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.

Candida albicans causes hematogenously disseminated and mucosal infections (7, 33). Despite current antifungal therapy, mortality and morbidity are unacceptably high (1, 16, 34). Therefore, new prophylactic and therapeutic strategies are urgently needed. The identification of virulence genes that can be targeted to alter pathogenicity is a necessary step in devising novel strategies to treat or prevent Candida infections.

Virulence factors in C. albicans include proteins that mediate adherence to and invasion of host tissues (43), morphological change from yeast to hyphae (29, 30), secretion of lytic enzymes (17, 27, 41), maintenance of cell wall integrity (55), and avoidance of the host immune response (39). Many of these virulence factors are glycosylphosphatidylinositol (GPI)-anchored proteins, which comprise 88% of all covalently linked cell wall proteins in C. albicans (23). Examples of GPI-anchored virulence factors are Phr1p and Phr2p, which mediate cell wall biogenesis and hyphal formation in response to changes in pH (15); Hwp1p, an epithelial adhesin and biofilm promoter (32) that is required for virulence during murine hematogenously disseminated candidiaisis (48); Als1p and Als3p, adhesins with broad substrate specificity (13, 22, 43); and Gpi7, an antivirulence factor that reduces candidal resistance to macrophages and virulence in mice (36). Numerous GPI proteins have been identified as virulence factors in C. albicans.

To identify new virulence factors in C. albicans, we used a conditional gene overexpression/suppression approach to screen a library of genes encoding GPI-anchored proteins for virulence phenotypes. We used the strong tetracycline-regulated (TR) promoter to control gene expression (31, 42). In the presence of the tetracycline analogue doxycycline (DOX), the expression of genes controlled by the strong TR promoter is suppressed. In contrast, in the absence of DOX, the gene is potentially overexpressed. By simultaneously screening for gain and loss of function, we identified a gene, IFF4, that had contrasting functions in distinct anatomical contexts, promoting epithelial cell adherence but also ameliorating virulence during disseminated infection in mice.

(This work was presented in part at the 47th Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, 17 to 20 September 2007, abstr. B-1445.)

MATERIALS AND METHODS

Strains, media, and plasmids.

The C. albicans strains used in this study are listed in Table 1. All strains were generated from strain THE1, a generous gift from Hironobu Nakayama (31). The HIS1-TR promoter cassette plasmid was constructed by replacing the URA3 gene in p99CAU1 (kindly provided by H. Nakayama) with a HIS1 fragment (encompassing 543 bp upstream and 355 bp downstream of its open reading frame [ORF]) amplified from THE1 genomic DNA, using 5′-His1 and 3′-His1 as primers (Table 2). All Candida strains were cultured in YPD medium (1% yeast extract [Difco], 2% Bacto peptone [Difco], 2% d-glucose) with or without 20 mg/liter DOX (Sigma-Aldrich, St. Louis, MO) at 30°C. The growth rates of strains whose virulence levels in animal models were compared were determined in YPD in the presence or absence of 20 mg/liter DOX (38). For strain construction, yeasts were grown in GMM (2% glucose, 0.5% ammonium sulfate, and 1× yeast nitrogen base without ammonium sulfate and amino acid) with or without nutrient supplementation. Fluoro-orotic acid (5-FOA) (Sigma-Aldrich; 0.1%) was used for selecting ura3 strains. Media were solidified with 1.5% Difco agar as necessary.

TABLE 1.

C. albicans strains in this study

Strain	Genotype	Source
THE1	ade2::hisG/ade2::hisG ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2	H. Nakayama
THE2	ade2::hisG/ade2::hisG HIS1/his1::URA3-dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2	This study
THE3	ade2::hisG/ade2::hisG HIS1/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2	This study
THE31	ade2::hisG/ade2::hisG HIS1/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434::URA3-IRO1 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2	This study
THE4	ade2::hisG/ade2::hisG his1::URA3-dpl200/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2	This study
CAA10	ade2::hisG/ade2::hisG his1::URA3-dpl200/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2 IFF4/HIS1-pTR-IFF4	This study
CAA10-1	ade2::hisG/ade2::hisG his1::dpl200/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2 IFF4/HIS1-pTR-IFF4	This study
CAA10-2	ade2::hisG/ade2::hisG his1::dpl200/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2 iff4::URA3-dpl200/HIS1-pTR-IFF4	This study
CAA10-3	ade2::hisG/ade2::hisG his1::dpl200/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2 HIS1-pTR-IFF4/HIS1-pTR-IFF4	This study
CAA10-31	ade2::hisG/ade2::hisG his1::dpl200/his1::dpl200 ura3-iro1::imm434/ura3-iro1::imm434::URA3-IRO1 ENO1/ENO1-tetR-ScHAP4AD-3XHA-ADE2 HIS1-pTR-IFF4/HIS1-pTR-IFF4	This study

TABLE 2.

Oligonucleotides used in this study

Oligonucleotide	Sequence
Oligonucleotides used for a wild-type HIS1 gene cloning and HIS1-pTR cassette
5′-His1	CGGGATCCAGAATGTGCCGTTGTGTTTG
3′-His1	CGGGATCCGTACCAGGTGAACTGTTTAATTG
Oligonucleotides used for generating his1 null mutant required for GPI gene overexpression library construction
HISKN-1A	AAGGGCAGATTATACGAAAAATGCTGTAACTTATTGAGTGGTGCCGATATACAGTTTAGAGTTTTCCCAGTCACGACGTT
HISKN-1B	AGATCTAATAGATTAGATATAGCACTTTCTACAAACTTGCCAATTGCATTAATCTTCTTGGTTTTCCCAGTCACGACGTT
HISKN-2A	TTAAAGGTGTGTACATCAAGGTGGTAGATAAAGATGGTATAAGACAGATTGAGTCAAATTTGTGGAATTGTGAGCGGATA
HISKN-2B	TCTACAATTTGATATCTCGAGTACCAATATATCGGTTGCACCAGCTTTCTTCAATTCGTCTGTGGAATTGTGAGCGGATA
HisConfirm1	CCAGACCGTTTGTTATTTGCTG
HisConfirm2	GTCAAGGAATTACAAACGAGAATGC
5-detect	GTTTTCCCAGTCACGACGTTGTAAAACGAC
3-detect	TGTGGAATTGTGAGCGGATAACAATTTCAC
Oligonucleotides used for making and confirming IFF4 conditional overexpression/suppression strain
P1	ACCACCATATTGCGCTTTTCTA
P2	ACAGCTTTATCTCAGAAAAACTAGTTCCGTTTTCTTTCACGTCAAC
P3	CGACAAACACAACGGCACATTCTGGCTTTAACGATTTGCAAAATTTAG
P4	ACCACTGTTGGTAATAGATCC
PH1	GTCGTCGCTGTGTTTGTC
PH2	CGTTGGAGAAGGTAATTGTGA
P5	AGTGGAATCACCATTAGCAGA
P6	CTAAATTTTGCAAATCGTTAAAGCC
IFF4KN1	GCACCAATCTCATCAAGTGAGTTTATTAGTTTGGAATCGTGTTTTCCCAGTCACGACGTT
IFF4KN2	GGATGTTTCAAACGGTAATTCGCTCCAACAAGAGGGTTCATGTGGAATTGTGAGCGGATA
IFF4CONF1	CACTTTAGCTAAAGTATCATCTACTG
IFF4CONF2	TTCGGAATAGGATGGTTTGAC
Oligonucleotides used for detecting IFF4 in vivo expression
IFF4 specific 1	CTGCACCAATCTCATCAAGTGA
IFF4 specific 2	GATCCACCAGCACTTGAGGA
EFB1a	ATTGAACGAATTCTTGGCTGAC
EFB1b	CATCTTCTTCAACAGCAGCTTG
Oligonucleotides used for confirming URA3-IRO1 in its original locus
URA3 Conf1	TGCTGGTTGGAATGCTTATTTG
URA3 Conf2	TGCAAATTCTGCTACTGGAGTT

Construction of a GPI gene overexpression library.

Oligonucleotides used in construction of the overexpressing/suppressing IFF4 and control strains are shown in Table 2. The library was constructed in THE1, a C. albicans strain that harbors a codon-optimized TR transactivator. Using a URA3 recyclable marker strategy (54), one allele of HIS1 in THE1 was disrupted to yield strain THE2. THE2 was used as a control strain for initial GPI gene overexpression library screening. The URA3 marker was eliminated by culturing THE2 on 5-FOA plates to generate THE3. Next, the second HIS1 allele was disrupted in THE3, yielding THE4.

C. albicans THE4 was utilized to generate a GPI gene overexpression library using the strategy outlined in Fig. 1. Briefly, two rounds of PCR were used to generate fragments C and D. In the first round of PCR, the flanking regions on both sides of the desired integration site were amplified using primers P₁ and P₃ to yield fragment A (an upstream sequence of the targeted gene), and primers P₂ and P₄ were used to yield fragment B (approximately 500 bp from the start of the targeted gene). Primers P₃ and P₂ contained a 5′ extension (25 bp) that was complementary to the 5′ HIS1 or 3′ TR promoter sequence, respectively (Fig. 1). The second round of PCR utilized primers P₁ and P_H1 to obtain fragment C, which contained an approximately 500-bp sequence upstream of the ORF of the targeted gene, as well as a 3′-truncated HIS1. Similarly, fragment D containing a 5′-truncated HIS1, an intact TR promoter, and approximately 500 bp from the start of the ORF of the targeted gene was generated by using primers P_H2 and P₄. Therefore, fragments C and D contained flanking sequences to target integration and overlapping, nonfunctional fragments of the HIS1 selection marker.

FIG. 1.

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

Fragments C and D were cotransformed into strain THE4, and His1⁺ prototrophs were selected. A triple-crossover event among fragments C and D and the chromosome resulted in integration of the HIS1-TR promoter cassette in front of the targeted GPI gene. The desired integration event was confirmed by PCR using a downstream, ORF-specific primer, P₅, and a universal upstream primer, P_H2. Finally, the expression of the gene without DOX or with DOX was confirmed by reverse transcription RT-PCR using P₂ and P₄ as primers.

Construction of strain CAA10-31, in which both IFF4 alleles are controlled by the TR promoter.

The C. albicans strain in which one allele of IFF4 was controlled by the TR promoter (strain CAA10) was cultured on plates containing 5-FOA, yielding strain CAA10-1. The resulting strain, lacking the URA3 marker, was utilized to construct strain CAA10-31, in which both alleles of IFF4 were controlled by the TR promoter, as outlined in Fig. 2. Briefly, a recyclable URA3 cassette (54) was used to disrupt the second URA3 allele. Clones were screened by PCR using primers P₆ and P₅ to confirm disruption of the second allele (i.e., the one not controlled by the TR promoter). After CAA10-2 was cultured on 5-FOA plates, clones in which interchromosomal recombination occurred, resulting in both alleles being controlled by the TR promoter, were selected, yielding strain CAA10-3. URA3 (from a 3.9-kb NheI-PstI fragment containing the URA3-IRO1 gene) was inserted into its original locus on the CAA10-3 chromosome, resulting in strain CAA10-31. The same URA3-IRO1 fragment was integrated into THE3, resulting in THE31, a control strain of CAA10-31.

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]).

Confirmation of IFF4 overexpression/suppression.

To confirm that IFF4 was regulated by the TR promoter in strains CAA10 (with one IFF4 allele controlled by the TR promoter) and CAA10-31 (with both IFF4 alleles controlled by the TR promoter), we performed semiquantitative RT-PCR. After an overnight culture, the strains were grown with DOX (20 μg/ml) and without DOX for 6 h in YPD. Total RNA was extracted by the hot-acidic-phenol method (3). RNA was also extracted from strain THE31 grown with DOX and without DOX as a control. To eliminate genomic-DNA contamination, total RNA was treated with RNase-free DNase I (TURBO DNA-free; Ambion, Texas) according to the manufacturer's instructions. Lack of genomic-DNA contamination in preparations was demonstrated by the absence of a 919-bp band containing the intron of EFB1 (50). DNA-free total-RNA samples were cleaned using the RNA Clean-up Kit (Zymo Research, Orange, CA). Next, an RT reaction procedure was performed with RETROscript (Ambion, Texas). Primers P₂ and P₄ were used for gene-specific amplification. EFB1 also served as an internal quantitative control, and its specific primers are shown in Table 2.

Effects of overexpression of 12 GPI-anchored proteins on Als1p, another GPI-anchored protein.

Twelve strains overexpressing RBT1, PGA8, SAP9, PLB3, PLB5, SOD6, HYR1, IFF2, IFF3, IFF4, FGR23, or PGA55 were constructed in the library and were grown in YPD in the presence or absence of DOX. Overexpression and suppression of these genes encoding GPI-anchored proteins were confirmed by RT-PCR. We then used direct immunofluorescence, as we have described previously (13), to quantify the degree of surface expression of Als1p in each of the 12 strains. In brief, the strains were incubated in RPMI 1640 medium with glutamine for 1 h at room temperature to induce Als1p expression. Als1p was detected and quantified by incubating intact organisms with a fluorescein isothiocyanate-labeled anti-Als1p monoclonal antibody for 1 h. As a negative control, the strains were labeled with a fluorescein isothiocyanate-labeled nonspecific mouse isotype-matched control antibody. Flow cytometry was then performed to determine the relative expression of Als1p in each strain grown in the presence or absence of DOX using a FACSCaliber (Becton Dickinson) flow cytometer. The mean fluorescence intensities of 10⁴ events were calculated using CELLQUEST software.

Adherence assays.

Adherence to plastic was tested by growing strains in YPD medium with or without DOX to early stationary phase in culture tubes (Falcon). The fungal cultures were discarded, and the tubes were gently washed three times with tap water. Adherence to plastic was assessed visually. No difference in the morphologies of overexpressing, suppressing, or wild-type strains was identified.

Adherence of IFF4 overexpression/suppression mutants to the FaDu oral epithelial cell line (ATCC) and endothelial cells was also assessed. FaDu cells were maintained in minimal essential medium-Earl's salts (Irvine Scientific) containing 10% fetal bovine serum, 1 mM pyruvic acid, 2 mM l-glutamine, 0.1 mM nonessential amino acids, 100 IU/ml penicillin, and 100 IU/ml streptomycin. The cells were grown to confluence in six-well tissue culture plates (Costar, Van Nuys, CA) prior to performance of the adherence assay. Endothelial cells were obtained from human umbilical veins as we have described previously (13, 21), and second-passage cells were grown to confluence in six-well tissue culture plates coated with 0.2% gelatin matrix (Collaborative Biomedical Products, Bedford, MA). After the cell monolayers were rinsed twice with prewarmed Hanks balanced salt solution (HBSS), blastospores (3 × 10² cells/ml of HBSS) grown in either YPD with DOX or YPD without DOX for 19 h at 25°C were added to each well. The plate was incubated in 5% CO₂ at 37°C for 30 min, after which the nonadherent organisms were aspirated and the FaDu or endothelial cell monolayers were rinsed twice with 10 ml of HBSS in a standardized manner. A 1.5-ml volume of YPD was added to each well and allowed to solidify. After the plate was incubated at 37°C for 24 h, the number of adherent organisms was determined by colony counting. Adherence was expressed as a percentage of the initial inoculum, which was confirmed by quantitative culturing in YPD agar. Each adherence assay was performed in triplicate on three separate occasions.

Endothelial cell damage assay.

The ability of IFF4 overexpression/suppression to modulate C. albicans-induced endothelial cell injury was determined with the chromium (⁵¹Cr) release assay in 96-well tissue culture plates as described previously (52).

Neutrophil-mediated killing assay.

In vitro human neutrophil-mediated killing was quantified by a modification of our previously described method (46). Briefly, 4 × 10⁴ purified neutrophils were cocultured with 2 × 10⁴ C. albicans cells (2:1 ratio) for 1 h at 37°C in RPMI plus 10% pooled human serum (Sigma-Aldrich). The cultures were sonicated to kill residual neutrophils, serially diluted, overlaid with YPD, and incubated overnight at 37°C. CFU were counted and compared to the number of CFU plated from C. albicans cultures without neutrophils to assess killing of C. albicans.

Murine models.

The effect of IFF4 overexpression/suppression on the virulence of C. albicans was assessed in the murine models of vaginal candidiasis (47) and hematogenously disseminated candidiasis. For the vaginal model, BALB/c mice (Charles River) were treated with estradiol valerate (30 μg subcutaneously) in peanut oil (both from Sigma-Aldrich) on day −3 relative to infection to induce pseudoestrus. On the day of infection, mice were sedated by intraperitoneal administration of 100 mg/kg of body weight of ketamine. The sedated mice were infected intravaginally with 10⁶ blastospores of C. albicans in 10 μl of endotoxin-free phosphate-buffered saline. On day 1 postinfection, the vaginas and approximately 1 centimeter of each uterine horn were dissected en bloc, homogenized, and quantitatively cultured.

For survival studies, male BALB/C mice were infected with 5 × 10⁵ blastospores of C. albicans through the tail vein (21). The mice were monitored three times daily, and moribund mice were euthanized. A second group of mice were infected and sacrificed 5 days postinfection to determine the tissue fungal burden. Kidneys, brain, and liver were removed, homogenized, and quantitatively cultured on YPD containing 50 μg/ml chloramphenicol. Values were expressed as log CFU per gram of tissue.

The mice were given food and water ad libitum throughout the course of the experiment. Mice infected with IFF4 conditional overexpression/suppression or control strains were given water with or without DOX (2 mg/ml) dissolved in 5% sucrose solution throughout the period of the experiment starting from day −2 relative to infection (42). All procedures involving mice were approved by the institutional animal use and care committee, following the National Institutes of Health guidelines for animal housing and care.

In vivo expression of IFF4.

IFF4 expression by wild-type C. albicans SC5314 was examined in the models of vaginal or hematogenous disseminated candidiasis. BALB/c mice were infected intravaginally or intravenously as described above. Twenty-four hours later, the mice were sacrificed and the vaginas or kidneys were collected as described above, cut into small pieces, and put into a red tube (Q-BIOgene) with a 1/4-inch ceramic bead (Q-BIOgene). Next, the samples were frozen in liquid nitrogen in the presence of 1 ml Tri reagent (Ambion) and homogenized twice with Fastprep FP120 (Bio 101 Thermo Electro corporation) at speed 4 for 25 seconds. Total-RNA samples were extracted according to the manufacturer's instructions. cDNA synthesis and RT-PCR were carried out as outlined above, except that IFF4-specific primers 1 and 2 were used to amplify IFF4-specific bands and primers EFB1a and EFB1b were used to amplify the housekeeping gene, EFB1 (Table 2). The PCR conditions were as follows: denaturing at 94°C for 2 min and amplification for 40 cycles at 94°C for 1 min, 52°C for 1 min, and 72°C for 2 min.

Statistical analysis.

Adherence and neutrophil-mediated killing were compared by the Mann-Whitney U test for unpaired comparisons. The nonparametric log rank test was utilized to determine differences in the survival times of the mice. Tissue fungal burdens among different groups were compared by the Steel test for nonparametric multiple comparisons (35) or the Mann-Whitney U test for unpaired comparisons, as appropriate. P values of <0.05 were considered significant.

RESULTS

Iff4p mediates C. albicans adherence to plastic.

Using a split-marker, triple-crossover strategy (Fig. 1), we made an incomplete GPI gene overexpression library that included the following members: RBT1, PGA8, SAP9, PLB5, SOD6, PLB7, HYR1, IFF2, IFF3, IFF4, FGR23, SOD5, ORF19.6420, ORF19.893, ORF19.4404, ORF19.13035, ORF19.3738, ORF19.3740, ORF19.6336, ORF19.2475, ORF19.5141, ORF19.5303, ORF19.5760, ORF19.2759, and ORF19.6321. The gene products were believed to be cell surface localized. The constructed strains had one allele of the targeted gene controlled by the TR promoter. Overexpression of these genes in the absence of DOX was confirmed by RT-PCR, comparing RNA samples extracted from mutants grown in the presence or absence of DOX (data not shown).

To determine if overexpression of a GPI-anchored protein affected the expression of other GPI proteins, we investigated the expression of Als1p (a known GPI-anchored protein) in 12 randomly picked strains, each of which overexpressed a different GPI-anchored protein. Als1p surface accumulation was quantified by direct immunofluorescence and flow cytometry (13). No effect on Als1p surface accumulation was detected among the 12 strains tested (Fig. 2).

To screen the library for an adherence function, we grew the overexpression strains in plastic tubes containing YPD with or without DOX. One clone, overexpressing the IFF4 gene, demonstrated enhanced adherence to plastic (Fig. 3). This clone was selected for further study. The other library strains showed no enhanced adherence to plastic.

FIG. 3.

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

IFF4 overexpression/suppression modulates the adherence of C. albicans to epithelial cells.

To completely control the expression of IFF4, we placed the second allele of IFF4 under the control of the TR promoter by interchromosomal recombination (Fig. 4). The generated strain (CAA10-31) had both IFF4 alleles controlled by the TR promoter. To avoid artifacts related to the chromosomal locus of the URA3 marker (4, 6, 8, 26, 49, 51), we also restored URA3-IRO1 to its original locus. After confirming the abrogation of IFF4 expression with DOX and IFF4 overexpression without DOX compared to the control (THE31) (Fig. 5A), we compared the abilities of these strains to adhere to human epithelial cells. Overexpression of IFF4 enhanced adherence to epithelial cells by more than 68% compared to IFF4 suppression (IFF4 without DOX), whereas suppression of IFF4 decreased adherence by 22% compared to the control strain grown with DOX (Fig. 5B).

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.

Overexpression of IFF4 increased the tissue fungal burden during vaginal candidiasis.

To determine the in vivo significance of IFF4-mediated adherence to epithelial cells, we used our murine model of candidal vaginitis (47). Since adherence is an early event during mucosal infection, we determined the vaginal fungal burden at 24 h in mice infected with the IFF4-conditional expressing strain (CAA10-31) or control C. albicans (THE31) grown without DOX or with DOX. Mice infected with C. albicans overexpressing IFF4 (without DOX) had a significantly higher vaginal fungal burden than those infected with IFF4-suppressed C. albicans (with DOX) or those infected with the control strain (Fig. 6A). There was no difference in vaginal fungal burdens between mice infected with C. albicans CAA10-31 grown under IFF4 suppression conditions and those infected with the control strain, THE31.

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.

We also sought to confirm that the IFF4 gene was expressed during murine vaginitis. Both wild-type C. albicans and IFF4-overexpressing C. albicans (the positive control) expressed IFF4 at 24 h during murine vaginitis, while no band was amplified from uninfected mice (negative control) (Fig. 6B). The constitutively expressed candidal EFB1 gene served as a quantitative control, as well as a control for lack of genomic contamination of C. albicans (50) (Fig. 6B).

IFF4 overexpression increases neutrophil killing of C. albicans.

To define potential interactions of IFF4 with host cells encountered during hematogenously disseminated infection, we investigated the conditional expression strain's ability to bind to or damage endothelial cells and its susceptibility to neutrophil-mediated killing. Overexpression/suppression of IFF4 did not have any effect on C. albicans adherence to or damage of endothelial cells (data not shown). However, C. albicans overexpressing IFF4 (without DOX) was more susceptible to neutrophil-mediated killing in vitro than suppressed cells (with DOX) or the control strain (Fig. 7A).

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.

Overexpression of IFF4 reduces the severity of hematogenously disseminated candidaisis in normal, but not neutropenic, mice.

To determine the in vivo significance of the enhanced neutrophil-mediated killing of IFF4-overexpressing C. albicans, we tested the contribution of IFF4 to C. albicans virulence during hematogenously disseminated candidiasis in immunocompetent and neutropenic mice. All C. albicans strains tested for virulence showed similar growth rates in vitro, with P values of >0.05 (doubling times were as follows: THE31 without DOX, 1.38 h; THE31 with DOX, 1.34 h; CAA10-31 without DOX, 1.45 h; and CAA10-31 with DOX, 1.35 h). In the immunocompetent-mouse model, overexpression of IFF4 (without DOX) significantly reduced the virulence of C. albicans CAA10-31 compared to mice infected with the same strain grown under IFF4 suppression conditions (with DOX) or those infected with the control strain, THE31 (Fig. 7B). In addition to enhanced survival, mice infected with the IFF4 overexpression strain had more than a 10-fold reduction in the kidney fungal burden compared to mice infected with the same strain grown under IFF4 suppression conditions (with DOX) or mice infected with the control strain. THE31 (Fig. 7C). In contrast, there was no difference in survival between neutropenic mice infected with C. albicans overexpressing IFF4 (without DOX) and those infected with the same strain grown under IFF4-suppressing conditions (with DOX) or those infected with the control strain (Fig. 7D).

Because overexpression of IFF4 decreased the virulence of C. albicans in mice infected via the tail vein, we hypothesized that this gene is likely not expressed in the wild-type strain during hematogenously disseminated candidiasis. Consistent with this hypothesis, we could not detect expression of IFF4 by RT-PCR in kidneys harvested from mice 24 h after intravenous infection with wild-type C. albicans (Fig. 8).

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.

DISCUSSION

We used a conditional gene overexpression/suppression approach to identify a candidal virulence factor, Iff4p. The IFF4 gene was found to promote epithelial adherence and mucosal infection in mice. Of note, a recent publication showed that the germ tubes of a C. albicans iff4 null mutant generated by gene disruption had decreased adherence to plastic surfaces (24). When we suppressed IFF4 expression in our URA3 repaired strains, we also found decreased adherence to epithelial cells using blastospores. However, the reduced epithelial adherence during suppression of IFF4 expression did not translate into a virulence phenotype in vivo, likely due to functional redundancy with other adhesins. It is also possible that our failure to detect a loss-of-function phenotype in vivo was due to low and undetectable IFF4 expression under suppression conditions (with DOX).

IFF4 belongs to the IFF gene family, which contains 12 members (IFF1 to IFF11 and HYR1) (37). The proteins have a conserved domain that does not display any significant homology to proteins with known functions (5). Most of the IFF family members encode proteins that exhibit the characteristic structure of GPI-anchored cell wall proteins, with the exception of Iff10p and Iff11p, which do not have any signal sequence for GPI anchor linkages (37). In addition, IFF7 appears to be located in the plasma membrane rather than in the cell wall (5). The function of this gene family is largely unknown, with the exception of IFF11, which encodes a secreted protein that is required for cell wall structure and virulence (5). In this study, we found that IFF4 promotes adherence to plastic and epithelial cells, but not endothelial cells. This epithelial adherence function was specific to IFF4, since overexpression/suppression of two other members of the IFF family, namely, IFF2 (HYR3) and IFF3, did not alter adherence (data not shown). The lack of altered adherence in these strains of IFF2 and IFF3 suggests that the adherence function of IFF4 is not localized in the conserved region.

Overexpression of IFF4 also increased the susceptibility of C. albicans to neutrophil-mediated killing in vitro. Susceptibility to neutrophil-mediated killing correlated with diminished virulence of C. albicans overexpressing IFF4 in the hematogenous model of infection using immunocompetent mice. The virulence of C. albicans overexpressing IFF4 was restored in neutropenic mice, emphasizing that the diminished virulence in normal mice was due to enhanced clearance of the yeast by neutrophils. Not surprisingly, wild-type C. albicans did not express the antivirulence factor, IFF4, during hematogenous seeding of the kidney. However, since forced expression of the gene caused a marked decrement in severity of infection, activation of IFF4 expression by small molecules is a potentially novel treatment for disseminated candidiasis.

The virulence factors of C. albicans are complex, and as demonstrated by transcriptional-profiling studies, different virulence factors may be operative in the same organism at different times, depending upon the anatomical context (2, 11, 12, 14). Our data therefore underscore the importance of testing C. albicans virulence factors that directly participate in host interaction in multiple models at different time points, reflecting the diversity of anatomical contexts that represent different in vivo niches of infection.

The advantage of our conditional-expression system is the ability to test both gain and loss of function in the same background strain simultaneously. In Saccharomyces cerevisiae, gene phenotypes and pathway mapping can be achieved by systematic gene overexpression (10, 45). However, in C. albicans, virulence gene function analysis using a similar approach is still in its infancy, despite the fact that gene dosage effects are well documented in the organism (9, 53). Multiple lines of research indicate that overexpression of a gene results in a detectable, authentic phenotype (13, 18-20, 25, 28, 40, 44). Our approach of conditional gene overexpression/suppression circumvents the need to construct a gene-complemented strain to verify virulence phenotypes, as is required in the gene disruption approach. Furthermore, insertion of the TR promoter into one allele is well suited to C. albicans, a diploid organism, because this genetic maneuver maintains the expression of any essential gene(s) even under suppression conditions (with DOX), allowing identification of the virulence function(s) of an essential gene. In contrast, gene disruption cannot be used to study the functions of essential genes.

One concern with using an overexpression strategy is so-called “capacity utilization.” For example, overexpression of a particular GPI-anchored protein could potentially result in limited space on the cell surface and restrict the expression of other GPI-anchored proteins. Alternatively, overexpression of a GPI-anchored protein could possibly alter the expression of other proteins because of limited enzyme activity to add GPI anchors. We found no evidence of a capacity utilization problem, as introduction of 12 randomly selected GPI genes, including IFF4, did not alter the surface accumulation of Als1p (another GPI-anchored protein).

In summary, we have described a conditional overexpression/suppression approach that can be broadly used to identify new genes encoding virulence-related phenotypes in C. albicans. This approach allows simultaneous evaluation of gain and loss of function. Using this new approach, we identified IFF4, a gene encoding a GPI-anchored protein, as an epithelial cell adhesin gene that modestly exacerbated mucosal infection in vivo. We also identified IFF4 as encoding an antivirulence factor during hematogenously disseminated infection. These data underscore the need to study virulence genes in multiple anatomical contexts to discern the full potential range of phenotypes and raise the possibility of targeting forced expression of IFF4 as a treatment for disseminated infection.