Flea-Associated Bacterial Communities across an Environmental Transect in a Plague-Endemic Region of Uganda

Abstract

The vast majority of human plague cases currently occur in sub-Saharan Africa. The primary route of transmission of Yersinia pestis, the causative agent of plague, is via flea bites. Non-pathogenic flea-associated bacteria may interact with Y. pestis within fleas and it is important to understand what factors govern flea-associated bacterial assemblages. Six species of fleas were collected from nine rodent species from ten Ugandan villages between October 2010 and March 2011. A total of 660,345 16S rRNA gene DNA sequences were used to characterize bacterial communities of 332 individual fleas. The DNA sequences were binned into 421 Operational Taxonomic Units (OTUs) based on 97% sequence similarity. We used beta diversity metrics to assess the effects of flea species, flea sex, rodent host species, site (i.e. village), collection date, elevation, mean annual precipitation, average monthly precipitation, and average monthly temperature on bacterial community structure. Flea species had the greatest effect on bacterial community structure with each flea species harboring unique bacterial lineages. The site (i.e. village), rodent host, flea sex, elevation, precipitation, and temperature also significantly affected bacterial community composition. Some bacterial lineages were widespread among flea species (e.g. Bartonella spp. and Wolbachia spp.), but each flea species also harbored unique bacterial lineages. Some of these lineages are not closely related to known bacterial diversity and likely represent newly discovered lineages of insect symbionts. Our finding that flea species has the greatest effect on bacterial community composition may help future investigations between Yersinia pestis and non-pathogenic flea-associated bacteria. Characterizing bacterial communities of fleas during a plague epizootic event in the future would be helpful.

Fig 1. Average relative abundances of bacterial phyla within flea species.

Proteobacteria were further divided based on Class: Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria. Ccc: Ctenophthalmus calceatus cabirus, Dl: Dinopsyllus lypusus, St: Stivalius torvus, Xb: Xenopsylla brasiliensis, Xc: Xenopsylla cheopis, Xn: Xenopsylla nubica. Xenopsylla nubica were not sexed.

Fig 2. Estimates of alpha diversity for flea-associated bacterial communities.

Alpha diversity was measured as the total number of observed OTUs detected in a subset of 1000 randomly chosen sequences from a sample (A) and as phylodiversity of bacteria within a sample (B). Diversity did not significantly differ between males and females of C. c. cabirus and sex was not determined for X. nubica. The number of observed species did not significantly differ in X. brasiliensis. In all other comparisons, male fleas harbored significantly more diversity based on student’s t-tests. Ccc: Ctenophthalmus calceatus cabirus, Dl: Dinopsyllus lypusus, St: Stivalius torvus, Xb: Xenopsylla brasiliensis, Xc: Xenopsylla cheopis, Xn: Xenopsylla nubica.

Fig 3. Principal coordinate analysis (PCoA) of flea-associated bacterial communities based on flea species.

PCoA was performed based on Bray-Curtis dissimilarities (A) and on weighted UniFrac distances (B). The percentage of variation explained by axes one and two are presented in parentheses. Green: Ctenophthalmus calceatus cabirus, Purple: Dinopsyllus lypusus, Yellow: Stivalius torvus, Blue: Xenopsylla brasiliensis, Red: Xenopsylla cheopis, Orange: Xenopsylla nubica.