Tracing Genetic Exchange and Biogeography of Cryptococcus neoformans var. grubii at the Global Population Level

Abstract

Cryptococcus neoformans var. grubii is the causative agent of cryptococcal meningitis, a significant source of mortality in immunocompromised individuals, typically human immunodeficiency virus/AIDS patients from developing countries. Despite the worldwide emergence of this ubiquitous infection, little is known about the global molecular epidemiology of this fungal pathogen. Here we sequence the genomes of 188 diverse isolates and characterize the major subdivisions, their relative diversity, and the level of genetic exchange between them. While most isolates of C. neoformans var. grubii belong to one of three major lineages (VNI, VNII, and VNB), some haploid isolates show hybrid ancestry including some that appear to have recently interbred, based on the detection of large blocks of each ancestry across each chromosome. Many isolates display evidence of aneuploidy, which was detected for all chromosomes. In diploid isolates of C. neoformans var. grubii (serotype AA) and of hybrids with C. neoformans var. neoformans (serotype AD) such aneuploidies have resulted in loss of heterozygosity, where a chromosomal region is represented by the genotype of only one parental isolate. Phylogenetic and population genomic analyses of isolates from Brazil reveal that the previously "African" VNB lineage occurs naturally in the South American environment. This suggests migration of the VNB lineage between Africa and South America prior to its diversification, supported by finding ancestral recombination events between isolates from different lineages and regions. The results provide evidence of substantial population structure, with all lineages showing multi-continental distributions; demonstrating the highly dispersive nature of this pathogen.

Keywords: Cryptococcus; genome sequence; hybridization; phylogeography; recombination; selection.

Figure 1

Phylogenetic analysis supports three major lineages of C. neoformans var. grubii. Using a set of 876,121 SNPs across the 159 nonhybrid isolates, a phylogenetic tree was inferred using RAxML. The tree was rooted with VNII as the outgroup (Hagen et al. 2015). The percentage of 1000 bootstrap isolates that support each node is shown for major nodes with at least 90% support. For each isolate, the geographic site of isolation is noted by colored boxes.

Figure 2

Ancestry characterization of three major groups highlights hybrid isolates. (A) The fraction of ancestry (k = 3) estimated by Structure is shown within a column for each isolate. (B) PCA separates the three major lineages, with the hybrid isolates showing a mix of VNB ancestry with either VNI or VNII.

Figure 3

Large blocks of ancestry suggest recent recombination between lineages. For each of the four isolates depicted (A, Bt125; B, Bt131; C, Ftc260-1, and D, CCTP51), the Structure-assigned ancestry for each site along each chromosome is depicted as a colored bar corresponding to VNI, VNII, and VNB ancestry. Locations of centromeres are marked with black bars.

Figure 4

Chromosome ancestry and ploidy variation of AD hybrids. For three AD hybrid isolates (RCT14, IFNR21, and CCTP50), the contribution and copy number of A (green) and D (blue) ancestry chromosomal regions was measured by aligning all sequence reads to a combined AD reference (A: H99, left, and D: JEC21, right). The copy number of each chromosome is depicted, with either full or partial chromosomal regions shown; see Figure S4 for detailed coverage plots for all AD hybrid isolates.

Figure 5

Lineage-specific gene clusters. Two large lineage-specific clusters were detected in the VNI genomes or VNII genomes; these are depicted using a representative genome from each lineage. (A) Insertion of CNAG_06649 to CNAG_06653 in H99 (blue, VNI). Syntenic genes in Bt85 (pink, VNB) and MW_RSA852 (green, VNII) are connected with gray bars. (B) Insertion of C358_04097 to C358_04102 in MW_RSA852.

Figure 6

VNB alleles in population subdivisions and across geography. (A) Phylogeny of VNB lineage showing major subdivisions (VNBI and VNBII) and inferred ancestral geography (South America or Africa, depicted as continent shapes). (B) Classification of all 445,193 private VNB alleles (present in at least one VNB isolate and no VNI or VNII isolates) by subdivisions and geography. Most VNB alleles are specific for each VNB subdivision and for the geographic subdivisions within each group. More alleles are shared between geographic locations in the same subdivision (VNBI or VNBII) than are shared within geographic locations across subdivisions.

Figure 7

Genome-sharing analysis of C. neoformans var. grubii using fineSTRUCTURE was performed on a SNP matrix using a representative of each clonal population within the VNI lineage. These genomes were reduced to a pairwise similarity matrix, which facilitates the identification of population structure based on haplotype sharing within regions of the genome. The x-axis represents the “donor” of genomic regions, while the y-axis represents the recipient of shared genomic regions. The scale bar represents the amount of genomic sharing, with black representing the largest amount of sharing of genetic material, and white representing the least amount of shared genetic material (no sharing). The geographic site of isolation is illustrated with colored boxes as in Figure 1, and lineages are also shown.

Figure 8

Genome-sharing analysis of the VNI lineage using fineSTRUCTURE on a SNP matrix of 111 genomes. The x-axis represents the donor of genomic regions, while the y-axis represents the recipient of shared genomic regions. The scale bar represents the amount of genomic sharing, with blue representing the largest amount of sharing of genetic material, and yellow representing the least amount of shared genetic material (no sharing).