Background

Bacteria of the order Rickettsiales are obligate intracellular parasites of eukaryotes. While some symbionts are known (for example, many Wolbachia species), most described species of Rickettsiales are best known as human pathogens that cause several diseases, including rickettsioses, anaplasmosis, and ehrichiosis [1]. Historically, rickettsial agents have been important causes of human morbidity and mortality, including R. prowazekii that caused several million deaths in the USSR [2], and it is estimated that Orientia tsutsugamushi is currently responsible for approximately one million cases of scrub typhus per year [2],[3]. The discovery of new pathogenic species or their associated diseases has attracted attention to the Rickettsiales as pathogens [4]-[8]. As their arthropod vectors often live at high densities and in close proximity to domestic animals and humans, Rickettsiales will continue to pose a risk for transmission to humans. Hence, the identification and characterization of novel Rickettsiales is of importance for both animal and human health.

The number of novel Rickettsiales associated with protists, arthropods, and mammals has increased rapidly through the application of molecular detection and phylogenetics [4]-[6],[9],[10]. Remarkably, analysis of the Trichoplax adhaerens genome also reveals novel species in the order Rickettsiales (for example, [11]). At present, this order contains three established families (Rickettsiaceae, Anaplasmataceae, and Holosporaceae) and one proposed family (Candidatus Midichloriaceae) [8],[11]-[14]. Additionally, some unclassified species warrant further attention to determine their phylogenetic and systematic positions [8],[11]. The intra- and inter-species genetic diversity and evolutionary relationships within genera of Rickettsiales bacteria have been characterized using 16S rRNA gene (rrs) sequences, especially in the case of those bacteria causing animal and human disease [5],[6],[9],[14],[15]. However, relatively little is known about their potential for cross-species transmission and emergence.

Compared with other zoonotic or vector-borne bacteria, Rickettsiales are associated with a more extremely diverse host range, including protists, hydra, annelids, arthropods, vertebrates, and even plants [5],[8],[15],[16]. While some Rickettsiales are specific to particular vectors and hosts [16],[17], others experience host-switching or regularly cycle between different hosts, typically a mammal (e.g. rodents, cattle and humans) and a blood-feeding arthropod (e.g. fleas, mites and ticks) [5],[16],[17]. However, the evolutionary associations between Rickettsiales and their hosts are not well understood [6],[16],[18],[19]. In particular, it is unclear whether Rickettsiales most often evolve by long-term bacteria-host co-divergence or cross-species transmission [20],[21]. As most emerging infectious diseases in humans are caused by spillover from animal hosts or vectors, a better understanding of the evolutionary relationships among Rickettsiales bacteria could provide important information on the likelihood of their emergence as agents of disease.

**njiang (one of five autonomous regions of China) is located in the northwestern part of China, and borders Russia, Mongolia, Kazakhstan, Kyrgyzstan, Tajikistan, Afghanistan, Pakistan and India (Additional file 1: Figure S1) and is one of the nation’s major grazing areas. Several important tick-borne diseases are endemic in **njiang [22]. The main aim of this study was to explore the diversity of Rickettsiales in **njiang, China, where their presence has only previously been shown by serological data [23],[24]. Accordingly, we screened ticks and identified bacteria by sequencing and analyzing three genes; rrs, citrate synthase (gltA), and heat shock protein (groEL). With these data in hand we explored key aspects of Rickettsiales biodiversity and evolution.

Results

Collection of ticks and detection of Rickettsiales bacterial DNA

In the spring of 2011, a total of 2062 adult ticks were collected from domestic animals (sheep and cattle) and grasslands in the border areas of the Bole and Tacheng regions of **njiang Uygur Autonomous Region, China (Additional file 1: Figure S1). The numbers, species, and geographic distributions of the adult ticks collected are shown in Table 1. After morphological examination and sequence analysis of mitochondrial 18S and 12S rDNA sequences as described previously [25], only Dermacentor nuttalli and Hyalomma asiaticum were found in **njiang.

Table 1 Detection of Rickettsiales bacteria from pooled tick samples
Full size table

A total of 388 tick pools (1862 ticks) were investigated in this study, 314 of which were from Bole and 74 from Tacheng. PCR was performed to detect Rickettsiales DNA based on rrs. PCR products of the expected size were amplified from 50 tick pools from Bole and 37 from Tacheng. Genetic analyses of these sequences indicated that all products belonged to Rickettisales (see below).

Genetic analysis of bacterial DNA sequences

The rrs, gltA, and groEL gene sequences amplified from the Rickettsiales DNA-positive tick-pool samples were sequenced (sequences are described in detail in Additional file 2: Table S1). Genetic analyses indicated that all sequences recovered from ticks from ** of morphological characters. Syst Biol. 2003, 52: 131-158." href="/article/10.1186/s12862-014-0167-2#ref-CR54" id="ref-link-section-d287326108e2303">54] methods implemented in the Mesquite package [55] to tentatively reconstruct the evolution of habitat among the Rickettsiales by treating “aquatic vs. terrestrial” habitats as discrete character states and map** their occurrence onto the phylogenies.

Analysis of co-divergence events

We tested the hypothesis of bacterial-host co-divergence using the ParaFit method [56] as implemented in the COPYCAT software package [57], which compares the patristic distance matrices derived from the bacteria and vector phylogenies. For this analysis we prepared three tick-only data sets including (i) tick-associated Rickettsia, (ii) Anaplasma, and (iii) Ehrlichia, as well as an overall data set including all Rickettsiales. The bacterial genetic distance matrices were derived from the rrs trees inferred by both BEAST and ML methods, while the vector genetic distance matrices were derived from the 18S rRNA gene trees generated using BEAST as described above. Significance testing was based of 9,999 randomizations of the association matrices. Additionally, to illustrate the association between bacteria (Additional file 6: Table S3, Additional file 8: Table S5) and their vectors, a tanglegram was generated by matching each bacterial species (or group) to their associated vectors using TreeMap 3.0 [58].

Additional files