STAINING BACTERIAL SMEARS WITH FLUORESCENT ANTIBODY

Transcription

1 STAINING BACTERIAL SMEARS WITH FLUORESCENT ANTIBODY III. ANTIGENIC ANALYSIS OF Salmonella typhosa BY MEANS OF FLUORESCENT ANTIBODY AND AGGLUTINATION REACTIONS BERENICE M. THOMASON, WILLIAM B. CHERRY, AND MAX D. MOODY U. S. Department of Health, Education, and Welfare, Public Health Service, Communicable Disease Center, Atlanta, Georgia Received for publication May 13, 1957 Moody et al. (1956) and Thomason et al. (1956) have applied fluorescent antibody techniques to the specific detection of Malleomyces pseudomallei in dried smears. This method already has proved successful for the detection and identification of several other pathogenic bacteria. It is potentially a valuable tool for the identification of pathogenic bacteria in clinical specimens and should be of interest to clinicians and bacteriologists in public health and hospital laboratories. Experience with the use of fluorescent antibody for the detection of bacteria has shown the need for more basic information regarding the nature and specificity of the reaction. The antigenic constitution of cells of a given species of bacteria may vary both quantitatively and qualitatively, depending upon the medium on which the cells have been cultured and upon the physicalchemical manipulations to which they have been subjected, and upon natural variation from unknown causes. In order to analyze the influence of the different classes of antigenic cell constituents upon staining by fluorescent antibody, a model system was sought. Salmonella typhosa seemed to qualify in this respect, since its serological characteristics are well-known and since it contains components representing three common classes of antigens, i. e., the K or envelope antigens (Vi), the flagellar or H antigens (d), and the somatic or 0 antigen (9, 12). The fact that each of the specific antigens, (Vi), (9, 12), and (d) is found in bacteria which have no other serological relationship to S. typhosa makes it possible to prepare individual sera containing antibody for each class of antigen found in the typhoid organism. By this approach it has been possible to show that three classes of antigens present in smears of cells of S. typhosa may be stained specifically with fluorescent antibody and that this specificity is roughly equivalent to that demonstrable by the agglutination test. Strains used. MATERIALS AND METHODS Salmonella typhosa, 0901 (9, 12) S. virginia, (8:d-) Bethesda-Ballerup paracolon (29, Vi) Salmonella typhosa, ND 1616 (9, 12, Vi:d-) Salmonella typhosa, 2V (9, 12, Vi:d-) Salmonella typhosa, Wis 1579 (9, 12, Vi:d-) Salmonella typhosa, Me 1328 (9, 12, Vi:d-) Salmonella typhosa, Me 1614 (9, 12, Vi:d-) Salmonella paratyphi A, 1015, (1, 2, 12:a-) Salmonella paratyphi A var. durazzo, (2, 12: d-) Salmonella paratyphi B var. odense, 8085 (4, 12:b-1, 2) Salmonella derby, (1, 4. 12:fg Salmonella typhimurium, 3842 (4, 5, 12:i-1, 2) Salmonella typhimurium, 0-form (4, 5, 12) Salmonella typhimurium, 3134 (4, 5, 12:i-1, 2) Salmonella typhimurium, 3523 (4, 5, 12:i-1, 2) Salmonella typhimurium, 3941 (4, 5, 12:i-1, 2) Salmonella bredeney, 1654 (4, 27, 12:1v-1, 7) Salmonella bredeney, 1051 (1, 4, 27, 12:1v-1, 7) Salmonella bredeney, 3799 (4, 27, 12:1v-1, 7) Salmonella bredeney, 3156 (4, 27, 12:]v-1, 7) Salmonella bredeney, 3528 (4, 12:1v-1, 7) Media. Edwards and Bruner's (1942) semisolid medium was used to test or enhance motility. Heart infusion broth and agar (Difco) were used for all culture work. Production and titration of antisera. Monospecific antibodies for 0, H, or Vi antigens of S. typhosa were produced in rabbits according to the methods described by Edwards and Ewing (1955). Thus, the 0 antigen consisted of a boiled suspension of the 0901 strain of S. typhosa; the Vi, of an alcohol-treated dried suspension of the Bethesda-Ballerup paracolon; and the H, of a 525

2 526 THOMASON, CHERRY, AND MOODY [VOL. 74 formalinized culture of S. virginia rendered highly motile by passage through semisolid agar. In addition, a single antiserum containing antibodies for 0, Vi, and H antigens was produced by selecting colonies from a recently isolated strain of S. typhosa strain ND 1616 which gave good slide agglutinations with both 0 (Salmonella gallinarum) and Vi (Ballerup) antisera and which was moderately motile in semisolid agar. Dried Vi antigen was prepared as described above. For immunization of rabbits a young (4 hr) formalinized broth culture containing 0, H, and Vi antigen was used. Since the Vi antigen content of cells in a broth culture is usually minimal, the suspension was fortified by the addition of a small amount of dried Vi cells of the same strain. Rabbits were each injected intravenously every fourth day with 0.2, 0.6, 1.2, 2.4, 4.8, and 4.8 ml of this antigen. The rabbits were bled and the serum collected seven days after the last injection. The 0 and H agglutinin titers of all sera were determined by tube agglutination tests using antigens of S. typhosa strain 0901, and S. virginia, respectively. Vi antibody prepared with the Ballerup paracolon was titered with S. typhosa strain 2 V, and that prepared with S. typhosa with the Ballerup paracolon. All of the above antigens consisted of living broth cultures which had been incubated at 37 C for 18 hr. Readings were made after incubation at 45 C for 2 hr TABLE 1 Characteristics of immune sera SeraImmunizing Antigen Agglutinating Agglutinin Sera ~~~~~Antigen Titer* O Salmonella (9, 12) 25,600 typhosa Vi 29, Vi 0 (0901) (9, 12) H (8: d-) 0 Vi Bethesda Ballerup Vi S. typhosa 1,600 paracolon (2 v) (9, 12, (29; Vi) Vi) H (8:d-) 0 H Salmonella virginia Vi 29, Vi 0 (8: d-) H (8:d-) 25,600 O, Vi, S. typhosa (9, 12) 50,000 H (ND 1616) Vi (29; Vi) 400 H (8:d-) 12,800 * Living broth cultures incubated at 37 C for 18 hr were used as antigens. Agglutination tests were incubated at 45 C for 2 hr followed by overnight refrigeration. followed by overnight refrigeration. The data are summarized in table 1. Slide agglutinations were performed with more concentrated cell suspensions using 1:10 dilutions of antisera. Preparation of globulins. Globulin fractions of each of the 4 sera shown in table 1 were obtained by fractionation with one-half saturated ammonium sulfate. The globulins were reconstituted to the original volume of the serum and retitered. There were no appreciable changes in titers. Conjugation of globulins with fluorescein isocyanate. Each of the four globulins were labeled with fluorescein isocyanate by the method of Coons and Kaplan (1950). Staining of smears with fluorescent antibody. Smears from saline suspensions of organisms grown overnight on agar at 37 C were prepared on glass slides. The smears were air dried, heatfixed, and stained with each of the labeled globulins according to the procedure described by Moody et al. (1956). Control smears stained with labeled normal globulin were prepared simultaneously. The results of the examination under the fluorescence microscope were recorded as positive (), negative (-), or doubtful (i). The Leitz Ortholux research microscope with a cardioid darkfield condenser and a Phillips CS-150 mercury lamp with the required filters was used in these studies. EXPERIMENTAL RESULTS Fluorescent staining and slide agglutination reactions obtained with 0, H, and Vi antibody solutions. Smears made from several recently isolated strains of S. typhosa, other Salmonella serotypes related to S. typhosa by the possession of a common antigen (12), the Bethesda-Ballerup paracolon, and S. virginia were stained separately with labeled globulin for each of the three antibody types or with a combination of all three. Slide tests were performed on the same suspensions using similar unlabeled antibody solutions. A comparison of the results obtained by the two types of tests may be made from the data given in table 2. Fifteen of the 20 strains containing a common antigen (12) were agglutinated with 0 (9, 12) antiserum, whereas 17 out of 20 were stained with 0 (9, 12) fluorescent antibody. The 5 strains not agglutinated possessed the Vi antigen which is known to mask the 0 agglutination reaction. However, the same 5 strains were stained with fluorescent 0 antibody. Three of 5 strains of S.

3 19571 STAINING BACTERIAL SMEARS WITH FLUORESCENT ANTIBODY 527 TABLE 2 Comparison of agglutination and fluorescent antibody reactions of Salmonella serotypes Immune Sera (Agglutination) or Labeled Globulin (Fluorescent Antibody)* Strains Tested Antigenic 0, Vi, d Normal Schema of Strains 0 (0901) Vi (B.-Ball.) d (S. virginia) (Nb 1616) Control Aggl. Fluor. Aggl. Fluor. Aggl. Fluor. Aggl. Fluor. Aggl. Fluor. Salmonella typhosa , V.9, 12, Vi; d- - - Me , 12, Vi; d- - - Me , 12, Vi; d- - - Wis , 12, Vi; d- - - ND , 12, Vi; d- - - Beth.-Ballerup paracolon. 29, Vi - - S. virginia.8: d- _ t S. paratyphi A , 2, 12: - - S. paratyphi A durazzo 2, 12: - - S. paratyphi B odense,4, 12: i _ - - S. derby , 4, 12: - - S. typhimurium 0-form 4, 5, 12: , 5, 12: ,5,12: :1 i , 5, 12: 14 i , 5, 12: - - S. bredeney 01.1, 4, 27, 12: 4_ ,2712: ,27,12: 1. i , 27, 12:... _ , 12: :4._4_ - * Sera for agglutination tests were diluted 1:10; labeled globulins for fluorescent staining were dit This reaction denotes cellular fluorescence since S. virginia was the homologous strain. luted 1:4. t The fluorescence observed consisted of pin-point spots over the periphery of the cells. bredeney were agglutinated but not stained. The appearance of a typical strain of S. typhosa stained with labeled 0 (9, 12) globulin is shown in figure 1. All strains containing Vi antigen were both agglutinated and stained by Vi antibody solutions. Only negative reactions were obtained with strains not possessing Vi antigen. The Vi cells of S. typhosa shown in figure 2 were stained with Vi globulin labeled with fluorescein. Results of tests with H antiserum (table 2) showed that agglutination occurred with those strains possessing the H antigen. The fluorescence reaction was negative, since the H antigen necessary for the demonstration of staining by labeled flagellar antiglobulin was lost during the preparation of smears. All cultures, with the exception of certain S. bredeney strains and S. virginia, reacted in both fluorescent antibody and agglutination reactions in which a single globulin solution containing 0, Vi, and H antibody was used. From the irregular results obtained with the strains of S. bredeney, it might be suggested that this serotype is less closely related serologically to the typhoid organism than are the other serotypes tested. The (d) antigen was the only component of S. virginia represented by specific antibody in the combined (0, Vi, H) globulin solution. Since the smear did not receive special handling, the flagella were lost and no staining could be demonstrated. Selective fluorescent staining of cellular and flagellar portions of Salmonella organisms. Since

4 528 THOMASON, CHERRY, AND MOODY [VOL. 74 Figure 1. Salmonella typhosa treated with labeled 0 (9, 12) antiglobulin. In figures 1, 2, 3, and 4 the photomicrographs were made through the Leitz Ortholux microscope using ultraviolet illumination and a dark-field condenser. Figure 2. Salmonella typhosa treated with labeled Vi (Bethesda-Ballerup paracolon) antiglobulin. See legend under figure 1. flagellar staining could not be demonstrated in the above experiment, all strains of S. typhosa and of S. virginia were passaged serially through semisolid agar to enhance motility. Actively motile, 4 to 6 hr broth cultures were prepared, formalin added to a final concentration of 0.3 per cent, the suspension centrifuged, and the organisms washed twice in distilled water. Since the flagella were easily dislodged from the cells, very careful handling of the cultures was required. Four sets of smears were made on acid-cleaned slides, air dried, and fixed in acetone. The first slide of each set was stained with 0, the second with Vi, the third with H, and the fourth with the combined (0, Vi, H) antibody. The distribution of fluorescent antibody on the cell body and flagella was noted under the fluorescence microscope. The results are shown in table 3. The cell body, but not the flagella, of all strains of S. typhosa was stained by both the 0 and Vi globulins but not by the H globulin. Cells of S. virginia (8:d-) did not fluoresce when treated with 0 and Vi globulin of S. typhosa but both the cells and flagella were stained by homologous globulin (8:d-). The flagella of all strains of S. typhosa stained brilliantly with the H antibody (8:d-); the cell body was unstained. This is illustrated in figure 3. Both flagella and cells fluoresced when treated with the labeled combined (0, Vi, H) globulin. Only the flagella of S. virginia were stained by the (0, Vi, H) globulin prepared with S. typhosa. The organisms in figure 4 were stained by combined (0, Vi, H) globulin (S. typhosa) resulting in fluorescence of both cells and flagella. After organisms containing Vi had been passed several times through semisolid medium the quantity of Vi antigen had decreased considerably as revealed by the weak fluorescence seen following treatment with Vi globulin. This change was reflected in agglutination tests in which suspensions previously inagglutinable in 0 serum became agglutinable. Inhibition tests. Further evidence of the specificity of staining with fluorescent antibody was obtained by the application of the one-phase inhibition test (Moody et al., 1956). This consisted of treating dried smears of cells with a mixture of appropriate dilutions of labeled and unlabeled antibody. In such a reaction the presence of unlabeled homologous antibody results in inhibition of fluorescent staining. The data given in table 4 summarize the results obtained. It should be noted that staining of each of the components of cells containing 0, Vi, and H antigen could be inhibited by the use of unlabeled homologous antibody without interfering with the staining of the remaining components. For example, when unlabeled 0 antibody was mixed with labeled combined (0, Vi, H) antibody, staining of only the 0 antigen was inhibited while Vi and H antigens reacted with their respective labeled antibodies as revealed by their fluorescence. Blockage of all cellular attachment sites occurred when the preparations were treated

5 19571 STAINING BACTERIAL SMEARS WITH FLUORESCENT ANTIBODY 529 TABLE 3 Results of staining highly motile Salmonella strains with fluorescent antibody Fluorescence Reactions of Cellular Components Stained with Fluorescent Antibody Organism 0 (0901)9,12 Vi ( B.-B.) H (8:d-) 0, Vi, H (9, 12, Vi:d-) Cells Flagella Cells Flagella Cells Flagella Cells Flagella Salmonella typhosa* 2V, (9, 12, Vi:d-)... Me 1328 (9, 12, Vi: d-)... Me 1614 (9, 12, Vi.d-)... Wis 1579 (9, 12, Vi:d-). - i - ND 1616 (9, 12, Vi: d-). Salmonella virginia (8:d-)... _ * Serial passage through semi-solid medium decreased Vi staining and agglutinability. Figure S. Salmonella typhosa treated with labeled H (8, d) antiglobulins. See legend under figure 1. Figure 4. Salmonella typhosa treated with labeled combined (9, 12, Vi, d) antiglobtulin. See legend under figure 1. TABLE 4 Inhibition studies with fluorescent antibody Antigens AntibodyApplied Applied Reactions Fluorescence with: Unable iunlabeled Labeled test Cells Fla- Conanhbtibod antibody Clsgella trols O, H, Vi O, H,Vi Vi Vi _ - O, H, Vi H H _ - O, H, Vi 0 O, H, Vi O, H,Vi Vi O, H, Vi O, H, Vi H O, H, Vi - O, H, Vi O, H, Vi O, H, Vi - - with a mixture of unlabeled and labeled antibodies directed against 0, Vi, and H antigens (table 4). Adsorption tests. Labeled 0, Vi, H, and combined (0, Vi, H) globulins were adsorbed with 0, Vi, or H antigens or with a combination of 0 and Vi antigen. 0 and Vi antibody was adsorbed with portions of dried antigens used for the production of the corresponding antisera. The H globulin was adsorbed with living cells of S. virginia harvested from plate cultures. The cells were added to the globulins and the mixture incubated in a water bath at 45 C for 2 hr followed by overnight refrigeration. After removal of the cells, the adsorbed globulins were titered by tube agglutinations and smears of organisms were prepared and stained with adsorbed globulins for the determination of residual antibody. The results given in table 5 show that, in each case, adsorption with heterologous components had no effect on the staining of the homologous

6 530 THOMASON, CHERRY, AND MOODY [vol. 74 Figure 6. Salmonella typhosa treated with labeled H (8, d) antiglobulin. Composite darkfield photograph. Left: Cells photographed with ultraviolet light revealing flagella stained with fluorescent antibody. Unstained cell bodies not visible. Right: Same field viewed by visible light showing blurred outline of cells and barely visible flagella. TABLE 5 Effect of the adsorption of agglutinins from fluorescein labeled globulins on staining reactions Labeled Globulins O (1:12,800) Vi (1:400) H (1:12,800) 0, Vi, H (0-) (Vi-1:400) (H-) O (0901) Vi (Beth.-Ballerup) H (S. virginia) 0 vi H 0 vi H 0 vi HO and Vi antigen. These results paralleled the agglutination tests, since, for the most part, homologous titers were reduced to a small fraction of their original figure while heterologous titers remained unchanged. Since residual antibody was present, the globulins were adsorbed completely by treating them with appropriate living cells and the tests were repeated. The staining results were unchanged. Agglutinin Titers after Adsorption Fluorescence of Salmonella typhosa after Exposure to Adsorbed Sera 0 Vi H Cells Flagella 1:160 1:12,800 1:12,800 1:160 1:80 1:400 1:80 1:400 1:400 1:80 1:400 1:40 1:12,800 1:12,800 1:8 1:20 DISCUSSION It has been demonstrated that the specificity of staining of S. typhosa and related bacteria by fluorescent antibody is roughly equivalent to that shown by agglutination tests. The slight deviations which occurred may be explained in terms of differences in the conditions under which the two tests are performed. For example, agglutination depends upon a minimal concentration

7 1957] STAINING BACTERIAL SMEARS WITH FLUORESCENT ANTIBODY 531 of cells for the formation of lattices which become detectable when they reach macroscopic size. It has been shown (Thomason et al., 1956) that positive staining of single organisms may be obtained by fluorescent antibody techniques in smears made with suspensions containing only a few dozen or, at most, a few hundred specific cells per ml. Thus when relatively few serologically specific cells are present, agglutination may not be observed. Inherent differences in the sensitivity of the two tests also may be a factor. The staining of flagella with fluorescent antibody presents the same problems as does staining with aniline dyes and mordants. Cleanliness of equipment and careful handling of flagellated bacteria is necessary if the flagella are to remain attached to the cells. The flagella must be numerous and well developed. For this reason flagella agglutination tests often succeed in the absence of specific staining by fluorescent antibody. Certain differences in staining characteristics were noted in antigenically distinct cells and in those treated in various ways. For example, cells stained with labeled 0 antibody are less brilliant but more evenly stained than those treated with Vi antibody (figures 1 and 2). The latter often appear spotted, particularly in older cultures, suggesting either that the antigen has not been produced evenly over the cell surface or that it has been lost by mechanical or physical-chemical means. The intensity of staining of somatic antigen (0) was not reduced by heating the suspension for one hour at 100 C prior to staining. In contrast, heating of the organisms containing Vi in aqueous suspension decreased significantly the degree of staining with Vi globulin. There was a striking difference in the appearance of cells of S. bredeney stained with labeled (9, 12) globulin and those of other Salmonella serotypes. The tiny spots of fluorescence around the periphery of the cell suggested localization of somatic (12) antigen, since this is the only expressed component common to S. typhosa and S. bredeney. These observations suggest the importance of fluorescent antibody techniques for the visual localization of surface antigens of bacterial cells. Repentigny and Frappier (1956) have recently reported on the study of localization of surface antigens of Hemophilus pertussis by means of fluorescent antibody. The specificity of staining shown by the labeled globulins in table 2 was amply confirmed by inhibition and adsorption. These tests substantiated the differences in somatic, envelope, and flagellar antigens and show that flagella are not extrusions of slime normally found on the cell surface. In no instance could staining with flagellar antibody be demonstrated without the presence of typical flagella. ACKNOWLEDGMENTS We wish to acknowledge the assistance given us in these studies by Dr. P. R. Edwards, Chief of the Enteric Bacteriology Unit of the Communicable Disease Center. He provided cultures, reference diagnostic sera, and invaluable suggestions. We are indebted to Dr. Morris Goldman of the Parasitology-Mycology Section of CDC for labeling the globulins. We wish to thank the personnel of the Audio-Visual Section of CDC for photographic assistance. SUMMARY The serological specificity of globulins labeled with fluorescein isocyanate has been studied with Salmonella typhosa. The three classes of antigenic components found in cells of this species could be stained specifically both collectively and individually with appropriate labeled sera applied to bacterial smears fixed on glass slides. The specificity of staining established by inhibition and adsorption tests with fluorescent antibody agreed closely with that of agglutination tests for all serotypes. No evidence was obtained to suggest that the application of a polyvalent globulin solution containing one or more types of heterologous antibody interfered in any way with specific staining of homologous antigen. REFERENCES COONS, A. H. AND KAPLAN, M. H Localization of antigen in tissue cells. II. Improvements in a method for the detection of antigen by means of fluorescent antibody. J. Exptl. Med., 91, EDWARDS, P. R. AND BRUNER, D. W Serological identification of Salmonella cultures. Circuilar 54, University of Kentucky Agr. Expt. Station, Lexington, Ky. EDWARDS, P. R. AND EWING, W. H Identification of Enterobacteriaceae. Burgess Publ. Co., Minneapolis, Minn.

8 532 THOMASON, CHERRY, AND MOODY [VOL. 74 MOODY, M. D., GOLDMAN, M., AND THOMASON, BERENICE M Staining bacterial smears with fluorescent antibody. I. General methods for Malleomyces pseudomallei. J. Bacteriol., 72, DE REPENTIGNY, J. AND FRAPPIER, A Studies on Hemophilus pertussis liquid cultures. III. Localization of surface antigens by means of fluorescent antibody. Can. J. Microbiol., 2, THOMASON, BERENICE M., MOODY, M. D., AND GOLDMAN, M Staining bacterial smears with fluorescent antibody. II. Rapid detection of varying numbers of Malleomyces pseudomallei in contaminated materials and infected animals. J. Bacteriol., 72,