Highly recombinant VGII Cryptococcus gattii population develops clonal outbreak clusters through both sexual macroevolution and asexual microevolution

Abstract

An outbreak of the fungal pathogen Cryptococcus gattii began in the Pacific Northwest (PNW) in the late 1990s. This outbreak consists of three clonal subpopulations: VGIIa/major, VGIIb/minor, and VGIIc/novel. Both VGIIa and VGIIc are unique to the PNW and exhibit increased virulence. In this study, we sequenced the genomes of isolates from these three groups, as well as global isolates, and analyzed a total of 53 isolates. We found that VGIIa/b/c populations show evidence of clonal expansion in the PNW. Whole-genome sequencing provided evidence that VGIIb originated in Australia, while VGIIa may have originated in South America, and these were likely independently introduced. Additionally, the VGIIa outbreak lineage may have arisen from a less virulent clade that contained a mutation in the MSH2 ortholog, but this appears to have reverted in the VGIIa outbreak strains, suggesting that a transient mutator phenotype may have contributed to adaptation and evolution of virulence in the PNW outbreak. PNW outbreak isolates share genomic islands, both between the clonal lineages and with global isolates, indicative of sexual recombination. This suggests that VGII C. gattii has undergone sexual reproduction, either bisexual or unisexual, in multiple locales contributing to the production of novel, virulent subtypes. We also found that the genomes of two basal VGII isolates from HIV(+) patients contain an introgression tract spanning three genes. Introgression substantially contributed to intra-VGII polymorphism and likely occurred through sexual reproduction with VGI. More broadly, these findings illustrate how both microevolution and sexual reproduction play central roles in the development of infectious outbreaks from avirulent or less virulent progenitors. Importance: Cryptococcus gattii is the causative agent responsible for ongoing infections in the Pacific Northwest of the United States and western Canada. The incidence of these infections increased dramatically in the 1990s and remains elevated. These infections are attributable to three clonal lineages of C. gattii, VGIIa, VGIIb, and VGIIc, with only VGIIa identified once previously in the Pacific Northwest prior to the start of the outbreak, albeit in a less virulent form. This study addresses the origin and emergence of this outbreak, using whole-genome sequencing and comparison of both outbreak and global isolates. We show that VGIIa arose mitotically from a less virulent clonal group, possibly via the action of a mutator phenotype, while VGIIb was likely introduced from Australia, and VGIIc appears to have emerged in the United States or in an undersampled locale via sexual reproduction. This work shows that multiple processes can contribute to the emergence of an outbreak.

FIG 1

Phylogenetic reconstruction shows evidence for clonal expansion in outbreak groups. Maximum likelihood phylogenies were constructed using a matrix of all available SNPs. Bootstrap values are based on 500 replicates. The scale bar indicates the number of substitutions per nucleotide position. (A) Phylogeny based on SNPs unique to VGIIa. ICB107 was also included as it differed at only one MLST marker in previous studies. (B) Phylogeny based on SNPs unique to VGIIb. 99/473-1 was also included as it was similar to VGIIb in previous studies. (C) Phylogeny based on SNPs unique to VGIIc.

FIG 2

VGIIa outbreak isolates are more closely related to each other than to historical isolates or South American VGII isolates. (A) Prediction of VGIIa population structure based on haplotype network inference. Only SNPs were used to eliminate alignment artifacts. All Pacific Northwest isolates are predicted to share a relatively recent last common ancestor. The South American isolate ICB107, while the closest among isolates included in this analysis, is more distantly related. (B) SNPs distinguishing ICB107 from the VGIIa isolate R265 plotted across the genome. SNPs are distributed across the genome, suggesting ICB107 was not separated from VGIIa by a recent meiotic event.

FIG 3

VGIIb outbreak isolates cluster more closely with Australian isolates than Caribbean and Florida isolates. Shown is a prediction of VGIIb population structure based on haplotype network inference. Pacific Northwest VGIIb isolates and Australian isolates cluster together, with the exception of the Pacific Northwest isolate B8554 sequenced by the CDC.

FIG 4

VGIIa-like low-virulence clade has a hypermutator mutation. (A) Read depth across an exon of CNBG_1161 from three VGIIa-like genomes, two VGIIa genomes, one VGIIb genome, and one VGIIc genome. All three VGIIa-like isolates have a single base deletion (denoted with a box) resulting in a frameshift mutation in the MSH2 gene. (B) Results from a conserved domain search for CNBG_1161 performed using the NCBI’s CD search program. CNBG_1161 has strong homology to the MutS or Msh2 domains involved in DNA damage repair.

FIG 5

VGII phylogenies are incongruent based on different loci. Shown is a pair of maximum likelihood phylogenies generated from the first 100 kb of supercontigs 6 and 7. Clonal clusters are shaded to facilitate comparison. Bootstrap values are based on 500 bootstraps. The scale bar indicates substitutions per nucleotide position. Crossing lines indicate lack of congruence.

FIG 6

Regions of reduced SNP density provide evidence for ancestral sexual reproduction. (A) SNP density is plotted relative to the R265 reference genome. Clonal groups are collapsed to the SNPs shared by all members of the group. Regions of unexpectedly low levels of polymorphism are boxed for emphasis. (B) Maximum likelihood phylogenies generated from regions of depressed SNP density demarcated in panel A. Bootstrap values are based on 500 bootstraps. The scale bar indicates the number of substitutions per nucleotide position. Outbreak groups are highlighted in larger font, as well as individual global isolates that contribute to the identity island.

FIG 7

Whole-genome allele compatibility test shows evidence for prolific recombination at the population level. (A) An example of a paired allele compatibility test from the VGII population. Alternative SNPs are depicted in red and the reference in white. Evidence for recombination is provided by any pairwise comparison of two loci in which strains are present where red-red, white-white, red-white, and white-red combinations are all found (AB, Ab, aB, and ab) satisfying the allele compatibility test, providing evidence for recombination. (B) One hundred random SNPs were selected from the VGII data set and collapsed into 46 unique allele patterns. The reference nucleotide is indicated by white and the variant by red. A pairwise comparison of all 46 unique loci is shown, with green shading indicating a positive result and evidence for recombination.

FIG 8

Genomic islands of high polymorphism on supercontig 13 are caused by two distinct VGII clades. (A) Linkage disequilibria (R²) among SNP loci on supercontig 13. Two regions located between positions 393 and 417 kb were in high linkage disequilibria. (B) Polymorphism among VGII isolates on supercontig 13. An island of high polymorphism colocalized with high linkage disequilibria between positions 393 and 417 kb. (C) The islands of high polymorphism were caused by the presence of two distinct groups of VGII isolates. Exclusion of isolates 2001/935-1 and IP96/1120-1 reduced the polymorphism within VGII to low levels (red area) between positions 393 and 408 kb compared to polymorphism among all VGII isolates (gray area). Exclusion of isolates NT3, NT7, NT8, RDH2, RDH7, and MMRL2647 reduced the polymorphism within VGII to low levels (blue area) between positions 414 and 417 kb compared to polymorphism among all VGII isolates (gray area). (D) Maximum likelihood phylogeny of VGII isolates based on all SNP on supercontig 13. Values indicate bootstrap support among 100 replicates. Isolates 2001/935-1 and IP96/1120-1 and isolates NT3, NT7, NT8, RDH2, RDH7, and MMRL2647 each grouped into a distinct clade of VGII.

FIG 9

Island of polymorphism on supercontig 13 shows evidence of introgression into VGII. (A) The region of high polymorphism between 393 and 408 kb on supercontig 13 contains three genes (CNBG_4871 to -4873). Variation in GC content in the region is shown in red. (B) Maximum likelihood phylogeny of C. gattii VGI, VGII, VGIII, and VGIV isolates. The tree was constructed based on 297 informative SNPs in the region of 393 to 408 kb on supercontig 13. VGII isolates 2001/935-1 and IP96/1120-1 most closely grouped with VGI. Values indicate bootstrap support among 100 replicates.

FIG 10

Aligned read depth on supercontigs of IP96/1120-1. Read depth was approximately 2-fold higher than the genome-wide mean coverage on supercontigs 6, 13, and 14. Variation in read depth on supercontigs 6 and 13 suggests partial duplications across the region of the chromosome mapping to these supercontigs. Read depth on supercontig 14 suggests duplication of the entire region represented by this supercontig.