Phylogenetic analysis of Orientia tsutsugamushi strains based on the sequence homologies of 56-kDa type-specific antigen genes


*Corresponding author. Tel. and Fax: +81 (25) 268-1210, E-mail address:


Close and distant relationship among 31 strains of Orientia tsutsugamushi (20, two, one and eight strains were isolated in Japan, Korea, China and southeast Asia, respectively) were clarified using phylogenetic analyses based on homologies of 56-kDa type-specific antigen genes. Isolates in Japan, Korea and China were located in eight separate clusters in the phylogenetic tree, and each was designated as JG (Japanese Gilliam type), JP-1 and JP-2 (Japanese Karp 1 and 2 types), Kato, Kawasaki, Kuroki, Shimokoshi and LX-1 types. All isolates originated in southeast Asia, including the prototype Gilliam and Karp strains isolated in Burma and New Guinea, respectively, were distantly located in the phylogenetic tree from those isolates in Japan, Korea and China, indicating that strains of O. tsutsugamushi distributed in northeastern and southeastern Asia are different types.


Scrub typhus, also called tsutsugamushi disease, is due to infection with Orientia tsutsugamushi, which is transmitted by the bite of trombiculid mite. O. tsutsugamushi, an obligate parasitic bacterium in the family Rickettsiaceae, contains many antigenic variants [1,2]. Antigenic differences among Gilliam, Karp and Kato strains of O. tsutsugamushi were well demonstrated by Shishido [3] in 1962, and he concluded that isolates obtained in Japan could be classified into these three types. However, other strain types were found in Thailand [4], and it was also reported that Shimokoshi [5], Kawasaki [6], and Kuroki [7] strains, which were isolated from patients in Japan, were antigenically distinguished from the prototype strains of Gilliam, Karp and Kato. This antigenic variation depends largely on the diversities of the immunodominant 56-kDa type-specific antigen located on the surface of this microorganism [8], and typing of newly isolated strains can be carried out using immunofluorescent (IF) testing using strain- or type-specific hyperimmune sera or monoclonal antibodies which recognize 56-kDa antigen, or by restriction fragment length polymorphism (RFLP) of 56-kDa protein genes amplified by polymerase chain reaction (PCR). Many newly isolated strains from patients and natural hosts in Japan and in Taiwan were tested using IF tests with monoclonal antibodies and RFLP analyses, and the isolates were classified not only into types but also into further subtypes [1,2].

To further clarify the relationships among individual strains, the 56-kDa protein gene of each strain was amplified using PCR and sequenced in an automated nucleotide sequencer, and the relation between each strain was determined from phylogenetic analysis based on the sequence homologies in this study.

2Materials and methods

Strains of O. tsutsugamushi used for analysis in this study are shown in Table 1. The sequences of 56-kDa genes of six strains, Gilliam, Karp, Kato, Kawasaki, Kuroki and Shimokoshi, were reported previously [9–11]. Sequences of the genes of Yonchon and Boryong strains isolated in South Korea have been reported in Korean studies [12,13], and the sequences of TA678, 686, 716 and 763 strains obtained in Thailand [4] and the Sxh 951 strain in China were cited from the GenBank. The sequencing of other strains in Table 1 was carried out in this study. Isolation and cultivation of most strains in Table 1 was reported previously [1,14], except the cases of Iwataki-1, Akita 7 and Omagari strains, which were isolated in this study using the same methods described previously. All strains used for sequencing were cultivated in L cell cultures and DNAs were prepared from pellets of infected cell homogenates obtained after differential centrifugation at 200×g for 5 min and 10 000×g for 5 min by incubation at 50°C for 1 h with 0.1% sodium dodecylsulfate and 100 μg ml−1 proteinase K and extraction twice with phenol-chloroform-isoamylalcohol (25:24:1) mixture. The DNAs were precipitated with ethanol, resolved in 10 mM Tris buffer, pH 8.0, containing 0.1 mM ethylenediamine tetraacetate, and used as a template for PCR.

Table 1.  Strains of O. tsutsugamushi used in this study
  1. aType and subtype of strains were determined in our previous study [1] or by the same method in this study.

  2. bThe strains were isolated by us in 1993 from laboratory reared mites shared with the USAMRU laboratory in Malaysia.

  3. cReaction patterns to the MAbs were different from the other types. RFLP analyses were not done.

TypeaSubtypea by MAbSubtypea by RFLPStrainSourceLocationsYear of isolation
 JG-1JGIwataki-1Microtus montebelliKyoto1996
 JG-2JGLP-1L. pallidumNiigata1986
KPKPKPKarpHumanNew Guinea1943
 KTKTAkita 7Apodemus speciosusAkita1989
 KTKTOmagariA. speciosusAkita1990
LXLXLXLX-1Leptotrombidium spp.Niigata1986
Non-typablecLF-1L. fletcheriMalaysia1993b
   LA-1L. arenicolaMalaysia1993b
Not examinedYonchonHumanSouth Korea1989
   Sxh 951HumanChina1998
   BoryongHumanSouth Korea1998
   TA678Rattus rattusThailand1963
   TA686Tupaia glisThailand1963
   TA716Menetes berdmoreiThailand1963
   TA763Rattus rajahThailand1963

For PCR, the first half and the latter half of 56-kDa gene were amplified separately. Primers used for PCR and sequencing are shown in Table 2, and the positions corresponding to the 56-kDa gene are illustrated in Fig. 1. PCR amplification for the first half was carried out in 50 μl reaction mixtures containing 10 μl of DNA sample, 200 μM (each) deoxynucleotide phosphate, 250 μM each of primers A and B, 1.0 unit of Taq DNA polymerase, 2 μM MgCl2, and 5 μl of 10-fold concentrated buffer (Promega Co., Madison, WI, USA), with a gene Amp PCR system 2400-R (the Perkin-Elmer Co., Norwalk, CT, USA). The reaction was started by incubation of the reaction mixture at 94°C for 2 min, then performed for 30 cycles of 94°C for 1 min (denaturation), 55°C for 1.5 min (annealing), 72°C for 2 min (extension), and finally once at 72°C for 7 min. Amplification of the latter half of the gene was carried out following the same method as above except that primers C and D were used and the annealing period in the cycle was shortened to 1 min.

Table 2.  Primers used in this study
  1. aPrimers B, E and F were from reference [15]. Primer D was the same used in our previous report[1].

 PrimersaPosition corresponding to 56-kDa gene of Gilliam (from start codon)
Primer A5′-TTTCGAACGTGTCTTTAAGC-3′ (forward)−266 to −285
Primer B5′-ACAGATGCACTATTAGGCAA-3′ (reverse)847 to 865
Primer C5′-ATGCTAATAAACCTAGCGCT-3′ (forward)731 to 749
Primer D5′-CTAGAAGTTATAGCGTACACCTGCACTTGC-3′ (reverse)1546 to 1575
Primer E5′-GTTGGAGGAATGATTACTGG-3′ (forward)124 to 143
Primer F5′-AGCGCTAGGTTTATTAGCAT-3′ (reverse)731 to 749
Primer G5′-TCCACATACACACCTTCAGC-3′ (reverse)1459 to 1478
Figure 1.

Position of primers used for 56-kDa gene amplification (white arrow heads) and sequencing (black arrow heads). Open reading frame of the gene is represented by the heavy line, and boxes I, II, III and IV show positions of variable domains.

The PCR product obtained was layered on a Microspin S-400HR column (Amersham Pharmacia Bioteck Inc., Piscataway, NJ, USA) for removal of salts and primers, the eluent was recovered by centrifugation of the column at 3000 rpm for 2 min, and the product in the eluent was sequenced by a cycle sequencing method using an ALFred DNA sequencer (Pharmacia LKB Biotechnology AB Marketing Department, Uppsala, Sweden). Primers E and F were used for the first half amplified gene, and primer G was used for the latter half gene. These primers were labeled with Cy5 and sequences were obtained according to the instruction of Thermo Sequenase fluorescent labeled primer cycle sequencing kit (Amersham Pharmacia Biotech. Inc., Piscataway, NJ, USA). From these methods, about 1435-bp sequences of the N-terminal side among 1576 bp of the open reading frame (ORF) region of Gilliam 56-kDa gene was determined which corresponded to 91% of the whole gene.

Sequencing analysis was performed at least twice for each sample, and if necessary, repeated until satisfactory results were obtained. For phylogenetic analysis based on the base-sequence homologies, alignment of sequences, calculation of evolutionary distance values and construction of a dendrogram using the neighbor-joining method were carried out using the GENETYX software package (SOFTWARE development Co., LTD., Tokyo, Japan). Bootstrap values were obtained after 100 resamplings at each bifurcation.

3Results and discussion

The phylogenetic tree constructed with 56-kDa gene base sequence homologies of 31 strains of O. tsutsugamushi is shown in Fig. 2. Significant findings are summarized as follows. (i) Strains Ikeda, 405S and LP-1, which were identified as Gilliam type in the reaction with monoclonal antibodies in our previous study [1], showed greater than 99.9% homologies and formed a cluster. Newly isolated Iwataki-1 strain, and Yonchon and Sxh 951 strains isolated in Korea and China, respectively, were included in this cluster showing a greater than 99.4% homology to each other, indicating that strains of this type are distributed in Japan, Korea and China. However, the percent homology between these strains and the prototype Gilliam strain was 88%, indicating that the strains in this cluster are different from the prototype Gilliam strain, which was isolated from a patient in Burma. Therefore, we designated the type of strains in this cluster as ‘Japanese Gilliam (JG) type’. (ii) In our previous studies [1], Matsuzawa and 402I strains were identified as Karp type from the findings of immunological tests, but were divided into two subtypes of JP-1 and JP-2 by RFLP analyses. Kamimoto, Mori and Okazaki strains were also identified as Karp type from immunological test in our previous study [14], and these three strains formed a cluster together with 402I strain in the phylogenetic tree of Fig. 2 (percent homology of these four strains was greater than 99.8%), indicating that these belong in JP-2 subtype. JP-1 (Matsuzawa strain) and JP-2 subtype strains were located in separate clusters with 95.7% homologies between each cluster. They also showed 94.5 to 95.3% homology with the prototype Karp strain, respectively, which was isolated in New Guinea, indicating that Karp, JP-1 and JP-2 strains are independent types. (iii) Kawasaki type strains of Kawasaki, Kanda, Taguchi and Oishi were divided into four subtypes from the reaction patterns with several monoclonal antibodies in our previous study [1], but the 56-kDa gene homologies of these strains were greater than 99.8% and the strains formed a cluster. (iv) Kuroki type strains of Kuroki and Nishino were also distinguished from each other using RFLP analyses of PCR-amplified 56-kDa genes, and were designated as KR-1 and KR-2 subtypes in our previous study [1]. The 56-kDa gene sequence homology between Kuroki and Nishino strains was 98.7%. The Boryong strain, isolated in South Korea, was located near Kuroki and Nishino strains showing homology percents of 99.2 and 98.1% to each strain, respectively. These strains can be grouped together as Kuroki type, and the findings obtained indicated that this type of strain is distributed both in Japan and Korea. (v) All strains from Thailand, TA686, 678, 716 and 763, showed similar values of less than 85% homology with each other, and also did not show high homology with any other members in Table 1, indicating that all are independent types. LF-1 and LA-1 strains were isolated from Leptotrombidium fletcheri and Leptotrombidium arenicola, respectively, which were harvested in Thailand. These strains were also located separately from others in the phylogenetic tree. (vi) The Shimokoshi strain showed the lowest homologies to other strains (63 to 69%), indicating that this is a peculiar strain. LX-1 strain showed less than 80% homology to other strains, indicating that this is also a new type.

Figure 2.

Phylogenetic tree of O. tsutsugamushi strains constructed based on base-sequence homologies of 56-kDa type-specific genes. The numbers at nodes indicate bootstrap values. Bar shows genetic distance of 0.1000. The strains originated from southeast Asia were shown in boldface. Accession numbers of published sequences: Karp (M33004), Gilliam (M33267), Kato (M63382), Kawasaki (M63383), Kuroki (M63380), Shimokoshi (M63381), Boryong (L04956), Yonchon (U19903), Sxh 951 (AF050669), TA686 (U80635), TA678 (U19904), TA716 (U19905), TA763 (U80636). GenBank accession numbers of other strains: Ikeda (AF173033), LP-1 (AF173034), Iwataki-1 (AF173035), 405S (AF173036), Oishi (AF173037), Taguchi (AF173038), Kanda (AF173039), Omagari (AF173040), Akita 7 (AF173041), LX-1 (AF173042), Matsuzawa (AF173043), Mori (AF173044), Okazaki (AF173045), Kamimoto (AF173046), 402I (AF173047), Nishino (AF173048), LA-1 (AF173049), LF-1 (AF173050).

Thus, the genetic similarity and distances between each strain were clarified using this phylogenetic analysis based on 56-kDa gene sequence homologies. Although distribution of the same type strains were recognized in Japan, Korea and China as described above, it is interesting that strains isolated in southeast Asia, such as Gilliam, Karp, LA-1, LF-1 and TA series strains (these strains were shown in boldface in Fig. 2), were different from the isolates in Japan. This may be correlated with the different species of trombiculid mites distributed in the areas which are the reservoirs of this microorganism. In southeast Asia, Leptotrombidium deliense, L. arenicola and L. fletcheri are considered the main vectors for human infection, but the main vectors in Japan are Leptotrombidium pallidum, Leptotrombidium scutellare and Leptotrombidium akamushi. Distribution of L. pallidum and L. scutellare in Korea and China has been reported [16,17].

In our previous study [1], JG, JP, Kawasaki and Kuroki type strains could be divided into several subtypes, as shown in Table 1, based on the findings of immunological tests with several monoclonal antibodies or from RFLP analyses. While strains of JP-1 and JP-2 subtypes formed separate clusters in this analysis, each subtype in JG, Kawasaki and Kuroki types did not separate in cluster formation. Since the C-terminal side is very A-T rich and common primers for many strains are difficult to design, the sequence determined in this study was 91% of the whole molecule of the N-terminal side of the 56-kDa gene, and the 9% of the C-terminal side sequence was not determined. Therefore, one possible consideration is that the difference among the subtypes observed using immunological tests may be due to the different structure of this undetermined C-terminal side. Another possibility is that a minor difference in the gene sequence produces a different tertiary structure of the 56-kDa protein molecules and this may correlate to the different reactivities with monoclonal antibodies.

In our phylogenetic analysis of six strains, Gilliam, Karp, Kato, Kawasaki, Kuroki and Shimokoshi, based on the homologies of 16S rRNA gene sequences [18], a close relationship between Karp and Kuroki strains, and distant location of Shimokoshi strain from other strains were demonstrated, of which results showed similarity with those of the present study. However, 16S rRNA genes were well conserved among strains and homologies among the six strains described above were higher than 98.4%. Therefore this phylogenetic analysis is useful for distinction of O. tsutsugamushi from microorganisms belonging to other species or genus, but not adequate for classification or typing of strains belonging to O. tsutsugamushi. Contrastingly, 56-kDa genes showed high varieties among strains as shown in this study, and details of close and distant relationships among many strains were demonstrated more clearly by the homologies of 56-kDa gene sequences than those of 16S rRNA genes.


We thank T. Kadosaka, Aichi Medical College, for sharing Akita 7 and Omagari strains from his collections, S. Nakajima, Kyoto Prefectural Institute of Hygienic and Environmental Sciences, for providing wild rodent organs from which we could isolate Iwataki-1 strain, and M. Takahashi, Kawagoe Sogo High School, for supplying L. fletcheri and L. arenicola reared in his laboratory. This study was supported in part by Grants-in-Aid from Ministry of Education, Science, Sports and Culture, and the Promotion and Mutual Aid Corporation for Private Schools of Japan.