INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Candida albicans is one of the most common human opportunistic pathogens and causes a variety of diseases, from superficial candidiasis to systemic infections in immunocompromised hosts (7). The adhesion of C. albicans to host tissues, mediated through binding to various extracellular matrix proteins such as fibronectin (FN) and laminin, is correlated with pathogenicity (4, 6, 12, 14, 19). FN is a major component of the host extracellular matrix that may play an important role in the initiation and dissemination of C. albicans infections (6, 14, 18). As has been found for mammalian cells (23) and some pathogenic bacteria (5, 20), several domains of FN are recognized by C. albicans. Klotz and Smith (13) suggested that the cell-binding domain of FN mediated its binding to C. albicans. Further evidence indicated that peptides containing the RGD sequence from this domain inhibit the binding of C. albicans to FN both in solution and in adhesion assays (19). Additional domains of FN, however, have been found to interact with C. albicans (18). Four domains of FN can interact with C. albicans grown in Sabouraud dextrose medium (18). Among them, a proteolytic fragment from the gelatin- and collagen-binding domain of FN exhibited the highest affinity, i.e., it was as potent as intact FN. However, all of these interactions of C. albicans with FN were studied by using cultures grown in complex media.

We have previously reported that hemoglobin induces a specific enhancement of FN-binding activity in C. albicans grown in defined medium (24). This induction is reversible and is not due to a bridge effect of hemoglobin between a receptor on the organism and FN. In addition, adhesion to immobilized FN was significantly increased for C. albicans grown in hemoglobin-containing medium. Because cells must be growing in the presence of hemoglobin for several hours to induce binding, expression of a specific FN receptor may be induced by hemoglobin. We have now used these defined growth conditions to determine which domain of FN is recognized by the receptors induced by hemoglobin. We report here that binding and adhesion to the cell-binding domain are specifically induced by hemoglobin. This binding was localized to the 9th through 13th type III repeats of the cell-binding domain of FN but does not require the RGD sequence in the 10th type III repeat. Therefore, hemoglobin in defined medium specifically induces a class of receptors in C. albicans with a specificity different from those previously reported for this organism.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and growth conditions. C. albicans ATCC 44807 was used throughout this study. Its binding to FN in Sabouraud medium as well as in the chemically defined media yeast nitrogen base (YNB) and Lee Buckley Campbell broth has been well characterized (18, 24). Cultures were routinely propagated in freshly prepared YNB broth for 48 h at 26°C on a rocking platform except where otherwise indicated. The stationary-phase blastoconidia were aliquoted and frozen at 70°C until used. For each experiment, organisms were thawed, inoculated into 6 ml of YNB medium with or without hemoglobin, and incubated as described above for 20 to 48 h. Under both growth conditions, no germination was found upon microscopic examination.

Proteolytic or recombinant fragments of FN and RGD synthetic peptides. FN was purified from frozen human plasma as previously described (18). Proteolytic fragments of FN were obtained from Telios Pharmaceuticals, Inc., San Diego, Calif., or Gibco BRL, Gaithersburg, Md. Recombinant fragments derived from several domains of FN, expressed in Escherichia coli, were purified and refolded as described previously (22). Construction of the cell-binding domain with the RGD sequence deleted was carried out by oligonucleotide site-directed mutagenesis using the 33-kDa expression vector (22). The origins in the FN sequences of the fragments used in this paper are presented schematically in Fig. 1.

View larger version (11K): [in this window] [in a new window]   FIG. 1.   FN and proteolytic and recombinant fragments derived from FN. I, II, and III represent, respectively, type I, type II and type III repeating motifs. Proteolytic fragments are indicated by the prefix p, and recombinant fragments are indicated by the prefix r. Numbers following p or r indicate the molecular mass in kilodaltons. The arrows indicate the positions of cell adhesion recognition sequences RGD, LDV, and REDV. The asterisk indicates the position of the RGD sequence deletion from the r33 fragment (r33).

FN-binding assay. FN was iodinated to a specific activity of 1 to 2 µCi/µg with Iodogen (Pierce, Rockford, Ill.), and unbound iodine was removed by being passed through a PD-10 column (18). In a typical binding assay, 2 × 10⁶ C. albicans cells were exposed to ¹²⁵I-FN at a final concentration of 0.5 to 1 µg/ml in a total volume of 200 µl of Dulbecco's phosphate-buffered saline (DPBS) without CaCl₂ and MgCl₂-0.1% bovine serum albumin (BSA), pH 6.0, in a polypropylene tube and were incubated for 3 h with shaking on a rotary plate at 160 rpm. The cell suspensions were transferred to microtubes, and blastoconidia were separated from unbound ¹²⁵I-FN by centrifugation through 100 µl of an oil mixture of dibutyl phthalate-dioctyl phthalate (2:1). Radioactive FN bound to the cell pellet was counted in a gamma counter (Packard Instrument Company, Downers Grove, Ill.). For inhibition assays, the binding of radiolabeled protein was determined in the presence of various concentrations of unlabeled FN, proteolytic or recombinant fragments of FN, or synthetic peptides.

Adhesion to immobilized FN and its fragments. Either FN or its fragments were used to coat glass slide wells by adding 300 µl of a solution of FN or FN fragments (0.01, 0.1, 1, or 10 µg/ml) in DPBS without Ca²⁺ and Mg²⁺ (pH 7.5) per well to Chamber slides. The slides were incubated at 4°C overnight (24). Candida cells prepared from cultures with or without hemoglobin in the YNB medium were added to each well at a concentration of 2 × 10⁶ CFU/ml and allowed to incubate at room temperature for 2 h; the wells were then washed three times with DPBS. Attached Candida cells were fixed, stained, and counted. Aggregates of Candida cells larger than four cells were not counted. The numbers of cells attached per square millimeter of surface were determined in triplicate and are presented as the means ± standard deviations (SD).

Biotinylation and extraction of surface proteins of C. albicans. To identify surface proteins of hemoglobin-induced C. albicans that bound FN and the recombinant 40-kDa fragment (r40), a modified biotinylation procedure was used (25). Briefly, 20 ml of Candida cultures grown in 4× YNB broth with hemoglobin at 26°C for 48 h were harvested by centrifugation and the cells were washed twice with DPBS. The cultures were divided into three parts. The cells were incubated with 1 ml of FN or r40 (both at the concentration of 100 µg/ml) in DPBS at room temperature for 2 h with shaking, and the remaining fraction was incubated in PBS without a ligand. The cells were centrifuged to remove excess unbound ligand, and the pellets were suspended in 1 ml of sulfosuccinimidyl-3-(4-hydroxyphenyl) proprionate (sulfo-SHPP; 0.5 mg/ml; Pierce) in 50 mM sodium bicarbonate buffer, pH 7.8, and incubated at room temperature for 1 h with shaking. The pellets were then washed with DPBS (four times, 15 min each) to remove bound FN or fragments. Finally, each fraction of cells was subjected to biotinylation to label surface proteins with lysyl residues that were protected by bound proteins.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

In contrast to the heterogeneous binding of FN to C. albicans grown in Sabouraud broth (18), an analysis of FN binding to C. albicans grown in the defined medium supplemented with 0.1% hemoglobin with the LIGAND program (17, 21) indicated the presence of only one class of binding sites, with a K_a = 2.2 × 10⁷ M¹ (24). Inhibition of hemoglobin-induced ¹²⁵I-FN binding to C. albicans with unlabeled FN was complete and dose dependent (Fig. 2) (24). The dissociation constant for FN binding to C. albicans grown in hemoglobin-supplemented medium is similar to that for FN binding to the low-affinity receptors induced by the complex Sabouraud medium, i.e., K_d values of 4.6 ± 0.5 × 10⁸ M (Table 1) and 1.2 × 10⁷ M (18), respectively. Thus, the low-affinity receptors are up-regulated following growth with hemoglobin.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Competitive displacement of FN and its derived fragments binding to C. albicans. Unlabeled FN (), the proteolytic 120-kDa cell-binding domain (), and proteolytic 30- () and 40-kDa fragments () were used as inhibitors of 125I-FN binding to C. albicans. C. albicans cells (2 × 106), prepared from a 48-h culture grown in YNB medium supplemented with 1 mg of hemoglobin per ml, were incubated with radiolabeled ligand and the indicated concentrations of unlabeled proteins at 26°C for 3 h. Binding is presented as the percentage of that measured in the absence of inhibitors; values are means ± SD (n = 3).

The cell-binding domain of FN mediates hemoglobin-induced binding to C. albicans. Because the binding of FN to C. albicans grown in defined medium in the presence of hemoglobin differed in affinity from that measured in complex medium, we examined whether different domains of FN might mediate these interactions. Several recombinant or proteolytic fragments of FN were used to inhibit the binding of radiolabeled FN to C. albicans (Table 2). Native FN inhibited the binding of radiolabeled ligand with a 50% inhibitory constant (IC₅₀) of 32.5 nM (Table 2). The most potent fragments of FN, the 120-kDa proteolytic fragment [p120] and r40, were derived from the cell-binding domain and had IC₅₀s of 210 and 320 nM, respectively. A proteolytic 30-kDa fragment derived from the collagen-binding domain in the N-terminal domain of FN, which showed the greatest inhibition of FN binding to C. albicans grown in complex medium (18), was less active (Fig. 2), and fragments from the fibrin-binding domain in the N terminus (r12, r18.5, and r31) responsible for FN binding to gram-positive bacteria (5) were also less active (Table 2) than the cell-binding domain.

Adhesion of C. albicans to immobilized FN and its fragments. Since hemoglobin also induced the adhesion of C. albicans to immobilized FN (24), we compared the adhesion of control and hemoglobin-induced cultures to the FN fragments. C. albicans cells grown with hemoglobin demonstrated a significantly greater adhesion to immobilized FN (P < 0.05 by two-tailed t test) than those grown in YNB medium alone (Fig. 3). Consistent with the results obtained from binding assays in solution, hemoglobin also promoted cell adhesion to the p120 and r40 fragments (P < 0.05 [compared to cultures without hemoglobin]) (Fig. 3). As reported previously (18), the 30-kDa fragment from the collagen-binding domain and the r31 fibrin I domain fragment of FN promoted the adhesion of C. albicans, but hemoglobin did not increase adhesion to these fragments. Adhesion to recombinant 12.5- and 18-kDa subfragments of the fibrin I domain was also not influenced by hemoglobin (results not shown). Adhesion of hemoglobin-induced cultures to the p120 and r40 fragments demonstrated a dose dependence similar to that of intact FN, whereas the collagen-binding p30 and fibrin I r31 fragments were less active (Fig. 4).

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Growth of C. albicans in hemoglobin enhances adhesion to the cell-binding domain of FN. Candida cells prepared from cultures with (shaded bars) or without (solid bars) hemoglobin were allowed to adhere to FN or FN fragments applied to Chamber slides as described in Materials and Methods. The numbers of cells attached per square millimeter of surface were determined in triplicate and are presented as the means ± SD. An asterisk indicates that the adhesion induced by hemoglobin significantly differs from that induced by the respective control, with P < 0.05 by a two-tailed t test.

View larger version (16K): [in this window] [in a new window]   FIG. 4.   Dose dependence of cell adhesion to immobilized FN or its fragments. FN () and FN fragments p120 (), p30 (), r31 (), and r40 () were adsorbed at 0.01, 0.1, 1, and 10 µg/ml in DPBS without Ca2+ and Mg2+ (pH 7.5) to Chamber slides by incubation overnight at 4°C. Unbound protein was removed by washing the wells three times with DPBS. Candida cells prepared from cultures with hemoglobin in YNB medium were added to each well. Attached Candida cells were fixed, stained, and counted (n = 3).

Identification of surface proteins that recognize both FN and the r40 fragment. Preliminary experiments demonstrated that the binding of FN requires amino groups on the Candida cell surface receptors, since modification of amino groups by an hydroxysuccinimide ester prevented the binding of ¹²⁵I-FN (25). Based on this observation, ligand protection from chemical modification was used to identify surface proteins that bind to FN and to the 40-kDa recombinant cell-binding domain. Candida cells were incubated with or without FN or the 40-kDa recombinant fragment to allow binding to their receptors on the cell surface. After the removal of excess ligand, unprotected amino groups of surface proteins were modified by using sulfo-SHPP in the presence of bound ligand. The bound ligands were then removed by thorough washing, and the cells were biotinylated. Biotinylation under these conditions was limited to those amino groups that were protected by bound ligand. Cell surface proteins were extracted, and biotinylated proteins were visualized by streptavidin blotting (Fig. 5). When cells were prebound with FN or the r40 fragment derived from FN, protection by 100 µg of either ligand per ml resulted in prominent labeling of a 55-kDa protein (lanes a and c) that was only faintly visible in the cells without protection (lane b) and increased labeling of a 30-kDa protein. This specific labeling contrasted with the prominent band at 110 kDa and diffuse background bands that were equally labeled in the cells with or without ligands.

View larger version (93K): [in this window] [in a new window]   FIG. 5.   Identification of FN-binding proteins by ligand protection. C. albicans cells at late-exponential phase were preincubated without ligand (lane b), with the r40 fragment of FN (lane a), or with FN (lane c) at 100 µg/ml for 2 h, and exposed amino groups were blocked by reacting them with sulfo-SHPP at a concentration of 0.5 mg/ml in 50 mM sodium bicarbonate, pH 7.8, for 1 h with shaking. The bound FN was removed by thorough washing and lysyl residues of surface proteins that were masked by FN binding were then labeled by biotinylation. Cell surface proteins were then extracted and separated, and biotinylated proteins were visualized as described in Materials and Methods.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Previous work in this laboratory showed that growth of C. albicans in Sabouraud broth induces multiple classes of receptors for FN (18). A p30 fragment from the collagen-binding domain near the N terminus of FN had the highest affinity for binding to FN, which was similar to the affinity of native FN. However, fragments from the cell-binding domain were also active as inhibitors of radiolabeled-FN binding to C. albicans grown in Sabouraud broth. The biphasic Scatchard plot obtained under these conditions suggested that lower-affinity receptors may recognize other domains of FN, including the cell-binding domain. In this study, we show that the hemoglobin-induced binding of FN to cultures grown in YNB broth, a defined medium, involves one class of receptors that preferentially recognize the cell-binding domain relative to the collagen-binding or fibrin I domains. Therefore, hemoglobin appears to induce increased expression or activity of a class of receptors specific for the cell-binding domain of FN. Based on the ligand protection data, this recognition of the cell-binding domain may be mediated by 55- and/or 30-kDa proteins on the surface of C. albicans. Because these putative receptors were identified by ligand protection, we cannot determine whether the proteins are newly expressed on the surface or are simply activated to bind FN following growth with hemoglobin. The induced cultures also show an increased adhesion to immobilized FN and to endothelial cells (24), suggesting that such enhancement may have significance in the pathogenesis of systemic candidiasis. The present data demonstrate that this hemoglobin-induced increase in adhesion is also mediated by the cell-binding domain of FN, but not by the collagen-binding or fibrin I domain.

FN is a dimeric protein and contains multiple ligand-binding domains. Recombinant and proteolytic fragments of FN provide powerful tools for locating the domain responsible for these interactions (18, 23). Unlike cells grown in Sabouraud medium, where the collagen-binding domain was recognized a higher affinity than that of the cell-binding domain, growth with hemoglobin in defined medium enhanced FN binding preferentially through the cell-binding domain. This finding is similar to the results of Penn and Klotz obtained with Sabouraud medium (19). As reported previously, however, FN-binding results with the complex Sabouraud medium vary from batch to batch and from laboratory to laboratory (13, 18, 24). Therefore, the binding data generated from cultures grown in complex medium may depend on multiple factors. The presence of multiple binding sites for FN in C. albicans may be characteristic of a successful pathogen, since the pathogenic bacterium Staphylococcus aureus also expresses receptors that interact with several domains of FN (5). Hemoglobin in defined medium consistently induces the expression of a specific receptor in C. albicans that interacts with the cell-binding domain of FN. Therefore, these defined growth conditions will provide a reproducible method to study the role of this domain in the interaction of C. albicans with the extracellular matrix.

It has been proposed that C. albicans interactions with FN are mediated by an integrin-like molecule on the cell surface (14). Lack of a requirement for the RGD sequence in the cell-binding domain for interactions of FN with C. albicans is clearly demonstrated by the equal activities of the 33-kDa FN fragments with and without the RGD sequences. The inability of RGD peptides to influence FN binding supports this conclusion. Although the ₅₁ integrin in mammalian cells requires the RGD sequence for FN binding, other FN-binding integrins such as ₄₁ do not require this sequence (23). Several mammalian ₁ integrins also promote RGD-independent adhesion to FN type III repeats (8). In addition, specific sequences flanking the RGD sequence, such as the PHSRN sequence in the ninth type III repeat, are required for high-affinity interactions of the ₅₁ integrin with FN (1) and may also directly interact with the integrin (16). Activity of the 28-kDa recombinant fragment to inhibit FN binding to C. albicans localizes a primary recognition sequence in repeats 10 to 12, although this recognition does not require the RGD sequence in repeat 10. The higher level of activity of the r40 fragment relative to the 28-kDa fragment, however, suggests that a synergy site for C. albicans may be present in repeats 12 and 13, amino acid residues 1723 to 1851. This putative synergy site is different from that identified for ₅₁ integrin binding (1), since the known synergy sequence is missing in the r28 fragment and the r33 fragment containing this sequence is less active. Folding of a given fragment of FN may also be important for recognition by C. albicans in addition to the possession of a specific peptide sequence. Alternatively, self-association of some of these fragments with the labeled intact FN may contribute to the apparent differences in their inhibitory activities (11, 15). Further work will be needed to identify the specific sequences responsible for hemoglobin-induced binding of FN to C. albicans.

The receptors interacting with the cell-binding domain induced by hemoglobin may also be different from those induced by Sabouraud medium. In the complex medium, r28 (type III_10-12) and p120 had equal potencies as inhibitors, whereas the r40 fragment showed very weak activity. In defined medium supplemented with hemoglobin, however, the r40 fragment is a stronger inhibitor than the r28 fragment. The reason for this discrepancy is not clear and may indicate the involvement of multiple receptors that are differentially induced by the different growth conditions.

The role of RGD peptides in the interaction between FN and C. albicans has been controversial. Comparing epithelial cell adhesion by C. albicans and Candida tropicalis, Bendel et al. (3) found that purified FN or RGD peptides failed to block C. albicans adhesion but that the epithelial adhesion of C. tropicalis was significantly inhibited by these peptides. They concluded that the pathogenic yeasts C. albicans and C. tropicalis recognize distinct RGD ligands containing distinct flanking sequences (2, 9). Other studies demonstrated that radiolabeled FN binding was inhibited by RGD, GRGESP, and GRGDTP, but not by GRGDSP (13). A separate study found that the binding of the p120 cell-binding-domain fragment to C. albicans was inhibited by RGD and an RGD-containing 23-mer FN peptide (19). Using complex medium, however, we failed to demonstrate any significant inhibition by RGD peptides, including an iC3b peptide (18). Furthermore, the deletion of the RGD sequence from the recombinant cell-binding domain did not alter its activity either in defined or complex medium (18). The specific induction of FN binding to C. albicans by hemoglobin in defined medium provides a unique model to clarify the role of RGD peptides and the cell-binding domain in the recognition of FN. As we expected, none of the RGD peptides used inhibited FN binding to C. albicans, and the recombinant domain with the RGD sequence deleted had the same activity as the native sequence for inhibiting FN binding.

The nature of the FN-binding receptors of C. albicans remains unclear. The putative receptor protein protected from chemical modification by both FN and the r40 fragment has a molecular mass lower than those of proteins that were identified as potential FN receptors in uninduced Candida cells by affinity chromatography on immobilized FN (14) or of the recently identified ALA1 adhesin (10). This difference could result from proteolytic degradation during extraction. Thus, the native molecular mass of the receptor identified with FN and the r40 fragment may be larger than 55 kDa. Purification on an FN affinity column also identified the 55-kDa protein as a potential FN receptor (25). Further characterization of these surface proteins at the molecular level should be helpful in defining their functions.