Genomic epidemiology of Cryptococcus yeasts identifies adaptation to environmental niches underpinning infection across an African HIV/AIDS cohort

Abstract

Emerging infections caused by fungi have become a widely recognized global phenomenon and are causing an increasing burden of disease. Genomic techniques are providing new insights into the structure of fungal populations, revealing hitherto undescribed fine-scale adaptations to environments and hosts that govern their emergence as infections. Cryptococcal meningitis is a neglected tropical disease that is responsible for a large proportion of AIDS-related deaths across Africa; however, the ecological determinants that underlie a patient's risk of infection remain largely unexplored. Here, we use genome sequencing and ecological genomics to decipher the evolutionary ecology of the aetiological agents of cryptococcal meningitis, Cryptococcus neoformans and Cryptococcus gattii, across the central African country of Zambia. We show that the occurrence of these two pathogens is differentially associated with biotic (macroecological) and abiotic (physical) factors across two key African ecoregions, Central Miombo woodlands and Zambezi Mopane woodlands. We show that speciation of Cryptococcus has resulted in adaptation to occupy different ecological niches, with C. neoformans found to occupy Zambezi Mopane woodlands and C. gattii primarily recovered from Central Miombo woodlands. Genome sequencing shows that C. neoformans causes 95% of human infections in this region, of which over three-quarters belonged to the globalized lineage VNI. We show that VNI infections are largely associated with urbanized populations in Zambia. Conversely, the majority of C. neoformans isolates recovered in the environment belong to the genetically diverse African-endemic lineage VNB, and we show hitherto unmapped levels of genomic diversity within this lineage. Our results reveal the complex evolutionary ecology that underpins the reservoirs of infection for this, and likely other, deadly pathogenic fungi.

Keywords: Cryptococcus neoformans; ecological genetics; fungi; microbial ecology; niche modelling.

Figure 1

(a) Variation of the Cryptococcus‐associated fungal community structure across the two main Zambian ecoregions during the dry and rainy seasons among the Zambezi Woodlands (ZM; green) and Miombo Woodlands (MW; red). Pie charts represent the profile of fungal phylafor each season: Ascomycota (blue), Basidiomycota (orange), Chytridiomycota (purple), Glomeromycota (yellow), Zygomycota (coral), other (red) and unidentified fungi (brown). (b) PCoA plots of fungal diversity for MW (red) and ZM (green) illustrating the difference in mycobiome across each ecoregion. (c) Environmental niche modelling for the two sister species showing the predicted distribution of C. neoformans (Cn) (i) is biased to the ZM ecoregion, whereas C. gattii (Cg) (ii) is biased to the MW. Each dot represents a positive sample for either species. (d) LDA effect size taxonomic cladogram comparing fungal community categorized by ecoregion and season. Branch areas are shaded according to the highest ranked variety for that taxon, and the significantly discriminant taxon nodes are coloured in green (ZM) or red (MW). Nonsignificant taxon appears in yellow. Highly abundant and select taxa are as follows: d9, Tremellales; b1, Pezizomycetes; a2, Leotiomycetes; j, Pleosporales. The complete list of discriminate taxa and ranks are listed on Fig. S7 (Supporting information). [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 2

Phylogenetic relationship between environmental and clinical isolates of C. neoformans compared against the H99 reference detailing association with the known lineages of C. neoformans. (a) Clinical isolates appear in red, whereas environmental isolates are in green. Phylogenetic analysis were performed using maximum‐likelihood‐based inference (raxml) (Stamatakis 2006) using SNP data; all branches had bootstrap values of 100 with 1000 generations. The tree was rooted using TempEst. Isolate references include two letters, Z – Zambia and either clinical – c, or environmental – e; (b) Distribution of the environmental or clinical C. neoformans isolates collected and sequenced throughout Zambia. The phylogenetic representation here was generated within a Microreact Project https://microreact.org/project/S1lkajtY. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 3

FineStructure analysis of C. neoformans population structure in Zambia for clinical (red) and environmental (green) isolates. On the x‐axis, each genome is considered as a recipient, and on the y‐axis, the C. neoformans isolate is considered a donor of genomic region. The VNB and VNI populations are clearly separated with very limited sharing of genomic regions between the VNI and VNB populations. The highest amount of shared genome regions between isolates appears in purple and the lowest in yellow. [Colour figure can be viewed at wileyonlinelibrary.com]