Human Exposure to Novel Bartonella Species From Contact With Fruit Bats

Ying Bai; Modupe O.V. Osinubi; Lynn Osikowicz; Clifton McKee; Neil M. Vora; Maria Rosales Rizzo; Sergio Recuenco; Lora Davis; Mike Niezgoda; Ajoke M. Ehimiyein; Grace S.N. Kia; Akin Oyemakinde; Olufunmilayo Sanni Adeniyi; Yemi H. Gbadegesin; Olugbon A. Saliman; Abiodun Ogunniyi; Albert B. Ogunkoya; Michael Y. Kosoy; Idanre Bat Festival Investigation Team


Emerging Infectious Diseases. 2018;24(12):2317-2323. 

In This Article


Bartonella in Egyptian Fruit Bats

We recovered Bartonella isolates from 22 of 177 Egyptian fruit bat blood clots, giving an overall prevalence of 12.4%. The gltA sequences of all Bartonella strains obtained from Egyptian fruit bats were identical or similar (>97% identity) to each other and represented 4 unique variants (GenBank accession nos. HM363764, MH069693–MH069695). A variant is defined when it differs by >1 nt from others. Together with Bartonella strains obtained from Egyptian fruit bats in Kenya,[25] these variants constitute a monophyletic genogroup that is distant from all other genotypes previously found in other bat species and any other described Bartonella species.

Multilocus sequence typing of the type strain (gltA; GenBank accession no. HM363764) with 7 additional genetic loci (ftsZ, nuoG, ribC, rpoB, ssrA, 16S rRNA, and ITS) further confirmed the uniqueness of this strain. Sequencing information for each genetic marker demonstrated that the Bartonella strain from the Egyptian fruit bats was distant from all other known Bartonella species and genotypes, including those reported from other bats from Africa. Sequences of all genetic loci obtained during the analyses were deposited in GenBank (accession nos. HM363769, KM387321, HM363779, HM363774, KM382247, HM363784, and KM382255). We compared the fragment sequences of each target with those from other Bartonella species/genotypes. The Egyptian fruit bat–associated Bartonella formed a separate genetic group that was distant from all other Bartonella species with >20% genetic distance and probably represents a novel Bartonella species, according to the definition of La Scola et al..[26] We proposed that this bacterial species be named Bartonella rousetti, to reflect the Egyptian fruit bat (Rousettus aegyptiacus) as the natural host. A phylogenetic tree based on the ITS locus illustrates the relationship of this proposed novel species to other Bartonella species (Figure).


Phylogenetic relationships of Bartonella rousetti (proposed name) obtained from Egyptian fruit bats (Rousettus aegyptiacus) collected in Nigeria, 2010 and 2013, and other Bartonella species and bat-associated Bartonella based on internal transcribed spacer sequences. The neighbor-joining method by the Kimura 2-parameter distance method and bootstrap calculation was conducted with 1,000 replicates for phylogenetic analysis. The internal transcribed spacer sequence obtained from the bat flies was closely clustered with B. rousetti. GenBank accession numbers are provided for the B. rousetti sequence and the comparison sequences.

Identification of bat Flies and Detection of Bartonella DNA

In 2013, we collected 51 ectoparasites from Egyptian fruit bats. With the exception of 1 unidentified mite, all arthropods were identified as the bat fly Eucampsipoda africana Theodor (Diptera: Nycteribiidae). The morphologic identification of every bat fly was confirmed by 1 or both mitochondrial markers (16S rRNA or COI). Representative 16S rRNA (accession nos. MH138030–MH138037) and COI (accession nos. MH151059–MH151066) sequences have been deposited in GenBank.

Of the 50 DNA extracts from bat flies, 21 (42%) produced >1 ITS or gltA sequence that was confirmed via BLAST ( as Bartonella. Positive samples yielded 19 ITS sequences and 18 gltA sequences; 16 samples yielded sequences for both loci and 5 samples yielded only 1 sequence. All but 1 of the 19 ITS sequences matched closely to the proposed B. rousetti (>95% sequence identity) (Figure); the remaining sequence was identical to B. tamiae (DQ395180). Of the 18 (66.7%) gltA sequences, 12 were close matches for B. rousetti (>98.3% sequence identity); all 12 of these samples also produced ITS sequences matching this strain.

The remaining 6 gltA sequences were identical to Bartonella sequences detected in a louse (Neohaematopinus sciuri) collected from a dead Eastern gray squirrel (Sciurus carolinensis) at a zoo in Greenville, SC, USA (GenBank accession no. EU368000) and an unidentified tick collected from a sheep in Peru (GenBank accession no. AF415209). These sequences were also closely (>99% sequence identity) related to other sequences from fleas (Ctenocephalides felis and C. canis) collected from dogs in Tunisia (GenBank accession nos. KP126468–74), a louse pool (Polyplax spp. and Hoplopleura spp.) collected from rodents in Thailand (GenBank accession no. KT324560), and an unidentified flea collected from a dog in Peru (GenBank accession no. GU583843). Of the specimens with this particular gltA sequence, 2 yielded no ITS sequence, 3 yielded ITS sequences matching B. rousetti, and 1 yielded the single B. tamiae sequence. Representative ITS (GenBank accession nos. MH14262–MH142639) and gltA (GenBank accession nos. MH151067–79) sequences for each novel sequence variant were submitted to GenBank.

Human Exposure to B. Rousetti

A total of 305 serum samples from 204 participants were tested for IgG against B. rousetti; 12 samples from different persons showed reactivity at an initial dilution of 1:32. Further 2-fold titration confirmed that 8 were positive, with titers >1:64 (Table). The positive samples were retested for 3 other Bartonella species—B. henselae, B. quintana, and B. elizabethae, all of which have been reported in Africa;[22–24] antibodies against these Bartonella species were not detected in any of the samples. Five seropositive participants reported having eaten bats and having either touched bats or been scratched or bitten by them, although not all reported having ever participated in the bat festival. Three seropositive participants reported never having eaten bats, touched bats, or been scratched or bitten by bats; in addition, these 3 participants claimed to have never participated in the bat festival. Of the 8 seropositive participants, only 1 reported having experienced a febrile illness since the bat festival that had taken place earlier in the year.