Anti-Biofilm Activity: A Function of Klebsiella pneumoniae Capsular Polysaccharide

Competition and cooperation phenomena occur within highly interactive biofilm communities and several non-biocides molecules produced by microorganisms have been described as impairing biofilm formation. In this study, we investigated the anti-biofilm capacities of an ubiquitous and biofilm producing bacterium, Klebsiella pneumoniae. Cell-free supernatant from K. pneumoniae planktonic cultures showed anti-biofilm effects on most Gram positive bacteria tested but also encompassed some Gram negative bacilli. The anti-biofilm non-bactericidal activity was further investigated on Staphylococcus epidermidis, by determining the biofilm biomass, microscopic observations and agglutination measurement through a magnetic bead-mediated agglutination test. Cell-free extracts from K. pneumoniae biofilm (supernatant and acellular matrix) also showed an influence, although to a lesser extend. Chemical analyses indicated that the active molecule was a high molecular weight polysaccharide composed of five monosaccharides: galactose, glucose, rhamnose, glucuronic acid and glucosamine and the main following sugar linkage residues [→2)-α-l-Rhap-(1→]; [→4)-α-l-Rhap-(1→]; [α-d-Galp-(1→]; [→2,3)-α-d-Galp-(1→]; [→3)-β-d-Galp-(1→] and, [→4)-β-d-GlcAp-(1→]. Characterization of this molecule indicated that this component was more likely capsular polysaccharide (CPS) and precoating of abiotic surfaces with CPS extracts from different serotypes impaired the bacteria-surface interactions. Thus the CPS of Klebsiella would exhibit a pleiotropic activity during biofilm formation, both stimulating the initial adhesion and maturation steps as previously described, but also repelling potential competitors.


Introduction
Biofilms are complex assemblages of microbial cells enclosed in a self synthesized polymeric matrix [1]. They are considered as the prevalent microbial lifestyle in nature and can form biofilm on a variety of surfaces such as metals, plastics, mineral surfaces and living tissue in human host [2][3][4]. In contrast to planktonic cells, sessile cells are subjected to intense interactions due to their concentration and proximity, consisting of either cooperative or competitive phenomena [5]. These interactions can influence the emergence or the disappearance of some species within the communities and thus play important roles in the development, composition and function of the microbial consortia [6,7]. Since biofilm formation is often considered as a major problem due to the ability of sessile bacteria to better tolerate exogenous stress than planktonic bacteria and therefore to persist, most studies have focused on antagonisms. Several bacterial non biocide biofilminhibiting molecules have been described so far; they impair either the initial adhesion step of the biofilm formation, its development and maturation, or the late detachment step [8]. A few anti-biofilm molecules have been isolated from monospecies biofilms, but most of them were discovered using mixed species biofilm experimental settings [9][10][11].
Indeed, mixed biofilms represent the ideal environment for discovering natural molecules that potentially influence the dynamics of bacterial populations [12]. Impairing bacterial communication systems such as the quorum sensing can affect bacterial ability to form biofilm, as in the case of Bacillus cereus production of an AHL lactonase inhibiting Vibrio cholerae biofilm settlement [13]. Cugini et al. (2007) also reported that C. albicans produces farnesol, a quorum sensing molecule, that inhibits the swarming mobility of P. aeruginosa, thereby enhancing its ability to form biofilm [14,15]. Some other bacteria act differently by producing extracellular enzymes that modify the aggregates, such as the Esp protease produced by Staphylococcus epidermidis that degrades specific proteins in the biofilm matrix and cell wall fractions of S. aureus [16]. Biofilm can also be affected by the degradation of the matrix polysaccharide components. Thus Kaplan et al. (2003) reported that Actinobacillus actinomycetemcomitans produces dispersin B that degrades poly-N-acetylglucosamine (PNAG), a major polysaccharide component of many bacterial extracellular matrices [17].
One of the most commonly shared strategies to avoid biofilm colonization by competitors, however, consists of secretion of polysaccharides with anti-biofilm non biocide effects against other species [18]. These exopolysaccharides likely act as biosurfactants, modifying the physicochemical properties of surfaces and altering bacterial interactions within mixed biofilms [9,[19][20][21][22]. Thus, Valle et al. (2006) showed that group 2 capsules of pathogenic Escherichia coli reduce bacterial adhesion of commensal E. coli strain by inhibiting cell surface and cell-to-cell interactions in biofilm development [9]. A recent study by Rendueles et al. (2011) showed that commensal and pathogenic strains of E. coli secrete high molecular weight polysaccharides, different from capsular components, with anti-biofilm activity only against Gram-positive bacteria and able to exclude Staphylococcus aureus from mixed biofilms [19]. In constrast, exopolysaccharides isolated from marine Vibrio were shown to inhibit initial adhesion of both Gram-positive and Gram-negative by impairing both the cell to surface adherence and the bacterial intercellular adhesion [21].
In this study, we investigated the capacity of synthesis of antibiofilm components by Klebsiella pneumoniae, a bacterium producing biofilms and heavily surrounded by exopolysaccharides. This work was conducted using cell-free extracts from K. pneumoniae and led to the isolation and characterization of a polysaccharide displaying anti-adhesion properties towards several bacterial species.

Bacteria and Growth Conditions
The bacterial strains used in this study are listed in table 1. All bacterial strains were stored at 280uC in Trypticase Soy (TS) broth containing 15% glycerol (v/v). They were subcultured from freezer stocks onto TS agar plates (TSA, Fisher Scientific) or in M63B1 broth supplemented with glucose (0.4% (w/v)). All subsequent liquid precultures (overnight cultures) were performed at 37uC in TS broth (TSB, Oxoid, Basingstoke, England) at 200 rpm in an orbital shaking and were derived from colonies isolated on TSA plates. For planktonic growth cultures, S. epidermidis RP62A and K. pneumoniae MGH 78578 were cultured in TSB at 37uC in aerobic conditions. Results of S. epidermidis numeration were expressed as the number of CFU.mL 21 of suspension. For quantification of capsular polysaccharides, K. pneumoniae strains were cultivated overnight in D.W. medium supplemented with 0.1% Casamino Acids, 200 mg of MgCl 2 per ml, 20 mg of CaCl 2 per ml, 1 mg of ZnCl 2 per ml, and 4 mg of FeCl 3 per ml [23].

K. pneumoniae Cell-free Extracts
Cell free supernatants from K. pneumoniae planktonic bacteria were obtained from 24 h-old cultures centrifuged for 10 min at 7,0006g at 4uC and the supernatants were filtered through a 0.2 mm filter (CA-membrane, Sartorius Stedim, North America).
Biofilm cell free extracts, i.e. biofilm supernatant and acellular matrix material, were prepared from biofilms grown in tissue culture dish (Falcon, Becton Dickinson, Franklin Lakes, USA) using an inoculum of 1610 6 CFU.mL 21 and incubated for 24 h at 37uC. The biofilm supernatants were recovered as described above for the planktonic supernatant. For acellular matrix material, the biofilms were then lifted by scraping, sonicated in a waterbath sonicator for 5 min in order to dissociate bacterial aggregates and thus free the extracellular material, vortexed and again sonicated for 5 min. To harvest the resulting 2 mL biofilm acellular material, the suspensions were centrifuged (7,0006g for 10 min) and filtered throughout a 0.2 mm filter.

Effect of K. pneumoniae Cell-free Extracts on Biofilm Formation
The effects of the three types of K. pneumoniae MGH 78578 extracts (planktonic supernatant, biofilm supernatant and acellular matrix biofilm) were assessed on biofilm formation using the three following procedures. To determine the spectrum of activity of the K. pneumoniae anti-biofilm component, biofilm formation assay was carried out in 96-well polystyrene microtiter plates (Falcon) and measured using the crystal violet staining procedure. Briefly, overnight cultures of each strain to be tested were diluted to about 1610 6 CFU.mL 21 with fresh TSB. Each well of the microtiter plates was filled with inoculum of the bacterial suspensions containing J (v/v) of K. pneumoniae MGH 78578 supernatant extract (concentration 2.654 mg equivalent galactose per milliliters). Control assays were performed by forming biofilm without any supernatant extract. The microtiter plates were incubated at 37uC for 8 h without shaking, and non-adherent bacteria removed by three washings with saline solution (9% NaCl (w/v)). Biofilm was stained by 0.5% (w/v) crystal violet solution for 10 min. Then, the plates were rinsed under running tap water, air-dried, and the crystal violet was resuspended in ethanol, and the OD 570 was determined. The composition of the biofilm matrix of the Staphylococcus strains tested (polysaccharide or protein) was assessed essentially as described in Wang et al. (2004), by treatment with proteinase K or sodium meta-periodate [24].
Biofilms were also measured using a magnetic beads-mediated agglutination assay measuring the immobilization of magnetic beads embedded in bacterial aggregates following biofilm formation (Ring testH, Biofilm Control, Saint-Beauzire, France). Consequently, the more beads are entrapped by cells, the less they are detectable after magnetization [25]. Three mm diameter magnetic particles (toner) were added to bacterial suspensions obtained after serial dilutions in TSB of an overnight culture or to TSB alone (control solution) at a final concentration of 10 mL.mL 21 . Each well was then inoculated with 200 mL of the mixture containing either S. epidermidis RP62A at concentration of 1610 6 CFU.mL 21 with or without (control) J (v/v) of either K. pneumoniae planktonic supernatant, biofilm supernatant or acellular matrix material. Three experiments with three repeats each were performed. After incubation at 37uC, 100 mL of contrast liquid (inert oil) was added to each well and the strip wells were immediately scanned and placed for 1 min on the block test for magnetization and re-scanned. This test is based on the concept of immobilization of beads by sessile bacteria that form aggregates with enough strength to overcome the magnetic attraction forces applied on them. Thus, in the absence of sessile cells all the beads gathered in the center of the wells and formed an easily detectable black spot. Images of each well before (I 0 ) and after (I 1 ) magnetization were compared with the Biofilm ControlH Software and the discrepancies between the two images gave rise to values named BioFilm Index (BFI) ranging from 0 to 30. High BFI values correspond to high mobility of beads under magnet action and therefore little or no immobilization of the beads by bacteria [25]. With TSB as incubation medium, a biofilm was considered installed when the value of BFI was less than or equal to 4.5, which corresponds to 10% untrapped beads.
For CFU determination, biofilms were formed in duplicate on ThermanoxH slides (Nalgène) incubated at 37uC in 24-well plates (Falcon, Becton Dickinson, Franklin Lakes, USA) in TSB for 8 hours. Coverslips were then gently removed and washed twice with sterile physiological water (9% NaCl (w/v)). Adhering bacteria were resuspended in 2 mL of physiological water and then sonicated three times for 5 minutes (S-LINE FisherBrand, 37 kHz) and vortexed between each sonication. The resulting biofilm suspensions were then serial diluted and plated onto selective agar media to determine the numbers of CFU.

Electron Microscopic Observation
For scanning electronic microscopy observation, biofilms were formed on ThermanoxH slides as described above and fixed overnight at 4uC with a solution of 0.2 M cacodylate buffered at pH 7.4 supplemented with glutaraldehyde at 1.6% (w/v) and ruthenium red at 0.05% (w/v). They were then rinsed in the same buffer. After post-fixation for 1 h with osmic acid in cacodylate buffer, biofilms were dehydrated using a graded ethanol series (10 min each in 25%, 50%, 75%, 100% absolute alcohol), and 100-150 mL of hexamethyldisilazine were added to each insert and allowed to evaporate (,2 h) in a fume hood. Supports were attached to 12 mm diameter aluminium SEM stubs using adhesive carbon tabs and they were gold coated using a sputter coater. The specimens were examined by a JSM-6060LV scanning electron microscopy (JEOL, Croissy-sur-Seine, France).

Bacterial Hydrophobicity Measurement
The Microbial Adhesion To Solvents (MATS) test was performed to evaluate the Lewis acid-base properties and the hydrophilic/hydrophobic nature of bacterial surfaces. This test was adapted from the method of Rosenberg et al. (1980) which is based on the adhesion of bacterial cells to four different solvents: chloroforme, hexadecane, decane and ethyl acetate [26]. 10 mL of bacteria grown overnight in LB were harvested by centrifugation (5,0006g, 8 min), washed twice with sterile PBS and resuspended in 10 mL of PBS. The OD 600nm of the suspension was measured and adjusted at 0.4 (OD 600nm initial). One mL of solvent was mixed to 3 mL of the cell suspension and vortexed for 1 min. The two phases were allowed to separate for 20 min. One mL of the aqueous phase was carefully removed and the OD 600nm was measured (OD 600 final). The microbial adhesion to each solvent was calculated using the formula: % affinity = 1006(12(OD 600 final/OD 600 initial). Each experiment was performed in triplicate.

Physical and Chemical Analyses of K. pneumoniae Planktonic Extract
For enzymatic treatments, planktonic extract was incubated for 1 h at 37uC with 100 mg.L 21 DNase I, RNase A, porcine pancreatic lipase or proteinase K (Sigma-aldrich). Controls consisted of mock-treated extracts, or enzyme alone without any extract. For sodium metaperiodate treatment, 0.1 vol of 100 mM sodium metaperiodate was added to the extract, and incubated at 37uC for 1 h. Controls consisted of mock-treated extract and sodium metaperiodate alone. Following all treatments, extracts and controls were incubated at 100uC for 10 min prior to testing in the S. epidermidis biofilm assay as described above. Temperature influence itself was assessed by heating the supernatant extract for 10 min at 100uC of without any other treatment.
Extraction, Purification and Chemical Characterization of the anti-biofilm Compound (1) Production and extraction of anti-biofilm compound. K. pneumoniae MGH 78578 was incubated for 6 h at 37uC in TSB. The culture was then centrifuged at 7,0006g for 10 min and the pellet was washed twice with M63B1 minimum medium. The bacteria were resuspended in M63B1 supplemented with glucose (0.4% w/v). After two washes in this medium, cells were incubated for 24 h at 37uC with orbital shaking. Cultures were centrifuged for 10 min at 7,0006g at 4uC and the supernatants were filtered through a 0.2 mm filter (Stericup, millipore). The supernatant was concentrated by evaporation with rotary evaporator (Heidolph Laborota 4000 Efficient) coupled to a vacuum pump (ILMVAC LVS 210 T) and extensively dialysed (CelluSep, H1 membrane, 3 kDa cut-off) with gentle agitation at 4uC during 8 days against ultrapure water (3 baths/day).   (4) Size Exclusion Chromatography (SEC). The planktonic extract from K. pneumoniae MGH 78578 at 10 g.L 21 in 50 mM phosphate buffer (pH 7) supplemented with NaCl (150 mM) was filtered (at 0.45 mm) and analyzed by SEC at room temperature. The column used was a Superdex 200 column (1.5 cm650 cm), (GE Healthcare, Sweden) coupled to an Ä KTA Purifier system (Amersham Pharmacia Biotech, Sweden) and eluted with a 50 mM phosphate buffer (pH 7) supplemented with NaCl (150 mM) at a flow rate of 0.5 mL.min 21 . Fractions of 0.5 mL were collected and the sugar content was determined by phenolsulphuric acid method. Each fraction collected was assayed by the magnetic beads-mediated agglutination assay to determine the anti-biofilm activities with dialyzed planktonic extract from K. pneumoniae.
(5) NMR analysis. The dried polysaccharide was dissolved in D 2 O (99.9% D) and freeze-dried to replace exchangeable protons with deuterium. For NMR analysis, the exchanged polysaccharide was dissolved in D 2 O at (40 g.L 21 ). The NMR spectra of the solutions were recorded at 60uC using a Brucker Advance 600 spectrometer of 300 MHz equipped with 13 C/ 1 H dual probe. The NMR experiments were recorded with a spectral width of 3000 Hz, an acquisition time of 1.36 s, a pulse width of 7 ms, a relaxation time of 1 s and a number of 256 scans. The HOD signal was presaturated by a presaturation sequence.
(6) Total sugar quantification. The quantity of total polysaccharides was assessed by the phenol sulphuric acid method [27] using galactose as standard.

Surface Coating Assay with Capsular Polysaccharide Extract from K. pneumoniae Strains
Capsular polysaccharides were extracted from bacteria grown overnight in DW, according to the method previously described by Domenico et al. in 1989 [23]. Briefly, 500 mL of bacterial cultures were mixed with 100 mL of 1% Zwittergent 3-14 detergent (w/v) (fluka, chemika) in 100 mM citric acid. This mixture was incubated at 56uC for 20 min. After centrifugation for 5 min at 20,0006g, 300 mL of the supernatant was transferred to a new tube and 1.2 mL of absolute ethanol was added. The mixture was placed at 4uC for 20 min. After centrifugation, the supernatant was decanted and the pellet was washed once with 70% ethanol, and then dissolved in 250 mL of distilled water.
A volume of 25 mL of capsular polysaccharide extract of K. pneumoniae strains (with equivalent quantities of CPS) or 25 mL of PBS as a control, was transferred to the center of a well of a 24well tissue-culture-treated polystyrene microtiter plate (Falcon no. 353047). The plate was incubated for 48 h at 4uC. The excess of extract was removed by pipetting and the plate was incubated at room temperature to allow complete evaporation of the liquid. The wells were then filled with 1 mL of TS broth containing 1610 6 CFU.mL 21 of S. epidermidis RP62A. After 8 h, wells were rinsed with water and stained with 1 mL of 0.5% (w/v) crystal violet solution. Then, the wells were rinsed under running tap water, air-dried and were photographed.

Statistical Analysis
For analysis of the significance of differences in bacterial biomass between biofilm and planktonic culture, two-tailed Student's t-tests were used to compare two groups of data. All experiments were done at least three times. A P-value of #0.05 was considered to be statistically significant.

Influence of K. pneumoniae Cell-free Extracts on Biofilm Formation
The anti-biofilm activity of K. pneumoniae supernatant was assessed on several bacterial species and quantification of the biofilm biomass performed by crystal violet staining indicated no significant effect on K. pneumoniae, P. aeruginosa biofilms and some S. aureus isolates. In contrary, the biomass of E. coli, S. epidermidis, P. mirabilis, E. aerogenes, B. cereus and L. monocytogenes biofilm were significantly reduced in the presence of K. pneumoniae supernatant extract (Fig. 1). All the Staphylococcus strains forming polysaccharide-dependent biofilm were inhibited by K. pneumoniae supernatant extract, whereas only two strains with protein-dependent biofilm characteristic were impaired (Fig. 1). Microbial adhesion to solvents method (MATS) was performed on two of the S. aureus strains forming protein-dependent biofilm CH937 (affected by supernatant) and CH939 (unaffected). Results indicated that the CH937 strain had an acid and basic character while the nonsusceptible isolate, CH939, harboured a more basic surface (data not shown), a difference that might account in the interactions with the supernatant active molecules.
To determine if the effect of K. pneumoniae was due to some released molecules specifically produced in planktonic growth conditions, we investigated the effect of two other types of K. pneumoniae cell-free extracts, biofilm supernatant and acellular Size exclusion chromatograph of planktonic extract from K. pneumoniae together with quantification of total polysaccharides assessed by the phenol sulphuric acid method were performed. Elution profile of standard markers (tyroglobulin, c-globulin, ovalbumin, myoglobulin, vitamin B12, respectively) is represented by the dotted curve. Anti-biofilm activity of each sample was monitored by magnetic bead-mediated agglutination assays and expressed as Biofilm Formation Indices (BFI). Only samples collected between 7 and 11 mL displayed anti-biofilm activity (gray area). doi:10.1371/journal.pone.0099995.g004 matrix biofilm, on the adhesive capacity of S. epidermidis RP62A, one of the highest attenuated strain. Monitoring of biofilm formation was performed over time using a magnetic beadsmediated agglutination assay. The addition of all cell free extracts from K. pneumoniae cultures modified biofilm installation by S. epidermidis, but the effect was more pronounced with planktonic supernatants (Fig. 2A). In addition, experiments carried out with serial dilutions of planktonic supernatant and S. epidermidis RP62A showed that the inhibition of agglutination was dose-dependent (Fig. S1).
SEM observations of S. epidermidis biofilm formed in the presence of K. pneumoniae planktonic supernatant showed small  clusters dispersed on the surface, in contrast to the dense aggregates observed in the control (without the extract) (Fig. 2B). Similar effects were observed using supernatant extracts from other K. pneumoniae isolates (data not shown), indicating that this anti-biofilm effect was not strain-dependent.
Determination of the biomass indicated that the number of S. epidermidis CFU in the biofilms recovered after 8 hours of incubation was significantly lower in the presence of planktonic supernatants of K. pneumoniae (2.39610 7 CFU.mL 21 ) compared to the control (S. epidermidis alone (6.77610 7 CFU.mL 21 ) (p,0.05), whereas no effect was observed on the growth of S. epidermidis in planktonic cultures (Fig. 3). Consequently, the lower number of CFUs observed in biofilms treated with K. pneumoniae supernatant was probably due to an anti-adhesion effect rather than a biocide effect.

Physico-chemical Characterization
Preliminary analysis of K. pneumoniae supernatant by size exclusion chromatography (SEC) using a Superdex 200 column indicated that the active fraction (samples between 7 and 11 mL of elution volume) contained compounds with molecular weight higher than 100 kDa (Fig. 4). Treatments of the K. pneumoniae supernatant with proteinase K, lipase, DNase or RNase or heat did not impair the capacity of S. epidermidis to form biofilm (Fig. 5). In contrast, treatment with carbohydrate-active agent sodium metaperiodate significantly reduced its anti-biofilm activity (Fig. 5), suggesting that the anti-biofilm activity was due to heat stable macromolecules mainly composed of sugars (polysaccharide).

K. pneumoniae Capsular Polysaccharide Cell-to-surface Interactions
The molecular weight of this polysaccharide was evaluated by SEC-MALLS analysis at 1.5610 6 Daltons and the low polydispersity index (1.3) confirmed the presence of polysaccharide macromolecule family. Further HPAEC analysis after acid hydrolysis of extracted and partially purified polysaccharide of K. pneumoniae indicated that it was composed of five monosaccharides: galactose (47.8%), glucose (27.6%), rhamnose (15.4%), glucuronic acid (6.1%) and glucosamine (3.1%) and an 1 H NMR analysis showed 6 specifics resonance peaks (Fig. 6) Table 2).
To investigate the mode of action of the anti-biofilm component, we used capsular extracts of K. pneumoniae MGH 78578 (serotype K52), but from also strains belonging to other capsular serotypes: K1 (NTUH-2044 and CF23), K2 (CF7 and CF8) and K35 (LM21). Previous coating of these CPS extracts onto the surface of polystyrene wells efficiently repelled S. epidermidis biofilm formation, (Fig. 7), indicating that the CPS inhibits the bacteriasurface interactions of the abiotic substrate, whatever the capsular serotype. A similar phenotype was observed with extract from a LPS O-antigen deficient mutant of LM21 strain (LM21DwecA) whereas extracts from its capsule isogenic deficient strain (LM21Dwzx) did not show any inhibition (Fig. 7), indicating that capsule extract but not LPS O-antigen was responsible for the anti-biofilm phenotype.

Discussion
Klebsiella pneumoniae is a widespread Enterobacteria, which lives in surface water, soil, on plant and as commensal resident of the mammalian nasopharynx and gastrointestinal tract. Its capacities to form biofilm have been largely investigated [28][29][30][31][32] and most of the isolates are heavily surrounded by capsular exopolysaccharides [33,34]. In this study, we isolated and characterized a K. pneumoniae exopolysaccharide with anti-biofilm activity against both Grampositive and Gram-negative bacteria. Several bacteria produce polysaccharides with anti-biofilm activity, such as Escherichia coli, Pseudomonas aeruginosa, Lactobacillus acidophilus, Bacillus licheniformis, Streptococcus phocae, Kingella kingae, Vibrio sp. and Actinobacillus pleuropneumoniae [9,10,[19][20][21][22][35][36][37]. These polymers, mainly composed of glucose, galactose, mannose, glucuronic acid or rhamnose, are either CPS components such as Ec111, Ec300 of E. coli and the CPS (serotype 5) of A. pleuropneumoniae, or non capsular polysaccharides such as the PAM galactan and the Pel/Psl molecules produced by K. kingae and P. aeruginosa PAO1, respectively. As with K. pneumoniae in this study, most of the previously described anti-biofilm polysaccharides were detected in both planktonic culture supernatants and in biofilm extracts [18,19] indicating that this synthesis was more related to the cells density than to the sessile condition. The higher inhibition observed in our study with planktonic supernatant compared to biofilm extracts ( Fig. 2A) is likely due to differences both in the amount of biological material in the initial samples and in the recovery procedures.
HPAEC analysis of dried supernatant from K. pneumoniae showed that the polysaccharide with anti-biofilm activity was composed of five monosaccharides (galactose, glucose, rhamnose, glucuronic acid and glucosamine), three of them, rhamnose, galactose and glucose, being specific of some K. pneumoniae CPS structures previously characterized [38]. In order to determine if the active macromolecules in K. pneumoniae supernatant extract were indeed CPS, 1 H NMR analysis (Fig. 5) was carried out and confirmed, based on the comparison with other NMR analysis of CPS described in literature [39], the presence of capsular polysaccharide structure in the dried extract. Indeed, the assignment of the 1 H NMR analysis was very similar to that of the capsular polysaccharide reported by Stenutz et al., with a resolved resonance in the anomeric region (4.5-5.5 ppm) and chemical shifts at 1.27 and 1.41 ppm, which is the characteristic signal of methyl protons of rhamnose residue ([R2)-a-L-Rhap-(1R]; [R4)-a-L-Rhap-(1R] respectively [39]. The signals in the anomeric region at 4.58 ppm, 5.21 ppm and 5.45 ppm have been attributed to galactose type sugar residues and more especially to [R3)-b-D-Galp-(1R], [a-D-Galp-(1R] and R2,3)-a-D-Galp-(1R] residues respectively [39]. Therefore, it is likely that the structure of the polysaccharide isolated in this study from K. pneumoniae MGH 78578 (K52 serotype) dried extract is similar to the CPS [39]. However, slight differences were observed. As described, the hydrolysis of K52 capsular polysaccharides with TFA gave rise to galactose and rhamnose with the molar ratio of 3:2 [39], whereas in the case of K. pneumoniae MGH 78578 dried extract, a molar ratio of 3:1 was observed. Moreover, the HPAEC analysis revealed the presence of additional sugars such as glucose and glucosamine compared to the published K52 CPS composition. The presence of these additional sugars could be attributed to contamination by the culture medium (glucose) and by traces of LPS fraction. Indeed, as described by Kubler-Kielb et al. (2013), many LPS cores from K. pneumoniae are composed of glucosamine residues (GlcN) [38]. Despite these differences, further characterization of the polydispersity and purity performed by SEC-MALLS analysis clearly indicated that the polysaccharides from K. pneumoniae MGH 78578 dried extract possessed high molecular mass and was therefore more likely capsular polysaccharides. Indeed, capsular polysaccharides such as amylovoran from Erwinia amylovora, or stewartan from Pantoea stewartii have been previously evaluated to be in the range of 1-2 MDa [40].
Several capsular polysaccharides have been associated with antibiofilm activities. Karwacki et al. recently showed that anti-biofilm activities of Actinobacillus pleuropneumoniae are due to capsular polysaccharide [37]. Some group 2 capsular polysaccharides of E. coli, composed of simple polysaccharides, also exhibited nonbiocidal anti-biofilm activity [9,19]. The E. coli capsules classification is based on their structure, organization and mechanisms of biosynthesis [41] and K. pneumoniae capsules are related to the E. coli group 1 capsule. These polymers are acidic, possess a low charge density and do not make colonic acid. To our knowledge, no study had yet demonstrated the anti-biofilm activity of this group of polysaccharides. The K. pneumoniae CPS anti-biofilm activity was not dependent on the capsular serotype, since capsular extracts from serotypes K1, K2, K35 and K52 exhibited this activity (Fig. 7). In addition and although traces of LPS could be present in capsule extracts, the role of LPS O-antigen was ruled out since capsule extract from a wecA deficient mutant showed anti-adhesion properties similar to that of the wild-type strain (Fig. 7). Nethertheless the role of other components of the LPS such as the core polysaccharide cannot be ruled out.
Several hypotheses have been formulated regarding the mechanism of action of anti-biofilm polysaccharides. Most of them inhibit biofilm formation by coating the abiotic surfaces and therefore modifying the initial adhesion of the bacteria to the substrate [9,19,37]. Some of them also impaired the bacterial aggregation, as described by Karwacki et al. [37]. In our study, precoating of surface with CPS extract indicated that K. pneumoniae CPS probably inhibits the S. epidermidis bacteria-surface interactions rather than disrupting the bacterial interactions. This was supported by the fact that incubation of a 6 h old S. epidermidis biofilm with K. pneumoniae CPS extract did not modify the biomass after 2 h of further incubation, compared to a control biofilm without any CPS addition (data not shown). We hypothesize that CPS modifies the physical properties of abiotic surfaces by increasing its hydrophobicity. However, differences were observed between isolates within the S. aureus species, owing probably to differences in individual bacterial cell surface characteristics. Using a genetic approach, Travier et al. recently showed that modifications in surface physicochemical properties of E. coli cells were able to modify the anti-biofilm activity of group 2 anti-biofilm capsule polysaccharides, probably due to changes in ionic charge and Lewis base properties induced by the CPS polysaccharides [42].
We previously showed that the capsular polysaccharide of K. pneumoniae is implied in surface adhesion, spacing and ordering of bacteria in the initial step of biofilm formation, and is required for late biofilm maturation step ( [31]; Muraglia et al. unpublished data). Though the initial adhesion of bacteria on the surface constitutes a key step in the formation of biofilm, the spread of bacteria on the surface is also another important factor in the formation of a biofilm, especially in a multispecies highly competitive environment. Some Gram-negative rods such as the highly motile Pseudomonas aeruginosa produce exopolysaccharides that promote their own surface movement during the early stages of biofilm formation [43]. K. pneumoniae is a non motile Gramnegative rod and may have developed different social strategies leading to surface exclusion of competitors by large CPS production and therefore allowing successful surface colonization. Figure S1 Dose-dependent effect of K. pneumoniae supernatant. S. epidermidis (inoculum, 10 6 CFU/mL) biofilm formation was measured in the presence of several concentrations of K. pneumoniae supernatant with the BioFilm Magnetic bead aggregation assayH: without supernatant (N); with undiluted supernatant (X); with supernatant diluted to 1/2 (m); supernatant diluted to 1/5 (&) and supernatant diluted to 1/10 (*). Quantitative data were expressed as BFI (means 6 SD of three determinations) and the dotted line represents the threshold of magnetic bead-mediated agglutination detection. (TIF)