Linkage of Adhesion, Filamentous Growth, and Virulence in Candida albicans to a Single Gene, INT1

See allHide authors and affiliations

Science  27 Feb 1998:
Vol. 279, Issue 5355, pp. 1355-1358
DOI: 10.1126/science.279.5355.1355


Adhesion and the ability to form filaments are thought to contribute to the pathogenicity of Candida albicans, the leading cause of fungal disease in immunocompromised patients. Int1p is a C. albicans surface protein with limited similarity to vertebrate integrins. INT1 expression in Saccharomyces cerevisiae was sufficient to direct the adhesion of this normally nonadherent yeast to human epithelial cells. Furthermore, disruption ofINT1 in C. albicans suppressed hyphal growth, adhesion to epithelial cells, and virulence in mice. Thus,INT1 links adhesion, filamentous growth, and pathogenicity in C. albicans and Int1p may be an attractive target for the development of antifungal therapies.

Candida albicansis the leading cause of invasive fungal disease in premature infants, diabetics, surgical patients, and hosts with human immunodeficiency virus infection or other immunosuppressed conditions. Despite appropriate therapy, mortality resulting from systemic C. albicans infection in immunocompromised patients approaches 30% (1). The pathogenesis of C. albicans infection is postulated to involve adhesion to host epithelial and endothelial cells and morphologic switching of yeast cells from the ellipsoid blastospore to various filamentous forms: germ tubes, pseudohyphae, and hyphae (2).

The C. albicans gene INT1 was originally cloned because of its similarity to vertebrate leukocyte integrins (3), adhesins that bind extracellular matrix proteins and induce morphologic changes in response to extracellular signals (4). INT1 expression in the budding yeastS. cerevisiae triggers a morphologic switch to filamentous growth (3). In C. albicans, multiple adhesins mediate attachment to epithelium, endothelium, or platelets (5-8). Because laboratory strains of S. cerevisiae have few adhesins (7), we investigated whether Int1p is present on the cell surface and can function as an adhesin when it is expressed in S. cerevisiae.

When intact S. cerevisiae cells expressingINT1 were treated with an impermeant biotinylation reagent, Int1p became biotin-labeled, indicating that at least one portion of Int1p was on the exterior cell surface (Fig.1A). Nonsurface proteins, such as Rap1p, an abundant nuclear protein, were not biotinylated (Fig. 1A).Saccharomyces cerevisiae cells expressing INT1(strain YCG101) adhered to monolayers of human cervical epithelial cells (HeLa), whereas S. cerevisiae cells carrying vector sequences (YCG102) and YCG101 cells grown in glucose [to repress Int1p expression from the GAL10 promoter (9)] did not adhere to HeLa monolayers (Fig. 1B). Furthermore, adhesion of YCG101 cells to HeLa monolayers was specific for Int1p epitopes: UMN13, a polyclonal antibody recognizing Int1p amino acids 1143 to 1157 [a region predicted to be extracellular (3)], inhibited adhesion, whereas nonimmune rabbit immunoglobulin G (IgG) did not (Fig.1B). Thus, the expression of Int1p alone was sufficient to confer adhesive capacity on S. cerevisiae.

Figure 1

Int1p is a surface protein that mediates adhesion to human epithelial cells. (A) Protein blots of S. cerevisiae proteins immunoprecipitated (IP) from cell lysates after labeling of cell surface proteins with biotin (25). Left panel, biotinylated proteins detected with HRP-avidin; right panel, proteins detected with the antibodies indicated. Lane 1 of each panel, S. cerevisiae expressing vector sequences only; lanes 2 and 3,S. cerevisiae expressing INT1. Numbers at the left are molecular size markers (in kilodaltons). (B) Expression of INT1 enables S. cerevisiae to adhere to human epithelial cells (26). Percent specific adhesion was determined for YCG101, S. cerevisiae YPH500 (MATα ura3 lys2 ade2 trp1his3 leu2) expressing INT1 under control of the GAL10promoter on the plasmid pCG01 (3); YCG102, YPH500 transformed with the GAL10 vector pBM272 (24); and JKY81-5-1, S. cerevisiae cdc12 -6 ts strain (10). YCG101 and YCG102 cells were grown to midexponential phase in minimal medium with 2% raffinose and then diluted to an optical density of 0.1 at 600 nm in minimal medium containing 8.7 mM methionine plus 2% galactose for the induction of INT1. JKY81-5-1 cells were grown in galactose at 25°C and then shifted to 30°C. (C) Morphology of each of the strains tested for adhesive ability. Cells were grown as in (A) and images were obtained with a Leitz Diaplan microscope with differential interference contrast optics, an MTI CCD72 camera, and Scion Image software.

INT1 expression induces the growth of highly polarized buds (3). To test the possibility that the increased surface area of polarized S. cerevisiae cells (Fig. 1C) influences cell adhesion, we performed adhesion assays with acdc12-6ts strain (JKY81-5-1) (10) that forms multiple elongated buds at the permissive temperature (11) (Fig. 1C). Adhesion of thecdc12-6ts strain to HeLa cell monolayers did not differ from that of YCG102 and was significantly less than the adhesion seen upon expression of INT1 (Fig.1B), indicating that filamentous morphology alone is not sufficient to explain the increased adhesion of INT1-expressing cells.

We next tested the hypothesis that INT1 is involved in adhesion and filamentous growth in C. albicans as well. Both copies of INT1 were disrupted sequentially in C. albicans strain CAI4 (12) by means of ahisG-CaURA3-hisG cassette (13) yielding a Ura+ int1/INT1heterozygote (CAG1) and a Ura+ int1/int1homozygote (CAG3). INT1-CaURA3 was reintegrated into the genome of a Ura int1/int1 homozygote (CAG4) to yield the int1/int1 + INT1 heterozygous reintegrant (CAG5) (14), which served as anint1/int1 + INT1 Ura+ control to ensure that CAG3 phenotypes could be attributed to disruption ofINT1.

The specific adhesion of the C. albicans int1/int1 strain (CAG3) to HeLa cells was reduced by 39% relative to that of theINT1/INT1 strain (CAF2) (12) (Fig.2A). Preincubation of the C. albicans INT1/INT1 strain (CAF2) with UMN13 antibodies reduced epithelial adhesion by 40% but did not eliminate it (Fig. 2A). In contrast, preincubation of S. cerevisiae expressingINT1 with UMN13 antibodies blocked virtually all adhesion to HeLa cells (Fig. 1B). These results suggest that although Int1p was the major adhesin expressed in S. cerevisiae, other attachment factors in addition to Int1p account for epithelial adhesion inC. albicans. The presence of a single copy ofINT1 (CAG1 and CAG5) did not restore wild-type adhesion; however, the single copy of INT1 in CAG1 and CAG5 was expressed because UMN13 significantly reduced adhesion in these strains by 39% and 28%, respectively (Fig. 2A). As expected, UMN13 did not significantly reduce adhesion in the int1/int1 strain (CAG3). These results indicate that Int1p is one of a number of adhesins that enable C. albicans to attach to epithelial cells, and that the remaining candidal adhesins in theint1/int1 strain (CAG3) bind HeLa cells by means of an epitope or epitopes not recognized by UMN13. Results with the heterozygote strains (CAG1 and CAG5) imply that the gene dosage ofINT1 is important for the full expression of the adhesive phenotype.

Figure 2

Disruption of INT1 inC. albicans reduces adhesion to human epithelial cells and filamentous growth. (A) Adhesion analysis (26) of INT1/INT1 (CAF2) (12), int1/INT1 (CAG1) (13),int1/int1 (CAG3) (13), andint1/int1 + INT1 (CAG5) (14) strains. (B) Hyphal growth of C. albicansstrains on milk-Tween agar and Spider medium (27). Left panels, INT1/INT1 (CAF2); center panels, int1/int1 (CAG3); right panels,int1/int1 + INT1 (CAG5). Colonies were photographed with a Zeiss Stemi DRC dissecting microscope and a Nikon 35- mm camera.

The effect of int1 mutations on the filamentous growth of C. albicans strains was monitored with isogenic Ura+ prototrophs on two different media that induce filamentation (Fig. 2B). The INT1/INT1 strain (CAF2) formed an extensive network of long, branching hyphae that overlay and penetrated milk-Tween agar. On Spider medium, the INT1/INT1strain (CAF2) formed wrinkled colonies, an indicator of filamentous growth (15). In contrast, the int1/int1 strain (CAG3) formed smooth-edged colonies with very few filamentous cells emanating from the colony edge on milk-Tween agar and primarily smooth colonies on Spider medium. Reintegration of INT1 (CAG5) restored the ability of the colonies to produce large numbers of hyphae on milk-Tween agar and to form wrinkled colonies on Spider medium, indicating that INT1 contributes to the filamentous growth of C. albicans.

On both media, filamentous growth of the reintegrant strain (CAG5) (Fig. 2B) and of the int1/INT1 strain (CAG1) (9) was similar, but not identical, to the growth of INT1/INT1colonies (CAF2). Similarly, strains heterozygous for the C. albicans homologs of STE7 (hst7/HST7),STE12 (cph1/CPH1), and STE20(cst20/CST20) exhibit intermediate defects in hyphal formation (16), suggesting that filamentous growth inC. albicans is sensitive to gene dosage.

Despite the altered morphology of the int1/int1 strain (CAG3) on milk-Tween and Spider media, this strain grew with a phenotype indistinguishable from that of the INT1/INT1strain (CAF2) in other liquid and solid media that induce hyphae (17), which shows that Int1p is not necessary for the growth of hyphae in C. albicans. Rather, the results suggest that Int1p may be a sensor that triggers the morphogenic decision process in response to a subset of environmental conditions. We propose that morphogenesis in C. albicans can occur through multiple pathways that include some discrete and some overlapping components. Consistent with this view, mutation of other C. albicansgenes that are involved in morphogenesis pathways suppress hyphal formation under some, but not all, growth conditions: Mutation of both alleles of the C. albicans mitogen-activated protein kinase components HST7, CPH1, and CST20suppresses hyphal growth on solid Spider medium but has no effect on hyphal growth in serum (16, 18). In addition,C. albicans phr1/phr1 strains, which lack a putative surface glycoprotein, affect the morphology of filamentous cells at high but not low pH (19).

We tested the virulence of the C. albicans int1 mutant strains in a mouse model of intravenous infection because both adhesion and hyphal growth are hypothesized to be important for the pathogenicity of C. albicans. Again, isogenicURA3 strains were used because ura3 strains have reduced virulence (18, 20). All of the mice injected with the INT1/INT1 strain (CAF2) died by day 11 (Fig.3). In contrast, the int1/int1homozygote (CAG3) was much less virulent; 90% of the mice were alive at the end of the experiment (Fig. 3). The virulence of theint1/int1 strain (CAG3) also was less than that of both heterozygous strains (CAG1 and CAG5) (Fig. 3). These results indicate that Int1p, a protein that functions in both adhesion and filamentous growth, is essential for the virulence of C. albicans in this murine model of intravenous infection.

Figure 3

Disruption of INT1 in C. albicans causes reduced virulence in a mouse model of systemic candidiasis (28) (n = 10 mice for each yeast strain). Curves are the compiled results of two replicate experiments (n = 5 mice for each yeast strain for each experiment). Although the first experiment was terminated at day 20, mice in the second experiment were followed until day 30; no additional deaths occurred between day 20 and day 30 in the mice injected with CAG3. The doubling times of all strains, grown in SD minus uracil at 30°C, were 72 ± 6 min, except in the first experiment, where CAG5 had a doubling time of 90 min.

A single copy of INT1 in the heterozygous C. albicans strains CAG1 and CAG5 restored the filamentation phenotype to nearly that of the wild type and restored virulence to an intermediate extent, but did not restore specific adhesion to the same degree. These results could be attributed to a number of possibilities: differences in the threshold amount of Int1p required for the given phenotype, different cell types assayed for each phenotype (yeast forms for adhesion versus filamentous forms for morphogenesis), or differences between the growth or assay conditions used to test each phenotype. The latter two situations may affect the expression of Int1p.

Candida albicans is a highly successful pathogen of immunocompromised hosts, most likely because it has several adhesins and multiple pathways for triggering the morphologic switch to filamentous growth (16, 18, 19, 21). Similar to the results with deletion of INT1, deletion of HST7(18) and combined deletion of CPH1 withEFG1 (21) suppress hyphal formation and virulence in C. albicans. Because Int1p is the first C. albicans protein shown to be involved in both adhesion and filamentation, it will be interesting to determine how INT1regulates, or is regulated by, other pathways of adhesion and morphologic switching in C. albicans. Unlike C. albicans homologs of S. cerevisiae genes involved in pseudohyphal growth (16, 18, 21), INT1 appears to be unique to C. albicans (3). Moreover,INT1 induces filaments in strains lacking STE12(3) and other genes required for pseudohyphal growth (22). The S. cerevisiae protein Bud4p has sequence similarity to Int1p in the COOH-terminal 300 amino acids; however, Bud4p is not required for Int1p-induced filamentous growth and high-copy BUD4 expression does not induce filamentous growth (22), indicating that Int1p and Bud4p are not functional homologs. Although Int1p has only limited similarity to vertebrate integrins (3), it clearly fits the integrin paradigm in its surface localization, its mediation of adhesion, and its ability to influence morphogenesis. Thus, Int1p is a candidal virulence factor that, unlike other reported virulence factors, is involved in both adhesion and morphogenesis. The surface location of Int1p makes it an attractive target for the design of preventive strategies and therapeutic agents.

  • * To whom correspondence should be addressed. E-mail: judith{at} and hoste001{at}


View Abstract

Navigate This Article