Human Antimicrobial Peptide LL-37 Inhibits Adhesion of Candida albicans by Interacting with Yeast Cell-Wall Carbohydrates

Candida albicans is the major fungal pathogen of humans. Fungal adhesion to host cells is the first step of mucosal infiltration. Antimicrobial peptides play important roles in the initial mucosal defense against C. albicans infection. LL-37 is the only member of the human cathelicidin family of antimicrobial peptides and is commonly expressed in various tissues and cells, including epithelial cells of both the oral cavity and urogenital tract. We found that, at sufficiently low concentrations that do not kill the fungus, LL-37 was still able to reduce C. albicans infectivity by inhibiting C. albicans adhesion to plastic surfaces, oral epidermoid OECM-1 cells, and urinary bladders of female BALB/c mice. Moreover, LL-37-treated C. albicans floating cells that did not adhere to the underlying substratum aggregated as a consequence of LL-37 bound to the cell surfaces. According to the results of a competition assay, the inhibitory effects of LL-37 on cell adhesion and aggregation were mediated by its preferential binding to mannan, the main component of the C. albicans cell wall, and partially by its ability to bind chitin or glucan, which underlie the mannan layer. Therefore, targeting of cell-wall carbohydrates by LL-37 provides a new strategy to prevent C. albicans infection, and LL-37 is a useful, new tool to screen for other C. albicans components involved in adhesion.


Introduction
Candida albicans is the major fungal pathogen that affects humans. Although C. albicans is a commensal organism on the cutaneous and mucosal surfaces of oral, gastrointestinal, urinary, and vaginal tracts of healthy individuals [1,2], it is also an opportunistic pathogen and can cause infections ranging from superficial mucosal infections to hematogenously disseminated candidiasis. In immunocompromised patients, C. albicans is responsible for a number of life-threatening infections [1,2]. Moreover, with the rapidly expanding use of medical devices (e.g., indwelling catheters) and increases in the number of patients receiving antibiotic and immunosuppressive therapies, there is an increased risk of fungal penetration through mucosal barriers with subsequent entry into the blood stream, which often leads to multiorgan infections. Consequently, C. albicans is the leading cause of nosocomial bloodstream infections and has a mortality rate of 40% [3,4].
C. albicans pathogenesis studies have suggested that several steps may lead to mucosal infection, i.e., early colonization, invasion, and late tissue disruption [5,6]. C. albicans first colonizes and proliferates on the mucosal surfaces of host epithelial cells; these events are followed by invasion and tissue damage [7,8]. Therefore, C. albicans adhesion is the first step in infection and allows the pathogen to persist on mucosal surfaces. As the outermost layer of C. albicans, the cell wall interacts with the host cells [9]. The cell wall of C. albicans contains many different carbohydrates and proteins that come into contact with epithelial cells and facilitate cell-cell interconnections [10,11].
Host cells defend against Candida infection by producing cytokines [12], chemokines [13], and antimicrobial peptides (AMPs) [14,15]. Human AMPs are the first line of mucosal immunity [16,17]. AMPs are generally short (10 to 50 amino acid residues), positively charged (generally +2 to +9), and contain $30% hydrophobic residues [18]. Consequently, when folded, many of these peptides exhibit amphiphilic, helical structures and can form pores in microbial membranes, which causes membrane rupture and eventual cell death [19]. Recent studies have indicated that AMPs can also inhibit the biosynthesis of microbial cell walls, nucleic acids, and proteins, and can inhibit the activity of certain microbial enzymes [19,20,21,22]. In humans, different types of AMPs are synthesized and secreted by various cells and tissues, including skin, mucosal surfaces, neutrophils, and epithelia [23].
LL-37 is the only member of the human cathelicidin family of AMPs [24]. LL-37 is stored in specific neutrophil granules as an inactive propeptide, which is cleaved extracellularly to yield the mature, active peptide [25]. In addition to exhibiting broad-spectrum antimicrobial activity against bacteria and fungi, LL-37 has other activities related to host defense, including chemotactic migration, endotoxin neutralization, angiogenesis, and wound healing [26,27]. To date, studies concerning the effects of LL-37 on C. albicans have been few in number. In a study of the candidacidal activity of AMPs, LL-37 was found to remain associated with the C. albicans cell surface, whereas other AMPs (e.g., histatin 5) translocated through the membrane and accumulated intracellularly [28].
For the study reported herein, we further investigated the effects of LL-37 on C. albicans. We show that LL-37 significantly reduced C. albicans attachment to an abiotic surface, to oral epidermis, and to murine urinary bladders. In addition, LL-37 was found to associate with C. albicans cell-wall carbohydrates, which caused the cells to aggregate and consequently may help protect host cells against infection. To our knowledge this is the first report showing that LL-37 can interfere with the adhesion of a fungal pathogen to human cells. These results suggest the potential for new therapeutic agents that can target the cell-wall carbohydrates of C. albicans and the use of AMPs to prevent C. albicans colonization and infection. Moreover, LL-37 might be used to screen for other adhesion molecules on the cell surface of C. albicans and/or other fungal pathogens.

LL-37 kills C. albicans
The ability of LL-37 to kill C. albicans was measured by a spot assay and a FUN-1 assay that incorporated flow cytometry (21). For the spot assay, cell viability after LL-37 treatment was visually compared with that of control cells (no LL-37 treatment). We found that cell viability was more sensitive to higher concentrations of LL-37 (20 to 40 mg/ml) as opposed to lower concentrations (5 and 10 mg/ml) or no treatment (Fig. 1A). To quantitatively measure cell mortality, FUN-1 staining was used to distinguish between dead and living cells. Concentrations of LL-37 $20 mg/ ml induced cell death. At 40 mg/ml LL-37, ,60% of the cells were killed (Fig. 1B). Therefore, LL-37 had a candidacidal effect on C. albicans at doses $20 mg/ml.

LL-37 affects adhesion of C. albicans to polystyrene
We found that after treatment with 5 mg/ml LL-37, cells were more easily centrifuged and did not adhere to centrifuge tubes to the extent that control cells did (Supplement Fig. S1). Cell adhesion is the first step during C. albicans colonization of host tissues and is required for infection. Adhesion is conferred by the binding of C. albicans adhesins to amino acid or sugar residues on the host cell surface [29]. Adhesins may also promote cell binding to abiotic substrates, e.g., plastic prostheses and catheters [29]. We therefore suspected that LL-37 might affect interactions between the C. albicans cells and the substratum upon which they can proliferate. To test this hypothesis, C. albicans were treated with sub-lethal doses of LL-37 (0.1-10 mg/ml), and the extent of C. albicans attachment to polystyrene dishes was measured using the 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5carboxanilide (XTT) reduction assay and the whole-cell enzymelinked immunosorbent assay (ELISA). The extent of attachment decreased with increasing LL-37 concentration ( Fig. 2A). Floating cells were collected and spotted onto YPD agar to assess cell viability. The number of colonies derived from these cells depended on the LL-37 concentration and increased between 0.1 and 10 mg/ml LL-37 (Fig. 2B), suggesting that the LL-37 concentrations used in the adhesion assays were not substantially candidacidal. Together, the results shown in Fig. 1 and 2A suggested that low doses of LL-37 inhibited the adhesion of C. albicans to polystyrene, that this effect was not directly related to the ability of LL-37 to kill the cells (Fig. 2B), and that mechanisms not involved in the candidacidal effect might be responsible for adhesion loss. We hypothesized that one such mechanism might involve a direct interaction between LL-37 and fungal cells, which was supported by the finding that the fungal cells aggregated after treatment with LL-37 (Fig. 2C). Morphological structures, e.g., the germ tube or hyphae, tend to cause aggregation of C. albicans cells. However, as seen in Fig. 2C, most aggregated cells were in the form of yeast. Therefore, LL-37, but not cell morphology, caused the cells to aggregate.
LL-37 induces cell aggregation and reduces cell attachment to polystyrene by directly binding to C. albicans To determine if LL-37 caused the cells to aggregate, which in turn caused them to detach from the plastic dishes, floating cells were collected after LL-37 treatment and characterized by optical microscopy. The extent of floating cell aggregation was directly proportional to LL-37 concentration (Fig. 3A).
Given the dose-dependent effect of LL-37 on cell aggregation, we hypothesized that LL-37 may directly bind to the C. albicans cell surface, thereby interfering with cell adhesion to polystyrene. To test this possibility, LL-37-binding assays were performed. C. albicans cells were treated with BA-LL37, and floating cells were collected. BA-LL37 bound to these cells as assessed by western blotting that used horseradish peroxidase-conjugated streptavidin (SA-HRP) for detection. Direct binding of LL-37 to the floating cells was obvious when the LL-37 concentrations were 10 and 20 mg/ml (Fig. 3B). However, smaller numbers of floating cells might have been the cause of the non-detectable LL-37 binding at the smaller doses. To test this possibility, we also assessed LL-37 binding using flow cytometry in conjunction with SA-4,6dichlorotriazinyl aminofluorescein (SA-DTAF) detection. For this system, the binding of BA-LL37 to C. albicans cells was observed at a dose of 1 mg/ml (Fig. 3C). These results strongly indicated that LL-37 directly bound to C. albicans and induced cell aggregation, which consequently removed the cells from the polystyrene.
To investigate if the inhibition of adhesion was LL-37 specific, two other types of AMPs, human b-defensin 3 (hBD-3) and histatin 5 (Hst 5), were tested using the adhesion assay. Hst 5 has a linear, non-helical structure, hBD-3 folds into a b-sheet, and the LL-37 conformation is helical [30]. All three AMPs have candidacidal activity [30]. Equivalent molar concentrations of these AMPs were used in all assays. After C. albicans cells were incubated with 2.12 mM BA-hBD3, BA-Hst 5, or BA-LL37 (2.12 mM BA-LL37 is ,10 mg/ml LL-37), XTT reduction assays and whole-cell ELISAs were performed. In comparison with the level of control-cell adhesion to polystyrene, BA-Hst 5 and BA-hBD3 reduced the adhesion level of C. albicans cells by ,10-30%, whereas BA-LL37 reduced adhesion by 60-80% (Fig. 4A). AMPinduced aggregation of floating cells was then examined by microscopy, and the number of aggregates was recorded. The extent of cell aggregation was greatest for BA-LL37-treated cells, and large aggregates were observed. Conversely, the BA-hBD3treated cells formed few aggregates, and those treated with BA-Hst 5 did not aggregate (Fig. 4B). Flow cytometry was also used to assess if the AMPs bound to the cells. According to the fluorescence intensities, cells bound by BA-hBD3 and BA-Hst 5 were only 10-15% of that found for BA-LL37-treated cells (Fig. 4C). To exclude the possibility that the inhibitory effects of the AMPs on cell adhesion were caused by cell death, cells were incubated with the same molar concentrations of the AMPs as used for the adhesion assay, and then cell viability was assessed using the spot assay. Cell death was not observed when cells were treated with BA-LL37 or BA-Hst 5, although almost half of the cell population was killed after BA-hBD3 treatment (Fig. 4D). In combination with the results of the cell-adhesion, cell-aggregation, and peptide-cell association assays (Fig. 4), these results indicated that, of the three AMPs tested, only LL-37 inhibited C. albicans adhesion to polystyrene by directly binding to cell surface and causing cell aggregation.

LL-37 binds to C. albicans cell-wall polysaccharides and thereby reduces C. albicans adhesion
The cell wall is the outmost layer of C. albicans and, as such, interacts with the environment and host cells [31]. The content of the C. albicans cell wall is ,80-90% carbohydrate and contains different carbohydrates [2]. LL-37 contains an XBBXBX motif (where X is a hydrophobic or uncharged residue, and B is a basic residue) that is responsible for the binding of LL-37 to glycosaminoglycans, e.g., heparin and dermatan sulfate [32]. To evaluate the role of carbohydrates in LL37-mediated inhibition of C. albicans adhesion, cell-wall carbohydrates were removed. Concanavalin A (Con A) specifically binds a-Dmannose and a-D-glucose, and BA-Con A was used to monitor the extent of deglycosylation. After deglycosylation, the amount of BA-Con A that bound C. albicans cells decreased by ,50%, which represents the removal of ,50% of the cell-wall carbohydrates. After removal of the carbohydrates, the binding of BA-LL37 to the C. albicans cells was reduced by ,30% compared with its binding to non-deglycosylated cells (Fig. 5A). Moreover, pull-down assays verified the LL-37/polysaccharide interaction. Since glucan and chitin in solutions will be insoluble, these two polysaccharides and mannan-agarose were used to associate with LL-37 and to pull LL-37 down. Mannan-agarose, glucan, and chitin were each incubated with BA-LL37, after which unbound peptide was separated from the carbohydrates by centrifugation. LL-37 bound directly to these polysaccharides (Fig. 5B). The interaction of each polysaccharide with LL-37 was dose dependent.
To identify the types of carbohydrates targeted by LL-37 binding, the three monosaccharide compositions of the C. albicans cell wall, D-glucose, N-acetyl-D-glucosamine, and mannose [33], were each included in a cell-adhesion assay. Monosaccharides were pre-incubated with LL-37 to form monosaccharide/LL-37 complex, the inhibition effect of LL-37 on adhesion was thereby suppressed. As shown in Fig. 5C, D-glucose significantly rescued LL37-mediated adhesion inhibition while N-acetyl-D-glucosamine and D-mannose also possessed some potential to neutralize the effect of LL-37. We thus concluded that LL-37 might bind to monosaccharides to prevent cell adhesion of C. albicans (Fig. 5C). However, polysaccharides, not monosaccharides, account for .90% of the C. albicans cell-wall carbohydrates [34], and monosaccharides can not properly reflect structure and function of the cell wall. Therefore, polysaccharides were also tested in adhesion assays (Fig. 5D). In the absence of LL-37, chitin and Saccharomyces cerevisiae mannan had no significant effect on cell adhesion, whereas S. cerevisiae glucan slightly decreased and C. albicans mannan largely increased the level of adhesion the level of adhesion (Fig. 5D). This data was consistent with the finding that C. albicans mannan enhances C. albicans adhesion to plastic [35]. Moreover, no significant difference was shown by comparing cells with LL-37 and without LL-37 treatment, indicating that LL37mediated adhesion inhibition of C. albicans was abolished in the presence of each polysaccharide. If we considered the effect caused only by LL-37 but not polysaccharedes (i.e. comparing only between the black columns of Fig. 5D), C. albicans mannan exhibited the highest extent to relieve LL37-mediated adhesion inhibition, followed by chitin (Fig. 5D). S. cerevisiae mannan slightly relieved the adhesion inhibition caused by LL-37 (Fig. 5D). Importantly, the aggregation of floating cells induced by LL-37 was largely reduced by C. albicans mannan (Fig. 5E). The aggregations were moderately reduced by chitin but S. cerevisiae mannan and glucan had only a minor effect on the floating cell aggregation (Fig. 5E). The dissociation constant (K d ) values for the C. albicans and S. cerevisiae mannan/LL-37 complexes were further determined using an Affinity Detection system. The K d values were the average of at least two independent experiments. The K d for the LL-37/C. albicans mannan complex was 0.460.1 mM, and that for the LL-37/S. cerevisiae mannan complex was 4.960.4 mM; this 12-fold decrease in binding strength indicated that LL-37 preferentially binds C. albicans mannan. Together, these results demonstrated that LL-37 binds the polysaccharides of the C. albicans cell wall, especially mannans, thereby promoting cell aggregation and reducing cell adhesion. Effects of LL-37 on C. albicans adhesion to oral epidermal cells and the urinary bladders of mice To assess the in vivo effects of LL-37 on C. albicans adhesion, we examined the ability of C. albicans to adhere to oral epidermal cells and to a murine urinary tract model [36]. C. albicans is involved in human oral infections [37], and LL-37 has been shown to diminish the effects of microbial infections in the oral cavity [14]. To determine the extent to which C. albicans can adhere to oral epidermal OECM-1 cells, a cell ELISA was performed. After 30 min pre-incubation with LL-37, the cells were added to the OECM-1 cells and incubated for half an hour. The adhesion of the LL37-treated cells was significantly reduced compared with control cells (Fig. 6A). For C. albicans treated with 20 mg/ml LL-37, cell adhesion was reduced by .70% (Fig. 6A).
Candida species are often isolated from the urine of patients suffering from microbial infection in the bladder or urinary tract [38]. To determine the effects of LL-37 on C. albicans adhesion within the urinary tract, we used a mouse model. LL-37 (5 mg/ml) and C. albicans were mixed and then injected into the urinary tract.
An hour after injection, the number of C. albicans cells that had adhered to the bladders was quantified as colony-forming units present on agar. The number of C. albicans cells isolated from the bladders when LL-37 was present was significantly less than that of the control (Fig. 6B; p = 0.0011). Therefore, LL-37 inhibited C. albicans adhesion both in vitro and in vivo.

Discussion
Human LL-37 is a multifunctional AMP commonly found at mucosal surfaces at a concentration of ,2-5 mg/ml [39,40]. However, its local concentration can be 30-1,500 mg/ml in infected or inflamed tissue [41,42]. Recently, the effects of LL-37 on bacterial infections have been the focus of much research. In the presence of pathogens, human lung epithelia respond by rapidly increasing LL-37 secretion onto airway surfaces, which subsequently kills nearby microbes [16]. Bacterial contact with urinary tract epithelial cells causes rapid production and secretion of human LL-37 into the urine [36]. Although epithelium-derived LL-37 substantially protects the urinary tract against infection  [36], the molecular basis of this host defense mechanism has not been thoroughly characterized.
For the study reported herein, in addition to its known ability to kill C. albicans, we demonstrated that human LL-37 can prevent C. albicans adhesion to abiotic and biotic substrates. Similarly, LL-37 has been shown to prevent Escherichia coli from adhering to the urinary tract of mouse [36] and to inhibit the attachment and biofilm formation of Pseudomonas aeruginosa and Staphylococcus epidermidis [41,43] on polystyrene. To our knowledge, however, our report is the first to show that LL-37 blocks the adhesion of a fungal pathogen to substrates. We found that physiological concentrations of LL-37 prevented the adhesion of C. albicans to abiotic surfaces in vitro and epithelial surfaces in vivo. The effects appeared to be specific to LL-37 because Hst 5 and hBD-3, two other candidacidal AMPs, did not inhibit cell adhesion or bind to C. albicans to the same degree ( Fig. 4A and 4C). LL-37, hBD-3, and Hst 5 have 6, 11, and 12 positive charges, respectively, at physiological pH [30]. In addition, recombinant LL-37 has a propensity to aggregate during purification [44]. Therefore, it is feasible that aggregation of the floating cells was initiated by the direct binding of LL-37 to the cell surfaces and allowed for interaction between LL-37 molecules on adjacent cells (Fig. 4B). At the same molar dose, BA-hBD3 much effectively killed C. albicans than BA-LL37 and BA-Hst 5 did (Fig. 4D). Together, these data suggest that the ability of LL-37 to inhibit C. albicans adhesion was unrelated to its positive charge or killing activity, but instead was caused by its interaction with the C. albicans cell surface, which resulted in cell aggregation and their subsequent inability to bind supporting substrata.
To identify the components of the C. albicans cell wall that interact with LL-37, we found that mannan, chitin or glucan and LL-37 associate. The importance of these associations in microbial infection has been reported. Foschiatti and colleagues showed that bacteria use exopolysaccharides to bind and segregate AMPs, thereby reducing the efficiency of the primary innate host defense [45]. Bergsson and colleagues showed that LL-37 binds to glycosaminoglycans from lungs afflicted with cystic fibrosis, which inhibits the ability of LL-37 to kill Pseudomonas [46]. In our study, LL-37 association with C. albicans cell-wall carbohydrates, in particular mannans, decreased cell adhesion to substrata both in vivo and in vitro. This ability is crucial for C. albicans biofilm formation on an abiotic surface and for colonization on host tissue during infection [10]. Interestingly, LL-37 bound to chitin, glucan, and S. cerevisiae mannan, but only C. albicans mannan completely rescued LL37-mediated cell aggregation and inhibition of cell adhesion (Fig. 5B, 5D, and 5E). The major difference between the two mannans is that C. albicans mannan contains b-1,2 linkages whereas S. cerevisiae mannan does not [47]. b-1,2-linked mannans act as the adhesins for C. albicans/buccal epithelium complex formation [35], and when the complex was recognized by galectin-3, C. albicans death occurred [48]. In an ex vivo assay, a monoclonal antibody that recognized b-1,2-mannans inhibited C. albicans attachment to murine splenic marginal-zone macrophages [49]. Dromer and colleagues demonstrated that synthetic b-1,2oligomannosides analogs prevented intestinal colonization by C. albicans [50]. Our results indicated that the reduced cell adhesion caused by LL-37 was specifically inhibited by C. albicans mannan, suggesting that the association between LL-37 and that particular mannan may involve the latter's b-1,2-linkages. This hypothesis requires further study. Because higher eukaryotic cells do not have a cell wall and because LL-37 exerts its effects by binding to cellwall components, these components may be drug candidates for Candida infection.
Although the polysaccharides tested in our study all interacted with LL-37 (Fig. 5B), when the cells were deglycosylated, BA-LL37 reduced its binding to cell by only ,30% even though ,50% of the carbohydrates had been removed (Fig. 5A). Therefore, we cannot exclude the possibility that LL-37 may also interact with other cell-wall components, e.g., proteins. Proteins represent ,20-30% of the fungal cell-wall mass [11]. Specifically, mannan is covalently bound to proteins (mannoproteins), and this form of mannan accounts for ,40% of the total cell-wall polysaccharides on the C. albicans exterior [51]. Mannoproteins are involved in cellcell recognition and trigger immune responses [52]. Several studies have showed that AMPs function via their interactions with microbial proteins. The protein, F. magna adhesion factor, of the anaerobic bacterium Finegoldia magna, blocks the cytocidal activity of LL-37 and promotes F. magna colonization and survival in humans [53]. hBD-3 binds to immobilized recombinant hemagglutinin B, a non-fimbrial adhesin from Porphyromonas gingivalis  [54,55], which may therefore prevent bacterial adhesion to host tissues [55]. The C. albicans proteins possibly targeted by LL-37 and the significance of such LL-37/protein interactions are currently under investigation in our laboratory.
The glucan and chitin layers of the C. albicans cell wall are buried beneath a thin but dense mannan layer [11]. We found that LL-37 interacts with mannan, chitin, and glucan (Fig. 5B). Moreover, compared with mannan, chitin and glucan were less able to rescue LL37-mediated cell aggregation and its ability to inhibit cell adhesion (Fig. 5D and 5E). Although this result supports previous findings that mannan is the most important C. albicans adhering glycan [56] and that loss of phosphomannan correlates with reduced levels of the AMP dermaseptin S3 bound to the cells [57], it also raises the possibility that LL-37 interferes with C. albicans cell-wall remodeling. It is thought that some proteins with enzymatic activity on the C. albicans cell surface could be bound by LL-37, leading to dysfunction of carbohydrate metabolism [58], impairment of cell-wall construction, and prevention of C. albicans/host adhesion. C. albicans cell-wall remodeling occurs during infection and drug (e.g., caspofungin) treatment, and consequently the inner glucan layer is exposed during remodeling [59,60]. Characterization of the effects caused by LL-37 on cell-wall remodeling will be investigated in future studies.
Finally, AMP-mediated cellular aggregation has been reported. The AMP L8 causes cellular aggregation of E. coli and Listeria monocytogenes, two important food-borne pathogens [61]. In addition, human neutrophil defensins induce aggregation of E. coli and Staphylococcus aureus and may, in part, increase bacteria uptake by neutrophils [62]. Similarly, human b-defensins bind bacterial components and thereby enhance microbial ingestion by dendritic cells [63]. We found that in addition to the association of LL-37 and C. albicans and the reduction in cell adhesion to polystyrene, floating cells aggregated-mediated by LL-37 bound to the cells (Fig. 2 and 3). Although the biological function of cell aggregation is not clear, we hypothesize that LL37-mediated aggregation may help C. albicans escape further attack by LL-37 or may allow phagocytes to easily recognize the pathogen.
In summary, our results revealed that LL-37 prevents C. albicans adhesion to polystyrene and tissues, in part by interacting with cellwall carbohydrates. Our study highlights the potential use of LL-37 as an effective therapeutic or preventive agent against C. albicans infection. C. albicans polysaccharides could also be targeted by monoclonal antibodies or short peptides to block cell adhesion during infection.
This culture was then sub-cultured in YPD broth for ,2.5 h until logarithmic-phase growth was reached. For LL-37 treatment, cells were washed twice with phosphate-buffered saline (PBS), collected by centrifugation, and suspended in Gibco RPMI-1640 medium (Invitrogen, Carlsbad, CA) or in PBS for LL-37-binding experiments. C. albicans cell-wall carbohydrates were removed with GlycoPro TM Enzymatic Deglycosylation kit reagents (Prozyme, San Leandro, CA) according to the manufacturer's instructions.

Assays for LL-37 candidacidal activity
To assess LL-37 candidacidal activity, a spot assay and the FUN-1 staining assay were used. After a 30-min incubation, LL-37-treated cells were collected and their solutions were subjected to 10-fold serial dilution. Cells (10 ml) were spotted onto YPD agar plates. Cell viability was assessed after incubation at 30uC for 24 h.
FUN-1 staining was performed using LIVE/DEADH Yeast Viability kit reagents (Molecular Probes, Eugene, OR) according to the manufacturer's instructions and (21). Briefly, after LL-37 treatment for 30 min, C. albicans cells were harvested by centrifugation and suspended in 500 ml of 10 mM HEPES containing 2% glucose and 5 mM FUN-1 for 30 min at 30uC in the dark. Samples were then examined using a FACSCalibur analyzer (BD Bioscience, San Jose, CA) equipped with a diode laser (excitation wavelength 488 nm). The intensities of the yellowgreen fluorescence emission from cells that had or had not been treated with LL-37 were recorded after the light had passed through an FL1 filter (515-545 nm). A total of 10,000 cells were acquired per experiment. The mean fluorescence index of the LL37-treated cells was compared with that of the control cells (no LL-37 treatment). The relative number of dead cells is reported as a percentage.

C. albicans adhesion to polystyrene
The adhesion of C. albicans was assayed using uncoated, flatbottom, 24-well plates (Orange Scientific, Braine-I'Alleud, Belgium) or uncoated, 96-well plates (Nunc TM , Rochester, NY) as described [65,66]. For the XTT reduction assay, C. albicans cells were harvested, washed, and suspended in RPMI-1640 medium at a concentration of ,6610 7 cells/ml. Next, 250 ml of each cell suspension was transferred into a well of a 24-well plate. Different concentrations of LL-37, BA-LL37, BA-hBD3 or BA-Hst 5 were independently added to the cell suspensions that were then incubated at 37uC for 30 min with shaking (100 rpm). The metabolic activities of the sessile cells were measured by detecting the reductive adduct of XTT [67]. Briefly, cells were washed three times with PBS to remove floating cells. Adherent cells were incubated with 300 ml XTT (1 mg/ml) and 0.6 ml menadione (1 mM) in PBS per well at 37uC for 20 min. Absorbance at 490 nm was determined using a VICTOR3 Multilabel Plate Reader (PerkinElmer). For the competition assays, LL-37 was premixed with 400 mM of a monosaccharide or 1 mg/ml of a polysaccharide in RPMI-1640 for 30 min at 4uC before incubating the LL-37/ polysaccharide mixture and the cells. All carbohydrate stock solutions were prepared in PBS. C. albicans cell-wall mannan was obtained from Takara Bio Inc. (Otsu, Japan). Mannan from S. cerevisiae (M7054), glucan from S. cerevisiae (G5011), and chitin from crab shells (C9752) were purchased from Sigma-Aldrich.
Cell ELISAs were performed using a similar procedure, except that 100 ml of 5610 7 cells/ml were used for each assay. The adherent cells were fixed with 3% paraformaldehyde and labeled with FITC-conjugated rabbit polyclonal antibody against C. albicans (Biodesign International, Saco, ME). Fluorescence was detected using a VICTOR3 Multilabel Plate Reader. The relative number of adherent cells was calculated as: [(mean absorbance or fluorescence for each treated cell sample)/(mean absorbance or fluorescence for the control sample)]6100%. All assays were performed in triplicate and repeated three or four times.

Microscopic visualization of floating C. albicans cells
After incubation with different LL-37 concentrations in the wells of 24-well plates, the floating C. albicans cells were collected and injected into m-Slides VI flat (Ibidi, Martinsried, Germany). Samples were analyzed with a Zeiss light microscope.
Assays for LL-37 binding to floating C. albicans cells C. albicans cells were suspended in PBS, mixed with BA-LL37 in the wells of 24-well plates, and incubated at 4uC for 30 min. Floating cells were collected by centrifugation, suspended in reducing sample buffer, and denatured by heating at 100uC for 10 min. The samples were separated through a 15% Tricine SDS-PAGE gel (prepared with a 40% acrylamide/Bis solution; MDBio, Inc., Taipei, Taiwan), and transferred onto a polyvinylidene difluoride (PVDF) membrane (Pall Corporation, Port Washington, NY). BA-LL37 binding to C. albicans was detected using SA-HRP according to the manufacturer's instructions (Zymed Laboratories, San Francisco, CA). The blots were processed using ECL kit reagents (PerkinElmer Life Sciences).

Binding of antimicrobial peptides to C. albicans cells
AMPs bound to C. albicans were detected as described [48]. A total of 6610 6 cells were mixed with different concentrations of BA-LL37 or with 2.12 mM BA-LL37, BA-hBD3, or BA-Hst 5 in 750 ml PBS and then incubated at 4uC overnight. Cell-bound peptide was assessed by flow cytometry with SA-DTAF detection (3 mg/reaction; Jackson ImmunoResearch, West Grove, PA). Reactions were quantitated using a FACSCalibur flow cytometer. Fluorescence data for 10,000 cells were acquired per experiment. The relative amount of an AMP bound to the cells (as a percentage) was obtained by comparing the mean fluorescence intensity emitted from the BA-hBD3-or BA-Hst 5-treated cells with that of BA-LL37-treated cells. Additionally, the relative levels of BA-LL37 and BA-Con A bound to cells that had or had not been deglycosylated were determined.

Binding of LL-37 to polysaccharides
Polysaccharide binding assays were performed as described [68] with modifications. Mannan-agarose (M9917; Sigma-Aldrich), glucan, and chitin were each incubated with 10 mg BA-LL37 in 750 ml PBS with gentle rotation at 4uC overnight. The samples were pelleted and washed twice with 1 ml PBS. The pellets were suspended in reducing sample buffer and denatured by heating at 100uC for 10 min. The polysaccharide-bound BA-LL37 samples were electrophoresed through a 15% Tricine SDS-PAGE gel, transferred to a PVDF membrane, and detected using SA-HRP and ECL kit reagents according to the manufacturer's instructions.

Measurement of dissociation constants for LL-37/ mannan complexes
Dissociation constants (K d ) for the LL-37/mannan complexes were measured using quartz crystal microbalance methodology [69]. The amine surfaces of gold electrodes on a quartz chip were purchased from ANT Technology Co., Ltd. (Taipei, Taiwan) and were activated with 2.5% glutaraldehyde. Mannans from S. cerevisiae and C. albicans (100 mg/ml each) were each cross-linked with glutaraldehyde and then coated onto a chip. The binding assay was performed using an Affinity Detection System (ANT Technology). Before each experiment, the system was equilibrated with PBS (pH 7.4) at the flow rate of 50 ml/min. For K d measurements, the LL-37 samples (0.1 to 1.6 mM for mannan from C. albicans; 1 to 10 mM for mannan from S. cerevisiae) were individually injected onto a chip surface that contained one of the yeast cell-wall mannans. All the experiments were performed in duplicate (for mannan from C. albicans) or triplicate (for mannan from S. cerevisiae). The K d values were calculated using Prism 5.0 (GraphPad Software Inc., La Jolla, CA).
To determine how well C. albicans adhered to the oral epithelial cells, a protocol similar to that of the cell ELISA experiment was used, except that 5610 4 OECM-1 cells were first grown to 95% confluency in wells of a 96-well plate. Thirty minutes prior to the start of the experiment, 100-ml samples of the cells (1610 7 cells/ ml) were prepared in RPMI-1640 that contained different concentrations of LL-37 at 4uC. Then, the treated C. albicans cells were incubated with the OECM-1 cells at 4uC for another 30 min. The percentage of adhered cells was calculated as follows: [(mean fluorescence for LL-37-treated cells)/(mean fluorescence for control cells)]6100%. All assays were performed in quadruplicate and repeated four times.

Mouse model for C. albicans adhesion
All animal studies were approved by the Institutional Animal Care and Use Committees of the National Tsing Hua University (approval number 09730) and the Animal Technology Institute Taiwan (approval number 97010). Adhesion of C. albicans to mouse bladder epithelia was assayed as described [36] with modifications. Briefly, 20-to 24-week-old female BALB/c mice (BioLasco Taiwan Co., Ltd.) were deprived of water overnight. Mice under anesthesia were each transurethrally catheterized with 100 ml of a LL-37 (5 mg/ml) and C. albicans mixture (2610 7 cells) using soft, sterile polyethylene tubing (PE 10, outside diameter 0.61 mm, inside diameter 0.28 mm, Clay Adams, Becton Dickinson) that had been lubricated. After 1 h of catheterization, the mice were sacrificed and their urinary bladders were removed. The bladders were washed four times with 1 ml PBS and homogenized using a tissue grinder. The number of colonyforming units of C. albicans cells that adhered to each bladder was determined by plating serially diluted samples (in PBS) onto YPD agar that contained 50 mg/ml ampicillin and incubating the plates at 30uC overnight. Relative cell adhesion was calculated as follows: [(number of colony-forming units of each bladder/(average number of colony-forming units of all bladders]6100%. Differences in the adhesion for C. albicans-injected mouse bladders were determined using GraphPad Prism 5.0 software.

Statistical analysis
Data were assessed for statistical significance by the two-tailed Student's t-test. Figure S1 Adhesion of LL-37-treated cells to centrifugation tubes. C. albicans was treated with 5 mg/ml LL-37 for 30 min, pelleted by centrifugation, and then the number of cells adhering to the tubes was compared with that of untreated cells by visual inspection. Upper panels: 50-ml centrifuge tubes; lower panels: 1.5-ml microcentrifuge tubes; left panels: cells after centrifugation; right panels: cells after removal of the supernatant; left tube in each panel: control cells (no LL-37 treatment); right tube in each panel: cells treated with 5 mg/ml LL-37. (TIF)