Use of immunoblotting as an alternative method for serogrouping Leptospira

Leptospirosis is one of the most important zoonotic diseases worldwide, and is caused by pathogenic Leptospira species, which are not readily distinguishable from saprophytic leptospire strains on the basis of morphology and biochemical characteristics. The disease varies from subclinical infection to a severe illness with multi-organ involvement (Bharti et al., 2003), which makes the diagnosis difficult sometimes it is diagnosed as other febrile illnesses. Therefore, confirmation of the diagnosis made by specific microbiological tests is necessary (Levett, 2001). Initially, the genus Leptospira was divided into two species, Leptospira interrogans, comprising all pathogenic strains, and Leptospira biflexa, containing saprophytic strains isolated from the environment (Johnson & Faine, 1984). Both L. interrogans and L. biflexa are divided into numerous serovars defined by agglutination after cross-absorption with homologous antigen (Dikken & Kmety, 1978; Johnson & Faine, 1984; Kmety & Dikken, 1993). Currently, the phenotypic classification of leptospires has been replaced by a genomic one, in which genomospecies, including L. interrogans sensu stricto and L. biflexa sensu stricto, do not correspond to the previous two species (L. interrogans and L. biflexa), and indeed, pathogenic and non-pathogenic serovars occur within the same species (Levett, 2001). However, the molecular classification is incompatible with the system of serogroups that has been familiar to clinicians and epidemiologists for a long time. Most clinical laboratories find it necessary to retain serological classification of pathogenic leptospires for epidemiological purposes and for clinical diagnosis, of which microscopic agglutination test (MAT) is the most widely used and meets the requirements (Ahmad et al., 2005). But, the performance of the MAT has been associated with some disadvantages. Here, the immunoblotting of whole-cell bacteria is attractive in comparison with the reference standard MAT, because it is simple, inexpensive, less burden and suitable for laboratory diagnosis. Bacterial culture. Twenty six leptospiral serovars, representing twenty serogroups (Table 1), were obtained from the National Leptospirosis Reference Center, National Institute of Health (NIH), Thailand and maintained by weekly subculture at 2830 °C in liquid Difco Leptospira medium base EMJH (Becton Dickinson). Burkholderia psuedomallei (K96243; kindly provided by the NIH, Thailand), Escherichia coli (DH5α; kindly provided by Department of Microbiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Thailand) and Helicobacter pylori (kindly provided by Dr Anuchai Niwetpathomwat, Department of Veterinary Medicine, Faculty of Veterinary Science, Chulalongkorn University, Thailand) were used as the non-leptospiral antigens (negative controls).

Table 1. Representative leptospires among pathogenic and non-pathogenic serogroups, serovars, strains and species used

Rabbit antisera (polyclonal antibodies) and mAbs against Leptospira. Each New Zealand white rabbit (810 weeks old) was immunized with an individual serovar of live leptospires by weekly intravenous injection for 46 weeks, as described elsewhere (Doungchawee et al., 2005; Sitprija et al., 1980). The serovar-specific antisera were then obtained and tested for MAT titre. All experimental procedures with animals were approved by the Animal Research Committee of the National Laboratory Animal Center, Thailand. For mAb, purified murine IgG specific to Bratislava and Bataviae serovars were kindly provided by Dr Pattama Ekpo, Department of Microbiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand.

MAT. MAT was performed according to a modified method (Adler & Faine, 1978). Briefly, 50 µl each antiserum was incubated at room temperature with an equal volume of a suspension of live leptospires (approx. 1x10⁸ cells ml¹) in separate wells of microtitre plates. After 2 h incubation, agglutination in each well was examined under a dark-field microscope (Olympus DP70 BX51; Shinjuku). The test was considered positive when >50 % agglutination was observed, and the most diluted titre with positivity was reported.

SDS-PAGE and immunoblotting. Leptospires were harvested at the mid-exponential phase and approximately 2x10⁷ cells were used for each strain. The bacteria were washed with PBS three times for 5 min each, and then lysed with a standard Laemmli buffer (1x) and heated in boiling water for 5 min. After removal of the remaining particulate matter using microcentrifugation, the supernatant was loaded onto a 12.5 % acrylamide gel. SDS-PAGE was performed in a Hoefer Mighty Small II mini-gel apparatus (Amersham Biosciences) using a constant voltage of 200 V for 1 h (Kelson et al., 1988). After completion, the resolved antigens were transferred onto a 0.45 µm thick polyvinylidene fluoride (PVDF) membrane using a semi-dry system (TE70; Amersham Biosciences) with a constant current density of 1.5 mA cm² for 60 min.

The blotted membrane was washed three times (5 min each) with PBST (PBS with 0.05 %, v/v, Tween 20), and then incubated with 1 : 1000 rabbit antisera or 1 : 300 murine mAbs diluted in 2 % skimmed milk in PBST. The membrane was then washed as above and subsequently incubated, for 45 min, with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulins (1 : 2000; Dakopatt) or sheep anti-mouse IgG (1 : 2500; Dakopatt), respectively. Immunoreactive bands were visualized using 3,3-diaminobenzidine (Sigma-Aldrich) as the substrate. Molecular masses of the immunoreactive bands were estimated using Amersham Bioscience standard protein markers.

This study was designed to identify immunogenic Leptospira protein antigens, valuable in serology, with a high test specificity. The reactivities of rabbit antisera raised against individual serovars were assessed in both MAT and immunoblot assays (Table 2). The sera reacted strongly to the corresponding leptospiral strains, as demonstrated in serovars Pomona, Pyrogenes and Canicola individually by a major smear-like banding (Fig. 1a, b, c; lanes 8, 9 and 6, respectively). Such smear-like banding was not observed when heterologous antisera were used. However, there was a cross-reaction producing the smear-like band when heterologous antisera against members within the same serogroup were used (as shown in Fig. 1b for the cross-reactivity in Zanoni, when anti-Pyrogenes antiserum was used Zanoni and Pyrogenes serovars are in the same serogroup, namely Pyrogenes). Similar findings were also observed for serovars Australis, Bangkok and Bratislava of the Australis serogroup (data not shown). This cross-reactivity among members within the same serogroup suggested that the serovar-specific epitope(s) might be similar within the same serogroup and that these antigenic determinants could be responsible for the agglutination when the MAT typing method was employed. Although there was a cross-reaction among the serovars within the same serogroup, molecular masses of such prominent smear-like bands were distinguishable among different serovars. Estimated molecular masses of the smear-like bands (within the range of 1930 kDa), which were specific for individual serovars, are shown in Table 2. Obviously, the immunoblot pattern obtained with the antisera to pathogenic strains was different from that of non-pathogenic L. biflexa (Fig. 1d), which shared a characteristic pattern of multiple reactive bands, ranging from 10 to 90 kDa. This multi-band pattern, which was characteristic for the pathogenic leptospiral strains, was not observed when the non-pathogenic L. biflexa (Patoc) antigen was probed with homologous or heterologous antiserum (Fig. 1, lane 11).

Table 2. Approximate molecular size of the smear-like band and MAT results of 11 representative antisera

(76K): Fig. 1. Immunoblotting of antigens derived from whole-cell lysates of 11 leptospiral serovars. The membranes were probed with rabbit antisera (pAb) raised against individual serovars. This figure shows representative immunoblots of 11 serovars reacted with anti-Pomona (a), anti-Pyrogenes (b), anti-Canicola (c) and anti-Patoc (d) pAb. The multi-band pattern (~1090 kDa) was unique for the pathogenic Leptospira antigens (lanes 110) and was not observed in the non-pathogenic L. biflexa (Patoc; lane 11). For pathogenic Leptospira species, a prominent smear-like band at approximately 1930 kDa was present when the antigens were probed with the homologous antisera. Proteins of the prominent band, although showing a cross-reaction between members within the same serogroup, had molecular sizes that differed among different serovars. M std, molecular mass standard marker.

Additional testing was carried out with antisera against serovars other than Pomona, Pyrogenes and Canicola, the lower molecular mass protein components at 1420 and the flagella proteins of 3536 kDa were found to be antigenically unique to Leptospira. Fig. 2 illustrates that immunoblotting using serovar-specific mAbs also provided the same prominent smear-like band, consistent with the results obtained from polyclonal antibodies (antisera). Evidence from a number of studies has suggested that the serovar-specific and/or serogroup-specific antigens might be outer-membrane glycolipids and lipopolysaccharides of L. interrogans (Barnett et al., 1999; Brown et al., 1991; Cho et al., 1992; Shinagawa & Yanagawa, 1972). Glycolipid antigens have been suggested to play a major role in immunity, and to contribute to the production of agglutinating and opsonic antibodies (Adachi & Yanagawa, 1977; Farrelly et al., 1987; Jost et al., 1986; Masuzawa et al., 1990; Midwinter et al., 1994). Although, several published reports have described various methods for assessing size variation of the lipopolysaccharide antigens of L. interrogans (Cho et al., 1992; Gitton et al., 1992; Masuzawa et al., 1990; Zuerner et al., 1991), the utility of these tests remains controversial, such as the variable degree of serovar specificity of 2131 kDa antigens determined by SDS-PAGE and ELISA (Cho et al., 1992), the detection of 2126 kDa as serovar-specific or serogroup-specific antigens among seven leptospiral strains by immunoblotting (Gitton et al., 1992), and the identification of 2330 kDa antigens of L. interrogans serovar Canicola with silver stain on SDS-PAGE (Masuzawa et al., 1990).

(42K): Fig. 2. Consistent results obtained from polyclonal antibodies (antisera) and mAbs. The prominent smear-like band (with a molecular mass of approximately 1930 kDa) was also detectable when homologous mAb were used. M std, molecular mass standard marker.

Fig. 3(a) shows that levels of the immunoreactivity were varied by the different amounts (10⁴, 10⁵ and 10⁶ cells per assay) of bacterial antigens used for blotting. We suggest using at least 10⁵ cells per assay to ensure the high quality of results using our method. Fig. 3(b) shows the results on non-leptospiral bacteria, i.e. E. coli, B. pseudomallei and H. pylori, which demonstrated minimal banding compared to that of leptospiral origin, which is indicative of the specificity of our technique for detecting Leptospira. Our immunoblot data show that the characteristics of some of the immunoreactive bands were similar, whereas the others differed from published data obtained using whole-cell and/or outer-envelope protein fractions (Brown et al., 1991; Cho et al., 1992; Gitton et al., 1992). It is likely that several Leptospira protein antigens have been characterized as minor bands on SDS-PAGE and these were clearly recognized when antiserum to a homologous strain was applied (Gitton et al., 1992; Zuerner et al., 1991). In addition, Leptospira species-associated antigens have been recognized and characterized, including flagellar components (35 or 3336 kDa bands) (Chapman et al., 1988; Kelson et al., 1988), outer-membrane proteins and carbohydrate components (14.426.5 kDa bands) (Chapman et al., 1988), outer-membrane-associated antigens (defined as LipL32, LipL36, LipL41 and LipL48) of leptospiral strains (Cullen et al., 2002), a novel 48 kDa outer-membrane lipoprotein (designated LipL48) (Haake & Matsunaga, 2002), and two non-agglutinating protein antigens (p12 and p20), which are conserved for the genus leptospira (Doherty et al., 1989).

(67K): Fig. 3. Feasibility of the assay. (a) Various amounts of Leptospira serovar Canicola (104, 105 and 106 cells per lane) were used for immunoblotting with anti-Canicola pAb (antisera). The degree of immunoreactivity was varied by the amount of antigens used. (b) Negative results were obtained when antigens derived from whole-cell lysates of E. coli, P. pseudomallei and H. pylori were used to react with anti-Canicola pAb. M std, molecular mass standard marker.

In our study, several antigens were predominantly detected in pathogenic leptospires (shown as multiple immunoreactive bands, ranging from 10 to 90 kDa when heterologous antisera were used) that could be used as the markers to discriminate from the non-pathogenic L. biflexa. Immunoblotting allows for the analysis of the immune response to a number of defined antigens and has confirmed that the concept of serovar specificity of Leptospira species is confined to the 1930 kDa epitopes. Characterizations of serovar-specific antigens, i.e. using MS, would be very interesting and deserves further studies. Extending the study to other reference strains and to human isolates may lead to further use of this test in epidemiological survey and/or in clinical diagnosis of leptospirosis as well. This study was supported by a grant from the Department of Communicable Disease Control, Ministry of Public Health, Thailand.