An α-Glucan of Pseudallescheria boydii Is Involved in Fungal Phagocytosis and Toll-like Receptor Activation*

The host response to fungi is in part dependent on activation of evolutionarily conserved receptors, including toll-like receptors and phagocytic receptors. However, the molecular nature of fungal ligands responsible for this activation is largely unknown. Herein, we describe the isolation and structural characterization of an α-glucan from Pseudallescheria boydii cell wall and evaluate its role in the induction of innate immune response. These analyses indicate that α-glucan of P. boydii is a glycogen-like polysaccharide consisting of linear 4-linked α-d-Glcp residues substituted at position 6 with α-d-Glcp branches. Soluble α-glucan, but not β-glucan, led to a dose-dependent inhibition of conidia phagocytosis. Furthermore, a significant decrease in the phagocytic index occurred when α-glucan from conidial surface was removed by enzymatic treatment with α-amyloglucosidase, thus indicating an essential role of α-glucan in P. boydii internalization by macrophages. α-Glucan stimulates the secretion of inflammatory cytokines by macrophages and dendritic cells; again this effect is abolished by treatment with α-amyloglucosidase. Finally, α-glucan induces cytokine secretion by cells of the innate immune system in a mechanism involving toll-like receptor 2, CD14, and MyD88. These results might have relevance in the context of infections with P. boydii and other fungi, and α-glucan could be a target for intervention during fungal infections.

Pseudallescheria boydii is a saprophytic fungus widespread in soil and polluted water and has recently emerged as an agent of localized as well as disseminated infections in both immunocompromised and immunocompetent hosts. Clinical observations and data obtained from experimental models indicate the essential role of innate immunity in resistance against infections caused by pathogenic fungi. Macrophages provide a major line of defense against fungal cells by ingesting and killing conidia by oxidative and non-oxidative processes (1)(2)(3)(4). Innate immunity recognition causes the release of pro-inflammatory mediators that induce recruitment of polymorphonuclear leukocytes (5). Failure of fungicidal activity of macrophages and neutrophils permits germination of conidia in hyphae with subsequent tissue invasion (6).
Innate immunity performs pathogen surveillance by detection of pathogen-associated components through germ lineencoded receptors that are expressed in resident leukocytes. Mammalian toll-like receptors (TLRs) 4 are a family of closely related transmembrane proteins, first identified as homologues of the Toll receptor in Drosophila (7,8). TLRs mediate the recognition of a large array of molecules present in pathogens, triggering the production of pro-inflammatory cytokines, activation of microbicidal mechanisms, and the induction of adaptive immunity (9 -11). TLR activation initiates a signaling cascade through a conserved pathway shared by IL-1R and IL-18R that requires adaptor proteins such as MyD88 leading to NFB activation and the induction of different pro-inflammatory genes (12). A great variety of pathogen molecules have been described to signal through TLRs, especially for TLR2 and TLR4, the best characterized TLRs. TLR2 recognizes lipoteichoic acid, bacterial lipopeptides, mycobacterial lipoarabinomannans, and glycosylphosphatidylinositol anchors of protozoan parasites (13)(14)(15)(16). TLR4 recognizes bacterial LPS and cytolysins of Gram-positive bacteria (17)(18)(19).
Recent studies have demonstrated the involvement of TLRs in the recognition of fungal pathogens such as Aspergillus fumigatus and Candida albicans. A. fumigatus induces cytokine release as well as NFB activation through TLR2 and TLR4 activation (20,21). Genetic deficiency of TLR4 makes mice more susceptible to an experimental A. fumigatus infection following immunosuppression, and TLR2-and TLR4-deficient mice present an increased fungal load in the lungs upon A. fumigatus intranasal challenge (22,23). However, the nature of the pathogen-associated molecular patterns expressed by A. fumigatus that trigger the synthesis of TNF-␣, IL-12, and macrophage inflammatory protein-2 remains to be established. Resistance to experimental infection with C. albicans requires TLR2 and TLR4 as observed by the increased fungal load on the kidneys of TLR4 mutant mice, C3H/HeJ, and the higher susceptibility of TLR2 Ϫ/Ϫ mice to C. albicans infection (24,25). In addition, C. albicans induces cytokine release through TLR2 and TLR4 (24,25). Although the results demonstrate that TLR2 and TLR4 participate in the recognition of pathogenic fungi, the molecules that trigger the activation of the TLR-associated signaling pathway leading to the induction of the innate immune response are largely unknown.
In the present study we describe the structural characterization of an ␣-glucan, a glycogen-like polysaccharide extracted from the P. boydii cell wall, and evaluate its role in the induction of innate immune response. The ␣-glucan from P. boydii is essential to conidial phagocytosis by macrophages and induces cytokine secretion by cells of the innate immune system in a mechanism involving TLR2, CD14, and MyD88.

EXPERIMENTAL PROCEDURES
Mice-C57/BL6, BALB/c, and C57BL/10 mice were obtained from the Fundação Oswaldo Cruz Breeding Unit (Rio de Janeiro, Brazil). C57BL/10ScN mice were obtained from the Universidade Federal do Rio de Janeiro, and C57BL/10ScCr mice were obtained from the Universidade Federal Fluminense (Rio de Janeiro, Brazil). MyD88, TLR2, and CD14 knock-out mice (on a C57BL/6 background) were kindly provided by Dr. Akira from the Research Institute for Microbial Diseases, Osaka University, Osaka, Japan, and Dr. Douglas Golenbock from the University of Massachusetts, Worcester, MA.
Reagents-LPS (O111:B4), zymosan, and thioglycolate were purchased from Sigma, and Pam 3 Cys-Ser-(Lys) 4 (Pam 3 Cys) was obtained from EMC Microcollections (Tübingen, Germany). Laminarin from Laminaria digitata was kindly supplied by Prof. Michael Noseda from the Departamento de Bioquímica, Universidade Federal do Paraná, Brazil. Polymixin B was obtained from Bedford Laboratories (Bedford, OH). A detoxifying polymixin column was purchased from Cambrex. RPMI medium for macrophage culture was obtained from Sigma and was supplemented with penicillin and streptomycin (100 IU/ml and 100 g/ml, respectively) obtained from Invitrogen.
Microorganism and Growth Conditions-P. boydii, isolated from eumycotic mycetoma, was kindly supplied by Bodo Wanke from Evandro Chagas Hospital, Instituto Oswaldo Cruz, Rio de Janeiro, Brazil. Cells were grown on Sabouraud solid slants, inoculated in liquid culture medium, and incubated for 7 days at 25°C with shaking. Cultures were then transferred to the same medium and incubated for 7 days at the same temperature with shaking; the mycelium was filtered, washed with distilled water, and stored at Ϫ20°C. For phagocytic assays, conidia were grown on Petri plates containing modified Sabouraud medium at 25°C. After 7 days in culture, conidia cells were obtained by washing the plate surface with phosphatebuffered saline and filtered through gauze to remove hyphae fragments and debris. Conidial suspensions were heat-killed by autoclaving the preparation at 120°C, for 15 min, washed, and counted.
Extraction and Fractionation of P. boydii Glucan-Hyphae of P. boydii were extracted with 2% KOH for 2 h at 100°C. The alkali extract was neutralized with glacial acetic acid and precipitated with 3 volumes of ethanol. The resulting precipitate was recovered by centrifugation, dialyzed against distilled water, and lyophilized. The crude polysaccharide was fractionated by gel-filtration chromatography on a Superdex S-200 column (30 ϫ 1.0 cm) coupled to an AKTA Purifier liquid chromatography (Amersham Biosciences) using phosphate buffer 0.01 M, 0.15 M NaCl, pH 7.0, as eluant pumped at 0.5 ml/min for 60 min. The elution profiles of the gel-filtration chromatography were monitored by refractive index detection, and the collected fractions were assessed for their carbohydrate content.
Sugar Analysis-Total carbohydrate was determined by the phenol-sulfuric acid method (26) and protein by the Folin phenol reagent method (27). Glucan (1 mg) was hydrolyzed with 2 M trifluoroacetic acid at 100°C for 3 h, and the solution evaporated to dryness. The resulting monosaccharides were characterized by high performance TLC and quantified by GC as alditol acetate derivatives (28) using a capillary column of OV-225 (30-m ϫ 0.25-mm inner diameter), with temperatures programmed from 50 to 220°C at 50°C/min.
Methylation Analysis-Methylation analysis was carried out by the method of Tischer et al. (29), using a modification of the method of Ciucanu and Kerek (30). Glucan (1 mg) was dissolved in a drop of H 2 O, which was diluted with Me 2 SO (ϳ1 ml) and then MeI (1 ml). Powdered NaOH (0.3 g) was added, and the mixture was agitated vigorously with a vortex for 30 min and then left overnight. After neutralization with HOAc, the product was then extracted with CHCl 3 and washed three times with H 2 O. Upon evaporation, the resulting per-O-methylated product was converted into partially O-methylated alditol acetates by successive treatments with 3% MeOH-HCl for 2 h at 70°C, 0.5 M H 2 SO 4 for 14 h at 100°C, reduction with NaBD 4 , and acetylation with Ac 2 O-pyridine. The products were analyzed by GC-MS on a capillary column of DB-225 (31), programmed from 50°C (1 min) at 40°C/min to 210°C (constant temperature).
Treatment with Yeast ␣-Amyloglucosidase-Glucan (1 mg) was incubated with 1.5 mg of yeast ␣-amyloglucosidase in 0.05 ml of 0.05 M sodium acetate buffer (pH 4.8) at 37°C for 22 h. At the end of the incubation period, the mixture was heated at 100°C for 5 min to inactivate the enzyme, centrifuged, and the supernatant was concentrated and applied to a TLC plate. This was eluted with n-BuOH-Me 2 CO-H 2 O (4:5:1, v/v) and developed with 0.05% (w/v) orcinol in 10% H 2 SO 4 at 100°C for 10 min, using 10 g of glucose as standard.
Endotoxin Removal-Glucan (2 mg) was dissolved in 1 ml of apyrogenic saline and then applied to an immobilized polymixin B gel column (Detoxi-Gel TM Endotoxin removing gel, Pierce), according to the manufacturer's instructions.
NMR Spectroscopy-NMR spectra at 500 MHz ( 1 H) and 125 MHz ( 13 C) were recorded, using a Varian INOVA spectrometer, for a sample (10 mg) of the glucan in D 2 O, at 60°C. Chemical shifts were relative to internal trimethylsilylpropionic acid-d 4 sodium salt (TPS) at 0 ppm ( 1 H) and external tri-methylsilane (TMS) at 0 ppm ( 13 C). Two-dimensional spectra (COSY, TOCSY, and heteronuclear single quantum coherence) were performed using the pulse sequences supplied by the instrument manufacturer.
Phagocytic Assay-Elicited peritoneal macrophages were obtained by the intraperitoneal instillation of 2 ml of 3% sterile thioglycollate. After 4 days, mice were sacrificed, and the peritoneal macrophages were harvested and washed with chilled HBSS, and plated. Elicited macrophages (2 ϫ 10 5 cells/ml) were cultured over round glass coverslips (13 mm) in 24-well flat bottom microtest plates. Adherent monolayers were challenged with 500 l of suspensions of heat-killed conidia containing 10 6 cells/ml. After incubation at 37°C in 5% of CO 2 for 60 min in RPMI 1640 medium, the cells were rinsed with HBSS for removal of non-internalized conidia. The preparations were fixed in Bouin's fixative and stained with Giemsa. The influence of ␣-glucan on conidia phagocytosis was evaluated by adding different concentrations of the polysaccharide (25, 50, and 100 g/ml) and of the polysaccharide (100 g/ml) after digestion with ␣-amyloglucosidase to the cultures simultaneously with the addition of conidia. The influence of a different purified glucans in the phagocytosis of conidia was also tested by adding 100 g/ml selected glucans to the cultures. To determine the phagocytic indexes (PIs), 200 cells were counted and the percentage of cells that ingested at least one particle was multiplied by the mean number of internalized particles (32).
Phagocytosis of Zymosan Particles-Macrophage monolayers were challenged with 500 l of suspensions of zymosan particles (10 6 particles/ml). After incubation at 37°C in 5% of CO 2 for 60 min in RPMI 1640 medium, the cells were rinsed with HBSS for removal of non-internalized particles. The preparations were fixed in Bouin's fixative and stained with Giemsa. The influence of ␣-glucan and laminarin on the phagocytosis of zymosan was carried out by adding 100 g/ml of each glucan, simultaneously to zymosan particles.
Macrophage Culture and Stimulation-Elicited peritoneal macrophages were obtained by the intraperitoneal instillation of 2 ml of 3% sterile thioglycollate. After 4 days, mice were sacrificed and the peritoneal macrophages were harvested, washed with chilled HBSS, and plated at a density of 2 ϫ 10 5 cells/well in a 96-well plate. The plate was incubated for 2 h at 37°C in 5% of CO 2 . Non-adherent cells were removed by washing with HBSS. Adherent cells were stimulated for 4 h, in RPMI medium, with the ␣-glucan, LPS, or Pam 3 Cys, at concentrations indicated in the figure legends. After this period the supernatant was recovered for TNF determination by ELISA according to the manufacturer's instructions. Polymixin B (1 or 10 g/ml) was added 5 min before the stimulation with ␣-glucan, to rule out the possibility that the stimulating activity was due to contaminating lipopolysaccharides.
Generation and Stimulation of Murine Bone Marrow-derived Dendritic Cells-Dendritic cells were generated as previously described (33), with some modifications. Briefly, bone marrow was harvested from the tibia and femur of C57/Bl6 or TLR2 Ϫ/Ϫ mice. The cells were resuspended at 10 6 per milliliter in RPMI 1640 (Sigma) supplemented with vitamins, amino acids, 50 M 2-mercaptoethanol, recombinant murine granulocyte macrophage-colony stimulating factor, and recombinant murine IL-4 at 10 ng/ml. After 5 days, fresh medium was added to culture and with 7 days of culture, the cells were collected and bone marrow-derived dendritic cells were separated by Optiprep gradient (Sigma) by centrifugation at 600 ϫ g for 30 min at 24°C. This protocol generated Ͼ75% of CD11cϩ cells. Bone marrow-derived dendritic cells were plated in 96 wells at a density of 2 ϫ 10 5 /well and incubated for 16 h with the stimuli, after which the supernatant was recovered for the determination of cytokines by ELISA.
Statistical Analysis-Statistical analysis was performed using the statistical software SPSS for Windows (Version 10.0.1, SPSS Inc., 1989 -1999). Statistical differences among the experimental groups were evaluated by analysis of variance with Newman-Keuls correction or with the t test. Values are expressed as the mean Ϯ S.E. The level of significance was set at p Ͻ 0.05.

RESULTS
Isolation and Characterization of the Glucan from P. boydii-Mycelia of P. boydii were extracted with hot 2% aqueous potassium hydroxide at 100°C followed by neutralization with acetic acid and precipitation with ethanol. The crude precipitate was applied to a Superdex 200 column, which was eluted with a discontinuous gradient of aqueous NaCl. Carbohydrate-containing fractions F-1, F-2, and F-3 were obtained. Hydrolysis of F-1 followed by TLC examination of the products showed only glucose, confirmed by GC-MS analysis of derived alditol acetates. These results indicate that the F1 fraction contains a glucan. Methylation-GC-MS analysis of the glucan gave rise to partially O-methylated alditol acetates, which corresponded to non-reducing end units of Glcp (13%), 4-O-(63%), and 4,6-di-O-substituted Glcp units (24%) ( Table 1). The P. boydii glucan yields 2,3-di-O-methyl glucose, which is characteristic of branch points with glucose units in (136) linkage, and a high proportion of 2,3,6-tri-O-methylglucose, indicating the presence of linear portions of (134)-linked glucopyranosyl units.
The 13 C NMR spectrum of the ␣-glucan of P. boydii was partially assigned by means of a heteronuclear correlated spec-  ␣-Glucan of Fungi Activates the Innate Immune System AUGUST 11, 2006 • VOLUME 281 • NUMBER 32 trum (not shown) and is compared with corresponding published values in Table 2. A DEPT-135 spectrum confirmed the assignment of the resonance at 61.3 ppm in the 13 C spectrum to C-6, as this methylene carbon signal was inverted with respect to the methine ring carbon signals ( Fig. 2A). A small, broad 13 C resonance at about 69 ppm was also inverted and may be assigned to C-6 of 6-linked glucose, although no other signals from the branch point, 4,6-linked residues could be identified. Assignments of resonances from 4)-␣Glc-(134)-(residue type A) and terminal ␣Glc-(134)-(residue type B) were in good agreement with published values (Table 2); limited assignments could also be made for 4)-␣Glc-(136)-(residue type C). The ratio C-4 (A ϩ C):C-4 (B) ϭ 89:11, by integration of signals in the 13 C NMR spectrum, indicates that ϳ11% of residues are not glycosylated at the 4-position. These results indicate that the ␣-glucan of P. boydii is a glycogen-like polysaccharide consisting of linear 4-linked ␣-D-Glcp residues substituted at position 6 with -␣-D-Glcp branches. ␣-Glucan Has an Important Role in the Phagocytic Process of P. boydii by Macrophages-To investigate whether the ␣-glucan is involved in the phagocytic process of P. boydii, macrophages were incubated with heat-killed conidia at a ratio of 5:1 for 1 h in the presence or absence of ␣-glucan. P. boydii conidia were endocytosed by macrophages and heat-killed conidia thoroughly, and similar phagocytic indexes were obtained by challenging macrophages with live conidia (Fig. 3, A and B). The addition of increasing concentrations of ␣-glucan led to a dosedependent inhibition of P. boydii phagocytosis (Fig. 3C). The concentration of 100 g/ml ␣-glucan consistently caused a 50% inhibition of conidia phagocytosis. To exclude the possibility of contaminants being responsible for the inhibition of conidial phagocytosis, an ␣-amyloglucosidase-treated ␣-glucan (100 g/ml) was added to the culture at time zero of interaction. The phagocytic index of conidia returned to the control level when macrophages were allowed to interact with the digested ␣-glucan (Fig. 3D). To further characterize the role of ␣-glucan in the phagocytosis of P. boydii, conidia were submitted to treatment with ␣-amyloglucosidase for 22 h at 37°C, and their phagocytic index was compared with that from untreated conidia (Fig. 3E). A significant decrease in the phagocytic index occurred when ␣-glucan from conidial surface was removed by enzymatic treatment. These results suggest that ␣-glucan present in the P. boydii surface plays an essential role in the internalization of conidia by macrophages.
To define the selectivity of ␣-glucan in the phagocytosis of P. boydii by macrophages, we performed the phagocytic assays in the presence of purified ␣or ␤-glucans (Fig. 4A). A significant decrease of P. boydii phagocytosis occurred for the cultures treated with different ramified ␣-glucans, but not with pullulan, a linear ␣-glucan from the lichen Teloschistes flavicans, or laminarin, a ␤-glucan isolated from L. digitata. Recent studies have shown that ␤-glucan plays a central role in the phagocytosis of zymosan, and this effect is dependent on the phagocytic receptor Dectin-1 (38). As expected, soluble laminarin strongly reduced the ingestion of zymosan, whereas purified ␣-glucan had only a minor effect on zymosan phagocytosis (Fig. 4B).
␣-Glucan Induces TNF Release by Macrophages through TLR2 and CD14-To study the role of ␣-glucan in cytokine production, peritoneal macrophages were stimulated with increasing concentrations of ␣-glucan in the presence of polymixin B (Fig. 5). ␣-Glucan was able to induce TNF secretion by macrophages at the concentration of 100 g/ml. To further certify that TNF release induced by the ␣-glucan was not due to any possible contaminant in its preparation, we digested the ␣-glucan with ␣-amyloglucosidase. The ␣-amyloglucosidase treatment completely abolished the ␣-glucan stimulation of macrophages but not that induced by Pam 3 Cys or LPS, as evaluated by TNF release by macrophages (Fig. 5A). Moreover, treatment of macrophages with other ␣-glucans such as nigeran, pullulan, amilopectin, and glycogen did not induce TNF secretion (Fig. 5B). These results indicate that highly purified ␣-glucan from P. boydii is recognized by macrophages triggering fungal phagocytosis and TNF release. To investigate the involvement of TLR on ␣-glucan induction of macrophage activation, we stimulated wild-type and MyD88-deficient macro-

␣-Glucan of Fungi Activates the Innate Immune System
phages with ␣-glucan and evaluated TNF release. The secretion of TNF induced by ␣-glucan was abolished from MyD88 Ϫ/Ϫ macrophages (Fig. 6A). As controls we stimulated wild-type and MyD88-deficient macrophages with LPS and Pam 3 Cys, TLR4, and TLR2 ligands, respectively. As previously reported, TNF release triggered by these ligands was completely dependent on MyD88 (39,40). The requirement for MyD88 on macrophage activation induced by ␣-glucan indicates that a TLR is involved in the recognition of ␣-glucan. TLR2 and TLR4 are the best studied TLRs, and a great variety of molecules are potential ligands for these receptors. We investigated the role of these TLRs in the recognition of ␣-glucan, by stimulating macro-phages from mice lacking TLR4 (C57BL/10ScN) or TLR2 (TLR2 Ϫ/Ϫ ) and their respective counterparts. TLR4-deficient macrophages showed a partial but significant reduction in TNF secretion induced by ␣-glucan, whereas the response was completely abrogated in the absence of TLR2 Ϫ/Ϫ (Fig. 6, B and C).
As controls LPS and Pam 3 Cys were included in the experimental protocol. Similar results were obtained with the TLR4 mutant strains C57BL/10ScCr and C3H/HeJ (data not shown). Previous studies have shown the involvement of CD14 in the recognition of fungi by macrophages (20,41). To investigate the contribution of CD14 to the macrophage activation induced by ␣-glucan, wild-type and CD14-deficient macrophages were  AUGUST 11, 2006 • VOLUME 281 • NUMBER 32 stimulated by ␣-glucan, and TNF release was evaluated. Wildtype macrophages responded to ␣-glucan by the secretion of TNF, whereas CD14 Ϫ/Ϫ macrophages were unable to release TNF in response to ␣-glucan (Fig. 6D). LPS, a well known ligand of CD14, did not induce optimal TNF production at low concentrations in the absence of CD14. These results indicate that TLR2 and CD14 are essential to the recognition of ␣-glucan by the innate immune system.

Activation of Dendritic Cells by ␣-Glucan Is Dependent on TLR-2-Dendritic cells play a central role in naïve T cell activation and Th1
versus Th2 differentiation (42). The ability of dendritic cells to exert these functions is largely dependent on its previous activation by TLR ligands (11). Treatment of bone marrow-derived dendritic cells with ␣-glucan caused the release of IL-12 and TNF (Fig. 7, A and B). Similar to macrophages, the effect of ␣-glucan on cytokine secretion by dendritic cells was abolished in the absence of TLR2.

DISCUSSION
In this work, we described the structure of a highly purified ␣-glucan, obtained from P. boydii, that mediates P. boydii conidial phagocytosis and triggers macrophage activation in a mechanism involving CD14, TLR2, and MyD88. The immune response to the infections caused by fungi like A. fumigatus and P. boydii requires phagocytosis and killing of conidia with induction of a strong inflammatory response, preventing the development of hyphae and tissue colonization. P. boydii glucan structure was determined based on a combination of several techniques, including gas chromatography, 1 H TOCSY, 1 H and 13 C NMR spectroscopy, and methylation analysis, to be a glycogen-like polysaccharide consisting of linear 4-linked ␣-D-Glcp residues substituted at position 6 with -␣-D-Glcp branches. Like oyster and A. fumigatus glucan (34,43), the P. boydii glucan yields 2,3-di-O-methyl glucose, which is characteristic of branch points with glucose units in (136) linkage, and a high proportion of 2,3,6-tri-O-methylglucose, indicating the presence of linear portions of (134)-linked glucopyranosyl units. The 1 H NMR spectrum of the purified glucan confirmed the similarity of the glucan of P. boydii with glycogen from other species, including A. fumigatus, Mycobacterium bovis, and rabbit liver (34 -37).

␣-Glucan of Fungi Activates the Innate Immune System
We have demonstrated that conidial phagocytosis depends to a significant extent on ␣-glucan recognition. The inhibition of conidial phagocytosis by soluble ␣-glucan or by amyloglucosidase treatment indicates that ␣-glucan is accessible on the conidial surface and mediates its interaction with macrophages. In this way, ␣-glucan recognition resembles the role of ␤-glucans in the interaction of other fungi with macrophages. The phagocytosis of yeast forms of C. albicans and Saccharomyces cerevisiae is critically dependent on the recognition of surfaceassociated ␤-glucans by Dectin-1 expressed by macrophages (44,45). The phagocytosis of P. boydii by macrophages was not significantly affected by treatment with soluble ␤-glucan laminarin. This result further supports the concept that ␣-glucan is an important molecule involved in the recognition and phagocytosis of P. boydii and suggests that Dectin-1 is not essential for the recognition of P. boydii ␣-glucan. However, our results do not formally exclude a possible role of Dectin-1 in ␣-glucan recognition by cells of the innate immune system. Moreover, the partial effect of ␣-glucan on phagocytosis clearly indicates that other ligands are involved in the recognition of conidia by macrophages. Future analyses are required to define the phagocytic receptor involved on ␣-glucan recognition, as well as the nature of other putative phagocytic ligands present on P. boydii conidia.
Several studies have demonstrated that TLR2, TLR4, and CD14 are involved in the recognition of fungi (20,21,41). Although many details of TLR recognition and signaling in response to different developmental forms of fungal pathogens are well known, the molecules expressed by fungal cells that trigger TLR signaling by fungi are largely unknown. We demonstrated that ␣-glucan induces cytokine secretion by macrophages via CD14, TLR2, and MyD88. Interestingly, in the absence of TLR4 we observed a significant reduction of TNF secretion induced by ␣-glucan but not by Pam 3 Cys. A major concern in the identification of new putative TLR ligands is the presence of undesirable contaminants. To rule out contamination as the explanation for these results several strategies were used in the present study. First, a highly purified molecule was used as assessed by the analytical methods; second, the endotoxin was specifically removed by using a polymixin B column; and third, the ␣-glucan was digested with an ␣-amyloglucosidase that completely abolished the ␣-glucan stimulation of macrophages but not that induced by Pam 3 Cys or LPS (data not shown). Jouault et al. (46) also showed that a phospholipomannan of C. albicans requires TLR2 and to a lesser extent TLR4 and TLR6 to induce TNF release by murine peritoneal macrophages. On the other hand, A. fumigatus and zymosan induce macrophage activation, respectively, by TLR2/4 (20,21) and TLR2/Dectin-1 (45,47), but the molecules responsible for TLR triggering are unknown. A. fumigatus presents ␣-glucans as well as ␤-glucans in its cell wall, so it is possible that ␣-glucans could be the TLR2-activating molecules representing typical PAMPs of filamentous fungi like P. boydii.
A number of ␣-glucans from lichens and oyster were ineffective in inducing TNF secretion. Interestingly, pullulan was unable to affect the phagocytosis of conidia or cause TNF secretion. On the other hand, glycogen caused inhibition of phagocytosis of conidia to a similar extent to ␣-glucan from P. boydii despite not inducing TNF. These results indicate different requirements of putative phagocytic and TLR2 receptors on  Thioglycollate-elicited peritoneal macrophages were plated on 96-well plates and stimulated with the indicated concentrations of ␣-glucan, ␣-glucan (100 g/ml) digested with ␣-amyloglucosidase, LPS (100 ng/ml), or Pam 3 Cys (100 ng/ml). B, macrophages were stimulated with several glucans at 100 g/ml as indicated. After 4 h, the supernatant was recovered, and TNF was evaluated by ELISA. Data represent mean Ϯ S.E. of three different experiments, and the stimulations were performed in duplicates. Asterisks denote values statistically different from control and from ␣-amyloglucosidase-treated ␣-glucan (p Ͻ 0.05).
recognizing ␣-glucans. Pullulan and nigeran are linear molecules, whereas glycogen, amylopectin, and ␣-glucan from P. boydii present different grades of ramification. Our results suggest that the degree of ramification is important for the recognition by phagocytic receptors. In contrast, the induction of TNF release was triggered only by P. boydii ␣-glucan, the ␣-glucan with the higher degree of ramification. These results suggest that extensive ramification is required for TLR2 recognition by ␣-glucans. Interestingly, curdlan, a linear ␤-glucan from fungal cell wall, shows MyD88-dependent macrophage activation (48). In fact, for ␤-glucans, increasing the degree of ramification caused a reduction in macrophage activation. The activation of macrophages by mannuronic acid polymers is preferentially performed by TLR4 but also requires TLR2 (49). Interestingly, the potency of high molecular mannuronic acid polymers is reduced by polymeric breakdown, whereas attaching oligomeric M blocks enhances particles potency (50).
In recent years the central role of dendritic cells as the link between innate and adaptative response has become clear. These cells are responsible for the activation of naïve T lymphocytes, presenting antigen, providing co-stimulation and secreting polarizing cytokines (11,31). The secretion of IL-12 is essential to the induction of a Th1 phenotype. An effective immune response to fungi such as A. fumigatus, C. albicans, and Paracoccidioides brasiliensis requires a strong Th1 response and consequent interferon-␥ production (51)(52)(53). The ability of dendritic cells to recognize ␣-glucan through TLR2, secreting IL-12, and TNF suggests a possible mechanism for the induction of a protective Th1-polarizing response during P. boydii infection. Future studies are necessary to define the involvement of this pathway in the in vivo host response to P. boydii infection.
Here, we described an ␣-glucan that represents a characteristic PAMP of filamentous fungi. We also demonstrated that this molecule participates in the phagocytosis of conidia and that TLR2 and CD14 are involved in the innate immune activation upon the recognition of ␣-glucan. These results might have relevance in the context of infections with P. boydii and other fungi. Recognition of ␣-glucan could be a target for immunomodulation during fungal infections increasing the host resistance through IL-12 secretion and Th1 induction; alternatively, ␣-glucan could also contribute to the pathology by inducing local and systemic TNF release, promoting tissue injury. . ␣-Glucan is a ligand of TLR2 and CD14 requiring MyD88. A, ␣-glucan requires MyD88 to induce TNF secretion by macrophages. Thioglycollate-elicited peritoneal macrophages, WT or MyD88 Ϫ/Ϫ , were stimulated with ␣-glucan (100 g/ml), LPS (100 ng/ml), or Pam 3 Cys (100 ng/ml). After 4 h, the supernatant was harvested and TNF was measured by ELISA. Results are expressed as mean Ϯ S.E. of three different experiments, and stimulation was performed in duplicates. B and C, TNF release induced by ␣-glucan is dependent on TLR2 but not TLR4. Peritoneal macrophages obtained from WT, TLR2 Ϫ/Ϫ , C57Bl/10 (TLR4ϩ/ϩ), or C57BL/10ScN (TLR4 Ϫ/Ϫ ) were stimulated with ␣-glucan (100 g/ml), LPS (100 ng/ml), or Pam 3 Cys (100 ng/ml). After 4 h, the supernatant was harvested for TNF determination. Data represent mean Ϯ S.E. of two independent experiments, and the stimulation was performed in duplicate. D, CD14 recognition is required to TNF release induced by ␣-glucan in macrophages. Thioglycollate-elicited peritoneal macrophages were obtained from WT or CD14 Ϫ/Ϫ mice. Macrophages were stimulated with ␣-glucan (100 g/ml), LPS (100 ng/ml), or Pam 3 Cys (100 ng/ml). After 4 h, the supernatant was harvested for TNF determination. Data represent mean Ϯ S.E. of two independent experiments, and the stimulation was performed in duplicate.

FIGURE 7.
A, ␣-glucan induces TNF and B, IL-12 release by dendritic cells in mechanism involving TLR2. Dendritic cells differentiated with IL-4 and granulocyte macrophage-colony stimulating factor were plated in 96-well at a density of 2 ϫ 10 5 cells/well and stimulated for 16 h in the presence of ␣-glucan (100 g/ml), LPS (100 ng/ml), or Pam 3 Cys (100 ng/ml). The supernatant was evaluated for TNF and IL-12 by ELISA. Data represent mean Ϯ S.E. of stimuli performed in duplicates of one representative experiment from two different experiments with similar results.