Fig 1.
Pyrimidine and 5-FC import and metabolism around some of the known and potential mechanisms of resistance.
5-Fluorocytosine (5-FC), cytosine, uracil and 5-fluorouracil (5-FU) are represented as blue, dark purple, light purple and light blue beads, respectively. For clarity, only two permeases known to import pyrimidines are shown: Fcy2 and Fur4. In prototrophic yeast strains, pyrimidines can be synthesized de novo from amino acids such as glutamine and aspartate [29]. The steps leading to the obtention of the nucleotide precursors uridine monophosphate (UMP) and uridine diphosphate (UDP) are catalyzed by Ura3 and Ura6, respectively, in the de novo pathway [30,31]. Alternatively, pyrimidines can be imported from the medium [32]. Uracil can also be obtained by deamination of cytosine by Fcy1 [33]. Finally, UMP can be obtained by the addition of a phosphate group by Fur1 [34]. 5-FC is a prodrug, which is metabolized the same way cytosine is, except its deamination leads to the obtention of the cytotoxic compound 5-FU [35]. Based on work done in laboratory strains or other fungal species, resistance to 5-FC can arise when either of the following three steps are compromised: import, conversion and activation, typically through mutations inactivating Fcy2 (for example), Fcy1 and Fur1, respectively [36]. We also hypothesize an alternative path to resistance in the pleiotropic drug response, wherein loss of mitochondrial function leads to derepression of the transcription factor Pdr3, which in turn overexpresses efflux pumps such as Pdr5. Figure created with BioRender, using the structures of Fcy1 (1P6O), Fur1 from C. albicans (7RH8), Ura3 (3GDL), Ura6 (1UKZ) and Pdr5 (7P04) and AlphaFold predictions of Fcy2 (P17064) and Fur4 (P05316) [37,38].
Fig 2.
Growth of individual mutants relative to the parental strains in various media.
Relative growth corresponds to the mean area under the curve (AUC, calculated on 22 h) from four replicate colonies, normalized by the WT for individual strains arrayed on solid media: YPD, SD, SD + 25 μg/mL 5-FC and SD + 6.25 μg/mL 5-FU. A two-way ANOVA followed by Tukey’s multiple comparison test was performed to compare the relative fitness of rho+ mutants to that of rho- mutants for each background in each condition. Statistical significance is as follows: ****, adj. p-value < 0.0001; ***, adj. p-value < 0.001; *, adj. p-value < 0.05; ns, not significant. Insets show the correlation between relative fitness in SD + 5-FC and relative fitness in SD + 5-FU, with the corresponding Spearman’s rank correlation coefficient (S. rho) and p-value.
Fig 3.
Cross-resistance of rho- mutants to 5-FC and fluconazole may be explained by efflux capacity.
A) Relative growth corresponds to the mean area under the curve (AUC, calculated for the first 22 h of incubation) from four replicate colonies, normalized by the WT for individual strains arrayed on solid media: YPD, SD, SD + 25 μg/mL 5-FC, SD + 0.5 μg/mL micafungin, SD + 2 μg/mL caspofungin, SD + 16 μg/mL nystatin and SD + 64 μg/mL fluconazole. A two-way ANOVA followed by Tukey’s multiple comparison test was performed to compare the relative fitness of rho+ mutants to that of rho- mutants for each background in each condition. Statistical significance is as follows: ****, adj. p-value < 0.0001; ***, adj. p-value < 0.001; *, adj. p-value < 0.05; ns, not significant. B) Rhodamine retention assay. Blue bars indicate the resulting detection thresholds (one negative control for each background i.e. cells not treated with rhodamine). abcΔ (s_012 in S1 Data) was used as the positive control. C, D) Rhodamine retention (data from panel B normalized with the WT) and relative growth in C) fluconazole or D) 5-FC (data from panel A) are compared, with the corresponding Spearman’s rank correlation coefficient (S. rho) and p-value.
Fig 4.
Genome sequences analysis and validation.
A) Upset plots showing the number of unique mutated genes (# genes) detected by gatk, samtools or both for each background. Categorical plots on top show the number of distinct genomes for which at least one variant was detected in a given gene (one dot per gene, colored according to the number of distinct mutations which have been detected). B) Distribution of the number of genes with variants per genome (as detected either by gatk or samtools) for each background. C) Data from Fig 2, where dots are colored by the mutated/deleted gene most likely to confer resistance. D) Growth assay in liquid medium for 35 resistant mutants for which a variant in FUR1 was detected. Cultures were inoculated in SD with or without 100 μg/mL 5-FC. Relative growth corresponds to the mean area under the curve from two replicates (AUC, calculated on 46h) normalized by the WT.
Fig 5.
Mutations observed in Fur1 lead to loss of function.
(A) Upset plots showing the number of unique mutations identified in Fur1 for each background. Categorical plots on top show the number of distinct genomes for which the same mutation was detected. Only one amino acid change resulted from two adjacent mutations (back-to-back in the same codon), all others corresponding to a single SNP or indel. B) Growth assay for 32 out of 35 resistant mutants tested in Fig 4D, with pMoBY expressing or not FUR1 from its native promoter. Relative growth corresponds to the area under the curve calculated on 25 h for the mutant with pMoBY-FUR1 divided by the same parameter for the mutant with pMoBY. Mutants were grown from single colonies in SD + 100 μg/mL 5-FC (+ G418 to maintain pMoBY). Colored dots indicate the same mutation was detected in the mutants from both backgrounds. Two mutations for which expression of the WT allele of FUR1 did not restore 5-FC sensitivity are highlighted. A closer look at the residues in the next panel suggests a putative role in assembly of the tetramer. C) Location of substitutions from our dataset on the predicted structure of Fur1, visualized using UCSF ChimeraX [40]. The tetramer (C. albicans Fur1 assembly 7RH8) is represented in light gray, with 1 chain in dark gray. Residues found to be mutated at least once in our dataset are colored in pink. UTP molecules are represented in sticks colored by atom. Two insets zoom in on one interface each (the slightly opaque one depicts the interface at the back of the represented structure). The highlighted residues (each one from a different monomer) appear to interact with one another to ensure proper assembly of the tetramer.
Fig 6.
FCY1-mediated resistance is context-dependent.
A) Predicted cross-feeding interactions between cells with different alleles of FCY1 and FUR1. Uracil (top) and 5-FU (middle) are expected to diffuse in the medium and be converted by Fur1 of neighboring cells, which would allow or inhibit growth, respectively. Notably, the growth of any FUR1+ cells, regardless of their FCY1 allele, will be inhibited because of 5-FU cross-feeding. On the contrary, any fur1- cells would grow despite 5-FU cross-feeding (bottom). Figure created with BioRender, using the structures of Fcy1 (1P6O) and Fur1 from C. albicans (7RH8). B) Medium conditioning assay. First, cultures of the deletion mutants were prepared in SD with or without 1.56 μg/mL 5-FC. After reaching an OD600 of 0.6, the cultures were filtered and the medium was transferred to a 96-well plate. Cultures were inoculated with either the WT or the other deletion mutant. Relative growth corresponds to the area under the curve (AUC) normalized by the WT and was obtained from biological triplicates. A two-way ANOVA followed by Tukey’s multiple comparison test was performed to compare the relative growth in SD with 5-FC to the one obtained in SD without 5-FC in each condition. Statistical significance is as follows: ****, adj. p-value < 0.0001; ns, not significant. C) Competition assay. Individual cultures of the deletion mutants were prepared in SD without selective pressure, then pooled in equal volumes, serially diluted and plated on large Petri dishes containing SD with 1.56 μg/mL 5-FC. For each counted colony, the genotype (fcy1Δ or fur1Δ) was confirmed by PCR. For each background, the relationship between CFU count of either mutant and the inoculum concentration was assessed by a Chi-square test of independence, performed on the cumulative sum of counts obtained from two independent experiments. D) Inhibition assay. For each background, a culture of the fcy1Δ mutant was inoculated in SD 1.7% agar with or without 25 μg/mL 5-FC and poured on top of thin 2% agar plates containing the matching medium. A culture of the fur1Δ mutant was then spotted on top. Uncropped pictures are provided in S8 Fig.