Anti-Salmonella lipopolysaccharide monoclonal antibodies: characterization of Salmonella BO-, CO-, DO-, and EO-specific clones and their diagnostic usefulness.

To facilitate the identification and serotyping of Salmonella species, we established a wide variety of murine monoclonal antibodies (MAbs) that were reactive with the lipopolysaccharides (LPSs) of Salmonella serogroups B to E. An effective approach for generating LPS-reactive hybridomas was used; this required immunization of mice with LPS-coated bacteria. To screen for diagnostically useful MAbs, the MAbs were tested by enzyme-linked immunosorbent assay against a set of purified LPSs from smooth and rough Salmonella strains. At least four major groups of antibody specificities were identified: Salmonella (i) BO specific, (ii) CO specific, (iii) DO specific, and (iv) EO specific. For a more detailed epitope analysis, a panel of eight different serogroup-specific MAbs which were shown to bind the O-antigenic polysaccharide chains, yielding characteristic ladder patterns in Western blots (immunoblots) against the LPS of Salmonella serogroups B to E, were selected. The availability of various chemically defined LPS structures and Salmonella O-antigen glycoconjugates permitted the definition of O-antigenic polysaccharide epitopes recognized by each MAb that serologically corresponded to factors O3, O4, O5, O6, O7, O8, O9, and O10 on the basis of the Kauffmann-White scheme for Salmonella classification. The diagnostic accuracy of these immunochemically defined O-specific MAbs for Salmonella serotyping was demonstrated by correct identification of all 167 salmonellae (including 72 serotypes from serogroups B to E) among the 294 bacterial strains in a slide agglutination test. No false-positive reactions were detected.     

Binding studies of a monoclonal antibody specific for 3-deoxy-D-manno-octulosonic acid with a panel of Klebsiella pneumoniae lipopolysaccharides representing all of the O serotypes.

A monoclonal antibody (MAb) raised against Salmonella minnesota R595 and specific for alpha-3-deoxy-D-manno-octulosonic acid (alpha-Kdo) of the inner core was tested for binding to lipopolysaccharides (LPS) of Klebsiella pneumoniae. The MAb was tested in several assay systems (enzyme-linked immunosorbent assay, passive hemolysis, and inhibition of passive hemolysis) with a large panel (n = 23) of K. pneumoniae LPS representing all nine currently known O serotypes. MAb 20 showed reactivity with almost all O serotypes of K. pneumoniae LPS, and this reactivity could be inhibited by synthetic Kdo. This suggests an epitope in the cores of these Klebsiella LPS much like that in the inner core of LPS of S. minnesota. Large differences in reactivity between LPS of different strains belonging to the same O serotype were observed. After sodium dodecyl sulfate-polyacrylamide gel electrophoresis of LPS followed by immunoblotting, reactivity of MAb 20 was observed only with the fast-moving fraction possibly representing the nonsubstituted core. No binding was seen with the high-molecular-weight fraction that contained core material substituted with several units of O-antigen building blocks. The chemical basis for these differences in reactivity remains to be established. As far as we know, this is the first report containing comprehensive immunochemical data on the LPS core of K. pneumoniae.     

A murine monoclonal antibody specific for the outer core oligosaccharide of Salmonella lipopolysaccharide.

Immunoglobulin G3 murine monoclonal antibody T6 specific for the lipopolysaccharide of Salmonella O serogroups A to E was established. By using R mutants of Salmonella spp., Escherichia coli, and Shigella spp., the major reactive epitope with T6 was tentatively identified as the terminal disaccharide, N-acetylglucosamine 1.2----alpha glucose, of the core oligosaccharide. T6 was reactive with 10 clinical isolates of each of the Salmonella O serogroups A to E but not with 58 isolates of other gram-negative bacteria. Its selective reactivity against Salmonella spp. renders T6 a potentially more useful reagent than the conventional polyvalent serum for the identification of Salmonella spp. It may also serve as a useful molecular tool for the study of the outer core structure of all Salmonella and related species.     

Blind comparison of traditional serotyping with three multiplex PCRs for the identification of Salmonella serotypes

Salmonella serotypes are defined on the basis of somatic (O) antigens which define the serogroup and flagellar (H) factor antigens, both of which are present in the cell wall of Salmonella. Most Salmonella organisms alternatively express phase-1 or phase-2 flagellar antigens encoded by fliC and fljB genes, respectively. Our group previously published two multiplex PCRs for distinguishing the most common first- and second-phase antigens. In this paper we describe a third multiplex PCR to identify the most common serogroups (O:B; O:C1; O:C2; O:D and O:E). The combination of these three PCRs enabled us to completely serotype organisms belonging to the Salmonella species. This multiplex PCR includes 10 primers. A total of 67 Salmonella strains belonging to 32 different serotypes were tested. Each strain generated one serogroup-specific fragment ranging between 162 and 615bp. Twenty-eight strains belonging to 21 serotypes, with a serogroup different from those tested in this work, did not generate any fragments. To compare molecular serotyping with traditional serotyping, 500 strains, received according to the order of arrival in the laboratory, were serotyped using both methods. The three multiplex PCRs were able to serotype 84.6% of the tested strains. This method was found to be very helpful in our laboratory as an alternative method for typing strains causing outbreaks, and it can be used to supplement conventional serotyping, since it is also applicable to motionless and rough strains.     

A monoclonal antibody reactive with a common epitope of Moraxella (Branhamella) catarrhalis lipopolysaccharides.

A hybrid cell line producing a monoclonal antibody (MAb) against Moraxella (Branhamella) catarrhalis lipopolysaccharide (LPS) was established. The specificity of the MAb 1B12 to purified rough LPSs from six strains of M. catarrhalis was ascertained by enzyme-linked immunosorbent assay (ELISA), competitive-inhibition ELISA, and immunoblotting. MAb 1B12 bound to live bacterial cells and culture supernatants from a total of 34 strains of M. catarrhalis, including 12 strains with different LPS serotypes. No cross-reactions with smooth and rough LPSs from selected enterobacterial and nonenterobacterial strains, with other respiratory pathogens, or with Neisseria species were observed. These data suggest that MAb 1B12 recognizes a common epitope of M. catarrhalis LPS which differs from serotype determinants.     

Characterisation and application of a murine monoclonal antibody specific for the serogroup C2 Salmonella

An IgG3 murine monoclonal antibody (designated MO8) specific for the serogroup C2 Salmonella lipopolysaccharide (LPS) was generated by fusing mouse myeloma cells NS1 with spleen cells of BALB/c mice immunised with heat-killed S. manhattan. MO8 reacted with purified LPS prepared from serogroup C2 Salmonella but did not react with that prepared from other O serogroups, and its reactivity was also specifically absorbed by serogroup C2 Salmonella only. Polyacrylamide gel electrophoresis of the serogroup C2 LPS and subsequent immunoblotting with MO8 yielded multiple reactive bands giving a characteristic ladder pattern. The specificity of MO8 was further demonstrated in the slide agglutination test with 223 bacteria, of which only 25 belonging to serogroup C2 Salmonella reacted with the MO8 ascitic fluid. The specificity of MO8 makes it useful not only for the serological identification of Salmonella but also for the epitope analysis of the serogroup C2 LPS.     

A murine monoclonal antibody defines a unique epitope shared by Klebsiella lipopolysaccharides.

A hybridoma secreting a monoclonal antibody (MAb) directed against Klebsiella lipopolysaccharide (LPS) was derived from spleen cells of mice immunized a smooth, nonencapsulated Klebsiella strain (Friedländer 201; serogroup O1). The MAb, called V/9-5 (immunoglobulin G2a), cross-reacted with LPS preparations produced from reference strains for the Klebsiella O serogroups O1, O2ab, O2ac, O3, O4, O5, and O12. Furthermore, the MAb reacted with LPSs from serogroup reference strains O6/O8, O9, and O11, which are regarded as being identical to O1, O2, and O4, respectively. When testing the supernatant of clinically isolated Klebsiella strains by means of an inhibition enzyme-linked immunosorbent assay, we found that 86 (92.4%) of 93 Klebsiella pneumoniae subsp. pneumoniae isolates and 24 (96.0%) of 25 K. oxytoca isolates harbored the cross-reactive epitope. By contrast, two laboratory strains of K. pneumoniae subsp. rhinoscleromatis did not react with MAb V/9-5. The MAb proved to be specific for the genus Klebsiella, since it did not react with any of a total of 73 strains belonging to other gram-negative bacterial genera. In conjunction with other LPS-specific MAbs, MAb V/9-5 might become a useful reagent for rapid identification of klebsiellae in clinical specimens. Furthermore, the epitope recognized by MAb V/9-5 might serve as a target epitope for the production of human MAbs for immunotherapeutic purposes.     

Demonstration of the Rb1 lipopolysaccharide core structure in Salmonella strains with the monoclonal antibody M105

The lipopolysaccharide (LPS) from 42 strains representing 19 Salmonella serogroups was differentiated into characteristic ladder-like profiles by SDS-PAGE analysis. The core-specific antibody M105 (Ra, Rb1 and Rb2) was used in an immunoblot assay of SDS-PAGE-separated LPS molecules. The M105 antibody bound to the R-type LPS of 18 of the 20 Salmonella strains tested. The results demonstrate that S. enterica serotype Godesberg, S. Adelaide (one of two strains), S. Milwaukee, S. Niarembe, S. Bere and S. Arizonae (serogroup 63) have an atypical LPS core structure which is Rb1 type.     

Production and characterization of monoclonal antibodies specific for the lipopolysaccharide of Escherichia coli O157.

Identification of the O157 antigen is an essential part of the detection of Escherichia coli O157:H7, which is recognized as a major etiologic agent of hemorrhagic colitis. However, polyclonal antibodies produced against E. coli O157:H7 lipopolysaccharide (LPS) may react with several other bacteria including Brucella abortus, Brucella melitensis, Yersinia enterocolitica O9, Escherichia hermannii, and Stenotrophomonas maltophilia. We produced eight monoclonal antibodies (MAbs) specific for the LPS of E. coli O157. Western blots (immunoblots) of both the phenol phase (smooth) and the aqueous phase (rough) of hot phenol-water-purified LPS indicated that three of the MAbs were specific for the O antigen and five were reactive with the LPS core. The eight MAbs could be further differentiated by their reactivities to Salmonella O30 LPS (group N), which is reported to be identical to the E. coli O157 antigen. All eight MAbs reacted strongly to all of the 64 strains of E. coli O157 tested, which included 47 isolates of O157:H7 and 17 other O157 strains. None of the eight MAbs cross-reacted with any of the 38 other E. coli serotypes tested, which consisted of 29 different O-antigen serotypes, or with 38 strains (22 genera) of non-E. coli gram-negative enteric bacteria.     

Hemagglutination activity and colonization factor antigens I and II in enterotoxigenic and non-enterotoxigenic strains of Escherichia coli isolated from humans.

We examined 205 enterotoxigenic strains of Escherichia coli for colonization factor antigens (CFA) I and II, using an immunodiffusion technique with specific antisera. A total of 36 strains of serogroups O63, O78, O114, O128, and O153 and 1 rough strain possessed CFA/I and gave a single precipitin line; 47 strains of serogroups O6, O8, O80, and O115 possessed CFA/II. The latter strains gave a major precipitin line (component 3) when tested with specific antisera prepared against strain E1392 or PB-176 (both E. coli O6.H16; biotype A). However, all 16 strains of E. coli O6.H16 belonging to biotype A gave a second precipitin line (component 1) when tested with both antisera. When CFA/II-positive strains were tested with a specific antiserum prepared against E. coli O6.H16 strains of biotype B or C, all strains gave component 3, but 16 of 17 strains of E. coli O6.H16 belonging to biotype B, C, or F gave a second precipitin line (component 2) not given by strains of biotype A. CFA/II-positive strains of serogroups other than O6 gave only component 3 in tests with all specific antisera. Nine enterotoxigenic strains of serotypes O7, O15, O25, O115, and O128 gave mannose-resistant hemagglutination of human or calf erythrocytes but lacked CFA/I or CFA/II. Although mannose-resistant hemagglutination was common in non-enterotoxigenic strains of E. coli, none of the non-enterotoxigenic strains possessed CFA/I or CFA/II; these strains included fecal strains of serogroups O6, O8, O63, and O78, fecal strains of enteropathogenic serogroups, and strains from extraintestinal sources.     

Serotypes of enterotoxigenic Escherichia coli in Thailand and the Philippines.

The serotypes of 386 enterotoxigenic Escherichia coli (ETEC) isolated from 82 individuals with and without diarrhea in Thailand and the Philippines were determined. The 136 strains producing both heat-labile toxin (LT) and heat-stable toxin (ST) belonged to 12 different O serogroups; however, 83% (113/136) were of one of four serogroups (O6, O8, O25, and O78), and 76% of (104/136) belonged to one of seven O:K:H serotypes. Only 14% (28/196) of LT-only-producing ETEC belonged to serogroups most common among LT and ST strains, and these 196 strains belonged to 35 different O:K:H serotypes. Three O serogroups (O20, O27, and O78) accounted for 94% (52/54) of strains producing only ST. Although only 4% (2/54) of ST-only ETEC belonged to the seven serotypes most commonly found among strains which produced LT and ST, 85% of ETEC belonged to three other serotypes, O20:K?:H21, O27:K?:H7, and O78:H-. A total of 46% (37/80) of ETEC of serotypes O6:H16, O8:H9, O25:H42, and O78:H12 were resistant to two or more antibiotics in comparison to 68% (208/306) of ETEC of other serotypes (P less than 0.001). In Thailand and the Philippines, E. coli which produced LT and ST or ST alone, but not those which produced LT alone, were restricted in their O:K:H serotypes.     

Structures and sugar compositions of lipopolysaccharides isolated from seven Actinobacillus pleuropneumoniae serotypes.

Highly purified lipopolysaccharide (LPS) preparations obtained from seven Actinobacillus pleuropneumoniae strains representative of seven different serotypes were used to determine the structure and monosaccharide composition of the polysaccharide components of each lipopolysaccharide. An indication of the structure of each LPS was obtained by procedures that included sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by silver staining and gel chromatographic fractionation of acetic acid-hydrolyzed LPS. The polysaccharide components of the LPSs were analyzed by gas-liquid chromatography. The LPSs of the strains of serotypes 2, 4, and 7 were of the smooth type, and those of the strains of serotypes 3 and 6 were of the rough type; the LPSs of the strains of serotypes 1 and 5 could be considered semirough. Rhamnose was present only in the O polysaccharide of the smooth-type and semirough-type LPSs, whereas galactose was present only in the O polysaccharide of the smooth-type LPS and in the core oligosaccharides of the rough-type and semirough-type LPSs. Glucoheptose and mannoheptose were present in the core oligosaccharides of all the LPSs except for the strain of serotype 3, in which only mannoheptose was detected. N-Acetylglucosamine was detected only in the O polysaccharides of the strains of serotypes 1 and 5.     

Improved O-serotyping method for Serratia marcescens.

In a previous study, we found that some O serotypes of Serratia marcescens, as defined by agglutination tests, were not based on lipopolysaccharide (LPS) O antigens. We developed a dot enzyme immunoassay with a high degree of LPS specificity and tested 104 distinct clinical strains. Only 7 of the 24 existing O antigens were found in more than one strain: O12/O14 (30.8% of strains examined), O21 (12.5%), O8 (8.7%), O6/O7 (5.8%), O4 (3.8%), O18 (2.9%), and O9 (2.9%). Two new antigens, S1254 (13.5%) and S3255 (3.8%), were also found. Agglutination tests with O antisera identified the LPS antigen in only 36 strains. Prodigiosin production was restricted to serotypes O8, O6, and S3255 and strains with a rough or semirough LPS phenotype. Dot immunoassay appears to offer greater accuracy than agglutination tests for serotype identification in S. marcescens.     

Examination of strains belonging to enteropathogenic Escherichia coli serogroups for genes encoding EPEC adherence factor and Vero cytotoxins

Four hundred and forty-nine strains isolated from patients with diarrhoea and belonging to 13 enteropathogenic Escherichia coli (EPEC) O serogroups were tested with a DNA probe for the EPEC adherence factor (EAF). Positive results were obtained with only 36 strains; they belonged to 10 O serogroups and flagellar typing showed they were usually of the "classical" EPEC serotypes. Thirty-four of the 36 EAF-positive strains showed localised adhesion to HEp-2 cells. The two remaining strains, of serotypes O114:H2 and O127:H4, showed low level or no adhesion to HEp-2 cells. No colonies hybridising with the EAF probe were identified in cultures from 115 faecal specimens from healthy children. Sixteen of the 449 strains hybridised with one or both probes for the Vero cytotoxin genes VT1 and VT2; 15 of the 16 strains belonged to serogroups O26 and O128. None of the strains hybridised with both the EAF and VT gene probes. These studies show that the great majority of strains belonging to EPEC O serogroups do not possess the EPEC adherence factor or carry VT genes.     

Comparisons of Pasteurella multocida lipopolysaccharides by sodium dodecyl sulfate-polyacrylamide gel electrophoresis to determine relationship between group B and E hemorrhagic septicemia strains and serologically related group A strains.

Lipopolysaccharides (LPSs) purified from 16 reference somatic serotypes of Pasteurella multocida were examined and compared by discontinuous sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Resolution of LPS patterns in a gel was optimum when sample wells were cast separately from the stacking gel and the running gel consisted of 15% T (total monomer) polyacrylamide and 4 M deionized urea. Band patterns of P. multocida LPSs in a gel differed from control Salmonella minnesota wild-type and core mutant LPSs. Although the band patterns and mobilities of LPSs from some P. multocida reference serotypes were similar, none were identical. Evidence for O antigens similar to those produced by enterobacteria was not observed. Proteinase K digestion of whole P. multocida cells resulted in LPS band patterns similar to those of purified LPS. The presence or absence of a capsule on a strain had no major influence on band patterns in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Comparisons of LPS patterns of group B and E hemorrhagic septicemia strains with those of serologically related group A strains of P. multocida indicated that they were similar. Typing antisera made with purified serotype 2 or 5 LPS reacted with electroblots of all these strains. However, the reactions did not distinguish strains as being serotype 2 or 5.     

Structural and serological specificities of Pasteurella haemolytica lipopolysaccharides.

Lipopolysaccharides (LPSs) from 16 serotypes of Pasteurella haemolytica were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and examined by silver staining and immunoblotting. Silver staining of proteinase K-digested cell lysates revealed two rough LPS serotypes (serotypes 2 and 8), which lacked demonstrable O-polysaccharide, while 14 others demonstrated a ladder pattern characteristic of smooth-type LPS. Purified LPSs from several serotypes yielded O-polysaccharide in addition to low-molecular-weight core oligosaccharide components when subjected to mild acid hydrolysis. Nuclear magnetic resonance spectroscopy revealed the O-chain polysaccharides of serotypes 1, 6, and 9 to be identical. Immunoblots using hyperimmune rabbit, mouse, bovine, and ovine sera from homologous and heterologous serotypes supported this finding and suggested that most of the A biotypes share common O-chain epitopes. Immunoblotting results also supported structural data which demonstrated that the O-polysaccharides of serotypes 3 and 15 and of serotypes 4 and 10 (T biotypes) are identical. Nuclear magnetic resonance analysis indicated that the core oligosaccharides of serotypes 1, 6, 8, 9, and 12 share similar structures, but that they are distinct from those of serotypes 3, 4, 10, and 15. Immunoblots with hyperimmune antisera and monoclonal antibody having specificity for the core region of serotype 1 LPS revealed shared epitopes in the core oligosaccharides of several A biotypes. Characterization of the molecular structure and antigenic specificities of LPS has been an important consideration in the development of purity and potency assays for veterinary vaccines which contain P. haemolytica.     

Specificity of rabbit antisera against the rough lipopolysaccharide of Salmonella minnesota R4 (chemotype Rd2P-)

Rabbit polyclonal antibodies against the rough mutant lipopolysaccharide (LPS) of Salmonella minnesota R4 (chemotype Rd2P-) were serologically characterized by using R4 LPS, deacylated LPS, dephosphorylated LPS, and synthetic partial structures, including compounds comprising the core region of Rd2P- LPS bound to the beta 1-->6-linked glucosamine disaccharide with two amide-linked 3-hydroxytetradecanoic acid residues or coupled to bovine serum albumin. By using a passive hemolysis assay and an enzyme immunoassay and absorption and inhibition experiments, the antibody specificities present could be determined. One group of antibodies required components of the core oligosaccharide (with or without the side chain 3-deoxy-D-manno-octulosonic acid [Kdo]) and the phosphorylated glucosamine disaccharide of the lipid A moiety for binding. The phosphate-independent antibodies were directed against the core oligosaccharide, recognizing an epitope consisting of one terminal heptose linked to Kdo or to the reducing moiety of the alpha 2-->4-linked Kdo disaccharide. Antibodies requiring the presence of acyl residues and those reacting with a single heptose or Kdo residue were not detected.     

A high-molecular-weight fraction of smooth lipopolysaccharide in Klebsiella serotype O1:K20 contains a unique O-antigen epitope and determines resistance to nonspecific serum killing

The lipopolysaccharide (LPS) O-antigen side chains of Klebsiella serotype O1 have been studied by using mutants selected by resistance to a Klebsiella bacteriophage designated O1-A. Two classes of LPS mutants were identified. The major group (90%) synthesized rough LPS. The remaining 10% of the mutants produced a novel LPS profile that lacked the highest-molecular-weight O-substituted molecules (HMW-LPS) but still produced lower-molecular-weight O-substituted species (LMW-LPS). By using antisera raised against mutant Klebsiella strains and antiserum specific for Pasteurella haemolytica serotype 4, it was demonstrated that HMW-LPS and LMW-LPS contain shared epitopes. HMW-LPS also contained an epitope absent in LMW-LPS. This unique epitope was recognized by a monoclonal antibody (O1-52.6) and appears to be responsible for the serological cross-reaction between the O antigens of Klebsiella O1 and Escherichia coli O19. This HMW-LPS epitope was present in eight other Klebsiella O1 isolates which were examined. Electron microscopy demonstrated that HMW-LPS excluded overlying capsular polysaccharide for a distance of 25 to 40 nm. The distance was reduced to 10 to 18 nm in strains which synthesized only LMW-LPS and to zero in rough LPS strains. The HMW-LPS of Klebsiella O1 was shown to be an important virulence determinant, since this molecule was responsible for the resistance of the bacterium to nonspecific, complement-mediated serum killing.     

Use of biochemical kinetic data to determine strain relatedness among Salmonella enterica subsp. enterica isolates

Classical biotyping characterizes strains by creating biotype profiles that consider only positive and negative results for a predefined set of biochemical tests. This method allows Salmonella subspecies to be distinguished but does not allow serotypes and phage types to be distinguished. The objective of this study was to determine the relatedness of isolates belonging to distinct Salmonella enterica subsp. enterica serotypes by using a refined biotyping process that considers the kinetics at which biochemical reactions take place. Using a Vitek GNI+ card for the identification of gram-negative organisms, we determined the biochemical kinetic reactions (28 biochemical tests) of 135 Salmonella enterica subsp. enterica strains of pig origin collected in Spain from 1997 to 2002 (59 Salmonella serotype Typhimurium strains, 25 Salmonella serotype Typhimurium monophasic variant strains, 25 Salmonella serotype Anatum strains, 12 Salmonella serotype Tilburg strains, 7 Salmonella serotype Virchow strains, 6 Salmonella serotype Choleraesuis strains, and 1 Salmonella enterica serotype 4,5,12:-:- strain). The results were expressed as the colorimetric and turbidimetric changes (in percent) and were used to enhance the classical biotype profile by adding kinetic categories. A hierarchical cluster analysis was performed by using the enhanced profiles and resulted in 14 clusters. Six major clusters grouped 94% of all isolates with a similarity of > or =95% within any given cluster, and eight clusters contained a single isolate. The six major clusters grouped not only serotypes of the same type but also phenotypic serotype variations into individual clusters. This suggests that metabolic kinetic reaction data from the biochemical tests commonly used for classic Salmonella enterica subsp. enterica biotyping can possibly be used to determine the relatedness between isolates in an easy and timely manner.     

Electrophoretic and serological characterization of the lipopolysaccharides of Legionella pneumophila.

Lipopolysaccharide (LPS) is a major constituent of the outer membrane of gram-negative bacteria. We used sodium dodecyl sulfate-polyacrylamide gel electrophoresis of proteinase-K-digested cell lysates to provide preliminary data on the LPS chemotypes for 20 strains of Legionella pneumophila (serogroups 1 to 8). The profiles of all strains except Chicago 2 (serogroup 6) were similar in the number, spacing, and size distribution of the bands visualized on silver-stained gels and were indicative of smooth LPS. However, compared with the bands from Salmonella minnesota smooth LPS, their banding pattern was much tighter, with three to four legionella bands for every salmonella band. The proteinase K digest of Chicago 2 was unique in that only two widely separated silver-stained bands were seen. LPS profiles of 10 serogroup 1 strains were identical, and the profile of Knoxville 1 was not altered by extensive in vitro passage. We used immunoblotting to investigate the serological specificities of the LPSs. When a rabbit antiserum prepared against a serogroup 1 strain was used to probe nitrocellulose sheets that bound LPS from strains belonging to eight different serogroups, it recognized only the LPS from the homologous serogroup. Similar results were observed with serogroup 2, 4, and 6 antisera. Our data indicate that L. pneumophila has a smooth-type LPS with an unusual banding pattern and that it is a serogroup-specific antigen.