Binding of Extracellular Matrix Proteins to Aspergillus fumigatus Conidia

INFECTION AND IMMUNITY, Dec. 1996, p. 5239–5247 0019-9567/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 64, No. 12

Binding of Extracellular Matrix Proteins to Aspergillus fumigatus Conidia ˜ ALVER,1 JOSE L. LOPEZ-RIBOT,2 M. LUISA GIL,1 M. CARMEN PEN JOSE E. O’CONNOR,3 AND JOSE P. MARTINEZ1* Departamento de Microbiologıá y Ecologıá, Facultad de Farmacia,1 and Departamento de Bioquı´mica y Biologıá Molecular, Facultad de Medicina,3 Universitat de Vale`ncia, Valencia, Spain, and Department of Microbiology and Immunology, Texas Tech University Health Sciences Center, Lubbock, Texas2 Received 4 June 1996/Returned for modification 3 July 1996/Accepted 30 September 1996

As detected by confocal immunofluorescence microscopy, binding of fibronectin and laminin appeared to be associated with the protrusions present on the outer cell wall layer of resting Aspergillus fumigatus conidia. Flow cytometry confirmed that binding of laminin to conidia was dose dependent and saturable. Laminin binding was virtually eliminated in trypsin-treated organisms, thus suggesting the protein nature of the binding site. Conidia were also able to specifically adhere to laminin immobilized on microtiter plates. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting (immunoblotting) with laminin and antilaminin antibody of whole conidial homogenates allowed identification, among the complex array of protein and glycoprotein species, of one polypeptide with an apparent molecular mass of 37 kDa which specifically interacts with laminin. The fact that binding of conidia to soluble or immobilized laminin or fibronectin was inhibited by fibronectin or laminin, respectively, suggests the existence of common binding sites for both ligands on the surface of conidia. Intact conidia were also able to adhere to type I and IV collagen immobilized on microtiter plates; adhesion was found to be dose dependent and saturable. Adhesion to immobilized type I and IV collagen was markedly inhibited by laminin and weakly inhibited by fibronectin. Coincubation of conidia with Arg-GlyAsp (RGD) peptides caused a dose-dependent decrease in binding of cells to immobilized or soluble fibronectin, yet interaction of cells with soluble or immobilized laminin and type I and IV collagen remained unaffected. Interactions described here could be important in mediating attachment of the fungus to host tissues, thus playing a role in the establishment of the disease. several types of collagen are abundant in the interstitium, particularly as part of the reticular fibers (7, 20). The mechanisms of adherence have been studied extensively in pathogenic bacteria (2, 20, 28) and Candida albicans (13, 20), but little is known about these mechanisms in A. fumigatus. Previous studies have shown that in vitro conidia of A. fumigatus are able to adhere to fibrinogen and laminin (1, 5, 6, 10, 26). More recently, we have described the characteristics and specificity of fibronectin binding to the surface of A. fumigatus conidia and reported the characterization of cell wall-related conidial proteins that may represent receptors for fibronectin (24). Accordingly, fibrinogen, laminin, and fibronectin are candidates for mediating adherence of conidia to host tissues and initiation of infection. It is now well established that the susceptibility of a host to A. fumigatus partly depends on the epithelial damage that leads to exposure of the subepithelial basement membrane, increasing the accessibility of laminin and type IV collagen and afterwards that of fibronectin and type I collagen present in the interstitial space (4, 7). The aim of the present study was to further characterize the interaction of A. fumigatus with ECM components (fibronectin, laminin, and type I and type IV collagen), using fluorescence confocal microscopy, flow cytometry, adherence experiments, and ligand affinity blotting. Binding phenomena described in this paper could play an important role in the invasion of host tissues by A. fumigatus and may help in understanding the mechanisms through which this fungal species infects humans.

Aspergillus fumigatus is considered to be the Aspergillus species that is the most pathogenic for humans. This fungus is ubiquitous and causes a variety of allergic or invasive (nonimmunologic) diseases after inhalation of its airborne conidia. The allergic disorders include allergic bronchopulmonary aspergillosis and asthma that affect atopic subjects, while the nonimmunologic diseases include aspergilloma and invasive aspergillosis and affect mainly patients with cystic fibrosis and immunocompromised individuals. Invasive aspergillosis is often fatal, and its prevalence has increased markedly over the past 30 years (3, 4, 9, 11). The development of aspergillosis in an immunodeficient host depends on interactions between fungal and host components. Thus, it has been suggested that adhesion of the conidia, the infectious propagules, to host cells and mucosal surfaces is a crucial step in the establishment of infection (6, 7). Adherence implies that the fungus recognizes ligands (protein or carbohydrate) on the surface of host cells or a constituent of extracellular matrix (ECM) or the fibrin and fibrinogen deposits formed in response to the inflammatory reactions at the surfaces of wounded epithelia. ECM components are intimately associated with host cell surfaces in tissues. There are two major types of matrices: interstitium and basement membrane. Laminin and type IV collagen are found mainly in the subendothelial basement membrane, while fibronectin and

* Corresponding author. Mailing address: Departamento de Microbiologıá y Ecologıá, Facultad de Farmacia, room 3-70, Universitat de Vale`ncia, Avda. Vicente Andre ´s Estelle´s, s/n, 46100-Burjasot, Valencia, Spain. Phone and Fax: 34-6-3864770. Electronic mail address: jose [email protected].

MATERIALS AND METHODS Microorganism and growth conditions. A. fumigatus (strain 2071) obtained from the Coleccio ń Espan õla de Cultivos Tipo was used throughout this work.

5239

5240

GIL ET AL.

INFECT. IMMUN.

FIG. 1. Binding of fibronectin and laminin to A. fumigatus cells as revealed by fluorescence confocal microscopy. Conidia (A) and germ tubes (B) were incubated with fibronectin (I) or laminin (II), rabbit antifibronectin antibody (I) or antilaminin antibody (II), and FITC-labelled goat anti-rabbit immunoglobulins. Serial sections at 0.5-mm intervals for a single conidium (A; panels 1 to 6) and at 0.75-mm intervals for a single germ tube (B; panels 1 to 8), and three-dimensional reconstruction of conidia (A; panels 7) and germ tubes (B; panels 9), were obtained by confocal microscopy, with associated software. Bars, 3 mm.

The organism was grown on Vogel’s N medium (27) for 30 days at 258C. From these cultures, mature conidia were harvested, washed, and quantified as described previously (24) and used directly in subsequent experiments or, alternatively, inoculated at a final concentration of 107 cells per ml into YPD (1% [wt/vol] Bacto Yeast Extract, 2% [wt/vol] Bacto Peptone, 2% [wt/vol] glucose) liquid medium to obtain mycelial growth. Incubation was carried out until the spores germinated (about 8 h at 378C) (23, 24). Fungal elements were harvested by centrifugation and washed twice in phosphate-buffered saline (PBS). Fluorescence confocal microscopy. Conidia (107 cells) were resuspended in 100 ml of PBS containing various amounts of fibronectin or laminin (see below). After incubation for 3 h at 378C with gentle agitation in a gyratory incubator, the conidia were washed three times with PBS, resuspended in rabbit antifibronectin or antilaminin antibody (1:10 dilution) in PBS plus 1% bovine serum albumin (BSA), and incubated for 1 h at 378C in fluorescein isothiocyanate (FITC)conjugated goat anti-rabbit immunoglobulin antibody (1:10 dilution) in PBS plus 1% BSA. Finally, the conidia were washed again with PBS and examined by

confocal microscopy or flow cytometry. Serial sections from fluorescent cells were obtained with an Olympus LSM GB200 laser scanning confocal microscope, using a 488 argon ion laser (Precision Instruments Div., Olympus Corp., Lake Success, N.Y.). The gain settings, etc., were optimized for each image. Serial sections (xy plane) were obtained at 0.5-mm intervals for conidia and at 0.75-mm intervals for young hyphae along the z axis. Three-dimensional reconstructions were obtained by the resident software. Flow cytometry analysis. Binding of laminin to A. fumigatus conidia was analyzed by flow cytometry. Aliquots (containing about 107 cells) of the conidial suspensions were incubated with various amounts of laminin in PBS, and immunofluorescence assays were performed by the procedure described above. Flow cytometry analyses were performed on an EPICS Elite Cell Sorter (Coulter Electronics, Inc., Hialeah, Fla.), using an air-cooled argon ion laser tuned at 488 nm and 15 mW. Cell debris and aggregates were excluded on the basis of the light scatter properties of the particles (12). The flow rate was kept at approximately 500 events (cells) per s. The filter settings for collecting fluorescein isothiocya-

VOL. 64, 1996

BINDING OF ECM PROTEINS TO A. FUMIGATUS CONIDIA

5241

FIG. 2. (A and B) Flow cytometry analysis of laminin binding to conidia; (C and D) influence of trypsin pretreatment of conidia on binding. (A and B) Conidia were incubated with PBS (nonspecific) or in the presence of various amounts of laminin, and immunofluorescence assays were performed as described in Materials and Methods. (C and D) Conidia treated with different trypsin concentrations were incubated with PBS (nonspecific) or in the presence of laminin (100 mg/ml) and assayed for immunofluorescence. The intensity of fluorescence at the surface of conidia was measured by flow cytometry. (A and C) Representative histograms. x axis, log of fluorescence intensity (LIGFL); y axis, number of fluorescent cells. (B and D) Fluorescence mean channel (less background) represented on a linear scale.

nate-related green fluorescence were as follows: 488-nm blocking filter, 550-nm dichroic filter, and 525 (6 5)-nm band pass filter. Green fluorescence was amplified logarithmically. Twenty thousand events were collected as monoparametric histograms of log fluorescence as well as list mode data files. When necessary, histograms of list mode files were analyzed off-line with the Elite Stand-alone software. Equal numbers of cells were processed similarly but with omission of incubation with laminin as negative controls. In some experiments, the conidial suspensions were incubated with increasing amounts (1 to 100 mg/ml) of trypsin for 30 min at 378C. Proteolytic treatment was stopped by three washes with PBS containing 3% BSA. The trypsin-treated cells were washed with PBS and assayed for laminin binding by flow cytometry as described above, using in this case 0.1 mg of laminin per ml for the initial incubation. Competition binding assays. A fixed concentration of fibronectin (500 mg/ml)

was added to tubes containing 107 conidia and various concentrations of laminin to act as the competitor ligand. After 3 h of incubation, binding of fibronectin was assayed by indirect immunofluorescence and flow cytometry. The same experiment was performed for laminin (100 mg/ml) binding, using in this case fibronectin as the competitor ligand. Adherence assays of conidia to immobilized ligands onto microtiter plates. Wells of microtitration plates (Nunc-Immunoplate I [AS Nunc]) were coated with solutions containing different amounts of ECM proteins. Fibronectin and laminin were dissolved in PBS, whereas type I and IV collagens were dissolved in 0.1 and 0.06 N HCl, respectively. Adherence of biotinylated, Extravidinperoxidase labelled conidia to the ECM proteins immobilized onto wells of the microtitration plates was assessed by following the experimental protocol previously reported (24). The intensity of the colored reaction that appeared after incubation with the chromogenic reagent (o-phenylenediamine) was determined

5242

GIL ET AL.

INFECT. IMMUN. aprotinin per ml, and 1 mM phenylmethylsulfonyl fluoride) to give a very dense suspension. An equivalent volume of Ballotini glass beads (0.5 mm in diameter) was then added to the suspension, and conidia were broken by shaking in a Vortex mixer for 30-s periods with 1-min cooling intervals. Cell breakage was assessed by examination in a phase-contrast microscope. After removal of the glass beads, the cell walls were sedimented (1,200 3 g for 10 min) from the cell-free homogenate; washed three times with chilled distilled water; resuspended in 10 mM phosphate buffer (pH 7.4) containing 1% (vol/vol) 2-mercaptoethanol (bME), 5 mg of aprotinin per ml, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride; and incubated for 30 min at 378C in a rotary shaker. The cell walls were subsequently sedimented, and the supernatant fluid was recovered, dialyzed against distilled water at 48C, and concentrated by freeze-drying (bME extract). bME-extracted walls were washed three times with chilled glassdistilled water and then boiled for 5 min with a 2% sodium dodecyl sulfate (SDS) solution in glass-distilled water. After treatment, the suspensions were centrifuged at 10,000 3 g for 15 min. The proteins eventually present in the supernatant were precipitated with 75% (vol/vol) (final concentration) ethanol at 48C for 16 h. The precipitates were recovered by centrifugation at 27,000 3 g for 30 min and resuspended in water (SDS extract). The total protein contents in the different samples were determined by the method of Lowry et al. (18) with BSA as the standard. PAGE and Western blotting (immunoblotting) techniques. SDS-polyacrylamide gel electrophoresis (PAGE) was performed basically as described by Laemmli (15) with minor modifications (8). Electrophoretic transfer (Western blotting) to nitrocellulose paper was carried out as described previously (8). Blotted proteins were assayed for laminin binding as follows. The nitrocellulose membranes were blocked with 3% BSA in 10 mM Tris-hydrochloride buffer (pH 7.4) containing 0.9% NaCl (TBS buffer) for 1 h at room temperature and then incubated for 6 h with agitation in PBS containing laminin (50 mg/ml). After being washed four times (10 min per wash) with TBS buffer containing 0.05% Tween 20 (TBST), the nitrocellulose sheets were incubated for 1 h with agitation with rabbit antilaminin antibody (1:500 dilution) in TBST plus 1% BSA. The blots were washed with TBST and incubated with peroxidase-labelled goat antirabbit immunoglobulin at a 1:3,000 dilution in TBST plus 1% BSA. Finally, the blots were washed again, and reactive bands were developed with hydrogen peroxide and 4-chloro-1-naphthol as the chromogenic reagent. Source of reagents. Human fibronectin and mouse laminin (isolated from a

FIG. 3. (A) Attachment of A. fumigatus conidia to immobilized laminin. Biotin-Extravidin-peroxidase-labelled conidia were allowed to adhere to wells of microtiter plates coated with different laminin concentrations (10 to 100 mg/ml). Nonadherent cells were removed by washing, and the amount of adherent conidia was estimated indirectly by measuring the optical density at 492 nm (OD492) of the colored reaction produced by peroxidase (see Materials and Methods). (B) Specificity of conidial attachment to immobilized laminin. Wells of microtiter plates were coated with a solution of laminin (50 mg/ml). Conidia were allowed to adhere without any inhibitor (experiment 1) or in the presence of soluble laminin at 500 mg/ml (experiment 2), soluble fibronectin at 500 mg/ml (experiment 3), specific antibodies at a dilution of 1:100 (experiment 4) or 1:50 (experiment 5), 100 mM glucose (experiment 6), 100 mM galactose (experiment 7), or 100 mM mannose (experiment 8). The amount of adherent conidia in the wells was estimated. The results shown in both panels are the means of quadruplicate determinations (less background) from three independent experiments with standard deviations. The values in the presence of laminin, fibronectin, antibodies, and galactose are significantly different from the value without inhibitors (P , 0.01).

at 492 nm with an automated plate reader (Labsystems Multiskan MCC/340). Results, expressed as the optical density at 492 nm, are the means for quadruplicate wells (less background) with standard deviations. Negative controls (i.e., attachment of cells to uncoated wells) did not exceed 18% of the corresponding values for adhesion to ligand-coated wells. Statistical analysis of data was performed by means of Dunnett’s t test for multiple comparisons. Preparation of conidial extracts. Conidia were mixed with lysis buffer (100 mM Tris-hydrochloride [pH 7.4], 1 mM EDTA, 5 mM dithiothreitol, 5 mg of

FIG. 4. Identification by ligand affinity blotting of polypeptides from A. fumigatus conidia with the ability to bind laminin. Whole conidial homogenates (lanes 1) and materials released by treatment of isolated conidial cell walls with bME (lanes 2) and with SDS after bME treatment (lanes 3) were analyzed by SDS-PAGE on 5 to 15% gradient gels followed by Coomassie blue staining (A) or, alternatively, were immunoblotted with laminin and antilaminin antibody (B). The positions of standard proteins with known molecular masses (expressed in kilodaltons) run in parallel are shown at the left of each panel. The arrowhead indicates the 37-kDa polypeptide species that bound laminin.


VOL. 64, 1996

5243

mouse Englebreth-Holm-Swarm sarcoma tumor) were obtained from Boehringer Mannheim. Human type I and type IV collagens were from Sigma Chemical Co. Culture medium components were purchased from Difco. Rabbit antifibronectin and antilaminin antibodies, fluorescein-conjugated goat anti-rabbit immunoglobulin antibodies, and RGD peptide were from Sigma Chemical Co. All other chemicals were from Sigma Chemical Co.

RESULTS Confocal microscopy. Binding of fibronectin and laminin to the surface of conidia and germ tubes was examined by immunofluorescence confocal microscopy (Fig. 1). Serial sections through a single resting conidium (Fig. 1A, panels 1 to 6) showed binding of both fibronectin (Fig. 1, part I) and laminin (Fig. 1, part II) to the surface of conidia. Serial sections and three-dimensional reconstructions (Fig. 1A, panel 7) indicated that binding was strongly associated with the protrusions present in the conidial outer cell wall layer. When observations were made with germinated conidia, no reactivity was detected at the surface of hyphal extensions (Fig. 1B, arrows). The mother swollen conidia from which germ tubes emanated showed a patchy fluorescent pattern once again mostly associated with the cell wall protrusions. This was evident both in serial sections (Fig. 1B, panels 1 to 8) and in three-dimensional reconstructions (Fig. 1B, panel 9) of cells. Fluorescence was dependent on the previous interaction of the cells with fibronectin and laminin, since no fluorescent elements were observed when the cells were incubated with the indicator antibodies only. Binding of fibronectin (Fig. 1, part I) and laminin (Fig. 1, part II) to the surface of A. fumigatus conidia appeared to be basically similar, as concluded from the fluorescence patterns observed. Flow cytometry analysis of laminin binding. Analysis of cells incubated in solutions of laminin at various concentrations ranging from 10 to 500 mg/ml revealed that the ligand bound to the cells in a dose-dependent manner. Representative histograms (Fig. 2A) showed that incubation of conidia with laminin resulted in uniform and intense labelling of the cells. The mean channel of fluorescence intensity (linear scale) is represented in Fig. 2B. The intensity of the fluorescence detected at the surface of conidia increased with the concentration of the ligand in the solution, attesting to the saturability of binding, which reached a plateau at a laminin concentration of 0.1 mg/ml. Conidia pretreated with trypsin were analyzed for their ability to bind laminin by flow cytometry (Fig. 2C and D), in an attempt to further characterize the biochemical nature of the binding sites. In this context, treatment with trypsin at concentrations of 1, 25, and 100 mg/ml inhibited binding of laminin to A. fumigatus conidia by 93, 98, and 99%, respectively (Fig. 2C and D). Attachment of conidia to immobilized laminin. To demonstrate the involvement of binding of laminin in the adherence process, an assay was developed to determine the ability of the cells to interact with the ligand immobilized onto wells of microtiter plates. As shown in Fig. 3A, biotinylated, Extravidin-peroxidase-labelled conidia (24) adhered to laminin in a dose-dependent manner. Thus, coating of the wells with increasing concentrations of laminin (from 10 to 100 mg/ml) resulted in an increase in the number of adherent conidia (Fig. 3A). Maximal adhesion was found at a coating laminin concentration of 50 mg/ml (Fig. 3A). Hence, the concentration of 50 mg/ml was selected for all subsequent experiments on attachment to immobilized laminin. We have previously shown (24) that labelling of conidia with the biotin-avidin-peroxidase complex does not affect binding of cells to immobilized ligands such as fibronectin. The adhesion of conidia to immobilized

FIG. 5. Competitive binding of fibronectin and laminin to the conidial surface, analyzed by flow cytometry. (A) Cells were first incubated with increasing concentrations of fibronectin (solid line) or increasing concentrations of laminin along with a fixed concentration (500 mg/ml) of fibronectin (dotted line), followed by incubation with antifibronectin antibodies and FITC-labelled goat anti-rabbit immunoglobulins. (B) Conidia were first incubated with increasing concentrations of laminin (solid line) or increasing concentrations of fibronectin along with a fixed concentration (100 mg/ml) of laminin (dotted line), followed by incubation with antilaminin antibodies and FITC-labelled goat anti-rabbit immunoglobulins. The intensity of fluorescence at the surface of conidia was measured by flow cytometry.

laminin was found to be specific. As shown in Fig. 3B, the number of adherent conidia decreased when binding experiments were done in the presence of 500 mg of soluble laminin per ml (77% inhibition) or in the presence of different concentrations (1:100 and 1:50 dilution) of specific antilaminin antibodies (89 and 92% inhibition). Adhesion to immobilized

laminin was also markedly inhibited by the presence of 500 mg of soluble fibronectin per ml in the assay (61% inhibition). Finally, the effect of several sugars (glucose, galactose, and mannose) which are found in the laminin molecule and which are also present in fungal cell wall (16, 22) on the ability of conidia to bind to the immobilized ligand was also assayed. Coincubation of the cells in 100 mM solutions of glucose and mannose did not significantly modify binding of conidia to laminin (Fig. 3B). Coincubation in 100 mM galactose reduced binding to immobilized laminin only by about 20% (Fig. 3B). Identification of cell components with the ability to bind laminin. Ligand affinity binding experiments with individualized protein components present in different conidial extracts, following the protocol previously reported to identify receptors for fibronectin in A. fumigatus conidia (24), were employed to identify species that were eventually able to bind laminin. Cellfree homogenates (Fig. 4A, lane 1) and cell wall bME (Fig. 4A, lane 2) and SDS (Fig. 4A, lane 3) extracts were analyzed by SDS-PAGE and stained with Coomassie blue. These samples were seen to contain a complex array of polypeptides (16 to 35 species, depending on the extract) with molecular masses ranging from 17 to 150 kDa, which is in agreement with previous observations from our group (24). Binding of laminin was observed with only one component, with a molecular mass of 37 kDa, which was present in the cell-free homogenate (Fig. 4B, lane 1). This was a specific reaction, as indicated by the absence of bands following detection with the specific antilaminin antibody, when previous incubation of the blots in the laminin solution was omitted. Polyclonal antilaminin antibodies added to the reaction mixture blocked binding of the ligand to the separated conidial polypeptides on the nitrocellulose blots, which additionally supported the contention of the specificity of the laminin affinity binding experiments mentioned above. Competitive binding of fibronectin and laminin to conidia. Either laminin or fibronectin was able to inhibit the adhesion of cells to the other immobilized counterpart ligand. We have previously reported that laminin inhibited adhesion to immobilized fibronectin by 74% (24) and fibronectin inhibited adhesion to laminin by 61% (see above). In order to investigate whether these glycoproteins may also compete in solution to bind to the surface of conidia, indirect immunofluorescenceand flow cytometry-based competition binding experiments were developed. Binding of a constant saturating concentration of fibronectin (500 mg/ml) (24) was inhibited by laminin in solution (Fig. 5A) in a concentration-dependent manner. On the other hand, binding of a saturating concentration of laminin (100 mg/ml) was inhibited by fibronectin also in a dosedependent fashion (Fig. 5B). Laminin (at a concentration of 500 mg/ml) inhibited by about 74% fibronectin binding (Fig. 5A), and fibronectin (at a concentration of 500 mg/ml) inhibited by 57% binding of laminin to conidia (Fig. 5B). Attachment of conidia to immobilized collagen. The ability of other proteins present in the ECM (i.e., type I and IV collagen) to promote the adhesion of conidia was investigated.

FIG. 6. Attachment of conidia to immobilized type I (A) and type IV (B) collagen. Cells were allowed to adhere to wells coated with collagen at different concentrations (from 10 to 75 mg/ml). The amount of adherent conidia in the wells was estimated as described in the legend to Fig. 3. (C) Inhibition experiments. Wells were coated with a solution of collagen (50 mg/ml), and conidia were allowed to adhere without any inhibitor (experiment 1) or in the presence of soluble fibronectin at 500 mg/ml (experiment 2) or soluble laminin at 500 mg/ml (experiment 3). The values in the presence of fibronectin and laminin are significantly different with respect to the corresponding value in the absence of inhibitors (P , 0.01). The results are the means of quadruplicate determinations (less background) from two independent experiments with standard deviations.

5244

VOL. 64, 1996


5245

Conidia bound to wells coated with type I and IV collagen (Fig. 6A and B). Coating wells of microtiter plates with protein solutions of increasing concentrations (10 to 75 mg/ml) resulted in an increase in the number of adherent conidia. The concentration of 50 mg/ml, which was selected for subsequent experiments, enabled us to demonstrate the saturability of the interaction with both types of collagen (Fig. 6A and B). Laminin (500 mg/ml) in solution intensely inhibited adhesion of cells to type I and type IV collagen by about 77 and 70%, respectively (Fig. 6C). However, soluble fibronectin assayed at the same concentration than laminin inhibited only very weakly (17% inhibition for both type I and IV collagen) the adhesion of conidia to these components of the ECM (Fig. 6C). Role of the RGD sequence in the interaction of conidia with ECM proteins. To investigate the contribution of the RGD sequence to the interaction of A. fumigatus conidia with ECM proteins, the ability of the synthetic peptide RGD to inhibit these interactions was investigated. RGD peptide at 0.5 and 1 mg/ml did not cause a significant reduction in conidial adherence to immobilized laminin or type I and IV collagen (data not shown). In contrast, the same RGD concentrations indicated above respectively inhibited by 34 and 41% adherence of conidia to immobilized fibronectin (Fig. 7A). The effect of coincubation of spores with RGD in their ability to bind soluble laminin and fibronectin was assayed by indirect immunofluorescence and flow cytometry. Once again, only binding of fibronectin was reduced by about 37 and 25% by the RGD peptide at concentrations of 0.1 and 1 mg/ml, respectively (Fig. 7B). DISCUSSION

FIG. 7. (A) Effect of RGD peptide on attachment of conidia to immobilized fibronectin. Wells were coated with a solution of fibronectin (50 mg/ml), and conidia were allowed to adhere without any inhibitor or in the presence of RGD solutions containing 0.5 or 1 mg of the peptide per ml. The amount of adherent conidia in the wells was estimated as described in the legend to Fig. 3. The results are the means of quadruplicate determinations (less background) from two independent experiments with standard deviations. The values in the presence of RGD are significantly different with respect to the corresponding value in the absence of RGD (P , 0.01). (B) Effect of RGD peptide in binding of soluble fibronectin to conidia. Cells were first incubated with RGD at 0.1 and 1 mg/ml along with fibronectin (500 mg/ml) and then with antifibronectin antibodies and FITC-labelled goat anti-rabbit immunoglobulins. The intensity of fluorescence at

Attachment of fungal elements to animal cells and tissues is mediated by complementary molecules exposed on both parasite and host cellular surfaces. In this context, interactions between A. fumigatus conidia (which are the infectious airborne form of the fungus) and components of the ECM are thought to be the first and crucial step for establishment of aspergillosis (6, 7, 26). In this communication binding of fibronectin and laminin to the surface of conidia was examined by fluorescence confocal microscopy. Binding of both ligands appeared to be exclusively associated with the conidial cells, while hyphae emanating from swollen conidia did not exhibit the ability to bind either fibronectin or laminin. Results presented here extend our previous single-focal-plane immunofluorescence observations on interaction of fibronectin (24) and confirm observations made by Tronchin et al. (26) using immunoelectron microscopy with respect to laminin binding to conidia, providing more detailed knowledge about distribution of binding sites for both ligands at the outer cell wall layer of resting conidia because of the serial sections obtained by the confocal microscopy technique used here. The similar cell surface-bound fluorescence intensity observed following incubation of conidia with the corresponding ligand-antiligand antibody combination indicated that binding sites for laminin and fibronectin are present over the entire conidial surface, yet the patchy fluorescent pattern that was readily discernible in the three-dimensional reconstructions from serial sections suggests that adhesins or receptors for both ligands are particularly abundant at the protru-

the surface of conidia was measured by flow cytometry. (B1) Representative histograms. x axis, log of fluorescence intensity (LIGFL); y axis, number of fluorescent cells. (B2) Fluorescence mean channel (less background) represented on a linear scale.

5246

GIL ET AL.

sions of the outer conidial cell wall layer, which are expected to be the first areas through which the fungal cells contact the substrate. These areas with a higher density of binding could arise either during synthesis of the cell wall by asymmetrical delivery of the receptors or as a result of rearrangement through the inner cell wall layers to the protrusions as a result of binding to the appropriate ligand (similar to the capping phenomena observed in higher eukaryotic cells). Clustering of receptors for ECM components such as laminin and fibronectin at the cell protrusions could increase the security of the interaction at the point of initial contact between A. fumigatus conidia and the host tissues. Binding of laminin to conidia was further investigated by flow cytometry analysis, which revealed that binding of laminin was saturable. The effect of trypsin treatment on the ability of conidia to bind laminin, as revealed by immunofluorescence and flow cytometry, suggested the protein nature of binding sites present at the conidial surface. However, several monosaccharides tested did not inhibit (glucose and mannose) or slightly reduced (galactose) binding of laminin to cells. Although these observations support the contention of a conidial protein receptor or binding site for laminin, the possibility that digestion with trypsin releases protein moieties or polypeptide fragments containing laminin binding motifs defined by sugars different from those tested cannot be completely ruled out. On the other hand, adhesion of conidia to immobilized laminin was found to be dose dependent and significant at low coating concentrations and to be specific, as demonstrated by the inhibition caused by soluble laminin and antilaminin antibodies. Intact conidia were treated with different chemical and/or enzymatic reagents in order to release cell surface-bound species that were eventually able to bind laminin, following the protocol previously used by our group to identify receptors for laminin in C. albicans (17). However, when analyzed by SDSPAGE, no polypeptide species were detected in the different extracts obtained, which may be a consequence of the extreme hydrophobicity and resistance to chemical degradation of the outermost cell wall layer of A. fumigatus intact conidia (19, 23, 25). Hence, the experimental procedure previously employed to characterize receptor-like moieties for fibronectin in whole conidial homogenates and bME extracts from isolated conidial cell walls of A. fumigatus (24) was used. In this case, a 37-kDa polypeptide that binds laminin was identified by ligand affinity blotting in whole conidial homogenates but not in the extracts from isolated conidial walls. Two possible explanations may be given for this observation: (i) the concentration of this polypeptide species present in the cell wall extracts is below the detection limit of the ligand affinity binding experimental procedure used, or (ii) the reagents used were unable to solubilize the polypeptide from the wall structure even in isolated cell wall preparations. In any case, since Tronchin et al. (26) have reported the presence of numerous laminin binding sites in the cytoplasm of conidia by means of immunoelectron microscopy, one may speculate on the possibility that the 37-kDa lamininbinding polypeptide detected in whole conidial homogenates is also exposed at the surface of conidial cell, a location which is consistent with a putative role as a receptor for the ligand. In this context, it has to be stressed that our group has reported the presence of a 37-kDa high-affinity laminin receptor-like molecule on the surface of C. albicans yeast cells (17). The possible relationship between these two 37-kDa laminin-binding species from A. fumigatus and C. albicans is currently under study in our laboratory. Since collagen is the main constituent of the ECM and represents a major target site for binding of many species of microorganisms (14, 20, 21, 28), it seemed of interest to deter-

INFECT. IMMUN.

mine whether A. fumigatus conidia were also able to interact with this ECM component. We found that actually conidia bound to purified type I collagen (which is the major form in the interstitium) and to type IV collagen (which is the main component of the basement membranes along with laminin). Inhibition experiments revealed that exogenous laminin blocked binding of soluble fibronectin to cells and also reduced the attachment of conidia to immobilized fibronectin. Likewise, coincubation of conidia with fibronectin inhibited binding of soluble laminin and attachment to immobilized laminin. These findings suggested that these ECM components share to some extent the same receptor-like molecule(s) or that binding sites for both ligands are in close proximity at the surface of conidial cells. However, it seems unlikely that they completely share the same receptors or binding sites. Thus, differentmolecular-mass components that may represent receptors for fibronectin (23- and 30-kDa polypeptides) (24) and laminin (the 37-kDa species described in this paper) have been characterized in conidial homogenates and bME extracts from isolated conidial cell walls. Moreover, soluble laminin reduced the attachment of conidia to immobilized collagen while soluble fibronectin did not, and attachment of A. fumigatus conidia to fibronectin was found to be partly mediated by the integrin recognition RGD sequence, since in our experimental conditions the RGD peptide inhibited both adhesion and binding to this ligand by approximately 40 and 30%, respectively. However, adhesion of conidia to other RGD-containing ECM components tested, such as laminin and type I and IV collagen, was not affected by coincubation of the cells with the RGD peptide. Taken together, these results suggest that A. fumigatus conidia have different adherence systems with more than one receptor for a single ECM protein or with a single receptor that recognizes multiple domains of the different ligands. Laminin and fibronectin and laminin and collagen may share some binding sites. It is well-known that establishment of aspergillosis partly depends on the epithelial tissue damage that frequently accompanies the susceptibility of a host to A. fumigatus infection (preceding viral, bacterial, or parasitic bronchopulmonary infections and neutropenic chemotherapies that also kill epithelial cells [4, 7]). Basement membranes beneath epithelial and endothelial cells may be exposed because of these tissue injuries, increasing the accessibility of laminin and type IV collagen. Thereafter, once the fungus transverses the basement membrane, it will encounter cellular fibronectin and type I collagen in the interstitial space. If the fungus gains access to the intravascular space, it encounters fluid fibronectin. So, elucidating the structural basis of ECM component recognition by conidia may provide new insights into mechanisms of adherence and its role in the pathogenesis of aspergillosis. ACKNOWLEDGMENT The support of grant SAF95-0595 from the CICyT (Plan Nacional de Salud y Farmacia), Ministerio de Educacio ń y Ciencia (Spain), is acknowledged. REFERENCES 1. Annaix, V., J. P. Bouchara, G. Larcher, D. Chabasse, and G. Tronchin. 1992. Specific binding of fibrinogen fragment D to Aspergillus fumigatus conidia. Infect. Immun. 60:1747–1755. 2. Beachey, E. H., C. S. Giampapa, and S. M. Abraham. 1988. Adhesin receptor mediated attachment of pathogenic bacteria to mucosal surfaces. Am. Rev. Respir. Dis. 138:S45–S48. 3. Bodey, G. P. 1988. The emergence of fungi as major hospital pathogens. J. Hosp. Infect. 11:411–426. 4. Bodey, G. P., and S. Vartivarian. 1989. Aspergillosis. Eur. J. Clin. Microbiol. Infect. Dis. 8:413–437. 5. Bouchara, J. P., A. Bouali, G. Tronchin, R. Robert, D. Chabasse, and J. M.

VOL. 64, 1996

6. 7. 8. 9.

10.

11. 12. 13. 14. 15. 16.


Senet. 1988. Binding of fibrinogen to the pathogenic Aspergillus species. J. Med. Vet. Mycol. 26:327–334. Bouchara, J. P., G. Tronchin, and D. Chabasse. 1994. Mechanisms and implications of the adherence phenomenons in Aspergillus fumigatus. Pathol. Biol. 42:640–646. Bouchara, J. P., G. Tronchin, G. Larcher, and D. Chabasse. 1995. The search for virulence determinants in Aspergillus fumigatus. Trends Microbiol. 3:327–330. Casanova, M., M. L. Gil, L. Carden ˜ oso, J. P. Martıńez, and R. Sentandreu. 1989. Identification of wall-specific antigens synthesized during germ tube formation by Candida albicans. Infect. Immun. 57:262–271. Cohen, J. 1991. Clinical manifestations and management of aspergillosis in the compromised patient, p. 117–152. In D. W. Warnock and M. D. Richardson (ed.), Fungal infection in the compromised patient, 2nd ed. John Wiley & Sons, Chichester, United Kingdom. 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. Fanti, F., S. Conti, L. Campani, G. Morace, G. Dettori, and L. Polonelli. 1989. Studies on the epidemiology of Aspergillus fumigatus infections in a university hospital. Eur. J. Epidemiol. 5:8–14. Guasch, R. M., C. Guerri and J. E. O’Connor. 1995. Study of surface carbohydrates on isolated Golgi subfractions by fluorescent-lectin binding and flow cytometry. Cytometry 19:112–118. Hostetter, M. K. 1994. Adhesins and ligands involved in the interaction of Candida albicans with epithelial and endothelial surfaces. Clin. Microbiol. Rev. 7:29–42. Kostrzynska, M., C. Schaleń, and T. Wadstro¨m. 1989. Specific binding of collagen type IV to Streptococcus pyogenes. FEMS Microbiol. Lett. 59:229–234. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680–685. Latge´, J. P., H. Kobayashi, J. P. Debeaupuis, M. Diaquin, J. Sarfati, J. M. Wieruszeski, E. Parra, J. P. Bouchara, and B. Fournet. 1994. Chemical and immunological characterization of the extracellular galactomannan of Aspergillus fumigatus. Infect. Immun. 62:5424–5433.

Editor: T. R. Kozel

5247

17. 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. 18. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265–275. 19. Parta, M., Y. Chang, S. Rulong, P. Pinto-DaSilva, and K. J. Kwong-Chung. 1994. HYP1, a hydrophobin gene from Aspergillus fumigatus, complements the rodletless phenotype in Aspergillus nidulans. Infect. Immun. 62:4389– 4395. 20. Patti, J. M., B. L. Allen, M. J. McGavin, and M. Ho ö¨k. 1994. MSCRAMMmediated adherence of microorganisms to host tissues. Annu. Rev. Microbiol. 48:585–617. 21. Patti, J. M., T. Bremell, D. Krajewska-Pietrasik, A. Abdelnour, A. Tarkowski, C. Rydeń, and M. Ho ö ¨k. 1994. The Staphylococcus aureus collagen adhesin is a virulence determinant in experimental septic arthritis. Infect. Immun. 62:152–161. 22. Paulsson, M. 1992. Basement membrane proteins: structure, assembly, and cellular interactions. Crit. Rev. Biochem. Mol. Biol. 27:93–127. 23. Pen ˜ alver, M. C., M. Casanova, J. P. Martıńez, and M. L. Gil. 1996. Cell wall protein and glycoprotein constituents of Aspergillus fumigatus that bind to polystyrene may be responsible for the cell surface hydrophobicity of the mycelium. Microbiology 142:1597–1604. 24. 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. 25. Thau, N., M. Monod, B. Crestani, C. Rolland, G. Tronchin, J. P. Latge´, and S. Paris. 1994. rodletless mutants of Aspergillus fumigatus. Infect. Immun. 62:4380–4388. 26. Tronchin, G., J. P. Bouchara, G. Larcher, J. C. Lissitsky, and D. Chabasse. 1993. Interaction between Aspergillus fumigatus and basement membrane laminin: binding and substrate degradation. Biol. Cell. 77:201–208. 27. Vogel, H. J. 1964. Distribution of lysine pathways in fungi: evolutionary implications. Am. Nat. 98:435–466. 28. Westerlund, B., and T. K. Korhonen. 1993. Bacterial proteins binding to the mammalian extracellular matrix. Mol. Microbiol. 9:687–694.