Each strain of O. tsutsugamushi was grown in mouse fibroblasts cells (L929), incubated at 34℃ in a humidified atmosphere of 5% CO₂, and purified by percoll density gradient centrifugation as described by Ohashi et al. (1). Standard techniques for DNA manipulation, such as plasmid DNA preparation, ligation, competent cell preparation and transformation were followed as described by Sambrook and Russell (16). The DNA to make the chimeric recombinant 56-kDa (cr56) protein was isolated from O. tsutsugamushi strains grown into the L929 cells as described previously (13). Primers were designed based on the DNA sequence coding for the 56-kDa outer membrane protein gene. Homological region which showed 30% or more amino-acid sequence homology with each other at 56-kDa outer membrane proteins of O. tsutsugamushi prototype Gilliam, Karp and Kato, were selected to prepare the chimeric antigen of which oligonucleotide primers were designed in order to include each selected sequence to make the chimeric 56-kDa protein (Table 1). A 21-kDa species-specific antigen in O. tsutsugamushi serotype Boryong strain for recombinant 21-kDa protein (r21) and a 56-kDa major antigen of O. tsutsugamushi Kangwon 87-61 strain for simple recombinant 56-kDa protein (kr56) were cloned in E. coli (Patent #, WO 2008/029981 A1) (15). Each plasmid carrying the inserts coding for cr56, r21, and kr56 were induced by 0.5 mM Isoproyl-b-D-thiogalactopyranoside (IPTG) (Sigma, St. Louis, MO, USA). The recombinant His tagged proteins were purified, using His binding resins (Novagen, Darmstadt, Germany) according to the manufacturer's instruction.

This study was approved by Institutional Review Board (IRB) of Hallym University. The control serum samples were obtained from healthy volunteers at Hallym University. Patient serum samples were referred from various hospitals of Hallym University Medical Center located in Chuncheon, Kangnam, Kangdong, Hangang and Sacred Heart Hospitals in Korea to diagnose scrub typhus-like symptoms from 1999 to 2000. Among them, 12 patients were confirmed with scrub typhus by the IFA test as described previously by Bozeman and Elisberg (20). All of these patients had at least two of the three symptoms of scrub typhus (fever, rash, and eschars), and all exhibited seroconversion or four-fold increment in antibody titers to O. tsutsugamushi by the IFA test. Samples were determined as scrub typhus positive by having IgM or IgG IFA titers of ≥ 1:10 or ≥ 1:40, respectively. An additional 32 serum samples were collected from scrub typhus patients who were diagnosed from only one serum sample having IgM or IgG IFA titers of ≥ 1:10 or ≥ 1:40, respectively (10, 11). For negative control, a total of 38 serum specimens were obtained from healthy blood donors, none of which reacted with O. tsutsugamushi by the IFA-IgM or -IgG test. In addition, 31 serum specimens from patients with other febrile diseases (12 specimens from hemorrhagic fever with renal syndrome [HFRS], 9 from murine typhus, and 10 from leptospirosis) were also obtained. HFRS and murine typhus were identified by the IFA test, and leptospirosis was identified by the microscopic agglutination tests as previously described (10, 17).

For RDT test, serum samples were prepared by two different hospitals separately. IFA-positive specimens were provided from Hallym University Hospital (n = 51) and Korea University Hospital (n = 58) as scrub typhus positive samples. Other serum specimens including murine typhus (n = 39) or HFRS (n = 55) infected patient serum specimens were provided by Hallym University Hospital (n = 79) and Korea University (n = 15).

SDS-PAGE was performed as described previously (5). The expression level of each antigen was confirmed by anti-his-horse-radish peroxidase (Abcam, Cambridge, MA, USA). The antigenicity to this antigen was determined by patient serum at 1;3,000 dilution and goat anti-human IgG (H&L)-HRP (Bio-Rad Laboratories, Richmond, CA) at a 1;10,000 dilution.

Serological diagnosis of scrub typhus was performed by the IFA as described by Kim et al. (10). Briefly, O. tsutsugamushi strains Gilliam, Karp, and Kangwon 87-61 were cultured on monolayer mouse L929 cells in a humidified 5% CO₂ atmosphere at 37℃ containing 2% fetal bovine serum (Gibco BRL, Grand Island, NY). When the cytopathic effect approached up to 70% of the cells, the cells were harvested and then dotted on teflon-coated spot slides, following fixation with acetone for 10 min. The slides were stored in a freezer at -70℃ without exposure to the air from outside until used. After applying two-fold serially diluted (1:10 to 1:1,280 in phosphate buffered saline [PBS]) patient serum to the antigen-coated spot on the slide for 30 min, fluorescein isothiocyanate-conjugated antibody, anti-human IgM or IgG (Cappel Laboratories, Cochranville, PA, USA), was applied to determine the positive signal. Endpoint titer against individual serum was represented as the reciprocal of the highest serum dilution at which rickettsiae exhibited clear positive fluorescence at any one of the three strains. From diagnostic study in Korea, cutoff values of titer at IgM and IgG IFA were ≥ 1:10 and ≥ 1:40 (10).

To develop the optimal dilutions of the recombinant protein antigens, the recombinant protein antigens were diluted with a 0.05 M carbonate buffer (pH 9.6) and added to wells at various concentrations (18). A 96 well plate (Corning, Corning, NY, USA) was coated with the diluted recombinant protein antigens and incubated at 4℃ for 18 hr. The plate was washed twice with PBS containing 0.05% Tween 20. The antigens were blocked with 100 µL of 3% bovine serum albumin (BSA) per well, and the plate was incubated at 37℃ for 2 hr. Patient serum was diluted 1:200 in PBS, and 100 µL of the diluted serum was used as a primary antibody. A peroxidase conjugated goat anti-human IgG (H&L) (Bio-Rad, Hercules, CA, USA) was diluted to 1:6,000 and used as a secondary antibody. A 100 µL of the prepared secondary antibody was added to the wells. After washing three times, 100 µL of 0.1 M citrate phosphate buffer (pH 4.9) containing 1 mg O-phenylene diamine dihydrochloride (OPD) (Sigma) per mL and 0.03% hydrogen peroxide was added to the wells. The plate was shielded from light at room temperature for 30 min, and then 50 µL of 1 M sulfuric acid was added to terminate the reaction. The plate was read at 490 nm using a Micro-ELISA reader (SPECTRA MAX 250, Molecular Devices, Sunnyvale, CA, USA).

One microliter of each purified recombinant protein antigen (0.1 µg/µL) was dropped onto a nitrocellulose membrane (GE healthcare Biosciences, Pittsburg, PA, Germany) and then dried at 37℃ for 1 hr. The dried membrane was blocked with a TBST (Tris-buffered saline containing 0.5% Tween 20) solution containing 5% skim milk for 1 hr. Twelve patients' serum samples were diluted to 1:3,000 as a primary antibody and then added to the membrane. The reaction was allowed for 1 hr and HRP conjugated goat anti-human IgG (Bio-Rad) was used to determine a positive reaction which appeared as a distinct dark dot on the X-ray film. If no dot was seen, the reaction was interpreted as negative. The results obtained were exhibited as - for non reactivity and 1+, 2+, 3+, 4+ indicating positive based on the degree of intensity.

A mixture of cr56, r21 and kr56 of O. tsutsugamushi as antigen was applied to RDT and this RDT was manufactured by Immunemed (15). The test procedure was briefly as follow: 300 µL diluent buffer including 3 µL serum was applied to the sample well of the test kit and then complex of antigen-serum antibody-gold conjugated anti-IgM or IgG was running on each IgM or IgG test strip. The result was read in 10 min. The red color appearing on control line (C) and test line (T) concurrently was regarded as positive. The test was considered as negative when only the control line appeared and invalid if no control line appeared.

KT clearinghouse (Center for Evidence-Based Medicine, Toronto, Canada) version was used to calculate all test performances. Variables indicate the number of true positives (TP), true negative (TN), false positive (FP), and false negatives (FN). Calculation of accuracy and reliability were conducted by (TP+TN)/number of all tests and ([TPXTN]-[FPXFN]/9[TP+FN][TN+FP]). Positive predictive value (PPV) indicates the value of TP/(TP+FP) and negative predictive value (NPV) is TN/(TN+FN).

The study protocol was approved by the institutional review board of Hallym University (IRB, HIRB-2008-002). Informed consent was waived by the board.

The amplified PCR products were electrophoresed and visualized by ethidium bromide staining as shown in Fig. 1 and 952-bp fragment in the Gilliam strain, 873-bp fragments in the Karp, and 886-bp fragment in the Kato strains, respectively, were fused to produce a 2,711-bp fragment. The PCR products of Boryong (549 bp) and Kangwon 87-61 (1,086 bp) were also observed. Protein expression of cr56 fusion protein, r21 protein and kr56 protein were confirmed by SDS-PAGE analysis, showing 101 kDa, 21 kDa, and 41 kDa, which are identical to the molecular mass predicted for the recombinant proteins, including His-tag (Fig. 2).

The His fusion proteins were purified with His resin and the reactivity of purified recombinant proteins with each antigen were analyzed on Western blotting, using the patient serum samples of scrub typhus (Fig. 3). Four serum samples from scrub typhus patients showed the strong and weak reactivity on the cr56 antigenic protein which includes each repetitive region of 56 kDa protein of prototype Gilliam, Karp, and Kato. On the other hand, three out of four serum samples showed the positive reaction with kr56 from Kangwon 87-61 and two out of four serum samples reacted with r21 protein from Boryong. Furthermore, some serum samples reacted to r21 or kr56, even though they reacted weakly or rarely to cr56 (data not shown).

The representative distributions of IFA titer, ELISA titer, and dot blot titer from twelve serum samples obtained from confirmed scrub typhus patients are shown in Table 2. Background reactivity was assessed by analyzing healthy serum specimens and the cut off values were generally set with comparing patients' and healthy person's serum samples. In our study, the cutoff value of IgM or IgG ELISA was simplified as 0.2 through adjustment of ELISA conditions. The mean of optical density (OD) from the healthy serum specimens was 0.085 ± 0.042 (to cr56/r21, 1:1), 0.078 ± 0.038 (to cr56/r21/kr56, 5:1:3), 0.081 ± 0.044 (to cr56/r21, 5:1) and 0.075 ± 0.035 (to kr56). No serum from healthy persons had OD greater than the cutoff value in the IgM or IgG ELISA.

In Table 3, when 44 serum samples from patients with scrub typhus, 38 serum samples from healthy person and 31 serum samples from patients with other febrile illness, were used for reactivity on recombinant antigen(s) at ELISA, mixture of cr56: r21 (1:1 or 5:1) showed 88.6-93.2% sensitivity and 91.3-92.8% specificity. In case of kr56 protein, when it was added to this mixture additionally, each positive and negative signal increased sensitivity to 95.5% and specificity to 97.1%.

The results obtained from the dot-blot analysis were exhibited as - for non reactivity and 1+, 2+, 3+, 4+ indicating positive based on the degree of reactivity (Table 4). In the dot-blot analysis, 41 (93.2%) of the 44 serum samples from patients with scrub typhus gave at least one positive to the cr56/r21 (1:1) antigen dot, 40 (90.9%) to the cr56/r21 (5:1), and 27 (61.4%) to the kr56, but 43 serum samples (97.7%) gave at least one positive to the cr56/r21/kr56 antigen dot. None of the negative control serum sample produced positive result. These results were comparable to the data obtained from ELISA.

Finally, the efficacy of RDT against scrub typhus to use the antigen mixture cr56/r21/kr56 (5:1:3) and another patients' serum samples is presented in Table 5. Its sensitivity was 99.1% and specificity was 98.9% from reactivity to IgM and IgG, and its accuracy was 99.0%.