|Home | About | Journals | Submit | Contact Us | Français|
Candida albicans is the most prevalent opportunistic fungal pathogen in the clinical setting, causing a wide spectrum of diseases ranging from superficial mucosal lesions to life-threatening deep-tissue infections. Recent studies provide strong evidence that C. albicans possesses an arsenal of genetic mechanisms promoting genome plasticity and that it uses these mechanisms under conditions of nutritional or antifungal drug stress. Two microarray-based methods, SNP (Single nucleotide polymorphism) and CGH (comparative genome hybridization) arrays, have been developed to study genome changes in C. albicans. However, array technologies can be relatively expensive and are not available to every laboratory. In addition, they often generate more data than needed to analyze specific genomic loci or regions. Here we have developed a set of SNP-RFLP (or PCR-RFLP) (restriction fragment length polymorphism) markers, two per chromosome arm, for C. albicans. These markers can be used to rapidly and accurately detect large-scale changes in the C. albicans genome including loss of heterozygosity (LOH) at single loci, across chromosome arms or across whole chromosomes. Furthermore, skewed SNP-RFLP allelic ratios are indicative of trisomy at heterozygous loci. While less comprehensive than array-based approaches, we propose SNP-RFLP as an inexpensive, rapid, reliable method to screen strains of interest for possible genome changes.
Candida albicans is the most prevalent opportunistic fungal pathogen in the clinical setting, causing a wide spectrum of diseases ranging from superficial mucosal lesions to life-threatening deep-tissue infections. Although host immunity is a major contributor to the development and progression of disease (Ashman et al., 2004; Tuite et al., 2004; Ashman, 2008) genome plasticity of the eukaryotic pathogen is an important factor in the adaptation of this fungus to its many niches in the warm-blooded host (Fradin et al., 2003; Hube, 2004; Forche et al., 2009). C. albicans is a diploid yeast and, although a parasexual cycle has been defined (Bennett and Johnson, 2003; Bennett and Johnson, 2005), no meiosis has been observed and the range of diversity observed in clinical strains is consistent with a predominantly clonal and asexual life cycle (Pujol et al., 1993; Graser et al., 1996; Forche et al., 1999). Thus, classical segregation analyses (such as those applied to Sacharomyces cerevisiae) are not used to study and detect genetic and genomic changes. The diploid assembly of the C. albicans SC5314 genome sequence (Jones et al., 2004; Van het Hoog et al., 2007) revealed a high level of natural heterozygosity, with approximately 55,000 single nucleotide polymorphisms (SNPs) (~3% of the 16Mb genome).
In many C. albicans isolates, segmental or whole chromosome aneuploidies arise in response to environmental stresses ((Janbon et al., 1998; Perepnikhatka et al., 1999; Selmecki et al., 2006; Rustchenko, 2007). For example, gain and loss of nutrient assimilation is associated with chromosomal alterations (Rustchenko et al., 1994), and the growth on specific carbon sources can result in loss of a specific chromosome (Janbon et al., 1998). A large proportion of C. albicans strains that acquire resistance to fluconazole, the most commonly used antifungal drug, form a segmental aneuploidy, isochromosome 5L (i(5L)), in which the left arm of Chr5 (i(5L) is duplicated. C. albicans that acquire i(5L) become highly resistant to the antifungal drug fluconazole due to extra copies of two genes on Chr5L: ERG11, which encodes the drug target and TAC1, a transcription factor that activates CDR1 and CDR2, ABC-transporters that lower the intracellular drug concentration (Selmecki et al., 2008). In vitro studies have detected loss of i(5L) (and reduction in FluR) when the drug pressure is relieved (Selmecki et al., 2006; Selmecki et al., 2008). Increased resistance to antifungal drugs can also arise by point mutations in ERG11 (White, 1997), TAC1 (Coste et al., 2004; Coste et al., 2006), or MRR1 (a transcriptional activator of MDR1, a major facilitator efflux pump) (Morschhauser et al., 2007; Dunkel et al., 2008). Once such mutations arise, loss of allelic variation via recombination causes increased drug resistance by duplicating the hyperactive allele (Coste et al., 2006; Holmes et al., 2006; Dunkel et al., 2008). These studies provide strong evidence that Candida albicans possesses an arsenal of genetic mechanisms promoting genome plasticity and that it uses these mechanisms under conditions of nutritional or antifungal drug stress.
Recently two microarray-based technologies have been developed to study genome changes in Candida albicans. Single nucleotide polymorphism (SNP) microarrays detect loss of heterozygosity for many loci across all eight chromosomes (Forche et al., 2004; Forche et al., 2005; Forche et al., 2008). Comparative genome hybridization (CGH) utilizes microarrays containing virtually all ORFs in the genome, to detect segmental and whole chromosome aneuploidies (Selmecki et al., 2005). These methods provide comprehensive large-scale analyses of the genome. However, CGH and SNP microarray technology can be relatively expensive and is not available to every laboratory. Furthermore, these approaches generate more data than needed to analyzed specific loci or defined regions in the genome, or for checking strains after laboratory manipulations, which can give rise to genome scale changes (Selmecki et al., 2005; Bouchonville et al, in preparation).
Here we have developed a set of SNP-RFLP (or PCR-RFLP) (restriction fragment length polymorphism) markers for C. albicans and demonstrate that they can be used to detect loss of heterozygosity (LOH) and aneuploidy for whole chromosomes and individual chromosome arms across the genome. SNP-RFLP analysis is based on restriction enzyme analysis of genomic regions that are heterozygous for restriction enzyme recognition sites (McEwen et al., 2000). While a limited number of SNP-RFLP markers has already been used successfully to analyze the genome status of strains after transformation (Legrand et al., 2008), for population genetics studies (Graser et al., 1996; Forche et al.; 1999, Xu et al., 1999), to analyze progeny after parasexual mating (Forche et al., 2008) or to analyze events on Chr5 (Wu et al., 2005; Coste et al., 2006), here we define a diagnostic set of 32 SNP-RFLP markers, with four markers per chromosome (usually two per chromosome arm). Each SNP includes a single polymorphic restriction enzyme site that yields digestion products that are readily distinguishable by size. We demonstrate that these markers can be used to rapidly and accurately detect large-scale changes in the Candida albicans genome including LOH at single loci, as well as across chromosome arms (concerted LOH of two markers on a single arm) or whole chromosomes (concerted LOH of all four markers for a given chromosome). Furthermore, skews in SNP-RFLP allelic ratios provide an initial indicator of trisomy at heterozygous loci. While less comprehensive than array-based approaches, we propose SNP-RFLP as an inexpensive, rapid, reliable and interpretable method to screen for chromosome changes in strains of interest.
Strains used in this study are listed in supplemental Table 1. Strains were stored in 50 % glycerol at −80 °C, and grown and maintained on YPAD medium (1 % yeast extract, 1 % peptone, 2 % glucose, 1.5 % agar). Genomic DNA was prepared as described previously (Selmecki et al., 2005)
The two original heterozygous Contig 19 sequences from each SNP marker used for SNP arrays (Forche et al., 2005) were analyzed using NEB webcutter software (http://tools.neb.com/NEBcutter2/index.php) to identify restriction enzymes (REs) that would digest at heterozygous loci and yield restriction fragments of different sizes that were easily separable on an agarose gel. For each chromosome, four SNP markers were selected, with the goal of two markers per chromosome arm, so that events involving long tracts of LOH on a single arm could be inferred. Additional criteria for SNP marker choice were that each RE should cut only once within the SNP marker and that a minimal number of REs should be used with exception of marker 1765/2519 on Chr3 in which there are two RE sites within the marker sequence (Table 2). The current set of markers requires the use of 13 different REs.
The diagnostic set of 32 markers was tested using strain SC5314. PCRs were performed in a final volume of 25 μl with 5 μl of e2TAK buffer (Mg2+ plus; TAKARA BIO INC., Japan), 2.5 mM each of dATP, dCTP, dGTP, and dTTP, 10 μM of each primer (Table 2), 0.25 μl of e2TAK polymerase and 1.0 μl of genomic DNA (30 ng/ul). PCR conditions were: initial denaturation at 98 °C for 3 min, 30 cycles each of denaturation at 98 °C for 10 sec, annealing at 55 °C for 10 sec, and extension at 72 °C for 1 min, and a final extension at 72 °C for 5 min. Amplification was verified by analysis of 3 μl of PCR product run in an 1 % agarose gel (Seakem LE agarose, LONZA, Rockland, ME; in 1x TBE). Gels were stained with ethidium bromide and photographed.
Each PCR product was digested at overnight with the relevant RE (see Table 2, Tables S2–9) in a total volume of 15 μl with 1 μl restriction enzyme, 10 × restriction buffer, distilled water and ~5 μl of the digested PCR product was run on a 3 % agarose gel (Seakem LE agarose, LONZA, Rockland, ME; 1.0 × TBE buffer) along with an undigested control PCR sample. Gels were stained with ethidium bromide and photographed. Genotypes were assigned based on banding patterns for each SNP marker.
The C. albicans genome is quite plastic, exhibiting a broad range of mitotic recombination events (Lephart et al., 2005; Lephart and Magee, 2006; Forche et al., 2008; Legrand et al., 2008)in addition to aneuploidies due to chromosome loss (Janbon et al., 1998) or acquisition of whole chromosomes or chromosome segments (Selmecki et al., 2005). To distinguish these different mechanisms, the minimal number of markers required is 4 per chromosome, with two markers per chromosome arm. A total of 152 SNP markers previously used on our SNP microarray ((Forche et al., 2005; Forche et al., 2008; Legrand et al., 2008), Figure 1) were screened for the presence of restriction enzyme sites that digest SNPs within the SNP marker sequence. These SNP markers are on average 200 bp in length and often include several SNPs (Forche et al., 1999; Forche et al., 2004; Forche et al., 2005). Of the 152 SNP markers analyzed, 134 (88 %) included one or more SNPs that were recognized by a restriction enzyme (RE) (Table 1, Tables S2–S8). Of these, 112 (74 %) included unique RE sites (those digested only once within the marker sequence) (Table 1).
We selected a set of 32 markers (8 chromosomes, 4 per chromosome) from the 112 SNP-RFLP markers with unique restriction enzymes recognition sites (Figure 1). PCR amplification of these markers, followed by digestion with the relevant RE results in 3 distinguishable fragments: one fragment from the uncut allele and two smaller fragments from the allele that carries the polymorphism recognized by the RE. Strain SC5314, which is heterozygous for all SNP markers, was tested for the correct restriction pattern (data not shown). Thirty one SNP markers gave the expected 3 restriction fragments (3 fragments; one fragment for uncut allele and 2 fragments for the cut allele), and restriction digest of marker 1765/2519 (Chr3) yielded the expected 4 restriction fragments (Figure 2).
In a previous study, haplotypes for each of the eight pairs of chromosomes for strain SC5314 were assigned based on SNP microarray analysis of strains with existing and induced aneuploidies (Legrand et al., 2008). For 9 out of 32 markers the SNPs analyzed by microarray and by SNP-RFLP were identical (Table 2) and thus the assignment of haplotype identity was straight-forward. For marker F12n4 on Chr1 no SNP-RFLP haplotype could be assigned (see note below Table 2). Based on the assumption that original trace sequences for strain SC5314 were approximately 400 bp in length, and based on the fact that array SNPs and RFLP SNPs within a specific marker were on average no further apart than approximately 80 bp, each SNP (nucleotide) analyzed by SNP-RFLP was matched up with the appropriate homolog ( a or b) containing the SNP (nucleotide) present on the SNP array (Legrand et al., 2008). Based on the restriction digest results (cut or uncut) haplotypes were assigned for the remaining 21 SNP-RFLP markers (Table 2).
By screening strains of interest, we asked if the SNP-RFLP markers would detect known genotype changes including LOH (the appearance of only a single uncut band or the two cut bands) at a single locus, LOH of a chromosome arm or LOH at all loci across an entire chromosome. In addition, we asked if trisomy of a chromosome arm or of an entire chromosome could be detected as a change in the relative ratio of the uncut band to the two cut bands (Legrand et al., 2008; Selmecki et al., 2008).
As a proof-of-principle, we performed SNP-RFLP analysis of strains known to have genotypic alterations when compared to their parental genotype profiles (Table S1, Figure 2). For example, strain YJB10019 (derived from SC5314) was known from a previous study (Forche et al., 2008) to have undergone changes on multiple chromosomes including a single LOH event on Chr1, a chromosome arm LOH event on Chr2, and a whole Chr LOH on ChrR. Along with strain SC5314, YJB10019 was subjected to SNP-RFLP analysis of markers from chromosomes R, 1, and 2. Figure 3 shows the results of the restriction digests. Strain SC5314 was heterozygous for all 4 markers on ChrR, Chr1, and Chr2, respectively (Figure 2). In strain YJB10019, a single LOH event on Chr1 was detected at the telomere-proximal 1799/2450 marker (Figure 2). As expected, the Chr2 arm event spanned both markers on Chr2L (2051/2483 and 1414/2481) (Figure 2). The most extensive LOH event was observed for ChrR, where all 4 SNP markers were homozygous. This indicates that LOH occurred across most parts or the entire chromosome. Whole chromosome LOH most likely arises via one or more chromosome non-disjunction events
Aneuploidy, especially trisomy is frequently observed in drug resistant C. albicans isolates (Selmecki et al., 2006; Selmecki et al., 2008). Array Comparative Genome Hybridization (aCGH) (Selmecki et al., 2005) is the most comprehensive method for detecting aneuploidies. However, aCGH can be expensive and requires microarray technology that may not be readily available to all labs. Because trisomy results in skewed allelic ratios (Legrand et al., 2008), we asked if it is possible to detect skewed allele copy number by SNP-RFLP. For this experiment we chose two sets of strains that had included chromosomes with segmental or whole chromosome aneuploidies.
We used a well documented set of strains derived, by transformation, from CAF-2 (diploid). These strains had become trisomic for Chr2 (CAI4-F2 and CAI4-F3) (Selmecki et al., 2005). The CAI4-F3 strain also became trisomic for Chr1 (Chen et al., 2004; Selmecki et al., 2005). SNP-RFLP analysis was carried out for 4 markers of Chr1 and Chr2 for the parental strain CAF-2 and for the two versions of CAI4. Strain CAF-2 was heterozygous for all 4 markers on Chr1 and Chr2 with relative band intensities appropriate for 1:1 amounts of the two different alleles (Figures 3D, E left images). In contrast, in strain CAI4-F3, allele ratio bias was evident for 4 markers on Chr2 (Fig. 3D right image) and for 1 marker on Chr1 (Fig. 3E right image). Furthermore, as expected, markers on Chr2, but not on Chr1 exhibited skewed allele ratios for strain CAI4-F2 (Figures 3D, E middle images). This demonstrates that SNP-RFLP analysis generates data consistent with the idea that these chromosomes are trisomic in the strain. When multiple markers on a single chromosome exhibit a skewed allelic ratio, it provides more confidence in the idea that the chromosome may be aneuploid. Of course, aCGH would be necessary to confirm this hypothesis.
We next compared strains FH1 and FH6 (White, 1997a; White, 1997b, Selmecki et al., 2008). From aCGH, it was known that FH6 carries i(5L), a segmental aneuploidy in which there are two extra copies of the left arm of chromosome 5, and SNP array analysis had found that markers on Chr5L were present in a 2:1 ratio (Selmecki et al., 2008). It is important to note that FH1 and FH6 are clinical isolates and likely possess a different SNP distribution across their genome. Thus, although we analyzed all 4 diagnostic SNP-RFLP markers from Ch5, only 2 of the Chr5 markers were heterozygous for both FH1 and FH6 (Figure 3F). Similarly, when markers are not heterozygous, it is difficult to infer changes in ploidy with high confidence. For example, while a skewed allelic ratio (2:1) for the first marker (10080A) on Chr5L can be detected for strain FH6 (Figure 3F), and this is consistent with trisomy of Chr5L detected in this strain (Selmecki et al., 2008), no skewed allelic ration (2:1) can be detected for strain FH6 for the second marker (SNF1–4) on Chr5L. Marker 2222A on Chr5R was homozygous in both strains and thus is not informative with regard to alterations in the genome (Fig. 3F).
In conclusion, we describe here a practical, inexpensive and simple approach to determining if chromosomes underwent loss of heterozygosity using a diagnostic set of SNP-RFLP markers to detect LOH events in the Candida albicans genome. This method provides the ability to distinguish short range events that generate LOH of a single SNP, LOH events that involve multiple SNPs on a chromosome arm as well as LOH of SNPs spanning entire chromosomes. Combined with our HapMap (Legrand et al., 2008), haplotype exchanges can also be detected. Furthermore, this SNP-RFLP approach can be used as a preliminary test for trisomy of chromosomes that remain heterozygous. Trisomy is more difficult to detect and thus the absence of a skewed ratio of RE fragment alleles is not definitive. However, the presence of a skewed ratio at multiple markers on a chromosome provides more confidence in the interpretation of a potential trisomic chromosome.
A major advantage is that this diagnostic set of SNP-RFLP markers provides a rapid and accurate method to detect genomic changes on all 8 chromosomes after strain manipulations such as transformation, which can lead to increased levels of aneuploidy and/or LOH (Bouchonville et al., in preparation). The use of 2 SNP markers per chromosome arm reveals distinct mechanisms of LOH. Importantly, this method relies upon simple techniques available in all molecular biology labs, enabling the generation of data that is easily compared between research groups. In this regard, it should be considered a simple, rapid and accessible alternative to SNP microarrays and MLST for testing strain integrity.
This work was supported by grant RO1-AI062427 to JB and AF from the National Institutes of Health/National Institute of Allergy and Infectious Diseases.
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
Supporting Information:Additional Supporting Information may be found in the online version of this article.