Detection of zoonotic bacterial pathogens in multiple hosts using a microbiome approach and molecular characterization of Anaplasma phagocytophilum

Abstract

The Mnisi community, a rural area, is nestled at the cusp of a human/livestock/wildlife interface in Bushbuckridge Municipality, Mpumalanga Province, South Africa. In this area, undifferentiated non-malarial acute febrile illness (AFI) is among the most common presenting signs in patients seeking healthcare at the community clinics. Recent research suggested that zoonotic pathogens either rodent-borne or tick-borne may be common aetiologies of febrile illness in the community. The study had shown that patients presenting with non-malarial AFI had prior exposure to Bartonella spp., spotted fever group Rickettsia, Coxiella burnetii and Leptospira spp. Low levels of West Nile and Sindbis, but no Rift Valley fever virus exposure were found. In a separate study, the molecular detection of a bacterium closely related to Anaplasma phagocytophilum in dog samples collected in the Mnisi community was also reported. The aim of this study was, therefore to investigate wild rodents, cattle and dogs as well as their associated ticks, as possible sources of zoonotic pathogen infection in the Mnisi community using a microbiome sequencing approach. We also screened AFI patient samples, rodents, dogs, cattle and ticks for the presence of A. phagocytophilum using a real-time PCR assay. The Anaplasma species detected were subsequently characterized using gene sequencing and phylogenetic analysis. The sample set consisted of 282 wild rodents trapped across three habitat types, 74 AFI patients, 56 domestic dogs, 100 cattle, 160 Rhipicephalus sanguineus ticks collected from dogs and 348 Amblyomma hebraeum ticks collected from cattle. Of these, the bacterial blood microbiome of a subset of samples was generated using circular consensus sequencing (CCS) performed on the Pacific Biosciences platform at Washington State University. The full sample set was then also screened for the specific presence of A. phagocytophilum using a Taqman real-time PCR assay, followed by the molecular characterization and phylogenetic analysis of A. phagocytophilum targeting the 16S rRNA, gltA, ankA and msp4 genes. The bacterial blood microbiome of 25 Mastomys rodent species collected from three habitat areas revealed a total of 65,060 bacterial sequences with 29% of the total sequence reads obtained corresponding to Bartonella grahamii, 23% to Bartonella sp. strain RF255YX and 12% of sequences to Bartonella spp. Overall, rodents from Hlalakahle (urban/periurban area) and Tlhavekisa (communal rangeland) had higher proportions of Bartonella species (~85%), while Gottenburg (urban/periurban area) and Manyeleti (protected area) had lower Bartonella spp. loads (~45%). Other organisms of zoonotic and veterinary significance detected included Ehrlichia sp. (~0.03%), C. burnetii (~0.02%), Anaplasma spp. (~0.5%), and Brucella spp. (~1%). Characterization of the blood microbiome of the dogs revealed 30,340 bacterial sequences, with 24% of the total sequences obtained from canine blood corresponding to Ehrlichia canis, 19.3% to A. platys, 14.8% to Anaplasma sp. ZAM dog, while 0.3% of sequences corresponded to A. phagocytophilum. The characterization of the blood microbiome from nine cattle revealed 34,559 bacterial sequences, of which 58% corresponded to A. marginale, 22.2% to Anaplasma sp. Mymensingh, 10.5% to Anaplasma spp. and 5.4% to Anaplasma sp. Dedessa. In addition, these species were detected in the following rates in cattle blood: Anaplasma sp. Hadesa: 2.7%, A. centrale: 1.4%, Bartonella spp.: 0.5%, A. platys: 0.2%, Anaplasma sp. Saso: 0.2% and A. phagocytophilum: 0.01%. Characterization of the bacterial microbiome from 24 pools of salivary glands and 23 pools of midgut tissues from the 348 A. hebraeum ticks produced a total of 86,308 bacterial sequence reads from the midgut pools as well as 84,420 sequences from the salivary gland pools. Of these, 81.7% of the bacterial sequences from the midgut pools and 83% of the sequences from the salivary gland pools were dominated by the zoonotic pathogen Rickettsia africae, the cause of African tick bite fever (ATBF). A further 6.8% of the sequences from the salivary gland pools and 6.9% of the sequences from the midgut pools corresponded to Rickettsia spp. Six percent of the total sequences from the salivary gland pools and 3.4% of the sequences from the midgut pools corresponded to E. ruminantium, while 1.4% of the sequence reads from the salivary gland pools and 1.2% of the reads from the midgut pools corresponded to Coxiella spp. symbionts. Characterization of the blood microbiome of nine AFI patients revealed 13,726 bacterial sequences. Of significance was the detection of R. africae from three AFI patients and Brucella melitensis from one AFI patient. DNA from 74 non-malarial acute febrile illness patients, 282 rodents, 100 cattle, 56 dogs, and 160 R. sanguineus ticks were screened using a quantitative real-time PCR designed to target the msp2 gene of A. phagocytophilum. However, the test was found to detect A. phagocytophilum and an Anaplasma sp. recently described from Zambian dogs (Anaplasma sp. ZAM dog). Sequencing of the 16S rRNA and gltA genes confirmed the presence of A. phagocytophilum DNA in humans, dogs and rodents; also highlighting its potential importance as a possible contributing cause of acute febrile illness in humans in this rural community in South Africa. Anaplasma sp. ZAM dog and Anaplasma platys were furthermore identified in dogs, while Candidatus Anaplasma boleense and Anaplasma sp. Mymensingh were identified in cattle. Anaplasma sp. ZAM dog was also identified in R. sanguineus ticks collected from dogs. Phylogenetic analyses grouped Anaplasma sp. ZAM dog into a distinct clade; providing sufficient divergence with the other Anaplasma species to warrant classification as a separate species. Until appropriate type-material can be deposited and the species can be formally described, we will refer to this novel organism as Anaplasma sp. SA dog (for Anaplasma sp. southern Africa dog). In conclusion, the detection of an array of zoonotic bacterial pathogens from wild rodents, domestic dogs, cattle and their associated ticks and humans in this study highlights their significance as possible contributing factors to non-malarial febrile illness in the Mnisi community area. We therefore recommend that healthcare practitioners in the community should consider them in the differential diagnosis of AFI.