New Genotypes of Orientia tsutsugamushi Isolated from Humans in Eastern Taiwan

Scrub typhus, an acute febrile illness, is caused by the obligate intracellular bacterium Orientia tsutsugamushi. In our study, O. tsutsugamushi was rapidly detected and typed by polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analysis of the 56-kDa type-specific antigen (TSA) gene. To investigate the genotypes of clinical variants of O. tsutsugamushi, we collected 3223 blood samples from eastern Taiwanese patients with suspected scrub typhus from 2002 to 2008. In total, 505 samples were found to be positive for scrub typhus infection by PCR, and bacteria were isolated from 282 of them. Four prototype genotype strains (Karp, Kato, Kawasaki and Gilliam) and eleven different Taiwanese genotype isolates (Taiwan-A, -B, -C, -D, -E, -G, -H, -J, -N, -O and -P) were identified by RPLF analysis. Taiwan-H, the major genotype in eastern Taiwan, exhibited prevalence and isolation rates of 47.3% (239/505) and 42.6% (120/282), respectively. We also assessed the genetic relatedness of the 56-kDa TSA gene among eight Taiwan-H isolates, thirteen other Taiwanese isolates and 104 DNA sequences deposited in the GenBank database using MEGA version 5.0 and PHYLIP version 3.66. We found that the Taiwan-H isolates formed into a new cluster, which was designated the Taiwan Gilliam-variant (TG-v) cluster to distinguish it from the Japanese Gilliam-variant (JG-v) cluster. According to Simplot analysis, TG-v is a new recombinant strain among Gilliam, Ikeda and Kato. Moreover, the Gilliam-Kawasaki cluster had the highest percentage of RFLP cases and was the most frequently isolated type in eastern Taiwan (50.1%, 253/505; 44.0%, 124/282). These findings shed light on the genetic evolution of O. tsutsugamushi into different strains and may be useful in vaccine development and epidemic disease control in the future.


Introduction
Scrub typhus (tsutsugamushi disease), an acute febrile illness, is caused by the obligate intracellular bacterium Orientia tsutsugamushi, which is transmitted by the bite of the trombiculid mite (or its larva, the chigger) [1]. The disease is a public health concern with approximately one billion people are at a risk of infection by the bacterium in the geographical triangle ( Figure 1) extending from northern Japan in the east to Pakistan and Afghanistan in the west and northern Australia in the south [2]. Scrub typhus is a reportable infectious disease in Taiwan [3], and thousands of patients are suspected of scrub typhus infection annually. The related syndromes include meningitis, eschar, disseminated intravascular coagulation and multiorgan failure [3]. The fatality rate can reach to 50% in the absence of suitable treatment; however, scrub typhus can be treated effectively with antibiotics [4].
No effective vaccine is currently available for the prevention of scrub typhus [5,6,7], and vaccine development has been impeded by the lack of availability of a determinant sequence or a systematic study of O. tsutsugamushi serotypes and genotypes, particularly clinical variants. Six antigenic variants of O.
Immunofluorescent antibodies have been the traditional method to analyze the variation in the 56-kDa TSA gene of O. tsutsugamushi in Taiwan [12,15,16]. Some studies have used the polymerase chain reaction (PCR) followed by the restriction fragment length polymorphism (RFLP) analysis to further characterize nucleotide sequence variation in the gene [12,15,17,21]. Six different Taiwanese genotypes, Taiwan-A, -B, -C, -D, -E and -F, were characterized by PCR-RFLP analysis of the variant domain I (VD-I) of the 56-kDa TSA gene at the Taiwan Centers for Disease Control (CDC) in 1999 [21]. Eastern Taiwan has the highest prevalence of scrub typhus in Taiwan. From 2002 to 2008, 3223 blood samples from patients with suspected scrub typhus were collected in eastern Taiwan to investigate the genotypes of clinical variants of O. tsutsugamushi. Nested PCR, performed by the Taiwan CDC-contracted laboratory at Tzu Chi General Hospital, was used to identify patients with scrub typhus infection. Scrub typhus infections were confirmed in 505 samples by PCR, and bacteria were isolated from 282 samples. Furthermore, the pathogen genotypes were characterized by RFLP. The 56-kDa TSA gene sequences of representative strains were examined with the aim of determining the genetic relatedness of the isolates in relation to other isolates in Taiwan or in other countries within the previously described geographical triangle [3,6,12,14,20,21].

PCR detection and isolation of O. tsutsugamushi from humans in eastern Taiwan
From 2002 to 2008, samples were collected from a total of 3223 blood eastern Taiwanese patients and subjected to PCR in this study. PCR was repeated several times with different primer pairs for the detection of O. tsutsugamushi infection [21,22]. Primers tsu-34 and tsu-11 were used for the first PCR and primers tsu-10 and tsu-11 were used for the semi-nested PCR ( Table 1). The molecular weights of the PCR products were 870 bp and 490 bp, respectively [22]. Of the blood samples, 505 were found to be PCR positive, and we proceeded with the PCR-RFLP assay for genotyping of the 56-kDa TSA gene. The molecular weights of the PCR products and RFLP genotypes are presented in Table 2 [21]. Bacteria were successfully isolated from 282 samples, and the RFLP genotypes of the isolated O. tsutsugamushi isolates are presented in Table 2. Karp, Taiwan-A and Taiwan-H were the three most commonly detected genotypes and isolates in our study, and Taiwan-H isolates were the most frequently isolates in eastern Taiwan (42.6%, 120/282).

Genotypes of O. tsutsugamushi strains
The results of RFLP analyses of the 56-kDa TSA gene PCR products from the 505 PCR-positive and 282 isolated samples are presented in Table 2. Previously, Tamura et al. described six prototype genotype strains: Karp, Kato, Kuroki, Kawasaki, Gilliam and Shimokoshi [3,10,11,12,13,17]. Six different Taiwanese genotyping isolates were also characterized by RFLP analysis by Taiwan CDC in 1999: Taiwan-A, -B, -C, -D, -E and -F [21]. In the present study, we analysed the genetic diversity of the 56-kDa TSA gene among the different cultured O. tsutsugamushi isolates by RFLP analyses. We detected six new RFLP genotypes among the clinical isolates cultured in eastern Taiwan: Taiwan-G, -H, -J, -N, -O and -P. However, we did not detect the three known RFLP types: Kuroki, Shimokoshi and Taiwan-F.
The molecular weights of the PCR fragments amplified from the prototype strains and clinical isolates cultured in eastern Taiwan using the primer pair tsu-a and tsu-b ranged from 140 bp to 210 bp ( Table 2). The differential molecular weights of the PCR products amplified from the new Taiwanese isolates using the primers tsu-a and tsu-b were as follows: 140 bp (Taiwan-G),  Phylogenetic analysis of O. tsutsugamushi The molecular weights of PCR products amplified using primers ST-A and ST-D for the entire 56-kDa TSA gene of O. tsutsugamushi are almost 1,600 bp. The twenty-one nucleotide sequences corresponding to the 56-kDa TSA gene of the cultured O. tsutsugamushi isolates in this study were consecutively processed and submitted to GenBank, and their accession numbers are provided in Table 3. The genotypes of the twenty-one isolates included Taitung-1 to Taitung-7 and Hualien-1 to Hualien-14 were analyzed by the PCR-RFLP genotyping method. Taitung-1, -3 and -6 were classified as the Taiwan-A RFLP genotype; Hualien-5 and -6 were classified as the Taiwan-G RFLP genotype and Hualien-2, -7, -8, -9, -10 and -13 and Taitung-2 and -5 were classified as the Taiwan-H RFLP genotype. Hualien-1, -3, -4, -11, -12 and -14 and Taitung-4 and -7 were classified as the Taiwan-C, Taiwan-J, Taiwan-B, Taiwan-O, Karp, Taiwan-P, Taiwan-N and Taiwan-E RFLP genotypes, respectively (Table 3).
Using the MEGA version 5.0 and PHYLIP version 3.66 software, we further analysed the genetic relatedness of the 56-kDa TSA gene of the twenty-one isolates used in this study as well as of 104 DNA sequences previously deposited in GenBank (Table 3) [3,4,5,6,8,14,21,23,24]. The results are shown in Figure 2. The phylogenetic tree of the 56-kDa TSA gene sequences of the O. tsutsugamushi isolates indicated that Taitung-1, -3 and -6 cluster together in the Taiwan-A branch and that Hualien-12 belongs to the Karp branch. These four aforementioned isolates and Taitung-7 are clustered in a Karp-related gene cluster. Taitung-4, Hualien-11 and -14 were in the TA763 cluster, and Hualien-1 was in the Kawasaki cluster. Hualien-3, 4, -5 and -6 were in the Kato cluster. Hualien-2, -7, -8, -9, -10 and -13 as well as Taitung-2 and -5 are grouped in the same branch. The eight aforementioned Taiwan-H isolates were the isolates most frequently cultured from eastern Taiwan and were clustered in the Taiwan Gilliam-variant (TG-v) cluster (Figure 2 and S1). In summary, phylogenetic and RFLP analysis of the O. tsutsugamushi genotypes indicated that Taiwan-A, -E and Karp are Karp-related isolates; Taiwan-B, -G and -J are Kato-related isolates; Taiwan-C is a Kawasaki isolate; Taiwan-N, -O and -P are TA763 isolates and Taiwan-H isolates are grouped in the TG-v cluster. Taiwan-H (TG-v), the most common O. tsutsugamushi isolate in eastern Taiwan, and genotype Taiwan-P are new recombinant isolates. The similarity of the 56-kDa TSA genes of the two new isolates and other prototype strains was analysed as described in the next paragraph.

Detection of possible recombinant isolates among various O. tsutsugamushi isolates and reference strains
BootScan analysis showed that the 56-kDa TSA gene sequences of eight O. tsutsugamushi TG-v isolates represented Gilliam, Ikeda, and Kato recombinants of identical structure, as illustrated by isolate Taitung-5 in Figure 3 [25]. A sliding window of 200 nt and a step of 20 nt were used to demonstrate the permuted tree of the 56-kDa TSA gene sequences of O. tsutsugamushi [26,27], and with signals of 70% or more in the observed permuted trees indicating potential recombination events. BootScan and phylogenetic analyses showed that the 56-kDa TSA gene sequence of O. tsutsugamushi Taitung-5 represents a mosaic gene constructed from the 59-terminal region (nt 1-262) and was similar to the corresponding sequence of the Kato strain ( Figure S3) [28].  Figure S4) [25].

Discussion
PCR-RFLP, multilocus sequence typing (MLST), and whole genome sequencing (WGS) are at present the most commonly used research methods in the study of bacterial evolution and classification [13,14,17,21,29,30,31,32]. A series of high-throughput DNA sequencing technologies have made WGS more rapid and economical. Although WGS offers the advantage of providing complete information on the evolution of antigens and is useful in antibiotic research [31,32], the existence of a large number of disease-causing infectious agents makes WGS more time-consuming and problematic than PCR-RFLP and MLST. MLST is a powerful method for the definitive assignment of isolates within bacterial populations to specific clones [29,30], and is able to reveal the underlying population genetic structure. The method was used successfully in 2010 by Sonthayanon et al. in their study of the relationship between host infection and the rates of homologous recombination of O. tsutsugamushi genes [29]. Because O. tsutsugamushi strains exhibit antigenic diversity and can be distinguished using the sequence of the 56-kDa TSA gene, which encodes a major outer membrane protein [14,20], RFLP-PCR analysis of the 56-kDa TSA gene is superior to MLST for immunological and vaccine studies. Furthermore, it is also more efficient than MLST analysis of seven housekeeping genes reported in the Sonthayanon et al. study [29].  The regions used to determine the RFLP genotypes are located in VD-I of the 56-kDa TSA gene of O. tsutsugamushi, and the length and restriction enzyme sites of the PCR-RFLP products vary (Table 2) [12]. Our study revealed the diversity of the 56-kDa TSA gene of the human O. tsutsugamushi pathogen in eastern Taiwan by RFLP analysis. The PCR method had sufficient sensitivity to detect O. tsutsugamushi at low levels in blood samples from infected humans. It is difficult to determine the entire 56-kDa TSA gene sequence or whole genome sequence until the bacterium is cultured. Furthermore, studies on this bacterium must be performed in expensive biosafety Level three facilities and this constraint adds to the difficulty [31,33]. Moreover, studies are time consuming and problematic for laboratory workers. Although the entire 56-kDa TSA gene sequence is necessary for a detailed study of O. tsutsugamushi antigens, our study demonstrated that the PCR-RFLP genotype may be used effectively to categorize clinical scrub typhus and that the PCR-RFLP method can be useful in public health investigations of O. tsutsugamushi.
Our study results support the findings of previous studies. RFLP analyses of ten Taiwanese O. tsutsugamushi isolates in Taiwan between 1986 and 1990 have been reported by Tamura et al. (Figure 1) [12,14]. RFLP revealed the ten isolates to be Karp-and Gilliam-related genotypes. The t10 isolate cultured from field rodents in Chengkung Harbour in eastern Taiwan was a Gilliamrelated isolate, and it differed from the other nine Karp-related isolates cultured from other regions of Taiwan. Qiang et al. analyzed the 56-kDa TSA gene sequences of the ten abovementioned isolates and four additional strains isolated from Kinmen Island in 1999 [15] They found that the fourteen isolates exhibited highly variable sequences of the 56-kDa TSA gene and that TW46-1 (the t10 isolate) differed from other Karp-related isolates in the phylogenetic tree [14,15].
Kelly et al. found that Karp-related isolates were the most common of all genotyped isolates by analyzing 271 sequences of the 56-kDa TSA gene at a global level [6]. Similar results were found in an analysis of 56-kDa TSA gene sequences of twenty-two O. tsutsugamushi isolates (TW-1 to TW-22) cultured from blood samples of scrub typhus patients in western Taiwan in 2006-2007 [3]. In our study, the distribution of RFLP genotypes and genotyped isolates in eastern Taiwan differed from that in western Taiwan and worldwide, in that Taiwan-H was the most common genotyping isolates in eastern Taiwan (Table 2). Our study revealed the geographic distribution of O. tsutsugamushi and correlation of genetic diversity of the 56-kDa TSA gene among eastern Taiwanese, southeast Asian, or northeast Asian isolates ( Figure 1).
Taiwan-H was clustered in the same branch of the phylogenetic tree as TW46-1 from Taitung Country in eastern Taiwan, and it exhibited high sequence similarity with other isolates from Thailand (UT125, UT144, UT196 and UT396) reported by Blacksell et al.(data not shown) [34]. These results indicated that the cultured Taiwan-H isolates were closely related to Gilliam strain in Myanmar or to other strains in southeast Asia. The Taiwan-H cluster was designated TG-v (Taiwan Gilliam-variant) cluster in this study to distinguish it from the Japanese Gilliamvariant (JG-v) cluster named by Kelly et al [6]. According to the geographic distribution of O. tsutsugamushi, TG-v, Kawasaki and TA763 cluster isolates were found mostly in eastern Taiwan (Hualien and Taitung), and Karp-related, Kuroki and Katorelated cluster isolates were found mostly in other parts of Taiwan ( Figure 1). TG-v, Gilliam, Kawasaki and TA763 cluster isolates were more closely related to isolates from Thailand and Myanmar (southeast Asia), and Kuroki, Karp-related and Kato-related cluster isolates in Taiwan were more closely related to strains from Japan, Korea and China (northeast Asia).
Eight isolates were sampled randomly from among 239 PCRpositive samples of the Taiwan-H RFLP genotype, and the fulllength 56-kDa TSA gene sequences were consecutively sequenced and analyzed. The percentage of similarity to the 56-kDa TSA gene sequences among the eight Taiwan-H strains exceeded 98%, and Taiwan-H was classified as a new cluster by phylogenetic analysis (Figure 2 and Figure S1). Based on phylogenetic tree analysis of the 56-kDa TSA gene, we deduced the DNA sequences of predicted PCR products using primers tsu-a and tsu-b from the TG-v cluster of O. tsutsugamushi isolates TW-16, TW-17, TW-18, TW-19, TW 46-1 and BB23 isolates, which had the same HhaI and SfaNI restriction enzyme sites. The other six O. tsutsugamushi isolates were grouped with eight Taiwan-H genotype isolates in the same TG-v cluster by phylogenetic tree analysis. The predicted PCR product using primers tsu-a and tsu-b and RFLP analysis corresponded to the Taiwan-H genotype (Figures S1 and S2).
TG-v displayed the highest percentage of RFLP and was the most frequently isolated type in eastern Taiwan (47.3%, 239/505; 42.6%, 120/282). BootScan analysis of the 56-kDa TSA gene demonstrated that TG-v was a new mosaic isolate derived from the Gilliam, Ikeda and Kato strains (Figure 3 and Figure S3). The geographical distribution of O. tsutsugamushi shows that the TG-v 56-kDa TSA gene originated in southeast and northeast Asia (Figure 1). However, the genetic diversity of the TG-v 56-kDa TSA gene corresponded to Taiwan's location in the center of the geographical triangle (Figure 1). For the same reasons, the geographical distribution of O. tsutsugamushi showed that the origin of 56-kDa TSA gene of the Taiwan-P was TA-763 from southeast Asia and Kato from northeast Asia (Figure 4 and Figure S4).
RFLP is a rapid and powerful tool for categorizing O. tsutsugamushi from blood or other specimens of scrub typhus patients [14,21]. Taiwan is located in the centre of the tsutsugamushi triangle [2], and the great diversity among O. tsutsugamushi pathogenic isolates found in eastern Taiwan has previously been described. Our protocol offers the advantages of using small sample amounts and avoiding the need to culture the organism before genotyping. Typing of the 56-kDa TSA gene by PCR-RFLP is valuable for classifying O. tsutsugamushi isolates as well as for determining the genetic relationship and epidemic evolution of O. tsutsugamushi isolates.
Investigations using the 56-kDa TSA gene of O. tsutsugamushi variants could provide much valuable information for vaccine development [5,6,7]. A clinical serological study showed that 20% of scrub typhus patients infected by Taiwan-H are anti-Karp antibody positive (data not shown) and that Karp or Karp-related strains may not provide immunological protection as vaccine strains if they are not used together with Taiwan-H isolates. Our findings provide a powerful method to classify the genotypes of O. tsutsugamushi and new insights into genetic and epidemic evolution among Taiwanese O. tsutsugamushi isolates, and this method may be useful in vaccine development and epidemic disease control in the Pan-Pacific region in the future.

Ethics statement
Ethical approval for this study was granted by the Buddhist Tzu Chi General Hospital Research Ethics Committee (IRB101-13). Scrub typhus is a notifiable infectious disease in Taiwan. The ethical standards and laboratory diagnosis of scrub typhus were based on regulations of the Taiwan CDC.

Sample collection
Heparinised venous blood samples collected 2002 to 2008 from from eastern Taiwanese patients with suspected scrub typhus were submitted to the Taiwan CDC-contracted laboratory at Tzu Chi General Hospital for laboratory diagnosis. Peripheral blood mononuclear cells (PBMC) were separated from whole blood samples by gradient centrifugation using the Ficoll-Hypaque (GE Healthcare Bio-Sciences Corp NJ USA) technique. The samples were examined for scrub typhus positivity by nested PCR.
Bacterial DNA samples extracted from PBMC of patients confirmed by Taiwan CDC-contracted laboratory to have scrub typhus infection were used in this study. The detailed procedures are described below.

DNA extraction and PCR amplification
Bacterial DNA was extracted by using the Macherey-Nagel NucleoSpinH Blood DNA Mini Kit (Macherey-Nagel GmbH & Co. KG, Düren, Germany) according to the manufacturer's instructions and stored at 280uC. The primers used for molecular detection, PCR-RFLP genotyping, and nucleotide sequencing are shown in Table 1. Primers were designed using the framework of the 56-kDa TSA gene of the O. tsutsugamushi Karp strain and recognized sequences located in the conserved region. They were able to detect the prototype strains Karp, Gilliam, Kato, Shimokoshi, Kawasaki, and Kuroki as well as other isolates from the geographical triangle [17]. We used BLASTN to verify that all primers were capable of detecting O. tsutsugamushi genotype isolates, and the primers used are described in the next paragraph.
Primers tsu-34 and tsu-11 were used for detection during the first PCR of the clinical samples to identify scrub typhus infection [22], and the first PCR products were purified using Macherey-Nagel PCR Clean-up Gel Extraction NucleoSpinH Extract II (Macherey-Nagel). The primers tsu-10 and tsu-11 were used to improve the detection sensitivity for detection during semi-nested PCR of the first PCR products [22]. PCR amplification was performed in 50-mL volumes using the FastStart High Fidelity PCR System (Roche GmbH, Mannheim, Germany) according to the manufacturer's protocol. The reaction mixture contained 0.2 mM of each of the primers, 0.5 mM of deoxyribonucleotide triphosphate mixture and 2.5 U of FastStart Taq DNA polymerase (Roche). Specific fragments were amplified using one cycle of 5 min at 94uC; 40 cycles of 30 s at 94uC, 45 s at 55uC and 1 min at 72uC and one cycle of 10 min at 72uC. After electrophoresis, gels were stained with ethidium bromide and amplicons were visualized using an ultraviolet transilluminator (UVP BioImaging Systems, Cambridge, UK). The molecular weights of the PCR products were calculated using the LabWorks TM Analysis Software UVP versions 3.0.02.00 software (Solutions for Science of Life) and the results are showed in the Table 2.

RFLP analysis
When blood samples were found to be positive for scrub typhus by PCR, the genotypes were analysed successively by PCR-RFLP. The primers Tsu-34 and Tsu-11 were used for detecting PCR-RFLP products in the clinical samples to identify the scrub typhus RFLP genotypes, and the primers tsu-a and tsu-b were used for nested PCR [21,22]. The PCR products were purified using MontageH PCR Centrifugal Filter Devices (Millipore Corporation, Billerica, MA, USA). Specific fragments were amplified using one cycle of 5 min at 94uC; 40 cycles of 30 s at 94uC, 45 s at 55uC and 1 min at 72uC and one cycle of 10 min at 72uC. The respective nested PCR products of different O. tsutsugamushi genotypes are shown in Table 2. Kawamori et al. digested the products of nested PCR using six restriction endonucleases (HhaI, SfaNI, MnlI, AlwNI, BglII, and MboI) and distinguished six prototype genotype strains (Karp, Kato, Kuroki, Kawasaki, Gilliam and Shimokoshi) using HhaI and SfaNI [17]. Chen et al. used these two restriction endonucleases to identify six new Taiwan-specific genotypes isolates (Taiwan-A, -B, -C, -D, -E, and -F) [21]. For RFLP analysis, the purified PCR products were digested individually with HhaI or SfaNI (New England Biolabs, USA) [21] at 37uC overnight. The HhaI reaction mixture contained 10 mM potassium acetate, 20 mM Tris-acetate, 10 mM magnesium acetate, 1 mM dithiothreitol, and 100 mg/mL BSA, and the SfaNI reaction mixture contained 100 mM sodium chloride, 50 mM Tris-HCl, 10 mM magnesium chloride, and 1 mM dithiothreitol. The digested products were electrophoresed on 6% NuSieve agarose gels (Molecular Station, USA) as markers. The gels were stained with ethidium bromide, and DNA fragments were visualized using an ultraviolet transilluminator (UVP Biomaging Systems). The molecular weights of the PCR-RFLP product fragments were calculated using the LabWorksTM Analysis Software UVP versions 3.0.02.00 software (Solutions for Science of Life) and the results are shown in the Table 2.
Cell culture and isolation of O. tsutsugamushi L929 mouse fibroblast cells (ATCC CCL 1 NCTC Clone 929) purchased from Bioresource Collection and Research Center (Food Industry Research and Development Institute, Taiwan) were cultured in MEM containing 10% fetal bovine serum (Biological Industries, Israel) in T-75 flasks. Peripheral blood mononuclear cells were collected from acute-phase blood containing heparin (10 U/mL) and used to isolate O. tsutsugamushi in the Taiwan CDC-contracted laboratory; the samples were derived from scrub typhus patients. Bacterial isolates from cell culture were propagated in a T-25 flask containing a monolayer of confluent L929 cells grown in maintenance medium (MEM supplemented with 1% fetal bovine serum) at 37uC in a 5% CO 2 incubator for 10-14 days and then detected by PCR and the indirect immunofluorescence assay; the results were confirmed by Taiwan CDC. After 14-20 days of incubation, bacteria-infected L929 cells were scraped off and stored at 280uC until use. The cultured O. tsutsugamushi isolates were collected from the Taiwan CDC-contracted laboratory and used in this study.

DNA sequencing
Primers ST-A and ST-D were used for PCR amplification of the entire 56-kDa TSA gene of O. tsutsugamushi [14]. The PCR products were purified using the High Pure PCR Product Purification Kit (Roche) and cloned into the TA vector plasmid DNA using the pGEMH-T Easy Vector Systems kit (Promega, WI, USA). Plasmid DNA sequencing of the entire length of the cloned 56-kDa TSA gene was performed using the ABI Prism 3700 DNA sequencer (Applied Biosystems, CA, USA). Two or three different plasmid DNAs of the same O. tsutsugamushi isolate were sequenced to determine the accuracy of sequencing of the 56-kDa TSA gene [12,14].

GenBank submission
Sequences of the entire 56-kDa TSA gene of O. tsutsugamushi isolates from eastern Taiwan were determined and deposited in DDBJ/EMBL/GenBank databases (http://www.ncbi.nlm.nih. gov/genbank). The GenBank nucleotide sequence accession numbers for the nucleotide sequence data generated in our study are presented in Table 3. The accession numbers of the gene sequences used in our study are presented in Figure 2.

Phylogenetic analysis and BootScan analysis
The nucleotide sequences were aligned using the ClustalW software. Phylogenetic and molecular analyses were performed using the MEGA version 5.0 [24] or PHYLIP version 3.66 software [23]. Phylogenetic trees were constructed from DNA sequences using the neighbour-joining and maximum likelihood methods, and genetic distances were calculated using the Kimura two-parameter distance algorithm. Phylogenetic relationships among the study and reference sequences were evaluated to estimate the node reliability of the phylogenetic trees constructed using the two methods, with 1000 bootstrap replicates. Bootstrap support values above 75 were considered significant.
A similarity plot was generated using SimPlot version 3.5.1 software which compares one 56-kDa TSA gene sequence from eastern Taiwan isolates used as query sequence against the 56-kDa TSA gene sequences from other strains used as reference sequences [25]. Sequences indicating possible recombination sites and crossover evens were identified sequentially. Taitung-5 (an isolate with genotype Taiwan-H) was used to investigate the distribution of recombinant forms in seven other O. tsutsugamushi strains (Gilliam, Karp, Kato, Kawasaki, Shimokoshi, Ikeda, and TA763). When aligning each query sequence of the above isolates to the reference isolate, it was possible to assign their particular genotypes unequivocally. The next step was the identification of closely spaced shifts among different genotype strains using the sliding window technique [26,27,28]. Segments of sufficient length for phylogenetic analysis and those that resulted in phylogenetic trees showing distinct clades with bootstrap values of 70% or greater were considered sufficient for subtype assignment [26,27,28].
The same analysis steps were performed for the Taiwan-P isolate. Hualien-14 (Taiwan-P) was examined against seven other recombinants O. tsutsugamushi strains of identical structure (Gilliam, Karp, Kato, Kawasaki, Shimokoshi, Ikeda, and TA763). Three neighboring O. tsutsugamushi strains (Karp, Kato, and TA763) were used as reference strains in order to remove excess sequence noise in the BootScan analysis. Segments of sufficient length were evaluated using 1000 Bootstrap replicates [26,27,28].