INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Hemorrhagic fever with renal syndrome (HFRS) and hantavirus pulmonary syndrome (HPS) are rodent-borne viral zoonoses caused by viruses in the genus Hantavirus, family Bunyaviridae (3). At least 14 virus species and 10 serotypes have been identified by genetic and antigenic characterizations, respectively. Each hantavirus appears to have a single predominant natural reservoir (19).

Four of the hantavirus species, Hantaan (HTN), Seoul (SEO), Dobrava/Belgrade, and Puumala (PUU), which are all different serotypes, are known to cause HFRS, while Sin Nombre virus causes HPS. In far eastern Asia, at least two serotypes of hantavirus, HTN and SEO, are spreading (19). Since the severity of infection depends on the viral serotype, a specific diagnosis of the causative virus is important, not only for rodent control but also for therapeutic purposes.

Currently, the plaque reduction neutralization test (PRNT) is the most specific serodiagnostic procedure for differentiating between HTN and SEO infections (4). However, PRNT takes more than 1 week to perform and requires a containment laboratory for virus manipulation. Therefore, a simple, safe, and rapid diagnostic method which is able to distinguish HTN from SEO infection serologically is required.

Hantavirus nucleocapsid protein (NP) possesses immunodominant, linear, cross-reactive epitopes in the first 100 amino acids of the N terminus (5, 7, 26). In addition, serotype-specific epitopes have also been detected by serotype-specific monoclonal antibodies (MAbs) in NP (15, 18, 28). We used truncated NP (trNP) expressed by a recombinant baculovirus and found that the HTN-specific antigenic site of the NP occupied about half of the C terminus of the NP (28).

In this study, we examined specific regions of the HTN and SEO serotypes on the NP in more detail and use specific regions to produce a diagnostic antigen for serotyping.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Viruses and cells. HTN virus strain 76-118 (14), SEO virus strain SR-11 (12), and PUU virus strain Sotkamo (2) were used as representative strains of the HTN, SEO, and PUU virus serotypes. They were propagated in the E6 clone of Vero cells (ATCC c11008, CRL 1586) grown in Eagle's minimal essential medium (Nissui, Tokyo, Japan) supplemented with 5% fetal bovine serum. Recombinant baculoviruses (Autographa californica nuclear polyhedrosis virus) containing coding information from the NPs of HTN and SEO viruses were supplied by C. S. Schmaljohn of the U.S. Army Medical Research Institute for Infectious Diseases (USAMRIID), Frederick, Md. (21). Recombinant baculovirus containing coding information from the NP of PUU virus was supplied by A. Vaheri of Helsinki University, Helsinki, Finland (24). The recombinant baculoviruses were propagated in High Five cells grown in Grace's insect cell culture medium (GIBCO BRL) supplemented with 10% fetal bovine serum.

Construction of truncated recombinant baculoviruses. Primers were designed from previously published sequences (1, 20). The portion of the gene coding for amino acids (aa) 155 to 429 of HTN NP was amplified from cDNA of the S segment of HTN virus strain 76-118 (a gift from C. S. Schmaljohn) (20) with the primers ATGCGGATTCGATTTAAGGATGA and TTAGAGTTTCAAAGGCTCTTGGT. The first methionine codon (ATG, underlined) was added as an initiation codon. The same region of SEO NP was amplified from cDNA of the S segment of SEO virus strain SR-11 (a gift from C. S. Schmaljohn) (1) with primers ATGAGGATCAGATTCAAGGA and TTATAATTTCATAGGTTCCTGGT. The DNA was amplified in 30 cycles of 97°C for 30 s, 55°C for 30 s, and 72°C for 1 min. Then PCR products were subcloned into pCRII plasmid (TA-cloning kit; Invitrogen) according to the manufacturer's instructions. The cDNA encoding aa 155 to 429 was excised from the pCRII plasmid by digestion with EcoRI (Takara, Tokyo, Japan) and ligated into the donor plasmid pFASTBAC1 (GIBCO BRL). The cDNAs encoding aa 155 to 288 of HTN NP and aa 155 to 295 of SEO NP were generated by digestion of plasmid pFASTBAC1 which contained aa 155 to 429 of HTN NP or SEO NP. Truncated genes encoding aa 1 to 103, 1 to 244, and 1 to 355 of HTN NP were generated from cDNA encoding the whole NP as described previously (28) and ligated into the donor plasmid pFASTBAC1. Five kinds of trNP from HTN and two kinds of trNP from SEO were expressed by using the BAC-TO-BAC baculovirus expression system according to the manufacturer's instructions (GIBCO BRL). The prepared recombinant baculoviruses were stored for use as seed viruses at 4°C.

Preparation of recombinant NP (rNP) and trNP. Monolayers of High Five cells cultured in 75-cm² flasks were inoculated with 1 ml of recombinant baculovirus culture fluid (2.2 × 10⁸ focus-forming units/ml). Six days later, the cells were pelleted by low-speed centrifugation (1,400 × g for 5 min). The cells were resuspended in Dulbecco's phosphate-buffered saline, pH 7.2 (PBS), and centrifuged again. Then the cells were suspended in 2 ml of PBS and sonicated four times for 15 s on ice. Proteinase inhibitors, EDTA (0.5 mg/ml), leupeptin (10 µg/ml), pepstatin A (10 µg/ml), aprotinin (1 µg/ml), and phenylmethylsulfonyl fluoride (1 µg/ml) were added to the extracts to prevent the degradation of the antigens, and the cell extracts were stored at 40°C.

MAbs, immune sera, and patient sera. MAb clones ECO2, ECO1, and BDO1, directed against the NP of hantavirus, were supplied by J. B. McCormick and C. J. Peters of the Centers for Disease Control and Prevention, Atlanta, Ga. (18). We prepared clone DCO3, directed against the NP of SEO virus. Clones E5/G6, C16D11, F23A1, and C24B4, directed against the NP of HTN virus, were prepared previously (28). Immune mouse serum was obtained from Slc/ICR mice infected with HTN virus by intraperitoneal inoculation of live virus (27). Immune rat serum was prepared by intraperitoneal inoculation of live viruses into Slc/Wistar rats (29). A total of 35 HFRS patient sera were provided by Y.-X. Yu of the National Institute for the Control of Pharmaceutical and Biological Products, Beijung, China. Thirty of these patients were believed to be infected with HTN virus, and the others were thought to be infected with SEO virus. Sera that were positive for SEO virus were kindly provided by Y. Nishimune of the Research Institute for Microbial Diseases, Osaka University (four serum specimens), and I. Kim of Seoul National University, Seoul, Korea (six serum specimens). These sera were thought to be infected with SEO type hantavirus during a laboratory rat-associated infection. One serum specimen from a nephropathia epidemica patient was provided by B. Niklasson of the Swedish Institute for Infectious Disease Control.

Focus reduction NT. The neutralization test (NT) was performed essentially as described by Tanishita et al. (23) with some modifications. In this study, the residual virus infectivity was determined by indirect immunofluorescence staining, described below. The NT titer was expressed as the reciprocal of the highest serum dilution resulting in a reduction greater than 80% in the number of infected cell foci.

IFA assay. An indirect immunofluorescence antibody (IFA) test was carried out by previously described methods (27). Acetone-fixed smears of Vero E6 cells infected with hantaviruses or of High Five cells infected with recombinant baculoviruses were used as antigens. Fluorescein isothiocyanate-conjugated anti-human immunoglobulin (heavy and light chains) goat immunoglobulin G (IgG) (Cappel Laboratories, Cochranville, Pa.) was used as the second antibody. IFA titers were expressed as the reciprocal of the highest serum dilution that caused characteristic fluorescence in the cytoplasm.

Western blotting. Western blotting was performed according to published methods (29). The denatured rNP or trNP antigen from infected High Five cells (3.0 × 10³ cells/lane) was separated by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (13). Mouse or rat immune sera to hantaviruses or a MAb to the NP (E5/G6) was used to detect antigen on the membrane. Binding antibodies were detected with horseradish peroxidase-labeled goat anti-rat IgG or horseradish peroxidase-labeled goat anti-mouse immunoglobulins (Zymed, South San Francisco, Calif.). 4-Chloro-1-naphthol (Sigma Chemical Co., St. Louis, Mo.) was used as the peroxidase substrate.

Capture ELISA. Microtiter plates were coated with MAb E5/G6 (5 µg/ml in PBS) overnight at 4°C as the capture antibody. The enzyme-linked immunosorbent assay (ELISA) was performed as described previously (28). rNP expressed in E. coli by the PinPoint Xa vector, which carries biotin as a tag, was used for an avidin capture ELISA. We used essentially the same procedure described previously (28).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Antigenic characterization of HTN and SEO NPs with MAbs. As shown in Table 1, HTN trNP which expressed aa 1 to 103 (trNP HTN 1-103) reacted with three cross-reactive clones (ECO2, ECO1, and GBO4). From the reactivity of clone E5/G6 to trNP HTN 1-244, 155-288, and 155-429, the binding region for clone E5/G6 should be located between aa 155 and 244. This agrees with our previous result, in which a synthetic peptide comprising aa 166 to 175 bound to E5/G6 (28). Clones C16D11 and F23A1 reacted with trNP HTN 1-335 but not with trNP HTN 1-244, indicating that the binding region for the two clones is located between aa 244 and 335. However, trNP HTN 155-429 and 155-288 lacked reactivity with these two clones. This difference probably results from the conformational dependency of the epitopes in the N-terminal region of NP. Nevertheless, trNP HTN 155-429 still reacted with HTN-specific clones (C24B4 and BDO1) but not with any cross-reactive clones except E5/G6. Therefore, trNP HTN 155-429 was used as a serotyping antigen. Since this antigen retains reactivity to E5/G6, this antibody was used as the capture antibody for the ELISA. trNP SEO 155-429 also retained reactivity with E5/G6 and a SEO-specific MAb (DCO3) and was used as a SEO-specific diagnostic antigen.

Reactivity of trNP in Western blotting. Expression of the entire NP, and of trNP HTN 155-429 and trNP SEO 155-429 by a baculovirus system, was confirmed by Western blot analysis. In the Western blots, the complete NP and the trNPs reacted with E5/G6, because MAb E5/G6 bound to the linear epitope at aa 166 to 175 (28) (Fig. 1a). The molecular sizes of the whole and truncated products were 50 and 35 kDa, respectively. The observed and expected sizes were the same. Although mouse and rat immune sera to HTN or SEO strongly cross-reacted with the whole rNPs of both HTN and SEO, the immune sera did not react with trNP HTN 155-429 or trNP SEO 155-429 (Fig. 1b and c, lanes 3 and 4). These results indicate that the truncated antigens lack the linear epitopes that react with antibodies produced after a natural infection. The sera from several HTN and SEO patients also gave the same pattern as these experimentally infected animal sera (data not shown). Since the baculovirus-expressed recombinant protein is highly sensitive to cellular protease (27), the additional protein bands in Fig. 1, lanes 1 and 2 (all panels), and the doublet bands in 35-kDa trNPs with MAb E5/G6 (Fig. 1a, lanes 3 and 4) were considered to be NP fragments after degradation.

View larger version (42K): [in this window] [in a new window]   FIG. 1.   Detection of rNP and trNP with Western blots stained with MAb E5/G6 (a), mouse sera infected with HTN virus (b), or rat sera infected with SEO virus (c). Lane 1, HTN rNP (whole); lane 2, SEO rNP (whole); lane 3, trNP HTN 155-429; lane 4, trNP SEO 155-429; lane 5, uninfected High Five cells. In each lane, 3.0 × 103 cells were applied.

IFA profiles of patient sera. The IFA profiles of patient sera with High Five cells expressing whole HTN or SEO rNP, trNP HTN 155-429, trNP SEO 155-429, or unrelated antigen are shown in Fig. 2. The whole NP showed strong cross-reactivity with heterologous patient sera, while both the HTN and SEO trNP antigens showed significantly higher reactivity with patient sera infected with the homologous virus rather than the heterologous virus. However, the intensity of fluorescence was weaker than with the whole antigen.

View larger version (68K): [in this window] [in a new window]   FIG. 2.   IFA profiles of patient sera with High Five cells expressing whole or truncated HTN or SEO rNP. Panels were stained with HTN (a through e) or SEO (f through j) patient serum. High Five cells expressed whole HTN rNP (a and f), whole SEO rNP (b and g), trNP HTN 155-429 (c and h), trNP SEO 155-429 (d and i), or recombinant p40 of Borna disease virus (16) (e and j).

Reactivity of rNPs by IFA assay and ELISA. To examine the usefulness of the truncated antigens for serotyping, the reactivities of patient sera to the truncated antigens were compared to the reactivities with whole NPs and authentic viral antigens by IFA and ELISA (Table 2). In the NT, patient sera showed significantly higher titers of antibody to the homologous virus. In the IFA assay, the HTN and SEO whole-NP antigens cross-reacted equally with HTN and SEO patient sera but to a lesser extent with PUU patient sera. These cross-reactive patterns were the same as those obtained with Vero cells infected with authentic viruses. trNP SEO 155-429 was able to distinguish SEO patient serum from HTN and PUU patient sera. Particularly, HTN patient serum showed IFA titers of <32 to trNP SEO 155-429, even though its homologous antibody titer was very high (16,384 to 32,768). On the other hand, trNP HTN 155-429 also reacted only with HTN patient serum, but the antibody titers were more than 1/100 lower than those with whole rNP. In the ELISA, patient sera showed the same cross-reactivity as was seen with the IFA test (Table 2). Serial dilutions of the sera of standard patients, results for which are shown in Table 2, were also examined (Fig. 3). Whole HTN and SEO rNP reacted with both HTN and SEO patient sera, with similar intensities, in a dilution-dependent manner (Fig. 3a and b). On the other hand, trNP SEO 155-429 reacted only with SEO patient sera (Fig. 3d), while trNP HTN 155-429 did not react with any sera (Fig. 3c). Thus, trNP SEO 155-429 was able to distinguish SEO infection from HTN infection serologically.

View larger version (39K): [in this window] [in a new window]   FIG. 3.   Serial dilutions of the sera of standard patients in the ELISA. Optical densities at 490 to 600 nm (OD490-OD600) are given for sera of patients infected with HTN () or SEO () and for negative human sera (+ and ×). High Five cells expressed the NPs shown above the panels.

Serologic diagnoses of patient sera by IFA assay. The trNP HTN 155-429 and trNP SEO 155-429 antigens were used to serotype groups of patient sera, 20 HTN virus patient sera from China and 10 SEO virus patient sera from Korea and Japan, in an IFA assay. Sera from China seemed to be obtained from patients infected with HTN virus, based on epidemiological data. In a prior study, an NT confirmed that the Korean and Japanese patients were infected with SEO virus by laboratory rats. The sera were screened at a single serum dilution (1:32). All the patient sera were positive for either whole HTN or whole SEO antigens. With the truncated antigens, the sera from all but one Chinese patient were positive with trNP HTN 155-429 (19 of 20), and all were negative with trNP SEO 155-429. Similarly, trNP SEO 155-429 reacted only with the Korean and Japanese sera (10 of 10). These results show that both truncated antigens can be used as diagnostic antigens for serotyping in an IFA assay.

Reactivities of patients' serial sera. An IFA assay with rNP was used to evaluate the reactivities of serial sera from Chinese patients. Patients from Anhui and Henan provinces were infected with HTN and SEO viruses, respectively (Table 3). All the sera from Anhui patients (Table 3) showed significant IFA titers to Vero cells infected with authentic virus, even in the febrile phase (acute phase), and reached high titers by the recovery phase. Interestingly, acute-phase sera showed relatively similar titers of antibody to the antigen in the Vero cells infected with authentic virus and the trNP HTN 155-429 antigen (1/2 to 1/8 IFA titer difference). However, in the recovery-phase sera, the IFA titers to the truncated antigen were markedly lower than those to the Vero cell antigen (1/64 to 1/256). Serial sera from Henan (Table 3), which were from SEO virus-infected patients, showed an apparent increase in the antibody titers with the time since the onset of the disease, although clinical records of the patients' symptoms were not available. These patients were diagnosed as having SEO virus infection by IFA with trNP SEO 155-429 and trNP HTN 155-429. In ELISAs using recombinant baculovirus expressing NP antigens, the whole-NP antigens cross-reacted strongly with heterologous sera (Fig. 4A). trNP SEO 155-429 was able to distinguish SEO virus infection from HTN virus infection. However, trNP HTN 155-429 showed only negligible reactivity, even with recovery phase sera (Fig. 4B). An ELISA was also performed with E. coli-expressed antigens. Whole NP of HTN virus (aa 1 to 429) and trNP HTN 1-103, expressed by an E. coli vector (28), were used as antigens. The reactivity with trNP HTN 1-103 antigen was similar to that obtained with whole-NP antigen (Fig. 4C).

View larger version (14K): [in this window] [in a new window]   FIG. 4.   Reactivities of serial patients' sera in ELISA. OD490-OD600 were plotted for acute-phase () and convalescent-phase () sera of patients from Anhui province and for acute-phase () and convalescent-phase () sera of patients from Henan province. Comparisons of ELISA ODs were made between whole HTN rNP and whole SEO rNP (A), between trNP HTN 155-429 and trNP SEO 155-429 (B), and between whole HTN rNP and trNP HTN 1-103.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

A baculovirus expressing trNPs of the HTN and SEO viruses, which lacked the major linear epitope located in the first third of the N-terminal part of the NP, was successfully used as a diagnostic antigen. An IFA assay and an ELISA using the truncated antigen expressed by the baculovirus were able to distinguish HTN and SEO virus infections serologically.

The application of E. coli-expressed trNP as a serotyping antigen was reported by Wang et al. (25). They used dot matrix comparisons of the amino acid sequences of HTN, SEO, and PUU NPs to select two unique regions. These two regions were combined and expressed by an E. coli vector. In this report, we identified a particular region from the reactivity with serotype-specific MAbs that was useful as a serotyping antigen. It has been reported that the first 100 amino acids of the N terminus of the NP in various hantaviruses contain several major linear epitopes (5, 7, 26). The antigenicity of this region is highly cross-reactive between HTN and SEO viruses. It is difficult to differentiate between HTN and SEO virus infections by measuring the titers of antibody against NP because the region is immunodominant (9, 22). However, the production of serotype-specific MAb to NP indicated the existence of a serotype-specific epitope on the NP. We determined that a trNP that lacked 154 amino acids from the N terminus of the NP retained reactivity to the serotype-specific antibody (Table 1). However, the reactivity of the serotype-specific antibody to baculovirus-expressed antigen was much stronger than that to E. coli-expressed NP (28). Therefore, we tried to use the trNP expressed by baculovirus for serotyping antigen. The different characteristics of NP antigen expressed by baculovirus and E. coli were reported for PUU virus NP. The PUU NP expressed by E. coli as a -galactosidase fusion protein lacked reactivity with the MAb that recognized the variable region of the NP (aa 251 to 260), but baculovirus-expressed NP retained this reactivity. The method of antigen preparation, differences in the way the bacterial product is folded, or fusion of a part of the protein that affects its structure are all possible reasons for this difference (22). In our experiment, trNP lacked reactivity with immune animal sera (Fig. 1) and patient sera (data not shown) by Western blot analysis, indicating that the serotype-specific epitopes on the NPs of HTN and SEO viruses were conformation dependent. As shown in Table 1, trNP HTN 155-288, which lacked a possible RNA-binding region (6), also lacked reactivity with serotype-specific MAbs. This result indicates that binding with the serotype-specific MAbs requires the C-terminal region of NP. Since the amino acid sequences of the C termini (RNA-binding regions) of HTN and SEO viruses are almost identical, the antigenicities of these regions should also be identical. Therefore, we propose that this region does not contain strain-specific epitopes. Rather, it is important for maintaining the structure of the NP as a whole.

The IFA assays with baculovirus-expressed trNP HTN 155-429 and trNP SEO 155-429 were able to differentiate the infections serologically with the same sensitivity as the NT. However, the IFA titers with the truncated antigens were lower than those with the native viral antigens (Table 3) or whole NP (data not shown). This trend was more noticeable in titrations of sera from HTN-infected patients (Table 3). With acute-phase sera, there was less than an eightfold difference in IFA titers between native and truncated antigens. In contrast, the difference with convalescent-phase sera was around 128-fold. The IFA titers against trNP 155-429 in the convalescent phase were relatively lower than those in the acute phase. Hedman et al. conducted IgG avidity assays and reported that high-avidity antibodies were produced in the convalescent phase of hantavirus infection (8). Therefore, the relative decrease of IFA titers to trNP 155-429 in convalescent-phase sera might be reflected by the increase in high-avidity antibodies against the N-terminal immunodominant and cross-reactive region.

Without further purification, the trNPs were also used as ELISA antigens with MAb E5/G6 as the capture antibody. Only trNP SEO 155-429 was able to differentiate SEO patient sera from HTN patient sera by ELISA. trNP HTN 155-429 seemed to lose its reactivity, even with sera from patients infected with HTN virus. Since this antigen retained its reactivity in the IFA assay, the procedure used for antigen preparation, perhaps sonication, may decrease the reactivity of its serotype-specific epitope. This has been reported for PUU rNP in a baculovirus vector (24). Additional studies of the use of milder conditions for the solubilization procedure are needed. The level of reactivity in an ELISA using the truncated antigen expressed by E. coli (aa 1 to 103) was similar to that with whole NP (aa 1 to 429), and both antigens were highly cross-reactive between HTN and SEO patient sera. These results confirmed that the 100 N-terminal amino acids of NP are a useful antigen for serological diagnoses.

Hantaviruses that are closely related to PUU virus antigenically and genetically have been reported in Asia and the Far East. These include Japanese PUU virus (11), Khbarovsk virus (10), and Lemmings virus (17). Epidemiological reports suggest that it is necessary to monitor for PUU virus infection even in areas where PUU infection has not been reported previously. Since the antigenicity of PUU virus is considerably different from that of HTN and SEO viruses, a combination of at least two different antigens, either HTN and PUU or SEO and PUU, is necessary for this serological surveillance.