Contribution of the FCY2 gene to flucytosine susceptibility in C.tropicalis

tropicalis

4.2.1 LOH at the FCY2 locus was associated with 5FC resistance

The 5FC susceptibility of the clinical isolate YM020291 and its genetic modified counterparts was examined in previous work. The MIC for isolate YM020291 and its FCY2^S/FCY2^S homozygous mutants YLO415 and YLO416 was 0.5 µg/ml, and the MIC for its FCY2^R/FCY2^R homozygous mutants YLO417 and YLO418 was higher than 64 µg/ml (Chia-Chen Wu, 2009). In this study, another clinical isolates YM060800 was chosen to create the FCY2 homozygous mutation strains for the purpose to confirm the results. Both strains carried the least amount of SNPs at the FCY2 locus, which made it easier to examine the factors that would influence the 5FC resistance to C. tropicalis. Isolates YM020291 and YM060800 were separately collected from middle and southern areas of Taiwan during the TSARY programs in 2002 and 2006. Also, they had different MLST diploid sequence types (DST 155 and DST 200, respectively) (Chou et al., 2007; Li et al., 2009). Thus, they were considered having different genetic backgrounds, and factors that contributed to the drug-resistance phenotype could be evaluated under different genomic backgrounds.

Here, 5FC susceptibilities were analyzed among both clinical isolates as well as their engineered derivatives to determine whether the homozygosity at the FCY2 locus responsible for the 5FC susceptibility in C. tropicalis. In agreement with previous result, the wild type strains YM020291 and YM060800 as well as their FCY2^S/FCY2^S homozygous mutants were susceptible to 5FC, while their FCY2^R/FCY2^R homozygous mutants were resistant to this drug. All in all, it has demonstrated unambiguously that loss of both FCY2^S alleles or the existence of the FCY2 homozygous resistant alleles was the reason for 5FC resistance in C. tropicalis, suggesting that the FCY2^S allele appeared to be dominant and indicating the linkage between 5FC resistant phenotype and the occurrence of loss of heterozygosity.

4.2.2 The nonsense mutation in the FCY2 gene, accompanied by LOH,

contributing to 5FC resistance in C. tropicalis

The effect of each polymorphic site on the 5FC susceptibility to C. tropicalis was further determined by evaluating the MICs of each single mutation strain. Although two kinds of solvents (DMSO and water) were used to dissolve the antifungal agent 5FC in the broth microdilution experiments, the results from each independent experiment has indicated that the usage of DMSO or water as solvents did not have obvious impact on the 5FC susceptibility to each strain. In fact, the strains that were resistant to 5FC in DMSO according to the MICs were still resistant to 5FC in water, and vice versa. Among the clinical isolate YM020291 and its genetically engineered strains, the results from E-test and broth microdilution have shown that only G145T single mutation strains and FCY2^R/FCY2^R homozygous mutants were resistant to 5FC.

Since they both exhibited a 145 T/T genotype, it implied the contribution of the SNP at position 145 in the FCY2 gene, resulting in a nonsense mutation, to 5FC susceptibility.

Additional evidence on the link between nonsense mutations in the FCY2 gene, coupled with LOH, and the 5FC resistance in C. tropicalis was demonstrated in 13 clinical isolate-derivative pairs based on the sequence analysis and their 5FC susceptibility testing results. Total 35 clinical isolates were divided into five groups according to the existence of nonsense mutations and the MICs for their derivatives.

The 13 previous selected clinical isolate-derivative pairs were separated into groups A, B and C, while the remaining pairs were in groups D and E (Tables 7 and 8). For groups A to C, their entire coding sequence of FCY2 gene as well as the flanking regions was confirmed by sequencing, which is different from groups D and E where only the 5’ end of the coding sequence was checked. In group B, the MICs for the derivatives were considered as resistance, but no noticeable nucleotide change that were known to mediate 5FC resistance was observed at the FCY2 locus as well as other three loci FCY1, FUR1 and URA3, which were analyzed previously. It indicated that there might be other mechanisms contributed to the 5FC resistance in C.

tropicalis. In group C, the derivatives were determined to be susceptible by broth microdilution. It might be caused by picking the colonies growing at the margin of the inhibition ellipse, which was, in fact, out of the inhibition zone. For strains that were classified in groups B and C, no LOH at the FCY2 locus was detected in the derivatives. In contrast, the linkage between genotype and phenotype was discovered among the rest seven clinical isolate-derivative pairs, which were classified as group

A. In this group, while the parental strains were susceptible to 5FC, their progeny were resistant to the drug, except YM060088 clinical isolate-derivative pair (one of the two progeny exhibited intermediate susceptibility). This intermediate phenotype might be contributed by other mechanisms involving in the 5FC toxicity. Sequence analysis has shown that the FCY2 locus was converted from the heterozygous state in parental strains to the homozygous state in their progeny. Most importantly, in six clinical isolate-derivative pairs, the SNP at position 145 was converted from the T/G genotype to the T/T genotype, while in the YM020715 clinical isolate-derivative pair, the SNP at position 201 was converted from the A/G genotype to A/A genotype. No matter which mutation (G145T or A201T) occurred in the FCY2 gene, they would result in a nonsense mutation. In conclusion, nonsense mutations in one of the FCY2 alleles, followed by the LOH, were responsible for the 5FC resistance observed in C.

tropicalis.

4.2.3 Isolates exhibiting 145 T/G genotype might represent a subgroup associated with the generation of 5FC resistant derivatives

The data from multilocus sequence typing (MLST) and pulse-field gel electrophoresis (PFGE) revealed the distribution of 145 T/G and 201 A/G genotype among various subgroups of the clinical isolates, which is shown in Table 10 (Chou et al., 2007). Total 97 clinical strains were divided into five groups according to their genotype and 5FC susceptibility phenotype. In brief, strains produced resistant progeny were put into groups 1, 2 and 3; in turn, each group represented the clinical isolates carrying heterozygous SNP (201 A/G), another heterozygous SNP (145 T/G), and homozygous SNPs (145 G/G and 201 G/G). Hyper-susceptible strains were classified as group 4 and the only 5FC resistant clinical isolate from the TSARY program was classified as group 5. In previous research, the clinical isolates were divided into six PFGE major types (I-VI). The strain YM020715 in group 1 belonged to type V, whereas the strains (YM020693, YM020274 and YM020291) in group 2 belonged to type III. It has suggested that strains between groups 1 and 2 were epidemiologically unrelated. In addition, while type II and IV were closely associated with the fluconazole susceptibility claimed by Chou et al. (2007), type III might relate to the 145 T/G phenotype, indirectly associated with the 5FC resistance. Similarly, three major clonal clusters (1-3) were identified in previous work (Chou et al., 2007).

Interestingly, clonal cluster 1 was corresponding to part of the group 2 because eight

strains with known DST number were included in this clonal cluster (4 in DST 134, 2 in DST 90, 1 in DST 153 and 1 in DST 155). Also, clonal cluster 1 was related to one major clonal cluster (clade 2) of isolates from the UK and the USA (Chou et al., 2007;

Tavanti et al., 2005), which implied the prevalence of the strains exhibiting 145 T/G genotype in the FCY2 gene. In addition, several PFGE subtypes were associated with the isolates in this cluster, indicated the existence of more than one clone possessed the heterozygous SNP at position 145 (T/G). One strain in group 3 belonged to clonal cluster 3 (DST 139). The reasons for the generation of 5FC resistant progeny in this group were supposed to be involved in other mechanisms. Two biggest populations of the clinical isolates were in DST 140 and DST 98, which belonged to clonal cluster 2.

They were composed of the clinical isolates that did not generate resistant progeny, and sequence analysis of five strains that belonged to DST 140 (YM060173, YM060509 and YM060828) or DST 98 (YM060547 and YM060647) has shown the homozygous SNPs (145 G/G and 201 G/G). Chen et al. declared the existence of another subgroup representing 5FC resistant clinical isolates, where the majority was DST 164. However, no clinical isolates with known DST in this study belonged to the subgroup (Chen et al., 2009). In conclusion, heterozygous SNP at position 145 (T/G) in FCY2 gene may be a common genotype associated with the generation of 5FC resistant progeny in C. tropicalis.