A Survival Strategy for Pseudomonas aeruginosa That Uses Exopolysaccharides To Sequester and Store Iron To Stimulate Psl-Dependent Biofilm Formation

ABSTRACT

Exopolysaccharide Psl is a critical biofilm matrix component in Pseudomonas aeruginosa, which forms a fiber-like matrix to enmesh bacterial communities. Iron is important for P. aeruginosa biofilm development, yet it is not clearly understood how iron contributes to biofilm development. Here, we showed that iron promoted biofilm formation via elevating Psl production in P. aeruginosa. The high level of iron stimulated the synthesis of Psl by reducing rhamnolipid biosynthesis and inhibiting the expression of AmrZ, a repressor of psl genes. Iron-stimulated Psl biosynthesis and biofilm formation held true in mucoid P. aeruginosa strains. Subsequent experiments indicated that iron bound with Psl in vitro and in biofilms, which suggested that Psl fibers functioned as an iron storage channel in P. aeruginosa biofilms. Moreover, among three matrix exopolysaccharides of P. aeruginosa, Psl is the only exopolysaccharide that can bind with both ferrous and ferric ion, yet with higher affinity for ferrous iron. Our data suggest a survival strategy of P. aeruginosa that uses exopolysaccharide to sequester and store iron to stimulate Psl-dependent biofilm formation.

IMPORTANCE Pseudomonas aeruginosa is an environmental microorganism which is also an opportunistic pathogen that can cause severe infections in immunocompromised individuals. It is the predominant airway pathogen causing morbidity and mortality in individuals affected by the genetic disease cystic fibrosis (CF). Increased airway iron and biofilm formation have been proposed to be the potential factors involved in the persistence of P. aeruginosa in CF patients. Here, we showed that a high level of iron enhanced the production of the key biofilm matrix exopolysaccharide Psl to stimulate Psl-dependent biofilm formation. Our results not only make the link between biofilm formation and iron concentration in CF, but also could guide the administration or use of iron chelators to interfere with biofilm formation in P. aeruginosa in CF patients. Furthermore, our data also imply a survival strategy of P. aeruginosa under high-iron environmental conditions.

INTRODUCTION

Pseudomonas aeruginosa is a ubiquitous environmental microorganism. It is also an opportunistic pathogen that is often found in the hospital environment. P. aeruginosa can cause severe acute nosocomial infections in immunocompromised individuals or chronic lung infections in cystic fibrosis (CF) patients (1). These infections generally persist despite the use of intensive antimicrobial therapy and have been linked to the formation of antibiotic-resistant biofilms, wherein bacterial cells were encased in an extracellular polymeric substance (EPS) matrix (2, 3). It is generally accepted that the biofilm matrix functions as both a structural scaffold and protective barrier for bacteria in biofilms (4, 5). The iron content of CF sputum is often highly increased compared to that in the sputum of healthy controls, which has also been considered a potential factor in the persistence of P. aeruginosa (6–8). Mineral elements are often found in the biofilm matrix, yet it is largely unknown how the metal ions associate and/or react with matrix components.

Exopolysaccharide is an important and main matrix component of P. aeruginosa biofilms. Until now, at least three types of exopolysaccharides (Psl, Pel, and alginate) have been identified in P. aeruginosa that play roles in structure maintenance and antibiotic resistance of biofilms (9–13). Psl, a repeating pentasaccharide containing d-mannose, d-glucose, and l-rhamnose (14), is critical to initiate and maintain biofilm architecture in both mucoid (alginate overproducer) and nonmucoid strains (15–17). It can anchor on a bacterial surface to facilitate bacterial cell-cell and cell-surface interactions (13). Psl is a key scaffolding matrix component in the biofilm of nonmucoid P. aeruginosa strains, which can form a fiber-like matrix that acts as a backbone allowing other matrix components, such as extracellular DNA (eDNA), to associate with it (18). In addition, Psl also has protective roles against immune attacks. It has been shown that P. aeruginosa Psl polysaccharide could reduce neutrophil phagocytosis and the oxidative response via limiting complement-mediated opsonization (19).

The 12 cotranscribed psl genes (pslA to pslL) are required for the synthesis of Psl exopolysaccharide in P. aeruginosa (14). Psl synthesis is under multiple levels of control (20). The alternative σ-factor RpoS is a positive transcriptional regulator of psl gene expression (21), and AmrZ acts as a repressor to inhibit the transcription of psl genes (22). As a transcriptional regulator, AmrZ also positively regulates twitching motility and alginate synthesis (23, 24). RsmA is a posttranscriptional regulator which binds to the 5′-untranslated region of psl mRNA to block the translation of Psl proteins (21). Besides, the synthesis of Psl, Pel, alginate, and rhamnolipid is coordinated through competition for common sugar precursors in their biosynthesis pathway (17). In addition, a high level of intracellular cyclic-di-GMP enhances Psl production (20).

Iron is an essential and scarce nutrient for bacteria, which can also be toxic for bacteria at a high level. Thus, iron uptake and sequestration are critical for bacterial survival in the environment and host. For most pathogens, including P. aeruginosa, there is intense competition for iron with the host (25). Under iron-limiting conditions, P. aeruginosa secretes two types of siderophores, pyoverdine and pyochelin, to scavenge iron (26). Iron chelating with lactoferrin has an inhibitory effect on P. aeruginosa biofilm formation (27), and a high iron level promotes biofilm formation (28, 29). These results suggest that iron serves as a signal for P. aeruginosa biofilm development. In support, P. aeruginosa mutants that are unable to scavenge adequate amounts of iron from their environment are defective in biofilm formation (30, 31). This effect appears to be under the control of the ferric uptake regulator Fur, a global repressor of gene transcription in iron-rich environments found in P. aeruginosa and other members of the Proteobacteria (32, 33). A recent report has also shown that iron regulates the expression of alginate, a main exopolysaccharide of mucoid P. aeruginosa (34). However, the mechanism remains unclear about how iron regulates P. aeruginosa biofilm development and the biofilm matrix component.

A quorum-sensing signal in P. aeruginosa, PQS [2-heptyl-3-hydroxy-4(1H)-quinolone], has been found to form a complex with iron to facilitate siderophore-mediated iron delivery (35, 36), yet whether a biofilm matrix component can bind or sequester iron has not been reported. In this study, we find that iron can bind with exopolysaccharide Psl in biofilms and in vitro. More strikingly, a high level of iron enhances Psl production to stimulate the biofilm formation of P. aeruginosa by reducing rhamnolipid production and inhibiting the transcription of AmrZ. Our data indicate a new strategy of P. aeruginosa to compete for iron and store iron for its community.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The P. aeruginosa strains used in this study are listed in Table 1. Unless otherwise indicated, P. aeruginosa was grown at 37°C in Luria-Bertani lacking sodium chloride (LBNS), Jensen's chemically defined medium (37), or M9 medium (Molecular Cloning). Biofilms of P. aeruginosa were cultured in Jensen's medium at 30°C.

TABLE 1.

Pseudomonas aeruginosa strains used in this study

Strain	Relevant characteristics	Reference or source
PAO1	Wild-type strain	68
PA14	Wild-type strain	69
WFPA801	psl-araC-pBAD promoter replacement strain of PAO1	15
FRD1	Mucoid CF isolate, mucA22 mutation	70
PDO300	Mucoid PAO1, mucA22 mutation	71
PAO1 ΔpvdS	pvdS in-frame deletion strain of PAO1	72
PAO1 Δpsl	Psl negative, the promoter of psl operon deletion strain of PAO1, also named WFPA800	15
PAO1 ΔpelA	pelA; polar mutant of the pel operon	73
PAO1 Δpsl Δpel	psl and pel operon promoter double-deletion mutant of PAO1	73
pslA-lacZ	pslA lacZ integrated into attB site of PAO1	21
amrZ-lacZ	amrZ lacZ integrated into attB site of PAO1	This study
rhlI-lacZ	rhlI lacZ integrated into attB site of PAO1	This study
rhlR-lacZ	rhlR lacZ integrated into attB site of PAO1	This study
lasI-lacZ	lasI lacZ integrated into attB site of PAO1	This study
lasR-lacZ	lasR lacZ integrated into attB site of PAO1	This study
rhlAB-lacZ	rhlAB lacZ integrated into attB site of PAO1	57
PAO1/pCdrA::gfps	PAO1 containing c-di-GMP reporter plasmid pCdrA::gfp	21
PAO1ΔpslpBAD-pel	pslBCD deletion mutant of PAO1 with an inducible pel operon	74

Iron chelation assay.

The iron-chelating abilities of Psl, Pel, and alginate were determined by the ferrozine method (38). By this method, iron is detected in the ferrous state as the ferric iron is reduced by ascorbate to ferrous iron. Briefly, different amounts of polysaccharide samples were added to 100 μM FeSO₄ or FeCl₃ solution. After 1 h of incubation at room temperature, 0.5 M ascorbate (10 μl) was added to each sample (600 μl). The iron-detecting reaction was initiated by the addition of 50 mM ferrozine (10 μl), and the mixture was shaken vigorously and left standing at room temperature for 3 min. To stop the reaction, 0.5 M EDTA (10 μl) was added. The absorbance at 562 nm was measured, and the corresponding iron concentration in the reaction system was quantified according to the corresponding standard curves.

Biofilm matrix staining.

The air-liquid interface biofilms (pellicles) were grown in glass chambers (chambered #1.5 German coverglass system; Nunc) with glass coverslips at the bottom. For confocal laser scanning microscope (CLSM) observation, buffers were gently sucked out from glass chambers to allow pellicles to drop down on the coverslips. The biofilms were stained by SYTO9 (Molecular Probes; Invitrogen). Psl matrix was stained by tetramethyl rhodamine isothiocyanate (TRITC) or fluorescein isothiocyanate (FITC)-labeled hippeastrum hybrid lectin (amaryllis) (HHA) at 100 μg/ml (EY Labs, Inc.), as described elsewhere (13). The Fe³⁺ in biofilms was stained by a synthesized iron-specific fluorescent probe (39). The biofilms were stained for 10 min in the dark by the iron-specific probe at 1 μM final concentration in phosphate-buffered saline (PBS) buffer.

Microscopy and image acquisition.

All fluorescent images were acquired by an FV1000 CLSM (Olympus, Japan). Images were obtained using a 63×/1.3 objective. CLSM-captured images were subjected to quantitative image analysis using the COMSTAT software, as previously described (40). The biofilm biomass was quantified from SYTO9-stained images. The colocalization of iron and Psl matrix was analyzed by using the Interactive 3D Surface Plot plugin of the ImageJ software (version 1.43u; NIH). The colocalization of iron and Psl in biofilm images was analyzed and calculated as previously described (41).

Swarming assay.

A swarming motility assay was performed as previously described by using nutrient broth with 0.5% Difco Bacto agar (42). After inoculation, swarming plates were incubated for another 12 to 24 h at 37°C until swarming zones formed. All data were obtained from three independent experiments.

Preparation of exopolysaccharides and immunoblot analysis.

Alginate was bought from Sigma-Aldrich. For the iron-binding assay, the Psl and Pel extracts were prepared from 24-h cultures of P. aeruginosa WFPA801 and PAO1 ΔpslpBAD-pel, respectively, as previously described (14). For immunoblot analysis, Psl extracts were obtained from 24-h cultures of WFPA801, and immunoblotting was done as previously described (14). The densitometry software Quantity One (Bio-Rad) was used to quantify the immunoblotting data.

Microtiter dish biofilm assay.

A 1:100 dilution of an overnight culture was inoculated into a microtiter dish (Falcon 3911). After incubation at 30°C for 24 h, the planktonic bacteria were sucked out. The wells were washed twice with distilled water, stained by the addition of 0.1% crystal violet, and solubilized in 30% acetic acid, and the A₅₆₀ was measured as described previously (15).

ITC.

Microcalorimetric measurements of the binding of iron ions to Psl were performed on a Nano isothermal titration calorimetry (ITC) 2G at 25°C (TA Instruments, USA), as described previously, with modifications (43). Psl was dissolved into deionized water. Psl (0.5 mg/ml) was dispensed into the microcalorimetric cell (volume, 1.3 ml), and the iron solution (10 mM) was used to fill the syringe compartment (volume, 250 μl). Iron was titrated in 10-μl portions (3.14 μl for the first injection) into the Psl-containing cell under constant stirring, and the heat of reaction was plotted versus time. Data analysis was executed by the NanoAnalyze software.

Construction of transcriptional lacZ reporter gene fusion strains.

For the amrZ lacZ strain, a 448-bp DNA fragment containing a promoter of amrZ from the PAO1 chromosome was obtained by PCR with primers PamrZF1 (5′-CCGGAATTCGATCTCCAGCCTCAGGTTG-3′) and PamrZR1 (5′-CGCGGATCCGGCGCATACATTGAACCTG-3′) (underlined bases denote restriction enzyme sites). For the lasI lacZ strain, a 487-bp DNA fragment containing a promoter of lasI from the PAO1 chromosome was obtained by PCR with primers pLasI-F (GGAATTCTGGCCGTTAATTTGGGTCT) and pLasI-R (CGGGATCCTTGCGCTCCTTGAACACT); for the lasR lacZ strain, a 591-bp DNA fragment containing a promoter of lasR from the PAO1 chromosome was obtained by PCR with primers pLasR-F (GGAATTCCCGGGCTCGGCCTGTTCT) and pLasR-R (CGGGATCCGGATGGCGCTCCACTCCA); for the rhlI lacZ strain, a 300-bp DNA fragment containing a promoter of rhlI from the PAO1 chromosome was obtained by PCR with primers pRhlI-F (GGAATTCTTCCACCACAAGAACATCCA) and pRhlI-R (CGGGATCCCCCTTCCAGCGATTCAGA); for the rhlR lacZ strain, a 430-bp DNA fragment containing a promoter of rhlR from the PAO1 chromosome was obtained by PCR with primers pRhlR-F (GGAATTCGCGCGAGCAGGAGTTGC) and pRhlR-R (CGGGATCCTAATCGAAGCCCAGGCGC). The PCR production and mini-CTX-lacZ vector were doubly digested with EcoRI and BamHI and ligated to generate a fusion plasmid. The plasmids were conjugated into the chromosomal attB sites of the P. aeruginosa strains used in this study, as described previously (44).

β-Galactosidase assays.

β-Galactosidase activity was quantitatively assayed as described elsewhere, with some modifications (45). P. aeruginosa strains were grown in Jensen's medium at 37°C with shaking at 200 rpm for 8 h (log-phase culture) or 24 h (stationary-phase culture). One-milliliter culture aliquots were resuspended in 200 μl of Z-buffer and frozen/thawed three times to lyse bacteria. Cell lysates were examined for β-galactosidase activity as well as for total protein content by a bicinchoninic acid (BCA) assay (Pierce, USA). All β-galactosidase activity units were normalized by total protein per milliliter of aliquot. All samples were done in triplicate.

Rhamnolipid assay.

After a 2-day incubation at 37°C in nutrient broth liquid medium, rhamnolipids were extracted as previously described (46). Briefly, cells were removed from the culture broth by centrifugation (10 min at 6,000 × g), and the supernatant was acidified with concentrated HCl until pH 2 was reached. An equal volume (50 ml) of chloroform-methanol was added to the acidified supernatant and then vortexed for 1 min. The lower organic phases were collected and evaporated to dryness and resuspended in 3 ml of methanol. After being filtered through a 0.22-μm-pore-size membrane, the samples were repeatedly evaporated to constant weight. The dry weights of rhamnolipids produced by different strains were quantified and normalized to the PAO1 level.

qRT-PCR.

Strains were cultured in Jensen's medium for 24 h with shaking. Total RNA was purified from a 1-ml culture using the RNeasy minikit (Qiagen), followed by immediate DNase I treatment using a DNA-free DNase kit (Ambion). Then, RNA was reverse transcribed to cDNA by Moloney murine leukemia virus reverse transcriptase (Promega). Quantitative reverse transcription-PCR (qRT-PCR) was performed on 4 ng/μl of cDNA using 2× SYBR green PCR master mix (Roche) and was carried out in the Roche LightCycler 480 system by using gene-specific primers. The primers used for amrZ and rpsL are amrZ-F (CGCTGACAAATTCGTCGTTC) and amrZ-R (CAGGCGAACACCGAGATTG), and rpsL-F (TACATCGGTGGTGAAGGT) and rpsL-R (CCCTGCTTACGGTCTTTG), respectively. Relative transcriptional levels were determined by the comparative standard curve method, as described previously (47). All samples were normalized to the constitutively produced rpsL transcript.

Alginate assay.

Alginates were collected from cultures grown in LBNS broth with rapid aeration at 37°C for 24 h, and levels were determined as previously outlined, with modifications (48). Briefly, samples (5 ml) of cultures were mixed with 5 ml of saline, and the cells were removed by centrifugation (12,000 × g for 30 min). The culture supernatant was mixed with 5 ml of 2% cetyl pyridinium chloride, and the precipitated alginate was collected by centrifugation (12,000 × g for 10 min at room temperature). The pellet was dissolved in 10 ml of 1 M NaCl, precipitated again with 10 ml of cold (−20°C) isopropanol, and dissolved in 10 ml of saline. The concentration of alginate in solution was determined by the carbazole method described by Knutson and Jeanes (49), in which a solution of alginate (30 μl) was mixed with 1.0 ml of borate-sulfuric acid reagent (10 mM H₃BO₃ in concentrated H₂SO₄) and 30 μl of carbazole reagent (0.1% in ethanol). The mixture was then incubated in a 55°C bath for 30 min, and the absorbance at 530 nm was determined spectrophotometrically. The alginate concentration was determined by extrapolation from a standard curve with various concentrations (0 to 50 μg/ml) of alginate.

RESULTS

Iron promotes biofilm formation via elevating Psl production in P. aeruginosa.

It was reported that the iron-replete condition (10 to 100 μM) promoted biofilm formation and a pvdS mutant (which lost the ability to produce pyoverdine, an iron chelator) had a defect in biofilm formation (28, 29). As Psl is crucial for the maintenance of biofilm biomass in P. aeruginosa, we hypothesized that iron might affect Psl-mediated biofilm formation. To test this hypothesis, we first examined the Psl production of the pvdS mutant, which did produce less Psl than that of isogenic wild-type PAO1 strain (half of the PAO1 level, Fig. 1). We then compared the biofilm and Psl production of PAO1 with pel or psl deletion mutants grown under iron-limiting (2 μM) and iron-replete (100 μM) conditions. High iron (100 μM) stimulated biofilm formation in PAO1 in a microtiter dish assay. Inconsistent with a biofilm phenotype, Psl production of PAO1 was increased accordingly, as revealed by anti-Psl immunoblotting (Fig. 2B). This phenomenon was more pronounced in the pel deletion mutant (Fig. 2A and B). In contrast, the biofilm biomass of Δpsl mutant strains (two PAO1-derived Δpsl mutant strains and a native Δpsl mutant strain, PA14) was not significantly affected by the iron concentrations (Fig. 2A), and all Δpsl mutant strains had little biofilm biomass. These results suggested that the effect of iron on biofilm formation was mediated mainly by Psl.

FIG 1.

Psl production of PAO1 and PAO1 ΔpvdS. A representative immune-dot blotting result is shown below each corresponding column. The Psl synthesized by PAO1 ΔpvdS is significantly lower than that of PAO1 (P < 0.01).

FIG 2.

Iron promotes biofilm formation via elevating the Psl level. (A) Iron stimulated the Psl-mediated biofilm formation in P. aeruginosa PAO1. PAO1 ΔpelA, Δpsl mutant strains, and PA14 were used for comparison. (B) Psl production of PAO1 and PAO1 pelA mutant at different iron levels. The amount of Psl was quantified by anti-Psl immunoblotting. Con., concentration. (C) The effect of iron on the production of Psl in nonmucoid strain PAO1 and mucoid strains. The amount of Psl was quantified by anti-Psl immunoblotting and has normalized to the corresponding Psl level under the 2 μM iron condition (for PAO1, 1 = 36 μg/ml; for PDO300, 1 = 15 μg/ml; for FRD1, 1 = 12 μg/ml). (D) The CLSM images of PAO1 biofilms grown under iron-depleted (2 μM) and -replete (100 μM) conditions. Biofilm bacteria were stained by SYTO9 (green fluorescence) and their Psl matrix was stained by lectin HHA-TRITC (red fluorescence). The selected horizontal sectioned images (square) and vertical sectioned images are shown. The biomass was indicated on the corresponding image. (E) Comparison of bacterial and Psl biomass. Shown are the ratio of Psl to biomass and the maximum (Maxim) thickness of corresponding biofilms (shown in panel D) at two iron concentrations. (F) Iron stimulated the biofilm formation of mucoid strains. Shown are the biofilm biomass of PAO1-derived mucoid strain PDO300 and mucoid CF isolate FRD1 grown under different iron conditions. **, P < 0.01.

To further confirm, we applied confocal laser scanning microscopy (CLSM) to visualize and examine the PAO1 biofilms under two iron concentrations (2 μM and 100 μM). The biofilm biomass was increased 2.5-fold under the 100 μM iron condition compared to that under the 2 μM iron condition. Accordingly, the amount of Psl in biofilms was enhanced 5-fold under the iron-replete condition compared to that under the iron-limiting condition. On the contrary, the biofilm biomass of the PAO1 Δpsl strain had only a 1.2-fold increase while grown under 100 μM iron compared to that under the 2 μM iron condition, which was inconsistent with the results from the microtiter dish biofilm assay (Fig. 2A, and data not shown). In the same growth medium, 100 μM iron also slightly increased the planktonic bacterial growth of both the PAO1 and PAO1 Δpsl strains compared to that under the 2 μM iron condition (the change in optical density at 600 nm [OD₆₀₀] was <1.2-fold). These results indicated that a high level of iron elevated Psl production, which led to a thicker biofilm with more biomass in P. aeruginosa (Fig. 2D and E).

The mucoid P. aeruginosa strains were often isolates from CF patients, which usually produce a large amount of alginate but little Psl. As shown above, the effect of iron on Psl of PAO1 was different from that on alginate, whose production was reduced by s high level of iron in mucoid strains (34). To know whether the iron could stimulate Psl production and biofilm formation in mucoid strains, we examined the PAO1-derived mucoid strain PDO300 and the CF isolated mucoid strain FRD1. Similar to nonmucoid strain PAO1, the amount of Psl produced by the two mucoid strains increased when grown under iron-replete conditions (≥10 μM) (Fig. 2C). Their biofilm biomass was enhanced accordingly under such iron conditions (Fig. 2F), which were the conditions shown to reduce the alginate production in mucoid strains (34). These data revealed that iron could also stimulate Psl production and Psl-mediated biofilm formation in both nonmucoid and mucoid strains.

Iron increases Psl production by repressing armZ transcription.

To investigate the molecular mechanisms underlying iron-stimulated Psl production, we first detected the c-di-GMP level of PAO1 under different iron concentrations by using the fluorescence-based reporter plasmid pCdrA::gfp^s as an indicator. The iron concentration did not appear to impact the c-di-GMP level in PAO1 (data not shown), which suggested that the c-di-GMP level was not involved in the iron-stimulated Psl production. We then detected the expression of psl genes by pslA transcriptional lacZ fusion. The psl expression of PAO1 was elevated under high-iron conditions, and the cultures grown with 50 μM iron showed the highest pslA transcription levels (Fig. 3A). Consistently, the highest Psl production was found under 50 μM iron conditions (Fig. 2B and 3C). These results suggested that iron stimulated Psl production by increasing the transcription of corresponding psl genes in P. aeruginosa.

FIG 3.

Iron influences on the transcription of psl and amrZ. (A) β-Galactosidase (β-gal) assays for transcriptional pslA-lacZ fusion at different iron levels. Shown are the results detected at early stationary phase (24 h) of bacterial growth. (B) β-Galactosidase assays for translational fusion of amrZ-lacZ at different iron levels on early stationary phase (24 h). (C) Relative amrZ expression of the stationary cultures of PAO1, PDO300, and FRD1 grown at different iron levels. Shown are the results of RT-PCR. The PAO1 level at 2 μM iron was used as a control. **, P < 0.01; *, P < 0.05.

RpoS is a positive transcriptional regulator of the psl gene in stationary phase (21). Thus, we detected Psl production in the PAO1 isogenic rpoS mutant under three iron concentrations (2 μM, 50 μM, and 100 μM). High iron increased Psl production in the rpoS mutant to the same extent as that in PAO1 (data not shown), which suggested that the iron-stimulated Psl production was not controlled by RpoS. The transcriptional factor AmrZ represses transcription of the psl operon (22). We then examined the impact of iron on the expression of amrZ by utilizing an amrZ-lacZ transcriptional fusion in PAO1. High iron indeed reduced the expression of amrZ in PAO1 at stationary phase (Fig. 3B) but not at mid-log phase (data not shown). In addition, 50 μM iron led to the lowest amrZ expression, which was consistent with the highest psl expression under such conditions (Fig. 3A and B). We also examined the effect of iron on the amrZ expression of mucoid strains, which showed a similar level of repression in PDO300 and FRD1 (Fig. 3C). Our data demonstrated that AmrZ was one of the regulating factors involved in the iron-stimulated Psl production and biofilm formation.

Iron stimulates Psl production by reducing the synthesis of rhamnolipids.

Bacterial motility impacts the architecture of biofilms (50, 51). We next examined three types of motility in P. aeruginosa under iron-replete conditions. Interestingly, we found that high iron levels impaired swarming motility only (Fig. 4A) but not flagellum-driven swimming or type IV pilus-mediated twitching motility (data not shown). Swarming motility is a result of complex multicellular activities that require flagella and type IV pili as well as biosurfactants, which are mainly rhamnolipids in P. aeruginosa (52, 53). We reasoned that high iron levels would impact rhamnolipid biosynthesis, because the functions of flagella and pili were not affected by high iron concentrations. In addition, the previous findings indicated that rhamnolipid production was induced under iron-limiting conditions (53–55). The rhamnolipids of P. aeruginosa grown under iron-replete conditions (both 50 and 100 μM) were significantly lower than those under the iron-limiting condition (2 μM) (Fig. 4B), indicating the inhibition of high iron concentration in the biosynthesis of rhamnolipids.

FIG 4.

High iron concentration impacts rhamnolipid production and the transcription of rhl genes. (A) Swarming motility of PAO1 at different iron levels. The swarming zone is indicated under each corresponding image. 0, no visible swarming zone. (B) Rhamnolipid production of PAO1 at different iron levels. (C) β-Galactosidase assays for transcriptional fusions of lasR-lacZ, lasI-lacZ, rhlAB-lacZ, rhlR-lacZ, and rhlI-lacZ at different iron levels. **, P < 0.01.

The biosynthesis of rhamnolipids requires RhlA and RhlB, whose transcription was directly activated by the RhlI/RhlR quorum-sensing (QS) system (56). We then investigated the influence of iron on the expression of rhlAB, rhlI, and rhlR, as well as lasI and lasR, genes encoding another QS signal synthase and corresponding receptor in P. aeruginosa that are also the positive regulators of rhlI and rhlR. Strikingly, high iron concentrations repressed the transcription of rhlAB, rhlI, and rhlR but not lasI and lasR (Fig. 4C). These results indicated that high iron levels inhibited the expression of RhlI/RhlR and RhlAB in order to reduce the rhamnolipid biosynthesis. This revealed another way that iron stimulated Psl production in P. aeruginosa at the posttranslational level because the decrease in rhamnolipid production can increase Psl production (57).

Iron associates with Psl matrix in P. aeruginosa biofilm.

In order to know whether iron can bind to Psl under physiological conditions, particularly in biofilms, we applied a highly selective and sensitive iron-specific fluorescent probe to label iron and to detect the localization of iron in the Psl matrix of P. aeruginosa biofilms (58). Biofilms after 1 day of growth under iron-replete (100 μM) or iron-limiting (2 μM) conditions were stained by Psl-staining lectin HHA-FITC (green fluorescence) and an iron-specific fluorescent probe (red fluorescence). The iron distribution pattern appeared similar to the Psl matrix, especially in the biofilm grown with a high concentration of iron (100 μM) (Fig. 5A). Colocalization analysis indicated that iron mostly associated with Psl matrix in biofilms under either the iron-replete condition (100 μM) (image at right, Fig. 5B) or the iron-limiting condition (2 μM) (image at left, Fig. 5B). Interestingly, compared to the iron-limiting condition, biofilms grown under iron-replete conditions appeared to have more clear Psl fiber structure and stronger Psl-staining signal (Fig. 5A). These data suggested that iron could associate with Psl matrix in biofilms, suggesting that Psl might serve as an iron pool for subsequent biofilm growth.

FIG 5.

The colocalization analysis of Psl matrix and iron in biofilms. (A) Pellicle biofilms after 1 day of growth under iron-depleted (2 μM) and -replete (100 μM) conditions were observed and imaged by CLSM. The representative images are shown. A highly selective and sensitive iron-specific fluorescent probe (red fluorescence) was used to label and locate the iron distribution in P. aeruginosa biofilms. Lectin FITC-HHA was used to stain Psl (green fluorescence). The deviation map clearly shows that iron was colocalized with Psl in P. aeruginosa biofilms. (B) The scatterplot analysis of fluorescent signals from the corresponding biofilm images shown above to evaluate the colocalization of iron and the Psl matrix. The scatterplot was created by defining the Psl signal as the x axis and the Fe³⁺ signal as the y axis. The information from the images was transformed to a plot chart to reveal the relationship between the two images. The deviation map shown above was created by the data set calculated from the corresponding scatterplot.

Iron binds to Psl in vitro.

To further confirm the iron-chelating or -binding activity of Psl, we established a biochemical method using ferrozine as an indicator to determine free iron ions (59). In this assay, purified Psl was mixed with FeCl₃ solution in vitro. After 1 h of incubation, free iron ions (those not bound or chelated by Psl) were detected by ferrozine. While increasing the concentrations of Psl, the free iron ions were reduced gradually, indicating that Psl binds or chelates iron in a dose-dependent manner (Fig. 6A). To further confirm this, we used cellulase to partially degrade Psl (13) and then tested the iron-binding ability of Psl posttreatment. As a result, cellulase-treated Psl significantly lost the iron-binding activity compared to that of untreated Psl (Fig. 6B). We also tested the iron-binding activity of genomic DNA of P. aeruginosa, because extracellular DNA is another main biofilm matrix component and is mainly derived from genomic DNA in P. aeruginosa (18). DNA did not show any iron-binding activity in this assay (Fig. 6B). Collectively, our data indicate that Psl can bind iron in vitro.

FIG 6.

Iron can bind to Psl in vitro. (A) Psl binds iron in a dose-dependent manner. Different concentrations of purified Psl were mixed with a 100 μM iron solution. Ferrozine was used as an indicator to determine free iron. (B) Cellulase-treated Psl significantly reduced the iron-binding activity of Psl. Cellulase (4 mg/ml) was incubated with 100 μg/ml purified Psl for 24 h at 25°C prior to mixing with iron solution. The P. aeruginosa genomic DNA was used as a control to test the binding activity. **, P < 0.01.

Psl is the only exopolysaccharide in P. aeruginosa that can sequester both ferrous and ferric ion.

To know whether iron can bind to other exopolysaccharides produced by P. aeruginosa, such as alginate and Pel, we performed the ferrozine-based iron-binding assay for Psl, alginate, and Pel and compared their affinities to ferrous (Fe²⁺) and ferric (Fe³⁺) ions. Among the three exopolysaccharides, Psl showed the best iron-binding ability with either Fe²⁺ or Fe³⁺. In addition, the same amount of Psl could bind more Fe²⁺ than Fe³⁺ (Fig. 7A), indicating that Psl had higher affinity to Fe²⁺ than Fe³⁺. Pel had the ability to bind Fe²⁺ but not Fe³⁺ (Fig. 7A). In contrast, alginate had the ability to bind Fe³⁺ but could not bind Fe²⁺ (Fig. 7A).

FIG 7.

Comparison of iron-binding ability of exopolysaccharides produced in P. aeruginosa. (A) The binding ability of Psl, Pel, and alginate to ferric (Fe³⁺) and ferrous (Fe²⁺) iron. (B) The interaction of Psl with Fe²⁺ examined by ITC. (C) The ITC examination of the interaction between Psl and Fe³⁺.

We applied isothermal titration calorimetry (ITC) to further examine the iron-Psl interaction. FeSO₄ or FeCl₃ was used to titrate Psl dispersion. FeSO₄ showed an exothermic reaction with a sigmoidal saturation profile (Fig. 7B). The observed enthalpy change of the Fe²⁺-Psl interaction was in the range of −45 to −5 kJ/mol (Fig. 7B, lower panel), which indicates a strong interaction. However, the Fe³⁺-Psl reaction cannot have a sigmoidal saturation profile (Fig. 7C), suggesting a weak interaction or unstable iron binding.

Iron stored with Psl can be reused by P. aeruginosa.

Iron was shown to enhance the growth of P. aeruginosa (29). We use this phenomenon to test whether the iron stored with Psl can be reutilized by bacteria. Psl was purified from the WFPA801 strain (Table 1) grown under 100 μM iron, and 80 μg of such Psl (named Psl¹⁰⁰) was supplied into M9 medium to test its effect on the growth of PAO1. The OD₆₀₀ of PAO1 in M9 medium with Psl¹⁰⁰ was similar to that of PAO1 grown with 20 μM iron, which was significantly higher than that of PAO1 grown in M9 medium without iron (Fig. 8A). To know whether the increase in bacterial growth was due to Psl or iron associated with Psl, we used the same amount of Psl from regular preparation (named Psl²; WFPA801 was grown in medium with 2 μM iron) to test the effect of Psl on bacterial growth. The OD₆₀₀ of PAO1 in M9 with Psl² was significantly lower than that grown with Psl¹⁰⁰, although it was higher than that grown in M9 medium without iron supplement (Fig. 8A). In addition, no free iron was detected from either Psl¹⁰⁰ or Psl² by the ferrozine method. These results suggested that iron stored with Psl was very likely reutilized by P. aeruginosa.

FIG 8.

Reutilization of iron stored by Psl and a schematic summary. (A) The optical density at 600 nm (OD₆₀₀) of PAO1 after 24 h of growth in M9 medium with different concentrations of iron or Psl prepared from the P. aeruginosa strain WFPA801 grown at 100 μM and 2 μM iron (named Psl¹⁰⁰ and Psl², respectively). (B) A schematic shows the mechanisms of the iron-Psl interplay in P. aeruginosa. Iron ions (both Fe³⁺ and Fe²⁺) are sequestered and stored with exopolysaccharide Psl. Iron can stimulate Psl production by repressing AmrZ expression and reducing rhamnolipid synthesis, which enhances biofilm formation.

DISCUSSION

As a bacterial community, biofilm must have the ability to respond to all kinds of environmental signals and to maintain sustainable growth of its community. Iron is an essential, yet toxic, nutrient for bacteria. It has been known that a high iron level promotes biofilm formation (28, 29). However, the molecular mechanism for how iron controls P. aeruginosa biofilm development is not clear. Exopolysaccharide Psl is an important scaffolding matrix component in P. aeruginosa biofilm, which also serves as a protective barrier and a signaling molecule (13, 20, 60). Here, we show that Psl can bind and sequester iron to serve as an iron pool or an iron storage channel in a biofilm, consequently stimulating biofilm formation. Furthermore, a high level of iron could stimulate Psl production via repression of rhamnolipid production or inhibition of the transcription regulator ArmZ, which represses the expression of the psl gene operon and activates alginate production (22). Our data not only promote our understanding on the new biological functions of bacterial exopolysaccharides but also provide a new strategy for P. aeruginosa to compete for iron and store iron for the long-term survival of its community (Fig. 8).

Iron is a scarce nutrient for bacteria. Iron uptake and sequestration are very important for bacterial survival in the environment and host. P. aeruginosa harbors several iron acquisition systems to take up soluble iron from environments, including siderophores, hemes, and PQS (33). Our data in this study suggest that Psl is also an important source for iron storage. Psl has shown the best iron-binding ability among the three exopolysaccharides produced in P. aeruginosa. More strikingly, Psl is the only exopolysaccharide in P. aeruginosa that can bind both ferric (Fe³⁺) and ferrous (Fe²⁺) iron. Psl has high affinity to sequester Fe²⁺, which can protect bacteria from the harm of a high level of Fe²⁺, because soluble Fe²⁺ can directly diffuse into bacterial cells. In addition, iron is colocalized with Psl matrix in biofilms, and iron bound with Psl can be utilized by bacteria (Fig. 8A). Thus, iron stored with Psl can be saved for future use by bacteria. Fe²⁺ can be readily oxidized to Fe³⁺ under aerobic conditions. Psl shows weak or instable interaction with Fe³⁺. This means that iron ions can be easily released from Psl once it is oxidized to the Fe³⁺ form. Fe³⁺ can be chelated by siderophores and transported into bacterial cells when iron is scarce in the environment. This reveals a new way for bacteria to sequester and store irons for long-term survival (Fig. 8). Furthermore, our elucidation of the iron-binding activity of exopolysaccharides further expands our understanding of iron acquisition systems in P. aeruginosa and adds another level of protection under iron-starved and iron-replete conditions. It has been reported that polysaccharide produced by Klebsiella oxytoca under anaerobic conditions has an iron-binding feature (61). It has been concluded that the formation of an iron-exopolysaccharide complex is a strategy for bacteria to survive in mine drainages and under other acidic conditions. Considering exopolysaccharides to be a common biofilm matrix component, it could be a common strategy that bacteria utilize exopolysaccharide to sequester and store iron for the long-term survival of its community.

It is of great interest that at least two different mechanisms were identified for iron-stimulated biofilm formation (Fig. 8B). Our findings that iron could inhibit swarming motility but not swimming or twitching of P. aeruginosa led us to find out that iron represses the expression of the RhlI/RhlR QS system to reduce the synthesis of rhamnolipids. It is in line with our previous finding that there is competition for common sugar precursors between the biosynthesis pathways of exopolysaccharides and rhamnolipids (17). The repression of rhamnolipid synthesis by high levels of iron will lead to the increase in Psl production. Additionally, through inhibiting AmrZ, high iron might derepress the psl operon to allow more Psl synthesis. The link between iron and alginate production has long been noticed (62–64). Recently, Wiens et al. (34) reported that a high iron level inhibits the alginate production in P. aeruginosa, and a low level of iron leads to alginate expression and mucoidy. Our finding that AmrZ is inhibited by high iron levels provides an underlying mechanism for the decreased production of alginate. How is the iron signal sensed to repress the expression of psl, rhl, and amrZ? Banin et al. (30) suggested that the ferric uptake regulator Fur has involvement in the iron-regulated biofilm development. Although the Fur box has not been found in promoter regions of psl, rhl, and amrZ, the indirect involvement of Fur cannot be ruled out.

The iron content of CF sputum is often highly increased compared to that in the sputum of healthy individuals (8), which has been proposed to be a potential factor in the persistence of P. aeruginosa in CF. In addition, Fe²⁺ has been found in the sputum samples from CF patients (65, 66). Our data have shown that Psl has high affinity to sequester Fe²⁺, and a high level of iron stimulates Psl production and Psl-mediated biofilm formation. This reveals a mechanism for how P. aeruginosa persists in the CF lung. Furthermore, our results could also guide the administration or use of iron chelators, such as the FDA-approved iron chelators deferoxamine and deferasirox, to interfere with biofilm formation of P. aeruginosa (67). It should be considered that exopolysaccharide Psl and other potential iron-binding molecules may compete for iron with deferoxamine or deferasirox. In the case that deferoxamine or deferasirox outcompetes Psl, the resultant iron limitation condition would induce the production of several virulence factors and alginate.