Within-Host Genomic Diversity of Candida albicans in Healthy Carriers

Abstract

Genomic variations in Candida albicans, a major fungal pathogen of humans, have been observed upon exposure of this yeast to different stresses and experimental infections, possibly contributing to subsequent adaptation to these stress conditions. Yet, little is known about the extent of genomic diversity that is associated with commensalism, the predominant lifestyle of C. albicans in humans. In this study, we investigated the genetic diversity of C. albicans oral isolates recovered from healthy individuals, using multilocus sequencing typing (MLST) and whole genome sequencing. While MLST revealed occasional differences between isolates collected from a single individual, genome sequencing showed that they differed by numerous single nucleotide polymorphisms, mostly resulting from short-range loss-of-heterozygosity events. These differences were shown to have occurred upon human carriage of C. albicans rather than subsequent in vitro manipulation of the isolates. Thus, C. albicans intra-sample diversity appears common in healthy individuals, higher than that observed using MLST. We propose that diversifying lineages coexist in a single human individual, and this diversity can enable rapid adaptation under stress exposure. These results are crucial for the interpretation of longitudinal studies evaluating the evolution of the C. albicans genome.

Figure 1

Protocol of the C. albicans genomic diversity analysis. (a) Genomic diversity between “carrier isolates”. In this context, 3 different “carrier isolates” were selected on the “primo culture” of the single oral swabbing of the individual. The whole genome of the 3 isolated colonies was analysed in order to determine the population genomic diversity within single oral sample. In total, 3 different single oral samples were selected from three independent individuals (A, D and G). (b) Genomic variability between “clonal isolates.” In this context, 3 different clonal isolates were selected on the “sub culture” of the different strains. The whole genome of the 3 clonal isolates was analysed in order to determine the basal genomic variability between clones. In total, 3 different independent strains from our collection of C. albicans clinical strains were analysed (X, Y and Z). (Part of the illustration was adapted with permission from).

Figure 2

Representation of LOH events between genomes from the different carrier isolates selected from individual (Ind.) A, D and G. Panels a: Detection of large LOH event by chromosome. For each genome from isolates, heterozygous SNPs density was mapped on the 8 chromosomes (1 Kb sliding windows). Homozygous regions are indicated in light or white colour. Appearance of large LOH event (MR/BIR) between genomes is indicated by a red square. The blue vertical line indicates the centromere of each chromosome. Panels b: Density of LOH events by chromosome. For each pair of genome comparisons the starting location of all LOH events was mapped on the 8 chromosomes. For each pair-wise comparison LOH events were screened in a symetric manner. (vs = versus, Chr = chromosome).

Figure 3

LOH characterisation (a). Definition of LOH event and size determination. Example of one short-range LOH event observed on chromosome 3 between the genome from carrier isolates A1 and A2. The first part represents the heatmap of the heterozygous SNPs density for the chromosome 3 of the 2 genomes (1 Kb sliding windows). The second part represents the associated diploid genome sequences. In this example the LOH MinS is 10 bps. (b) Distribution of MinS LOH event from the 9 pair-wise carrier isolates comparisons from individual sample A, D and G. The x-axis corresponds to the classes of LOH size (MinS in bp). The y-axis corresponds to the number of events observed by MinS classes.

Figure 4

Comparison of MinS LOH event distribution between carrier isolates and Clonal isolates. The majority of LOH events (95% of the total number) identified between carrier isolates had MinS ≤ 3000 bp (blue arrow) while for clonal isolates they had a MinS ≤ 300 bp (red arrow).

Figure 5

Distribution of the LOH event by MinS classes and the number of SNPs differences involved by event for the 9 pair-wise carrier isolates comparisons from individual sample A, D and G. The x-axis corresponds to the classes of LOH size (MinS in bp). The y-axis is in log scale and corresponds to the number of SNPs differences involved by LOH event. (vs = versus).

Figure 6

Statistical analysis of the genomic diversity between carrier and clonal isolates. (a) Comparison of the number of LOH events observed from genomes of clonal and carrier isolates. The number of LOH events observed between the genomes from carrier isolates was significantly higher than thus observed between the genome from clonal isolates (One way ANOVA test, p = 0.0013). (b) Comparison of the number of SNPs between genomes from clonal and carrier isolates. The number of SNPs detected between the genomes from carrier isolates was significantly higher than thus observed between the genome from clonal isolates (One way ANOVA test, p = 0.007). (c)Mutation frequency comparisons for carrier and clonal isolates by genomic regions. The mutation frequencies were represented for 10.000 bps of the different genomic regions. Repeat regions were significantly more mutated than other features whatever group considered (carrier or clonal isolates) (**p < 0.01; Post-hoc test: Tukey HSD). Carrier isolates were significantly more variable in intergenic regions than clonal isolates (**p < 0.01, Post-hoc test: Tukey HSD).

Figure 7

Comparison of the allelic ratio distribution of all heterozygous polymorphic positions. (a) Comparison from carrier isolates. (b) Comparison from clonal isolates. The allelic ratio (ABHet = number of reads for reference allele / total number of reads) was determined with GATK tools and histograms were built based on the number of SNPs with AbHet values in a given interval (bin = 0.02). For each heterozygous SNP differences between pair of genomes the ABHet ratio was plot on the histogram. Red color represents the ABHet ratios from the positions located in the repeated regions, retrotransposons and LTRs and blue color the positions located in the other regions of the genome. For carrier isolates (a), we observed a bimodal distribution with the majority achieving a Gaussian distribution centred on a ABHet value of 0.5, which is the expected value for a diploid genome, and a minority achieving a Gaussian distribution centred on a ABHet value of 0.85, which is the limit cut-off for the heterozygous SNPs definition in our pipeline. This last distribution (around 0.85) was the only one found for clonal isolates (b).