|Home | About | Journals | Submit | Contact Us | Français|
Brominated furanones from marine algae inhibit multicellular behaviors of gram-negative bacteria such as biofilm formation and quorum sensing (QS) without affecting their growth. The interaction of furanone with QS in gram-positive bacteria is unknown. Staphylococci have two QS systems, agr and luxS, which lower biofilm formation by two different pathways, RNAIII upregulation and bacterial detachment, and polysaccharide intercellular adhesin (PIA) reduction, respectively. We synthesized natural furanone compound 2 [(5Z)-4-bromo-5-(bromomethylene)-3-butyl-2(5H)-furanone] from Delisea pulchra and three analogues to investigate their effect on biofilm formation in gram-positive bacteria. Compound 2, but not the analogues, enhanced the biofilms of Staphylococcus epidermidis 1457 and 047 and of S. aureus Newman at concentrations between 1.25 and 20 μM. We show the growth inhibition of S. epidermidis and S. aureus by free furanone and demonstrate bactericidal activity. An induction of biofilm occurred at concentrations of 10 to 20% of the MIC and correlated with an increase in PIA. The biofilm effect was agr independent. It was due to interference with luxS, as shown by reduced luxS expression in the presence of compound 2 and independence of the strong biofilm formation in a luxS mutant upon furanone addition. Poly(l-lysine)-grafted/poly(ethylene glycol)-grafted furanone was ineffective on biofilm and not bactericidal, indicating the necessity for free furanone. Free furanone was similarly toxic for murine fibroblasts as for staphylococci, excluding a therapeutic application of this compound. In summary, we observed a biofilm enhancement by furanone in staphylococci at subinhibitory concentrations, which was manifested by an increase in PIA and dependent on luxS.
Implant-associated infections are commonly caused by microorganisms growing in biofilms (10, 31). Staphylococcus aureus and S. epidermidis are the major causative agents of implant infections. Upon implantation, the surface of the implant is rapidly covered with host proteins and cells. If bacteria are present, they compete with host cells in a “race for the surface” (15, 41). Many approaches have been made to prevent bacterial colonization of surfaces, e.g., by coating with antifouling substances such as poly(l-lysine)-grafted-poly(ethylene glycol) copolymers (PLL-g-PEG) (17, 33, 42). Bacterial adherence to implant surfaces occurs in two phases (7, 32). First, staphylococci adhere via exopolysaccharides and microbial surface components recognizing adhesive matrix molecules in the wound. Then, staphylococci proliferate and accumulate in multilayers of exopolysaccharides, what is commonly described as the biofilm. The major component of staphylococcal biofilms is polysaccharide intercellular adhesin (PIA), which is encoded by the ica genes (13, 16). PIA production causes resistance against antibiotics (13) and makes bacteria less vulnerable by shielding them from immune defense (28). Global regulators, including the alternative sigma factor B (σB) (26) and the quorum-sensing (QS) system agr, modulate biofilm turnover by regulating a whole set of genes. It was found that agr upregulates RNAIII, which is associated with a reduction of biofilm (44). Clinical isolates of S. epidermidis often have mutations of agr, which is linked to more biofilm on the foreign body and less dissemination (27). Another QS system encoded by luxS is required for autoinducer 2 synthesis, which negatively affects biofilm by reducing ica gene transcription in S. epidermidis (46). Thus, substances acting on biofilm, e.g., via QS are drug candidates for prevention and treatment of gram-positive implant infections.
Halogenated furanones are natural compounds secreted by the alga Delisea pulchra. They are structurally similar to bacterial acyl homoserine lactones (AHL). AHL are released upon changes in cell population density, cross membranes and bind to transcription factors of the LuxR family, thereby inducing QS controlled genes. Gram-negative bacteria use AHL-dependent relationships for interference with eukaryotes and plants. Thus, marine algae have developed halogenated furanones as AHL antagonists, most likely in response to a negative impact of bacterial colonization (14). Delisea pulchra produces more than 30 furanones with a variable potential to prevent swarming of gram-negative bacteria with an AHL system, without altering their general metabolism or growth (14, 34). Furanones suppress virulence factor production and pathogenesis in Pseudomonas (20) and biofilm formation in Escherichia coli (40). In Bacillus subtilis this effect was associated with a reduction in viability (39).
Several natural furanones were found to inhibit the growth rate of selected clinical ocular strains of S. aureus and S. epidermidis (11, 24, 25) by an unknown mechanism. Hume et al. described biofilm inhibition in S. epidermidis by covalently bound furanone in vitro and a reduced bacterial load on furanone-coated catheters in sheep in vivo (23).
A better understanding of the effect of free and surface-bound furanone on gram-positive bacteria is crucial for a potential clinical application of this class of substances in implant infections. Therefore, we assessed the bactericidal and biofilm-modulating activity of free and surface-bound furanone and compared it to its eukaryotic cytotoxicity. We were able to show that furanone is similarly toxic for staphylococci and eukaryotic cells, rendering a clinical application improbable. In addition, we found a biofilm-enhancing effect of furanone on several staphylococci associated with enhanced PIA production. This effect in staphylococci was related to a downregulation of the QS system luxS, whereas σB, agr, and RNAIII were not involved.
Tryptic soy broth (TSB) and Mueller-Hinton broth and agar were obtained from Becton Dickinson (Allschwil, Switzerland). Luria-Bertani broth, crystal violet, and poly-l-lysine (PLL) were purchased from Sigma-Aldrich (Buchs, Switzerland). A CytoTox96 kit, proteinase K, RNasin, and random primers were obtained from Promega (Dübendorf, Switzerland), and lysostaphin was purchased from Genmedics (Reutlingen, Germany). Taq DNA polymerase was purchased from Invitrogen (Lucerne, Switzerland). The RNeasy minikit, RNAprotect bacterial reagent, Omniscript reverse transcriptase kit, and DNase were obtained from Qiagen (Hombrechtikon, Switzerland). The rabbit immunoglobulin G (IgG) isotype antibody was purchased from Vector (Geneva, Switzerland). Horseradish peroxidase-conjugated donkey anti-rabbit IgG was from Jackson (Magden, Switzerland), and the ECL Western blotting analysis system was obtained from Amersham Biosciences/GE Healthcare (Otelfingen, Switzerland). HEPES buffer and all reagents for cell culture were obtained from Gibco/Invitrogen.
Several furanone compounds were used in the present study: 4-bromo-5-(bromomethylene)-3-butyl-2(5H)-furanone (molecular mass, 310.0 g/mol), known as compound 2; 4-bromo-5-(bromomethylene)-3-(1-bromobutyl)-2(5H)-furanone (molecular mass, 388.9 g/mol [compound 3]); 4-bromo-5-(bromomethylene)-3-(1-(−2-propynyloxy)butyl)-2(5H)-furanone (molecular mass, 364.1 g/mol [compound 4]); and 4-bromo-5-(bromomethylene)-3-(1-(2-(2-(2-azidoethoxy)ethoxy)ethoxy)butyl)-2(5H)-furanone (molecular mass, 483.2 g/mol [compound 5]) (Fig. (Fig.1).1). The synthesis of furanones, as well as the addition of functional groups on the side chain was achieved according to published methods (2, 29, 35, 38). Free furanones were dissolved in ultrapure ethanol and applied to the wells of 96-well plates. Excess ethanol was evaporated under sterile conditions, and growth medium was added to resolubilize the furanone for experiments with bacteria or fibroblasts.
The surface bound analogue of furanone compound 2 is compound 5 (which was also used in a free, unlinked form), which was grafted by click-chemistry to the polyethylene glycol (PEG) and assembled on a PLL backbone. Therefore, furanone was sequentially reacted with N-bromosuccinimide and 2-(2-(2-azidoethoxy)ethoxy)ethanol to a terminal azido functionalized furanone. The azido group was then reacted with a terminal acetylene function on a 3.4-kDa PEG polymer (Laysan Bio, Inc.), which was finally grafted to the 20-kDa PLL backbone. A grafting density of 80% PEG-furanone and 20% PEG was used here. For immobilization purposes, 100 μl of a compound composed of 0.1 mg of PLL-g-PEG-furanone/ml in HEPES buffer (10 mM in sterile NaCl 0.9%) was applied to the wells of a 96-well plate, followed by incubation for at least 30 min at room temperature under sterile conditions. Subsequently, the surfaces were rinsed with HEPES buffer followed by ultrapure water, blow drying under a stream of nitrogen, and storage in sterile containers until use. The surfaces were rehydrated for at least 30 min in HEPES buffer before use.
In our studies, we used S. epidermidis 1457 and its isogenic ica::ermB (32) and luxS::ermB (46) mutants, S. epidermidis 047 (18) and its isogenic ica::ermB mutant (kindly provided by F. Götz), S. aureus SA113 (ATCC 36665), S. aureus Newman (kindly provided by F. Götz) and its agr::tet mutant (ALC355 Δagr ), B. subtilis (kindly provided by R. Frei), and Pseudomonas aeruginosa PA01 (ATCC 9027).
Bacterial strains were freshly grown in TSB for 7 h at 37°C without shaking. A 1:100 dilution was used to inoculate the overnight culture, which was further diluted into a fresh culture. After reaching the log phase, the culture was diluted to 105 CFU/ml in fresh medium before being used in experiments with free furanone. In experiments with covalently linked furanone, bacteria were diluted to 106 to 102 CFU/ml. Then, 100-μl portions of bacterial inocula were seeded into 96-well plates, followed by incubation for 18 h at 37°C without shaking. Bacterial numbers were estimated by optical density at 600 nm and assessed by plating serial dilutions on Mueller-Hinton agar.
Biofilm assays were performed in 96-well plates using TSB containing 0.25% glucose, with modifications as previously published (4). After incubation for 18 h at 37°C without shaking, supernatants were removed, and the plates were gently washed three times with 0.9% NaCl. For fixation of biofilm, plates were incubated for 60 min at 60°C. Each well was incubated with 100 μl of a 0.5% crystal violet solution for 20 min and washed under running tap water, and 100 μl of 33% acetic acid was added. Samples were transferred into enzyme-linked immunosorbent assay plates for reading of absorbance at 590 nm in a Molecular Devices reader (Applied Biosystems, Rotkreuz, Switzerland).
MIC were determined according to Clinical and Laboratory Standards Institute standards by the macrodilution method as described previously in document M7-A7 (5).
A modification of the method was used as published by Cramton et al. (6). Briefly, S. epidermidis 1457 (105 CFU/ml) was grown in TSB with 2.5 to 20 μM compound 2. After 18 h of incubation, the supernatants were removed, adherent CFU were counted by plating, and bacteria were collected by centrifugation for 10 min at 4,000 × g. The bacterial pellets containing 3 × 108 to 4 × 108 CFU were resuspended in 0.5 M EDTA (pH 8.0), incubated for 5 min at 99°C, and then further incubated with 6 U of proteinase K overnight at 37°C.
Lysates corresponding to 5 × 105 CFU were transferred to a nitrocellulose membrane (Macherey-Nagel, Oensingen, Switzerland) using a Bio-Dot SF microfiltration apparatus (Bio-Rad, Reinach, Switzerland). Membranes were washed twice with Tris-buffered saline (TBS) and blocked with 10% nonfat milk for 4 h at room temperature. After being washed with TBS with 0.05% Tween (TTBS), the membranes were incubated with rabbit anti-PIA antibody (1 μg/ml; kindly provided by D. Mack, Swansea, Great Britain) or isotype-matched control antibody in 1% milk-TTBS overnight at 4°C. After another wash and incubation with the secondary horseradish-peroxidase donkey anti-rabbit IgG (0.1 μg/ml), membranes were developed with the ECL Western blotting analysis system and visualized on films (Kodak and Sigma-Aldrich).
S. epidermidis 1457 was grown in 96-well plates with subinhibitory concentrations of free furanone compound 2 (0, 2.5, 10, and 20 μM, 48 wells per concentration). After incubation for 18 h, bacteria were resuspended, and 2 volumes of RNAprotect bacterial reagent were added. Tubes were mixed by inverting them several times and incubated 5 min at room temperature before centrifugation for 15 min at 3,800 × g. RNA isolation and purification was performed according to the instructions of the RNeasy minikit. Bacterial cells were disrupted with 50 U of lysostaphin, further digested with 5 U of proteinase K and mechanically disrupted for 30 s at a speed of 6.5 m/s using a lysing matrix B tube in the FastPrep 120A instrument (Bio 101 Systems/Lucernachem, Lucerne, Switzerland). RNA was treated with DNase and quantified with a NanoDrop apparatus (ND-1000; Witec-AG, Littau, Switzerland).
For reverse transcription, 1 μg of RNA was mixed with 18 μl of Omniscript transcriptase reagent mix containing RNasin and random primers according to the instructions of the Omniscript reverse transcriptase kit. Reverse transcription was performed for 1 h at 37°C, and synthesized cDNA was chilled on ice. cDNA was kept at −20°C until use.
The following primers were used: RNAIII forward, 5′-TGAAGTTATGATGGCAGCAGAT-3′; RNAIII reverse, 5′-GTTGGGATGGCTCAACAACT-3′; gyrase B forward, 5′-TTATGGTGCTGGACAGATACA-3′; gyrase B reverse, 5′-CACCGTGAAGACCGCCAGATA-3′; σB forward, 5′-TTGGTATGGTTGGTCTAATAGGTGC-3′; σB reverse, 5′-CTGAAACTTCTAAGCGTTGTGCG-3′; luxS forward, 5′-TCCTATGGGTTGTCAAACTGG-3′, luxS reverse, 5′-CCTTCTCCGTAGATGTCATTCC-3′; 16S RNA forward, 5′-ACTTTCTGGTCTGTAACTGACGCTG-3′; and 16S RNA reverse, 5′-ACCCAACATCTCACGACACGAG-3′. PCR was performed using Taq DNA polymerase with 25 amplification cycles of 94°C for 30 s, 50°C for 30 s, and 72°C for 30 s. RNA bands were visualized on a 1.2% agarose-Tris borate-EDTA gel. Quantification of signals was performed with QuantityOne (Bio-Rad) software and expressed as arbitrary units calculated from the intensity/mm2.
L929 fibroblasts were cultured in RPMI with 5% heat-inactivated fetal bovine serum, and 5 × 104 cells/ml were seeded in a 96-well plate (Falcon) containing different concentrations of free furanone. Lactate dehydrogenase (LDH) release was measured as published previously (8) in the supernatants of treated cells after 24 h and analyzed with a CytoTox96 kit according to the manufacturer's instructions (Promega).
Data were analyzed by using Prism 5.0a (GraphPad Software, Inc.), and a Mann-Whitney test was used for statistical analyses. A P value of <0.05 was considered statistically significant.
In our study, we used one of the well-studied natural furanones, called compound 2 (12), as well as three derivatives of this substance, which were synthesized in our laboratory (Fig. (Fig.1).1). We first studied the effect of free compound 2 on biofilm production of several gram-positive bacteria (Fig. (Fig.2).2). Interestingly, free compound 2 significantly enhanced biofilm formation of the two S. epidermidis strains 1457 and 047, which are strong biofilm producers (19, 32), at concentrations between one-tenth and one-fifth of their MICs (Table (Table1)1) . Isogenic S. epidermidis strains 1457 and 047, lacking the ica genes (Δica), were used as controls. Both strains did not form biofilm with or without furanone treatment (Fig. (Fig.2).2). Compound 2 also strongly enhanced biofilm of S. aureus Newman at subinhibitory concentrations but left the weak biofilm former S. aureus SA113 unaffected. Furthermore, the strong biofilm of B. subtilis was not influenced by furanone compound 2 (Fig. (Fig.22).
The increase of biofilm at subinhibitory furanone concentrations did not affect the viability of bacteria as shown by CFU counting (Fig. (Fig.22).
Furanone compounds 3 to 5, which represent structural analogues of compound 2, did not affect biofilm of S. epidermidis 1457 (Fig. (Fig.3).3). These results indicate that minor modification of compound 2, i.e., by an additional bromide functionality (compound 3) or by varying the side chains (compounds 4 and 5), profoundly altered the interaction with bacteria.
Furanone compound 2 was bactericidal for staphylococci at concentrations greater than 10 μM (Fig. (Fig.2).2). Therefore, we determined the MICs according to Clinical and Laboratory Standards Institute standard methods. S. aureus strains had MICs of 15 μM and were highly susceptible to soluble furanone compound 2. MICs for S. epidermidis strains and B. subtilis were higher, ranging from 30 to 65 μM and 100 μM, respectively. The minimum bactericidal concentrations (MBCs) were higher than 130 μM for all tested species except for B. subtilis (Table (Table1).1). Compounds 3 to 5 were only bactericidal for S. epidermidis 1457 at concentrations greater than 50 μM, compound 5 was the least potent (Fig. (Fig.33).
Staphylococcal biofilm consists mainly of PIA (13). Therefore, we quantified the effect of furanone compound 2 on biofilm by measuring PIA. In slots containing lysate from equal CFU numbers of S. epidermidis 1457 wild type, the amount of PIA was increased after incubation of bacteria with subinhibitory concentrations of compound 2 (Fig. (Fig.4a).4a). Synthesis of PIA requires the enzymes encoded by the intercellular adhesion (ica) operon, which was found to be regulated by σB (26). We observed, however, unaltered σB expression, along with the increase in PIA (Fig. (Fig.4b4b).
Staphylococcal biofilm is negatively regulated by the QS system agr (44) with RNAIII as an important effector gene. To study whether furanone compound 2 interferes with the agr system, we evaluated RNAIII expression in response to compound 2. RNAIII expression of S. epidermidis 1457 remained unaltered at furanone concentrations ranging from 2.5 to 20 μM (Fig. (Fig.4c).4c). In addition, we quantified biofilm formation and growth of an agr deletion mutant (Δagr) of S. aureus Newman in the presence of furanone compound 2. Free compound 2 similarly enhanced biofilm formation in S. aureus Newman wild-type and Δagr strains (Fig. (Fig.4d).4d). The bactericidal effect of compound 2 was similar upon wild-type and mutant S. aureus strains (data not shown).
Alternatively, luxS is known to repress ica gene expression and thereby PIA biosynthesis (46). Therefore, we quantified luxS mRNA under biofilm-enhancing concentrations of furanone compound 2. LuxS expression was clearly reduced at 20 μM (Fig. (Fig.4e,4e, arbitrary units of luxS/16S RNA: untreated, 0.783; 2.5 μM treatment, 0.762; 10 μM treatment, 0.749; and 20 μM treatment, 0.653). We next evaluated the growth and biofilm of a luxS deletion mutant (ΔluxS) of S. epidermidis 1457. Biofilm formation in the ΔluxS mutant was strong, confirming the described causal repressive relationship between the luxS and PIA production. Of note, the pronounced biofilm in the ΔluxS mutant was not influenced by furanone addition, lending further support to the idea that furanones impact biofilm formation via luxS (Fig. (Fig.4f).4f). The bactericidal effect of furanone compound 2 was similar on the wild type and the ΔluxS mutant (data not shown).
To use furanone as antibacterial coating for implants, furanone may either be incorporated in a coating that releases the antibacterial compound over time or, alternatively, it has to be linked covalently to the surface. We have chosen the second option and linked furanone compound 5 covalently to the biopassive (nonfouling) surface PLL-g-PEG polymer with the aim of achieving a bactericidal activity due to a potential cell wall action of furanone. Surfaces were prepared by self-assembly of the PLL-g-PEG-furanone and subsequently exposed to S. epidermidis 1457. On PLL-g-PEG-coated surfaces, S. epidermidis was unable to adhere and form biofilms (Fig. (Fig.5).5). The grafted furanone abolished the biopassive characteristics of PLL-g-PEG and allowed adhesion, as well as biofilm formation of S. epidermidis to a similar level, as found on control surfaces with the nonbiopassive PLL (Fig. (Fig.55).
Furanones have previously been described to inhibit quorum sensing in gram-negative bacteria and reduce biofilm through interference with the AHL system by accelerating degradation of LuxR and by decreasing the DNA-binding activity of LuxR (9, 21). In view of the unexpected effects of free compound 2 in gram-positive bacteria, we tested biofilm and MIC of P. aeruginosa PA01 after incubation with furanone compound 2. In contrast to staphylococcal biofilms, P. aeruginosa biofilm was unaffected and decreased only at high furanone concentrations from 322 μM to 2.58 mM without any bactericidal activity. The observed tolerance of P. aeruginosa to furanone compound 2 was further confirmed with high MICs and MBCs of more than 130 μM.
The cytotoxicity of an antimicrobial compound is important, if the compound is to be used to prevent bacterial implant infections. Therefore, we quantified cytotoxicity of the different furanone compounds in eukaryotic L929 fibroblasts by LDH release. After 24 h, 30% cytotoxicity was reached at a concentration of 28.4 ± 8.4 μM with compound 2, at 18.6 ± 12.3 μM with compound 3, and at 38.35 ± 15.0 μM with compound 4; for compound 5 this value was 12.1 ± 5.7 μM (Fig. (Fig.6).6). In contrast, PLL-g-PEG-coated furanone compound 5 was not toxic for L929 fibroblasts (data not shown). These results indicate that murine fibroblasts and gram-positive bacteria are similarly susceptible to free furanones.
The present study was designed to investigate the effect of furanone upon biofilm in gram-positive bacteria. In many biofilm-forming bacteria, the differentiation from planktonic to sessile exopolysaccharide-producing cells is associated with the activation of a complex regulatory network in response to quorum sensing signals and/or environmental stress factors such as high osmolarity, detergents, urea, ethanol, and oxidative stress (37).
In S. epidermidis, the production of biofilm may be due to upregulation by the global regulator σB (26) or by interference with the luxS QS systems (46). We did not observe any change of σB expression levels but found luxS downregulated by subinhibitory concentrations of furanone. The involvement of LuxS was further confirmed by using a ΔluxS mutant, which generally formed more biofilm than the isogenic wild type but remained unaffected by subinhibitory furanone concentrations. Recent studies on LuxS function suggest that, besides its transcriptional repression of ica genes, it has strong metabolic effects and thus the mechanism of interaction between luxS and PIA under furanone treatment remains to be determined in a future study (30).
Alternatively, furanone may act by repression of agr, which is a global regulator of QS in S. aureus and S. epidermidis and enhances biofilm detachment via protease activation (3) and via altering biofilm structure (43). However, our data do not support an action of furanone on agr, since subinhibitory concentrations similarly enhanced biofilm in wild-type and Δagr staphylococci. Furthermore, an agr-independent mechanism is supported by the fact that RNAIII expression, which is upregulated by agr (36) and is associated with a reduction of biofilm (44), was not modified by furanone. Finally, our observation that furanone acted similarly on biofilm in S. epidermidis 1457 and in S. epidermidis 047, which is a natural Δagr mutant (43), excludes a contribution of agr in furanone-induced biofilm formation.
The furanone effects on staphylococcal biofilm resemble results obtained with tetracycline and quinupristin-dalfopristin (37). These substances, which were previously found to increase S. epidermidis biofilm at subinhibitory concentrations, are protein synthesis inhibitors acting at the ribosome. The bactericidal mechanism of action of furanone upon staphylococci is unknown; our results indicate that furanone has to enter the bacterial cell, since the PLL-g-PEG-grafted furanone remained without effect on staphylococci. The results of an earlier study describing biofilm inhibition in S. epidermidis by covalently bound furanone in vitro and a reduced bacterial load on furanone-coated catheters in sheep in vivo (23) are discrepant with our results. However, it should be pointed out that the authors of that study used a different furanone compound and a different immobilization strategy. Their results may be explained by a bactericidal effect of high furanone concentrations upon S. epidermidis, since [3H]thymidine uptake—taken as a measure of biofilm, but truly a growth indicator—was found to be reduced by furanone (23). We found a similar biofilm effect of furanone on S. epidermidis and S. aureus at least in strain Newman, which is a strong biofilm producer. In contrast, B. subtilis biofilm was not affected, and bactericidal activity was observed only above furanone concentrations of 50 μM, a finding that is in agreement with a reduction in viability found in a previous study (39).
The potential application of furanone as a bactericidal substance against staphylococci depends on the therapeutic window between the MIC for bacteria and toxicity in eukaryotic cells. We found a potent bactericidal activity for the compounds 2 and 3 and thus confirmed previous results showing antibacterial activity (24, 25). In addition, our data demonstrate that S. epidermidis is less susceptible than S. aureus. We also found no variability of MIC among S. epidermidis strains and no difference of MIC between ica-positive and -negative strains. These findings indicate that the bactericidal action of furanone is independent of biofilm and uses an unknown mechanism, which is common to several types of gram-positive bacteria and absent in gram-negative bacteria. We could confirm the previously described biofilm-reducing effect for Pseudomonas in rich growth medium without bactericidal activity (22) by using high furanone concentrations. The bactericidal action of the different furanone compounds on gram-positive bacteria occurred, however, at a similar concentration as cytotoxicity for murine fibroblasts. Interestingly, compound 5, which was the least bactericidal, was the most toxic furanone compound in fibroblasts. Therefore, these particular furanones are not suitable for further development to drugs with systemic application, as similarly shown for other furanones (24, 25). However, a topical application, e.g., in the treatment of burns, is conceivable. Alternatively, further chemical modifications may yield compounds that have a therapeutic window. Baveja et al. found human granulocyte viability unaltered after incubation with a polystyrene-bound furanone (1). This may indicate that covalently bound furanone was inactive or the concentration of immobilized furanone was very low. In addition, human myeloid cells may be less susceptible to furanone than murine fibroblasts.
In summary, we show that free, but not surface-bound furanone is toxic for staphylococci and eukaryotic cells to a similar extent. Furthermore, subinhibitory concentrations of furanone compound 2, but none of its analogues, enhance biofilm formation of staphylococci via a luxS-suppressing mechanism, leading to increased PIA production.
We thank the RMS Foundation for supporting R. Kuehl. We are grateful for the advice of M. Schmaler, N. Jann, and W. Zimmerli and for the technical help of F. Ferracin and Z. Rajacic. We thank D. Mack and L. Harris, Swansea, Great Britain, for providing the anti-PIA antibody.
Published ahead of print on 20 July 2009.