Dissimilar attenuation of Candida albicans virulence properties by human immunodeficiency virus type 1 protease inhibitors.

The secreted aspartyl proteinase (Sap) of Candida albicans, which is believed to represent an important virulence factor of this opportunistic yeast, and the human immunodeficiency virus type 1 (HIV-1) protease, which is obligatory for the production of infectious virions, both belong to the same family of aspartyl proteinases. We have previously shown that the HIV-1 protease inhibitor Indinavir directly inhibits secretion and proteinase activity of Sap in a dose-dependent manner. Furthermore, at very high concentrations, viability of C. albicans is markedly reduced by Indinavir, indicating that HIV-1 protease inhibitors may possess antifungal activity. We thus proposed that these drugs may add to the resolution of mucosal candidiasis in HIV-1 infected subjects. We have now compared three different HIV-1 protease inhibitors. The rank order of Sap inhibition, already significant at 0.1 mg/ml for all protease inhibitors, was Ritonavir > Indinavir > Saquinavir. However, the cross-reactivity of Ritonavir to pepsin was also more pronounced compared with the other two. Indinavir did not affect Candida viability at concentrations up to 1 mg/ml, in line with our previous study. In contrast, at this concentration Saquinavir was even fungicidal as assessed by three different viability assays (colony formation assay, MTT assay, propidium iodide staining) whereas Ritonavir significantly affected the mitochondrial activity only (MTT assay). No influence on Candida viability was observed for any of the three at concentrations of 0.1 mg/ml or lower. It remains to be examined whether HIV-1 protease inhibitors or derivatives thereof may be suitable for in vivo therapy of subjects suffering from mucosal candidiasis resistant to current antimycotics.     

Critical role of germ tube formation in the pathogenesis of candidal vaginitis.

A variant strain of Candida albicans incapable of hyphal production at 37 degrees C was used to study the role of germ tube formation in the pathogenesis of experimental vaginal candidiasis in rats. No difference was observed in the in vitro adherence at 25 degrees C of blastoconidia of the variant strain to vaginal epithelial cells when compared with the parent wild-type, germ tube-producing strain and multiple clinical isolates of C. albicans. However, after exposure to conditions favoring germ tube production, the adherence of the variant strain to epithelial cells was significantly less than that of germinated strains (P less than 0.01). In vivo animal studies revealed that the variant strain was less likely to result in vaginal colonization and infection than the wild-type strain and the other clinical isolates. Furthermore, infection, when established, was milder, often transient, and with significantly lower titers of cultured vaginal microorganisms obtained by lavage. Electron microscopic studies confirmed the failure of the variant strain to produce hyphae in vivo. The capacity of C. albicans to produce hyphae appears to be an important but nonessential virulence factor in the pathogenesis of candidal vaginitis.     

Intravaginal and intranasal immunizations are equally effective in inducing vaginal antibodies and conferring protection against vaginal candidiasis

Oophorectomized, estrogen-treated rats were immunized by the intravaginal or intranasal route with a mannoprotein extract (MP) or secreted aspartyl proteinases (Sap) of Candida albicans, with or without cholera toxin as a mucosal adjuvant. Both routes of immunization were equally effective in (i) inducing anti-MP and anti-Sap vaginal antibodies and (ii) conferring a high degree of protection against the vaginal infection by the fungus. These data suggest that appropriate fungal antigens and adjuvant can be used to protect against candidal vaginitis, by either route.     

A triple deletion of the secreted aspartyl proteinase genes SAP4, SAP5, and SAP6 of Candida albicans causes attenuated virulence.

Secreted aspartyl proteinases (Saps) from Candida albicans are encoded by a multigene family with at least nine members (SAP1 to SAP9) and are considered putative virulence factors important for the pathogenicity of this human pathogen. The role of Sap isoenzymes in the virulence of C. albicans has not yet been clearly established, and therefore, using recent progress in the genetics of this yeast, we have constructed a panel of isogenic yeasts, each with a disruption of one or several SAP genes. We focused on the construction of a C. albicans strain in which three related SAP genes (SAP4, SAP5, and SAP6) were disrupted. Growth of the delta sap4,5,6 triple homozygous null mutant DSY459 in complex medium was not affected, whereas, interestingly, growth in a medium containing protein as the sole nitrogen source was severely impaired compared to the growth of the wild-type parent strain SC5314. Since the presence of Sap2 is required for optimal growth on such medium, this suggests that Sap4, Sap5, or Sap6 plays an important role for the process of induction of SAP2. When guinea pigs and mice were injected intravenously with DSY459, their survival time was significantly longer than that of control animals infected with the wild-type SC5314. Attenuated virulence of DSY459 was followed by a significant reduction of yeast cells in infected organs. These data suggest that the group of Sap4, Sap5, and Sap6 isoenzymes is important for the normal progression of systemic infection by C. albicans in animals.     

Protective role of antimannan and anti-aspartyl proteinase antibodies in an experimental model of Candida albicans vaginitis in rats.

The role of antibodies (Abs) in the resistance to vaginal infection by Candida albicans was investigated by using a rat vaginitis model. Animals receiving antimannoprotein (anti-MP) and anti-aspartyl proteinase (Sap) Ab-containing vaginal fluids from rats clearing a primary C. albicans infection showed a highly significant level of protection against vaginitis compared to animals given Ab-free vaginal fluid from noninfected rats. Preabsorption of the Ab-containing fluids with either one or both proteins MP and Sap sequentially reduced or abolished, respectively, the level of protection. A degree of protection against vaginitis was also conferred by postinfectious administration of anti-Sap and anti-MP monoclonal antibodies (provided the latter were directed against mannan rather than protein epitopes of MP) and by intravaginal immunization with a highly purified, polysaccharide-free Sap preparation. Postinfectious administration of pepstatin A, a potent Sap inhibitor, greatly accelerated the clearance of C. albicans from rat vagina. No anti-MP or anti-Sap Abs were elicited during a C. albicans vaginal infection of congenitally athymic nude rats. Although they were as able as their euthymic counterparts to clear the primary infection, these animals did not show increased resistance to a rechallenge, demonstrating that induction of anticandidal protection in normal rats was a thymus-dependent Ab response. Overall, our data strengthen the concept that Abs against some defined Candida antigens are relevant in the mechanism of acquired anticandidal protection in vaginitis. The T-cell dependence of this protection may also provide a link between cell-mediated and humoral immunity in vaginal infection.     

Phenotypic switching and its influence on expression of virulence factors by Candida albicans causing candidiasis in human immunodeficiency virus-infected patients

PURPOSE: The purpose of the present study was to determine the degree of expression of virulence factors such as adherence, cell surface hydrophobicity (CSH) and production of proteinase by different morphological forms of Candida albicans causing oral candidiasis in human immunodeficiency virus (HIV)-infected individuals. METHODS: C. albicans 3153A and two strains isolated from oral thrush in HIV infected individuals were induced to undergo phenotypic switching by exposure to UV light and the degree of expression of virulence factors by the different morphological forms was studied. RESULTS: Three different morphological forms of C. albicans were obtained namely, star (S), wrinkled (W) and ring (R) types from the original smooth (O) variety. It was found that proteinase production was greatest with the W type followed by the R type and O type. The S type produced the least proteinase. Expression of cell surface hydrophobicity and adherence was greatest in the O type followed by the R and then the W type and finally the S type. CONCLUSIONS: The differential expression of virulence factors occurs with different phenotypic forms of C. albicans and this may provide a particular morphological type with a distinct advantage over other types in causing candidiasis.     

Clonal and spontaneous origins of fluconazole resistance in Candida albicans

The genotypes and susceptibilities to fluconazole of 78 strains of the human pathogenic yeast Candida albicans were compared. The strains comprised two sets of samples from Durham, N.C.: one from patients infected with the human immunodeficiency virus (HIV) and the other from healthy volunteers. For each strain, the MIC of fluconazole was determined by the standard National Committee for Clinical Laboratory Standards protocol. Genotypes were determined by PCR fingerprinting with five separate primers. The analysis revealed little evidence for genotypic clustering according to HIV status or body site. However, a small group of fluconazole-resistant strains isolated from patients infected with HIV formed a distinct cluster. In addition, two fluconazole-resistant strains were isolated from individuals who never took fluconazole, one from a patient infected with HIV and the other from a healthy person. The results suggest both clonal and spontaneous origins of fluconazole resistance in C. albicans.     

Vaginopathic and proteolytic Candida species in outpatients attending a gynaecology clinic.

Non-pregnant, non-diabetic outpatients were examined for the presence of pathogenic vaginal yeasts to determine if a correlation existed between a specific yeast and clinical disease. Yeasts were isolated as single vaginal species from 186 of 228 subjects with clinically diagnosed candidal vaginitis, as well as from 122 out of 380 asymptomatic, age-matched controls. Apart from Candida albicans and C glabrata, other prevalent species were C krusei, C parapsilosis and Saccharomyces cerevisiae which accounted for 9.2%, 6.0% and 5.4%, and 9.0%, 2.4% and 19.7%, of yeasts from patients and carriers, respectively. Only C albicans and C parapsilosis were significantly more common in those with vaginitis. Only the isolates of these two species secreted aspartyl proteinase in vitro, and the amount of the enzymes secreted by the isolates from patients was significantly higher than that secreted by the isolates from carriers. These two species consistently produced vaginal infection in pseudoestrus rats, whereas none of the non-proteolytic species tested (C glabrata, C krusei, and S cerevisiae) colonised the vagina in these rats. Proteinase secretion correlated with experimental vaginal infection; it could also be a reliable factor for distinguishing clinically active infection from asymptomatic fungal carriage.     

Germ tubes and proteinase activity contribute to virulence of Candida albicans in murine peritonitis

Peritonitis with Candida albicans is an important complication of bowel perforation and continuous ambulatory peritoneal dialysis. To define potential virulence factors, we investigated 50 strains of C. albicans in a murine peritonitis model. There was considerable variation in their virulence in this model when virulence was measured as release of organ-specific enzymes into the plasma of infected mice. Alanine aminotransferase (ALT) and alpha-amylase (AM) were used as parameters for damage of the liver and pancreas, respectively. The activities of ALT and AM in the plasma correlated with invasion into the organs measured in histologic sections and the median germ tube length induced with serum in vitro. When the activity of proteinases was inhibited in vivo with pepstatin A, there was a significant reduction of ALT and AM activities. This indicates that proteinases contributed to virulence in this model. Using strains of C. albicans with disruption of secreted aspartyl proteinase gene SAP1, SAP2, SAP3, or SAP4 through SAP6 (collectively referred to as SAP4-6), we showed that only a Deltasap4-6 triple mutant induced a significantly reduced activity of ALT in comparison to the reference strain. In contrast to the Deltasap1, Deltasap2, and Deltasap3 mutants, the ALT induced by the Deltasap4-6 mutant could not be further reduced by pepstatin A treatment, which indicates that Sap4-6 may contribute to virulence in this model.     

Aspartyl proteinases of Candida albicans and their role in pathogenicity.

Among the putative virulence factors of Candida albicans, secreted aspartic proteinases (Sap, encoded by a family of at least nine genes) continue to attract the attention of many investigators studying the pathogenesis of candidiasis. Several early studies documented a correlation between the levels of Sap secretion and the virulence of different strains, but much stronger support for this role has been provided by more recent data on differential SAP gene(s) expression in ex vivo and in vivo models, the outcome of infections with SAP-deleted mutants, and use of Sap immunogens. In particular, some SAP-deleted strains suffered a substantial loss of virulence, and, more interestingly, this was specifically associated with selected gene products and selected experimental pathologies. Moreover, anti-Sap antibodies have been shown to mediate a degree of protection in an experimental, mucosal candidiasis model. There is now initial evidence that distinct Saps are differentially produced in various Candida illnesses or stages of them. The exact mechanisms of each Sap involvement in any particular Candida disease, with special regard to human infections, and how the immune system deals with Sap, are critical issues for future research. An answer to these questions will possibly facilitate the generation of Sap-based anticandidal drugs or immunotherapeutics.     

Elevated aspartic proteinase secretion and experimental pathogenicity of Candida albicans isolates from oral cavities of subjects infected with human immunodeficiency virus.

Isolates of Candida albicans from the oral cavities of subjects at different stages of human immunodeficiency virus (HIV) infection or uninfected controls were examined for (i) production of aspartic proteinase(s), a putative virulence-associated factor(s); (ii) the presence in the fungal genome of two major genes (SAP1 and SAP2) of the aspartic proteinase family; and (iii) experimental pathogenicity in a murine model of systemic infection. It was found that the fungal isolates from symptomatic patients secreted, on average, up to eightfold more proteinase than the isolates from uninfected or HIV-infected but asymptomatic subjects. This differential property was stably expressed by the strains even after years of maintenance in stock cultures. Moreover, representative high-proteinase isolates were significantly more pathogenic for mice than low-proteinase isolates of C. albicans. The characters high proteinase and increased virulence were not associated with a single molecular type or category identifiable through DNA fingerprinting or pulsed-field electrophoretic karyotype, and both SAP1 and SAP2 genes were present in both categories of isolates, on the same respective chromosomes. In conclusion, our data suggest that during HIV infection more-virulent strains or biotypes of C. albicans which are identifiable by direct analysis of virulence determinants are selected. It also appears that the biotype switch to increased aspartic proteinase and virulence properties occurs before the HIV-infected subject enters the symptomatic stage and overt AIDS.     

Heterogeneity of oral isolates of Candida albicans in HIV-positive patients: correlation between candidal carriage, karyotype and disease stage

Opportunist infections involving Candida albicans often develop in HIV-positive patients and oral lesions tend to become more frequent as the disease progresses. Previous studies have shown contrasting results concerning the variability of the pulsed-field gel electrophoresis (PFGE) subtypes of C. albicans observed in HIV-positive patients. Carriage of C. albicans was determined by an oral rinse technique; 41 strains of C. albicans (78% serotype A and 22% serotype B) were isolated. There was a direct correlation between candidal load (cfu/ml) and the blood HIV load, whereas there was an inverse correlation with the stage of disease and the CD4 cell counts. The PFGE patterns of isolates were variable with regard to the number and positions of bands. The variability of the band sizes in some run positions showed a Gaussian distribution. Generally, the most frequent size variants were associated with the strains with the highest cfu/ml and lowest CD4 counts (< or =200 cells/microl). These findings suggest a possible strain selection over time during disease progression, especially in HIV-positive subjects with low CD4 counts.     

Protection against systemic candidiasis in mice immunized with secreted aspartic proteinase 2

Secreted aspartic proteinases (Sap) have been described as virulence factors implicated in the mechanisms of host colonization by the yeast Candida albicans in different types of candidiasis. Intraperitoneal inoculation of C. albicans into BALB/c mice rapidly leads to systemic candidiasis, with significant colonization of the kidneys measurable in the following week. In this study we assessed the potential of vaccination with C. albicans secreted aspartic proteinase 2 (Sap2) in preventing systemic candidiasis in BALB/c mice. Intradermal injection of highly purified native Sap2 protein incorporated in alum adjuvant provided efficient immune protection, as indicated by a 20-fold decrease in the colonization of kidneys. The protective effect of Sap2 immunization with alum adjuvant was also observed in mice infected with a lethal inoculum of C. albicans. Immunization with the native Sap2 alone, as well as with a denatured recombinant form of the protein, also conferred protection, albeit to a lesser level. In all cases, protection correlated with an increase in serum antibodies to Sap2. Moreover, passive transfer of anti-Sap2 immunoglobulin G (IgG) significantly decreased the yeast burden in kidneys of C. albicans-infected mice. This result shows that immune protection against systemic candidiasis in mice immunized with Sap2 is antibody-mediated. Taken together, these analyses demonstrate that Sap2 can be successfully used as a vaccination target in systemic candidiasis and reveals the potential immunomodulatory role of Sap2 on C. albicans infection.     

Characterization of binding of Candida albicans to small intestinal mucin and its role in adherence to mucosal epithelial cells

In order to approximate and adhere to mucosal epithelial cells, Candida must traverse the overlying mucus layer. Interactions of Candida species with mucin and human buccal epithelial cells (BECs) were thus investigated in vitro. Binding of the Candida species to purified small intestinal mucin showed a close correlation with their hierarchy of virulence. Significant differences (P < 0.05) were found among three categories of Candida species adhering highly (C. dubliniensis, C. tropicalis, and C. albicans), moderately (C. parapsilosis and C. lusitaniae) or weakly (C. krusei and C. glabrata) to mucin. Adherence of C. albicans to BECs was quantitatively inhibited by graded concentrations of mucin. However, inhibition of adherence was reversed by pretreatment of mucin with pronase or C. albicans secretory aspartyl proteinase Sap2p but not with sodium periodate. Saturable concentration- and time-dependent binding of mucin to C. albicans was abrogated by pronase or Sap2p treatment of mucin but was unaffected by beta-mercaptoethanol, sodium periodate, neuraminidase, lectins, or potentially inhibitory sugars. Probing of membrane blots of the mucin with C. albicans revealed binding of the yeast to the 66-kDa cleavage product of the 118-kDa C-terminal glycopeptide of mucin. Although no evidence was found for the participation of C. albicans cell surface mannoproteins in specific receptor-ligand binding to mucin, inhibition of binding by p-nitrophenol (1 mM) and tetramethylurea (0.36 M) revealed that hydrophobic interactions are involved in adherence of C. albicans to mucin. These results suggest that C. albicans may both adhere to and enzymatically degrade mucins by the action of Saps, and that both properties may act to modulate Candida populations in the oral cavity and gastrointestinal tract.     

Serum and saliva antibodies do not inhibit Candida albicans Sap2 proteinase activity using a BSA hydrolysis assay.

The opportunistic fungal pathogen Candida albicans possesses 10 members of a secreted aspartyl proteinase (Sap) family, which are associated with fungal virulence. The C. albicans proteinases are known to induce antibody responses in humans, but the direct inhibition of Sap activity by antibody has not yet been demonstrated. The aim of this study was to determine whether antibodies in saliva or serum could inhibit C. albicans proteinase activity. A two-step Sap-inhibition assay based on bovine serum albumin (BSA) hydrolysis was developed. First, serum and saliva were incubated with Sap2 to allow antibodies to bind to the enzyme, and then a Sap activity assay was performed to determine whether or not the bound antibodies were capable of inhibiting Sap activity. Inhibition of Sap2 activity was investigated using nine sources of sera or saliva: mouse Sap1, Sap2 and Sap3 antisera; rabbit Sap2 antiserum; two pooled human serum samples from HIV-positive and HIV-negative patients with oral C. albicans infection; and three pooled saliva samples from patients with oral C. albicans infection, C. albicans carriers, and Candida-culture-negative individuals. Pooled saliva samples did not inhibit Sap2 activity, whereas mouse, rabbit, and human sera demonstrated inhibition of Sap2 activity by 20-45%. Further analysis of different serum fractions, purified total IgG, and Sap2-specific antibodies demonstrated that non-protein, non-antibody components of serum appeared to be responsible for the partial inhibition of Sap2 activity. Therefore, no evidence was found to demonstrate that specific or non-specific antibodies in serum or saliva could inhibit C. albicans Sap2 activity.     

Reduced virulence of HWP1-deficient mutants of Candida albicans and their interactions with host cells

The Candida albicans gene HWP1 encodes a surface protein that is required for normal hyphal development in vitro. We used mutants lacking one or both alleles of HWP1 to investigate the role of this gene in virulence. Mice infected intravenously with the homozygous hwp1 null mutant, CAL3, survived a median of >14 days, whereas mice infected with a control strain containing two functional alleles of HWP1 survived only 3.5 days. After 1 day of infection, all strains produced similar levels of infection in the kidneys, spleen, and blood. However, after 2 and 3 days, there was a significant decrease in the number of organisms in the kidneys of the mice infected with CAL3. This finding suggests that the hwp1 homozygous null mutant is normal in its ability to initiate infection but deficient in its capacity to maintain infection. CAL3 also germinated minimally in the kidneys. The ability of the heterozygous null mutant to germinate and cause mortality in mice was intermediate to CAL3, suggesting a gene dosage effect. To investigate potential mechanisms for the diminished virulence of CAL3, we examined its interactions with endothelial cells and neutrophils in vitro. CAL3 caused less endothelial cell injury than the heterozygous hwp1 mutant. We conclude that the HWP1 gene product is important for both in vivo hyphal development and pathogenicity of C. albicans. Also, the ability to form filaments may be critical for candidal virulence by enabling the fungus to induce cellular injury and maintain a deep-seated infection.     

Inhibition of growth and secreted aspartyl proteinase production in Candida albicans by lysozyme.

Lysozyme (muramidase) is a non-specific, antimicrobial protein ubiquitous in human mucosal secretions such as saliva. Although its antibacterial and antifungal activities are well recognised, there are no data on the specific concentrations necessary to affect the growth of Candida albicans or about the effect of lysozyme on the production of secreted aspartyl proteinase (Sap), a putative virulence factor of C. albicans. Five Sap-producing isolates of C. albicans were cultured in YCB-BSA medium with various concentrations of lysozyme to examine its effect on yeast cell growth, ultrastructural cellular topography and extracellular and intracellular Sap concentration and activity. Lysozyme was candidacidal at high concentrations and decreased significantly the extracellular Sap concentration at sublethal doses, accompanied by intracellular accumulation of the enzyme. At low concentrations of lysozyme (c. 10 microg/ml), Sap activity decreased more than two-fold and Sap concentration decreased five-fold without any appreciable effect on cell growth or viability. Ultrastructural investigations showed ballooned cells and cells with invaginations (especially present near bud scars), indicating that cell-wall components may be possible targets for this enzyme. All concentrations of lysozyme tested were well within physiologically attainable levels. These data suggest that lysozyme has, at least, a bimodal action on C. albicans, killing the organism at higher concentrations and modulating Sap metabolism at lower concentrations.     

Disruption of each of the secreted aspartyl proteinase genes SAP1, SAP2, and SAP3 of Candida albicans attenuates virulence.

Secreted aspartyl proteinases (Saps), encoded by a gene family with at least nine members (SAP1 to SAP9), are one of the most discussed virulence factors produced by the human pathogen Candida albicans. In order to study the role of each Sap isoenzyme in pathogenicity, we have constructed strains which harbor mutations at selected SAP genes. SAP1, SAP2, and SAP3, which are regulated differentially in vitro, were mutated by targeted gene disruption. The growth rates of all homozygous null mutants were similar to those of the isogenic wild-type parental strain (SC5314) in complex and defined media. In medium with protein as the sole source of nitrogen, sap1 and sap3 mutants grew with reduced growth rates but reached optical densities similar to those measured for SC5314. In contrast, sap2 null mutants tended to clump, grew poorly in this medium, and produced the lowest proteolytic activity. Addition of ammonium ions reversed such growth defects. These results support the view that Sap2 is the dominant isoenzyme. When sap1, sap2, and sap3 mutants were injected intravenously in guinea pigs and mice, the animals had increased survival rates compared to those of control animals infected with SC5314. However, reduction of proteolytic activity in vitro did not correlate directly with the extent of attenuation of virulence observed for all Sap-deficient mutants. These data suggest that SAP1, SAP2, and SAP3 all contribute to the overall virulence of C. albicans and presumably all play important roles during disseminated infections.     

Assessment of Candida albicans genes expressed during infections as a tool to understand pathogenesis.

Candida albicans is the most common fungal opportunistic pathogen of humans and causes mucocutaneous, bloodstream and deep organ infections. Screening for C. albicans genes that are preferentially expressed within infected hosts represents a strategy to identify novel virulence factors and define global expression patterns relevant to pathogenesis. Until recently, C. albicans has not been amenable to screening using existing technologies. This has begun to change with the development of new molecular genetic tools and the sequencing of the C. albicans genome. In this paper, we review studies using recently developed techniques to identify genes expressed by C. albicans during infections, as well as work from our laboratory using a human antibody-based strategy. Along with others, we have shown that selected in vivo expressed genes encode known and previously unrecognized candidal virulence factors. Future studies in this area will identify additional novel virulence factors, as well as advance our understanding of pathogenesis.     

Evidence implicating phospholipase as a virulence factor of Candida albicans.

Three different approaches were used to investigate the role of extracellular phospholipases in the pathogenicity of Candida albicans. First, we compared 11 blood isolates of this yeast with an equal number of commensal strains isolated from the oral cavities of healthy volunteers. Blood isolates produced significantly more extracellular phospholipase activity than the commensal strains did. Second, two clinical isolates of C. albicans that differed in their levels of virulence in a newborn mouse model were compared for their ability to secrete phospholipases. The invasive strain produced significantly more extracellular phospholipase activity than the noninvasive strain did. Third, nine blood isolates were characterized for their phospholipase and proteinase production, germ tube formation, growth, and adherence to and damage of endothelial cells in vitro. These factors were analyzed subsequently to determine whether they predicted mortality in a mouse model of hematogenously disseminated candidiasis. By proportional hazard analysis, the relative risk of death was 5.6-fold higher (95% confidence interval, 1.672 to 18.84 [P < 0.005]) in the mice infected with the higher-phospholipase-secreting strains than in the low-phospholipase secretors. None of the other putative virulence factors predicted mortality. Characterization of phospholipases secreted by three of the blood isolates showed that these strains secreted both phospholipase B and lysophospholipase-transacylase activities. These results implicate extracellular phospholipase as a virulence factor in the pathogenesis of hematogenous infections caused by C. albicans.