Purification and Partial Characterization of a Paracoccidioides ...

INFECTION AND IMMUNITY, Apr. 2005, p. 2486–2495 0019-9567/05/$08.00⫹0 doi:10.1128/IAI.73.4.2486–2495.2005 Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Vol. 73, No. 4

Purification and Partial Characterization of a Paracoccidioides brasiliensis Protein with Capacity To Bind to Extracellular Matrix Proteins Angel Gonza´lez,1* Beatriz L. Go ´mez,1,2 Soraya Diez,1,2 Orville Hernańdez,1 1 Angela Restrepo, Andrew J. Hamilton,2† and Luz E. Cano1,3† Medical and Experimental Mycology Group, Corporacio ń para Investigaciones Biolo ´gicas,1 and Molecular Microbiology Group, Bacteriology and Clinical Laboratory School, Universidad de Antioquia,3 Medellıń, Colombia, and Dermatology Department, Guy⬘s and St. Thomas’ Hospital, King’s College, London University, London, United Kingdom2 Received 28 July 2004/Returned for modification 2 September 2004/Accepted 7 December 2004

Microorganisms adhere to extracellular matrix proteins by means of their own surface molecules. Paracoccidioides brasiliensis conidia have been shown to be capable of interacting with extracellular matrix proteins. We aimed at determining the presence of fungal proteins that could interact with extracellular matrix protein and, if found, attempt their purification and characterization. Various extracts were prepared from P. brasiliensis mycelial and yeast cultures (total homogenates, ␤-mercaptoethanol, and sodium dodecyl sulfate [SDS] extracts) and analyzed by ligand affinity assays with fibronectin, fibrinogen and laminin. Two polypeptides were detected in both fungal forms. SDS extracts that interacted with all the extracellular matrix protein were tested; their molecular masses were 19 and 32 kDa. Analysis of the N-terminal amino acid sequence of the purified 32-kDa mycelial protein showed substantial homology with P. brasiliensis, Histoplasma capsulatum, and Neurospora crassa hypothetical proteins. Additionally, a monoclonal antibody (MAb) produced against this protein recognized the 32-kDa protein in the SDS extracts of both fungal forms for immunoblot. Immunofluorescence analysis revealed that this MAb reacted not only with mycelia and yeast cells, but also with conidia, indicating that this protein was shared by the three fungal propagules. By immunoelectron microscopy, this protein was detected in the cell walls and in the cytoplasm. Both the 32-kDa purified protein and MAb inhibited the adherence of conidia to the three extracellular matrix proteins in a dose-dependent manner. These findings demonstrate the presence of two polypeptides capable of interacting with extracellular matrix proteins on the surface of P. brasiliensis propagules, indicating that there may be common receptors for laminin, fibronectin, and fibrinogen. These proteins would be crucial for initial conidial adherence and perhaps also in dissemination of paracoccidioidomycosis. vessels (macrophages, polymorphonuclear leukocytes, and human or animal tumor cells) can also exhibit these receptors (4). Fibronectin is a disulfide-linked dimeric glycoprotein present in a soluble form in blood plasma and other body fluids and in a fibrilar form in ECM. The major function of fibronectin is probably related to its ability to mediate adhesion to mammalian cells, a process that involves the binding of specific cell surface receptors to discrete domains in the fibronectin molecule (36, 42). Fibrinogen, the major plasma glycoprotein, plays a key role in inflammatory reactions, and the recognition of fibrin (or fibrinogen) deposits in the surface of wounded epithelia constitutes a mechanism by which microbial attachment to mucosal surfaces takes place (4). Several fungi of clinical importance such as Candida albicans (10, 15, 16, 31); Aspergillus fumigatus (4, 11, 18, 37, 49); Histoplasma capsulatum (33); Cryptococcus neoformans (41); Pneumocystis carinii (27, 34); Sporothrix schenckii (29); and Penicillium marneffei (22, 23) are known to attach to ECM proteins. The dimorphic fungus Paracoccidioides brasiliensis is the etiological agent of paracoccidioidomycosis, an important endemic mycosis in Latin America, including Colombia (40). The clinical manifestations of paracoccidioidomycosis are diverse, ranging from asymptomatic pulmonary lesions to systemic generalized infections. Most patients (about 50%) develop fibrotic

Adhesion of pathogenic microorganisms to host tissues is considered indispensable for initial colonization and further dissemination. The extracellular matrix (ECM) is a complex mixture of molecules containing several components, including fibronectin, vitronectin, collagens, and proteoglycans. ECM composition varies in different tissues and during phases of injury, inflammation, and repair (27). Laminin is an ECM glycoprotein present in basement membranes and in the lungs; this glycoprotein can be exposed after tissue damage resulting from either inflammatory processes or lytic activity by bacterial toxins or drugs (22). Interactions with laminin are crucial for a number of biological processes requiring cell adhesion (such as diapedesis, cellular cohesion inside tissues, and metastasis of cancer cells). Indeed, laminin receptors have been reported on cells that normally interact with basement membranes, such as epithelial or endothelial cells, muscle cells, and neuronal cells. Cells migrating from blood

* Corresponding author. Mailing address: Medical and Experimental Mycology Group, Corporacio ń para Investigaciones Biolo ´gicas (CIB), Carrera 72 A, No. 78B 141, A. A. 73 78 Medellıń, Colombia. Phone: 57-4-441 08 55. Fax: 57-4-441 55 14. E-mail: agonzalezm@cib .org.co. † Luz E. Cano and Andrew J. Hamilton share senior authorship on the manuscript. 2486

VOL. 73, 2005

ADHERENCE IN PARACOCCIDIOIDES BRASILIENSIS

sequelae that may severely hamper respiratory function (9, 14, 45). Infection is presumed to be acquired by inhalation of the air-borne conidia produced by the mycelial form of the fungus. These structures are sufficiently small to reach the alveoli (9, 40). At the moment, very little information is available on the mechanisms underlying the pathogenesis of paracoccidioidomycosis or on the means by which the fungus persists in the lungs and disseminates to other organs (40). It has been suggested that the ability of P. brasiliensis propagules to adhere to host cells and tissues could play an important role in the establishment of infection (40). Recently, our group has demonstrated that P. brasiliensis conidia interact with ECM proteins (laminin, fibronectin, and fibrinogen) in a dose-dependent manner (7). In this study, we demonstrate for the first time the presence of two proteins in the cell wall of P. brasiliensis that interact with ECM proteins. Additionally, the purification and partial characterization of a 32-kDa protein that binds to the ECM are described.

active bands were developed with nitro blue tetrazolium and 5-bromo-4-chloro3-indolylphosphate as the chromogenic reagents. Purification of the 32-kDa protein. Purification was performed by preparative gel electrophoresis with the Prep-Cell system (model 491; Bio-Rad). A 12% (vol/vol) acrylamide resolving gel (with 0.1% [wt/vol] SDS) was prepared and loaded with 20 mg of P. brasiliensis mycelium cell wall extract treated with SDS (made up in loading buffer with ␤-mercaptoethanol and boiled in water for 2 to 3 min). Two hundred sequential 2-ml fractions were collected, and then 200 ␮l from each fraction was precipitated with 3 volumes of chilled acetone and left at ⫺20°C for 2 h; samples were then centrifuged at 12,000 ⫻ g for 15 min at 4°C. The resultant pellet was air dried, resuspended in loading buffer, and then run on SDS-12% (vol/vol) PAGE. Proteins were visualized by Coomassie blue staining (0.1% wt/vol) and/or silver staining. Ligand affinity assays with fibronectin identified those fractions containing the 32-kDa protein, which were then selected and concentrated with assisted evaporation. Protein concentration was measured by the Bradford technique. N-terminal amino acid sequence. Purified samples of the 32-kDa protein (5 ␮g/␮l), pretreated with loading buffer containing ␤-mercaptoethanol and boiled for 3 min, were loaded onto a SDS–12% (vol/vol) PAGE gel (with 2 mM thioglycolic acid in the upper buffer chamber) and electrophoresed at 160 V for 2 h. The gel was then blotted onto a PVDF membrane as previously described (12, 20). The blot was stained for 1 min with 0.1% (wt/vol) Coomassie blue R and then destained. The purified band was then subjected to N-terminal amino acid sequencing with an Applied Biosystems (Warrington, United Kingdom) Procise sequencer (Protein and Nucleic Acid Chemistry facility, Department of Biochemistry, University of Cambridge, Cambridge, United Kingdom). Carbohydrate analysis. Cell wall SDS extracts or the 32-kDa purified protein was used to determine the presence of associated carbohydrates with a digoxigenin glycan differentiation kit (Roche Molecular Biochemicals, Mannheim, Germany), in which specific lectins recognize different groups of sugars (or terminal carbohydrates) associated with proteins. The methodology was carried out following the manufacturer’s instructions. Additionally, 12% polyacrylamide gels were stained with the Schiff reagent to determine the presence of sugars. MAb production. On day 1, six female BALB/c mice (8 to 10 weeks old; Harlan Olac, Oxon, United Kingdom) were immunized intraperitoneally with 50 ␮g per 100 ␮l of an equal mixture of P. brasiliensis purified 32-kDa protein made up 1:1 in Freund’s incomplete adjuvant (Difco, East Molesey, United Kingdom). The same dose was repeated each week three times. In order to determine which mouse had the highest antibody response animals were tail-bled on day 22 and response to the P. brasiliensis purified 32-kDa protein was quantified by ELISA as previously described (13, 21). The chosen mouse received a further intravenous dose of 50 ␮g of the 32-kDa protein in 100 ␮l of sterile PBS, and its spleen was then used for the fusion protocol 3 days later. Cells of the murine myeloma line Sp2/0 were fused with spleen cells from the donor mouse as previously described (20). Hybridomas were screened 7 days later by enzyme-linked immunosorbent assay, and colonies showing specificity against the 32-kDa purified protein were expanded onto 24-well plates and subcloned twice by limiting dilution. The supernatants from the different hybridoma lines were collected and concentrated 100-fold by ammonium sulfate precipitation (25). Characterization of MAbs. The specificity of MAbs was assessed by enzymelinked immunosorbent and immunoblot assays, as described elsewhere (13, 20). MAbs subclasses were determined with a subclassing kit (Serotec, Kidlington, United Kingdom) as previously described (20). Immunoenzyme development of Western blots. For analysis of purified protein and extracts from cell walls treated with SDS, SDS-12% (vol/vol) PAGE gels and immunoblotting assays were done as previously described (13, 20). Isoelectric focusing analysis. Two-dimensional gel electrophoresis (Mini-Protean II Cell, Bio-Rad) was used in order to determine the isoelectric point of P. brasiliensis cell wall extracts treated with SDS. In the first dimension, 15 ␮g of the SDS-cell wall proteins were run on an SDS-PAGE gel with an ampholyte gradient of pH 3 to 10 (Pharmacia) plus ␤-mercaptoethanol, urea (9.2 M), and Triton X-100. In the second dimension, the first product was run on an SDS-12% PAGE gel, followed by silver staining or transfer onto a PVDF membrane and development for immunoblotting as described above. Indirect immunofluorescence. The conidia obtained from Percoll gradients, the yeast cells, and the mycelium fragments were separately blocked with 5% BSA for 2 h. Cells were then centrifuged at 6,000 ⫻ g for 5 min and resuspended in PBS–1% BSA containing MAb 2G4 directed against the 32-kDa protein (1:10 dilution) and incubated for 2 h at 37°C with gentle shaking. Cells were centrifuged and washed three times with PBS and incubated for 1 h at 37°C with fluorescein isothiocyanate-conjugated rabbit anti-mouse immunoglobulin G Fc, diluted 1:20 (Jackson Immunochemicals, West Grove, Pa.) in PBS–1% BSA with shaking. Cells were then washed again three times and resuspended in a 50%

MATERIALS AND METHODS Fungus culture and conidia production. P. brasiliensis isolate ATCC 60855, previously known to sporulate freely on special medium, was employed. The techniques used to grow the mycelial form and collect and dislodge conidia have been reported previously (39). Briefly, the stock mycelial culture was grown in a liquid synthetic medium, modified McVeigh-Morton (SMV) broth (38), at 18°C (⫾ 4°C) with shaking. Growth was homogenized and portions were used to inoculate water agar plates; the latter were incubated at 18°C (⫾ 4°C) for 2 to 3 months for adequate fungal sporulation. Cultures were periodically checked for mold or bacterial contamination. Conidia used in both adherence assays and immunofluorescence studies were obtained by the discontinuous Percoll gradient method (26). Conidia were counted in a hemacytometer and their viability was evaluated by fluorescein diacetate and ethidium bromide staining (6). Absence of contamination was assessed in brain heart infusion (BHI) broth incubated at 37°C. P. brasiliensis yeast cells (ATCC 60885) were grown for 4 to 5 days in liquid BHI plus glucose (1%) medium at 36°C with shaking. Yeast cells were harvested by centrifugation, washed twice in phosphate-buffered saline (PBS), and disrupted by ultrasound to obtain as many single yeast cells as possible; their viability was assessed as described before. Preparation of fungal extracts. Extracts were prepared according to Pen ãlver et al. (37). A very dense suspension (equivalent to 20 mg wet weight) of both mycelia grown in liquid SMV medium (38) (incubated at 18°C in a gyratory shaker at 150 rpm for 2 weeks) and yeast forms grown in BHI medium for 5 days were used. Cells were washed twice with PBS and then treated with a protease inhibitor cocktail for mycelia and yeast extracts (Sigma, St Louis, Mo,). This cocktail was added at a ratio of 1 ml per 20 g of wet weight. Three extracts were produced: (i) a cell-free homogenate, (ii) a cell wall preparation made via ␤mercaptoethanol treatment, and (iii) a cell wall preparation via SDS extraction. The protein content of samples was determined by the Bradford method (5). Affinity ligand assays. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out as described by Laemmli (28), with 12% resolving gels. Gels were stained with silver stain (Bio-Rad, Hercules, Calif.) to detect the presence of proteins. Electrophoretic transfer of proteins from polyacrylamide gels to polyvinylidene difluoride membranes (Immobilon P; Millipore Corporation, Watford, United Kingdom) was carried out as described previously (8, 37). Blotted proteins were assayed for extracellular matrix protein binding as follows. The polyvinylidene difluoride membranes were blocked with 5% bovine serum albumin (BSA) in Tris-buffered saline (TBS) for 2 h at room temperature and then incubated overnight with gentle shaking in PBS containing 50 ␮g of laminin per ml (derived from Engelbreth-Holm Swarm mouse sarcoma; Sigma), 100 ␮g of human fibronectin per ml (Sigma), or 50 ␮g of bovine fibrinogen (Sigma) per ml. After being washed four times (10 min per wash) with TBSTween 0.05% (TBST), the membranes were incubated for 1 h with agitation with rabbit anti-mouse laminin (1:100 dilution); rabbit anti-human fibronectin antibodies (1:100), or goat anti- fibrinogen (1:500) (all of them from Sigma) in TBST plus 1% BSA (TBSTB). The blots were washed with TBST and incubated with alkaline phosphatase-labeled goat anti-rabbit or donkey anti-goat immunoglobulin (Jackson Immunochemicals, West Grove, Pa.) at a 1:10,000 or 1:20,000 dilution, respectively, in TBSTB. Finally, the blots were washed again, and re-

2487

2488

´ LEZ ET AL. GONZA

INFECT. IMMUN.

FIG. 1. Identification of P. brasiliensis proteins able to bind to ECM proteins. A, SDS-PAGE gel stained with silver; B to D, ligand affinity assays with laminin (B), fibronectin (C), and fibrinogen (D). Tracks MW, molecular size markers (in kilodaltons). Different extracts from mycelium (tracks 1 to 3) and yeast (tracks 4 to 6) are shown. Total homogenates (tracks 1 and 4), cell wall ␤-mercaptoethanol extracts (tracks 2 and 5), and cell wall SDS extracts (tracks 3 and 6) are shown. The arrows indicate two proteins with molecular masses of 19 and 32 kDa that recognize the three ECM proteins which were present solely on cell wall SDS extracts.

(vol/vol) glycerol-PBS (pH 7.5) buffer and observed by fluorescence microscope. A negative control in which the primary antibody was replaced with PBS buffer was also used. Immunoelectron microscopy. Cells were processed for postembedding analysis as described elsewhere (19) with minor modifications. Briefly, cells were fixed in 0.5% vol/vol glutaraldehyde–4% vol/vol formaldehyde for 30 min at 4°C and washed three times for 3 min each with phosphate buffer, 0.1 M (pH 7.4). After fixation, the cells were processed for embedding in LR-white resin. Ultrathin sections mounted on Formvar-carbon-coated nickel grids were floated on droplets containing a solution of PBS plus 1% BSA, and then MAb 2G4 was added and incubated for 1 h at room temperature. Cells were washed with 0.02 M TBS (pH 7.2) and incubated for 1 h at room temperature with TBST containing goat anti-mouse immunoglobulin G-gold complex (12-nm average particle size, 1:10 dilution; Jackson Immunochemicals). After washing, ultrathin sections were stained with 6% wt/vol uranyl acetate and examined with a Zeiss electron microscope. Surface labeling of conidia with biotin. Surface labeling was performed as previously described by Pen ãlver (37) with minor modifications. Briefly, conidia were incubated for 15 min with PBS containing 1% (vol/vol) H2O2. Conidia (at 7 a concentration of 10 per ml) were added to 100 mM phosphate buffer (pH 8.0) containing 2 ␮g of n-hydroxysuccinimido-biotin per ml (n-hydroxysuccinimidobiotin was previously dissolved in dimethyl sulfoxide). After incubation for 1 h at 28°C with gentle shaking in a gyratory plate, the conidia were recovered and washed once with 50 mM phosphate buffer (pH 6.0) and then once with 10 mM phosphate buffer (pH 7.4). Biotinylated conidia were incubated with streptavidin-peroxidase conjugate (1 ␮g per ml) in TBSTB. After incubation for 1 h at

room temperature with agitation, the labeled conidia were washed in PBS and used for adherence assays. Adherence assays. Adherence assays were performed as described elsewhere (22, 37); 96-well microtiter plates (Maxisorp; Nunc A/S, Kamstrup, Denmark) were initially coated with 100 ␮l per well of laminin (50 ␮g/ml), fibronectin (100 ␮g/ml), or fibrinogen (50 ␮g/ml) dissolved in PBS and incubated overnight at 4°C and then for 1 h at 37°C. Plates were blocked by adding 1% (wt/vol) BSA solution in PBS and incubated at 37°C for 1 h. Plates were then washed three times with PBS and different concentrations of the purified 32-kDa protein were added (0, 1, 5, 10, 20, 50, and 100 ␮g/ml) and incubated for 2 h at 37°C. Plates were washed three times with PBS, and labeled conidia were added as appropriate (105 conidia per well, 100 ␮l in total) and incubated for 1 h at 37°C. In other experiments, the conidia were previously treated with different concentrations of MAb 2G4 (dilutions 1:10, 1:100, and 1:500). Nonadherent cells were removed by washing in PBS containing 0.05% (vol/vol) Tween 20. Negative control wells were coated with 1% wt/vol BSA in PBS only. The color reaction was developed with 3,3⬘,5,5⬘-tetramethylbenzidine (TMB) (BD Biosciences, San Diego, Calif.); after the addition of the substrate, plates were incubated in the dark, and the color reaction was stopped by adding 50 ␮l of 2 M H2SO4 to each well. Color intensity was determined at 450 nm with an automated plate reader. Results expressed as optical density (OD) at 450 nm, represented as the means of three independent experiments done in duplicate. Statistical analysis. Results are expressed as the mean ⫾ standard error of the mean of two or three independent experiments. Data were analyzed by one-way analysis of variance with GraphPad Prism version 3.02 for Windows (GraphPad Software, San Diego, Calif., www.graphpad.com).


VOL. 73, 2005

FIG. 2. Recognition of the 32-kDa protein by ECM proteins with ligand affinity assays. A, SDS-PAGE gel stained with silver; B to D, ligand affinity assays with laminin (B), fibronectin (C), and fibrinogen (D). Tracks MW, molecular size markers (shown in kilodaltons); track 1, cell wall SDS extracts (control); track 2, 32-kDa purified protein. The arrow indicates the 32-kDa protein that recognizes the three ECM proteins.

RESULTS Identification of cell components with the ability to bind extracellular matrix proteins. In order to identify which components from the fungal cells bound to ECM proteins, different extracts from mycelial and yeast forms were used. On silverstained SDS –12%PAGE gels, we observed a complex array of proteins in all fungal extracts (total homogenates, cell wall ␤-mercaptoethanol extracts, and cell wall SDS extracts) (Fig. 1A). However, binding of the three ECM proteins (laminin, fibronectin, and fibrinogen) used was detected only in the SDS cell wall extracts in two components having relative molecular masses of 19 and 32 kDa (Fig. 1B to D). The reaction was specific, as indicated by the absence of bands following exposure to the specific anti-ECM protein antibodies when previ-

2489

ous incubation of the blots in the ECM protein solution had been omitted (data not shown). Purification of the 32-kDa protein and N-terminal amino acid sequencing. A total of 30 of the 200 fractions (numbers 101 to 130) collected from the Prep-Cell system were found to contain the 32-kDa protein when assessed by SDS-PAGE and affinity ligand assay with fibronectin. Pure protein was found in fractions 106 to 118 (data not shown); the total protein content of these fractions was 0.3 mg, which was obtained from an initial quantity of 20 mg of cell wall SDS extracts (giving an approximate yield of 1.5%). The purification of this protein was repeated five times with no significant variation, demonstrating the reproducibility of the method. Ligand affinity assays incorporating the three ECM proteins and the purified 32-kDa protein revealed that the latter recognized laminin, fibronectin, and fibrinogen (Fig. 2A to D). The N-terminal amino acid sequence of the 32-kDa purified P. brasiliensis mycelium protein led to the identification of 19 amino acids (TKITTLLFDCDNTLVLSEE). A BLAST search revealed the substantial homology of this sequence with hypothetical proteins of P. brasiliensis (unpublished sequence data kindly communicated by Gustavo H. Goldman; http://143.107 .203.68/est/default.html), Neurospora crassa (NCBI accession AL513467), and Histoplasma capsulatum (http://www.genome .wustl.edu/projects/hcapsulatum) with an identity of 100, 94, and 86%, respectively; the same hypothetical protein appears to be conserved also in Gibberella zeae (NCBI accession XM_38530) and Aspergillus fumigatus (TIGR, http://www.tigr .org). Additionally, lower homology was observed with serine phosphate proteases (P-Ser-Hpr) of Bacillus subtilis and Listeria monocytogenes (58%) (Table 1). Carbohydrate analysis. Sugar groups (or terminal carbohydrates) associated with the P. brasiliensis 32-kDa protein were not observed when specific lectins or the Schiff reagent was used. However, when the presence of sugars on the cell wall SDS extracts was analyzed, species with molecular masses of about 38 and 82 kDa recognizing terminal mannose, and ␣(1– 3)-, ␣(1–6)-, or ␣(1–2)-linked mannose were observed, different from the 32-kDa protein (data not shown).

TABLE 1. N-terminal amino acid sequence comparison of 32-kDa protein with the most-homologous proteins from the BLAST search Protein

P. brasiliensis 32-kDa protein P. brasiliensis hypothetical protein Neurospora crassa hypothetical protein Histoplasma capsulatum hypothetical protein Gibbirella zeae hypothetical protein Aspergillus fumigatus hypothetical protein Listeria monocytogenes P-Ser-HPrc Bacillus subtillis P-Ser-HPr a b c

Identical residues are underlined. Percent sequence identities in amino acid overlap. P-Ser-HPr, serine phosphatase proteinase.

Sequencea

TKITTLLFDCDNTLVLSEE 1 10 19 TKITTLLFDCDNTLVLSEE 2 20 GKITTLLFDCDNTLVLSEE 2 20 ITTLLFDCDNTLVLSEE PEINTTLLFDCDNTLVLSEE 6 20 QITGIFFDCDNTLVLSEE KITTLLFDLDGTL 4 16 KITTLLFDLDGTL 4 16

% Identityb

100 94 86 79 74 58 58

2490

´ LEZ ET AL. GONZA

FIG. 3. P. brasiliensis 32-kDa protein as visualized on an SDS–12% PAGE gel stained with silver (A) and immonoenzyme reactivity on Western blot of MAb 2G4 with the 32 kDa purified protein (B). Tracks MW, molecular size markers (in kilodaltons); track 1, cell wall SDS extract from mycelium; track 2, cell wall SDS extract from yeast; track 3, 32-kDa purified protein.

MAb production and immunoenzyme recognition of the 32kDa protein by immunoblot. After serial subcloning, a panel of different hybridoma lines reactive with the 32-kDa protein were produced, including MAb 2G4, which possessed a particularly high reactivity against this protein. MAb 2G4 belongs to the immunoglobulin G1 subclass (data not shown) and was used in the immunodetection of the 32-kDa protein by immu-

INFECT. IMMUN.

noblot, indirect immunofluorescence, and immunoelectron microscope studies. This MAb recognized the 32-kDa protein in its purified form (Fig. 3) and was also reactive with cell wall SDS extracts from P. brasiliensis mycelial and yeast forms. Additional assays demonstrated that MAb was unreactive when total homogenates and ␤-mercaptoethanol cell wall extracts were used (data not shown). MAb 2G4 recognized a protein with a molecular mass of 32 kDa and an isoelectric point of 6.6 by isoelectric focusing (Fig. 4). Immunofluorescence identification of the 32-kDa protein in fungal cells. The surface of the different propagules of P. brasiliensis (mycelium, mycelial fragments, yeast, and conidia) demonstrated immunofluorescence labeling when incubated with the anti-32-kDa protein MAb, indicating the presence of this protein on all fungal forms (Fig. 5D, H, and L). In contrast, no reactivity was evident when the cells were incubated in the absence of MAb (Fig. 5C, G, and K), demonstrating that fluorescence was dependent on the previous interaction of the cells with MAb 2G4. Detection of the 32-kDa protein by immunoelectron microscopy. Mycelial and yeast cells were examined by immunoelectron microscopy with MAb 2G4. In mycelial and yeast cells processed by the postembedding method, gold particles were present in both the cytoplasm and the cell wall (Fig. 6A and B). Labeling intensity varied in different cells, both within the cytoplasm and in the cell wall. Control sample not exposed to MAb 2G4 prior to incubation with the gold-conjugated antibody were free of label (data not shown). Inhibition of adherence to extracellular matrix proteins with the 32-kDa purified protein or the monoclonal antibody to the anti-32-kDa protein. Biotinylated, streptavidin-peroxidase-labeled conidia adhered to immobilized laminin, fibronectin, and fibrinogen (Fig. 7 and 8). However, when the adhesion experiments were performed after treatment of laminin, fibronectin, and fibrinogen with different concentra-

FIG. 4. Fractionated P. brasiliensis cell wall SDS extract (15 ␮g) by two-dimensional gel electrophoresis. A, SDS-PAGE gel stained with silver. B, Immunoenzyme reactivity of the 32-kDa protein by Western blot with MAb 2G4. To the left of both figures are molecular size standards (in kilodaltons), and at the top are the isoelectric point determinations performed in the first dimension.

VOL. 73, 2005


2491

tions of 32-kDa purified protein (0, 1, 5, 10, 20, 50, and 100 ␮g/ml), the attachment to three extracellular matrix proteins decreased significantly (P ⬍ 0.02) in a dose-dependent manner (Fig. 7). The same results were observed when the conidia were treated with different concentrations of MAb 2G4 (1:10, 1:100, and 1:500) (P ⬍ 0.001). A low but significant inhibition of adherence to laminin and fibronectin was also observed in the presence of an irrelevant antibody (1:10 dilution). However, a significant difference in inhibition was found when MAb 2G4 was used at dilutions of 1:10 and 1:100 compared with the irrelevant antibody (P ⬍ 0.05) (Fig. 8). These results indicate that the 32-kDa protein mediates adhesion to laminin, fibronectin, and fibrinogen. DISCUSSION Adhesion of microorganisms to host tissues is now considered an essential event in the infection process. The recognition of host cells by the pathogen requires the presence of complementary molecules at the surface of the former. Various host proteins such as laminin, collagen, fibronectin, fibrinogen, and the C3 component of complement have been proposed as the microbial cell ligands (35). Normally, these glycoproteins remain unexposed at the surface of epithelial and endothelial cells. However, any type of trauma that damages the host tissue may lead to the exposure of basement membranes and therefore to the accessibility of components such as laminin, collagen, and heparan-sulfate proteoglycans. Moreover, the subsequent inflammatory reaction results in deposition of fibrinogen and of some components of the complement system (4). In this study, we demonstrated for the first time that the dimorphic fungus P. brasiliensis presents on its surface two proteins with molecular masses of 19 and 32 kDa that interact with different ECM proteins such as laminin, fibronectin, and fibrinogen, as shown by affinity ligand assays. These two proteins may be common receptors capable of participating in fungal adherence through their binding to ECM components of the host cells. In several fungi of clinical importance, the presence of different molecules that participate in the interaction with these ECM proteins has been demonstrated. In addition, these molecules also appear to have common receptors. Thus, in P. marneffei, ligand affinity assays have revealed a 20-kDa protein that binds to laminin and fibronectin (22, 23). In A. fumigatus, the existence of receptors with the capacity to recognize more than one ligand has also been demonstrated (4). These molecules exhibit a diverse array of sizes, with the 23-, 30-, 37-, and 72-kDa species recognizing ECM proteins such as laminin and fibrinogen (3, 4, 18, 37, 46). In Candida albicans, molecules of 37, 58, and 67 kDa that bind to laminin and fibrinogen have also been described (31, 32). In addition, integrin-like and lectin-like receptors have been shown in yeast cells and germinal tubes of this fungus, which are involved in adherence to vitronectin, collagen type IV, and also human endothelial cell lines (1, 43, 44). In H. capsulatum, a molecule of 50 kDa that confers the capacity to bind to laminin has also FIG. 5. Immunofluorescence reactivity of P. brasiliensis cells with MAb 2G4. A to D, mycelium or mycelial fragments. Bar, 20 ␮m. E to H, yeast cells. Bar, 20 ␮m. I to L, conidia. Bar, 5 ␮m. Conventional optical microscopy (A, B, E, F, I, and J) and fluorescence microscopy

(C, D, G, H, K, and L). Negative controls without monoclonal antibody 2G4 (C, G, and K) did not show fluorescence.

2492

´ LEZ ET AL. GONZA

INFECT. IMMUN.

FIG. 6. Immunoelectron microscopy detection of the 32-kDa protein in P. brasiliensis mycelium and yeast cells by postembedding methods. Labeling was observed mainly in the cytoplasm and cell walls of the mycelium (A) and the yeast cells (B). Bars, 0.5 ␮m. CW, cell wall. The arrowheads indicate gold particles. Negative controls not exposed to MAb 2G4 prior to incubation with gold-conjugated antibody were free of label (data not shown).

been identified (33). Recently, it has been demonstrated that Cryptococcus neoformans interacts with fibronectin through two polypeptides of 25 and 35 kDa present on total homogenates and cell wall preparations, respectively; additionally, it was observed that a serine proteinase of 75 kDa was capable of

degrading this ECM protein in order to gain access to host tissues (41). In P. brasiliensis, the 43-kDa glycoprotein, which is the major antigenic component of the fungus, appears to be a lamininlike receptor (30, 48). In addition, it has been demonstrated

FIG. 7. Inhibition of P. brasiliensis conidial binding to ECM proteins with purified 32-kDa protein. Biotin-streptavidin-peroxidase-labeled conidia were added to wells coated separately with 50 ␮g of laminin per ml (A), 100 ␮g of fibronectin per ml (B), or 50 ␮g of fibrinogen per ml (C) and various concentrations of the 32-kDa purified protein. Nonadherent conidia were removed by washing with PBS, and the adherent conidia were indirectly estimated by measuring the optical density at 450 nm. Negative controls consisted of wells covered with BSA. The results represent duplicate determinations of three experiments and are expressed as means ⫾ standard error of the mean. *, P ⬍ 0.02, comparisons made with conidia already adhered to ECM that had been treated with various concentrations of the 32-kDa purified protein and with conidia adherent to ECM but untreated.

VOL. 73, 2005


2493

FIG. 8. Inhibition of P. brasiliensis conidial binding to ECM proteins with MAb 2G4 against the 32-kDa protein. Biotin-streptavidin-peroxidaselabeled conidia previously treated with different concentrations of MAb 2G4 at 1:10, 1:100, and 1:500 dilutions or treated with irrelevant antibody (IRREL at 1:10 dilution) were added to wells coated separately with 50 ␮g of laminin per ml (A), 100 ␮g of fibronectin per ml (B), or 50 ␮g of fibrinogen per ml (C). Nonadherent conidia were removed by washing with PBS, and the adherent conidia were indirectly estimated by measuring the optical density at 450 nm. Negative controls consisted of wells covered with BSA. The results represent duplicate determinations of three experiments and are expressed as means ⫾ standard error of the mean. *, P ⬍ 0.001, comparisons made with conidia already adhered to ECM that had been previously treated with the various concentrations of MAb 2G4 with conidia adhered to ECM but untreated; **, P ⬍ 0.05, comparison between conidia treated with irrelevant antibody and conidia adhered to ECM but untreated; ⫹, P ⬍ 0.05, comparisons between conidia treated with MAb and conidia treated with irrelevant antibody.

that the laminin-binding property of gp43 enhanced the pathogenicity of laminin-coated yeast cells in a hamster model of testicle infection (48). Additional studies demonstrated that patient sera, anti-gp43, and monoclonal antibodies against this glycoprotein all inhibited the adherence of yeast cells to different cell lines (17, 24). Recently, Andre´ et al. (2) showed that treatment with laminin does not enhance P. brasiliensis pathogenicity in a pulmonary model of infection, even when low infecting doses of the virulent yeast (Pb18) or a low-virulence isolate (Pb265) were used. However, treatment with laminin led to a less severe pathology, as reveled by histopathological analyses of the Pb18-infected group and diminished CFU counts in the lungs of mice infected with the low-virulence isolate. These investigators speculated that previous in vitro treatment with laminin covered laminin-binding epitopes on gp43, avoiding their in vivo interaction with laminin molecules present in damaged alveolar spaces, resulting in lower adherence of fungal cells, and suggested that isolates with different degrees of virulence could express a diverse pattern of adhesins or even different concentrations of the same fungal adhesin (2). In our study, we identified two novel proteins of 19 and 32 kDa that were present on cell wall SDS extracts from both mycelial and yeast forms. These proteins showed the capacity to bind not only to laminin, but also to fibronectin and fibrinogen. The characterization of the P. brasiliensis cell wall SDS extracts and the purified 32-kDa protein revealed no associated carbohydrates, indicating that these molecules are not glycoproteins. In studies with other fungal pathogens, some of the molecules interacting with ECM proteins are associated with carbohydrates, as in the cases of a 43-kDa glycoprotein of P. brasiliensis that binds to laminin (30, 48), a 72-kDa glyco-

protein of A. fumigatus that binds to laminin (46), a neutral fraction of cell wall from Sporothrix schenkii that binds to fibronectin (29), and a mannoprotein of 58 kDa on Candida albicans that binds to fibrinogen (31). Recently, a specific lectin to sialic acid residues of a 32-kDa protein from A. fumigatus has been purified; this lectin was preferentially located on the conidium cell wall, suggesting that this molecule could participate in adherence to both epithelial cells and ECM proteins (47). Previous studies by our group have indicated that the adherence of P. brasiliensis conidia to ECM proteins is a sialic acid-dependent event (7). The N-terminal amino acid sequence revealed a variable identity with different fungi, especially with hypothetical proteins of P. brasiliensis, H. capsulatum, and N. crassa (Table 1). This hypothetical protein appears to be conserved in Giberrella zeae and A. fumigatus and would have a predicted molecular mass of approximately 27 kDa for these proteins. The apparent difference (about 5 kDa) could be explained by posttranslational modifications or possibly by alternative splicing of the two-exon hypothetical gene. No information on the probable function of this putative protein is available. Although it is reported to contain a haloacid dehalogenase-like hydrolase domain, this domain family is very large, and no close matches of the hypothetical domain were found in other proteins. In Neurospora, Histoplasma, and Aspergillus spp., the predicted gene always has one intron near the beginning of the sequence (after the first three amino acids). These results suggest that we are in the presence of a new function of this protein (adhesin?), which could participate in the adherence of P. brasiliensis. More studies are needed to confirm and identify other possible functions of this protein.

2494

´ LEZ ET AL. GONZA

An MAb has also been produced against the 32-kDa protein; this MAb recognized the 32-kDa protein on cell wall SDS extracts from mycelium and yeast extracts via Western blot. In addition, this MAb demonstrated the presence of the protein on the surface of the different fungal propagules by immunofluorescence and also its subcellular localization on the cell wall and in the cytoplasm by immunoelectron microscopy. The differences in the lack of reactivity in the cytoplasmic extract and its presence in the cytoplasm by electron microscopy could have three possible explanations: (i) the protein is produced in the cytoplasm but stored in small vesicles that are then transported to the surface, (ii) the vesicles containing the protein remain intact during the protein extraction process, and (iii) the quantity of this protein in the cytoplasm would be so small that it cannot be detected by the method used. However, these findings suggest that this protein is expressed in all morphological forms of the fungus, which in turn implies that any of these propagules could interact with different ECM proteins. Perhaps the clearest evidence for the role of the 32-kDa protein comes from the inhibition assay with P. brasiliensis conidia binding to different immobilized ECM proteins. This demonstrated that the 32-kDa purified protein or MAb against this protein significantly inhibited the conidial-ECM protein interaction in a dose-dependant manner. Conceptually, it would seem likely that attachment of conidia to the bronchoalveolar epithelium is a crucial step in the establishment of initial infection and later on the dissemination process. The ability to adhere to epithelial cells may represent the means by which conidia avoid entrapment by respiratory tract mucus and removal by the action of ciliary cells (22). In fact, the pronounced similarities in the laminin, fibronectin, and fibrinogen binding process under our experimental conditions strongly indicate that they are mediated through the same receptor molecule, as shown in the ligand affinity assay for the ECM. Taken together, these results suggest that P. brasiliensis has different adherence systems with more than one receptor for a single ECM protein or with a single receptor that recognizes multiple domains in the different ligands. Laminin, fibronectin, and fibrinogen may share some binding sites. More studies are necessary to clarify this mechanism, together with analysis of the 19-kDa protein to determine its role in the adherence of P. brasiliensis to ECM proteins and host cells. ACKNOWLEDGMENTS This work was supported by Wellcome Trust Project No. 062247/Z/ 00Z, the Corporacio ń para Investigaciones Biolo ´gicas, and the Universidad de Antioquia. The National Doctoral Program of COLCIENCIAS supported Angel Gonza´lez. We thank Oliver Clay for helpful discussion and Maria L. Caldas and Ladys Sarmiento from the Microscopic and Imageneology Department, Instituto Nacional de Salud, Bogota´, Colombia, for assistance with immunoelectron microscopy. REFERENCES 1. Alonso, R., I. Llopis, C. Flores, A. Murgui, and J. Timoneda. 2001. Different adhesins for type IV collagen on Candida albicans: identification of a lectinlike adhesin recognizing the 7S(IV) domain. Microbiology 147:1971–1981. 2. Andre´ D. C., J. D. Lopes, M. F. Franco, C. A. C. Vaz, and V. l. G. Calich. 2004. Binding of laminin to Paracoccidioides brasiliensis induces a less severe pulmonary paracoccidioidomycosis caused by virulent and low-virulent isolates. Microbes Infect. 6:549–558.

INFECT. IMMUN. 3. Annaix, V., J. P. Bouchara, G. Larcher, D. Chabasse, and G. Tronchin. 1992. Specific binding of human fibrinogen fragment D to Aspergillus fumigatus conidia. Infect. Immun. 60:1747–1755. 4. Bouchara, J. P., M. Sa ńchez, A. Chevailler, A. Marot-Leblond, J. C. Lissitzky, G. Tronchin, and D. Chabasse. 1997. Sialic acid-dependent recognition of laminin and fibrinogen by Aspergillus fumigatus conidia. Infect. Immun. 65:2717–2724. 5. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248–252. 6. Calich, V. L., A. Purchio, and C. R. Paula. 1978. A new fluorescent viability test for fungi cells. Mycopathologia 66:175–177. 7. Caro, E. 2003. Paracoccidioides brasiliensis conidia: Interaction with extracellular matrix proteins and type II alveolar cells. M.Sc. thesis. Universidad de Antioquia, Medellıń, Colombia. 8. Casanova, M., M. L. Gil, L. Cardenoso, J. P. Martinez, and R. Sentandreu. 1989. Identification of wall-specific antigens synthesized during germ tube formation by Candida albicans. Infect. Immun. 57:262–271. 9. Cock, A. M., L. E. Cano, D. Ve´lez, B. H. Aristiza ´bal, J. Trujillo, and A. Restrepo. 2000. Fibrotic sequelae in pulmonary paracoccidioidomycosis: histopathological aspects in BALB/c mice infected with viable and non-viable Paracoccidioides brasiliensis propagules. Rev. Inst. Med. Trop. S. Paulo 42: 59–66. 10. Cotter, G., R. Weedle, and K. Kavanagh. 1998. Monoclonal antibodies directed against extracellular matrix proteins reduce the adherence of Candida albicans to HEp-2 cells. Mycopathologia 141:137–142. 11. Coulot, P., J. P. Bouchara, G. Renier, V. Annaix, C. Planchenault, G. Tronchin, and D. Chabasse. 1994. Specific interaction of Aspergillus fumigatus with fibrinogen and its role in cell adhesion. Infect. Immun. 62:2169– 2177. 12. Diez, S., B. L. Go ´mez, A. Restrepo, R. J. Hay, and A. J. Hamilton. 2002. Paracoccidioides brasiliensis 87-kilodalton antigen, a heat shock protein useful in diagnosis: characterization, purification, and detection in biopsy material via immunohistochemistry. J. Clin. Microbiol. 40:359–365. 13. Figueroa, J. I., A. J. Hamilton, M. A. Bartholomew, T. Harada, L. Fenelon, and R. J. Hay. 1990. Preparation of species-specific murine monoclonal antibody against the yeast phase of Paracoccidioides brasiliensis. J. Clin. Microbiol. 28:1766–1769. 14. Funari, M., J. Kavakama, M. A. Shikanai-Yasuda, L. G. Castro, G. Bernard, M. S. Rocha, G. C. Cerri, and N. L. Mu ¨ller. 1999. Chronic pulmonary paracoccidioidomycosis (South American blastomycosis): high resolution CT findings in 41 patients. AJR Am. J. Roentgenol. 173:59–64. 15. Gaur, N. K., S. A. Klotz, and R. L. Henderson. 1999. Overexpression of the Candida albicans ALA1 gene in Saccharomyces cerevisiae results in aggregation following attachment of yeast cells to extracellular matrix proteins, adherence properties similar to those of Candida albicans. Infect. Immun. 67:6040–6047. 16. Gaur, N. K., and S. A. Klotz. 1997. Expression, cloning, and characterization of a Candida albicans gene, ALA1, that confers adherence properties upon Saccharomyces cerevisiae for extracellular matrix proteins. Infect. Immun. 65: 5289–5294. 17. Gesztesi, J. L., R. Puccia, L. R. Travassos, A. P. Vicentini, J. Z. de Moraes, M. F. Franco, and J. D. Lopes. 1996. Monoclonal antibodies against the 43,000 Da glycoprotein from Paracoccidioides brasiliensis modulate lamininmediated fungal adhesio ń to epithelial cells and pathogenesis. Hybridoma 15:415–422. 18. Gil, M. L., M. C. Pen ˜ alver, J. L. Lopez-Ribbot, J. E. OⴕConnor, and J. P. Martıńez. 1996. Binding of extracellular matrix proteins to Aspergillus fumigatus conidia. Infect. Immun. 64:5239–5247. 19. Gozalbo, D., I. Gil-Navarro, I. Azorıń, J. Reanu-Piqueras, J. P. Martıńez, and M. L. Gil. 1998. The cell wall-associated glyceraldeyde-3-phosphate dehydrogenase of Candida albicans is also a fibronectin and laminin binding protein. Infect. Immun. 66:2052–2059. 20. Hamilton, A. J., M. A. Bartholomew, L. E. Fenelon, J. I. Figueroa, and R. J. Hay. 1990. A murine monoclonal antibody exhibiting high species specificity for Histoplasma capsulatum var capsulatum. J. Gen. Microbiol. 136:331–335. 21. Hamilton, A. J., M. A. Bartholomew, L. E. Fenelon, J. I. Figueroa, and J. R. Hay. 1990. Preparation of monoclonal antibodies that differentiate between Histoplasma capsulatum variant capsulatum and H. capsulatum variant duboissii. Trans. R. Soc. Trop. Med. Hyg. 84:425–429. 22. Hamilton, A. J., L. Jeavons, S. Youngchim, N. Vanittanakom, and J. R. Hay. 1998. Sialic acid-dependent recognition of laminin by Penicillium marneffei conidia. Infect. Immun. 66:6024–6026. 23. Hamilton, A. J., L. Jeavons, S. Youngchim, and N. Vanittanakom. 1999. Recognition of fibronectin by Penicillium marneffei conidia via a sialic aciddependent process and its relationship to the interaction between conidia and laminin. Infect. Immun. 67:5200–5205. 24. Hanna, S. A., J. L. Monteiro da Silva, and M. J. Mendes-Giannini. 2000. Adherence and intracellular parasitism of Paracoccidioides brasiliensis in Vero cells. Microbes. Infect. 2:877–884. 25. Harlow E. D., and D. Lane. 1998. Antibodies: a laboratory manual, p. 283– 318. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

VOL. 73, 2005 26. Jimeńez M. del, P., A. Restrepo, L. F. Garcıá, and L. E. Cano. 2004. Separation of Paracoccidioides brasiliensis conidia through Percoll gradients. Med. Mycol. 42:349–353. 27. Kottom, T. J., J. R. Kohler, C. F. Jr. Thomas, G. R. Fink, and A. H. Limper. 2003. Lung epithelial cells and extracellular matrix components induce expression of Pneumocystis carinii STE20, a gene complementing the mating and pseudohyphal growth defects of ste20 mutant yeast. Infect. Immun. 71: 6463–6471. 28. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the heat of bacteriophage T4. Nature 227:680–685. 29. Lima, O. C., C. C. Figueiredo, J. O. Previato, L. Mendonca-Previato, V. Morandi, and L. M. Lopes Bezerra. 2001. Involvement of fungal cell wall components in adhesion of Sporothrix schenkii to human fibronectin. Infect. Immun. 69:6874–6880. 30. Lopes, J. D., M. C. R. Moura-Campus, A. P. Vicentini, J. L. Gesztesi, W. de Souza, and Z. P. Camargo. 1994. Characterization of glycoprotein gp43, the major laminin-binding protein of Paracoccidioides brasiliensis. Braz. J. Med. Biol. Res. 27:2309–2313. 31. Lo ´pez-Ribot, J. L., C. Monteagudo, P. Sepu ´lveda, M. Casanova, J. P. Martıńez, and W. L. Chaffin. 1996. Expression of the fibrinogen binding mannoprotein and the laminin receptor of Candida albicans in vitro and in infectious tissues. FEMS Microbiol. Letters 142:117–122. 32. Lo ´pez-Ribot, J. L., M. Casanova, C. Monteagudo, P. Sepu ´lveda, and J. P. Martıńez. 1994. Evidence for the presence of a high-affinity laminin receptor-like molecule on the surface of Candida albicans yeast cells. Infect. Immun. 62:742–746. 33. McMahon, J. P., J. Wheat, M. E. Sobel, R. Pasula, J. F. Downing, and W. J. II Martıń. 1995. Murine laminin binds to Histoplasma capsulatum. A possible mechanism of dissemination. J. Clin. Investig. 96:1010–1017. 34. Narasimhan, S., M. Y. K. Armstrong, K. Rhee, J. C. Edman, F. F. Richards, and E. Spicer. 1994. Gene for an extracellular matrix receptor protein from Pneumocystis carinii. Proc. Natl. Acad. Sci. USA 91:7440–7444. 35. Patti, J. M., B. L. Allen, M. J. McGavin, and M. Hook. 1994. MSCRAMMmediated adherence of microorganism to host tissues. Annu. Rev. Microbiol. 48:585–617. 36. Pearlstein, E., L. I. Gold, and A. Garcia-Pardo. 1980. Fibronectin: a review of its structure and biological activity. Mol. Cell. Biochem. 29:103–128. 37. Pen ˜ alver, M. C., J. E. O’Connor, J. P. Martıńez, and M. L. Gil. 1996. Binding of human fibronectin to Aspergillus fumigatus conidia. Infect. Immun. 64: 1146–1153. 38. Restrepo, A., and B. E. Jimeńez. 1980. Growth of Paracoccidioides brasiliensis

Editor: T. R. Kozel


39.

40.

41.

42. 43.

44.

45.

46.

47.

48.

49.

2495

yeast phase in a chemical defined culture medium. J. Clin. Microbiol. 12: 279–281. Restrepo, A., M. E. Salazar, L. E. Cano, and M. M. Patin ˜ o. 1986. A technique to collect and dislodge conidia produced by Paracoccidioides brasiliensis mycelial form. J. Med. Vet. Mycol. 24:247–250. Restrepo–Moreno, A. 2003. Paracoccidioidomycosis, p. 328–345. In W. E. Dismukes, P. G. Pappas, and J. D. Sobel (ed.), Clinical mycology. Oxford University Press, New York, N.Y. Rodrigues, M. L., G. dos Reis, R. Puccia, L. R. Travassos, and C. S. Alviano. 2003. Cleavage of human fibronectin and other basement membrane-associated proteins by a Cryptococcus neoformans serine proteinase. Microb. Pathog. 34:65–71. Ruoslahti, E. 1988. Fibronectin and its receptors. Annu. Rev. Biochem. 57: 375–413. Santoni, G., E. Spreghini, R. Lucciarini, C. Amantini, and M. Piccoli. 2001. Involvement of ␣v␤3 integrin-like receptor and glycosoaminoglycans in Candida albicans germ tube adhesion to vitronectin and to a human endothelial cell line. Microb. Pathog. 31:159–172. Spreghini, E., A. Gismondi, M. Piccoli, and G. Santoni. 1999. Evidence for ␣v␤3 and ␣v␤5 integrin-like vitronectin (VN) receptors in Candida albicans and their involvement in yeast cell adhesion to VN. J. Infect. Dis. 180: 156–166. Tobo ń, A. M., C. A. Agudelo, M. L. Osorio, D. L. Alvarez, M. Arango, L. E. Cano, and A. Restrepo. 2003. Residual pulmonary abnormalities in adult patients with chronic paracoccidioidomycosis: prolonged follow-up after itraconazole therapy. Clin. Infect. Dis. 37:898–904. Tronchin, G., K. Esnault, G. Renier, R. Filmon, D. Chabasse, and J. P. Bouchara. 1997. Expression and identification of a laminin-binding protein in Aspergillus fumigatus conidia. Infect. Immun. 65:9–15. Tronchin, G., K. Esnault, M. Sanchez, G. Larcher, A. Marot-Leblond, and J. P. Bouchara. 2002. Purification and partial characterization of a 32kilodalton sialic acid-specific lectin from Aspergillus fumigatus. Infect. Immun. 70:6891–6895. Vicentini, A. P., J. L. Gesztesi, M. F. Franco, W. de Souza, J. Z. Moraes, L. R. Travassos, and J. D. Lopes. 1994. Binding of Paracoccidioides brasiliensis to laminin through surface glycoprotein gp43 leads to enhancement of fungal pathogenesis. Infect. Immun. 62:1465–1469. Wasylnka, J. A., and M. M. Moore. 2000. Adhesion of Aspergillus species to extracellular matrix proteins: evidence for involvement of negatively charged carbohydrates on the conidial surface. Infect. Immun. 68:3377–3384.