Phenotypic analysis of the single mutation strains

The 5FC susceptibilities of the single mutation strains each carrying a point mutation G273T (G486T), T273G (T486G), G-69T (G145T), T-69G (T145G), C1518T (C1731T) or T1518C (T1731C) at the FCY2 locus were determined by broth microdilution method. The 5FC MICs of the clinical isolate YM020291, its homozygous resistant mutants (FCY2^R/ FCY2^R) and homozygous sensitive mutants (FCY2^S/FCY2^S) were also measured as a control to examine the consistency of their susceptibilities comparing to previous results. The 5FC was dissolved in DMSO in the first three experiments, but then the drug was dissolved in sterile distilled water in the

last four experiments; thus, the 5FC susceptibility testing results of each experiment under different conditions were separately given in Tables 4 and 5. The MICs of each strain including clinical isolate YM020291, its homozygous mutants and its single mutation strains were determined after 48 h of incubation at 35°C. According to CLSI guidelines, the MIC was defined as the lowest drug concentration that inhibited growth by 50% compared to the drug-free growth control. The 5FC susceptibility testing was conducted at least three times in 11 strains (i.e., YM020291, YLO417, YLO418, YLO415, YLO454, YLO456, YLO466, YLO467, YLO468, YLO470 and YLO472); thus, the MIC values for these strains from each experiment are summarized in Table 6. The MIC values were listed as follows: (1) For the clinical isolate YM020291 (FCY2^R/FCY2^S), its homozygous sensitive mutant YLO415 (FCY2^S/FCY2^S), as well as its single mutation strains YLO468 [T-69G (T145G)] and YLO472 [T1518C (T1731C)], the MICs were less or equal to 0.5 µg/ml; (2) For the single mutation strain YLO456 [T273G (T486G)], the MICs were less or equal to 1 µg/ml; (3) For the single mutation strains YLO454 [G273T (G486T)] and YLO470 [C1518T (C1731T)], the MICs were less or equal to 2 µg/ml; (4) For the homozygous resistant mutants YLO417 and YLO418 (FCY2^R/FCY2^R), the MICs were higher or equal to 32 µg/ml; (5) For the single mutation strains YLO466 and YLO467 [G-69T (G145T)], the MICs were higher or equal to 64 µg/ml. Susceptibility testing of one reference strain and two quality control strains were used in parallel. All MIC results were within the range given by CLSI except in strain ATCC22019, which gave an increased MIC to 5FC. Based on the criteria of CLSI (1997), yeast cells with MIC ≤ 4 µg/ml were considered susceptible, while those with MIC ≥ 32 µg/ml were considered resistant. Accordingly, the FCY2^R/FCY2^S heterozygous strains and the FCY2^S/FCY2^S homozygous mutants as well as the G273T (G486T), T273G (T486G), T-69G (T145G), C1518T (C1731T) and T1518C (T1731C) single mutation strains were susceptible to 5FC; on the contrary, the FCY2^R/FCY2^R homozygous mutants and the G-69T (G145T) single mutation strains were resistant to 5FC.

The susceptibility of the clinical isolate YM020291 and its mutant strains to 5FC was also assessed with E-test assay (Figure 15). The MICs were determined after 48 h of exposure to 5FC E-test strip and the values were read at the point where the ellipse intersects the strip. The MIC for homozygous sensitive mutant YLO415 (FCY2^S/FCY2^S) was 0.008 µg/ml; for parental strain YM020291 (FCY2^R/FCY2^S) and

the G273T (G486T) single mutation strain YLO454 was 0.064 µg/ml; for single mutation strains YLO456 [T273G (T486G)], YLO468 [T-69G (T145G)], YLO470 [C1518T (C1731T)] and YLO472 [T1518C (T1731C)] was 0.023 µg/ml. In contrast, the MIC for homozygous resistant mutants YLO417 and YLO418 (FCY2^R/FCY2^R), the G-69T (G145T) single mutation strains YLO466 and YLO467 was higher than 32 µg/ml. Based on the criteria of CLSI (1997), yeast cells with MIC ≤ 4 µg/ml were considered susceptible, while those with MIC ≥ 32 µg/ml were considered resistant.

As a result, the FCY2^R/FCY2^S heterozygous strains and FCY2^S/FCY2^S homozygous mutants as well as the G273T (G486T), T273G (T486G), T-69G (T145G), C1518T (C1731T) and T1518C (T1731C) single mutation strains were susceptible to 5FC, whereas the FCY2^R/FCY2^R homozygous mutants and the G-69T (G145T) single mutation strains were resistant to 5FC. Among the 5FC susceptible strains, the MICs for YLO415, YLO456, YLO468, YLO470 and YLO472 were less than that for the parental strain YM020291.

The results of MICs from broth microdilution and E-test indicated that a guanine to thymine substitution at the nucleotide -69 (145) in the FCY2^S genes has contributed to the 5FC resistance in C. tropicalis.

3.6 Contribution of the polymorphic nucleotide in the promoter region of the FCY2 gene

To evaluate whether 5FC resistance observed in the FCY2^R/FCY2^R homozygous strains and the G-69T (G145T) single mutation strains was caused by the differential mRNA levels of FCY2, the mRNA levels of this gene were examined in the clinical isolate YM020291, its single mutation strains YLO466 [G-69T (G145T)] and YLO468 [T-69G (T145G)], and its homozygous mutants YLO415 (FCY2^S/FCY2^S) and YLO417 (FCY2^R/FCY2^R) by real-time quantitative PCR. The effect of the drug on the mRNA levels of FCY2 gene was also examined. All strains were treated under two conditions (the absence or presence of 0.5 µg/ml of 5FC) at 30°C for an hour. The FCY2 mRNA levels of the clinical isolate YM020291, its homozygous mutants and its single mutation strains conducted in two independent experiments are shown in Figure 16. For the clinical isolate YM020291, the homozygous resistant strain YLO417, homozygous susceptible strain YLO415, the G-69T (G145T) single

mutation strain YLO466 and the T-69G (T145G) single mutation strain YLO468, the FCY2 mRNA levels were 1.00±0.00, 0.52±0.02, 1.53±0.07, 0.50±0.11 and 1.70±0.52 folds, respectively, when the strains were not treated with 5FC. For the same strains, the FCY2 mRNA levels were 1.36±0.05, 0.51±0.01, 2.27±0.12, 0.50±0.20 and 2.70±0.50 fold, respectively, when the yeasts were treated with 0.5 µg/ml of 5FC.The analysis showed that the FCY2 mRNA existed not only in the wild type strain YM020291 but also in its engineered mutants (YLO415, YLO417, YLO466 and YLO468), and the mRNA levels were in a similar pattern among different treatments (with or without 5FC). That is, a slight increase of the mRNA level in the presence of 0.5 µg/ml of 5FC compared to that in the absence of the drug for all strains except in YLO466. In addition, under both conditions (the treatment with or without 5FC), the highest- and the lowest-levels of FCY2 mRNA were observed in the homozygous sensitive mutant YLO415 and the homozygous resistant mutant YLO417, respectively, among the clinical isolate YM020291 and its homozygous mutants. Similarly, the highest- and the lowest-levels of FCY2 mRNA were observed in the single mutation strain YLO468 [T-69G (T145G)] and the single mutation strain YLO466 [G-69T (G145T)], respectively, among the clinical isolate YM020291 and its single mutation strains. The tendency of the expression data was similar between the homozygous mutants and the single mutation strains.

In this analysis, it seemed that the FCY2 mRNA levels might be influenced by a low concentration (0.5 µg/ml) of 5FC. Besides, the similar mRNA level profiles between the single mutation strains and the homozygous mutant strains implied that the SNP at position -69 (145) in the FCY2 gene has contributed to the differential mRNA levels observed in this study.

3.7 Nucleotide sequence analysis of the FCY2 gene

When a BLASTN search was executed to determine if the polymorphic nucleotide at position -69 (145) was located in a crucial regulatory site, it has revealed that the polymorphic site was located in the coding region of the FCY2 genes. Amino acid alignment of MYA-3404 Fcy2p from C. tropicalis database with the C. albicans and S. cerevisiae purine-cytosine permeases was performed accordingly. The alignment result is given in Figure 17. It has shown that the starting amino acid

methionine of MYA-3404 Fcy2p in the second row was corresponding to the 93rd amino acid of the S. cerevisiae Fcy2p in the fifth row and corresponding to the 77th amino acid of the C. albicans Fcy21p in the third row. As a consequence, the nucleotide sequences of both FCY2alleles in YM020291 were used to examine the FCY2 coding region by ORF Finder, available on the NCBI website; it showed that the largest frame identified in a 2258 bp DNA segment was 1530 bp long because the 5’ end of the newly identified ORF was extended up to 213 bp preceding the former predicted start codon (Figure 18).

Nucleotide analysis of the redefined FCY2 open reading frame has indicated the new locations of the three selected SNPs as follows: one was shifted from position -69 to position 145; another was from position 273 to position 486; the other was from position 1518 to position 1731 (at the 201 nucleotide downstream of the stop codon of the FCY2 gene) (Figure 18). Most importantly, the polymorphic nucleotide at position 145 in the FCY2^R gene was a guanine-to-thymine substitution, which resulted in a nonsense mutation. There can be no doubt that the truncation at Glu 49 could result in the loss of purine-cytosine permease, which led to the 5FC resistant to C. tropicalis. It also revealed a guanine-to-adenine substitution at the nucleotide position 201 in the FCY2 ORF of the strains MYA-3404, which also led to a nonsense mutation. Clearly, because of the inframe stop codon at Trp 67 in PCP, Met 72 was mistakenly recognized as the start codon in C. tropicalis database (Figure 18).

Other than MYA-3404, the nucleotide sequences of the FCY2 gene in three C.

tropicalis strains, ODL4-302, ODL3-231 and CBS94, were available on the NCBI database as well. Although they have both 145 G/G and 201 G/G genotypes, their deposit sequences were not completed comparing to the redefined FCY2 ORF. That is, for strains ODL4-302, ODL3-231 and CBS94, their DNA sequences of FCY2 separately began at nucleotides 136, 129 and 131, downstream of the new start codon.

3.8 Genotypic characterization of clinical isolates

Nucleotide sequence analysis of the FCY2 gene in 22 parental strains from 35 clinical isolate-derivative pairs as well as seven previously selected clinical isolate- derivative pairs was performed to determine whether the 145 G/T genotype in the FCY2 gene was related to the generation of 5FC resistant progeny observed among

the clinical isolates. The analysis of eight other hyper-susceptible clinical isolates was included as reference. Moreover, the nucleotide sequence of FCY1, FCY2, FUR1 and URA3 of the 5FC resistant clinical isolate was examined. The results of 145 and 201 genotypes in the parental strains from 35 clinical isolate-derivative pairs are summarized in Tables 7 and 8, and those in the additional reference strains as well as the 5FC resistant isolate are summarized in Table 9. For 30 flucytosine susceptible clinical isolates that generated 5FC resistant derivatives in their E-test inhibition ellipse, 23 of which carried the heterozygous SNP at position 145 (T/G), yet one of the strains YM020715 carried the heterozygous SNP at position 201 (A/G). On the other hand, for the eight hyper-susceptible strains (YM020367, YM020649, YM060302, YM060509, YM060547, YM060559, YM060647 and YM060828), none of them carried the nonsense mutation at position 145 or position 201. As to the 5FC resistant clinical isolate YM060607, no nucleotide change was detected in the FCY1 gene, and no noticeable nucleotide change that might associate with the 5FC resistance was identified in the FCY2 gene. In contrast, nucleotide sequence analysis revealed a heterozygous SNP (C/T) at the position 431 of the FUR1 gene, where the cytosine-to-thymine substitution resulted in a threonine (ACT) to isoleusine (ATT) change at amino acid 144 in UPRT. It also revealed a heterozygous SNP (A/G) at the position 775 of the URA3 gene, where the guanine-to-adenine substitution resulted in an alanine (GCC) to threonine (ACC) change at amino acid 259 in ODCase. In conclusion, for clinical isolates that could generate progeny in inhibition ellipse, 76.7% (23/30) carried the 145 T/G genotype in the FCY2 gene. It implied the 145 T/G genotype exhibited in the FCY2 gene of the clinical isolates was responsible for the generation of their drug resistant progeny.

3.9 Characterization of loss of heterozygosity events in clinical isolate-derivative pairs

Two clinical isolates (YM020291 and YM060800), and their derived strains (YM0202091-1, YM020291-2, YM060800-1 and YM060800-2) were included in this study. According to previous work, the clinical isolates and their derivatives were differed by at least one LOH event at the FCY2 locus located on the chromosome (corresponding to supercontig 2 on the C. tropicalis database). Therefore, SNP

mapping were conducted to determine the extent of LOH events on the chromosome containing FCY2 gene that differentiated the strains from each analyzed pair. The strategy to search for the boundary regions was described in Materials and methods.

In brief, regions containing SNPs in the parental strain YM020291 or YM060800 were identified by sequencing, where the heterozygous SNP was determined based on the presence of two coincident peaks. Those SNP markers were subsequently used to indicate the heterozygous or homozygous state in the corresponding area in their derivatives. Since the LOH was first observed in the FCY2 gene, the border was approached from the ORF toward both ends of the chromosome.

The SNP map of the chromosomes of two clinical isolates and their derivatives is shown in Figure 19. Location of the selected loci and the amount of identified SNPs as well as the status of each locus (heterozygosity or homozygosity) are detailed in Figure 20. Among 89 regions of the chromosome analyzed in the clinical isolates (YM020291 and YM060800), 64 were polymorphic (i.e., at least one heterozygous SNP present in one of the clinical isolates), and total 233 SNP markers were identified.

Other regions were excluded because of their poor sequence quality or the lack of polymorphic nucleotide. Comparison of the chromosome of the clinical isolate YM020291 and its derivatives revealed the following results: first, of the 34 polymorphic regions identified in YM020291, 23 consecutive regions located on the left end side were homozygous in YM020291-1, while the remaining 11 consecutive regions located on the right end side were heterozygous. Second, of 35 polymorphic regions identified in YM020291, 28 consecutive regions located in the left end side were homozygous in YM020291-2, while seven consecutive polymorphic regions located on the right end side were heterozygous. As a result, derivatives YM020291-1 and YM020291-2 separately differed from the isolate YM020291 by an LOH that encompassed approximately 2.19 and 2.25 Mb of the chromosome.

For clinical isolate YM060800, 21 consecutive regions on the left end side of the chromosome preceding the gene CTRG_01804, which located at the position 1565480 to 1572091, were devoid of the SNP markers, when compared to the relevant polymorphic sites found in YM020291. The nucleotide sequence close to the telomere of the chromosome (i.e., a 3080 bp DNA fragment corresponding to nucleotides from position 20 bp to 3100 bp) were analyzed in clinical isolates (YM020291 and YM060800) and both derivatives of YM020291 (YM020291-1 and YM020291-2).

Six SNP markers recognized in YM020291 have confirmed that homozygosity was

retained at the end of the chromosome in the strains YM020291-1, YM020291-2 and YM060800 (Figure 21). Accordingly, comparison of the chromosome of the clinical isolate YM060800 and its derivatives began at the region approximately 1.56 Mb from the telomere and revealed the results as follows: first, of the 32 polymorphic regions identified in YM060800, 31 consecutive polymorphic regions located in the left end side were homozygous in YM060800-1, while one region (including 43 SNP markers) located on the right end side were heterozygous. Second, of 30 polymorphic regions identified in YM060800, two heterozygous segments (12 consecutive regions were on the left, while 9 were on the right) were flanking a homozygous segment (including 9 consecutive regions) in YM060800-2. As a result, YM060800-1 differed from its parental strain YM060800 by an LOH that encompassed approximately 0.7 Mb of the chromosome. As to YM060800-2, the LOH event was restricted to an extent smaller than 181 kb. In conclusion, four of the five LOH events (including one observed in the clinical isolate YM060800, which might be a result of LOH events) covered a large region from 2.19 to 2.23 Mb in the chromosome containing FCY2 gene, while one was remained heterozygous of the chromosome except a 181 kb internal homozygous segment.

3.10 Characterization of the loss of heterozygosity boundaries

Six LOH borders were determined and further pinpointed to a region flanked by the adjacent homozygous and heterozygous SNPs. They included four at the right end side of the chromosome downstream of the FCY2 gene, and two preceding the FCY2 gene. Since these chromosome fragments probably underwent a recombination event that resulted in the observed LOH, a DNA segment overlapping the LOH boundary region was sequenced to investigate features that may be associated with the recombination events, such as the location of the boundary (at intragenic or intergenic region) and the frequency of SNP around that area. The boundary region was referred to the area flanked by adjacent heterozygous and homozygous SNPs. The SNP maps and the nucleotide sequences of the six LOH boundaries are depicted in Figures 22, 23, 24, 25, 26, 27 and 28. For the strain YM020291-1, its LOH boundary (within position 2194213 and 2194621 of supercontig 2) fell in the 5’ UTR of CTRG_02080, covering a 409 bp segment. Eight homozygous SNPs were found clustered in the 5’

end of this gene as well as its promoter region within 510 bp long (Figure 22). For the second strain YM020291-2, the homozygous SNP at position 2253701 was located in the promoter region of the gene CTRG_02104, whereas the heterozygous SNP at position 2255494 was in the coding sequence of the gene CTRG_02105. Since no SNP maker was present within a 1794 bp segment, this LOH boundary (within position 2253701 and 2255494) was across the intergenic region containing the gene CTRG_02104 (Figure 23). As to the third strain YM060800-1, its LOH boundary (within position 2300598 and 2300901) was located in the gene CTRG_02123, covering a 304 bp segment. 47 SNPs were identified in a 2500 bp segment of this gene (Figure 24). For the fourth strain YM060800-2, two LOH boundaries were either located in an intergenic region (between position 2028007 and 2028323; 317 bp) preceding the FCY2 gene, or between two genes CTRG_02086 and CTRG_02087 on the right end side of the chromosome downstream of the FCY2 gene (within position 2208445 and 2209266; 822 bp). Three SNP clusters were observed around the left border. More specifically, eight SNPs were in the gene CTRG_02009 within a 559 bp segment and five SNPs located within 680 bp region adjacent the LOH boundary, followed by 10 SNPs located in another gene CTRG_02010 within a 239 bp segment (the data does not show on the map). On the contrary, one cluster was observed on the right border, which was in the CTRG_02087 with five SNPs within a 471 bp segment (between position 2209266 and 2209736 of supercontig 2) (Figures 25 and 26). For the clinical isolate YM060800, the homozygous SNP at position 1561562 was located in the gene CTRG_01802, whereas the heterozygous SNP at position 1564848 was in the promoter region of the gene CTRG_01803. Since no SNP maker was present within a 3287 bp segment, this LOH boundary (within position 1561562 and 1564848 of supercontig 2) was across the intergenic region containing one gene CTRG_01803 (Figure 27). Furthermore, a BLASTN search has performed to search for recombination hot spots of the six LOH boundaries. However, no detectable hot spots were found (Figure 29). Taken together, for the six LOH boundaries reported in this study, two were located in intergenic promoter-containing intervals (for the border of YM020291-1 and for the left border of YM060800-2) and one existed in a gene with abundant SNPs (the border for YM060800-1). Three were narrowed down to a region between two genes including at least one of these intervals (i.e., 821 bp for the left border of YM060800-2, 1794 bp for YM020291-2 and 3287 bp for YM060800).

Chapter 4. Discussion

4.1 The redefinition of FCY2 open reading frame in C. tropicalis

Multiple sequence comparison revealed a new 5’ end to the FCY2 ORF. The amino acid sequences of both C. tropicalis Fcy2p (one was available in the database and the other was deduced from the FCY2^S allele in YM020291) were aligned with C.

lucitaniae and S. cerevisiae PCPs. These yeasts were chosen because their FCY2 coding sequence was available and the length of each ORF was comparable to their mRNA length (Chapeland-Leclerc et al., 2005; Schmidt et al., 1984; Weber et al., 1990). The result suggested that there was very likely a new 5’ end to the FCY2 ORF, which is 213 nt upstream of the previous one presented in MYA-3404 from the C.

tropicalis database.

Another evidence consistent with the existence of a new 5’ end was given by the comparison of the sequence context around both start codons in C. tropicalis FCY2 gene. According to the study in S. cerevisiae, a consensus sequence of 5’-(A/U)A(A/C) A(A/C)A(A/C)AUGUC(U/C)-3’ (the position of the start codon is underlined), was found around transcriptional start codons (Hamilton et al., 1987). Kozak and Liu et al.

have reported when two AUG codons were present in the 5’ proximal region, the translation of mRNA most frequently initiated at the AUG codon that located in the most optimal context (i.e., the sequence most resembling the consensus). In their works, the initiation of translation would rather choose a downstream AUG in an

have reported when two AUG codons were present in the 5’ proximal region, the translation of mRNA most frequently initiated at the AUG codon that located in the most optimal context (i.e., the sequence most resembling the consensus). In their works, the initiation of translation would rather choose a downstream AUG in an