Deletion of the Uracil Permease Gene Confers Cross-Resistance to 5-Fluorouracil and Azoles in Candida lusitaniae and Highlights Antagonistic Interaction between Fluorinated Nucleotides and Fluconazole

Abstract

We characterized two additional membrane transporters (Fur4p and Dal4p) of the nucleobase cation symporter 1 (NCS1) family involved in the uptake transport of pyrimidines and related molecules in the opportunistic pathogenic yeast Candida lusitaniae. Simple and multiple null mutants were constructed by gene deletion and genetic crosses. The function of each transporter was characterized by supplementation experiments, and the kinetic parameters of the uptake transport of uracil were measured using radiolabeled substrate. Fur4p specifically transports uracil and 5-fluorouracil. Dal4p is very close to Fur4p and transports allantoin (glyoxyldiureide). Deletion of the FUR4 gene confers resistance to 5-fluorouracil as well as cross-resistance to triazoles and imidazole antifungals when they are used simultaneously with 5-fluorouracil. However, the nucleobase transporters are not involved in azole uptake. Only fluorinated pyrimidines, not pyrimidines themselves, are able to promote cross-resistance to azoles by both the salvage and the de novo pathway of pyrimidine synthesis. A reinterpretation of the data previously obtained led us to show that subinhibitory doses of 5-fluorocytosine, 5-fluorouracil, and 5-fluorouridine also were able to trigger resistance to fluconazole in susceptible wild-type strains of C. lusitaniae and of different Candida species. Our results suggest that intracellular fluorinated nucleotides play a key role in azole resistance, either by preventing azoles from targeting the lanosterol 14-alpha-demethylase or its catalytic site or by acting as a molecular switch for the triggering of efflux transport.

INTRODUCTION

Candida lusitaniae is an opportunistic pathogenic yeast that can be responsible for deep and invasive fungal infections, particularly in neonatology and in hematology and oncology units (1, 2). Although much less frequent than Candida albicans, C. lusitaniae is feared because of its ability to acquire antifungal resistance during treatment, noticeably to polyene (3–5) and azole antifungals (1, 6), and it is considered to have decreased susceptibility to echinocandins in vitro (7).

In a preceding work, we demonstrated cross-resistance between flucytosine (5FC) and fluconazole (FLC) in several genetically unrelated clinical isolates of C. lusitaniae (8). These isolates were resistant to flucytosine, susceptible to fluconazole, and resistant to both antifungals, but only when they were used simultaneously. Such cross-resistance between flucytosine and fluconazole is intriguing in several regards, first because the two molecules have very different targets in the cell. Flucytosine is a prodrug that uses the salvage pathway of pyrimidine to exert its toxic activity (Fig. 1). The molecule enters the fungal cell using the specific transporter cytosine permease, then it is deaminated by a cytosine deaminase to 5-fluorouracil (5FU), which is finally converted to toxic fluorinated UMP (5FUMP) through the activity of uracil phosphoribosyl transferase (UPRTase). The diverse fluorinated derivatives that are generated further inhibit either the enzyme thymidylate synthetase and affect the synthesis of DNA or are incorporated into the mRNA and block the translation process (9). It is noteworthy that fluorinated UMP also can be generated through the de novo pathway of pyrimidine synthesis when fed with 5-fluoroorotic acid (Fig. 1), a toxic analog of orotate that is widely used for the selection of ura3 auxotrophs. Fluconazole is thought to enter the cell via facilitated transport, which has not been characterized so far (10), and then reach the endoplasmic reticulum, where its target is located (11), the lanosterol 14-alpha-demethylase (Erg11p), a key enzyme of the ergosterol biosynthetic pathway (12). Our ongoing goal is to decipher the underlying mechanisms of the flucytosine-fluconazole cross-resistance phenotype. We already know that a single genetic event supports both resistance to flucytosine and cross-resistance to fluconazole, either a mutation in the FCY2 gene encoding purine-cytosine permease (13, 14) or a mutation in the FCY1 gene encoding cytosine deaminase (14, 15). The mutation of FUR1, encoding UPRTase, confers cross-resistance to flucytosine and 5-fluorouracil but, surprisingly, does not support cross-resistance to fluconazole (15). Because of the structural similitude between flucytosine and 5-fluorouracil and of their different properties to support or not support cross-resistance to fluconazole, we decided in this study to complete the characterization in C. lusitaniae of the membrane transporters of the NCS1 (nucleobase cation symporter 1) family that are involved in the uptake transport of pyrimidine nucleobases, focusing on the ortholog of Fur4p, the plasma membrane permease which specifically transports uracil in Saccharomyces cerevisiae (16), and to determine its possible contribution to the fluoropyrimidine-fluconazole cross-resistant phenotype in C. lusitaniae. The results obtained allow us to conclude that fluorinated nucleotides derived from 5-fluorouracil are the key molecules responsible for resistance to triazole and imidazole antifungals through a mechanism that remains to be elucidated, either by preventing azoles from reaching the target, Erg11p, at the intracellular level or by activating efflux pumps, which leads to decreased azole concentrations at the enzyme target within the yeast cell. We also show that this mechanism is not restricted to C. lusitaniae but that it is widespread in the genus Candida.

FIG 1.

Salvage pathway and de novo biosynthesis pathway of pyrimidines in Candida lusitaniae. 5FC, flucytosine; 5FU, 5-fluorouracil; 5FUMP, 5-fluorouridine monophosphate; 5-FOMP, 5-fluoroorotidine monophosphate; 5FOA, 5-fluoroorotic acid. The nomenclature of the genes used is the same as that for Saccharomyces cerevisiae: URA1, dihydroorotate dehydrogenase; URA2, carbamoylphosphate synthetase-aspartate transcarbamylase; URA3, orotidine-5′-phosphate decarboxylase; URA4, dihydroorotase; URA5, orotate phosphoribosyltransferase; FCY1, cytosine deaminase; FCY2, purine-cytosine permease; FCY21, pseudogene paralog of FCY2; FUR1, uracil phosphoribosyltransferase; FUR4, uracil permease; DAL4, allantoin permease. The genes with asterisks have been characterized in C. lusitaniae.

MATERIALS AND METHODS

Candida lusitaniae strains and media.

The yeast strains used and constructed in this study are listed in Table 1. The Candida lusitaniae wild-type 6936 MATa strain (Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands) was used for gene cloning and as a susceptible reference strain for antifungal agents susceptibility tests. The clinical isolate CL38 MATα fcy2, which has a defective purine-cytosine permease resulting from a C505T nonsense mutation in FCY2, is resistant to flucytosine, susceptible to 5-fluorouracil and to fluconazole, and cross-resistant to the flucytosine-fluconazole association (8). This strain was used for genetic crosses. The 6936 MATa fur1Δ::URA3 strain, defective for uracil phosphoribosyl transferase, and the 6936 MATa fcy2Δ::URA3 strain, defective for purine-cytosine permease, were used as control strains for antifungal testing (15). The auxotrophic 6936 MATa ura3Δ (bearing a deletion of 360 bp in the orotidine 5′-phosphate decarboxylase gene) and 6936 MATa trp1Δ (bearing a deletion of 798 bp in the phosphoribosyl anthranilate isomerase gene) strains, along with the double auxotrophic 6936 MATa ura3Δ trp1Δ strain, were used for transformation experiments. These strains were described or constructed as already described elsewhere (17).

TABLE 1.

Candida yeast strains used and constructed in this study

Species and strain	Genotype	Origin	Reference or source
Candida lusitaniae
CL38 fcy2	MATα fcy2[C505T]	Clinical isolate, 5FCr	8
42720	MATα	Clinical isolate, ATCC	4
6936	MATa	Wild type, CBS
6936 ura3Δ	MATa ura3Δ360	6936	17
6936 trp1Δ	MATa trp1Δ798	6936	This study
6936 ura3Δ trp1Δ	MATa ura3Δ360 trp1Δ798	6936	This study
6936 fur1Δ	MATa fur1Δ::URA3 ura3[D95V]	6936 ura3[D95V]	15
6936 fcy2Δ	MATa fcy2Δ::URA3 ura3[D95V]	6936 ura3[D95V]	13
6936 dal4Δ	MATa dal4Δ::URA3 ura3Δ360	6936 ura3Δ	This study
6936 fur4Δ	MATa fur4Δ::TRP1 trp1Δ798	6936 trp1Δ	This study
6936 dal4Δ fur4Δ	MATa dal4Δ::URA3 fur4Δ::TRP1 ura3Δ360 trp1Δ798	6936 ura3Δ trp1Δ	This study
6936 dal4Δ fur4Δ fcy2	MATa dal4Δ::URA3 fur4Δ::TRP1 fcy2[C505T]	Progeny of CL38 fcy2 × 6936 dal4Δ fur4Δ	This study
6936 dal4Δ::DAL4	MATa dal4Δ::DAL4 ura3Δ360::URA3	6936 dal4Δ	This study
6936 fur4Δ::FUR4	MATa fur4Δ::FUR4 trp1Δ798::TRP1	6936 fur4Δ	This study
Other Candida species
C. albicans SC5314		Clinical isolate, ATCC	37
C. tropicalis IP1275		Institut Pasteur, Paris, France
C. parapsilosis 22019		ATCC
C. guilliermondii 6260		ATCC
C. glabrata 90030		ATCC
C. krusei 6258		ATCC

The yeast cells were cultivated in liquid YPD medium (1% yeast extract, 2% peptone, 2% glucose) at 35°C under agitation (250 rpm). After transformation, selection was accomplished on synthetic YNB medium (0.67% yeast nitrogen base without amino acids [Difco Laboratories] and 2% glucose) supplemented with 1 M sorbitol and, when needed, 25 μg/ml uracil, 0.5 mg/ml 5-fluorootic acid (5FOA), or 0.5 mg/ml 5-fluoroanthranilate. Other media used in this study were thiamine-free YNB medium (US Biological) for phenotypic characterization of the mutants, YNB medium without amino acids and without ammonium sulfate (0.17%; Difco Laboratories), YCB medium (1.17% yeast carbon base; Difco Laboratories) for mating and genetic crosses, and RPMI 1640 (Sigma) for antifungal testing. Solid media were obtained with 2% agar (Sigma).

Genetic crosses and ascospore isolation.

The dal4Δ::URA3 fur4Δ::TRP1 fcy2 triple mutant strain was obtained in the progeny of a genetic cross involving the 6936 MATa dal4Δ::URA3 fur4Δ::TRP1 double mutant strain and the CL38 MATα fcy2 sexually compatible clinical isolate (8). Genetic crosses were performed by mixing cells of opposite mating types on solid YCB as described previously (18). Ascospores were isolated after spheroplast formation and osmotic lysis of the parental yeast cells and were germinated on solid RPMI medium supplemented with 4 μg/ml 5-fluorouracil and 8 μg/ml 5-fluorocytosine at 35°C for 3 to 4 days. The genotype of the selected progeny then was verified using PCR assay and Southern blot analysis. Nucleotide sequencing was used to verify the presence of the C505T nonsense mutation in the fcy2 mutant allele inherited from the clinical isolate CL38 (14).

Antifungal agents and susceptibility tests.

Stock solutions of 2 mg/ml fluconazole (ICN Biomedicals Inc.), 2 mg/ml flucytosine (Sigma), 2 mg/ml 5-fluorouracil (Sigma), and 2 mg/ml 5-fluorouridine (5FUI) were prepared in sterile distilled water. A solution of 5-fluoroorotic acid was prepared at 100 mg/ml in dimethylsulfoxide (DMSO). Microdilution assays were performed according to CLSI standards (19) in RPMI 1640, pH 7.0, buffered with 0.165 M morpholinepropanesulfonic acid (MOPS). Yeast suspensions were diluted with RPMI 1640 medium at a final cell count of 10³ cells/ml. The 96-well plates were incubated for 48 h at 35°C, and growth was measured with an automated microtiter reader at 450 nm. All experiments were done at least in triplicate, and growth variations did not exceed 10%. Cross-resistance between fluoropyrimidines and azole antifungals were determined using Etest (bioMérieux, France) with concentration ranges of 0.002 to 32 μg/ml for itraconazole (ITC), voriconazole (VRC), posaconazole (POS), and ketoconazole (KTC) and 0.016 to 256 μg/ml for fluconazole. Etest strips were deposited onto solid RPMI medium left unsupplemented or supplemented with different concentrations of flucytosine or 5-fluorouracil (e.g., 8, 16, or 32 μg/ml) and inoculated with the yeast strains at the cellular density recommended by the supplier. The checkerboard method used for drug combination studies was also based on microdilution CLSI standards (19) using RPMI medium. Fluconazole concentrations ranging from 0.06 to 64 μg/ml were combined with flucytosine, 5-fluorouracil, or 5-fluorouridine concentrations ranging from 0.0075 to 0.5 μg/ml. The fractional inhibitory concentrations (FICs) of the drugs used in combination were calculated to obtain the FIC indices (∑FIC) as described previously (20) and are interpreted as stated in Instructions to Authors of Antimicrobial Agents and Chemotherapy (synergy for ∑FIC, ≤0.5; indifferent for ∑FIC, >0.5 and ≤4; antagonism for ∑FIC, >4) (21).

DNA extraction, PCR amplifications, and nucleotide sequencing.

Genomic DNA was extracted by a glass bead method for yeast cell disruption as previously described (17). The high-fidelity DNA polymerase Pfu Turbo (Stratagene) was used to amplify DNA fragments by PCR for cloning steps and for overlapping PCR. Routine PCRs were performed with Hot-Start Taq DNA polymerase (Qiagen) as recommended by the supplier. All primers used in this study (see Table S1 in the supplemental material) were synthesized by Eurofins MWG Operon (Europe). When required, PCR products were sequenced using the BigDye Terminator v1.1 kit (Applied Biosystems) at the Genotyping–Sequencing Pole of the Functional Genomic Platform of Bordeaux.

Yeast transformations.

Auxotrophic strains were transformed by an electroporation procedure previously described (22), slightly modified by adding 1 M sorbitol to the lithium acetate buffer (17), using 1 to 2 μg of transforming DNA per experiment. Prototrophic transformants were selected on YNB selective medium after 3 days of incubation at 35°C.

Identification of the C. lusitaniae FUR4 and DAL4 genes.

The open reading frames (ORFs) of the C. lusitaniae FUR4 and DAL4 genes were retrieved from the C. lusitaniae genome database of the Broad Institute (http://www.broad.mit.edu/annotation/fungi/candida_lusitaniae/index.html) and were identified as CLUG_05588.1 on supercontig 1.7 and CLUG_04321 on supercontig 1.5, respectively. Similarity searches in the database were performed with the basic local alignment search tool (BLAST) algorithm (23) using the amino acid sequence encoded by the genes FUR4 and DAL4 of Saccharomyces cerevisiae available at the Saccharomyces Genome Database (http://www.yeastgenome.org/).

Genetic constructions and gene deletions.

The genes DAL4 and FUR4 were amplified from the C. lusitaniae wild-type strain 6936 with the primer pairs F1.5/R1.5 and F1.7/R1.7 (see Table S1 in the supplemental material), respectively, and were cloned into the vector pGEM-T (Promega) to give the plasmids pGDAL4 and pGFUR4. Because of the occurrence of synthetic lethality between the ura3Δ and fur4Δ mutations, the FUR4 gene was deleted with the TRP1 marker and DAL4 was deleted with the URA3 marker. A deletion cassette was obtained for DAL4 by replacing an 800-bp Mfe1/Ale1 restriction fragment located within the ORF of DAL4 cloned in pGDAL4 with the 2.1-kbp GUN fragment made of the URA3 gene of C. lusitaniae flanked by two 300-bp noncoding repeated sequences, called NPT, derived from the bacterial neomycin phosphotransferase NPT1 gene, as described previously (15). Likewise, a deletion cassette for FUR4 was obtained by replacing a 997-bp Mfe1/NheI restriction fragment located within the ORF of FUR4 cloned in pGFUR4 with the 2.0-kbp GTN fragment made of the TRP1 gene of C. lusitaniae flanked by the two repeated NPT sequences. The deletion cassettes then were amplified from the plasmids and used in yeast cell transformation experiments to obtain, by homologous recombination, the simple mutant strains 6936 dal4Δ::URA3 and 6936 fur4Δ::TRP1 and the 6936 dal4Δ::URA3, fur4Δ::TRP1 double mutant strain. Reintegrant strains were obtained directly by transformation of the dal4Δ::URA3 and fur4Δ::TRP1 strains with a linear DNA fragment carrying the wild allele DAL4 or FUR4, respectively. The dal4Δ::DAL4 and fur4Δ::FUR4 reintegrant strains that had lost the genetic markers URA3 and TRP1 were selected onto YNB minimal medium supplemented with either 500 μg/ml of 5-fluorooroctic acid plus 25 μg/ml uracil or 500 μg/ml of 5-fluoroanthranilic acid plus 25 μg/ml tryptophan, respectively. Subsequently, the wild alleles URA3 and TRP1 were reintroduced to their own locus in the reintegrant strains. The genotype of all strains was verified by PCR and Southern blot hybridizations (see Table S1 and Fig. S1 in supplemental material).

Southern blotting.

Approximately 10 μg of C. lusitaniae DNA was digested with the appropriate restriction enzyme, separated by electrophoresis in a 1% agarose gel, and transferred onto nylon membranes (Hybond N+; Roche Molecular Biochemicals). Hybridization was carried out with digoxigenin (DIG)-labeled probes synthesized with a PCR DIG probe synthesis kit (Roche Molecular Biochemicals), as recommended by the supplier.

Uptake measurements.

Uptake measurements were performed as described previously (8, 24) by using [2-¹⁴C]uracil (3.81 GBq/mmol) from Moravek Biochemicals (Brea, CA, USA). Yeast cells were grown in YPD medium at 35°C under agitation, harvested by centrifugation at mid-log phase (5 × 10⁷ to 7 × 10⁷ cells/ml), and resuspended at a density of 4 × 10⁶ to 5 × 10⁶ cells/ml in 50 mM sodium citrate (pH 5.0) containing 2% glucose at 30°C. The reaction was started by adding [2-¹⁴C]uracil at concentrations ranging from 5 to 150 μM. After 4, 8, and 12 s of incubation, aliquots (0.3 ml) were filtered through 0.8-μm-pore-size cellulose acetate filters (Sartorius, Germany), washed twice with 3 ml of cold water, dried, and assayed for radioactivity. As already reported (24), we verified that the initial rate of uptake was constant over the first 10 s of base accumulation. The maximal rate of uptake (V_max) and the apparent Michaelis constant of transport (Kt_app) were calculated by nonlinear regression analysis of the initial rates of uptake versus the solute concentrations using the Michaelis-Menten equation.

Nucleotide sequence accession numbers.

The FUR4 and DAL4 sequences of the C. lusitaniae strain 6936 have been deposited in the GenBank database under accession numbers KC832301 and KC832300, respectively.

RESULTS

In silico identification of the FUR4 and DAL4 nucleobase transporters of the NCS1 family in the genome of C. lusitaniae.

In Saccharomyces cerevisiae, the nucleobase cation symporter 1 (NCS1) family is constituted by several uptake transporters of the salvage pathway for purines and pyrimidines and related molecules. The purine-cytosine permease gene FCY2 has been characterized already in C. lusitaniae (8, 13), and its paralog, FCY21, was shown to be unexpressed (13). Our aim was to characterize other genes of the NCS1 family, such as the uracil permease Fur4p (16), and to define its possible role in azole resistance. In S. cerevisiae, Fur4p has similarities with five other transporters (85% and 70% similarity with allantoin permease Dal4p and uridine permease Fui1p, respectively, and 50% similarity with the nicotinamide transporter Nrt1p and the thiamine transporters Thi7p and Thi72p). A tblastn analysis of the C. lusitaniae genome database (http://www.broadinstitute.org), using S. cerevisiae Fur4p as the query, allowed us to identify only two very similar paralogs in C. lusitaniae, CLUG_05588.1 and CLUG_04321.1, with the same percentages of amino acid identity and similarity to Fur4p (48% and 68%, respectively). The two ORFs are highly similar, exhibiting 49% identity and 80% similarity. Bioinformatics comparison to each member of the NCS1 transporters of S. cerevisiae did not allow us to predict a specific function for either one of the transporters. This could be achieved only from the phenotypic characterization of the null mutants, whose results for this study are provided below. For further clarity in this report, we use the following correct names for the ORFs: CLUG_05588.1 corresponds to FUR4, an intronless gene of 1,746 nucleotides (nt), encoding a predicted protein of 581 amino acids, and CLUG_04321.1 corresponds to DAL4, an intronless gene of 1,809 nt, encoding a predicted protein of 602 amino acids.

Characterization of NCS1 transporter-defective mutant strains of C. lusitaniae.

In S. cerevisiae, the Fur4p protein family mainly includes transporters of uracil, uridine, allantoin, and thiamine. C. lusitaniae is auxotrophic for thiamine and exhibits growth failure for thiamine concentrations as low as 1 × 10⁻³ μg/ml. The growth ability of the different fur4 and dal4 mutant strains of C. lusitaniae constructed in this study was tested in thiamine-free YNB, in both liquid and solid medium, supplemented with thiamine concentrations ranging from 2 × 10⁻⁵ to 0.012 μg/ml. Growth of the mutant strains was no more impaired in the presence of limiting concentrations of thiamine than that of the wild strain, indicating that Fur4p and Dal4p were not involved in thiamine uptake (Fig. 2). Similar experiments were performed with allantoin, a nitrogen-rich product of purine catabolism, in a concentration range from 1 × 10⁻³ mg/ml to 10 mg/ml. All of the strains which bore a dal4Δ deletion were unable to grow normally on YNB supplemented with 0.1 mg/ml allantoin (Fig. 2), indicating that Dal4p was involved in allantoin uptake, as confirmed by the wild-type phenotype restored in the dal4Δ::DAL4 reintegrant strain.

FIG 2.

Drop tests showing the growth of wild-type and mutant strains of C. lusitaniae on thiamine-free YNB supplemented with thiamine and YNB supplemented with allantoin as the sole source of nitrogen.

The ability of Fur4p and Dal4p to transport uracil and/or uridine then was investigated. First, we observed that growth of the ura3Δ auxotrophic strain on YNB medium could be restored with 25 μg/ml of uracil only and not with 25 μg/ml of uridine. We then measured the susceptibility of the different strains of C. lusitaniae to 5-fluorouracil, which is known to be transported by uracil permease in S. cerevisiae (25). The previously characterized 5-fluorouracil-resistant fur1Δ strain was used as a positive control (15). The results showed that the wild-type strain, the dal4Δ mutant, and the fur4Δ::FUR4 reintegrant strain all were susceptible to 5-fluorouracil, even at the lowest concentration used (Fig. 3). All of the mutant strains harboring the fur4Δ mutation were resistant to 5-fluorouracil, which suggested that Fur4p was responsible for 5-fluorouracil uptake and that it very likely was involved in uracil uptake in C. lusitaniae.

FIG 3.

Growth of wild-type and mutant strains of C. lusitaniae in RPMI supplemented with increasing concentrations of 5FU. Growth is expressed as a percentage of growth observed in RPMI without 5-fluorouracil. The wild-type strain 6936, the dal4Δ strain, and the fur4Δ::FUR4 reintegrant strain all are susceptible to 5-fluorouracil.

Uptake measurement of [¹⁴C]uracil in C. lusitaniae.

Uracil permease activity was assayed by the use of [¹⁴C]uracil uptake transport measurements in the wild-type and different mutant strains, and the kinetic parameters were determined (Table 2). All of the strains susceptible to 5-fluorouracil (MIC, ≤0.25 μg/ml) showed detectable transport activity, with V_max and Kt_app values for uracil not being statistically different, even for the strains defective for purine-cytosine permease (CL38 fcy2 strain) or defective for uracil phosphoribosyl transferase (6936 fur1Δ strain). These assays confirmed that only Fur4p, and not Dal4p, was involved in uracil uptake, as also shown by the restoration of both 5-fluorouracil susceptibility and uracil transport in the reintegrant strain 6936 fur4Δ::FUR4.

TABLE 2.

MICs of 5FU and kinetic parameters of [¹⁴C]uracil uptake for the C. lusitaniae strains^a

Strain	5FU MIC (μg/ml)	Uracil uptake parameterb
		Ktapp (μM)	Vmax (nmol/min/107 cells)
6936 wild type	≤0.25	23.04 ± 5.85	0.34 ± 0.03
CL38 fcy2	≤0.25	30.75 ± 16.5	0.17 ± 0.05
6936 fur1Δ	>256	33.30 ± 4.7	0.16 ± 0.02
6936 dal4Δ	≤0.25	22.63 ± 15.35	0.27 ± 0.05
6936 fur4Δ	128	NM	NM
6936 dal4Δ fur4Δ	128	NM	NM
6936 fur4Δ::FUR4	≤0.25	49.01 ± 10.9	0.23 ± 0.07

^aResults shown are MICs of 5FU and kinetic parameters of [¹⁴C]uracil uptake for the C. lusitaniae wild-type strain, the fur1Δ mutant, and mutants defective for nucleobase transporters in RPMI medium.

^bUracil uptake was measured over a concentration range from 5 to 150 μM. If transport could not be detected by using the highest concentration of uracil, it was noted as nonmeasurable (NM). Kt_app and V_max were calculated by nonlinear regression of the saturation curves obtained with at least five different solute concentrations, and all experiments were done in triplicate. The kinetic parameters of uracil uptake for the wild-type strain 6936 in YPD were a Kt_app of 21.64 ± 4.81 μM and a V_max of 1.92 ± 0.18 nmol/min/10⁷ cells.

Cross-resistance to azole antifungals.

In previous studies (8, 13, 15), we reported that mutant strains of C. lusitaniae, which were affected in different genes of the pyrimidine salvage pathway, had specific resistance patterns with regard to the fluoropyrimidines flucytosine and 5-fluorouracil and to the triazole antifungal fluconazole. For testing the ability of the fur4Δ mutant to support cross-resistance to 5-fluorouracil and fluconazole, we developed a new assay based on the use of fluconazole Etest strips and RPMI medium left unsupplemented or supplemented with increasing concentrations of 5-fluorouracil (Fig. 4). Using unsupplemented RPMI, the fluconazole MIC of the fur4Δ strain was 0.25 μg/ml. Using RPMI supplemented with 8, 16, and 32 μg/ml 5-fluorouracil, the fluconazole MIC increased in a dose-dependent manner to reach 12, 64, and 256 μg/ml, respectively. As a control strain, we used the UPRTase-deficient 6936 fur1Δ mutant strain, which was also 5-fluorouracil resistant (MIC, >256 μg/ml) but remained susceptible to fluconazole, whatever the 5-fluorouracil concentration used. We took advantage of the simplicity of the test to demonstrate the occurrence of the cross-resistance phenotype toward all other azole antifungals, i.e., the imidazole ketoconazole as well as the triazoles itraconazole, voriconazole, and posaconazole (Table 3), for the fur4Δ strain and the fcy2 clinical and laboratory mutant strains. The increase of MIC was much more marked for fluconazole, to which the mutant strains became fully resistant (with MICs ranging from 64 to 256 μg/ml).

FIG 4.

Etest determination of the fluconazole MIC of the C. lusitaniae fur4Δ and fur1Δ mutant strains on RPMI medium supplemented with increasing concentrations of 5FU or left unsupplemented.

TABLE 3.

MICs of azole antifungals for wild-type and mutant strains of C. lusitaniae^b

Antifungal	MIC by strain and culture mediuma
	6936 in RPMI-0	CL38 fcy2		6936 fcy2Δ		6936 fur4Δ
		RPMI-0	RPMI + 5FC	RPMI-0	RPMI + 5FC	RPMI-0	RPMI + 5FU
Fluconazole	0.5	0.75	96 (×128)	0.5	64 (×128)	0.5	256 (×512)
Itraconazole	0.006	0.023	1.5 (×65)	0.004	1.0 (×25)	0.008	0.38 (×48)
Voriconazole	0.012	0.016	0.38 (×24)	0.012	0.38 (×32)	0.012	0.5 (×42)
Posaconazole	0.008	0.012	0.19 (×16)	0.008	0.19 (×24)	0.008	0.125 (×16)
Ketoconazole	0.032	0.032	0.19 (×6)	0.032	0.25 (×8)	0.032	0.19 (×6)

^aSolid RPMI medium (RPMI-0) was supplemented with 32 μg/ml 5FC or 5FU according to the genotype of the strain. Numbers in parentheses indicate the increase in MIC observed on RPMI supplemented with 5-fluoropyrimidine compared to the level with unsupplemented RPMI.

^bMICs were determined with Etest strips on RMPI solid medium supplemented with 5-fluoropyrimidines or left unsupplemented.

Pyrimidine transporters are not involved in fluconazole uptake.

In our previous works, we suggested that cross-resistance between flucytosine and fluconazole was due to the competitive inhibition of fluconazole uptake by extracellular flucytosine (8, 13, 15). The data obtained with 5-fluorouracil and the fur4Δ mutant in the present work are consistent with this hypothesis, showing, notably, that the MIC of fluconazole increases as the 5-fluorouracil concentration increases. To determine whether the pyrimidine transporters could mediate fluconazole import, triple mutants were selected in the progeny of a genetic cross involving an fcy2 parental strain having a nonfunctional purine-cytosine permease and the fur4Δ dal4Δ double mutant constructed in this study. None of the triple mutants exhibited an increase of MIC for the different azole antifungals compared to the parental strains, indicating that these transporters were not involved in azole uptake.

Attempts to induce fluconazole resistance with other chemical compounds.

In another set of experiments, RPMI medium containing 8, 16, or 32 μg/ml fluconazole was supplemented with different concentrations of the pyrimidine nucleobase cytosine or uracil, with the amino acid histidine, which contains an imidazole ring, or with thiamine and thiamine pyrophosphate, both harboring an aminopyrimidine and a thiazole ring. None of these molecules was able to promote resistance to fluconazole like fluoropyrimidines did in the fur4Δ strain (data not shown). This result pointed out the particular importance of the fluorine atom in azole resistance. Nevertheless, we failed to trigger azole resistance by the use of NaF (concentration range, 0.7 to 200 mM; the MIC of NaF measured in this work for the fur4Δ and wild-type 6936 strains was 400 mM for both), 2-amino-5-fluorobenzoic acid (concentration range, 0.2 to 50 mM; the MIC of 5-fluorobenzoate measured in this work was 100 mM for fur4Δ and wild-type 6936 strains), and 5-fluorodeoxyuridine (concentration range, 0.5 to 256 μg/ml; the MIC measured in this work was ≥256 μg/ml for the fur4Δ and wild-type 6936 strains).

Subinhibitory doses of fluorinated pyrimidines mediate resistance to fluconazole in a susceptible wild-type strain.

We previously reported that a fur1 mutant strain, which had lost the ability to convert 5-fluorouracil to phosphoribosylated derivatives, was resistant to both flucytosine and 5-fluorouracil but not cross-resistant to fluconazole (15), indicating that the fluorinated nucleobase 5-fluorouracil, even when it accumulated into the fungal cell, was unable to promote fluconazole resistance by itself. Only mutants of the pyrimidine salvage pathway, which kept functional uracil phosphoribosyl transferase activity, such as fcy1, fcy2, and fur4 mutant strains, exhibited cross-resistance to azole in the presence of certain concentrations of fluoropyrimidines. We hypothesized that small amounts of 5-fluorouracil, present in the cytoplasm of the mutant fungal cells, could be converted by Fur1p into fluorinated nucleotides that then would be responsible for the triggering of azole resistance.

To test this hypothesis, we studied the effect of the supplementation of RPMI medium with subinhibitory doses of 5-fluorouracil or flucytosine (from 0.19 × 10⁻⁴ μg/ml to 1 μg/ml) on the growth of the fluconazole-susceptible wild-type strain 6936 (fluconazole MIC, 1 μg/ml) in the presence of either 8 μg/ml or 16 μg/ml fluconazole. Under these conditions, the growth of wild-type yeast cells was inhibited by fluconazole, except in the presence of 5-fluorouracil concentrations ranging from 0.06 to 0.25 μg/ml (i.e., from 0.48 to 1.92 μM) or flucytosine from 0.015 to 0.06 μg/ml (i.e., from 0.12 to 0.48 μM) (Fig. 5). The difference of active concentrations between flucytosine and 5-fluorouracil for triggering azole resistance can be explained by the difference of affinity of the respective flucytosine and 5-fluorouracil permeases, with the purine-cytosine permease having about 4 times more affinity for cytosine (6.2 μM) (8) than uracil permease for uracil (21.6 μM) in the 6936 wild-type cells of C. lusitaniae.

FIG 5.

Growth of the antifungal-susceptible wild-type strain 6936 in RPMI supplemented with inhibitory concentrations of fluconazole along with various low concentrations of 5-fluorouracil (A) or flucytosine (B). Growth is given as a percentage of the growth of the wild-type strain in antifungal-free RPMI. The lack of bars indicates a complete inhibition of growth.

The de novo pathway of pyrimidine synthesis can be used to generate fluorinated pyrimidine molecules and to trigger fluconazole resistance.

5-Fluoroorotic acid is a fluorinated analog that can enter the de novo pathway of pyrimidine biosynthesis to generate the toxic 5-fluoro-UMP through two enzymatic activities, orotate phosphoribosyl transferase (URA5) and orotidine 5′-phosphate decarboxylase (URA3) (Fig. 1). We observed that the wild-type strain of C. lusitaniae cultivated in the presence of uracil, 5-fluoroorotic acid, and inhibitory concentrations of fluconazole could exhibit, at certain concentrations of 5-fluoroorotic acid, growth corresponding to 40% of that observed in inhibitor-free RPMI medium (Fig. 6). Growth of a ura3Δ mutant strain under the same conditions was much lower, indicating that the fluorinated orotidine phosphate, which is probably accumulated in the mutant strain, was less efficient at triggering azole resistance than fluorinated uridylic acid.

FIG 6.

Growth of wild-type and ura3Δ strains of C. lusitaniae in RMPI supplemented with 8 μg/ml fluconazole and different concentrations of 5FOA. Growth is expressed as the percentage of growth observed in inhibitor-free RPMI. The lack of bars indicates the complete inhibition of the growth.

Resistance to fluconazole can be induced in other species of Candida using flucytosine, 5-fluorouracil, and 5-fluorouridine.

We then investigated whether the interaction between fluoropyrimidines and fluconazole was specific to C. lusitaniae or if it could be observed in other strains of different Candida species. First, the MICs of fluoropyrimidines and fluconazole were determined for a representative panel of strains of different species, including the strain ATCC 42720 of C. lusitaniae, used as a control (Table 4). All of the strains were susceptible to flucytosine except C. krusei, which was susceptible dose dependent (SDD), indicating that the metabolic pathway involved in the uptake and the intracellular conversion of flucytosine to toxic derivatives of fluorouracil was functional in each strain of each species. Species of the CTG clade of Candida are divided in two groups, those which were susceptible to 5-fluorouracil and resistant to 5-fluorouridine (C. tropicalis and C. lusitaniae) and those which were resistant to 5-fluorouracil and susceptible to 5-fluorouridine (C. albicans, C. parapsilosis, and C. guilliermondii), but all were susceptible to fluconazole. The two other species, C. glabrata and C. krusei, were SDD for 5-fluorouracil, 5-fluorouridine, and fluconazole.

TABLE 4.

MICs of 5FC, 5FU, 5FUI, and FLC for strains of different Candida species determined by the CLSI method

Strain	MICa (μg/ml)
	5FC	5FU	5FUI	FLC
C. albicans SC5314	≤0.12	64	≤0.12	≤0.25b
C. tropicalis IP1275	≤0.12	≤0.12	>64	2
C. parapsilosis 22019	0.25	>64	≤0.12	2
C. lusitaniae 42720	≤0.12	≤0.12	>64	2
C. guilliermondii 6260	0.12	>64	0.25	8
C. glabrata 90030	0.25	2	16	16
C. krusei 6258	16	8	2	32

^aThe value corresponding to the IC₅₀ (concentration inhibiting at least 50% of the growth of the control) is given only for fluconazole and for C. albicans. Other MICs are given for 100% growth inhibition.

^bValue corresponding to IC₅₀ (concentration inhibiting at least 50% of the growth of the control) is given only for fluconazole and for C. albicans. Other MICs are given for 100% growth inhibition.

Each strain then was cultivated in liquid RPMI containing fluconazole at a concentration corresponding to 4 times the MIC (e.g., from 8 μg/ml for C. tropicalis to 128 μg/ml for C. krusei) and a fluoropyrimidine (flucytosine, 5-fluorouracil, or 5-fluorouridine) at various concentrations, from 0.002 μg/ml to 1 μg/ml or from 0.25 μg/ml to 128 μg/ml, depending on the susceptibility or resistance of the strain to the relevant fluoropyrimidine. Results for three species (C. guilliermondii, C. glabrata, and C. tropicalis) are shown in Fig. 7, and results for the others are shown in Fig. S2 in the supplemental material (C. parapsilosis, C. krusei, and C. lusitaniae). With the exception of C. albicans SC5314 (data not shown), antagonism between fluoropyrimidines at certain concentrations and fluconazole could be observed for each strain of each species, allowing growth in the medium with inhibiting concentrations of fluconazole up to 70% of the growth of the control without inhibitor. The results of an RPMI microdilution checkerboard analysis confirmed strong antagonism (∑FIC, >4) between fluconazole and flucytosine and between fluconazole and 5-fluorouracil or 5-fluorouridine for strains of C. lusitaniae and for strains of C. guilliermondii, C. tropicalis, and C. glabrata (Table 5). Antagonism results specifically from an elevation of the MIC of fluconazole in the presence of fluoropyrimidine, whereas the MIC of fluoropyrimidine is invariable and indifferent to the presence of fluconazole.

FIG 7.

Growth of C. guilliermondii ATCC 6260, C. glabrata ATCC 90030, and C. tropicalis IP1275 in RPMI medium supplemented with inhibitory concentrations of fluconazole (4× the MIC for each strain) in the presence of different concentrations of the fluoropyrimidines 5FC, 5FU, or 5FUI. Growth is expressed as a percentage of the growth observed in inhibitor-free medium. The lack of bars indicates the complete inhibition of growth.

TABLE 5.

FIC index determined by checkerboard analysis of fluconazole and fluoropyrimidine combinations in different Candida species

Strain	Antifungal combinationa	FIC variation	∑FIC	Interpretationb
C. lusitaniae 6936	FLC + 5FC	1–128.5	28	ANT
C. lusitaniae 6936	FLC + 5FU	1–64.5	15	ANT
C. lusitaniae 42720	FLC + 5FC	1–64.5	14	ANT
C. lusitaniae 42720	FLC + 5FU	1–64.5	12.8	ANT
C. tropicalis IP1275	FLC + 5FC	1–32.5	8.6	ANT
C. tropicalis IP1275	FLC + 5FU	1–32.5	10	ANT
C. guilliermondii 6260	FLC + 5FC	1–16.5	5.6	ANT
C. guilliermondii 6260	FLC + 5FUI	1–16.5	6.3	ANT
C. glabrata 90030	FLC + 5FC	1–16.5	4.4	ANT
C. glabrata 90030	FLC + 5FU	1–16.5	4.4	ANT
C. krusei 6258	FLC + 5FC	1–>4	2	IND
C. krusei 6258	FLC + 5FU	1–>4	1.5	IND
C. parapsilosis 22019	FLC + 5FC	1–4.5	1.5	IND
C. parapsilosis 22019	FLC + 5FUI	1–4.5	1.8	IND

^aThe fluoropyrimidines selected for this analysis were 5FC and either 5FU or 5FUI according to the susceptibility of the Candida strain to one or the other of these two molecules. The higher starting antifungal concentrations were 64 μg/ml for FLC and 0.5 μg/ml for fluoropyrimidines.

^bANT, antagonism; IND, indifferent.

DISCUSSION

This work completes the characterization of uptake transporters of the nucleobase cation symporter 1 family, which are involved in the pyrimidine nucleobases salvage pathway, and of related molecules in the opportunistic yeast C. lusitaniae. After having characterized the purine-cytosine permease in clinical isolates and in laboratory mutants of C. lusitaniae (8, 13), we were interested in the uptake transport of uracil. In S. cerevisiae, members of the NCS1 family generally possess 12 transmembrane α-helical spanners and are thought to function in uptake by a substrate-H⁺ symport mechanism. More specifically, uracil uptake is mediated by the protein Fur4p (26), a permease that has similarity to a group of five other transporters specialized in the uptake of uridine, allantoin, nicotinamide, and thiamine. In contrast, the Fur4p family in C. lusitaniae, as well as in the CTG clade of Candida species, consists of only two very similar transporters. Using gene deletion and loss-of-function analysis, we showed that one of them, named Dal4p, was involved in the uptake transport of allantoin, a nitrogen-rich molecule generated from purine catabolism, that can be used as a nitrogen source by yeast cells (27). The second one was called Fur4p and was shown to be the permease for uracil. [¹⁴C]uracil uptake transport was abolished in the fur4Δ mutant but not in the dal4Δ mutant, indicating that Fur4p specifically transported uracil. In wild-type strains of S. cerevisiae, the kinetic parameters of the transport of cytosine by Fcy2p and uracil by Fur4p are very similar, with Kt_app values around 2 μM and V_max values in the range of 8 to 10 nmol/min/10⁷ cells (16, 28). In the C. lusitaniae wild-type strain, Fcy2p shows a greater affinity for cytosine [Kt_app, 6 μM] (8) than Fur4p for uracil [Kt_app, 20 to 24 μM], which itself has 10 times less affinity for uracil than Fur4p of S. cerevisiae (2.5 μM) (16). These differences can be interpreted as a gain of substrate specificity of the transporters during evolution of S. cerevisiae following whole-genome duplication, while the nucleobase transporters of the more ancient clade of Candida would have retained certain plasticity for different substrates. This can be illustrated by comparing C. lusitaniae to C. albicans. Growth of a ura3Δ strain of C. lusitaniae is restored with uracil but not with uridine, suggesting that Fur4p is unable to transport the nucleoside form, which is consistent with the resistance of C. lusitaniae to 5-fluorouridine. Inversely, it is known that growth of a ura3 mutant strain of C. albicans can be restored with uridine but not with uracil. In parallel, we observed that C. albicans SC5314 was susceptible to 5-fluorouridine (MIC, ≤0.12 μg/ml) and had a high MIC for 5-fluorouracil (64 μg/ml), although the wild-type strain possesses a functional uracil phosphoribosyl transferase (29), suggesting that resistance to 5-fluorouracil resulted from a defect in uptake transport (30). Like C. lusitaniae, C. albicans possesses only two transporters that are very similar to Fur4p and Dal4p. It can be inferred that Fur4p transports uracil and not uridine in C. lusitaniae, while it transports uridine and not uracil in C. albicans. This shows that an ortholog protein like Fur4p may have evolved differently with regard to its substrate specificity, even in two species belonging to the same evolutionary clade.

Interestingly, the fur4Δ null mutant of C. lusitaniae is resistant to 5-fluorouracil and cross-resistant to 5-fluorouracil and fluconazole, but also to all other azole antifungals used in this study, when both the fluoropyrimidine and the azole are used simultaneously. This phenotype contrasts completely with that of the fur1Δ mutant, which is also resistant to 5-fluorouracil but has no cross-resistant phenotype to fluconazole (15). Indeed, in previous works with C. lusitaniae, we demonstrated that mutations affecting the purine-cytosine permease Fcy2p and the cytosine deaminase Fcy1p, but not uracil phosphoribosyl transferase Fur1p, conferred cross-resistance to flucytosine and fluconazole in a dose-dependent manner. It was speculated that flucytosine, but not 5-fluorouracil, was responsible for competitive inhibition of the uptake transport of fluconazole (8, 13–15). Although it was shown more recently that flucytosine itself did not compete for [³H]fluconazole import (10), it cannot be completely excluded that the nucleobase transporters in C. lusitaniae could transport both fluoropyrimidines and azole antifungals. It has long been known that the nucleoside transporter of human cells can translocate a large variety of toxic anticancer and antiviral nucleoside analogs and also structurally unrelated pharmacologically active molecules (for a review, see reference 31). Closer to the field of microbiology, an adenosine transporter of Trypanosoma brucei was shown to be implicated in the uptake transport of two structurally unrelated antiparasitic molecules, melarsoprol and pentamidine, which are used to treat sleeping sickness in humans and Nagana disease in animals (32), with a null mutant of this transporter being cross-resistant to both drugs (for a review, see reference 33). Little is known about the uptake transport of azole antifungals in Candida yeasts, except that it is mediated by facilitated diffusion with a very low V_max in C. albicans through transporters that remain to be identified (10). Unfortunately, our results indicate that the uptake transport of fluconazole and other azoles probably is not mediated, or not exclusively mediated, by nucleobase transporters in C. lusitaniae, because a triple mutant harboring the fur4Δ and dal4Δ null mutations, associated with the clinical nonsense mutation fcy2_[C505T], showed no reduced susceptibility to fluconazole.

From previous and present data, we observed that only some mutations leading to the accumulation of the fluorinated nucleobases flucytosine and 5-fluorouracil, either at the intracellular (fcy1Δ mutant) or at the extracellular (fcy2Δ and fur4Δ mutants) level, were able to confer cross-resistance to fluoropyrimidines and azole antifungals. All of these mutants possessed a functional FUR1 gene. On the other hand, mutation of the FUR1 gene itself is the only one that results in resistance to flucytosine and 5-fluorouracil but not in cross-resistance to azole antifungals. Thus, we hypothesized that some intracellular 5-fluorouracil, derived from spontaneous deamination of flucytosine in the fcy1Δ mutant and from the entry by diffusion of fluoropyrimidines in the fcy2Δ and fur4Δ mutants, can be converted by Fur1p to fluorinated nucleotides, which in turn are responsible for cross-resistance to azoles. This hypothesis was verified by showing that subinhibitory doses of fluorinated pyrimidines were able to trigger cross-resistance to fluconazole in susceptible wild-type strains of C. lusitaniae by both the salvage and the de novo pathways of pyrimidine synthesis. The demonstration was enlarged to other Candida species by showing that fluconazole resistance could be induced by flucytosine, 5-fluorouracil, and 5-fluorouridine in a panel of strains of C. parapsilosis, C. tropicalis, C. guilliermondii, C. krusei, and C. glabrata but not in C. albicans SC5314, maybe because of the strong trailing growth of that strain in the presence of fluconazole alone. The exact mechanism by which the fluorinated nucleotides do trigger cross-resistance is not known. A competition for transport at the intracellular level between fluorinated nucleotides and azole antifungals cannot be ruled out, because Erg11p, the target of azoles, is located at the membrane of the endoplasmic reticulum, and access for fluconazole could require a membrane transporter. Another possibility is that fluorinated nucleotides block the entry or the fixation of azole antifungals into the catalytic pocket of Erg11p. A third possibility relies on the activation of an efflux system. In C. glabrata, the recently reported antagonism between flucytosine and azole antifungals was explained in part as the result of a Pdr1p-dependent overexpression of the Cdr1p efflux pump (34). Alternatively, it was recently described in bacteria that a fluoride-responsive riboswitch class of noncoding RNA (35) could be involved in fluoride export or detoxification. As riboswitches also were described in eukaryotes (36), and even though we failed to induce cross-resistance to azoles in C. lusitaniae with NaF or 5-fluorobenzoate, it would be interesting to investigate if a fluoride export mechanism does exist in Candida yeasts and if it is involved in the export of fluorinated azole antifungals.