Gene overexpression/suppression analysis of candidate virulence factors of Candida albicans

Abstract

We developed a conditional overexpression/suppression genetic strategy in Candida albicans to enable simultaneous testing of gain or loss of function in order to identify new virulence factors. The strategy involved insertion of a strong, tetracycline-regulated promoter in front of the gene of interest. To validate the strategy, a library of genes encoding glycosylphosphatidylinositol (GPI)-anchored surface proteins was screened for virulence phenotypes in vitro. During the screening, overexpression of IFF4 was found to increase the adherence of C. albicans to plastic and to human epithelial cells, but not endothelial cells. Consistent with the in vitro results, IFF4 overexpression modestly increased the tissue fungal burden during murine vaginal candidiasis. In addition to the in vitro screening tests, IFF4 overexpression was found to increase C. albicans susceptibility to neutrophil-mediated killing. Furthermore, IFF4 overexpression decreased the severity of hematogenously disseminated candidiasis in normal mice, but not in neutropenic mice, again consistent with the in vitro phenotype. Overexpression of 12 other GPI proteins did not affect normal GPI protein cell surface accumulation, demonstrating that the overexpression strategy did not affect the cell capacity for making such proteins. These data indicate that the same gene can increase or decrease candidal virulence in distinct models of infection, emphasizing the importance of studying virulence genes in different anatomical contexts. Finally, these data validate the use of a conditional overexpression/suppression genetic strategy to identify candidal virulence factors.

FIG. 1.

Diagram detailing our split-marker strategy to insert the TR promoter in front of the gene encoding putative GPI-anchored proteins.

FIG. 2.

Overexpression of GPI-anchored proteins does not affect the expression of Als1p, a known GPI-anchored protein. Twelve overexpression strains and one control strain were screened by flow cytometry. A representative strain is shown; all other strains demonstrated equivalent characteristics. The strains were grown under conditions in which Als1p was either known not to be expressed or known to be expressed (13). Overexpression of the GPI protein had no effect on Als1p surface accumulation (no difference between the results without DOX [−DOX] and with DOX [+DOX]).

FIG. 3.

Overexpression of IFF4 induces adherence of C. albicans to culture plastic tubes. The control strain used was C. albicans THE2.

FIG. 4.

Strategy to generate a homozygous mutant in which both alleles of the desired gene were controlled by the TR promoter.

FIG. 5.

Overexpression of IFF4 results in enhanced adherence of C. albicans to FaDu epithelial cells. (A) RT-PCR results for IFF4 demonstrating overexpression of the gene without DOX medium and lack of expression with DOX medium. The P₂ and P₄ primers (Table 2) were used to amplify IFF4. The EFB1 fragment was coamplified and served as a control. Lack of genomic-DNA (gDNA) contamination in cDNA preparations was demonstrated by the absence of a 919-bp band containing the intron of EFB1. THE31 was the control strain, and CAA10-31 was the strain overexpressing IFF4. (B) Adherence of strains THE31 and CAA10-31 grown without DOX (overexpression condition) or with DOX (suppression condition) to FaDu epithelial cells. The data are displayed as the median ± the interquartile. *, P = 0.01 compared to control plus DOX; **, P < 0.002 versus all others.

FIG. 6.

IFF4 overexpression increases the tissue fungal burden in the mouse vagina. (A) Mice (n = 8 per group) infected with C. albicans overexpressing IFF4 had an increased vaginal fungal burden 24 h postinfection compared to the same strain gown with DOX or to a control strain. The data are displayed as the median ± the interquartile. *, P < 0.037 compared to IFF4 plus DOX or the control strain cultured without DOX or with DOX. (B) RT-PCR results demonstrating expression of IFF4 in the mouse vagina infected with wild-type C. albicans. RT-PCR of RNA samples extracted from a mouse vagina infected with C. albicans overexpressing IFF4 was included as a positive control, whereas RNA extracted from uninfected mice was included as a negative control. The EFB1 fragment was coamplified and served as a loading control. Furthermore, lack of genomic DNA (g-DNA) contamination in cDNA preparations was demonstrated by the absence of a 919-bp band containing the intron of EFB1.

FIG. 7.

IFF4 overexpression increases neutrophil killing of C. albicans and reduces the severity of hematogenously disseminated candidiasis. (A) In vitro neutrophil-mediated killing of C. albicans strain THE31 or CAA10-31 grown under overexpression or suppression conditions. The data are displayed as the median ± the interquartile. ‡, P < 0.001 compared to IFF4 plus DOX. (B) Survival of mice (n = 8 per group) infected with a control strain or C. albicans IFF4 grown under overexpression or suppression conditions. *, P < 0.002 compared to mice infected with C. albicans IFF4 with DOX or the control strain. The experiment is representative of two studies with similar findings. (C) Burden of C. albicans in kidneys of immunocompetent mice (n = 8 per group) infected with C. albicans IFF4 or a control strain grown under overexpressing (−DOX) or suppressing (+DOX) conditions. Kidneys were harvested 5 days postinfection. The data are displayed as the median ± the interquartile. *, P < 0.025 versus no expression of IFF4 or the control strain. (D) In the absence of neutrophils, the virulence of IFF4-overexpressing C. albicans was restored. Shown is the survival of neutropenic mice (n = 8 per group) infected with a control strain or C. albicans IFF4 grown under overexpression or suppression conditions.

FIG. 8.

RT-PCR of RNA samples extracted from mouse kidneys infected with wild-type C. albicans for 24 h. RNA extracted from uninfected mice was included as a negative control, whereas RNA extracted from wild-type C. albicans SC5314 in vitro was included as a positive control. The EFB1 fragment was coamplified and served as a loading control. Lack of genomic-DNA contamination in cDNA preparations was demonstrated by the absence of a 919-bp band containing the intron of EFB1.