Mosquitoes provide a transmission route between possums and humans for Buruli ulcer in southeastern Australia

Abstract

Buruli ulcer, a chronic subcutaneous infection caused by Mycobacterium ulcerans, is increasing in prevalence in southeastern Australia. Possums are a local wildlife reservoir for M. ulcerans and, although mosquitoes have been implicated in transmission, it remains unclear how humans acquire infection. We conducted extensive field survey analyses of M. ulcerans prevalence among mosquitoes in the Mornington Peninsula region of southeastern Australia. PCR screening of trapped mosquitoes revealed a significant association between M. ulcerans and Aedes notoscriptus. Spatial scanning statistics revealed overlap between clusters of M. ulcerans-positive Ae. notoscriptus, M. ulcerans-positive possum excreta and Buruli ulcer cases, and metabarcoding analyses showed individual mosquitoes had fed on humans and possums. Bacterial genomic analysis confirmed shared single-nucleotide-polymorphism profiles for M. ulcerans detected in mosquitoes, possum excreta and humans. These findings indicate Ae. notoscriptus probably transmit M. ulcerans in southeastern Australia and highlight mosquito control as a Buruli ulcer prevention measure.

Fig. 1. Dominant mosquito species distribution across the Mornington Peninsula.

Map showing the proportional distribution of the two dominant mosquito species trapped during 2019 and 2020. The pie charts are an aggregation of the 180 different trap sites. Trap groups containing mosquitoes that were PCR positive for M. ulcerans are also indicated. The mesh-block statistical areas are also shown, with those in red containing at least one human Buruli ulcer case diagnosed in 2019–2020. Map boundaries from Australian Bureau of Statistics under a Creative Commons license CC BY 4.0.

Fig. 2. M. ulcerans genome from human Buruli ulcer cases compared with sequences recovered from possum excreta and mosquitoes on the Mornington Peninsula.

a, Summary of IS2404 qPCR screening of primary samples and the sequence-capture libraries pre- and post-enrichment. Note that possum excreta were enriched as barcoded sequence library pools so they share the same pre- and post-enrichment Ct values. b, Artemis coverage plots depicting sequence-capture reads mapped to the M. ulcerans JKD8409 chromosome from possum excreta samples and three qPCR-positive mosquitoes (labels are given as numbers of mapped reads; percentage of total chromosome bases mapped; number of core variable nucleotide positions (VNPs) covered). The grey horizontal bar above the chromosome map shows the sites of 117 core SNPs (black inverted triangles). c, Maximum likelihood phylogeny inferred from an alignment of 117 core genome SNPs using reads mapped to the JKD8049 reference chromosome, and with tips aligned with environmental sample origin or patient origin. The dataset includes reads from the 5 environmental samples listed in a with >21 VNPs, and a reference collection of 36 M. ulcerans genomes representing the genomic diversity of the M. ulcerans population in southeastern Australia. The shortest vertical branch length represents a single SNP difference, as per the scale bar. Map boundaries in c from Google Maps under a Creative Commons license CC BY 4.0.

Fig. 3. Mosquito blood-meal analysis.

Summary of cytb gene sequences from the 36 blood-fed mosquitoes to identify host blood-meal sources. A blue circle indicates positive for host blood source; the larger the circle, the more individual mosquitoes with an identical blood-meal profile. Red boxes indicate individual mosquitoes that have dual blood meals for both humans and ringtail possum sources.

Fig. 4. Spatial clustering of trapped mosquitoes, possum excreta and human Buruli ulcer cases on the Mornington Peninsula.

Map showing the clusters identified by the three separate SaTScan analyses: (1) trapped mosquitoes (177 traps screened for IS2404; collected November 2019 to March 2020; denoted M1 and M2), (2) M. ulcerans detected in possum excreta collected during the summer of 2019 (December 2018 to February 2019; P1–P4) using data from a previous study, and (3) notified human Buruli ulcer cases from the study area in the years 2019–2020 (H1–H6). The inset shows an instance where all three analyses had overlapping clusters in the suburb of Rye. Map boundaries from Australian Bureau of Statistics under a Creative Commons license CC BY 4.0.

Extended Data Fig. 1. Location of the Mornington Peninsula and the Bellarine Peninsula of Victoria, Australia.

Population density is represented on the map from low to high (white to red) based on the human population per square kilometre. The study area is shown within an elipse (inset). Map boundaries in a from Australian Bureau of Statistics under a Creative Commons license CC BY 4.0.

Extended Data Fig. 2. Validation of allele imputation approach to assess robustness to level of missing data and genome selection.

(a) Flow diagram of imputation method. (b) Boxplots showing the distribution of imputation accuracy across 100 replicates for varying numbers of M. ulcerans chromosome SNP sites masked as missing data for five randomly selected M. ulcerans clinical isolate genomes. The line inside each box represents the median accuracy and the annotated numbers above each box indicate the mean accuracy in percentage. The whiskers extend to 1.5× the interquartile range. Source data

Extended Data Fig. 3. Tanglegrams analysis showing impact of imputation on phylogenetic inference.

Depicted are tanglegrams (Dendroscope) showing phylogenies inferred with imputed SNP alignments containing each of the five, individual sequence enrichment and 36 clinical isolate M. ulcerans genome sequences compared to phylogenies of the individual sequence enrichment genomes and 36 clinical isolates without imputation. Yellow highlighted tips show the sequence enrichment genomes. Branch lengths transformed and presented as cladograms.

Extended Data Fig. 4. Mosquito bloodmeal analysis of all detected bloodmeal sources.

A blue dot indicates positive for host blood source; the larger the dot, the more individual mosquitoes with an identical bloodmeal profile. Red boxes indicate individual mosquitoes which had dual bloodmeals from both humans and ringtail possum sources.