Separation of Soluble Brucella Antigens by

Transcription

1 INFECTION AND IMMUNITY, JUIY 1970, p Copyright (D 1970 American Society for Microbiology Vol. 2,. 1 Printed in U.S.A. Separation of Soluble Brucella Antigens by Gel-Filtration Chromatography JERRY R. McGHEE AND BOB A. FREEMAN Department of Microbiology, University of Tennessee Medical Units, Memphis, Tennessee Received for publication 6 February 1970 Soluble precipitating antigens of Brucella suis have been, in various degrees, purified by filtration on Sephadex gels. The most useful gels employed were Sephadex G-150, Sephadex G-200, and Sepharose 4B. Although not all fractions proved to be immunologically pure, some crude molecular-size estimates of most of the 13 soluble antigens of the Brucella cell could be given. In addition, monospecific antisera to three purified Brucella antigens have been prepared. By using purified preparations, physical and chemical data were obtained on two major antigens, E and 1, and a minor antigen, f. Antigen E is not an agglutinogen and may be toxic. Antigen I is of low molecular weight and is neither toxic nor agglutinogenic. The minor antigen f is an agglutinogen as well as a precipitinogen and is found on the cell surface. Both major antigens, when purified, were immunogenic in rabbits. Although soluble Brucella antigens have been studied in detail by gel diffusion (16, 17) and immunoelectrophoresis (2, 5, 9a), relatively little has been accomplished toward isolation and chemical characterization of the antigens. Glenchur et al. (10) separated soluble extracts of B. melitensis into several antigenic fractions by column chromatography on diethylaminoethyl (DEAE)-cellulose. Most of these fractions elicited either agglutinating or precipitating antibodies in rabbits. These investigators concluded that almost all constituents of the bacterial cell can induce agglutinins and precipitins as well as dermal hypersensitivity in rabbits. Ellwood et al. (9) used Sephadex G-200 and DEAE-cellulose to separate and purify an immunogen from the culture filtrate of B. abortus. This purified material had lost most of its capacity to induce protection in mice and guinea pigs and had no specifically precipitable antigens, but it still retained a portion of its original activity in the antibactericidal test. We concerned ourselves with the separation of several individual antigens of Brucella by use of gel filtration and the characterization of these antigens by physical, chemical, and immunological means. MATERIALS AND METHODS Cultures. The cultures used were smooth B. suis 1330 and B. suis 32P. Stock cultures were grown on 2% Tryptose-agar slants, and 18-hr logarithmic cultures were harvested in Tryptose broth and stored at -78 C. Cultures for regular use were maintained on 2% Tryptose-agar slants and stored at 5 C. Preparation of soluble antigens. Brucella for preparation of soluble antigens were grown on 2% tryptoseagar for 48 hr at 37 C. The cells were suspended in saline, centrifuged at 5,500 X g, and washed twice with saline. After the last wash, the sedimented cells were resuspended in distilled water to a concentration of approximately 5 X 1010 cells/ml. Samples of this suspension were placed in the closed gas chamber of a Sonifier (model S75; Branson Instruments, Danbury, Conn.), and full power was applied in 1-min bursts with 2-min cooling intervals. A total of 20 min of sonic treatment resulted in complete disruption of cells as determined by stained smears. The solubilized preparation was centrifuged at 8,500 X g for 15 min, and the supernatant was sterilized by membrane filtration. This soluble material was concentrated to small volume by flash evaporation at low temperature or to complete dryness by lyophilization. In general, sonic extracts required concentration to about onethird the original volume for maximum expression of precipitin lines by immunoelectrophoresis. This usually resulted in a protein concentration of 10 to 15 mg/ml (14). These preparations were stored at 5 C without loss of activity. Animal immunization. Immunizing antigens of B. suis 32P and 1330 were prepared, and rabbits were immunized by methods and schedules previously described (2). Seven days after the last injection, the animals were exsanguinated by cardiac puncture. The sterile, immune serum was stored at -20 C in small portions without preservation. On occasion, soluble antigens were used for immunization. When whole-cell sonic extracts were employed for this purpose, the protein concentration was determined, and the antigen was lyophilized and mixed with complete Freund's adjuvant to give a total of 2.3 mg of protein per ml. Two rabbits were immunized by the above schedule so that each rabbit 48

2 VOL. 2, 1970 SEPARATION OF BRUCELLA ANTIGENS 49 received a total of 10 to 15 mg of protein during the immunization. When column fractions were used, the immunization doses were adjusted to compare to that of whole-cell extract. Gel-filtration chromatography. Soluble Brucella antigens were prepared as described above and separated by column chromatography on Sephadex and Sepharose gels (Pharmacia Fine Chemicals, Piscataway, N. J.). The gels were allowed to swell in either distilled water or 0.1 M sodium phosphate buffer (ph 7.2) as recommended by the manufacturer. Cooled and jacketed columns (Pharmacia Fine Chemicals, Piscataway, N. J.) of either analytical (2.5 cm diameter) or preparative (5.0 cm diameter) size were packed by pouring a thin slurry of gel into the column partially filled with the eluting fluid. After settling for 24 hr, the gel was washed with a minimum of three bed volumes of the eluting solvent. Flow was maintained by gravity feed by use of a constant-pressure flask with a head not exceeding 10 to 15 cm of water. For preparative columns, the bed volume usually ranged from 800 to 1,200 ml, whereas the smaller, analytical columns had volumes of 150 to 200 ml. The homogeneity of the bed and the void volume of the column were determined by application of blue dextran (type 2000; Pharmacia Fine Chemicals, Piscataway, N. J.). Samples were applied to the top of the column and did not exceed 1% of the total volume. Effluent was monitored by an ultraviolet analyzer at 254 nm and collected by fraction collector. Fractions were pooled where required, dialyzed against distilled water at 5 C, and concentrated by low-temperature flash evaporation. Since the rate of elution from Sephadex gels is largely dependent on molecular size, we have recorded molecular sizes of the antigens as equivalent to the molecular weight of a standard globular protein (1). Although the molecular weights determined would then be accurate only for globular proteins, they are indicative of relative sizes and are useful for present purposes. Immunological methods. The procedures employed for immunoelectrophoresis of soluble Brucella antigens were given in detail by Baughn and Freeman (2). After electrophoretic separation and development with immune serum, the slides were washed overnight in saline to remove unprecipitated protein; permanent records of precipitation lines were made by dark-field photography of the wet preparations. Immune sera were titrated for agglutinins by use of formalinized B. suis in a standard tube test with a final density of 3 X 109 cells/ml. Titrations were incubated at 56 C for 4 hr followed by overnight incubation at 4 C; titers are expressed as the highest serum dilution showing complete agglutination. To demonstrate the presence of surface antigens and to prepare monospecific antisera, certain sera were absorbed with formalinized Brucella. The antiserum and cells were mixed at a ratio of 1.0 ml of serum to 5 X 1010 Brucella and incubated at 37 C for 2 hr with intermittent shaking. This was followed by incubation at 4 C for 18 hr, after which cells were removed by centrifugation, and the serum was stored at 4 C. Chemicals tests. Chemical tests were carried out on each of the pooled column-eluate fractions. Deoxyribonucleic acid (DNA) was determined by the microchemical method of Ceriotti (4), ribonucleic acid (RNA) was determined by the orcinol method of Dische (7), and pentose was determined by the Lampen (12) modification of Mejbaum's method. Protein was estimated by the procedure of Lowry et al. (14). Keto sugars, as assayed by the method of Dische and Borenfreund (8), were not found in any of the antigen fractions. Rabbit dermal toxicity tests. Each fraction was tested for rabbit dermal toxicity by intradermal injection of 0.1 ml of the fraction or appropriate dilutions. Reactions were read at 24, 48, and 72 hr after injection, and maximum readings were recorded as the diameter of the reactive site (13), or as the volume of the reactive site measured by the technique of Lurie (15). By use of the latter method, an arbitrary volume of 10 mm3 was chosen as a standard, and toxicity units have been calculated for each fraction on this basis. RESULTS For interpreting the antigenic analyses in this and previous studies, the precipitating antigens are identified by a code in which negatively charged antigens are given number designations; positively charged and nonmigrating antigens are assigned arabic numerals. Further, strongly precipitating antigens are designated "major," and their numbers or letters appear in either italics or upper case (2). Separation of antigens by gel filtration. Figure 1A shows the profile of soluble B. suis antigens eluted from Sephadex G-150. All chemical, immunological, and toxicity tests were performed on the pooled fractions represented by the vertical lines and are listed in Table 1. The first peak contained most of the DNA and pentose; it also had considerable protein and RNA and most of the dermal toxicity. Immunoelectrophoresis of this fraction showed that it contained the major antigen E and minor antigen f of Baughn and Freeman (2). Since antigens E and f appeared in the void volume, the molecular sizes could not be determined with this gel. The middle peak contained several antigens which had molecular sizes in the range of 5 to 10 X 104. These included major antigens A, C, and I and minor antigens 4, b, d, g, and h. The last peak, which appeared near the end of the bed volume, contained only antigens I and 4. Antigens 2, 3, and 5 are lost by filtration on this gel. When Brucella sonic extracts were chromatographed on Sephadex G-200, the elution pattern shown in Fig. lb was obtained. The data from this gel are not markedly different from those obtained with Sephadex G-150. With this more porous gel, the majority of the first peak appears

3 50 McGHEE AND FREEMAN INFEC. IMMUN. E U-) wc- 0z m 0 mcr, c.5 c9t U- LL z m 0 co Cl) In z, cl Z05 0 cr 0.25[ I.00r 075h 0o50F 025F A ELUATE VOLUME (ml) ELUATE VOLUME (ml) ELUATE VOLUME (ml) c antigen E by electrophoresis; this antigen was, therefore, assigned a molecular size of 1.8 X 106 calculated as globular protein. The other peaks contained several antigens each, and the last peak had most of the antigens which could be eluted from this gel. The chemical data on fractions from Sepharose 4B do not differ significantly from those obtained with Sephadex gels. Separation of Brucella antigens was attempted on Sepharose 2B, but all antigens were eluted in 150 a single broad peak, indicating that they were all well below the sizes excluded by this gel. It should be pointed out that no differences could be seen between B. suis 1330 and B. suis 32P on any of the gels employed or when subjected to B immunoelectrophoresis; therefore, B. suis 1330 alone was used in the remainder of these experiments. Immunization with purified antigens. Rabbits were immunized with purified antigens I and E and with the mixed fraction containing E and f. Fractions which contained only a single antigen, either E or 1, induced monospecific antiserum to their respective antigens, as detected by immunoelectrophoresis with whole-cell sonic extracts of B. suis. This indicates, at least in two instances, that solubilized antigens of Brucella can elicit antibody formation in the absence of other bacterial antigens. The column fractions, which contained both antigens E and f, induced antibodies to both antigens. When purified antigen E was mixed with this antiserum, antibody to E was removed, thus creating a monospecific antiserum to antigen f. We have, therefore, prepared three monospecific antisera, to major antigens E and I and to minor antigen f, to be used in further studies. 200 Studies with monospecific antiserum. The monospecific antiserum against antigen I showed FIG. 1. Elution profile of soluble JBrucella suis agglutination only at the lowest dilution (1:10), extracts on gel filtration columns. A, Sepihadex G-150; B, Sephadex G-200; and C, Sepharose 41 antiserum to antigen E showed only a 1:80 titer, which we consider negligible, and antiserum to antigen f showed a titer of 1:1,280 which is in the void volume and contains antigen E. considered quite high, since this is a minor antigen Antigen f is found in the trailing side of this peak, in the Brucella and is present in low quantity in indicating partial penetration of the gel. It is the cell extracts. These data show that antigen f calculated by the method of Andrews (1) to have is probably an agglutinogen. It must be rememis a shift in bered, however, that other antigens, such as an a molecular size of 5 X 105. There protein to the first peak and some toxicity appears agglutinogen, may occur in the same chromat- 5, which ographic fraction as antigen f and not be shown in the second peak (Table 2). Antiglen was not found in the Sephadex G eluates, is by the precipitin reaction. The agglutinins induced found in the second peak, and onl3y antigen I by the antigen E preparations may be due to appears in the third peak. contamination with other unknown antigens. When the very porous gel Sephar ose 4B was In the first peak from Sepharose 4B, which employed (Fig. 1C, Table 3), three ultraviolet contains only antigen E, no antigens were shown absorbing peaks were observed. The fil rst appeared which precipitate in the well of immunoelectro- only phoresis slides or which, when partially just after the void volume and contained hydro-

4 VOL. 2, 1970 SEPARATION OF BRUCELLA ANTIGENS 51 TABLE 1. Chemical, toxic, and immunological properties of soluble Brucella suis fractions eluted from Sephadex G-150 Column fraction Protein Pentose DNA RNA Toxicity Amt Per Amt Per Amt Per Amt Per Units Per (mg) cent (mg) cent (Ag) cent (Mg) cent cent Antigens present Fractioni , E, f Fraction ne A, b, C, d, g, 1, 4 Fraction Trace , 4 Totalsa , a Totals in per cent column indicate per cent of original sample recovered from column. TABLE 2. Chemical, toxic, and immunological properties of soluble Brucella suis fractions eluted from Sephadex G-200 Protein Pentose DNTA RNA Toxicity Column fraction t Antigens present Amt Per Amt Per Amt Per Amt Per Unit Per (mg) cent (Mg) cent (;Mg) cent (JAg) cent cent Fraction , E, f Fraction A,b,C, d, h, 1, 4, 5 Fraction Trace ne 1 Totalsa , a Totals in per cent column indicate per cent of original activity recovered from column. TABLE 3. Column fraction Chemical, toxic, anld immunological properties of soluble Brucella suis fractions eluted from Sepharose 4B Protein Pentose DNA RNA Toxicity Antigens Amt Per Amt Per Amt Per Amt Per I t Per present (me) cent (g) cent (MUg) cent (JAg) cent Units cent Fraction , E Fraction ne f, h Fraction Trace ne A, C, d, I Totalsa , a Totals in per cent column indicate per cent of original sample recovered from column. lyzed, give the precipitin lines reported by Hinsdill and Berman (11) as characteristic of Brucella endotoxin. Further, the endotoxin fraction 5a prepared by their procedure does not not inhibit agglutination of whole Brucella by monospecific antiserum to antigen f. It is believed, therefore, that preparations which contain antigen f do not have the classical endotoxin and that f is, indeed, an agglutinogen. To establish the location of these antigens in the Brucella, the monospecific antisera were absorbed with whole cells. With this technique, it was found that both major antigens E and 1 are subsurface antigens, i.e., their antibodies are not absorbed. It appears that antigen f, however, is primarily a surface antigen since whole cells will remove its antibody from monospecific serum. The technique of absorption was also used to show the location of other precipitating antigens. The results are shown in Table 4 along with a summary of other observations. The first column shows the Brucella antigens, by using the convention of capitalizing or underscoring to designate major antigens (2). The next three columns compare the relative position of these antigens in the peaks from Sephadex and Sepharose gels. The molecular sizes shown next are based on gel exclusion and, except for antigens E and f, are only very crude approximations. The next

5 52 McGHEE AND FREEMAN INFEC. IMMUN. TABLE 4. Characteristics of soluble antigents of Brucella suis separated by gel filtratioli Antigen designation Antigens chromatographed on Sephadex G-150 Sephadex S Sepharose 4B Approximate molecular weight I Found on cell surface Immuno-! genic A,,,,Ilutinin titer A bc d E h a and 1 and I 5X X 104b 5-10 X 104 5X X X 10' 5-10 X X 101 Low 5-10 X X 104 a Ultraviolet-absorbing peak in which antigen is located. b Indicates 5 X 104 to 10 X 104. column describes the location of the antigen in the Brucella, insofar as these methods apply. The apparent location of antigens is based on the absorption, by whole Brucella, of antibody induced to purified antigen. The next column states whether the antigens purified by gel fractionation are still capable of inducing precipitating antibodies. Of all antigens recovered, only 4 and b are no longer antigenic. Presumably, they are altered in some way by the fractionation process, or they are not sufficiently antigenic to induce antibody in the concentration available. Finally, agglutinin titers of monospecific antisera to antigens E, f, and I are listed. DISCUSSION Soluble Brucella antigens can be separated by gel filtration, and in some instances an estimate of their molecular size can be established. Until the individual antigens are separated with some degree of purity, their chemical composition is still uncertain. It is realized that the chromatographic separation presented here is only preliminary to further and more complete antigen separation by such techniques as polyacrylamide gel electrophoresis or ion exchange chromatography. For the most part, the largest as well as the smallest antigen molecules have been purified by this technique. Antigen E could be separated from the minor antigen f only by Sepharose 4B chromatography. An estimate of the molecular size of this antigen was established by its elution pattern from Sepharose 4B. Considerable protein and DNA are also associated with this material. If this antigen contains DNA as a part of its i 1:80 1: :10 molecular structure, it may be related to the DNA-associated antigens of others (3, 18, 19). Although chromatographic data from Sephadex G-150 and G-200 gels were similar, it was always necessary to use the latter gel to separate antigen 1. This antigenhas the lowest molecular size of any of the antigens studied. It is probably a protein and is either contaminated with RNA and DNA fragments or contains these as a part of its molecular structure. It appears to have no net charge at ph 8.6, since it does not migrate upon electrophoresis and it does not possess dermal toxicity. Monospecific antiserum to antigen f could only be prepared by injecting animals with material from the first peak from Sephadex G-150 or G-200. This antiserum, which contained antibodies against both antigen f and E, was then absorbed with antigen E contained in the first peak from Sepharose 4B. Agglutination tests on the monospecific antisera indicated that minor antigen f is an agglutinogen as well as a precipitinogen. This suggests that more than one surface antigen is present in the Brucella and that they are capable of eliciting agglutinins. In fact, Baughn and Freeman (2) reported, and it is confirmed here, that at least three minor antigens are present on the cell surface. On the other hand, Diaz et al. (6) have stated their belief that only one antigen is responsible for agglutination and that this antigen is not one of the migrating antigens such as those described here. The high agglutinin titer of monospecific antiserum to antigen f, along with its absorption by whole Brucella, strongly suggests that it is a

6 VOL. 2, 1970 SEPARATION OF BRUCELLA ANTIGENS 53 surface antigen. This antiserum does not yield a precipitate with endotoxin preparations; therefore, it is not believed that antigen f is a part of the classical endotoxin in these organisms. This is supported by the observation that fractions 5 and 5a prepared by the technique of Hinsdill and Berman (11) will not significantly alter the agglutinin titer of the monospecific antiserum to antigen f by the inhibition reaction. It should be pointed out that, with only a few exceptions, the column fractions were better immunogens than either whole cells or whole-cell extracts, and the antisera produced gave stronger and less diffuse lines in immunoelectrophoresis. Although the reasons for this are uncertain, it may be that poorly immunogenic or interfering material is separated from Brucella antigens by the chromatographic procedures. The successful separation and purification of several soluble Brucella antigens suggest that these methods can be used to prepare more efficient protective antigens for man and animals and can lead to more definitive studies on the biological activity of these materials. ACKNOWLEDGMENTS This research was supported by U.S. Army Medical Research and Development Command contract DA MD-2724 and by Public Health Service fellowship 5-F01-GM-34,140 (to J. R. M.) from the National Institute of General Medical Sciences. LITERATURE CITED 1. Andrews, P Estimation of the molecular weights of proteins by Sephadex gel filtration. Biochem. J. 91: Baughn, R. E., and B. A. Freeman Antigenic structure of Brucella suis spheroplasts. J. Bacteriol. 92: Braun, W., J. W. Burrous, and J. Phillips A phenolextracted bacterial deoxyribonucleic acid. Nature (London) 180: Ceriotti, G A microchemical determination of deoxyribonucleic acid. J. Biol. Chem. 198: Diaz, R., and A. Chordi Immunoelectrophoretic analysis of Brucella suis. Experientia 22: Diaz, R., L. Jones, D. Leong, and J. Wilson Differences between Brucella antigens involved in indirect hemagglutination tests with normal and tanned red blood cells. J. Bacteriol. 94: Dische, Z Spectrophotometric method for the determination of free pentose and pentose in nucleotides. J. Biol. Chem. 181: Dische, Z., and E. Borenfreund A new method for the detection and determination of keto sugars and trioses. J. Biol. Chem. 192: Ellwood, E., J. Keppie, and H. Smith The chemical basis of the virulence of Brucella abortus. VIII. The identity of purified immunogenic material from culture filtrate and from the cell-wall of Brucella abortus grown in vitro. Brit. J. Exp. Pathol. 48: a. Freeman, B. A., J. R. McGhee, and R. E. Baughn Some physical, chemical, and taxonomic features of the soluble antigens of the Brucellae. J. Infec. Dis. 121: Glenchur, H., U. S. Seal, H. H. Zinneman, and W. H. Hall Antigenicity of some Brucella melitensis cell fractions. J. Bacteriol. 85: Hinsdill, R. D., and D. T. Berman Antigens of Brucella abortus. I. Chemical and immunoelectrophoretic characterization. J. Bacteriol. 93: Lampen, J Formation of ribose phosphate from xylose by extracts of Lactobacillus pentosus. J. Biol. Chem. 204: Larson, C. L., E. Ribi, K. C. Milner, and J. E. Lieberman A method for titrating endotoxic activity in the skin of rabbits. J. Exp. Med. 111: Lowry, 0. H., J. N. Rosebrough, A. L. Farr, and R. J. Randall Protein measurements with the Folin phenol reagent. J. Biol. Chem. 193: Lurie, M. B I. Inherited native and natural respiratory contagion. Amer. Rev. Tuberc. (Suppl.) 44: Olitzki, A. L., and D. Sulitzeanu Studies on the antigenic structure of B. suis with the aid of the agar gel precipitation technique. I. The resistance of antigens to physical, chemical, and enzymatic treatments. Brit. J. Exp. Pathol. 39: Olitzki, A. L Studies on the antigenic structure of virulent and nonvirulent Brucellae with the aid of the agar gel precipitation technique. Brit. J. Exp. Pathol. 40: Phillips, J. H., W. Braun, and 0. J. Plescia Immunochemical study of a bacterial DNA. J. Amer. Chem. Soc. 80: Phillips, J. H., W. Braun, and 0. J. Plescia Antigenicity of a bacterial deoxyribonucleic acid. Nature (London) 18: