Production of monoclonal antibodies to Legionella pneumophila serogroups 1 and 6

To better define the surface antigens of Legionella pneumophila for clinical and experimental purposes, we have produced monoclonal antibodies to L. pneumophila serogroups 1 and 6. Two hybridomas were produced in serogroup 1. One antibody, LP-I-17, recognized a serogroup-common antigen. The second antibody, LP-I-81, was specific for serogroup 1. This antibody was able to agglutinate bacterial cells belonging to the serogroup 1 reference strains. Philadelphia and Knoxville. Microagglutination assays of environmental and clinical isolates revealed a subgroup of serogroup 1 environmental isolates which were not agglutinated by LP-I-81. This subset of isolates was segregated to certain buildings in the medical complex. Immunodiffusion studies showed identity between the LP-I-81 antigen and the serogroup-specific antigen of serogroup 1 organisms. This antigen could be absorbed out of the serogroup 1 organism extract with LP-I-81-coated Staphylococcus aureus, leaving the serogroup-common antigens. Three hybridomas were produced to serogroup 6. All three produced antibodies which were serogroup 6 specific and agglutinated serogroup 6 bacteria.     

Outer membrane proteins from Legionella pneumophila serogroups and other Legionella species.

Outer membranes were isolated from eight serogroups of L. pneumophila and five other Legionella species. The protein composition of the membranes was characterized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A single, disulfide stabilized protein with a molecular size of 29,000 to 30,000 daltons was found to be the major outer membrane protein (MOMP) of all the serogroups. The equivalent of the L. pneumophila MOMP was not observed in any of the other Legionella species examined. Silver staining of sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels revealed distinctive patterns for each serogroup and other Legionella species that were not observed by staining with Coomassie blue and may result from the presence of lipopolysaccharide in the membrane preparations. The MOMP from serogroup 1 was isolated by exposing crude peptidoglycan to detergent in the presence of heat and reducing agent and was found to be tightly associated with lipopolysaccharide. Antibodies to this complex were used to probe the outer membranes of the remaining, L. pneumophila serogroups and other Legionella species by Western blotting. Serogroup 1 anti-MOMP antibodies were found to react with the MOMP from the remaining seven serogroups examined, whereas antibodies directed against the lipopolysaccharide of serogroup 1 only reacted with lipopolysaccharide from two of the remaining seven serogroups.     

Detection of cell-associated or soluble antigens of Legionella pneumophila serogroups 1 to 6, Legionella bozemanii, Legionella dumoffii, Legionella gormanii, and Legionella micdadei by staphylococcal coagglutination tests.

Current methods used for the detection of whole-cell isolates of Legionella or for the detection of Legionella soluble antigens are technically impractical for many clinical laboratories. The purpose of this study was to explore practical alternatives. The results showed that whole cell isolates of Legionella pneumophila serogroups 1 to 6, Legionella bozemanii, Legionella dumoffii, Legionella gormanii, and Legionella micdadei were identified specifically by a simple slide agglutination test or slide coagglutination test in which the reagent antisera are first bound to staphylococcal protein A. Soluble antigens were also identified specifically by the slide coagglutination test and by a sandwich immunofluorescence assay. The latter test may be useful in detecting antigen in body fluids of patients with legionellosis or in environmental samples.     

Evaluation of an indirect hemagglutination test for Legionella pneumophila serogroups 1 to 4

Parallel testing of 895 sera by indirect hemagglutination and indirect fluorescent-antibody techniques showed 97.3% agreement. Although the indirect hemagglutination technique usually showed more cross-reactivity among serogroups than the indirect fluorescent-antibody technique with Formalin-fixed antigens and a conjugate which detected primarily immunoglobulin G antibodies, heterologous serogroup reactions were significantly lower than homologous serogroup titers and the etiological serogroup could be easily defined. The indirect hemagglutination techniques showed no cross-reactivity with a crude extract of Escherichia coli O13:K92:H4. Since the indirect hemagglutination technique was shown to detect both immunoglobulin M and immunoglobulin G antibodies and was found to be rapid, simple, and inexpensive, it appears to be an excellent alternative to the indirect fluorescent-antibody test for serodiagnosis of legionellosis.     

Identification of protein antigens of Legionella pneumophila serogroup 1.

Growth of Legionella pneumophila serogroup 1 in yeast extract broth was standardized, and protein profiles of detergent-solubilized cells were determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Over 30 protein bands were identified, 6 of which were more prominent both in Coomassie brilliant blue-stained profiles and in fluorograms with intrinsically radiolabeled bacteria. These major proteins were 22,000 dalton (22K), 24K, 43K, 63K, 76K, and 78K. We found that the 24K and 63K major proteins were antigenic, as demonstrated both by radioimmunoprecipitation (RIP) of [35S]methionine-labeled organisms and by immunoblotting with rabbit hyperimmune sera. In addition, both techniques detected antigens migrating at 58K, 72K, 76K, and 78K. The major 22K and 43K major proteins and antigens migrating at 25.5K, 29K, and 46K were only detected by radioimmunoprecipitation, whereas antigens of 19K, 48K, 53K, and 68K were detected only by immunoblotting. Cell-surface localization of the proteins was determined by a modified radioimmunoprecipitation assay designed to react specifically with surface antigens and by the use of hyperimmune sera that had been extensively preabsorbed with intact cells to deplete the sera of antibodies directed against surface components. The 22K, 24K, 43K, 63K, and 78K major proteins and several minor proteins were found to be located on the surface of L. pneumophila cells.     

Crossed immunoelectrophoretic analysis of Legionella pneumophila serogroup 1 antigens.

By crossed immunoelectrophoresis, 85 different antigens were demonstrated in sonicated preparations of Legionella pneumophila serogroup 1 (Lp1). The precipitin patterns of 82 anodic-migrating antigens were numbered and were designated the Lp1 reference system. Eleven antigens were stable to boiling, and seven of these were shown to be surface antigens. One heat-stable surface antigen (antigen no. 61) was highly reactive with limulus amoebocyte lysates and formed a precipitin resembling lipopolysaccharide. Serum from an isolation confirmed case of Lp1 infection and serogroup-specific rabbit antiserum reacted specifically with antigen no. 61, which was designated the serogroup-specific antigen. Normal human and rabbit sera commonly had antibodies to antigen no. 66 of the Lp1 reference system. This antigen is antigenically related to the "common antigen" of Pseudomonas aeruginosa.     

Serogroup specificity of Legionella pneumophila is related to lipopolysaccharide characteristics.

We studied the lipopolysaccharide (LPS) of Legionella pneumophila and six other Legionella species to determine whether strain differences were apparent. The LPS was purified by a cold ethanol extraction procedure, and total carbohydrates represented 10 to 20% of LPS weight. 2-keto-3-deoxyoctonate represented 1 to 13% of the total carbohydrate present in the LPS. By sodium dodecyl sulfate-polyacrylamide gel electrophoresis, all strains except L. dumoffi showed smooth-type LPS with multiple high-molecular-weight complexes. Proteinase K-treated, whole-cell lysates showed profiles similar to those of purified LPS. Each serogroup of L. pneumophila and each Legionella species had a distinct sodium dodecyl sulfate-polyacrylamide gel electrophoresis profile. L. pneumophila lipid A is antigenically related to the lipid A of Enterobacteriaceae. In immunoblot assays with the LPS of L. pneumophila serogroups 1 to 6 as antigens, serogroup-specific immune monkey sera recognized homologous purified LPS, but not the LPS of the five heterologous serogroups. These studies indicate that LPS composition may be a determinant of serogroup specificity as defined by the immunofluorescence-based serogrouping schema for L. pneumophila and other Legionella species.     

Legionella pneumophila lipopolysaccharide activates the classical complement pathway.

Legionella pneumophila is a gram-negative bacterium capable of entering and growing in alveolar macrophages and monocytes. Complement and complement receptors are important in the uptake of L. pneumophila by human mononuclear phagocytes. The surface molecules of L. pneumophila that activate the complement system are unknown. To identify these factors, we investigated the effects of L. pneumophila lipopolysaccharide (LPS) on the classical and alternative complement pathways of normal human serum by functional hemolytic assays. Although incubation of LPS in normal human serum at 37 degrees C resulted in the activation of both pathways, complement activation proceeded primarily through the classical pathway. Activation of the classical pathway by LPS was dependent on natural antibodies of the immunoglobulin M class that were present in various quantities in sera from different normal individuals but were absent in an immunoglobulin-deficient serum obtained from an agammaglobulinemic patient. Additional studies using sheep erythrocytes coated with LPS suggested that the antibodies recognized antigenic sites in the carbohydrate portion of LPS. The ability of LPS to interact with the complement system suggests a role for LPS in the uptake of L. pneumophila by mononuclear phagocytes.     

Background prevalence of microagglutination antibodies to Legionella pneumophila serogroups 1, 2, 3, and 4.

The background prevalence of microagglutination antibodies to Legionella pneumophila was determined by testing the sera of 517 individuals who lived or worked in a small Iowa town. In this population, the upper limit of normal microagglutination titer for serogroups 1, 3, and 4 was 1:16, and that for serogroup 2 was 1:8. The prevalence of microagglutination titers of greater than or equal to 1:32 against any serogroup of L. pneumophila was only 7.4% and did not vary significantly with age or sex.     

Characteristics of anti-Legionella antibodies in patients infected with Legionella pneumophila serogroups 1, 6, and 10.

Nosocomial infections with Legionella pneumophila serogroups 1 and 10 in the Leiden University Hospital and infections with L. pneumophila serogroup 6 in neighboring hospitals gave us an opportunity to study the development of opsonizing antibodies against L. pneumophila serogroups 1, 6, and 10 in the serum of 13 patients. Seven of these patients were infected with L. pneumophila serogroup 1, two were infected with serogroup 6, and four were infected with serogroup 10. The opsonic cross-reactivity of antibodies against these serogroups of L. pneumophila and complement involvement in opsonization were also investigated. Convalescent-phase sera from patients infected with L. pneumophila serogroup 1 or 6 were able to promote ingestion of these serogroups by polymorphonuclear leukocytes, whereas ingestion of L. pneumophila serogroup 10 was enhanced only in the presence of convalescent-phase sera from patients infected with this serogroup. Opsonization of L. pneumophila serogroups 1, 6, and 10 was complement dependent.     

Cloning and expression of Legionella pneumophila antigens in Escherichia coli.

To isolate and characterize Legionella pneumophila antigens, we constructed a genomic library of L. pneumophila serogroup 1 (strain 130b). L, pneumophila DNA fragments (2.5 to 7.5 megadaltons) obtained by partial digestion with Sau 3A endonuclease and size fractionation on a sucrose density gradient were inserted into the dephosphorylated BamHI site of vector pBR322; CaCl2-treated Escherichia coli cells of strain HB101 were transformed with hybrid plasmids. To detect expression of antigens, 2,559 ampicillin-resistant transformants were transferred to nitrocellulose paper, lysed in situ, and screened by enzyme immunoassay (EIA) with E. coli-absorbed rabbit anti-L. pneumophila sera. A total of 77 (3%) of the colonies were reactive by EIA; 31 (1.2%) were strongly reactive, and 6 were strongly reactive by EIA without colony lysis. Analysis of 29 stable, strongly reactive clones by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electroblotting showed antigenic bands in 18 clones by EIA with E. coli-absorbed antisera. Absorption of antisera with heat- and Formalin-killed L. pneumophila antigen eliminated or diminished the reactivity of the antigenic bands in representative clones. These studies confirm that several L. pneumophila antigens can be cloned and expressed in E. coli.     

Detection of Legionella pneumophila capsular-like envelope antigens by counterimmunoelectrophoresis.

The capsular-like envelope of Legionella pneumophila strains Togus 1 (serotype 2) and Philadelphia 1 (serotype 1) was isolated and purified by column chromatography on Sepharose 6B. Antibody raised in rabbits to these two antigenic materials did not cross-react in gel diffusion. Upon electrophoresis followed by gel diffusion, the majority of both envelope materials was found to migrate towards the cathode. A minor antigenic component of each envelope only migrated slightly towards the anode. Using the envelope antigens and the two anti-envelope sera in a counterimmunoelectrophoresis (CIE) assay, positive results were only obtained when the antigenic materials were placed in the cathodal well. The Togus 1 and Philadelphia 1 antigens did not cross-react in CIE. The sensitivity of the CIE assay was poor (15.6 micrograms/ml by carbohydrate content) compared to its sensitivity in other microbial systems. Although CIE may not be a useful diagnostic aid in identifying Legionella species due to its low sensitivity, it may be of value in serotyping the microorganism since we did not see cross-reactivity between the two strains when anti-envelope sera were used.     

Similar DNA restriction endonuclease profiles in strains of Legionella pneumophila from different serogroups

DNA of strains of Legionella pneumophila serogroups 1, 3, 4, and 6, isolated from patients and environmental sources, was examined by restriction endonuclease analysis (REA). Major differences in profiles enabled subtyping in many strains with the same serogroup antigen. However, a cluster of L. pneumophila strains, originating from all the examined serogroups, had similar restriction endonuclease profiles, sometimes with minor differences. This suggests that the genetic similarity between strains of L. pneumophila of different serogroups is sometimes closer than in strains with the same serogroup antigen. Seven environmental sources harbored two L. pneumophila strains with various serogroup antigens; six sources had similar restriction endonuclease profiles. The resolution of small differences in profiles is hampered in REA by the great magnitude of DNA fragments; even upon extensive analysis, these differences are not always readily visualized. Double digestions with the restriction enzymes HpaI and HpaII showed the best results and sometimes revealed differences not evident by digestions with a single endonuclease. REA has a great capacity for accurate epidemiological typing of L. pneumophila, in addition to classical serogrouping; it appeared that the results of the two techniques do not necessarily correlate. On the other hand, it should be stressed that small differences in profiles are not easily detected by REA.     

Immunological and biochemical relationships among flagella isolated from Legionella pneumophila serogroups 1, 2, and 3.

Flagella were isolated from virulent Legionella pneumophila serogroups 1, 2, and 3. Antiserum made against purified serogroups 1 flagellin agglutinated live, flagellated serogroups 1, 2, and 3 but not heat-killed or nonflagellated bacteria. A single line of identity was seen in immunodiffusion slides between the flagella isolated from the three serogroups and antibody to flagellin isolated from serogroups 1, 2, and 3. Indirect immunoperoxidase staining showed that antibody to flagellin isolated from serogroup 1 organisms reacted with flagella on serogroup 1, 2, and 3 bacteria. Indirect immunoperoxidase staining was also showed that antibody to flagellin isolated from serogroup 1 L. pneumophila did not react with the serogroup-specific cell surface antigen, thus demonstrating that the flagella- and the serogroup-specific antigen are separate antigens. The amino acid content of the flagella from the three serogroups was essentially the same, with aspartate, glutamate, alanine, and threonine comprising 41% of the total. Thirty-five percent of the amino acids were hydrophobic, and there were not detectable amounts of cysteine, tryptophan, or tyrosine.     

Differences in lipopolysaccharide profiles of serologically identical Legionella pneumophila serogroup 6 strains.

Over five years 18 strains of Legionella pneumophila serogroup 6 were isolated in Amsterdam from the hot water supply in three hospitals and from one patient. Immunodiffusion and immunoblot procedures showed that these strains were identical. Profiles of isolated lipopolysaccharides from the 18 strains and the reference serogroup 6 strain were visualised in polyacrylamide gels stained with silver. Four strains from hospital A, isolated in 1982, 1984, and 1985 displayed similar lipopolysaccharide profiles which were different in relative mobility from those of hospitals B and C. Those from hospital B (12 strains isolated in 1983 and 1986) and C (one strain) were similar in relative mobility but different in colour. The strain from a patient with acquired immune deficiency syndrome (AIDS) in hospital A displayed a lipopolysaccharide profile characteristic of hospital A. These reproducible profiles were all different in relative mobility from the reference serogroup 6 strain. They can be used as a marker system in epidemiological surveys of serologically identical serogroup 6 strains. Lipopolysaccharide patterns from strains isolated throughout the years in the same hospital were similar. This suggests an outgrowth from organisms inhabiting the plumbing system rather than reseeding from the Amsterdam mains supply.     

Detection of antibodies against Legionella pneumophila serogroups 1 to 6 and the Leiden-1 strain by micro ELISA and immunofluorescence assay

One hundred and seventy-five human sera were tested for the presence of type-specific antibodies against L pneumophila serogroups 1 to 6 and the Leiden-1 strain by means of an enzyme linked immunosorbent assay (ELISA) and compared with the results obtained by the indirect immunofluorescence assay (IFA). A high correlation (correlation coefficient 0.92) between both methods was found. No consistent pattern of IgG, IgA and IgM classes of antibody titres against L pneumophila were found. In the sera of 15 of 17 patients with a proven L pneumophila pneumonia, IgM class antibodies against L pneumophila could be detected. A "polyvalent" ELISA was developed which permits rapid routine screening of human sera for antibodies against L pneumophila serogroups 1 to 6 and the Leiden-1 strain.     

Legionella pneumophila serogroup 14 isolated from patients with fatal pneumonia.

Two Legionella-like organisms, one isolated from postmortem lung tissue and the other from a bronchial aspirate, were shown by growth, physiologic, and genetic characteristics to belong to the species Legionella pneumophila. Subsequent serologic testing indicated that both strains belonged to a new serogroup, serogroup 14.     

Isolation of Legionella pneumophila serogroup 14 from a human source.

A strain of Legionella pneumophila serogroup 14 was isolated during a retrospective study, after death from the sputum of a patient who had had acute leukaemia and pneumonia. This is the third strain of that serogroup to be isolated from a human source. This event emphasises the importance of performing culture as well as serological tests, so as to detect cases of legionellosis caused by strains which rarely cause fatal clinical illness.     

Lymphoid cell blastogenesis as an in vitro indicator of cellular immunity to Legionella pneumophila antigens.

The lymphocyte blastogenic transformation assay was adapted to study responsiveness of lymphoid cells from animals and humans to Legionella pneumophila antigens in vitro. Spleen cells from guinea pigs after active immunization with Legionella vaccine, but not from normal animals, responded by blast cell transformation when stimulated in vitro with killed Legionella whole-cell vaccine, sonic extracts thereof, or a purified somatic antigen. The response was dose dependent. Similar lymphocyte blastogenesis occurred with spleen cells from mice sensitized to Legionella by sublethal infection with the bacteria. In addition, blastogenesis occurred with peripheral blood leukocytes from human volunteers tested in vitro with whole-cell vaccine, sonic extracts, or purified somatic antigen. Maximum responsiveness generally occurred 4 to 5 days after stimulation of human peripheral blood leukocytes, but a day or two earlier with spleen cells from normal or sensitized mice. Guinea pig spleen cells generally showed peak responses at the same time as human peripheral blood leukocytes after stimulation in vitro. Blastogenic responses with purified antigen were comparable to those with the whole-cell vaccine or sonic extract. Such antigens from Legionella provide a useful material for inducing responses in vitro as a correlate of cellular immunity to these bacteria.