Found throughout much of the world, cattle are the most common type of large domesticated ungulate. They comprise hundreds of breeds that are recognized worldwide. Previously described as two separate subspecies (Bos taurus subsp. taurus and B. taurus subsp. indicus), all cattle belong to the sub-tribe Bovina under tribe Bovini according to a very recent classification based on multiple autosomal introns [1]. Water buffaloes (Bubalus bubalis) resemble cattle in many characteristics but belong to Bubalina, another sub-tribe under Bovini [1]. These animals are agriculturally and economically important to humans as they are widely used for a variety of purposes, such as meat, dairy products, and draught power. However, a variety of infections in cattle, such as bovine spongiform encephalopathy, foot-and-mouth diseases, brucellosis, anthrax, infectious endocarditis, and many others, have caused either economical loss of these animals or big public health concerns as some of these infections are transmissible to humans and potentially can cause death.

The genus Bartonella contains B. bovis and many other species and/or subspecies. These fastidious Gram-negative bacteria infect and persist in mammalian erythrocytes and endothelial cells and are found in a wide range of wild and domesticated mammals, including rodents, insectivores, carnivores, ungulates, and others. A number of Bartonella species have been associated with human illnesses and are responsible for a growing spectrum of emerging diseases, including endocarditis [2-10]. Knowledge of the transmission of Bartonella bacteria between mammalian hosts is incomplete. However, hematophagous arthropods, such as fleas, flies, lice, mites, and ticks, have been found naturally infected and are frequently implicated in transmitting Bartonella species [11-16].

Infection with Bartonella spp. (mainly B. bovis) in cattle has been reported in a number of studies [17-20], with B. bovis originally described in cows from France [21]. For water buffaloes, there are no previous reports of infections with Bartonella spp. The prevalence of B. bovis in cattle is generally high but varies widely across studies from different countries, i.e., 50-89%, 70%, 36%, 24%, and 20%, in the USA, French Guyanna, France, Italy, and West Africa, respectively [18-23]. In a very recently study from Poland, the prevalence of B. bovis in cattle was much lower (6.8%) [24]. In addition to B. bovis, two more Bartonella spp. (B. chomelii and B. schenobuensis) are occasionally found in cattle [17,25]. Although Bartonella spp. do not usually cause clinical signs in cattle, B. bovis is associated with bovine endocarditis [26,27]. Understanding whether any specific genetic variations are responsible for the pathogenic outcome in animals and people is important and requires separate investigation.

In the present study, we first investigated the prevalence of Bartonella spp. in cattle from five countries across the world (Kenya, Thailand, Japan, Georgia, and Guatemala) and in water buffaloes from Thailand. Following that, we characterized the B. bovis strains obtained from the present study and from reference collections (France and USA), and then compared the genetic difference among these strains. We employed multi-locus sequence typing (MLST) that is based on comparison of nucleotide sequences derived from multiple loci. This approach has been applied to study genetic diversity of different agents, including Bartonella spp. [28-30]. It has been shown to provide high discriminatory power in epidemiological and genetic analysis of strain populations while retaining signatures of longer-term evolutionary relationships or clonal stability. This in turn can help to enhance our understanding of population structure of the bacteria and the relationships between sequence type and animal host. Our aims were: (1) to assess the apparent prevalence of bartonella culture-positive infections in cattle and water buffaloes from different countries; (2) to identify Bartonella spp. in positive samples using the citrate synthase gene (gltA); and (3) to develop a MLST scheme for B. bovis, and to validate this scheme against a diverse sample of strains representing each country in this study to allow comparison of their sequence diversity.

In this study, we investigated infections with Bartonella spp. in water buffaloes from Thailand and cattle from five countries across the world. We provided the first evidence that water buffaloes are also infected with Bartonella spp., with a prevalence of 7%. In cattle, bartonella prevalence varied widely between the studied countries and between different regions in the same country. In Japan, Kenya, and the Mestia district of Georgia, the cattle were free from the infection; in contrast, infection prevalence reached as high as 90% in the Marneuli district of Georgia, and a moderate prevalence was observed in other areas. It is of interest to note that the prevalence of infection in cattle varied widely across districts in Georgia, from apparent absence to nearly 100%. The observed variation in bartonella prevalence in cattle between different countries/regions can be the result of multiple factors. Since bartonellae are mainly vector-transmitted [11-16], we speculate that the distribution and abundance of specific arthropods play a major role in this matter. It may be that a heavy level of infestation is required for transmission. Cattle ticks (Rhipicephalus microplus) and biting flies (Diptera spp.) are implicated as potential vectors that may transmit bartonella between cattle [31,32]. Interestingly, low ectoparasite infestation on cattle in Japan was noticed (Maruyama, personal communication). Such a fact could provide a plausible explanation for the absence of bartonella infection observed in cattle from Japan. In addition to ectoparasites, environmental factors, such as geographic characters, landscape, etc., may also have influences on bartonella prevalence. Of the three districts in Georgia, Mestia district (with no bartonella infection in cattle) is located in a mountainous area in the northwest of the country, while the other two districts (with high bartonella infection in cattle) are located in lower-lying areas in the northeast and southeast.

The cultures obtained from the study belonged to three Bartonella spp. based on gltA sequences. The majority of strains (172/182), identified as B. bovis, were the most common in cattle from all investigated places. This is concordant with previously reported observations. The other two species, B. chomelii and B. schoenbuchensis, are not common, although both have been previously reported in cattle [17,25]. Interestingly, both of these species were only found in cattle from Georgia. This may suggest that the bartonella community associated with cattle in Georgia is more diverse comparing to other places. All nine B. chomelii isolates were obtained from cattle in Dusheti District located in northeastern Georgia. These observations may be associated with the composition of the local ectoparasite community. Future studies of cattle ectoparasites should test hypotheses about whether any particular arthropod species act as vectors for bartonella transmission between cattle. Finally, all isolates obtained from water buffaloes in Thailand also belong to B. bovis, indicating that water buffaloes can also serve as hosts for B. bovis.

Although B. bovis is known to be widely distributed in bovines, no study has demonstrated how strains from different geographic areas and hosts vary. As a potential pathogen for domestic animals, information about genetic diversity of B. bovis is important. In this study, we developed an MLST scheme for B. bovis based on nine loci to characterize B. bovis strains among geographically diverse populations. Resolving into 22 STs among the 28 strains from different geographical regions, our MLST data demonstrated B. bovis strains can be genetically different. Although some STs can be quite close, the same ST was never found in strains from different regions.

Multi-locus sequencing typing has been applied in a number of studies previously, in which the MLST was based on the comparison of multiple housekeeping gene sequences [28-30,33]. In this study, we included ITS. Although non-functional, sequence comparison of the ITS region is widely used in taxonomy and molecular phylogeny for its ease of amplification and high degree of variation, even between closely-related species. Our data showed either insertion or deletion has occurred in some of the tested strains, which likely are associated with host species. Such results suggested that ITS might be more powerful in differentiating genetic diversities among strain populations, compared to some regularly-used housekeeping genes.

Identification of three close but distinct lineages among the STs suggests a clonal population structure for the species. All STs of cattle-originated strains fall into either lineage I or II, while STs of water buffalo-originated strains exclusively fall in lineage III, showing the specific host relationship. Meanwhile, we suspect that lineage I and lineage II each is associated with a particular lineage of Bos taurus. By general understanding, cattle from US, Europe, and Georgia belong to the ‘taurine’ lineage (former B. taurus subsp. taurus); while those from Thailand belong to the ‘zebu’ lineage (former B. taurus subsp. indicus); and in Guatemala, cattle were of mixed breeds. Based on this information, we hypothesize that lineage I is associated with cattle of ‘taurine’ lineage and lineage II is associated with cattle of ‘zebu’ lineage. Further studies are required to confirm this hypothesis.

MLST analysis also showed B. bovis strains retain geographical particularity. This observation can be explained by the association of B. bovis with specific breed of cattle as well. For example, all cattle-originated strains from Thailand belong to lineage II; while all strains from Georgian cattle together with strains from the US and France, belong to lineage I. Contrastingly, strains from Guatemalan cattle fall into both lineages I and II. Mixed breed of cattle in Guatemala likely is the cause of crossing lineages.

Lineage I was the most common group of B. bovis, containing most STs (13/22) that are distributed in different regions across the world. The isolate that caused cattle endocarditis in the US (B37080, ST2) was also within this lineage. Although none of the other tested strains is of the same ST, all STs in this lineage are in fact very close. While clinical symptoms were not recorded for cattle in these studies, B. bovis strains have high potential as a pathogen, being widely distributed in cattle populations globally.