Manganese Acts as an Environmental Inhibitor of Pseudomonas aeruginosa Biofilm Development by Inducing Dispersion and Modulating c-di-GMP and Exopolysaccharide Production via RbdA

ABSTRACT The opportunistic human pathogen Pseudomonas aeruginosa causes chronic infections that involve multicellular aggregates called biofilms. Biofilm formation is modulated by the host environment and the presence of cues and/or signals, likely affecting the pool of the bacterial second messenger cyclic diguanylate monophosphate (c-di-GMP). The manganese ion Mn2+ is a divalent metal cation that is essential for pathogenic bacterial survival and replication during the infection in a host organism. In this study, we investigated how Mn2+ alters P. aeruginosa biofilm formation via the regulation of c-di-GMP levels. Exposure to Mn2+ was found to temporally enhance attachment but impair subsequent biofilm development, apparent by reduced biofilm biomass accumulation and lack of microcolony formation due to the induction of dispersion. Moreover, exposure to Mn2+ coincided with reduced production of the exopolysaccharides Psl and Pel, decreased transcriptional abundance of pel and psl, and decreased levels of c-di-GMP. To determine whether the effect of Mn2+ was linked to the activation of phosphodiesterases (PDEs), we screened several PDE mutants for Mn2+-dependent phenotypes (attachment and polysaccharide production) as well as PDE activity. The screen revealed that the PDE RbdA is activated by Mn2+ and is responsible for Mn2+-dependent attachment, inhibition of Psl production, and dispersion. Taken together, our findings suggest Mn2+ is an environmental inhibitor of P. aeruginosa biofilm development that acts through the PDE RbdA to modulate c-di-GMP levels, thereby impeding polysaccharide production and biofilm formation but enhancing dispersion. IMPORTANCE While diverse environmental conditions such as the availability of metal ions have been shown to affect biofilm development, little is known about the mechanism. Here, we demonstrate that Mn2+ affects Pseudomonas aeruginosa biofilm development by stimulating phosphodiesterase RbdA activity to reduce the signaling molecule c-di-GMP levels, thereby hindering polysaccharide production and biofilm formation but enhancing dispersion. Our findings demonstrate that Mn2+ acts as an environmental inhibitor of P. aeruginosa biofilms, further suggesting manganese to be a promising new antibiofilm factor.

encased in an extracellular polymeric matrix (EPS), are the dominant mode of growth in nature and during infections (6). The biofilm mode of growth is beneficial for bacteria as it allows cells to maintain close to nutrients, promotes the exchange of genetic material, and confers protection from a variety of chemical and environmental stresses (e.g., nutrient limitation, desiccation, and shear forces) (7). Moreover, biofilms are associated with persistent and chronic infections, including cystic fibrosis (CF), that are refractory to treatment by conventional antibiotics (8,9). The opportunistic human pathogen Pseudomonas aeruginosa has been classified as one of the ESKAPE pathogens (i.e., Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, P. aeruginosa, and Enterobacter species) that are highly recalcitrant to antibiotic treatment, and its biofilm formation is considered the primary cause of mortality in patients with CF (9,10).
Biofilm formation by P. aeruginosa is a sequential and highly regulated process involving several distinct stages, including reversible and irreversible attachment, biofilm maturation stages I and II (defined by cluster and microcolony formation, respectively), and dispersion (11). The process is governed by several regulatory proteins, factors, and cues, including the surface sensing systems Pil-Chp and Wsp (12)(13)(14), SagS (15,16), and environmental stimuli such as Psl polysaccharide (17,18) or nitric oxide (NO) (19), leading to attachment and dispersion, respectively, as well as phosphodiesterases that contribute to dispersion, including NbdA (19,20), RbdA, and DipA (19,21). These various factors, cues, and/or signals have in common that they affect the intracellular level of the c-di-GMP that inversely regulates biofilm formation and motility (22)(23)(24).
Biofilm formation is furthermore affected by the presence and availability of metal ions. The effect of metal ion availability on biofilm formation has been investigated in several pathogenic bacteria, including P. aeruginosa. For example, magnesium limitation inhibits the expression of retS encoding the sensor kinase protein, which leads to increased EPS production and biofilm formation in P. aeruginosa (25). In contrast, levels of production of the biofilm matrix components CdrA and Psl are decreased in both mucoid and nonmucoid P. aeruginosa strains when grown under high-calcium conditions (26). The effect of iron on P. aeruginosa biofilm formation has also been explored. Elevated levels of iron reduce motility by P. aeruginosa cells and enhance biofilm microcolony formation, whereas iron-limiting conditions, obtained by the presence of the iron-binding glycoprotein lactoferrin, has been reported to coincide with reduced P. aeruginosa biofilm formation (27,28). Exposure of P. aeruginosa to ferric citrate, which is transported actively into the cells, restored biofilm formation even in the absence of the iron scavenger pyoverdine, indicating that iron uptake is the determinant of biofilm formation (28). Moreover, under iron-rich condition, Psl is increased as a result of inhibition of amrZ expression and reduced rhamnolipid production, with Psl acting as an iron reservoir by sequestering iron (29). Iron affecting biofilm formation is consistent with sputum of CF patients harboring relatively high concentration of iron (398 to 1,292 mg/L) relative to healthy lungs (30).
Another metal ion, manganese (Mn 21 ), has been reported to be essential for pathogenic bacterial survival during the infection in a host organism as well as various cellular processes, such as metabolism, transcriptional regulation, and resistance to oxidative stress (31)(32)(33). Interestingly, however, Mn 21 concentrations in the sputum of CF patients have been shown to be relatively low, ranging from 4 to 17 mg/L or 0.1 to 0.8 mg/L (30,34) and not differing significantly from the Mn 21 concentration present in the lungs of healthy patients or those suffering from chronic obstructive pulmonary disease (COPD) (34). Manganese has also been known to compete with iron for binding to enzymes such as Mn-superoxide dismutase due to similar sizes and coordination geometries with iron (31,35). However, little is known about the role of Mn 21 /manganese ions in P. aeruginosa biofilm development despite the structural similarity to iron and its physiological importance. The goal of this study was to determine the effect of manganese ions on P. aeruginosa biofilm formation and elucidate its mode of action.

RESULTS
Exposure of P. aeruginosa to Mn 2+ enhances attachment. Little is known about the effect of manganese on P. aeruginosa biofilm formation's contribution to the chronicity of infections. To explore the role of manganese in the development of P. aeruginosa biofilms, we first examined whether manganese affects the attachment capabilities of P. aeruginosa. Attachment was assessed using 96-well plates and growing P. aeruginosa PAO1 in Lennox broth (LB) lacking or supplemented with manganese for 24 h. The attached biomass was quantitated by crystal violet (CV) staining. Cells grown in the presence of 0.1 and 0.2 mM MnCl 2 showed increased CV-stainable biomass compared to its absence (Fig. 1A). However, elevated MnCl 2 concentrations exceeding 0.1 mM MnCl 2 (e.g., 0.2 mM) showed significant reduction in the CV-stainable biomass relative to that in attachment assays performed in the presence of 0.1 mM MnCl 2 (Fig.  1A). To determine if the enhanced attachment was linked to enhanced growth, growth curve measurements in the absence and presence of 0.1 mM MnCl 2 were performed. Under the conditions tested, no differences in the growth rates or growth behaviors were noted (Fig. 1B), suggesting that the increase in attachment by P. aeruginosa in LB supplemented with manganese was not the result of a general increase in growth rate.
To exclude the possibility that the complex LB may contain an unknown factor or factors involved in manganese-mediated attachment and determine whether the enhanced attachment was specific to manganese, we repeated the attachment assay using the defined Vogel and Bonner citrate minimal medium (VBMM) in the presence of increasing concentrations of MnCl 2 (0.05, 0.1, 0.2, and 1 mM). In contrast to LB, addition of 0.1 and 0.2 mM MnCl 2 did not coincide with a significant increase in attachment (Fig. 1C). However, addition of 1 mM MnCl 2 significantly enhanced attachment by wildtype P. aeruginosa PAO1 in VBMM (Fig. 1C), suggesting that manganese-mediated increase in attachment is independent of the growth medium used. Moreover, no growth difference in VBMM supplemented with and without 1 mM MnCl 2 was noted (Fig. 1D).
We next assessed relative attachment by wild-type P. aeruginosa PAO1 grown in VBMM supplemented with or without other divalent metal ions. Specifically, we tested ZnCl 2 and NiCl 2 . However, no significant difference in attachment was noted (Fig. 1E), suggesting that manganese ions specifically, and not (divalent) metal ions in general, enhance attachment by P. aeruginosa.
We furthermore asked whether the manganese-dependent enhanced attachment is also apparent under conditions mimicking the host environment. Therefore, the attachment was assessed using the San Diego State University artificial sputum medium (SDSU ASM) in the presence/absence of 0.1 mM MnCl 2 . Similar to LB (Fig. 1A), the presence of 0.1 mM MnCl 2 in SDSU medium coincided with increased attachment (Fig.  1F). The finding suggested that manganese ions also enhance attachment under conditions mimicking in vivo conditions. We furthermore explored whether the response was limited to the P. aeruginosa PAO1 laboratory strain. We therefore selected two P. aeruginosa clinical strains, CF1-13 and PA215, that were previously isolated from the CF lung and burn wounds, respectively. Their attachment capabilities were determined using LB supplemented with or without 0.1 mM MnCl 2 . Like P. aeruginosa PAO1 (Fig.  1A), the clinical isolates exhibited enhanced attachment capabilities in the presence of manganese relative its absence (Fig. 1G), suggesting that this effect is not limited to the laboratory strain.
Exposure of P. aeruginosa to Mn 2+ reduces biofilm biomass accumulation and the formation of structured biofilms. Since the exposure of P. aeruginosa to manganese enhanced attachment, we assumed that manganese also stimulates subsequent biofilm formation. To verify this assumption, we grew P. aeruginosa biofilms in 24-well plates in 5-fold-diluted LB supplemented with or without 0.1 mM MnCl 2 for up to 96 h and assessed the accumulation of the biofilm biomass by CV staining at 6,16,24,48,72, and 96 h postinoculation. As expected, exposure to manganese coincided with a significantly high biomass accumulation of P. aeruginosa PAO1 at early time points (6 and 16 h) relative to LB alone ( Fig. 2A). However, manganese enhancement of biomass accumulation was temporary, as continued incubation in the presence of Mn 21 resulted in significantly reduced biofilm biomass, with the inhibitory effect first noticeable following 3 days of biofilm growth ( Fig. 2A). A similar inhibitory effect on biofilm biomass was noted for 5-day-old biofilms formed by P. aeruginosa PAO1 and the two clinical isolates CF1-13 and PA215 when grown in 5-fold-diluted VBMM supplemented with 1 mM MnCl 2 (Fig. 2B).
Our observations indicated that Mn 21 enhanced attachment, but reduced subsequent biofilm biomass accumulation upon continued exposure. This prompted us to FIG 1 Effect of manganese ions on P. aeruginosa attachment. Attachment of P. aeruginosa PAO1 grown for 24 h in 96-well plates in LB (A and G), VBMM (C and E), or SDSU artificial sputum medium (ASM) (F). Attachment was assessed using crystal violet (CV) staining. Experiments were conducted in triplicate (two technical replicates each). Statistical significance relative to strain grown in the medium alone was assessed using one-way analysis of variance (ANOVA), followed by a Dunnett's post hoc test. *, P , 0.01; **, P , 0.001. (A) Attachment of P. aeruginosa PAO1 grown in LB in the presence of 0 to 0.2 mM MnCl 2 . (B and D) Growth curve obtained for P. aeruginosa PAO1 supplemented with or without MnCl 2 at the concentrations indicated. Absorbance was determined at 600 nm. The mean and standard deviation (SD) from four measurements are shown. P. aeruginosa was grown in (B) LB or (D) VBMM. (C) Attachment of P. aeruginosa PAO1 grown in VBMM in the presence of 0 to 1 mM MnCl 2 ; (E) attachment by wild-type P. aeruginosa PAO1 grown in VBMM alone or supplemented with 1 mM MnCl 2 , ZnCl 2 , or NiCl 2 ; (F) attachment of P. aeruginosa PAO1 grown in SDSU ASM in the presence of 0 and 0.1 mM MnCl 2 ; (G) attachment of wild-type P. aeruginosa PAO1 and the clinical isolates CF1-13 and PA215 grown in LB in the presence or absence of 0.1 mM MnCl 2 . explore how Mn 21 affects the development of P. aeruginosa biofilms. Daily monitoring over a period of 5 days by bright-field microscopy indicated that P. aeruginosa forms structured biofilms composed of microcolonies within 3 and 5 days in 5-fold-diluted LB and VBMM, respectively (Fig. 2C). In contrast, P. aeruginosa grown in the presence of Mn 21 failed to develop the typical wild-type biofilm architecture (Fig. 2C), indicating that exposure to Mn 21 impairs the formation of structured biofilms. Exposure to Mn 21 affecting the biofilm architecture was confirmed by confocal microscopy, apparent by untreated, 5-fold-diluted LB-grown 3-day-old biofilms featuring a structured biofilm architecture composed of large microcolonies ;100 mm in diameter, while Mn 21 -treated biofilms appeared to be thin and unstructured (Fig. 2D). COMSTAT analysis confirmed biofilms grown in the presence of manganese ions to be composed of 3 to 5 times less biomass and reduced biofilm height relative to the untreated biofilms ( Fig. 2E and F).
The microscopic analysis combined with COMSTAT suggested exposure to Mn 21 not only impairs microcolony formation but also affects the overall biofilm biomass. To The level of biofilm biomass by 5-fold-diluted LB-grown P. aeruginosa PAO1 at different time points was determined by CV staining. Experiments were conducted in triplicate using two technical replicates. Statistical significance was assessed using Student's t test. *, P , 0.01. (B) Level of biofilm biomass of P. aeruginosa PAO1 and the clinical isolates CF1-13 and PA215 grown for 5 days in 5-fold diluted VBMM medium supplemented with or without 1 mM MnCl 2 , as determined using CV staining. Experiments were performed in triplicate using at least two technical replicates. Statistical significance was assessed using Student's t test. **, P , 0.01; *, P , 0.05. (C) Representative bright-field microscopy images of P. aeruginosa PAO1 biofilms grown for 3 or 5 days in either 5-fold-diluted LB or VBMM in the presence or absence of Mn 21 . Bars, 20 mm. (D) Representative confocal microscopy images demonstrating the architecture of P. aeruginosa PAO1 biofilms grown for 3 days in 5-fold-diluted LB in the presence or absence of Mn 21 . Biofilms were stained prior to microscopy using a LIVE/DEAD BacLight viability stain. Bars, 100 mm. The total biofilm biomass (E) and biofilm thickness (F) were determined by COMSTAT analysis. *, statistically different (P , 0.0001) by unpaired, two-tailed t test relative to biofilms grown without Mn 21 . (F and G) P. aeruginosa biofilms were grown for 3 days in tube reactors in 20-fold-diluted LB in the absence or presence of 0.1 mM MnCl 2 . (F) Representative image of biomass accumulation of the biofilms; (G) average number of viable cells present in the biofilms as determined using CFU count. *, significantly different (P , 0.01) by Student's t test relative to biofilms grown in the absence of manganese. Shown are the mean and standard deviation (n = 3 independent measurements).
confirm that exposure to Mn 21 indeed affects the overall biofilm biomass accumulation, biofilms were grown under flowing conditions in tube reactors in the presence or absence of Mn 21 , and the resulting biofilm biomass was harvested after 3 days of biofilm growth. Visible differences in the biofilm biomass were noted (Fig. 2G), with viability counts confirming exposure to Mn 21 coinciding with an overall reduction in the CFU/biofilm relative to untreated biofilms (Fig. 2H).
Manganese affects attachment and biofilm formation in a manner dependent on SagS. The effect of Mn 21 on attachment and biofilm formation was reminiscent of the role of SagS in biofilm development, with the inactivation of sagS coinciding with enhanced attachment but the formation of thin and unstructured biofilms (15). We therefore asked whether SagS is required for Mn 21 to affect both attachment and biofilm formation. We reasoned that if SagS is required, the inactivation of sagS would render the effect of Mn 21 negligible. To address this question, we made use of a DsagS mutant (DsagS::CTX) and an sagS complemented strain (DsagS::CTX-sagS). Similar to P. aeruginosa PAO1, exposure of the DsagS::CTX-sagS strain to Mn 21 significantly enhanced attachment, while the same treatment had little to no effect on DsagS::CTX (Fig. 3A). Likewise, exposure to Mn 21 failed to affect the biomass accumulation or architecture of biofilms formed by DsagS::CTX relative to untreated mutant biofilms, but significantly reduced the biofilm biomass accumulation and architecture of biofilms formed by the complemented sagS mutant strain ( Fig. 3A and B). Our visual observations by confocal microscopy were confirmed by the quantitative analysis of the biofilm biomass by COMSTAT (Fig. 3C). Our findings suggested Mn 21 affects attachment and biofilm formation in a manner dependent on SagS.
Manganese is not a cue for SagS. In order to elucidate whether SagS contributed to Mn 21 -dependent attachment and biofilm formation via sensing Mn 21 , we investigated the role of SagS domains and amino acid residues previously reported to contribute to SagS sensory function (16,36,37) (Fig. 3D). SagS harbors three domains, including an N-terminally located HmsP sensory domain, a centrally located histidine kinase (HisKA) domain and a C-terminal receiver (Rec) domain (Fig. 3D). We generated truncated SagS variants that either (i) comprised only the periplasmic HmsP sensory domain or (ii) lacked the sensory HmsP domain, with both SagS variants retaining the transmembrane helices, ensuring proper localization in the inner membrane (Fig. 3D). In addition, SagS variants harboring amino acid substitutions D105A and L154A were investigated (Fig. 3D). The amino acid residue L154 has been reported to be important in biofilm development (37), while residue D105 has been linked to biofilm antibiotic tolerance but not biofilm development (37). DsagS mutant strains harboring the respective SagS variant constructs were subsequently evaluated for their response to manganese using 96-well plate attachment assays (Fig. 3E). P. aeruginosa PAO1 and the complemented sagS strain were used as positive controls. With the exception of the SagS variant composed of only the periplasmic HmsP sensory domain, variants resulted in the DsagS mutant, demonstrating enhanced attachment in the presence relative to the absence of Mn 21 (Fig. 3E). Our findings suggested Mn 21 was not perceived by the periplasmic sensory domain of SagS. SagS has been known to promote biofilm formation through hierarchical phosphotransfer to the two-component system BfiSR (15). To further explore the role of SagS, we determined whether exposure to Mn 21 affects the phosphorylation status of BfiR. We therefore made use of P. aeruginosa harboring expressing V5-tagged bfiR grown under planktonic and biofilm growth conditions in the absence and presence of Mn 21 and evaluated the phosphorylation status of BfiR by immunoblot analysis using Phos-tag acrylamide. In agreement with previous findings (38), BfiR was found to be more phosphorylated under biofilm than under planktonic growth conditions ( Fig. 3F and G). However, no difference in the BfiR phosphorylation state was noted in the absence or presence of Mn 21 ( Fig. 3F and G), further confirming SagS is not activated by or does not perceive Mn 21 .
Exposure to Mn 2+ affects c-di-GMP levels. The second messenger c-di-GMP has emerged as a key modulator of the lifestyle transition between motility and sessility (22)(23)(24). High intracellular levels of c-di-GMP enhance biofilm formation, whereas its low levels promote a motile lifestyle, leading to biofilm dispersion and returning to a planktonic state (22)(23)(24). To further elucidate the mechanism by which Mn 21 temporally enhances attachment, but impairs the formation of structured biofilms, we next asked whether Mn 21 affects the pool of c-di-GMP. We, therefore, made use of the P cdrA :: gfp(ASV) reporter, for which the fluorescence intensity is directly proportional to the intracellular c-di-GMP concentration (39). As expected, when wild-type cells were grown planktonically to the late stationary phase in the presence of Mn 21 , green fluorescent protein (GFP) fluorescence significantly decreased compared to in its absence, indicating that Mn 21 reduces c-di-GMP levels (Fig. 4A). The findings suggested Mn 21 is capable of affecting c-di-GMP production. as determined using CV staining. Statistical significance relative to strain grown in LB alone was assessed using Student's t test (*, P , 0.01). Shown are the mean and standard deviation (n = 3 independent measurements). (B) Representative confocal microscopy images of the biofilm architecture by DsagS::CTX and DsagS::CTX-sagS cells. Biofilms were grown for 3 days in 24-well plates in 5-fold-diluted LB supplemented with or without 0.1 mM MnCl 2 and stained prior to microscopy using a LIVE/DEAD BacLight viability stain. Bars, 100 mm. (C) Quantification of the total biomass as determined by COMSTAT analysis. *, significantly different (P , 0.05) by Student's t test compared to untreated biofilms. (D) Schematic presentation of the wild-type and a truncated version of SagS constructed to test the attachment capability. The wild-type or truncation mutations of sagS were integrated into the PAO1 DsagS chromosome at the attB site. H 315 and D 713 are SagS conserved phosphorelay sites, and D 105 and L 154 are residues associated with antibiotic tolerance and biofilm formation, respectively. (E) Effect of manganese on attachment. Attachment capabilities of each strain were determined by CV staining assay following 24 h of growth in 96-well plates in LB supplemented with or without 0.1 mM MnCl 2 . *, significantly different (P , 0.001) by Student's t-test from each strain tested in the absence of manganese. Shown are the mean and standard deviation (n = 3 biological replicates). (F) Phos-tag Western blot analysis of cell extracts prepared from wild-type P. aeruginosa harboring a plasmid expressing the BfiR-V5 grown supplemented with or without manganese in either the planktonic or biofilm mode of growth. Biofilms were grown for 3 days in tube reactors, and planktonic cells were grown to exponential phase. Coomassie-stained SDS gels showing total cell extracts after transfer were used as loading controls. Data are representative of three independent experiments, which produced similar results. (G) Determination of the percentage of phosphorylated BfiR over total BfiR, based on the Phos-tag Western blot analysis. Band intensities were analyzed using ImageJ software. Experiments were performed using biological triplicates. ns, not significant by Student's t test.
Mn 2+ stimulate overall PDE activity by P. aeruginosa PAO1. The cellular pool of c-di-GMP is controlled by two classes of enzymes with opposing activities, diguanylate cyclases (DGCs) and phosphodiesterases (PDEs) (22,23), with reduced cellular levels of c-di-GMP likely being due to attenuated DGC activity and/or increased PDE activity (22,23). We first assessed whether exposure of P. aeruginosa cells to Mn 21 has an overall effect on the PDE activity. We therefore prepared total cell extracts of P. aeruginosa PAO1 cells grown planktonically to the late stationary phase in LB alone or supplemented with Mn 21 . The resulting cell extracts were subsequently used to determine the overall PDE activity assays using the synthetic PDE substrate bis(p-nitrophenyl) phosphate (bis-pNPP) (Fig. 4B). Relative to cell extracts obtained from cells grown in LB alone, addition of Mn 21 to the growth medium coincided with a 5.6-fold increase in the PDE activity by P. aeruginosa PAO1, suggesting that exogenous Mn 21 indeed stimulates the PDE activity, with the increase in PDE activity likely leading to reduced c-di-GMP levels ( Fig. 4A and B).
We next sought to identify the Mn 21 -responsive PDE. We reasoned that inactivation of such a PDE would coincide with a reduction in PDE activity in the presence of Mn 21 to levels comparable to those in the absence of Mn 21 . However, if the PDE contributes to reduced c-di-GMP levels in an Mn 21 -independent manner, the respective mutant strain would show similar PDE activity to wild-type cells in the presence of Mn 21 . We focused on several PDEs known to contribute to early biofilm formation stages and/or the overall pool of c-di-GMP in biofilms. These included BifA, RmcA, MorA, RbdA, and DipA. BifA has been reported to contribute to biofilm formation in early stages, by inhibiting the synthesis of c-di-GMP and Pel in P. aeruginosa PA14 (40), while the PDEs RmcA and MorA have been reported to contribute to the maintenance of the mature biofilm structure in response to nutrient limitations (41). Moreover, RbdA and DipA have been shown to contribute to the formation of the mature biofilm architecture, not only by contributing to the overall pool of c-di-GMP but also by promoting dispersion and, thus, the return to a planktonic mode of growth by reducing intracellular c-di-GMP levels (19,21). Therefore, we made use of two P. aeruginosa mutants, the DbifA and DmorA strains, harboring in-frame deletions of bifA and morA, respectively, and mutants harboring transposon insertion elements (ISs) in rmcA, dipA, or rbdA. The latter mutants are referred to as the rmcA::IS, dipA::IS, and rbdA::IS strains. PDE activity assays were performed using total cell extracts obtained from PDE mutants grown planktonically to the late stationary phase in the presence or absence of 0.1 mM MnCl 2 . Under the conditions tested, the PDE activity by P. aeruginosa PAO1 in response to Mn 21 increased ;6-fold (Fig. 4B). Similarly, the PDE activity in cell extracts obtained from DbifA, DmorA, rmcA::IS and dipA::IS mutants increased on average 5.6-fold in response to Mn 21 (Fig. 4B). In contrast, however, a much reduced response to Mn 21 was noted by the rbdA mutant (rbdA::IS) strain (Fig. 4B). While the PDE activity was somewhat increased relative to the rbdA mutant grown in the absence of Mn 21 , the PDE activity increased on average only 2.5-fold (Fig. 4B). The findings strongly suggested the PDE activity of RbdA is enhanced by Mn 21 .
Mn 2+ stimulates the PDE activity of RbdA. PDEs have been reported to use a twometal mode of catalysis, with likely metal ions including Mn 21 , Co 21 , Ni 21 , and Mg 21 (42). To determine whether the PDE activity of RbdA is indeed responsive to Mn 21 , we next purified C-terminally V5/6ÂHis-tagged RbdA from P. aeruginosa wild-type cells (Fig. 5A). Purified C-terminally V5/6Â His-tagged DipA was used as a negative control (Fig. 5A). Purified enzymes were subsequently subjected to PDE activity assays. In the absence of Mn 21 , purified enzymes harbored an average specific activity of 40 to 75 mU/mg (Fig. 5B). No significant increase in the activity of DipA was noted upon addition of 1 mM MnCl 2 (Fig. 5B). In contrast, the PDE activity of RbdA increased ;6-fold to a specific activity of ;460 mU upon addition of 1 mM MnCl 2 (Fig. 5B).
To ensure that the activity of RbdA is indeed enhanced by addition of 1 mM MnCl 2 , we subjected RbdA to chelating conditions to remove divalent ions, followed by addback of specific metal ions. To do so, purified RbdA (purified from Escherichia coli [ Fig.  5C]) was exposed for 10 min to EDTA (0.1 and 0.5 mM) to remove divalent ions. Following chelation, the enzymatic activity of RbdA in the presence of EDTA in the absence or presence of 2 mM MnCl 2 was evaluated (Fig. 5D). Untreated RbdA in the absence of MnCl 2 served as a positive control (Fig. 5D). Under the conditions tested, untreated RbdA in the absence of MnCl 2 had a specific PDE activity of 87 mU/mg, which increased 6.3-fold to 550 mU/mg in the presence of MnCl 2 (Fig. 5D). Pretreatment of RbdA for 10 min with 0.1 and 0.5 mM EDTA resulted in a significant, 10-fold reduction in the specific activity relative to untreated RbdA (Fig. 5D). Addition of 2 mM MnCl 2 almost fully restored the PDE activity of EDTA-treated RbdA to that noted for untreated RbdA in the presence of Mn 21 (Fig. 5D). It is of interest to note that the addition of 2 mM MgCl 2 failed to restore the RbdA activity posttreatment with EDTA (Fig. 5D). Our findings strongly suggested that not only is the PDE RbdA responsible for the overall increase in PDE activity upon exposure to Mn 21 , but the activity of RbdA is also significantly enhanced by Mn 21 . Our findings are in agreement with previous reports of Mn 21 enhancing PDE activity (42). However, while Mn 21 and Mg 21 can frequently be exchanged without loss of enzyme activity, our findings clearly indicate that Mn 21 cannot be substituted by Mg 21 to activate RbdA.
RbdA is composed of an N-terminal PAS-PAC domain, a GGDEF (possessing DGC activity) domain, and a C-terminal EAL (possessing PDE activity) domain (43). The PDE activity of RbdA has been reported to be stimulated when GTP binds to its GGDEF domain (43,44). We therefore asked whether GTP would affect the Mn 21 -stimulated PDE activity of RbdA. To explore the GTP effect, we examined the PDE activity of purified RbdA (purified by E. coli [ Fig. 5C]) by adding an excess amount of GTP in the absence and presence of Mn 21 . Addition of GTP, however, did not significantly enhance the PDE activity of RbdA, regardless of the absence and presence of Mn 21 (Fig. 5E).
RbdA contributes to the Mn 2+ dependency of biofilm formation. Our findings indicated RbdA to be responsive to Mn 21 . We therefore reasoned that if manganese ions indeed affect biofilm formation by stimulating the PDE activity of RbdA, and thus, the overall level of c-di-GMP, inactivation of rbdA would eliminate or at least significantly diminish the effect of Mn 21 on biofilm formation. To examine the role of RbdA in the biofilm formation in response to Mn 21 , we grew biofilms formed by the wildtype and rbdA-deficient mutant strains in 24-well plates for 3 days in the presence or absence of Mn 21 and determined the level of biofilm biomass accumulations by CV staining. While the biofilm biomass by the P. aeruginosa wild-type strain was significantly reduced in the presence of Mn 21 relative to its absence, no significant difference in the CV-stainable biomass was noted in biofilms formed by the rbdA::IS mutant strain (Fig. 6A). The findings indicated that the overall reduction in the biofilm biomass in response to Mn 21 is due to RbdA and its responsiveness to Mn 21 .
RbdA contributes to biofilm dispersion in response to Mn 2+ . RbdA has previously been reported to be contribute to biofilm dispersions in response to exogenous dispersion cues such as glutamate and nitric oxide (19,21). Given that Mn 21 contributes to the loss of biofilm biomass and that RbdA is responsive to Mn 21 , we next asked if exposure to Mn 21 induces biofilm dispersion in a manner dependent on RbdA. To address this question, we performed a dispersion assay using MnCl 2 as a cue to induce dispersion. Biofilms formed by the P. aeruginosa wild-type or rbdA::IS mutant strain were grown under flow conditions in a tube reactor for 5 days and subsequently exposed to a sudden change of growth medium supplemented with 0.5 mM MnCl 2 . Biofilm effluents were collected after exposure to MnCl 2 -containing medium, and the absorbance of the effluents was subsequently determined at 600 nm. Dispersion events have previously been reported to be apparent by a sharp increase in the absorbance in the effluent upon induction of dispersion compared to untreated biofilms (20,21,45,46).
Exposure of wild-type biofilms to growth medium supplemented with 0.5 mM MnCl 2 resulted in a sharp increase in the absorbance of the biofilm effluents, indicative of dispersion events, a response that was absent in biofilms that were left untreated and exposed to growth medium alone ( Fig. 6B and C). However, biofilms formed by the rbdA::IS mutant strain failed to disperse in response to Mn 21 ( Fig. 6B and C). It is of interest to note that biofilms formed by rbdA::IS mutant have been reported to be comparable to wild type in architecture but demonstrating enhanced biofilm biomass accumulation (43). The finding strongly suggested Mn 21 to be a likely cue capable of inducing biofilm dispersion, with Mn 21 -induced dispersion to be the likely cause for the significant loss of biomass of biofilms grown in the presence of Mn 21 ( Fig. 2 and Fig. 6A). Moreover, our data indicated an RbdA dependency of both dispersion and loss of biofilm biomass in response to Mn 21 (Fig. 6).
Mn 2+ affects expression of pel and psl genes and abundance of Pel and Psl polysaccharides. We next explored the mechanism by which RbdA affects biofilm biomass in a Mn 21 -dependent manner. We took into account the finding that the second messenger c-di-GMP has been reported to stimulate Psl and Pel production and to be elevated by Psl exopolysaccharide, establishing a positive-feedback loop (22,24). Moreover, previous reports indicated that Pel and Psl exopolysaccharides are degraded upon dispersion (46) and that matrix exopolysaccharides, including alginate, Pel, and Psl, affect attachment and microcolony formation (47)(48)(49). Specifically, lack of Psl has been reported to coincide with significantly reduced initial attachment (47). In addition, while both Psl and alginate are required for the formation of characteristic mushroom-like structures (47)(48)(49), Pel appeared to play a role in biofilm cell density and/or the compactness of the biofilm (47)(48)(49). In contrast, mutants deficient in both Psl and Pel production (but capable of producing alginate) were severely impaired in their ability to form biofilms (48). As our findings indicated exposure to Mn 21 to coincide with reduced intracellular c-di-GMP levels via stimulation of RbdA activity (Fig. 4 and Fig. 5) and the formation of thin biofilms lacking microcolonies or cellular aggregates (Fig. 2), we asked whether Mn 21 affects the expression of genes involved in the biosynthesis (and degradation) of the polysaccharides Pel and Psl. We therefore determined the transcript abundance of pelA and pslG, parts of pel and psl operons, respectively, by quantitative reverse transcription-PCR (qRT-PCR). PelA, encoded by pelA, is a periplasmic modification enzyme with an N-terminal glycoside hydrolase domain and a C-terminal deacetylase domain (50), whereas PslG, encoded by pslG, has been characterized to harbor Psl glycoside hydrolase activity (51). Under planktonic growth conditions, the transcript abundance of pelA and pslG was significantly reduced upon exposure to Mn 21 (Fig. 7A). Likewise, biofilms exposed to Mn 21 for 24 h demonstrated significantly decreased transcript abundance of pslG, but not pelA, relative to untreated biofilms (Fig. 7A). It is noteworthy that, under the conditions tested, the expression level of brlR, encoding the biofilm resistance locus regulator, showed no significant differences regardless of the absence or presence of Mn 21 (Fig. 7A). To determine whether differential expression of pelA and pslG transcript in response to Mn 21 correlated with differences in the abundance of Pel and Psl polysaccharides, the respective polysaccharides were purified from P. aeruginosa PAO1 grown planktonically to the exponential phase in the presence or absence of Mn 21 . The Pel and Psl polysaccharides were subsequently detected by immunoblot analysis using anti-Pel and anti-Psl antibodies, respectively (Fig. 7B). In agreement with our qRT-PCR analysis, the quantitative analysis of Pel and Psl abundance indicated exposure to Mn 21 coincides with the reduced abundance of the Psl and Pel polysaccharides (Fig. 7B). Psl abundance is dependent on RbdA. RbdA has been known to negatively regulate exopolysaccharide production (43). Considering that our findings so far suggested Mn 21 affects the expression of pel and psl as well as the abundance of the Pel and Psl exopolysaccharides, we next asked whether Mn 21 -mediated inhibition of Pel and Psl requires RbdA.
Relative to P. aeruginosa PAO1, insertional inactivation of rbdA had no effect on Pel abundance (Fig. 7B). In contrast, insertional inactivation of dipA correlated with an up to 1.5-fold increase in Pel abundance relative to the wild type (Fig. 7B). Exposure of the rbdA::IS and dipA::IS strains to Mn 21 resulted in a reduction of Pel abundance in a manner comparable to the overall reduction in Pel abundance by wild-type cells (Fig. 7B). Insertional inactivation of dipA and rbdA affected the overall abundance of Psl, with the abundance of Psl being increased 2-and 2.5-fold in dipA-and rbdA-deficient mutant strains, respectively, relative to the wild type (Fig. 7B). Similar to Pel, the abundance of Psl decreased in the dipA-deficient mutant strains in response to Mn 21 in a manner similar to that of the wild-type strain (Fig. 7B). In contrast, however, no difference in Psl abundance was noted in the rbdA-deficient mutant strain regardless of the absence or presence of Mn 21 (Fig. 7B), suggesting RbdA contributes to the abundance of the Psl response to Mn 21 . Taken together, our data demonstrate that Mn 21 acts through the PDE RbdA to impede not only Psl production but also the formation of structured biofilms.

DISCUSSION
Several studies have explored the effects of Mn 21 on biofilm formation. A combination of Mn 21 and glycerol has been found to promote biofilm formation of Bacillus species such as B. subtilis, B. licheniformis, and B. cereus via the histidine kinase KinD (52). Hussain et al. (53), demonstrated that the expression levels of genes involved in motility and EPS formation were positively correlated with B. cereus biofilm formation in brain heart infusion (BHI) medium supplemented with Mn 21 and heme. The same medium induced B. cereus biofilm formation (53). However, Guo et al. (54) showed that Mn 21 inhibits E. coli biofilm formation in a concentration-dependent manner through the regulation of phenotypic morphology and metabolic reprogramming (54). It has also been shown that exposure of P. putida MnB1 to Mn 21 for 12 h coincided with an upregulation of bifA, accompanied by biofilm suppression (55). While the studies indicated Mn 21 to be a potential cue to modulate biofilm formation, no mechanism has been elucidated. The goal of this study was to determine whether Mn 21 affects biofilm formation by P. aeruginosa and, if so, elucidate the mechanism by which Mn 21 modulates P. aeruginosa biofilm formation. Here, we showed that Mn 21 initially enhances attachment ( Fig. 1) but subsequently inhibits P. aeruginosa biofilm formation (Fig. 2) by reducing intracellular c-di-GMP levels (Fig. 4A) and Psl polysaccharide abundance (Fig.  7). Our findings of Mn 21 reciprocally affecting attachment and biofilm formation (as well as inducing dispersion) provide an explanation for the inconsistency of the role of Mn 21 in the literature. Our mechanistic study revealed Mn 21 alters P. aeruginosa biofilm development via modulation of the second messenger c-di-GMP. Specifically, we demonstrate that Mn 21 acts through the PDE RbdA to reduce c-di-GMP levels ( Fig. 4  and 5), thereby reducing Psl abundance and biofilm formation ( Fig. 6 and 7).
Our findings are furthermore in agreement with reports by An et al. (43) characterizing RbdA as an active PDE capable of affecting the production of exopolymeric substances, determined by Congo red staining and pellicle formation assays. However, despite providing several lines of evidence demonstrating a link between EPS production and RbdA activity, An et al. (43) did not provide direct evidence of EPS abundance being affected by RbdA activity nor specified the EPS (Pel, Psl, or alginate). In this study, we show that RbdA contributes to the abundance of the exopolysaccharide Psl, not Pel (Fig. 7). To our knowledge, this is the first report of Psl production being dependent on RbdA.
Enzymatic activities, especially PDE activities, being stimulated by metal ions are not unusual. For example, the EAL domain from the c-di-GMP specific-PDE HmsP of Yersinia pestis exhibited high PDE activity in the presence of 1 mM manganese relative to the presence of other divalent metals such as Zn and Mg, indicating that the PDE activity depends on Mn 21 (56). Similarly, the metallophosphoesterase PlcP by Phaeobacter sp. strain MED193 and Pelagibacter sp. strain HTCC7211 is essential for lipid remodeling, with its phosphomonoesterase and phosphodiesterase activity being dependent on manganese (57). The pGpG-specific PDE PggH of Vibrio cholerae showed the highest PDE-B activity in the presence of Mn 21 even though the enzyme also showed activity in the presence of divalent cations such as Mg 21 , Ca 21 , and Zn 21 (58). In contrast, the divalent cations Mg 21 and Mn 21 were necessary for V. cholerae VieA activity, while Ca 21 and Zn 21 inhibited the PDE activity of VieA (59). Other c-di-GMP-modulating enzymes are magnesium dependent. For instance, the c-di-GMP-specific PDE activities of CC3399, a GGDEF-EAL protein from Caulobacter crescentus, and RocR, an EAL domain protein from P. aeruginosa PAO1, require Mg 21 (60,61). Our findings of EDTA-treated RbdA being restored in its activity by Mn 21 but not Mg 21 (Fig. 5D) indicate that RbdA is a manganese-inducible PDE. However, the mechanism by which manganese affects the PDE activity of RbdA remains to be elucidated. Liu et al. (44) suggested that a signal detected by the N-terminal putative periplasmic sensor domain of RbdA or GTP binding to the GGDEF domain may lead to a conformational change and, thus, RbdA dimerization and PDE activity activation. It is thus likely that manganese may be such a signal, either inducing conformational change of RbdA to activate its PDE function or modulating the PDE activity by binding to one of its domains (e.g., GGDEF).
Our findings clearly indicated that exposure of mature biofilms formed by P. aeruginosa to Mn 21 coincided with reduced biofilm biomass accumulation, reduced microcolony formation, and induction of dispersion via the modulation of c-di-GMP and that the effect by Mn 21 is dependent on RbdA, leading to the modulation in Psl abundance. As the Psl polysaccharide has been reported to protect P. aeruginosa from host defenses in an acute murine pulmonary model of infection (62), our findings underscore the possibility that Mn 21 may be able to modulate the P. aeruginosa virulence repertoire via Psl. Therefore, our study demonstrates that Mn 21 acts as an environmental inhibitor of P. aeruginosa biofilm development, suggesting Mn 21 to be a promising new antibiofilm factor.

MATERIALS AND METHODS
Bacterial strains, plasmids, and culture conditions. The bacterial strains, plasmids, and oligonucleotides used in this study are listed in Table 1 and Table 2. The complex medium, Lennox Broth (LB), was used as a growth medium for P. aeruginosa and E. coli. The minimal medium, Vogel and Bonner citrate minimal medium (VBMM) (63), and the artificial sputum medium (ASM), SDSU (64,65), were used to grow P. aeruginosa PAO1 and the clinical isolates. All planktonic cells were grown at 37°C under shaking conditions (220 rpm). Biofilms were grown as indicated below. The following antibiotics were added for plasmid maintenance: for P. aeruginosa, 60 mg mL 21 tetracycline, 75 mg mL 21 gentamicin, and 250 mg mL 21 carbenicillin; for E. coli, 15 mg mL 21 tetracycline, 20 mg mL 21 gentamicin, and 100 mg mL 21 ampicillin. Additionally, 1 mM isopropyl-b-D-1-thiogalactopyranoside (IPTG) or 0.1% L-arabinose was added to the growth medium where indicated.
Biofilm growth. P. aeruginosa biofilms were grown in 20-fold-diluted LB or 5-fold-diluted VBMM using a continuous-flow tube reactor system with a size of 13 for biofilm biomass accumulation for 3 days and 14 Masterflex silicone tubing for dispersion assays for 5 days (Cole Parmer, Inc.) at flow rates of 0.1 and 0.2 mL min 21 , respectively (11). Tube reactors were inoculated with 1 mL of P. aeruginosa overnight cultures. For a 24-well plate biofilm cultivation, biofilms were grown in 5-fold-diluted LB or VBMM, as previously described with some modifications (66). Briefly, overnight cultures of P. aeruginosa were adjusted to an optical density at 600 nm (OD 600 ) of 0.1, and 10 mL of the OD-adjusted suspension was transferred into wells of 24-well flat-bottom polystyrene microplates (BD Falcon), with each well containing 250 mL medium. (The wells can hold a total volume of 1 mL but only 250 mL medium was added to prevent spills and cross-contamination.) The plate was incubated at 37°C and 220 rpm at a 45°a ngle. The 45°angle promotes biofilm formation by increasing the air-liquid interface. The medium was exchanged every 12 h. Adherent cells were quantitated using crystal violet (CV) staining. Briefly, 300 mL 0.1% CV solution was directly added to each well, followed by incubation for 15 min at 37°C with shaking. The plate was washed three times with 500 mL of water and then allowed to dry prior to the addition of 500 mL of 95% ethanol to each well. After incubation for 15 min at 37°C with shaking, the OD 570 was determined. Quantitative analysis of the confocal laser scanning microscopy (CSLM) images of 24well plate-grown biofilms, acquired by a Leica TCS SP5 (Leica Microsystems, Inc.), was performed using COMSTAT (5,15). For confocal image acquisition, biofilms were stained with the LIVE/DEAD BacLight bacterial viability kit (Life Technologies). For plasmid maintenance, the growth medium was supplemented with carbenicillin (10 mg mL 21 ) or gentamicin (2 mg mL 21 ).
Attachment assays. Overnight-grown P. aeruginosa cells were washed using 0.85% saline and inoculated into LB, VBMM, or SDSU medium supplemented with or without MnCl 2 , ZnCl 2 , or NiCl 2 at a final OD 600 of 0.05. Then, 100 mL of the suspension was transferred into a 96-well plate and incubated with shaking at 220 rpm at 37°C for 24 h. The wells were subsequently rinsed with 0.85% saline to remove planktonic cells, and the remaining adherent cells were stained with 0.1% CV solution for 15 min. After being washed with 200 mL of water two times, the remaining crystal violet was solubilized with 200 mL of 95% ethanol and the absorbance measured at 570 nm.
Biofilm dispersion assays. Biofilm dispersion assays were performed using biofilms grown in continuous-flow tube reactors as previously described (45,46). In brief, the dispersion of 5-day-old biofilms was induced by the sudden addition of 0.5 mM MnCl 2 to the growth medium. Then, the biofilm effluents were collected into 96-well plates at 1-min intervals and the optical density if the collected effluent measured at 600 nm. An increase in the turbidity of the effluent indicated dispersion events. For the quantitative analysis of the dispersion response, the absorbance of effluents collected 20 to 23 min after exposure to 0.5 mM MnCl 2 and that of untreated biofilms were evaluated.
Determination of the phosphorylation state of BfiR. Planktonic and 3-day old biofilm cells were cultured as described above and normalized by measuring the OD 600 . The cell pellets were resuspended and 100 mM MnCl 2 as described by the manufacturer. Electrophoresis was performed at 4°C and 120 V for 3 h. Before immunoblotting, the Phos-tag gel was incubated for 10 min in transfer buffer containing 10 mM EDTA and for another 10 min in transfer buffer twice to remove EDTA. Subsequent immunoblotting was performed using anti-V5-horseradish peroxidase (HRP) antibody (1:5,000) (Invitrogen). The blots were developed with Immun-Star WesternC chemiluminescent reagents (Bio-Rad). The densities of protein bands were determined by quantification using ImageJ software version 1.48 (NIH). Quantification of c-di-GMP. Quantification of relative-c-di-GMP levels was performed using cells harboring plasmids expressing GFP(ASV) from the c-di-GMP-responsive cdrA promoter [pCdrA::gfp(ASV)] and expressing mCherry from a constitutive promoter as previously described, with some modifications (5,39). P. aeruginosa cells biofilms grown for 3 days in LB supplemented with or without 0.1 mM MnCl 2 were harvested, resuspended into 200 mL phosphate-buffered saline (PBS), and the suspension was transferred into a 96-well black clear-bottom microtiter plate (Greiner Bio-One). The fluorescent emission of GFP (excitation 485 nm/emission 535 nm) and mCherry (580 nm/620 nm) was measured every 30 min using a SpectraMax i3x plate reader (Molecular Devices). Fluorescence units from GFP were normalized to mCherry.
Phosphodiesterase activity assay. Phosphodiesterase activity was determined using the synthetic substrate bis(p-nitrophenyl) phosphate (bis-pNPP) (Sigma-Aldrich), as previously described (21,40). Briefly, either 100 mg of crude extract or 18 mg of purified protein was incubated with 5 mM bis-pNPP in reaction buffer (50 mM Tris-HCl [pH 8.0], 50 mM NaCl, and 5 mM MgCl 2 ) at 25°C for 4 h. The release of pnitrophenol (pNP) was quantified at OD 405 every 30 min. Crude extracts were obtained from cells grown planktonically in the absence of presence of MnCl 2. Controls without extracts or purified protein were included to account for any nonenzymatic bis-pNPP hydrolysis.
Purification of His-tagged proteins. For protein purification, cells were grown planktonically to the exponential phase. Proteins harboring C-terminal V5/His tags cloned into pET vectors (E. coli) and pMJT-1 vectors (P. aeruginosa) were overproduced using 1 mM IPTG and 0.1% L-arabinose, respectively, and purified using Ni-nitrilotriacetic acid (NTA) metal-affinity resin (Thermo Fisher Scientific) according to the manufacturer's instruction. Briefly, harvested cells were resuspended in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride), lysed by sonication, and centrifuged at 10,000 Â g at 4°C for 10 min to remove cell debris. After removal of cell debris from the lysate, membrane-bound proteins were solubilized by 1% (wt/vol) n-dodecyl-b-D-maltoside (DDM) and loaded onto Ni-NTA metal-affinity resin. Proteins bound to the resin were eluted with 200 mM imidazole and then desalted using VivaSpin centrifugal concentrator columns and a Tris buffer (50 mM Tris-HCl [pH 7.5], 100 mM NaCl, 5% glycerol) containing 0.05% DDM.
RNA isolation and qRT-PCR. Total RNA was isolated from cells grown planktonically and as a 3-day biofilm using a E.Z.N.A. Total RNA kit (Omega Bio-Tek), and DNA was removed using Turbo DNase (Thermo Fisher Scientific) as previously described (5). The same amount of RNA (1 mg) from each cell was converted to cDNA using the iScript Select cDNA synthesis kit (Bio-Rad). qRT-PCR was carried out using the Bio-Rad CFX Connect real-time PCR detection system and SsoAdvanced SYBR green supermix (Bio-Rad). Specific primers for amplification of the cDNA are listed in Table 2. To normalize the transcript level, the cysD gene was used as a reference.
Pel and Psl polysaccharide dot blot analysis. Crude polysaccharide extracts were obtained by resuspending 10 OD 600 equivalents of cells from each growth condition in 100 mL of 0.5 M EDTA and  (67). The supernatant was treated with proteinase K at 60°C for 1 h (final concentration 0.5 mg mL 21 ), followed by proteinase K inactivation at 80°C for 30 min. For immunoblotting, 5 mL of polysaccharide extract was spotted on a nitrocellulose membrane and probed as described previously (21,67). The Pel and Psl productions were quantified using ImageJ software version 1.48 (NIH).