Differentiation of Rickettsiae by groEL Gene Analysis

Abstract

The nucleotide sequences (534 to 546 bp) of the groEL gene, which encodes the 60-kDa heat shock protein GroEL, from 15 rickettsial strains were determined and compared. In the phylogenetic tree created by the unweighted pair group method with arithmetic averages and the neighbor-joining method, rickettsial strains could be distinguished from Ehrlichia strains. Five spotted fever group strains, four typhus group strains, and six scrub typhus group (STG) strains were differentiated as distinct entities. Unlike gltA and ompA gene analyses, differentiation between members of the genus Rickettsia and the STG rickettsiae by groEL gene analysis was possible. In comparison with 16S rRNA gene analysis, the groEL gene has a higher degree of divergence among the rickettsiae. We therefore successfully developed rapid differentiation methods, PCR-restriction fragment length polymorphism analysis and a species-specific PCR, based on the groEL gene sequences. Four Korean isolates were identified by these methods and groEL gene analysis. The results suggest that the groEL gene is useful for the identification and characterization of rickettsiae.

Rickettsiae, a group of highly fastidious gram-negative bacteria, are divided into three groups. These are the typhus group (TG), which consists of Rickettsia typhi, R. prowazekii, and R. canada; the spotted fever group (SFG), which includes about 20 different species; and the scrub typhus group (STG), which includes R. tsutsugamushi, which is to be transferred into the new genus Orientia (15).

Rickettsiae have been characterized conventionally and identified by serotyping and protein analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (15, 17). Since rickettsiae are difficult to grow because of their strictly intracellular nature, these methods are time-consuming and expensive. Also, cross-reactions with different species and different genera by serotyping have been reported (8, 14, 18).

Various DNA-based techniques have been developed for the species identification of rickettsiae. Phylogenetic analyses based on the 16S rRNA gene sequence have frequently been used, but since the sequences are highly conserved, significant inferences about intragenus phylogeny are not possible (15, 16). Phylogenetic analyses based on the citrate synthase-encoding gene (gltA) and the outer membrane protein-encoding gene (ompA) have been developed to find a more sensitive and significant phylogenetic relationship among the rickettsiae (4, 16). Comparison of gltA gene sequences is more sensitive than comparison of 16S rRNA gene sequences, but reliable phylogenetic interrelationships could be drawn only among the rickettsiae that diverged early from the common ancestor of SFG (4). The interspecies differences of the ompA gene are greater than those of both the 16S rRNA and the gltA genes among the SFG, but the gene sequences of some species (R. akari, R. australis, R. helvetica, R. belii, R. canada, R. typhi, R. prowazekii) were not available for testing (4, 15). Therefore, a new identification method that counteracts the disadvantages of these methods is needed.

Recently, the groEL gene was used to identify Streptococcus (5), Enterococcus (6), Staphylococcus (7), Bartonella (12), Mycobacterium (20), and Ehrlichia (21). The groEL genes, which encode the 60-kDa heat shock protein GroEL, are ubiquitous in both prokaryotes and eukaryotes and encode highly conserved housekeeping proteins which are essential for the survival of these cells. The groEL genes provide the defining evolutionary relationships among the members of the eubacterial lineage (7, 12).

In this study, we developed a new method for the identification of the rickettsiae based on groEL gene analysis. The nucleotide sequences (534 to 546 bp) of the groEL genes from 15 rickettsial strains were determined and compared. Korean isolates were identified by comparison of their groEL gene sequences with those of reference strains. Also, we developed a rapid identification method using PCR-restriction fragment length polymorphism (RFLP) analysis and primer sets specific for species of the STG, SFG, and TG rickettsiae based on their groEL gene sequences.

MATERIALS AND METHODS

Bacterial strains and DNA extraction.

The rickettsial strains used in this study are listed in Table 1. The following strains were obtained from the American Type Culture Collection (ATCC; Manassas, Va.): Orientia tsutsugamushi Karp (VR-150) and Kato (VR-609), R. typhi Wilmington (VR-144), R. prowazekii Breinl (VR-142), R. akari MK (VR-148), R. japonica YH (VR-1363), R. conorii Indian Tick Typhus (VR-597), and R. sibirica 246 (VR-151). O. tsutsugamushi Kawasaki was supplied by O. Ogawa (National Institute of Infectious Diseases, Tokyo, Japan).

TABLE 1.

Rickettsial strains used in this study

Species and strain	Source	Geographic location	RFLP patterna	GenBank accession no.
				groEL gene	16S rRNA gene
O. tsutsugamushi
Karp	Human	New Guinea	A	M31887	D38623
Kato	Human	Japan	A	AY191586	D38624
Kawasaki	Human	Japan	A	AY191587	D38625
Gilliam	Human	Burma	NDb	AY191585	D38622
Boryong	Human	Korea	A	AY059015
Youngworl	Human	Korea	A	AY191588
Hwasung	Human	Korea	A	AY191589
R. typhi
Wilmington	Human	United States	C	AY191590	U12463
87-91	Blood (human)	Korea	C	AY191591
87-100	Blood (human)	Korea	C
R. prowazekii Breinl	Human	Poland	B	Y15783	M21789
R. rickettsii R	Dermacentor andersonii	USA	ND	U96733	L36217
R. akari MK	Blood (Human)	USA	B	AY059013	L36099
R. japonica YH	Blood (Human)	Japan	C	AF432181	L36213
R. conorii
Indian Tick Typhus	Rhipicephalus sanguineus	India	D	AY059012	U12460
R. sibirica 246	Dermacentor nuttalli	Russia	D	AY059014	D38628

^aRFLP patterns of the amplified groEL DNA were determined with the restriction enzyme DdeI.

^bND, not done.

Strains of the SFG and TG rickettsiae were cultivated in L929 cell monolayers (ATCC CCL NCTC clone 929) in tissue culture dishes (100 by 20 mm; Falcon; Becton Dickinson, Paramus, N.J.) supplemented with Earle's minimum essential medium (Gibco BRL, Life Technologies Ltd., Paisley, Scotland), 4% fetal bovine serum (Gibco BRL), and 2 mM l-glutamine. Cultures of the SFG rickettsiae were incubated at 32°C, whereas cultures of the TG rickettsiae were incubated at 35°C. When the cells were heavily infected, they were harvested and stored at −70°C. The STG rickettsiae were cultivated in L929 cell monolayers overlaid with Dulbecco modified Eagle's minimal essential medium (Gibco BRL) in the presence of 1% glucose and 2 mM l-glutamine at 35°C. Heavily infected cells were harvested and stored at −70°C.

DNA was extracted with a High Pure PCR Template Preparation kit (Roche Diagnostics Co., Indianapolis, Ind.).

Nucleotide sequencing.

PCR was performed with a set of primers (primer RF [5′-GTTGAAGTT/AGTTAAAGG-3′; positions 650 to 666 in the R. rickettsii R numbering] and primer RR [5′-TTTTTCTTTT/ATCATAATC-3′; positions 1183 to 1166]) to amplify the groEL gene (534 to 546 bp). Template DNA (50 ng) and 20 pmol of each primer were added to a PCR mixture tube (AccuPower PCR PreMix; Bioneer, Daejeon, Korea), which contained 1 U of Taq DNA polymerase, 250 μM each deoxynucleoside triphosphate, 50 mM Tris-HCl (pH 8.3), 40 mM KCl, 1.5 mM MgCl₂, and gel loading dye. The volume was adjusted to 20 μl with distilled water. The reaction mixture was subjected to 30 cycles of amplification (30 s at 94°C, 45 s at 44°C, and 45 s at 72°C), followed by a 5-min extension at 72°C (model 9600 thermocycler; Perkin-Elmer Cetus). The PCR products were electrophoresed on a 1.5% agarose gel and purified with a QIAEX II Gel Extraction kit (QIAGEN, Hilden, Germany). The purified PCR product (20 ng) was ligated with 50 ng of the pGEM-T vector (Promega) at 16°C for 18 h and used for the transformation of Escherichia coli XL-1 Blue. Recombinant DNA was extracted with a High Pure Plasmid Isolation kit (Boehringer Mannheim, Indianapolis, Ind.). At least two clones of each strain were used for nucleotide sequence determination. The purified DNA was sequenced with the CEQ L DNA Analysis system and the CEQ 2000 Dye Terminator Cycle Sequencing kit (Beckman Coulter Inc., Fullerton, Calif.) with primers RF and RR and a forward and reverse sequencing primer (M13). For the sequencing reaction, 20 ng of purified DNA, 100 pmol of primer, 2 μl of 10× sequencing reaction buffer, 1 μl of the deoxynucleoside triphosphate mixture, 2 μl each of ddUTP, ddGTP, ddCTP, and ddATP, and 1 μl of DNA polymerase enzyme were mixed together; and the final volume was adjusted to 20 μl by adding distilled water. The reaction was performed for 30 cycles of 20 s at 96°C, 20 s at 50°C, and 4 min at 60°C.

Sequence analysis.

Sequences were aligned by using the multiple-alignment algorithm in the MegAlign software package (Windows version 3.12e; DNASTAR, Madison, Wis.). Phylogenetic trees were constructed by the neighbor-joining method and the unweighted pair group method with arithmetic averages (UPGMA) by using the MEGA program (11). A bootstrap analysis (100 repeats) was performed to evaluate the topology of the phylogenetic tree.

PCR-RFLP analysis with DdeΙ digestion.

PCR was performed with the same set of primers described above (primers RF and RR) to amplify the groEL DNA (534 to 546 bp), as described above. Endonuclease DdeΙ (New England Biolabs, Beverly, Mass.) was used to cleave the PCR products, as recommended by the manufacturer. The restriction fragments were electrophoresed on a 3% Amplisize agarose gel (Bio-Rad Lab., Hercules, Calif.) for 15 min at 100 V.

PCR for species-specific identification of rickettsiae.

Primer sets specific for species of the TG, the STG, and the genus Rickettsia (SFG and TG) were designed on the basis of the groEL DNA sequences determined. Template DNA (50 ng) and 20 pmol of each primer were added to a PCR mixture tube (AccuPower PCR PreMix; Bioneer), and the volume was adjusted to 20 μl with distilled water. See Table 3 for the PCR amplification conditions.

TABLE 3.

Intragroup variations of groEL and 16S rRNA gene sequences of rickettsial strains

Group	No. of strains	% Identity
		groEL gene	16S rRNA gene
SFG	5	95.3-99.6	97.4-99.4
TG	2	97.0	99.4
STG	4	95.4-99.1	99.2-99.7

Nucleotide sequence accession numbers.

The groEL gene sequences determined for the rickettsial strains have been deposited in GenBank, and their accession numbers are listed in Table 1. The groEL sequences of R. rickettsii R (GenBank accession no. U96733), R. prowazekii Breinl (GenBank accession no. Y15783), and O. tsutsugamushi Karp (GenBank accession no. M31887) were used for comparison (Table 1). The 16S rRNA gene sequences of O. tsutsugamushi Karp (GenBank accession no. D38623), O. tsutsugamushi Kato (GenBank accession no. D38624), O. tsutsugamushi Kawasaki (GenBank accession no. D38625), O. tsutsugamushi Gilliam (GenBank accession no. D38622), R. typhi Wilmington (GenBank accession no. U12463), R. prowazekii Breinl (GenBank accession no. M21789), R. akari MK (GenBank accession no. L36099), R. japonica YH (GenBank accession no. L36213), R. conorii Indian Tick Typhus (GenBank accession no. U12460), and R. sibirica 246 (GenBank accession no. D38628) were also used for comparison (Table 1).

RESULTS

Comparison of 16S rRNA and groEL nucleotide sequences.

The nucleotide sequences of the groEL gene (534 to 546 bp) and the 16S rRNA gene from 11 reference strains of rickettsiae were compared to determine pairwise similarities.

More than 91.6% similarity was observed among the groEL gene sequences of strains of the genus Rickettsia (SFG and TG). The groEL gene sequence similarities of strains of the genus Rickettsia (SFG and TG) to STG strains and Ehrlichia strains were 55.6 to 61.8 and 52.4 to 55.6%, respectively (Table 2). SFG strains showed 91.6 to 93.3% similarities to TG strains, but they showed lower levels of similarity (55.6 to 60.7%) to STG strains. The sequence similarities of STG strains to SFG strains, TG strains, and Ehrlichia strains were 55.6 to 60.7, 57.5 to 61.8, and 44.1 to 47.0%, respectively.

TABLE 2.

Levels of similarity between groEL sequences and between 16S rRNA gene sequences

Species and strain	% Similaritya
	R. conorii Indian Tick Typhus	R. sibirica 246	R. japonica YH	R. rickettsii R	R. akari MK	R. typhi Wilmington	R. prowazekii Breinl	O. tsutsugamushi Kato	O. tsutsugamushi Gilliam	O. tsutsugamushi Karp	O. tsutsugamushi Kawasaki	Ehrlichia sennutsu Miyaymab	Ehrlichia phagocytophila FG
R. conorii Indian Tick Typhus		99.2	99.6	99.5	98.2	98.5	98.2	88.3	88.2	88.2	88.2	76.7	79.2
R. sibirica 246	99.6		98.9	98.8	97.4	97.5	97.3	89.0	88.6	88.9	88.9	76.8	79.3
R. japonica YH	99.4	99.1		99.4	98.1	98.2	98.0	88.2	87.9	88.0	88.0	77.0	79.4
R. rickettsii R	99.4	99.1	99.6		98.1	98.2	98.0	88.2	87.9	88.0	88.0	76.5	79.1
R. akari MK	95.7	95.3	96.3	95.9		97.5	97.1	88.1	87.7	87.8	87.9	76.2	78.2
R. typhi Wilmington	93.1	92.7	92.9	93.3	91.9		99.4	88.5	88.2	88.3	88.2	76.5	79.4
R. prowazekii Breinl	92.3	91.9	92.1	92.5	91.6	97.0		88.7	88.4	88.6	88.5	76.5	79.3
O. tsutsugamushi Kato	56.0	55.8	55.8	55.8	59.6	61.4	57.7		99.4	99.7	99.4	76.0	79.0
O. tsutsugamushi Gilliam	56.7	56.4	56.7	56.7	60.7	61.8	59.2	95.4		99.6	99.2	75.7	78.7
O. tsutsugamushi Karp	55.8	55.6	55.6	55.6	59.4	61.6	58.2	99.1	96.0		99.4	75.8	78.8
O. tsutsugamushi Kawasaki	59.0	58.6	58.8	58.8	59.7	61.0	57.5	95.4	96.7	96.0		76.4	79.1
Ehrlichia sennutsu Japan	53.9	53.2	53.9	53.9	53.9	55.6	54.1	44.1	44.3	44.1	44.9		80.7
Ehrlichia phagocytophila FG	52.8	52.4	52.4	52.4	53.7	53.7	53.0	46.7	47.0	46.7	46.9	52.5

^aThe values on the upper right are the levels of similarity between 16S rRNA gene sequences, and the values on the lower left are the levels of similarity between groEL gene sequences.

^bE. sennutsu Miyayma (GenBank accession no. M73219) was used in 16S rRNA gene sequence analysis but not in groEL gene sequence analysis.

The 16S rRNA gene nucleotide sequences (1420 to 1425 bp [positions 11 to 1430 in the R. rickettsii R numbering]) were aligned and compared. More than 97.1% similarity of the 16S rRNA gene sequences was observed among strains of the genus Rickettsia (SFG and TG). The 16S rRNA gene sequence similarities of strains of the genus Rickettsia (SFG and TG) to STG strains and Ehrlichia strains were 87.7 to 89.0 and 76.2 to 79.4%, respectively (Table 2). SFG strains showed similarities to TG strains of 97.1 to 98.5%, but they showed lower levels of similarity (87.7 to 89.0%) to STG strains. The sequence similarities of STG strains to SFG strains, TG strains, and Ehrlichia strains were 87.7 to 88.9, 88.2 to 88.6, and 75.7 to 79.1%, respectively.

The similarities of the groEL and 16S rRNA gene sequences of the SFG, TG, and STG rickettsiae are shown in Table 3. The groEL gene sequence similarities among the SFG, TG, and STG rickettsiae were 95.3 to 99.6, 97.0, and 95.4 to 99.1%, respectively; and the 16S rRNA gene sequence similarities among the SFG, TG, and STG rickettsiae were 97.4 to 99.4, 99.4, and 99.2 to 99.7%, respectively.

Comparison of phylogenetic trees determined from the 16S rRNA and groEL sequences.

Phylogenetic trees based on the groEL gene sequences were constructed by UPGMA and the neighbor-joining method (data not shown) to investigate the relationships among the rickettsial strains (Fig. 1). The rickettsial strains could be distinguished from Ehrlichia strains. Five SFG strains (R. rickettsii, R. japonica, R. conorii, R. sibirica, and R. akari) and two TG strains (R. typhi and R. prowazekii) formed a cluster separate from the STG strains. Also, four STG strains (O. tsutsugamushi Kato, Karp, Kawasaki, and Gilliam) formed a cluster separate from strains of the genus Rickettsia (SFG and TG). SFG strains were closer to TG strains than to STG and Ehrlichia strains. Phylogenetic trees based on 16S rRNA gene sequences were constructed by UPGMA and the neighbor-joining method (data not shown) to investigate the relationships among the rickettsial strains (Fig. 2). Results similar to those found with the phylogenetic tree based on the groEL gene sequences were obtained with the phylogenetic tree based on the 16S rRNA gene sequences. However, R. akari was branched independently and formed a distinct cluster clearly separate from the other SFG strains (Fig. 2), as was shown in a previous report (16).

FIG. 1.

Phylogenetic tree based on groEL gene sequences of the rickettsial strains tested. The phylogenetic tree was constructed by UPGMA with MEGA software. Bootstrap analysis was performed with 100 replicates.

FIG. 2.

Phylogenetic tree based on 16S rRNA gene sequences of the rickettsial strains tested. The phylogenetic tree was constructed by UPGMA with MEGA software. Bootstrap analysis was performed with 100 replicates.

Nucleotide sequences and deduced amino acid sequences of groEL DNA from Korean isolates.

The groEL DNA (534 to 546 bp) was successfully amplified from four Korean isolates. The nucleotide sequences of the amplified DNAs were determined and compared for pairwise similarities (Table 4). Isolates Youngworl and Hwasung showed high levels of similarity (94.7 to 99.6%) to the STG strains. The sequence similarities of Youngworl and Hwasung to SFG, TG, and Ehrlichia strains were 56.6 to 60.1, 57.9 to 62.4, and 43.0 to 46.7%, respectively. Isolates 87-91 and 87-100 showed high levels of similarity (99.8%) to R. typhi Wilmington and showed relatively higher levels of similarity (97.2%) to R. prowazekii Breinl than to the other rickettsial strains tested. The sequence similarities of isolates 87-91 and 87-100 to SFG, STG, and Ehrlichia strains were 92.1 to 93.4, 61.2 to 62.4, and 53.7 to 55.6%, respectively.

TABLE 4.

Levels of similarity between groEL sequences and between GroEL amino acid sequences

Species and strain	% Similaritya
	R. rickettsii R	R. japonica YH	R. conorii Indian Tick Typhus	R. sibirica 246	R. akari MK	R. prowazekii Breinl	R. typhi Wil- ming- ton	R. typhi 87-91	R. typhi 87-100	O. tsutsuga- mushi Hwasung	O. tsutsuga- mushi Youngworl	O. tsutsuga- mushi Kato	O. tsutsuga- mushi Boryong	O. tsutsuga- mushi Karp	O. tsutsuga- mushi Kawasaki	E. sennetsu Japan	E. phago- cytophila FG
R. rickettsii R		99.6	99.4	99.1	95.9	92.5	93.3	93.4	93.4	59.0	56.6	55.8	59.2	55.6	58.8	53.9	52.4
R. japonica YH	100		99.4	99.1	96.3	92.1	92.9	93.1	93.1	59.0	59.7	55.8	59.2	55.6	58.8	53.9	52.4
R. conorii Indian Tick Typhus	100	100		99.6	95.7	92.3	93.1	93.3	93.3	59.2	59.9	56.0	59.4	55.8	59.0	53.9	52.8
R. sibirica 246	98.9	98.9	98.9		95.3	91.9	92.7	92.9	92.9	58.8	56.4	55.8	59.0	55.6	58.6	53.2	52.4
R. akari MK	98.9	98.9	98.9	97.8		91.6	91.9	92.1	92.1	59.9	60.1	59.6	60.1	59.4	59.7	53.9	53.7
R. prowazekii Breinl	98.9	98.9	98.9	97.8	97.8		97.0	97.2	97.2	60.7	57.9	57.7	57.7	58.2	57.5	54.1	53.0
R. typhi Wilmington	98.9	98.9	98.9	97.8	97.8	100		99.8	99.8	61.2	62.2	61.4	61.4	61.6	61.0	55.6	53.7
R. typhi 87-91	98.9	98.9	98.9	97.8	97.8	100	100		100	61.4	62.4	61.6	61.6	61.8	61.2	55.6	53.7
R. typhi 87-100	98.9	98.9	98.9	97.8	100	100	100		61.4	62.4	61.6	61.6	61.8	61.2	55.6	53.7
O. tsutsugamushi Hwasung	56.7	56.7	56.7	55.6	57.9	56.7	56.7	56.7	56.7		95.6	94.9	99.6	95.8	99.1	44.1	46.5
O. tsutsugamushi Youngworl	56.2	56.2	56.2	55.1	57.3	56.2	56.2	56.2	56.2	96.2		94.7	96.0	95.2	95.8	43.0	46.7
O. tsutsugamushi Kato	55.6	55.6	55.6	54.5	56.7	55.6	55.6	55.6	55.6	98.4	96.7		95.2	99.1	95.4	44.1	46.7
O. tsutsugamushi Boryong	56.2	56.2	56.2	55.1	57.3	56.2	56.2	56.2	56.2	98.9	97.3	99.5		96.2	99.5	44.5	46.5
O. tsutsugamushi Karp	56.2	56.2	56.2	55.1	57.3	56.2	56.2	56.2	56.2	98.9	97.3	99.5	100		96.0	44.1	46.7
O. tsutsugamushi Kawasaki	56.2	56.2	56.2	55.1	57.3	56.2	56.2	56.2	56.2	98.9	97.3	99.5	100	100		44.9	46.9
E. sennetsu Japan	60.1	60.1	60.1	59.0	59.6	60.1	60.1	60.1	60.1	49.2	48.6	48.6	49.2	49.2	49.2		52.5
E. phagocytophila FG	56.7	56.7	56.7	55.6	56.2	56.7	56.7	56.7	56.7	48.9	47.8	47.8	48.3	48.3	48.3	54.7

^aThe values on the upper right are the levels of similarity between groEL sequences, and the values on the lower left are the levels of similarity between GroEL amino acid sequences.

The deduced amino acid sequences of amplified groEL DNA comprised 178 amino acid residues (V₁₈₇ to K₃₆₄ in the R. rickettsii R numbering). The deduced amino acid sequences were compared for pairwise similarities (Table 4). Isolates Youngworl and Hwasung showed high levels of similarity (96.7 to 98.9%) to STG strains. The deduced amino acid sequence similarities of Youngworl and Hwasung to SFG, TG, and Ehrlichia strains were 55.1 to 57.9, 56.2 to 56.7, and 47.8 to 49.2%, respectively. The deduced amino acid sequences of isolates 87-91 and 87-100 were identical to the amino acid sequences of R. typhi Wilmington and R. prowazekii Breinl. The deduced amino acid sequence similarities of isolates 87-91 and 87-100 to SFG, STG, and Ehrlichia strains were 97.8 to 98.9, 55.6 to 56.7, and 56.7 to 60.1%, respectively.

Phylogenetic analysis with groEL genes from Korean isolates.

Phylogenetic trees were constructed by the neighbor-joining method (Fig. 3) and UPGMA (data not shown). In the phylogenetic tree, isolates Youngworl and Hwasung formed a cluster with the STG strains separate from the other rickettsial strains. Isolates 87-91 and 87-100 formed a cluster with R. typhi Wilmington, were closer to R. prowazekii Breinl, and were separate from the other rickettsial strains tested (Fig. 3).

FIG. 3.

Phylogenetic tree based on the groEL gene sequences of Korean isolates. The phylogenetic tree was constructed by the neighbor-joining method with MEGA software. Bootstrap analysis was performed with 100 replicates.

PCR-RFLP analysis with DdeΙ digestion.

RFLP analysis with DdeΙ digestion easily distinguished the four groups. O. tsutsugamushi stains had the same restriction pattern (407 and 139 bp). Korean isolates Youngworl and Hwasung were readily distinguished from the other rickettsial strains. The restriction patterns of R. prowazekii and R. akari (322 and 212 bp) were well distinguished from those of the other rickettsial strains tested. Although the restriction pattern of R. typhi Wilmington (322, 114, and 98 bp) was not distinguished from that of R. japonica, it differed from that of R. conorii and R. sibirica (322, 114, 83, and 15 bp) (Table 5; Fig. 4). Also, Korean isolates 87-91 and 87-100 had the same restriction pattern as R. typhi Wilmington.

TABLE 5.

DdeI restriction fragments of amplified groEL DNA^a

Species	Restriction fragment sizes (bp)b	DNA size (bp)	RFLP pattern
O. tsutsugamushi	407, 139	546	A
R. prowazekii, R. akari	322, 212	534	B
R. typhi, R. japonica	322, 114, 98	534	C
R. conorii, R. sibirica	322, 114, 83, 15	534	D

^aThe groEL fragment amplified was 534 to 546 bp.

^bThe exact restriction fragment sizes were determined from the sequences.

FIG. 4.

DdeI restriction patterns of the amplified groEL DNA (534 to 546 bp) from rickettsial strains. Lanes: M, marker DNA (25- and 100-bp mixed DNA ladder); 1, O. tsutsugamushi Karp; 2, O. tsutsugamushi Kato; 3, O. tsutsugamushi Boryong; 4, O. tsutsugamushi Kawasaki; 5, O. tsutsugamushi Youngworl; 6, O. tsutsugamushi Hwasung; 7, R. prowazekii Breinl; 8, R. akari MK; 9, R. typhi Wilmington; 10, R. typhi 87-91; 11, R. typhi 87-100; 12, R. sibirica 246; 13, R. conorii Indian Tick Typhus; 14, R. japonica YH.

Species-specific PCR for identification of rickettsiae.

PCR with species-specific primers sufficiently differentiated the rickettsial species (Table 6; Fig. 5). STG-specific primers could amplify only the strains of STG. Korean isolates Youngworl and Hwasung were readily distinguished from the other rickettsial strains tested with the STG-specific primers. TG-specific primers specifically amplified groEL DNA only from the strains of TG. Korean isolates 87-91 and 87-100 were not distinguished from R. prowazekii but were readily distinguished from the other rickettsial strains tested by TG-specific primers. Rickettsia (TG and SFG)-specific primers amplified groEL DNA from the strains of the TG and SFG rickettsiae and were able to distinguish Rickettsia strains from STG strains.

TABLE 6.

Oligonucleotide sequences of species-specific primer sets

Group	Primera	PCR condition	DNA size (bp)
STG	TTGCTGATGATGTAGACGGA (F)	94°C for 30 s, 58°C for 30 s, 72°C for 30 s; 30 cycles	307
	TGTTCACAACGAGAATTAACTT (R)
TG	GGTGAAGCACTTGCGACG (F)	94°C for 10 s, 68°C for 5 s, 72°C for 20 s; 30 cycles	72
	AGGAGCTTTTACTGCTGC (R)
Rickettsia (TG and SFG)	GATAGAAGAAAAGCAATGATG (F)	94°C for 30s, 56°C for 30 s, 72°C for 30 s; 30 cycles	217
	CAGCTATTTGAGATTTAATTTG (R)

^aF, forward primer; R, reverse primer.

FIG. 5.

Amplification of groEL DNA from rickettsial strains with the primers specific for the species of STG (A), TG (B), and the genus Rickettsia (C). Lanes M, marker DNA (25- and 100-bp mixed DNA ladder). (A) Lanes: 1, O. tsutsugamushi Karp; 2, O. tsutsugamushi Kato; 3, O. tsutsugamushi Boryong; 4, O. tsutsugamushi Kawasaki; 5, O. tsutsugamushi Youngworl; 6, O. tsutsugamushi Hwasung; 7, R. prowazekii Breinl; 8, R. typhi Wilmington; 9, R. sibirica 246; 10, R. conorii Indian Tick Typhus; 11, R. japonica YH; 12, R. akari MK; 13, Coxiella burnetii Henzerling; 14, Ehrlichia sennetsu Miyayama. (B) Lanes: 1, R. prowazekii Breinl; 2, R. typhi Wilmington; 3, R. typhi 87-91; 4, R. typhi 87-100; 5, O. tsutsugamushi Karp; 6, O. tsutsugamushi Kato; 7, O. tsutsugamushi Boryong; 8, O. tsutsugamushi Kawasaki; 9, R. sibirica 246; 10, R. conorii Indian Tick Typhus; 11, R. japonica YH; 12, R. akari MK; 13, C. burnetii Henzerling; 14, E. sennetsu Miyayama. (C) 1, R. sibirica 246; 2, R. conorii Indian Tick Typhus; 3, R. japonica YH; 4, R. akari MK; 5, R. prowazekii Breinl; 6, R. typhi Wilmington; 7, R. typhi 87-91; 8, R. typhi 87-100; 9, O. tsutsugamushi Karp; 10, O. tsutsugamushi Kato; 11, O. tsutsugamushi Boryong; 12, O. tsutsugamushi Kawasaki; 13, C. burnetii Henzerling; 14, E. sennetsu Miyayama.

DISCUSSION

Scrub typhus infection, caused by O. tsutsugamushi, accounts for up to 23% of all febrile episodes in areas of the Asia-Pacific region where scrub typhus is endemic and has a mortality rate of up to 35% if it is left untreated (2). Diagnosis of scrub typhus is generally based on clinical presentation, the history of the patient, and serological assays. However, the differentiation of scrub typhus from other acute febrile illnesses is difficult because the symptoms are very similar (2, 15). In particular, the differentiation of scrub typhus from other rickettsial diseases, such as spotted fever diseases, murine typhus, and epidemic typhus, is very difficult. Serological assays are also time-consuming and expensive, since rickettsiae are highly fastidious (15). Therefore, a rapid method for the identification of rickettsiae is required. The ability to distinguish between the STG and SFG rickettsiae is especially important in the Asia-Pacific region.

GroEL is a highly conserved heat shock chaperonin system of proteins in bacteria, thus establishing a phylogenetic relationship between rickettsiae and bacteria (7, 12, 20). The groEL gene provides better topological and statistical support than the gltA and 16S rRNA genes for the determination of similarity within Bartonella species (12). The groEL gene can also be used as a tool for the rapid identification of Streptococcus (5), Enterococcus (6), Staphylococcus (7), and Mycobacterium (20) by dot blot hybridization, PCR-RFLP analysis, etc.

To date, the groEL sequences of rickettsiae available from GenBank are those of R. rickettsii R (GenBank accession no. U96733), R. prowazekii Breinl (GenBank accession no. Y15783), and O. tsutsugamushi Karp (GenBank accession no. M31887). We designed and synthesized a set of primers whose sequences were based on regions of groEL gene sequences, which were previously shown to be conserved between the rickettsiae and other, different microorganisms, and that could successfully amplify groEL DNA (534 to 546 bp).

The interspecies differences of the groEL gene sequences among the SFG rickettsial strains could not be compared with those of the gltA and ompA gene sequences because the groEL gene sequences of only five SFG strains were compared.

Analysis of the groEL gene has several advantages over the analysis of other genes. Differentiation between members of the genus Rickettsia and the STG rickettsiae by gltA and ompA gene sequence analyses was not possible because the sequences of the gltA and ompA genes of the STG rickettsiae had not been available until recently. Analysis of the groEL genes of the STG rickettsiae could be done with the same primers used for amplification of the groEL genes of members of the genus Rickettsia. In comparison with the 16S rRNA gene sequences, the groEL gene sequences have higher degrees of divergence among the rickettsiae. More than 91.6% similarity was observed among the groEL gene sequences of strains of the genus Rickettsia (SFG and TG). More than 97.1% similarity was observed among the 16S rRNA gene sequences of strains of the genus Rickettsia (SFG and TG). The similarities of the groEL gene sequences among SFG, TG, and STG strains were 95.3 to 99.6, 97.0, and 95.4 to 99.1%, respectively, whereas the similarities of the 16S rRNA gene sequences among SFG, TG, and STG strains were 97.4 to 99.4, 99.4, and 99.2 to 99.7%, respectively. Using the higher level of divergence of the groEL gene sequence, we successfully developed rapid differentiation methods: PCR-RFLP analysis and a species-specific PCR based on the groEL gene sequences. These methods will be especially useful in areas of the Asia-Pacific region where scrub typhus is endemic. These results show that groEL gene analysis is useful for the differentiation of STG strains and strains of the genus Rickettsia.

The sequences of some species (R. akari, R. australis, R. helvetica, R. belii, R. canada, R. typhi, and R. prowazekii) were not available for ompA gene analyses. However, the sequences of R. akari, R. typhi, and R. prowazekii were available for groEL sequence analysis, although other species (R. australis, R. helvetica, R. belii, and R. canada) could not be used for groEL sequence analysis. R. akari clustered with the SFG rickettsiae in the tree based on the groEL sequence, although it was considered an outlying member of SFG (16). R. akari also clustered with the SFG rickettsiae in the tree based on the gltA sequence (16) and was readily distinguished from the STG and TG rickettsiae by the species-specific PCR based on groEL sequences. These results show that groEL sequence analysis is also very useful for the differentiation of the SFG and TG rickettsiae.

We characterized four Korean isolates (isolates Youngworl, Hwasung, 87-91, and 87-100). The four isolates were identified as O. tsutsugamushi and R. typhi only by their reactivities with monoclonal antibodies. In the phylogenetic tree, Youngworl and Hwasung formed a cluster with STG strains separate from the other rickettsial strains (Fig. 3). Isolates 87-91 and 87-100 formed a cluster with R. typhi Wilmington and were closer to R. prowazekii Breinl and were separate from other rickettsial strains (Fig. 3). Youngworl and Hwasung had the same DdeI restriction pattern (407 and 139 bp) as the other O. tsutsugamushi strains and were readily distinguished from the other rickettsial strains tested. Isolates 87-91 and 87-100 had the same DdeI restriction pattern (322, 114, and 98 bp) as R. typhi Wilmington and were readily distinguished from the other rickettsial strains tested.

The results suggest that groEL DNA is useful for the identification and characterization of rickettsiae. groEL gene analysis is able to facilitate the rapid diagnosis of rickettsial diseases and to differentiate rickettsial diseases from other acute febrile diseases.

Scrub typhus is one of the most prevalent febrile illnesses in Korea (19). R. typhi and R. prowazekii have also been isolated in Korea, although there has been no recent documentation of this (1). Although no case of SFG rickettsiosis had been detected in Korea since the isolation of R. akari from Korean voles in 1957 (9), the possibility is strong that the SFG rickettsiae exist in Korea. There are two reasons for this. First, new SFG rickettsiae have been isolated in all parts of the world (15). Since the isolation of R. japonica in Japan, new SFG rickettsiae have been reported in China, Thailand, and the eastern part of Russia (3, 13). Second, serological surveys of wild rodents suggest that SFG rickettsiae exist in Korea (1, 10). groEL gene sequence analysis is able to facilitate the rapid detection and isolation of SFG rickettsiae in Korea.