Search

Isolates of Legionella pneumophila that are serologically different from strains of serogroups 1 through 5 were obtained from lung biopsy tissue or pleural fluid from three renal transplant recipients in Chicago, Ill. These strains were placed in a newly designated L. pneumophila serogroup, serogroup 6, on the basis of fluorescent-antibody staining characteristics. An L. pneumophila strain obtained from Bethesda, Md., one from Houston, Tex., and one from Oxford, England, also belong to this new serogroup. L. pneumophila serogroup 6 appears to be widely distributed geographically.

It is technically impractical for many clinical laboratories to use the direct immunofluorescence assay for identifying and serogrouping clinical isolates of Legionella. We compared the results obtained with the direct immunofluorescence assay with the results of a simple and less-demanding slide agglutination test for identifying 15 serogroups representing seven Legionella species. The slide agglutination test was in complete agreement with the direct immunofluorescence assay, and the serogroup to which 64 clinical isolates of Legionella belonged was correctly identified. With polyvalent, pooled antisera and absorbed, serogroup-specific antisera, the slide agglutination test is a useful alternative to the direct immunofluorescence assay in the diagnosis of Legionella infections and for studying the serological relationships of Legionella-like organisms.

Major protein-containing antigens of Legionella pneumophila serogroup 1 were were identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblot analysis with rabbit antisera to 14 different Legionella species or serogroups. Fourteen bands were observed in immunoelectropherograms of whole-cell, sonicated cell, and heated cell preparations, seven of which appeared in the supernatant fluid from the heated cells and three of which were shown in an outer membrane fraction. Immunoblots of whole-cell antigen preparations of 14 Legionella species or serogroups revealed seven major Legionella proteins: antigens with molecular weights of 58,000, 79,000, and 154,000 were present in all Legionella sp. strains, antigens with molecular weights of 44,000 and 97,000 occurred in multiple species, and antigens with molecular weights of 14,000 and 25,000 were present only in L. pneumophila strains. All sera from 15 patients with culture-confirmed L. pneumophila serogroup 1 disease and 14 of 18 (78%) sera from serologically diagnosed patients reacted with the 58-kilodalton (kDa) common antigen. In contrast, less than one-half of the sera reacted with the L. pneumophila-specific proteins (14 and 25 kDa). Absorption of sera with Escherichia coli cells had no effect on their reactivity with the 58-kDa antigen, whereas absorption with L. pneumophila serogroup 1 cells removed reactivity. These data suggest that the 58-kDa antigen may prove useful in serodiagnostic tests for legionellosis.

In a previous study, all convalescent-phase sera from patients with culture-confirmed legionellosis reacted on immunoblots with a Legionella genus-wide 58-kilodalton (kDa) protein antigen (J.S. Sampson, B.B. Plikaytis, and H.W. Wilkinson, J. Clin. Microbiol. 23:92-99, 1986). The present study was done to immunologically characterize and determine the diagnostic relevance of this purified antigen. The antigen was precipitated from enriched cell extracts with ammonium sulfate and purified by high-pressure liquid chromatography. High-titered rabbit antiserum produced to the purified protein was used to show its presence on immunoblots in the 60-kDa range in 38 Legionella serogroups, representing 23 species, and in 39 non-Legionella bacteria. The antiserum was made specific for Legionella strains by sequential absorptions with Bordetella pertussis, Pseudomonas aeruginosa, and Pseudomonas fluorescens whole cells. Serum from legionellosis patients reacted with both specific and nonspecific epitopes. Results of indirect immunofluorescence experiments showed that neither specific nor nonspecific epitopes of the 60-kDa protein were surface exposed on Legionella cells and that cross-reactive epitopes were variably exposed on non-Legionella bacteria. The 60-kDa protein antigen should be useful in diagnostic tests for legionellosis if care is taken to expose cryptic epitopes and if the tests use or measure only the Legionella-specific epitopes.

A semiautomated, kinetic-dependent, enzyme-linked immunosorbent assay (K-ELISA) was adapted to detect serum antibodies to Legionella pneumophila. In a comparative study, 158 human serum samples (79 pairs) were tested by K-ELISA and the standard indirect immunofluorescence assay for determination of antibody levels to L. pneumophila serogroup 1. K-ELISA determinations were made by using a serogroup-specific antigen or a preparation (unfractionated antigen) which contained both common antigen and serogroup-specific reactivity. There was good correlation between the immunofluorescence assay and the K-ELISA by using either antigen, although greater correlation was achieved with the unfractionated antigen (coefficients of correlation, 0.894 with unfractionated antigen and 0.841 with serogroup-specific antigen). These results indicate that the K-ELISA is a reliable alternative to the immunofluorescence assay for serologically diagnosing legionellosis.

Confirmation of a culture as Legionella when it is unreactive with available serologic reagents involves tests that are impractical in most clinical laboratories. A nucleic acid probe that hybridizes only to members of the genus Legionella was recently prepared for marketing by Gen-Probe, Inc., San Diego, Calif. We tested 215 Legionella strains, representing 22 species, and 84 non-Legionella strains, representing 17 bacterial genera, with the Gen-Probe kit. All but four Legionella strains (L. bozemanii, less than 2% of total) and no heterologous strains gave positive test results. We conclude that the Legionella gene probe is a valuable addition to existing diagnostic tests for Legionella organisms.

The modification of the CAMP test for goup B streptococci involved substituting a paper disk impregnated with partially purified beta-hemolysin for the staphylococcal culture that was the source of beta-hemolysin in the original test. The disk is placed onto a sheep blood agar plate beside the streak of Streptococcus being tested. The plate is then incubated aerobically at 35 degrees C. A positive reaction consists of a lunar-shaped clear zone that appears within 24 h in the dark beta-hemolysin zone surrounding the disk. A double-blind study of 135 randomly coded streptococcal isolates showed complete agreement between the CAMP-disk test and the standard Lancefield precipitin test. All group B streptococci tested had positive reactions, and all strains tested from streptococcal groups A, C, D, and G were negative. The CAMP-disk test is a simple and convenient way to identify presumptively group B streptococci.

We used the slide agglutination test to determine the serologic relationships of 22 Legionella spp. representing 33 serogroups. Antisera prepared against 14 of the Legionella spp. contained cross-reactive antibodies (1+ or greater) at their working dilutions. Numerous cross-reactions were observed for the blue-white fluorescing Legionella spp. With only three exceptions in the latter group, cross-reactive antibodies were removed by absorption, thereby producing serogroup-specific antisera. For screening tests or for identification only to the genus level, nine polyvalent antiserum pools were prepared. Routine use of slide agglutination test reagents should expand the number of Legionella spp. that can be identified in the clinical laboratory and, at the same time, provide a simpler, less costly test procedure.

Sera from six outbreaks of legionellosis and four outbreaks of pneumonia of other etiologies were tested with the indirect immunofluorescence assay (IFA) as currently performed. The current IFA is at least as sensitive as the original test in detecting cases of Legionnaires disease (78 to 91%). By using Center for Disease Control criteria for a positive (fourfold increase in titer during convalescence to greater than or equal to 128) or presumptive (single titer greater than or equal to 256) serological test, the specificity exceeded 99%. No cross-reactions against Legionella pneumophila antigens were observed among sera from epidemic cases of Q fever, tularemia, and psittacosis; the only positive L. pneumophila IFA titer among the epidemic Mycoplasma pneumonia sera was reduced to a negative titer with an immunosorbent extracted from Escherichia coli strain O13:K92:H4. The slight increase in specificity (to 100%), however, was offset by a slight decrease in sensitivity. The sensitivity of the IFA was maximal when a conjugate that detected immunoglobulins G, M, and A was used. IFA titers were not significantly altered by replacing the monovalent serogroup 1 antigen with a polyvalent antigen (serogroups 1 through 4) nor by the presence of rheumatoid factor or heat-labile serum factors.

A Legionella-like organism (strain 1087-AZ-H) was isolated from a pleural-fluid specimen from a renal transplant patient undergoing immunosuppressive therapy. Growth characteristics and gas-liquid chromatography profiles of the isolate were consistent with those for Legionella spp. The isolate fluoresced blue-white under long-wave UV light. Strain 1087-AZ-H was serologically distinct in the slide agglutination test with absorbed antisera. DNA hybridization studies placed it in a new Legionella species, Legionella tucsonensis (ATCC 49180).

Evidence obtained by others who used direct immunofluorescence staining to demonstrate serological differences among strains of Legionnaires disease bacterium prompted this study of parameters influencing the ability of the indirect immunofluorescence test to detect human antibodies to Legionnaires disease bacterium. A total of 25 Legionnaires disease bacterium strains, representing four serogroups, were used as immunofluorescence antigens to test selected human sera. The use of diethyl ether in preparing the antigens was discontinued when it was found that titers against ether-killed group 2 (Togus 1-like) antigens were impossible to determine. Instead, heat-killed suspensions of Legionnaires disease bacterium in 0.5% buffered normal chicken yolk sac were used to show the serogroup diversity of the strains and the serogroup specificity of the antibody response of some, but not all, patients with serological evidence of Legionnaires disease. These studies suggest that multiple antigens should be used in serological tests for Legionnaires disease. Furthermore, the fact that some sera contain antibodies that bind equally well to strains of all four serogroups implies that demonstration of a fourfold increase in titer of paired sera when tested with a single antigen should not be interpreted as evidence of infection with a strain of the same serogroup.

It is important to resolve existing differences of opinion regarding group B streptococcal type distribution in human disease because of the relevance of type prevalence to future programs of prevention. This report compares data obtained from typing 392 group B streptococci isolated from systemic infections in both infants and adults in the United States from 1972 through 1975. The data showed a substantial predominance of type III among strains isolated from cases of infant meningitis and from "late-onset" septicemia but did not confirm a prior report that type Ia causes most cases of "early-onset" infant septicemia. Type II was the predominant serotype among 11 cerebrospinal fluid isolates from adults. The fact that over one-fourth of the isolates were types other than Ia or III means that future epidemiological studies, including definition of immunological factors, must include all five group B types.

Legionella jordanis sp. nov., as found in two cultures, is described. One isolate was from river water in Indiana and the other isolate was from sewage in DeKalb County, Ga. The former is the type strain of the species, and is designated BL-540 (ATCC 33623). L. jordanis had a partial relationship to L. bozemanii by direct fluorescent-antibody tests but was unrelated to L. pneumophila, L. dumoffii, L. micdadei, L. gormanii, or L. longbeachae. Legionella phenotypic characteristics, including large amounts of branched-chain cellular fatty acids, were shown by the isolates. Studies of DNA relatedness showed that the two cultures of L. jordanis were only slightly related to the six previously described species of Legionella but were more than 90% related to each other. Indirect fluorescent-antibody tests with human sera suggested that unrecognized human infections with L. jordanis may be occurring.

A semiautomated solid-phase immunofluorescence technique (FIAX) was compared with the standard indirect immunofluorescence assay (IFA) for the determination of antibody levels to Legionella pneumophila serogroup 1 in paired human serum samples. The FIAX method was in agreement with the IFA test for 91.8% of the serum pairs but gave evidence of a recent Legionella infection significantly fewer times than did the IFA. These results suggest that the FIAX technique may eventually be a useful alternative test for measuring Legionella antibodies. However, further study will be required to determine its efficacy in providing a serodiagnosis of legionellosis.

While preparing slide agglutination test antisera and immunofluorescence conjugates for the identification of Legionella species and serogroups, we found that several of the reagents cross-reacted with Bordetella pertussis strains. To determine the extent of this problem and to estimate the specificity of Legionella reagents, we tested slide agglutination test antisera against 22 species and 35 serogroups with 92 bacterial strains representing 19 genera. The only cross-reactions observed were with Legionella pneumophila serogroup 10, L. maceachernii, L. gormanii, and L. feeleii serogroup 1 antisera and 4 of 10 B. pertussis strains. Nineteen conjugates, previously available from the Centers for Disease Control but no longer distributed as reference reagents, were tested with the four cross-reactive B. pertussis strains. Two conjugates, L. micdadei and L. wadsworthii, stained three of the B. pertussis strains at a fluorescence intensity of greater than or equal to 3+. All cross-reactions were removed from the antisera and conjugates by absorption with the cross-reacting strain without diminishing the homologous reaction. Special emphasis should be placed on the identification and removal of cross-reactions in Legionella reagents with strains that have similar morphologic and growth characteristics.

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.

Legionella maceachernii, previously isolated only from the environment, was shown to be a cause of fatal pneumonia in an immunocompromised patient.

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.

Current interest in determining the possible protective role of antibodies against group B streptococcal disease prompted this study of the feasibility of using a radioimmunoassay to measure type-specific immunity in humans. The radioimmunoassay was standardized as a quantitative test for antibodies against the carbohydrate (CHO) antigens of all five group B types. The data showed that the CHO antigens extracted by a cold trichloroacetic acid-sonification method measure more antibodies than do the corresponding CHO antigens extracted by hot hydrochloric acid; that the Ia CHOs extracted from two different types, Ia and Ic, measure the same quantity of Ia antibodies; and that human sera contain antibodies reactive with all five type-specific CHOs. No evidence of "protective" antibody was found in the serum samples studied, although there was evidence of and antibody response in adults to prolonged colonization by group B streptococci. The wide ranges of antibody concentration in a serum bank collection, the broad reactivity of all human sera tested, and the mixed populations of antibodies in human sera that react with different determinants on the same type-specific CHO antigen (type III) indicate that further studies must be done to better define normal and susceptible populations and to determine antigenic components important in protection.

A simple, inexpensive slide agglutination test correctly identified the serogroup to which 38 Legionella pneumophilia strains belonged, and distinguished them from two atypical, Legionella-like organisms.

Many clinical laboratories are currently using counterimmunoelectrophoresis (CIE) as an aid in the rapid diagnosis of bacterial meningitis. Because cross-reactions among causative agents have been reported, the present study was undertaken to explore the problems that might occur when reference and commercial antisera are used in CIE. Broth cultures of 35 bacterial strains were tested with 76 reference and commercial antisera by CIE. Some of the antisera tested failed to react with their homologous strains. Furthermore, several cross-reactions between genera, as well as within species, were noted. These findings suggest that precautions must be taken to insure that all materials used in CIE tests are of high quality. If properly performed and interpreted, CIE may be a valuable adjunct in the identification of organisms causing bacterial meningitis, but it is, nevertheless, a presumptive test and should not be used to replace the Gram stain and culture techniques.

Three strains of Legionella feeleii from patients with pneumonia (425-MI-H, 691-WI-H, and 693-WI-H) and one environmental strain (713-MI-E) received at the Centers for Disease Control for reference diagnostic testing were compared with the type strain WO-44C-C3 (ATCC 35072) by DNA hybridization, chemical analysis of cellular fatty acids and ubiquinones, biochemical tests, and serological characteristics. All four isolates were assigned to the L. feeleii species on the basis of DNA hybridization results. However, strains 691-WI-H and 693-WI-H were serologically distinct from strain WO-44C-C3, as shown by their minimal reactivity (1 to 2+) with a direct immunofluorescence conjugate prepared against L. feeleii serogroup 1 (strain WO-44C-C3). Therefore, strains 691-WI-H and 693-WI-H were placed in a new L. feeleii serogroup (serogroup 2). The reference strain of L. feeleii serogroup 2 is 691-WI-H (ATCC 35849).

A Legionella-like organism (strain 797-PA-H; ATCC 43130) was isolated from a specimen taken from an endotracheal tube of a patient 4 days before death and from the left lung at autopsy. Growth characteristics were consistent with those for Legionella species. Strain 797-PA-H gave negative test results with available direct immunofluorescence assay conjugates and with slide agglutination test antisera prepared against the 22 Legionella species and 35 serogroups now recognized. Minimal reactivity (1 to 2 +) was observed with both tests by using reagents prepared against the Legionella-like organism Lansing 3. Reciprocal absorption studies, however, showed that the cross-reactive antibodies could be removed easily. Physiologic, gas-liquid chromatographic, and DNA hybridization tests revealed that the strain belonged to the species Legionella pneumophila. Therefore, strain 797-PA-H was designated as the type strain of a new L. pneumophila serogroup, serogroup 11.

Legionella gormanii, previously isolated only from the environment, was grown from the bronchial brush specimen of a patient with pneumonia. The organism was characterized by serologic, biochemical, and DNA hybridization studies.

A Legionella-like organism (strain 82A3105; ATCC 43736) was isolated from a lung aspirate taken from a patient with pneumonia. Results of physiologic, gas-liquid chromatographic, genetic, and serologic tests showed that strain 82A3105 and four additional clinical isolates belong to a new Legionella pneumophila serogroup, serogroup 13.