The aspartate proteinase inhibitor pepstatin A has been shown previously to reduce the adherence of Candida albicans yeast cells to human surfaces. This suggests that in addition to their presumed function facilitating tissue penetration, the secreted aspartate proteinases (Saps) of this fungal pathogen may have auxiliary roles as cellular adhesins. We therefore examined the relative adherence of yeast cells of a parental wild-type strain of C. albicans in relation to yeast cells of strains harbouring specific disruptions in various members of the SAP gene family in an otherwise isogenic background. The adhesiveness of Δsap1, Δsap2 and Δsap3 null mutants and a triple Δsap 4–6 disruptant was examined on three surfaces – glass coated with poly-l-lysine or a commercial cell-free basement membrane preparation (Matrigel) and on human buccal epithelial cells. Pepstatin A reduced adherence to all surfaces. Adherence of the each of the single SAP null mutants to these three substrates was either reduced or not affected significantly compared to that of the parental strain. The adherence of the Δsap4–6 mutant was reduced on poly-l-lysine and Matrigel, but increased on buccal cells. The results suggest that in addition to a primary enzymatic role, various SAPs may also act singly or synergistically to enhance the adhesiveness to C. albicans cells to certain human tissues.
Candida albicans is a commensal organism commonly found in the oral cavity, gastrointestinal tract, female genital tract and occasionally on the skin. It can cause infections ranging from superficial mucosal lesions to life-threatening systemic diseases in immunocompromised patients . Putative virulence factors of C. albicans include cell wall adhesins, phenotypic switching, hypha formation, thigmotropism and the secretion of proteinases and other hydrolytic enzymes [1–4]. Production of extracellular proteinases was first reported by Staib  and has been a major focus for studies of the virulence of Candida species . These secreted aspartate proteinases (Saps) are characterised by an acidic optimum pH [7, 8], a wide substrate specificity and sensitivity to the inhibitor pepstatin A, a hexapeptide from Streptomyces. To date, eight gene sequences of putative secreted proteinases have been reported [6, 10]. These encode a family of isoenzymes whose members are expressed differentially according to the strain, growth medium, temperature, switch phenotype, and cellular morphology [6, 11]. Experiments in vitro suggest that yeast cells express at least three genes (SAP1, SAP2 and SAP3) while SAP5 and SAP6 of the highly homologous SAP4–6 genes were co-expressed in hyphal cells cultivated on serum-containing medium [11, 12]. Proteinase-deficient mutants, created by chemical mutagenesis, were found to have attenuated virulence in mice [13, 14] suggesting these enzymes may be involved in the process of host invasion. The role of SAPs as virulence factors was confirmed recently by the demonstration that the virulence of null mutants of SAP1, SAP2, SAP3 and a triple mutant of SAP4–6 were all attenuated in systemic mouse and guinea pig model infections [15, 16].
Most studies have emphasised the likely role of these proteinases in facilitating penetration via tissue necrosis or in the destruction of immunoglobulins. However, the Sap inhibitor pepstatin A has been reported to block invasion of host tissue by hyphal cells at neutral pH  and to reduce the adherence of yeast cells to various cell types [18, 19], suggesting a possible non-enzymatic role for the enzymes in adherence. Therefore we used various SAP gene null mutants to investigate the putative role of Saps in the adhesion of C. albicans to three different substrates: glass coated with poly-l-lysine, or a layer of endothelial cell basement membrane components (Matrigel) and to human buccal epithelial cells (BECs). The results support the view that certain Sap isoenzymes may have auxiliary roles as cellular adhesins.
2Materials and methods
The C. albicans clinical isolate SC5314 was supplied by Bill Fonzi  and was maintained by weekly sub-culture on Sabouraud dextrose agar (Oxoid) at 30°C. The null mutant strains, Δsap1, Δsap2, Δsap3, and Δsap4–6 triple mutant , were prepared by gene disruption using the hisG-URA3-hisG‘ura-blaster’ method , in the ura3−, CAI-4 strain that was derived from SC5314 (Table 1).
Single colonies of C. albicans were inoculated into 10-ml aliquots of medium, either 0.67% (w/v) YNB (yeast nitrogen broth, Difco) containing 50 mM glucose or 0.67% (w/v) YNB with 500 mM galactose and incubated for 24 h with shaking (300 rpm) at 30°C. Cultures were diluted 1:10 into fresh medium and incubated for a further 24 h to obtain stationary phase cultures. Cells were harvested by centrifugation (3600 rpm, 5 min, 4°C), washed twice in 10 ml PBS (phosphate-buffered saline, Sigma, Poole, UK), pH 7.4, and resuspended in 10 ml PBS. Cell densities were then determined by counting using an improved Neubauer haemocytometer and diluted to a final concentration of 1×108 cells ml−1 in PBS .
2.3Adherence to poly-l-lysine- or Matrigel-coated slides
Glass microscope slides were degreased and sterilised by dipping in ethanol and flaming. To prepare poly-l-lysine-coated surfaces the slides were incubated in 0.01% (w/v) filter-sterilised poly-l-lysine (MW 150 000–300 000, Sigma, Poole, UK) for 15 min at room temperature and the slides dried horizontally for 30 min in a sterile air flow cabinet. To prepare Matrigel-coated surfaces the sterile slides were cooled to 5°C and coated evenly with 50 μl of Matrigel (Collaborative Research, Lexington, KY, USA) and then allowed to set at room temperature for 30 min.
Suspensions of yeast cells, prepared as described above, were added to the slide (0.5 ml for poly-l-lysine surfaces, 50 μl for the Matrigel surface) for 30 min at room temperature. Non-attached and loosely attached cells were washed off with PBS and the number of cells in 10 microscope fields was counted (each field of view being equivalent to 1.52 mm2 at a magnification of ×400).
2.4Adherence to buccal epithelial cells
The method of Kimura and Persall  was used with slight modifications . Suspensions of human buccal epithelial cells (1×105 cells ml−1 in PBS, 0.1 ml) and yeast cells (1×108 cells ml−1 in PBS, 0.1 ml) were mixed and incubated at 37°C for 45 min with gentle shaking. After adding 2 ml of PBS per sample the mixture was filtered through Nuclepore polycarbonate filters (12 μm pore size, 25 mm diameter, Costar, High Wycombe, UK) and the filter washed with 30 ml PBS to remove any unattached yeast cells. The filters were air-dried and stained by the Gram procedure. The number of adherent yeasts per 100 epithelial cells was counted for each filter. Triplicate filters were prepared for each assay, and the assay repeated on subsequent days to test for day-to-day variation in the epithelial cells. Control slides were incubated with PBS only to verify the absence of natural carriage of Candida cells on donor BECs. In experiments in which aspartate proteinase activities were inhibited with pepstatin A stock solutions of 10 mg ml−1 pepstatin A (Sigma, Poole, UK) were prepared in 0.005 N NaOH and filter sterilised. This was diluted by a factor of 1/1000 into the yeast suspension and incubated at 37°C for 30 min before proceeding with the adhesion assays as described previously.
The mean numbers of adherent yeast cells to various substrates were compared by Student's t-test. The human buccal epithelial cell assay was performed on two consecutive days, therefore a two-way analysis of variance was used to determine whether significant variations in adherent cell populations occurred (at the 5% level of significance) between days and for different strains.
3.1Inhibition of adherence of wild-type C. albicans by pepstatin A
The effect of the aspartate proteinase inhibitor pepstatin A on the adhesion of the parental C. albicans strain SC5314 to poly-l-lysine- and Matrigel-coated microscope slides and to human BECs was determined (Fig. 1). Poly-l-lysine-coated slides represent a homogeneous, positively charged surface. Matrigel is a solubilised extract of basement membrane from Engel-Horm Swarm (EHS) transplantable mouse tumour and therefore contains a variety of important components of endothelial cell layers. Human buccal epithelial cells are colonised readily by C. albicans in pathological and sub-clinical conditions. Strain SC5314 is known to have a full complement of at least eight SAP genes .
Yeast cells that had been grown in YNB with 50 mM glucose or galactose as the carbon source were used as an inoculum. Cells were pre-incubated in 10 μg ml−1 pepstatin A for 30 min prior to exposing the yeast cells to the surfaces. The results confirmed previous observations that galactose-grown yeast cells were significantly more adherent to BECs . Galactose-grown cells were however only marginally more adherent to Matrigel and not significantly more adherent to poly-l-lysine (Fig. 1 and Table 2).
Table 2. Adherence of yeast cells of wild-type and various Sap mutants of C. albicans to three substrates
aField of view at 400× magnification; one F.O.V. is equivalent to 1.52 mm2.
bHuman buccal epithelial cell assay as described in ; assays were performed in triplicate on consecutive days, data are average values from both days from a total of six replicates.
cYeast cells grown as the inoculum were grown to stationary phase in YNB+glucose or galactose as the source of carbon.
dParental strain and progenitor of all Δsap mutant strains.
All values are means±S.D. *Values differing significantly (P≤0.05) from adherence values to the parental strains grown on the same carbon source.
Buccal epithelial cells
Adhering cells/10 F.O.Va
Adhering cells/10 F.O.Va
Adhering cells/100 BECsb
Pre-treatment of cells with pepstatin A led to a significant decrease in the numbers of cells adhering to poly-l-lysine and Matrigel, but did not affect adherence to BECs (Fig. 1). The largest percentage decrease was observed on poly-l-lysine surfaces where adherence of galactose- and glucose-grown cells was reduced to 18.5 and 38% of the control values. Therefore pepstatin A reduced the number of yeast cells which adhered to two of the three substrates, suggesting that Saps may play a role in C. albicans adhesion to some, but not all surfaces.
3.2Adhesion of Sap null mutants
The contribution of individual Sap isoenzymes to adhesion was investigated with a panel of SAP null mutant strains –Δsap1, Δsap2, Δsap3, and Δsap4–6. The adherence of each strain to poly-l-lysine, Matrigel and BECs was determined for yeast cells grown either in 50 mM glucose or 500 mM galactose (Table 2). Galactose-grown cells were either equally adherent or more adherent than glucose-grown cells. Increased adherence of galactose-grown cells relative to glucose-grown cells was greatest for adherence to BECs. On Matrigel, galactose-grown cells were approximately 50% more adherent than glucose-grown cells for all strains and there was no significant difference in the relative adherence of the parental strain and any of the mutants grown on either carbon source, with the exception of glucose-grown Δsap3, which was less adherent. On poly-l-lysine the parental galactose-grown strain was marginally less adherent than glucose-grown cells and Δsap1, sap3 and sap4–6 mutants were marginally less adherent than the parent when grown on glucose (Table 2). There was no significant difference in the adherence of the various Δsap mutants to poly-l-lysine-coated microscope slides when cells were grown on galactose (Table 2).
Adhesion assays to human BECs were repeated on two subsequent days to accommodate variations in the donor cell status. However, two-way analysis of variance showed no significant differences in adhesion on different days for all strains. Donor epithelial cells did not show any natural Candida colonisation. Fewer cells of the Δsap1, Δsap2 and Δsap3 strains adhered to BECs compared to the parental strain (60±22%, 86±55% and 65±8% respectively) when grown in YNB+glucose. However, the Δsap4–6 mutant was significantly more adherent (316±68%) (Table 2). This trend was also observed for galactose-grown cells where the adherence relative to the parental strain for Δsap1, Δsap2 and Δsap3 was 86±29%, 91±23% and 88±15% respectively while the relative adherence of Δsap4–6 was 232±36% (Table 2).
Our data, based on examination of the adherence of Sap null mutant strains of C. albicans and the effects of the Sap inhibitor pepstatin A, suggest that secreted Sap proteins may have an auxiliary role to their function in proteolysis in serving as adhesins to certain surfaces. Adhesion of C. albicans is known to be a complex, multifactorial property dependent on a multiplicity of recognition systems and surface receptors such as surface integrin-like molecules [23, 24]. In addition, general biophysical properties of the cell such as cell surface hydrophobicity  and electrostatic charge [26, 27] can affect adhesion. Therefore it is not surprising that pepstatin A treatment and the disruption of individual SAP genes had only a partial effect on yeast cell adherence. Adhesion is also known to vary significantly between different strains and to be dependent on growth state, growth form, and the carbon source for growth [21, 28]. These factors presumably affect the expression of these various surface adhesins . We used an established protocol for testing cellular adherence to BECs and modified this to examine adherence to the basement membrane preparation Matrigel, and to the positively charged polymer poly-l-lysine. Our analysis was also limited to uniform inocula of stationary phase yeasts cells. Sap production in stationary phase yeast cells is likely to occur at a basal level [11, 29] and other aspects of the prevailing conditions in the adhesion assays are likely to suboptimal for the expression of the various Sap isoenzymes. Several previous reports have demonstrated pepstatin A inhibition of adherence for cells grown in media that would not be expected to be proteinase-inducing [17, 19, 30] again suggesting that some degree of proteinase secretion occurs even in non-protein-containing media. In the present study the changes in relative adherence measured for the various Sap isoenzymes also provides indirect evidence that SAP gene expression does occur under the conditions examined.
In yeast cells SAP2 mRNA is normally the major SAP transcript and activity while SAP1 and SAP3 are expressed to lesser levels. In contrast, SAP5 and SAP6 genes are expressed preferentially in hyphal cells grown in the presence of exogenous protein [11, 12]. The Δsap4–6 deletions were found to affect yeast growth on bovine serum albumin suggesting that SAP4–6 may play a role in the induction of SAP2 expression in yeast cells . Therefore, all of the SAP1–6 genes are possible candidate adhesins for the yeast form. It is not known if all Sap isoenzymes are equally sensitive to pepstatin A. We showed that pepstatin A antagonises adherence of C. albicans to each of the surfaces examined and demonstrated that no one isoenzyme has a dominant role in mediating adhesion. It is possible that various Sap isoenzymes act synergistically in mediating adhesion. The differential adhesion effects observed on the three surfaces examined also suggest that the isoenzymes do not contribute equally and that the significance of proteinase-mediated cell attachment varies according to the underlying surface.
Pepstatin A has been shown elsewhere to reduce the number of cells adhering to mucosal surfaces and various cell types [11, 18, 19, 30]. Pepstatin A is thought to be an inhibitor of the transition state aspartate proteinases  suggesting that enzymatic activity of the Saps is important in the formation of adhesive bonds under the conditions tested. In our analysis the largest reduction in numbers was observed with the positively charged poly-l-lysine surface. It is possible that adhesion to this material is normally promoted by proteolytic modification of the mannoproteins on the fungal cell surface. This could result in the unmasking of charged moieties such as phosphomannan leading to an increase in adherence via non-specific ionic interactions.
A somewhat smaller reduction in adherence in the presence of pepstatin A was observed for cells attaching to Matrigel and human BECs. It is likely that interactions between C. albicans yeast cells and these surfaces involves specific receptor-ligand interactions such as those mediated via specific mannan residues, RGD peptides and integrin interactions  as well as hydrophobic interactions , all of which are likely to be unaffected by pepstatin A.
We attempted to examine the role of individual Sap isoenzymes of C. albicans using strains harbouring specific SAP gene disruptions [15, 16]. Similar trends were observed on the poly-l-lysine- and Matrigel-coated surfaces. For these surfaces there was no significant difference between adherence of any of the galactose-grown null mutants and the parental strain. Growth on galactose enhances adherence by increasing the fibrillar wall mannoprotein content and consequently hydrophobicity . It is possible that the substantial increase in adhesiveness of cells grown on galactose masked any changes in adhesion due to the deletion of SAP genes. A decrease in adherence was, however, noted when the mutant strains were grown in YNB-glucose. The Δsap1, Δsap3 and Δsap4–6 null mutant strains had significantly reduced adherence to the poly-l-lysine surface while the Δsap3 strain had decreased adherence on Matrigel.
On human buccal epithelial cells Δsap1, Δsap2 and Δsap3 mutants were all less adherent. However, adherence of the Δsap4–6 strain was increased substantially for both glucose- and galactose-grown cells compared to the parental strain. This marked phenotype reinforces the observation made previously that expression of these genes is not restricted to hyphal cells . The enhanced adherence of the Δsap4–6 mutant may relate to the absence of a proteolytic activity that normally removes or degrades cell surface components on the yeast or epithelial cells that are inhibitory to the host-fungus recognition and adhesion mechanism. Alternatively the SAP4–6 gene products may bind to proteins that would otherwise promote adhesion. What ever the mechanism, this effect was only found for interaction with buccal cells.
Overall the results suggest that Sap isoenzymes are one component of a multiplicity of factors that influence the adhesion of C. albicans to surfaces.
We thank the BBSRC (ROPA Grant 1/CEL 04556) and the Wellcome Trust (039643/Z/93/Z/1.27) for financial support. N.A.R.G. acknowledges gratefully the support of a Royal Society/Leverhulme Senior Research Fellowship.