Nonsense and missense mutations in FCY2 and FCY1 genes are responsible for flucytosine resistance and flucytosine-fluconazole cross-resistance in clinical isolates of Candida lusitaniae

Abstract

The aim of this work was to elucidate the molecular mechanisms of flucytosine (5FC) resistance and 5FC/fluconazole (FLC) cross-resistance in 11 genetically and epidemiologically unrelated clinical isolates of Candida lusitaniae. We first showed that the levels of transcription of the FCY2 gene encoding purine-cytosine permease (PCP) in the isolates were similar to that in the wild-type strain, 6936. Nucleotide sequencing of the FCY2 alleles revealed that 5FC and 5FC/FLC resistance could be correlated with a cytosine-to-thymine substitution at nucleotide 505 in the fcy2 genes of seven clinical isolates, resulting in a nonsense mutation and in a putative nonfunctional truncated PCP of 168 amino acids. Reintroducing a FCY2 wild-type allele at the fcy2 locus of a ura3 auxotrophic strain derived from the clinical isolate CL38 fcy2(C505T) restored levels of susceptibility to antifungals comparable to those of the wild-type strains. In the remaining four isolates, a polymorphic nucleotide was found in FCY1 where the nucleotide substitution T26C resulted in the amino acid replacement M9T in cytosine deaminase. Introducing this mutated allele into a 5FC- and 5FC/FLC-resistant fcy1Delta strain failed to restore antifungal susceptibility, while susceptibility was obtained by introducing a wild-type FCY1 allele. We thus found a correlation between the fcy1 T26C mutation and both 5FC and 5FC/FLC resistances. We demonstrated that only two genetic events occurred in 11 unrelated clinical isolates of C. lusitaniae to support 5FC and 5FC/FLC resistance: either the nonsense mutation C505T in the fcy2 gene or the missense mutation T26C in the fcy1 gene.

FIG. 1.

Expression analysis of C. lusitaniae FCY2 and FCY21 genes in clinical isolates and in the wild-type strain 6936. Shown is Northern blot analysis of total RNA from the clinical isolates CL29, CL31, CL38, and CL42 and from strain 6936 hybridized with FCY2 and FCY21 RNA probes. Membranes were stripped and hybridized with the 18S rRNA probe for RNA loading control.

FIG. 2.

Southern blot hybridization (B) and schematic representation (A) of a resident fcy2 locus and of molecular events that occurred in transformants. (A) The signals revealed by the labeled probes (each marked with an asterisk) correspond to those expected from the genomic restriction map. The hybridization patterns were visualized with the Fi, URA3, and BLA probes of BamHI-digested genomic DNA from CL38-5 (a) and from the reintegrant strain CL38-5 FCY2 (b). (B) Hybridization of genomic DNA with the URA3 probe revealed an additional fragment of 19.4 kb in both strains, which corresponds to the resident locus URA3. DNA fragment sizes are indicated in kilobases.

FIG. 3.

Expression levels of the FCY1 gene in the wild-type strain 6936 and in the clinical isolates CL26, CL119, CL174, and CL216. A representative semiquantitative RT-PCR analysis is shown (from three independent experiments). For each target gene, the amount of transcription was compared with that of the actin 1 (ACT1) gene.

FIG. 4.

Amino acid sequence alignment of cytosine deaminases encoded by the FCY1 gene from the C. lusitaniae wild-type strain 6936 (GenBank accession no. DQ372926), C. albicans (GenBank accession no. AJ616007), S. cerevisiae (GenBank accession no. NC001148), and C. glabrata (GenBank accession no. CR380950). The polymorphic amino acid residue that has been identified in Fcy1p sequences of C. lusitaniae at position 9 is framed. The regions involved in metal coordination and substrate protonation, described as signature sequences of the deaminase, family are shaded in gray. The amino acid residues that are thought to be essential for the binding of cytosine and 5FC are in boldface. Absolutely conserved residues are marked with asterisks, highly conserved residues are marked with colons, and weakly conserved residues are marked with periods.