In 2014–2015, ticks (larvae, nymphs, and adults) were collected from vegetation from forests in Chaiyaphum Province, Thailand (16°16′25.2″N 101°29′02″E; 16°22′23.0″N 101°46′38.7″E; and 16°12′18.9″N 101°52′22.9″E). The areas being studied are evergreen forests dominated by plateaus at elevations of approximately 1,000 m. We placed the ticks into tubes that were held in a container with liquid nitrogen and stored them in a freezer after returning to the laboratory. Morphological identification of all tick stages was performed under a stereomicroscope (26, 27). Tick identification was performed using molecular methods on the ticks that were positive for either Coxiella or Rickettsia bacteria; previously published primers (16S+1/16S-1) for the amplification of mt 16S rDNA were used (28). There were seven pools of Haemaphysalis ticks at the immature stage, including nymphs (four pools: three pools of five and one pool of four) and larvae (three pools: one pool of six, one pool of seven, and one pool of 15). For the Amblyomma immature stage (nymphs), there were five pools of three nymphs.

Before DNA extraction, ticks were cleaned three times with 70% ethanol, 10% sodium hypochlorite, and sterile distilled water. DNA extraction was conducted using a DNeasy Blood and Tissues Kit (Qiagen) according to the manufacturer's protocol. The DNA products were stored at −20°C until use as templates for the PCR assay.

All extracted DNA samples were used as templates for PCR assays with specific bacterial primers for detecting the presence of Coxiella, Rickettsia, Francisella, and Borrelia. Coxiella-positive bacteria were identified through the 16S rRNA, groEL (60-kDa chaperone heat shock protein B), and rpoB genes (β subunit of bacterial RNA polymerase). Rickettsia species were screened by PCR, targeting the 17-kDa antigen, citrate synthase (gltA), outer membrane protein A (ompA), outer membrane protein B (ompB), and cell surface antigen (sca4) genes. PCR primer pairs for the detection of bacterial species, primer name, target genes, and size of the amplicons (bp) are shown in Table 1. The PCR product of the expected size from each corresponding primer pair (gene) was cloned and sequenced as a positive control. A sterile distilled water negative control was also included.

The positive PCR products were purified with a Nucleospin Gel and PCR Clean-up Kit (Düren, Germany) and sequenced. The DNA sequence alignment of representative positive samples generated by this study was created using the CLUSTALW program. The nucleotide sequences were analyzed and blasted with the National Center for Biotechnology Information BLASTn database. Afterward, phylogenetic analyses were carried out using the maximum likelihood (ML) (38) and neighbor-joining (NJ) (39) methods (evaluated by bootstrap analysis with 1,000 replicates) for Coxiella and Rickettsia spp., respectively.

A total of 94 ticks were collected and identified. The species, number, and life stage of the ticks are shown in Table 2. The ticks belonged to two genera, Haemaphysalis and Amblyomma. Of the adult ticks, 13 males and 12 females belonged to H. lagrangei. The remaining ticks belonged to A. testudinarium (two female ticks). In the immature stage, Haemaphysalis sp. (22 nymphs and 28 larvae) was the most commonly collected, followed by Amblyomma sp. (17 nymphs). The species of all ticks included in the phylogenetic trees were confirmed by molecular methods, and their sequences were submitted to GenBank with the accession numbers shown in Tables 3, 4.

Coxiella and Rickettsia bacteria were detected in two genera of ticks (Haemaphysalis and Amblyomma), whereas Francisella sp. and Borrelia sp. were not identified in this study (Table 2). Single infection with Coxiella was detected in 16 of 25 H. lagrangei adult ticks (six of 13 males and 10 of 12 females), and single infection with Rickettsia was detected in one of 25 H. lagrangei females. Adult A. testudinarium ticks were infected with Coxiella (two of two females) and Rickettsia (one of two females). Single infection with Coxiella was detected in three of three nymphs and four of seven pools of Haemaphysalis ticks (one pool of four nymphs, two pools of five nymphs, and one pool of 15 larvae). In addition, two of two nymphs and five of five pools of Amblyomma ticks were positive for Coxiella and two of two nymphs and four of five pools were positive for Rickettsia. However, infection with Rickettsia was not detected in the immature stage of Haemaphysalis. Moreover, coinfection with these two bacteria was detected in infected adults of H. lagrangei. In addition, adults of A. testudinarium were coinfected with Coxiella and Rickettsia. Coinfection with these two bacteria was also present in Amblyomma nymphal ticks.

A phylogenetic tree based on the partial sequences of 16S rRNA revealed that all Coxiella sequences detected in this study belonged to two endosymbiotic groups (Figure 1A). Coxiella sequences detected in H. lagrangei and Haemaphysalis obesa were in the first group. PK16, PK17, and PK146-150 clustered with CLB of HLSD3 found in H. lagrangei, whereas PK104-118 grouped with CLB of TPSD8 detected in H. obesa. These sequences were closely related to those of CLB in H. hystricis S002 (LT009432), H. longicornis 47 (AY342035), and H. longicornis A (AB001519). Coxiella sequences in A. testudinarium and Amblyomma sp. formed a monophyletic clade and clustered together with CLB found in Amblyomma sp. S027 and A. testudinarium AMTKK2.1 from Malaysia and Thailand, respectively (Figure 1A). BLAST analysis of the Coxiella 16S rRNA and groEL and rpoB gene sequences from Haemaphysalis and Amblyomma ticks is shown in Table 3.

Interestingly, the BLAST results of the groEL gene sequences from this work showed DNA sequence identities of <92% compared to the existing sequences. Phylogenetic trees based on the partial sequences of groEL revealed that CLB in this study were clustered into three groups (Figure 1B). All sequences of Coxiella spp. in Amblyomma (except PK48 and PK138-140) grouped together as sister clades to the CLB of I. ricinus. CLB in Haemaphysalis grouped as sister clades with CLB detected in H. wellingtoni TRG32 (MG874471). Interestingly, the groEL sequences of PK48 and PK138-140 clearly formed separate clades from the other Coxiella spp. in A. testudinarium obtained in the first group, although they still formed a sister clade with the Amblyomma genus from other geographical regions. CLB in Haemaphysalis (PK16, PK17, PK104-118, and PK146-150) from this study (Figure 1B) formed a distinct clade from those of CLB in Amblyomma. Importantly, the BLAST results of the rpoB gene sequences from this work exhibited DNA sequence identities <95% compared to the existing sequences for Haemaphysalis (PK16, PK17, PK146-150, and PK104-118). Phylogenetic trees based on the partial sequences of rpoB revealed that all CLB sequences detected in this study were also divided into two groups. The first group included Coxiella spp. detected in Amblyomma ticks that formed a monophyletic clade and grouped together with CLB in Amblyomma sp. S027 (LT174617) from Malaysia. The second group included Coxiella sequences found in Haemaphysalis that formed an independent clade and grouped with CLB in H. hystricis S002 (LT174612) from Malaysia (Figure 1C).

BLAST analysis of the Rickettsia 17-kDa antigen, gltA, ompA, ompB, and sca4 gene sequences from Amblyomma and Haemaphysalis ticks is shown in Table 4. The bands for ompA and ompB from PK16 were faint and could not be sequenced. Interestingly, the BLAST search results of all ompB genes studied herein showed <97% DNA sequence similarity compared to the existing ompB genes in Rickettsia spp. In addition, the BLAST search results of PK100-102 showed that the Rickettsia ompB gene sequences from A. testudinarium nymphal ticks had 96% sequence similarity to Rickettsia rhipicephali 3-7-female 6-CWPP (CP003342). Moreover, the BLAST search results of the sca4 gene of PK100-102 showed 97.3% identity to Candidatus Rickettsia thierseensis (MT424980) and Rickettsia fournieri AUS118 (KF666473). Phylogenetic trees based on the partial sequences of the 17-kDa, gltA, ompB, and sca4 genes from Rickettsia sp. are shown in Figures 2A–D. Phylogenetic trees based on the partial sequences of the 17-kDa (Figure 2A) and gltA (Figure 2B) genes indicated that the Rickettsia spp. detected in this study formed three groups. The first group (PK16) was clustered with Rickettsia sp. HOT2 and Rickettsia sp. 315 (17-kDa gene), and this group was within the Rickettsia massiliae subgroup of Rickettsia. Phylogenetically, the gltA PK16 sequence grouped with the Rickettsia raoultii clade (R. massiliae subgroup) and was closely related to the clades containing Rickettsia japonica and Rickettsia heilongjiangensis. The second group included PK100-102, which clustered together with Rickettsia sp. in A. testudinarium from Laos. The third group (PK48, PK51, PK67-69, PK81-83, PK138-140, and PK168) formed a sister clade with R. tamurae based on their 17-kDa and gltA gene sequences.

However, phylogenetic analysis based on the partial sequence of the ompA gene was not included in this study because the ompA amplicons might be the products of contaminated reactions. In addition, phylogenetic trees based on the partial sequences of the ompB gene revealed that Rickettsia spp. formed two groups: the first group (including PK100-102) formed a sister clade to R. raoultii, R. rhipicephali, and R. massiliae (R. massiliae subgroup), and the second group (including PK48, PK51, PK67-69, PK81-83, PK138-140, and PK168) grouped within the clade containing R. tamurae (Rickettsia helvetica subgroup) (Figure 2C). A phylogenetic tree of the sca4 gene (not including sequence PK138-140) showed the presence of two groups within SFG Rickettsia (Figure 2D), similar to the results for the ompB gene.