Genome-wide fitness test and mechanism-of-action studies of inhibitory compounds in Candida albicans

Abstract

Candida albicans is a prevalent fungal pathogen amongst the immunocompromised population, causing both superficial and life-threatening infections. Since C. albicans is diploid, classical transmission genetics can not be performed to study specific aspects of its biology and pathogenesis. Here, we exploit the diploid status of C. albicans by constructing a library of 2,868 heterozygous deletion mutants and screening this collection using 35 known or novel compounds to survey chemically induced haploinsufficiency in the pathogen. In this reverse genetic assay termed the fitness test, genes related to the mechanism of action of the probe compounds are clearly identified, supporting their functional roles and genetic interactions. In this report, chemical-genetic relationships are provided for multiple FDA-approved antifungal drugs (fluconazole, voriconazole, caspofungin, 5-fluorocytosine, and amphotericin B) as well as additional compounds targeting ergosterol, fatty acid and sphingolipid biosynthesis, microtubules, actin, secretion, rRNA processing, translation, glycosylation, and protein folding mechanisms. We also demonstrate how chemically induced haploinsufficiency profiles can be used to identify the mechanism of action of novel antifungal agents, thereby illustrating the potential utility of this approach to antifungal drug discovery.

Figure 1. An Overview of the CaFT

Individual heterozygous deletion strains contain two unique barcodes, up-tags (red) and down-tags (blue), flanked by two pairs of common primer sequences (black for all the up-tags, grey for all the down-tags) at the deleted allele. The CaFT pool contains 2,868 strains, representing ∼45% of the C. albicans genome. Aliquots of the pool are treated with an inhibitory compound (at different concentrations) or a mock treatment over 20 population doublings. The relative abundance of each strain is subsequently monitored by DNA microarrays competitively hybridized with amplified and labeled tags (using the common primer pairs) from the two treatments. The response of each strain to the compound is appraised by a normalized z-score, with a positive value indicating hypersensitivity, and a negative value relative resistance.

Figure 2. Characterization of Fluconazole-Induced HI by Spot Tests and CaFT Profiling

(A) Specificity of chemically induced HI by fluconazole as determined by spot tests. The heterozygous deletion strains corresponding to genes involved in the ergosterol biosynthesis pathway were tested against fluconazole at 0.75 μg/ml. Note that the two underlined strains, CYB5 (orf19.7049) and ERG4 (orf19.5379), were not present in the CaFT pool and that an ERG9 (orf19.3616) strain was not constructed. The HIS3 strain is used throughout this study as the wild-type control, since the HIS3 gene, as an auxotrophic marker, was used to construct all the heterozygous deletion strains (see Materials and Methods). (B) CaFT profile of fluconazole at 0.162 μg/ml (with fitness, F, value of 0.80; i.e., IC₂₀). The x-axis represents the strains in the CaFT pool, and y-axis the normalized z-scores of the corresponding strain in this experiment. (Although the z-scores of both up- and down-barcodes are assessed independently for each strain, the higher of the two was selected for display.) Significant outliers are highlighted by filled symbols, as their z-scores deviate significantly from the population. The z-scores of two groups of strains, corresponding to other genes involved in ergosterol biosynthesis or potential efflux pumps, are highlighted by open symbols. (C) CaFT profiles of fluconazole at different concentrations, with corresponding F values and ICs (in parentheses). In order to simplify a given CaFT profile, all the z-scores are displayed one-dimensionally according to the values but regardless of strain identities. Selected strains are marked so that their z-scores can be compared with the population and between experiments. Note that clotrimazole-induced HI of CDR1 and PDR16 was independently observed in the ScFT [13]. (D) Spot tests of heterozygous deletion strains identified in the CaFT (CDR1, PDR17, and ERG6) and the CDR2 and MDR1 strains (for their relevance, see the text) against multiple fluconazole concentrations. (E) Spot tests of homozygous deletion strains against fluconazole. (The orf19 designations are as follows: ERG11 = orf19.922, NCP1 = orf19.2672, CDR1 = orf19.6000, PDR17 = orf19.5839, ERG6 = orf19.1631, CDR2 = orf19.5958, MDR1 = orf19.5604, FCR1 = orf19.6817.)

Figure 3. CaFT Profiling and Characterization of Cerulenin-Induced HI

(A) CaFT profiles of cerulenin, with highlighted strains as follows: FAS1 (orf19.979), FAS2 (orf19.5949), and MDR1 (orf19.5604). A homozygous deletion of MDR1 is hypersensitive to cerulenin [21]; however, the heterozygous deletion strain showed no specific hypersensitivity at the ICs tested in the CaFT. (B) The spot tests of selected heterozygous deletion strains against cerulenin. Note that the hypersensitivity of the MDR1 strain was only seen at the highest concentration of cerulenin tested. (C) A model for regulating stoichiometry of the FAS complex in S. cerevisiae. As shown by Wenz et al. [20], the expression of ScFAS2 is repressed by an unknown transcription repressor (rep. X), which is in turn derepressed by free β subunit (Fas1p). According to this model, ScFAS2 expression is dependent on free Fas1p to control the normal stoichiometry of the FAS complex. A similar regulatory mechanism in C. albicans may explain the observed cerulenin-induced HI of FAS1 but not FAS2 (see text for details). (D) FAS1 is haploinsufficient under the standard growth conditions. In order to determine HI, cultures of the selected strains were first incubated for 6 h (reaching exponential growth) and then diluted to an OD₆₀₀ of 0.005. The fresh cultures were incubated for another 12 h, after which the OD was monitored for an additional 4 h.

Figure 4. CaFT Profiles of Microtubule Inhibitors

(A) CaFT profiles of nocodazole, benomyl, mebendazole and thiabendazole. Note that the ScFT experiments with these compounds have not been reported. Highlighted, in addition to TUB1 (orf19.7308) and TUB2 (orf19.6034), are BUB1 (orf19.2678), MAD2 (orf19.1040) and ESP1 (orf19.3356) strains that showed modest but reproducible hypersensitivity to nocodazole. The corresponding proteins are involved in mitotic checkpoint regulation and spindle pole body assembly [27]. Their relevance becomes apparent in the results described in Figure 7. (B) Spot tests of heterozygous deletion strains highlighted in (A) against benomyl. Note the lack of hypersensitivity of TUB2 in spot tests is consistent with the CaFT results, implying that the failure to detect hypersensitivity of the TUB2 strain in the CaFT is not due to poor performance of the barcodes. (C) TUB1 and TUB2 strains are haploinsufficient under the standard growth conditions (see Figure 3 legend for experimental details).

Figure 5. CaFT Profiles and Spot Tests of Radicicol

(A) CaFT profiles of radicicol with genes encoding co-chaperones of Hsp90p highlighted, including SGT1 (orf19.4089), CDC37 (orf19.5531), CNS1 (orf19.6052), orf19.7602 (an ortholog of ScAHA1), and CPR6 (orf19.7654) ([30] and references therein). Radicicol also elicited HI of ERG27 (orf19.3240), CDR1 (orf19.6000, an efflux pump), and PRS1 (orf19.1575, involved in cell integrity/stress signaling as reported [32]). (B) Spot tests of radicicol against heterozygous deletion strains corresponding to HSP90 and its co-chaperones.

Figure 6. Genetic Characterization of 5-FC and 5-FU in C. albicans

(A) Pyrimidine salvage pathway, and transport and metabolism of 5-FC (see text for details). (B) A hypothesis that fluorinated uracil blocks the formation of pseudouridine in rRNA. Note that pseudouridines are a prerequisite for ribosomal RNA processing [34]. (C and D) Functional characterization of potential permeases or transporters involved in the uptake of 5-FC (C) and 5-FU (D). A homozygous deletion strain for FCY2 (distinguished by the asterisk) and conditional shut-off (tetracycline repressible promoter, GRACE) strains [6] for genes indicated were tested in the absence (i.e., the non-repressing conditions, top panels) or the presence (i.e., the repressing conditions, bottom panels) of 100 μg/ml tetracycline against both compounds at the concentrations indicated. Note that 1) in (C) the FCY2 homozygous deletion strain was resistant to 5-FC, and 2) in (D) only the FCY21 GRACE strain exhibited slightly increased susceptibility to 5-FU under the non-repressing condition, suggesting that overexpression of FCY21 (from the tetracycline promoter, as observed in other cases, unpublished observations) may marginally facilitate the uptake of 5-FU. (E) Suppression of 5-FC (top) and 5-FU (bottom) antifungal activity by genetic depletion of Fur1p. The FUR1 and HIS3 (the control) GRACE strains were grown in the presence (100 μg/ml, +TET) or absence (−TET) of tetracycline with 5-FC or 5-FU at the concentrations indicated. Individual growth was normalized with growth of the HIS3 strain without antifungal drug in the presence or absence of tetracycline. Note that when the expression of FUR1 was repressed, the strain was markedly resistant to both compounds, whereas it was modestly hypersensitive under the non-repressing conditions, likely due to overexpression of FUR1 from the tetracycline promoter. (The orf19 designations are as follows: FCY2 = orf19.1357, FCY21 = orf19.333, NNT1 = orf19.4118, DIP5 = orf19.2445, and FUR1 = orf19.2640.)

Figure 7. CaFT Profiling and Characterization of Novel Antifungal Compounds

(A and B) Chemical structures (A) and the CaFT profiles (B) of novel and structurally related compounds that are predicted to affect microtubule dynamics. (C) Spot tests of representative compounds (ECC220, ECC284, and ECC275) confirm hypersensitivity detected in the CaFT. For orf19 designation, see Figure 4.

Figure 8. MOA Characterization of Compounds That Affect Microtubule Dynamics

(A) The phenotypic defects associated with a TUB1 conditional shut-off strain [6] examined under the repressing conditions (with time after the switch indicated on each photo). Under the non-repressing conditions, this strain was indistinguishable from the mock-treated wild-type cells (unpublished data). Nuclear migration was visualized by DAPI straining with morphology viewed under Nomarski optics. In (B–G), a strain carrying a Tub1p-GFP fusion was used. (B) The microtubule/spindle structures (as visualized by GFP-Tub1 fusion), nuclear migration/division, and morphology of the mock-treated cells. (C–G) The microtubule/spindle structures, nuclear migration, and morphology of cells treated with benomyl (C), nocodazole (D), ECC85 (E), ECC248 (F), and fluconazole (G). The concentrations (μg/ml, in parentheses) and incubation times are indicated on each photo.