Use of Monoclonal Antibodies to Lipopolysaccharide for Antigenic Analysis of Coxiella burnetii

Coxiella burnetii, an obligate intracellular bacterium, is the etiologic agent of Q fever. It is widely distributed in nature and is responsible for infection in various animals and humans (5, 8). Many C. burnetii strains have been isolated from milk, ticks, and human patients with acute and chronic Q fever. Although all strains so far studied belong to the same serotype, it has become apparent that C. burnetii strains differ in their antigenic (3, 17) and genetic (4, 10, 12, 18, 19) properties. The remarkable antigenic differences among strains are due to lipopolysaccharide (LPS) (3). However, the components that differ among strains have not been elucidated because the LPS ladder-like bands are close together and share many antigenic epitopes (9, 16). In addition, the differences among strains are confounded by an LPS change that occurs during serial passages in eggs or tissue culture—a process called phase variation (8). During phase variation, phase I cells, with full-length LPS O-chains, change to intermediate phases with decreasing LPS O-chain lengths and then to phase II, with truncated LPS. This LPS change is an irreversible transition and is used as one of the criteria for distinguishing the phase state of C. burnetii (11). Recently, we analyzed this LPS change by the use of monoclonal antibodies (MAbs) against LPS O-chains and the LPS outer core (7). In this report, to define the LPS component that differs among C. burnetii strains, the reactions of the MAbs against 21 strains were analyzed. Our results suggest that C. burnetii strains could be divided into two groups immunologically. This conclusion is not confounded by the LPS change that occurs during phase variation.

C. burnetii Nine Mile strain phase I cells and 20 other strains were immunologically compared with respect to their reactivities with 10 MAbs. The name, original source, and plasmid type for each strain are listed in Table 1. Other characteristics of each strain are described elsewhere (6, 16, 17). The samples of strains KAV and PAV used were purified LPSs (2), and the samples of other strains used were prepared from whole-cell lysates by digestion with proteinase K as described previously (6). The MAbs used in this study were previously divided into three groups according to their reactions in Western blots following sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (7). The MAbs of groups I (H5A, H45, and H83) and II (H21), all of which recognize the LPS O-chains of Nine Mile strain phase I, mainly produced ladder-like bands above 20 and 15 kDa, respectively. The MAbs of group III (H73, H76, H78, H91, H99, and H100) recognized the 14-kDa LPS outer core of Nine Mile strain phase I and the Crazy variant (one of the intermediate-phase cells). None of the MAbs reacted with Nine Mile strain phase II LPS (7). SDS-PAGE was carried out by using a 15% separating gel. The LPS profiles were observed by silver staining and Western blotting as described previously (16).

The immunochemical properties of the LPSs of the 21 strains were compared by Western blotting following SDS-PAGE. The MAbs of groups I (H5A, H45, and H83) and II (H21) reacted with all 21 strains. Their reaction patterns were slightly different from those against Nine Mile strain phase I. The MAbs of group III (H73, H76, H78, H91, H99, and H100) did not react with the KAV, PAV, Priscilla, G Q212, Ko Q229, and S Q217 strains but reacted with the other 15 strains with a 14-kDa dense band. The mouse antiserum against Nine Mile strain phase I cells reacted with all strains with ladder-like bands above 14 kDa. Figure 1 shows the reaction patterns of the representative MAbs of groups II (H21) and III (H78) against the El Tayeb, Ohio 314, California 76, Henzerling, Bangui, MAN, ME, Ko Q229, S Q217, and Priscilla strains. The reactions of the MAbs are summarized in Table 2. With silver stain, all 21 strains showed ladder-like LPS profiles (data not shown).

The reactions of the MAbs of group III, which recognized the 14-kDa LPS outer core of the Nine Mile strain, indicate that the LPS outer cores of KAV, PAV, Priscilla, G Q212, Ko Q229, and S Q217 differ antigenically from those of the other 15 strains (Table 2 and Fig. 1B). This antigenic difference is not due to the phenomenon of phase variation because the ladder-like LPS profiles indicate that none of the strains examined are pure phase II cells and are in good agreement with those observed by silver staining in the study by Hackstadt (3). This suggests that the KAV, PAV, Priscilla, G Q212, Ko Q229, and S Q217 strains lack the epitopes recognized by the MAbs of group III. On the other hand, it is unclear whether the difference in the ladder-like banding above 15 kDa is due to the phenomenon of phase variation (Fig. 1A). The 14-kDa LPS outer-core oligosaccharide of the Nine Mile strain contains mainly dihydrohydroxystreptose and galactosaminuronyl-α(1-6)-glucosamine (2) and seems to be the dominant antigenic and immunogenic determinant (1, 13, 14). In C. burnetii, these properties have been analyzed in detail with the Nine Mile and Priscilla strains. These strains differ in their distributions of dihydrohydroxystreptose in LPS (15) and their immunogenic properties and virulence to experimental animals (9). Considering these earlier findings, the antigenic difference detected in this study could be related to the chemical and immunogenic differences between the Nine Mile and Priscilla strains. Furthermore, this antigenic difference might be related to the virulence difference, since host immune responses are associated with the infection of this intracellular bacterium. It is likely that the KAV, PAV, G Q212, Ko Q229 and S Q217 strains also differ from the Nine Mile strain in their chemical and immunological properties and virulence. The six strains differentiated from the other strains have similar genetic properties (QpRS or plasmidless types) and were isolated in the same part of the world (Canada or the United States). Therefore, the serotyping system based on the reactions of the MAbs may be useful in diagnostic and epidemiologic investigations of Q fever. Using this serotyping system would allow C. burnetii strains to be differentiated into two antigenic groups without confusing the phenomenon of phase variation. Further studies of the LPS outer core in a large number of strains from various sources may help us to understand the differences in immunogenic properties and virulence among C. burnetii strains.