Insights into the origin, hybridisation and adaptation of Candida metapsilosis hybrid pathogens

Abstract

Hybridisation is a source of genetic diversity, can drive adaptation to new niches and has been found to be a frequent event in lineages harbouring pathogenic fungi. However, little is known about the genomic implications of hybridisation nor its impact on pathogenicity-related traits. A common limitation for addressing these questions is the narrow representativity of sequenced genomes, mostly corresponding to strains isolated from infected patients. The opportunistic human pathogen Candida metapsilosis is a hybrid that descends from the crossing between unknown parental lineages. Here, we sequenced the genomes of five new C. metapsilosis isolates, one representing the first African isolate for this species, and four environmental isolates from marine niches. Our comparative genomic analyses, including a total of 29 sequenced strains, shed light on the phylogenetic relationships between C. metapsilosis hybrid isolates and show that environmental strains are closely related to clinical ones and belong to different clades, suggesting multiple independent colonisations. Furthermore, we identify a new diverging clade likely emerging from the same hybridisation event that originated two other previously described hybrid clades. Lastly, we evaluate phenotypes relevant during infection such as drug susceptibility, thermotolerance or virulence. We identify low drug susceptibility phenotypes which we suggest might be driven by loss of heterozygosity events in key genes. We discover that thermotolerance is mainly clade-dependent and find a correlation with the faecal origin of some strains which highlights the adaptive potential of the fungus as commensal.

Fig 1. C. metapsilosis environmental strains are hybrids.

K-mer profile plots of the five new C. metapsilosis isolates showing the frequency of 27-mers in sequencing reads. The presence or absence of the k-mers in the reference genome (BP57) is shown in red and black, respectively.

Fig 2. Strains belonging to newly described subclade 1.3 show increased thermotolerance.

(A) Neighbour net splits tree network based on alignments harbouring genomic variants. (B) Blocks of LOH (green) and heterozygous regions (purple) across Scaffold 9 of representative C. metapsilosis strains from each clade highlight genomic differences between clades. (C) Inference of the Jaccard index computing pairwise comparisons between LOH blocks of all samples of C. metapsilosis strains supports the notion of subclade 1.3 being a distinct clade. Subclade 1.3 is highlighted with a green square. (D-E) Area under the growth curve of a set of representative C. metapsilosis strains grown at increasing temperatures ranging from 30 to 40°C (n = 4 biological replicates). Growth measurements were taken every 15 minutes during a period of 24 hours in solid YPD rich medium. The growth of each strain at each temperature is relative to the average growth of the four replicates of that strain at 30°C. In the left panel each line represents a strain. The right panel shows the average of all strains belonging to the same clade. Data are presented as mean values +/- SD.

Fig 3. C. metapsilosis clades differ in their tolerance to antifungal drugs.

(A) Area under the growth curve of a selected subset of C. metapsilosis strains in the presence of fluconazole (FLZ) or anidulafungin (ANI) at two different concentrations relative to their growth on rich media in the absence of the drug (n = 4 biological replicates). Marine isolates are marked by a green dot. The network splits tree on top shows phylogenetic relationships between all strains belonging to Subclade 1.1. Subgroups 1.1a (teal) and 1.1b (purple), both belonging to subclade 1.1, are shown separated by a dashed line. The centre line in the boxplot indicates the median value, the boxes contain the Q1 and Q3 quartiles (IQR). The whiskers extend up to 1.5 × IQR. (B) Graph showing IGV coverage tracks (top) and LOH blocks (bottom) of representative strains belonging to subclade 1.1 subgroup 1.1a and subgroup 1.1b. SNPs are shown as vertical lines on the tracks. Homozygous variants appear in a single-coloured line whereas heterozygous SNPs are represented by two-coloured vertical lines. Strains belonging to less-susceptible subgroup 1.1a have undergone LOH in gene PDR16 whereas strains from subgroup 1.1b remain heterozygous in that locus. (C) Graph showing IGV coverage tracks of representative strains belonging to clade 1.1, 1.2, 1.3 and 2. Fluconazole-tolerant tested strains are highlighted in a yellow box. All strains from subclade 1.2 have undergone LOH in gene ERG11. Tolerant strain MSK606 from subclade 1.3 has also undergone LOH in the ERG11 locus although favouring the alternative allele. Strains PL29 and 319–2, which show reduced susceptibility show partial LOH in the ERG11 locus.

Fig 4. C. metapsilosis pathogenic potential.

Survival curves of Galleria mellonella injected with 3×10⁶ cells per larva (n = 15) in two replicate experiments. Larvae survival was monitored daily for a period of 7 days. Marine strains (DMKU_13-A and INY121) are depicted with dashed lines whereas clinical strains (BP57, CBS2916 and MSK414) are depicted with a solid line. An avirulent strain of S. cerevisiae and PBS were used as negative controls. Difference between strains was assessed by the log rank (Mantel-Cox) test and was significant with a P < 0.005.