NO derived from nitrite transporterregulated NIR activated DGCs to produce c-di-GMP for biofilm formation.
Nitrite transporter mutant shows increased survival of Caenorhabditis elegans compared to that of wild-type PA14 and the QS mutants. Biofilm extracellular matrix plays a role in P. aeruginosa virulence by improving microbial attachment for invasion and avoiding phagocytosis (4,5). P. aeruginosa mediates pathogenesis via biofilm-mediated production of pyoverdine, a virulence factor, in a C. elegans animal model (42). Thus, we first checked if complestatin was able to attenuate P. aeruginosamediated pathogenesis in vivo using C. elegans, a well-established and practical model for studying P. aeruginosa virulence (43)(44)(45)(46). C. elegans rapidly died when fed P. aeruginosa PA14, as evidenced by only 20% survival 30 h postinfection (Fig. 7A). Treatment with complestatin (0.01 to 1 M) and P. aeruginosa at the same time resulted in a dose-dependent increase in nematode survival, up to 67% survival at the highest  (Fig. 7A). We then tested whether the protective effects of complestatin on killing of C. elegans by PA14 were also the result of inhibition of the nitrite transporter NasA by comparing the virulence of the nasA mutant with that of the wild-type PA14. Indeed, the nasA mutant showed increased survival of C. elegans compared to that of the wild-type PA14 (Fig. 7B). Complestatin treatment had no effect in C. elegans infected with the nasA mutant strain, as expected. In contrast, the QS inhibitor FC further reduced the virulence of the nasA mutant. These results confirmed that the nitrite transporter is the target of complestatin in C. elegans infection.
Next, we compared the virulence of the nitrite transporter mutant with that of the QS mutants. The nasA mutant exhibited increased survival of C. elegans compared to that with the lasR and rhlR mutants (Fig. 7C). The virulence-reducing property of the nasA mutant was reversed by C-PTIO treatment, whereas that of the QS mutants was not, suggesting that the nasA virulence in vivo was NO dependent similarly to that in vitro (Fig. 7D). These results indicated that NasA contributes more to P. aeruginosa virulence via NO production than QS.
Overall, these data suggest that NasA is a potential therapeutic target against P. aeruginosa infection in vivo.
Inhibition of biofilm formation by targeting the nitrite transport is enhanced in the presence of nitrate. Because complestatin produced NO via NIR activation of the denitrification pathway using nitrate/nitrite, complestatin is expected to increase NO production and consequently biofilm formation inhibition in the presence of nitrate. Indeed, complestatin increased NIR (nirS) induction and decreased biofilm in PA14 supplemented with 15 mM nitrate (see Fig. S7A). Complestatin inhibited biofilm formation three times more potently in nitrate-supplemented medium than in control medium, while the QS inhibitor FC did not display this discrepancy (Fig. S7B). Similarly, complestatin reduced cellular c-di-GMP levels with increased potency in the presence of nitrate (Fig. S7C). This finding was confirmed by comparing the nitrite transporter mutant to QS mutants. The increased NIR expression and reduced biofilm formation and c-di-GMP levels in the nitrite transporter mutant (ΔnasA) in P. aeruginosa were more severe in the presence of nitrate than in its absence, whereas they did not change in the QS mutants, i.e., lasR and rhlR mutants ( Fig. S6A and S8). These results indicated that biofilm formation inhibition by targeting the nitrite transporter was potentiated in the presence of nitrate.
Combined inhibition of two different mechanisms, nitrite transport and QS, increases biofilm formation prevention. Nitrite transporters enhanced biofilm formation via a completely different mechanism than the QS system in P. aeruginosa and E. coli. It was suggested that combined inhibition of the two mechanisms could prevent biofilm formation, because biofilms were not inhibited completely in either nitrite transporter or QS mutants. Indeed, FC, a QS inhibitor, inhibited the residual biofilm formation in the nasA mutant in a dose-dependent manner (49.5% more reduction by 100 M FC) (Fig. 8A). Conversely, complestatin effectively eradicated the residual biofilms of the lasR and rhlR mutants (36.5% and 43.4% more reduction by 10 M complestatin, respectively) (Fig. 8A). The profound combined effects were also verified in the C. elegans in vivo model (Fig. 6B and 8B). These results indicated that combined inhibition of nitrite transporter and QS more effectively inhibited biofilm formation and prevented P. aeruginosa virulence in vivo.

DISCUSSION
NIR, which produces NO, is reported to be involved in both biofilm formation and dispersal, conflicting processes, in P. aeruginosa. The mechanism by which NIR regulates biofilm dispersal is relatively understood, but there are no reports about how NIR is involved in biofilm formation (23). In this study, we found that complestatin inhibited c-di-GMP production and biofilm formation by targeting the nitrite transporters in E. coli and P. aeruginosa. We then demonstrated that the nitrite transporter partially suppressed NIR, which produced the limited amount of NO to activate DGCs, not PDEs. The activated DGCs produce c-di-GMP for biofilm formation. These results provided a novel mechanism for the NO requirement in biofilm formation (Fig. 9A). We have further shown that nitrite transporter contributes to P. aeruginosa virulence in vivo more than the QS receptors, LasR and RhlR, and combined inhibition of nitrite transporter and the QS targets enhances both antibiofilm activity and antivirulence effects.
In other reports, preestablished biofilms were dispersed by NO via enhancement of PDE activity and subsequent reduction in c-di-GMP levels (31), and biofilm formation was inhibited by terrein, a QS inhibitor, via decreases in DGC activity and the subsequent reduction in c-di-GMP levels (17). In this study, complestatin inhibited biofilm formation at 1 to 10 M without affecting cell viability by lowering cellular c-d-GMP levels in PAO1 and PA14 without affecting QS systems. Instead, complestatin enhanced c-di-GMP-specific PDE activity in PAO1 and PA14 without dispersing preestablished biofilms. To our knowledge, this is the first description of the inhibition of biofilm formation by a drug that enhances PDE activity and subsequently reduces c-di-GMP levels.
NO is an important factor regulating biofilm formation and mediating changes in the biofilm life cycle through c-di-GMP and/or QS in a wide variety of bacteria (23,24). NIR, which produces NO, is reported to be involved in both biofilm formation and dispersal, conflicting processes, in P. aeruginosa. The mechanism by which NIR regulates biofilm dispersal is relatively understood, but there are no reports about how NIR is involved in biofilm formation (23). A P. aeruginosa NIR (nirS) mutant produces poorly dispersing biofilms and partially regains dispersal ability upon exogenous NO addition, indicating that endogenous NO produced by NIR is essential for P. aeruginosa biofilm dispersal (30,31). On the other hand, a P. aeruginosa NIR (nirS) mutant failed to form biofilms under anaerobic conditions, indicating the requirement of NIR and NO production for biofilm formation (32). Additionally, biofilm formation and c-di-GMP levels were considerably reduced in a P. aeruginosa NIR (nirS) mutant under aerobic conditions (28). In this study, complestatin inhibited biofilm formation through the activation of NIR expression, and the nitrite transporter enhanced biofilm formation through the suppression of NIR expression in P. aeruginosa, which contrasts with the previous results (28,32) that showed that NIR is necessary for biofilm formation under both anaerobic and aerobic conditions. However, interestingly, this study also showed that nitrite transporter overexpression greatly enhanced biofilm formation through increased suppression of NIR expression. These results strongly suggested that the nitrite transporter partially suppressed NIR expression in the wild type so that NIR was expressed to specific levels to produce particular low-level NO during biofilm formation, which might be a requirement for biofilm formation.
In this study, the relatively low levels of NO produced when NIR was suppressed by nirC overexpression led to increased biofilm formation and c-di-GMP, whereas the high levels of NO produced when NIR was desuppressed in the nirC mutant led to a reduction in biofilm formation and c-di-GMP. The phenotype that biofilm formation is inhibited by NO has been reported with high levels of exogenous NO (approximately nanomolar to micromolar) in P. aeruginosa (30) (Fig. 2C), but stimulation of biofilm formation by relatively low levels of exogenous NO has not yet been reported. Notably, it is reported that the high levels of exogenous NO not only inhibit biofilm formation but also disperse biofilms in P. aeruginosa (30). Interestingly, complestatin inhibited biofilm formation and c-di-GMP production via NasA inhibition, NIR desuppression, and subsequent high NO production but had no effects on dispersal and c-di-GMP levels of This study suggested that NO-mediated biofilm formation occurred via the appropriate suppression of NIR by a nitrite transporter (NasA and NirC of P. aeruginosa and E. coli, respectively). The controlled NO production activated DGC but not PDE activity, which led to an increase in intracellular c-di-GMP levels and subsequent biofilm formation. (B) NOmediated biofilm dispersal, which is relatively well understood and reviewed (23,31). Complestatin inhibited biofilm formation by the following sequential processes: nitrite transporter inhibition, NIR desuppression, high NO production, PDE activity stimulation, and c-di-GMP level reduction.
preestablished biofilms (Fig. 2D). These results suggested that the NasA gene is active in planktonic cells or early-stage biofilms, not matured biofilms ( Fig. 9A and B).
Given that at low concentrations, endothelial NO synthase-derived NO binds to soluble guanylyl cyclase (sGC) to produce cyclic-GMP (c-GMP) in mammals (41), we hypothesized that a limited amount of NO produced by a partially suppressed NIR activates DGCs to produce c-di-GMP for biofilm formation (Fig. 9A), whereas a relatively large amount of NO produced under normal NIR activity conditions, which is somehow derepressed after biofilm maturation, activates PDEs to degrade c-di-GMP for biofilm dispersal or inhibition of biofilm formation (Fig. 9B). Here, we present evidence that the NO produced while NIR was suppressed by a nitrite transporter (NirC) activated DGCs in E. coli to produce c-di-GMP for biofilm formation. First, nirC overexpression enhanced DGC activity but did not affect PDE activity, which was prevented by C-PTIO treatment. This result indicated that NIR, partially suppressed by NirC, produced endogenous NO that in turn activated DGCs. Second, using a PA14 NIR mutant (ΔnirS), it was confirmed that NIR is necessary for biofilm formation in this system. Indeed, biofilm formation and c-di-GMP levels were dramatically reduced in the ΔnirS mutant (data not shown), indicating that NIR is essential for biofilm formation, which is consistent with the previous reports of Zhou et al. (28) and Yoon et al. (32). Thus, taken together, these results indicate that NO derived from nitrite transporter-regulated NIR activated DGCs to produce c-di-GMP for biofilm formation. Although the detailed mechanisms of NO-mediated DGC activation remain to be clarified, this result could explain the unanswered question of how NIR plays roles in both biofilm formation and dispersal (Fig. 9).
Nitrite is a central intermediate in the nitrogen metabolism of microorganisms and plants. The bacterial membrane nitrite transporter protein catalyzes nitrite uptake and export across the cytoplasmic membrane; this nitrite is subsequently reduced by cytoplasmic or periplasmic NIRs to produce NO or NH 4 ϩ for nitrogen metabolism and cytoplasmic detoxification (47,48). Here, we present evidence supporting a critical role for nitrite transporters in biofilm formation in E. coli and P. aeruginosa. First, complestatin inhibited biofilm formation and potently prevented P. aeruginosa virulence in C. elegans, and the complestatin target was identified as a nitrite transporter through a genome-wide screen of the Keio collection, which was validated using a nitrite transporter-overexpressing strain. Second, E. coli and P. aeruginosa nitrite transporter mutants exhibited reduced biofilm formation and c-di-GMP levels, whereas nirC overexpression increased biofilm formation. Furthermore, it was demonstrated that the nitrite transporter partially suppressed NIR transcription, which caused the following sequential processes: low NO production, followed by DGC activation, c-di-GMP production, and, ultimately, biofilm formation stimulation.
P. aeruginosa produced NO by NIR of the denitrification pathway using nitrate/ nitrite. Given the fact that complestatin and the nitrite transporter mutant produced NO via NIR activation, complestatin treatment and the nitrite transporter mutant are expected to increase NO production and consequently biofilm formation inhibition in the presence of nitrate. Indeed, complestatin more potently decreased biofilm and c-di-GMP levels in the presence of nitrate. The nitrite transporter mutant also presented greater reductions in biofilm formation and c-di-GMP levels in the presence of nitrate. Of note, NIR expression in the nitrite transporter mutant was elevated in the presence of nitrite, whereas that in the wild-type and QS mutants (lasR and rhlR mutants) was not. Considering that sufficient nitrite is present in CF airway surface liquid and sputum (27,49), these data suggest that nitrite transporter inhibitors could have important clinical implications and advantages over QS inhibitors.
Denitrification genes such as nar, nir, nor, and nos are induced under anaerobic or low-oxygen conditions in the presence of nitrate or nitrite in P. aeruginosa (50,51). Recently, it was reported that denitrification can also be activated under aerobic conditions in P. aeruginosa (23,28). Expression of denitrification genes is tightly controlled by the arginine nitrate regulator (ANR) and dissimilative nitrate regulator (DNR) transcription factors. The master regulator ANR, a homologue of E. coli FNR (a well-characterized oxygen-sensing regulator), activates another CRP/FNR-related transcriptional DNR under anaerobic or low-oxygen conditions (51,52), which in turn activates transcription of all denitrification genes (50). The two-component nitrate sensor-response regulator NarX/NarL in cooperation with ANR induces the expression of the genes narK, dnr, nirQ, and nar (51). Another anaerobic NIR regulator, NirQ, is predicted to be involved in fine-tuning the expression and activation of NIR and nitric oxide reductase (51). In E. coli, FNR activates transcription of the denitrification genes. In this study, among the denitrification genes, mainly NIR genes (nirS) were activated dose dependently in complestatin-treated PAO1 and PA14, although narG (plasmic nitrate reductase) was a somewhat activated in PA14 (see Fig. S6B and C in the supplemental material). Among transcriptional regulators, nirQ was somewhat activated in PAO1 and PA14 ( Fig. S6B and C). However, given the fact that mainly NIR was activated, it was suggested that other regulation might be involved. In the nasA mutant of P. aeruginosa, transcription of mainly NIR was dramatically activated, although narG was somewhat activated in PA14, which is consistent with that in complestatin-treated cells (Fig. S6A). Additionally, in E. coli, transcription of only NIR was affected in complestatin-treated cells, the nirC mutant, and the nirC-overexpressing strain, whereas that of other genes and FNR were not affected (Fig. 5). These results indicate that the nitrite transporter suppressed transcription of NIRs, which could occur via an unknown mechanism. Also, the possibility of the nitrite transporter itself as a transcriptional factor cannot be excluded.
By investigating the mechanism of biofilm formation with complestatin, this study describes a novel mechanism governing biofilm formation in E. coli and P. aeruginosa. Nitrite transporter was identified as a protein target of complestatin. Nitrite transporter partially suppressed NIR in E. coli and P. aeruginosa. NO, produced by the partially suppressed NIR, activated DGCs, not PDEs, and subsequently produced c-di-GMP that is essential for biofilm formation. This finding indicates that the partial suppression of NIR by nitrite transporter is a prerequisite for biofilm formation in E. coli and P. aeruginosa, and NO plays roles in both biofilm formation and dispersal, conflicting processes, via differential regulation of NIR. Thus, this study provides nitrite transporters as new antibiofilm targets. In this respect, our findings provide new insight into how the biofilm life cycle is regulated through NO and how biofilm can be prevented.

MATERIALS AND METHODS
See Text S1 in the supplemental material for additional details regarding the materials and methods. Materials. Complestatin was isolated from Streptomyces chartreusis AN1542 mycelia as stated in our previous study (35). Vancomycin, FC, rifampin, DAF-2DA, and C-PTIO potassium salt were purchased from Sigma-Aldrich (St. Louis, MO, USA).
Quantitative analysis of EPS in biofilms. EPS in biofilms of P. aeruginosa was evaluated as previously described (54).
Confocal laser scanning microscopy for biofilm visualization and intracellular NO detection. Confocal laser scanning microscopy for biofilm visualization (55) and intracellular NO detection (56) was performed as previously described with some modifications.
Measurement of QS signaling molecules. The effects of complestatin on the production of QS signaling molecules were determined as previously described (57).
Quantitative cellular c-di-GMP analysis by liquid chromatography-tandem mass spectrometry. c-di-GMP in P. aeruginosa and E. coli biofilms was analyzed using a previously described method (17).
PDE and DGC activity assays. The DGC and PDE activities in P. aeruginosa and E. coli were evaluated as described previously (17).
Screening the E. coli Keio collection library. Keio mutants representing 3,801 genes were screened to identify the specific target protein of complestatin.
Overexpression assay. Overexpression E. coli strains were constructed using the pBAD-TOPO TA expression vector.
Expression and RT-qPCR of the nitrite reductase gene and related genes. The transcription of the nitrite reductase gene and related genes was determined as previously described with some modifications (17).